Jo U R N a L O F P Et R O L E U M T E C H N O L O

9/13/16 3:10 PM Oct16_JPT_Cover.indd 1

OCTOBER 2016 • VOLUME 68, NUMBER 10 JOURNAL OF PETROLEUM TECHNOLOGY WellWatcher Flux MULTIZONAL RESERVOIR MONITORING SYSTEM

Event detected

A single wellhead penetration along with inductive couplers enable multidropping of PT gauges into the lower completion, and arrays of digital temperature sensors provide the highest-resolution measurements in the industry.

Maximize reservoir drainage while reducing uncertainty and minimizing operating costs. Early detection is critical to understanding complex reservoir behaviors and reducing risk and remedial costs. By acquiring continuous, real-time data across the entire sandface, the WellWatcher Flux* multizonal reservoir monitoring system enables operators to recognize events immediately and take action before problems occur. In addition, the system tracks fl ow profi les, identifi es underperforming zones, and pinpoints hotspots—all without impeding fl ow. Data is analyzed at surface to enable operators to manage well strategies, fi ne-tune performance, and optimize production.

Find out more at slb.com/Flux

14 GUEST EDITORIAL • RISK MANAGEMENT AT NASA AND ITS APPLICABILITY TO THE OIL AND GAS INDUSTRY At NASA, the qualitative approach to risk assessment is augmented by a quantitative technique to mitigate low-probability, high- consequence outcomes. 24 SHIFT TOWARD PERMIAN DRIVES RISING RIG COUNT The US rig count has inched up recently, and the driver has been the old reliable of onshore oil production, the Permian Basin. With surging production, the Permian is the only major US oil shale play to grow since crude prices began to fall. 28 OILFIELD DATA ANALYTICS ARRIVE The use of intelligent software is on the rise in the industry and it is changing how engineers approach problems. A series of articles explores the potential benefits and limitations of this emerging area of data science. DEVON ENERGY RISES TO THE TOP AS A DATA-DRIVEN The source code of a software PRODUCER application displayed on a The North American shale producer is turning in best-in-class results thanks monitor. Such intricate coding to being an early adopter of advanced analytics. work is delivering a new breed of advanced algorithms that oil and FOUR ANSWERS TO THE QUESTION: WHAT CAN I LEARN gas companies are using to predict FROM ANALYTICS? the future and analyze the past. Recent technical papers consider whether it is better to drill a lateral well Source: Getty Images. up-slope or down-slope; what makes a better fracture; and how old, slow- producing shale wells can temper declines in large portfolios of wells. ANALYTICS FIRMS EXPLORE OIL AND GAS MARKET A host of new software developers have set their sights on solving the industry’s big data issues. DEPARTMENTS ACCELERATING THE UPTAKE CYCLE THROUGH COLLABORATION, OUTSOURCING 6 Performance Indices A new technology consultancy is playing matchmaker between operators and entrepreneurs in an effort to speed up the industry’s adoption rate of 8 Regional Update commercial-ready technologies. 10 President’s Column 42 MANAGEMENT • THE REASON TO EXPECT 12 Comments PROLONGED USD 30–60/BBL OIL 16 Technology Applications A fairly stable set of conditions coalesce to make a strong reason to 20 Technology Update expect the oil price to range between USD 30/bbl to USD 60/bbl 81 People for the foreseeable future. 82 Professional Services 46 FOUR HONORED AS LEGENDS OF ARTIFICIAL LIFT 83 Advertisers’ Index The 2016 SPE Legends of Artificial Lift are announced. Recipients 84 SPE Events will be honored at the 2016 SPE Artificial Lift Conference and Exhibition for North America to be held this month.

An Official Publication of the Society of Petroleum Engineers. Printed in US. Copyright 2016, Society of Petroleum Engineers. WE DELIVER WHAT OTHERS CAN’T. C&J CONTINUES TO PRODUCE new technologies to improve efciencies, increase production and reduce costs.

Re-ConSM Frac-less Re-Stimulation for EFFICIENT, INCREASED PRODUCTION

SM Lateral Science Process PRECISELY ENGINEERED COMPLETIONS

USBS Directional Motor THE INDUSTRY’S SHORTEST bit-to-bend drilling motor

To learn more, contact your C&J representative. Or contact us at [email protected]. EXCELLENCE DELIVERED. CJENERGY.COM archerwell.com/point

TECHNOLOGY FOCUS

48 FIELD DEVELOPMENT PROJECTS Barrier Maria A. Capello, SPE, Executive Adviser, Kuwait Oil Company 49 Sapinhoá Field, Santos Basin Presalt: From Design to Execution and Results failure is 52 Improving Megaproject Economics Through the Use of Early Production 54 Completion and Well-Spacing Optimization for Horizontal Wells certain. in Pad Development

57 NEW-FRONTIER RESERVOIRS II Leonard Kalfayan, SPE, Global Production Engineering Adviser, Hess Corporation 58 The Role of Induced Unpropped Fractures in Unconventional Oil and Gas Wells 60 Injection in Shale: Experience on the Norwegian Continental Shelf Well integrity 62 Model for a Shale-Gas Formation With Salt-Sealed Natural Fractures is at a tipping 66 PETROLEUM DATA ANALYTICS Luigi Saputelli, SPE, Senior Reservoir Engineering Adviser, ADNOC, and Frontender Corporation point. 67 Functional Approach to Data Mining, Forecasting, and Uncertainty Quantification Most integrity surveys only 69 Mitigating Drilling Dysfunction With a Drilling Advisory System evaluate barrier condition. 71 Big-Data Analytics for Predictive Maintenance Modeling: Archer’s new Point® system Challenges and Opportunities evaluates barrier performance to locate leaks and flowpaths 73 SAND MANAGEMENT AND SAND CONTROL eficiently, accurately and R.J. Wetzel, SPE, Drilling and Completions Senior Adviser and comprehensively. Team Lead, Chevron Energy Technology Company Better well integrity means 74 Re-Evaluation of Gravel-Pack-Sizing Criteria improved profitability and 76 Evaluating Sand-Screen Performance With Sand-Retention Tests reduced integrity risk. and Numerical Modeling It’s time to talk to Archer. 79 Factors Governing the Performance of Multilayered Metal-Mesh Screens

The Point® system

The complete SPE technical papers featured in this issue are available free to SPE members for two months at www.spe.org/jpt. OFFSHORE COMPLETIONS DEEPWATER COMPLETE SOLUTIONS FOR RELIABILITY

Modern technology with a consultative approach

Rather than adapting older technologies for new applications, we design our extensive portfolio of completion technologies for today’s complex, deep wells. Our systems are manufactured to the highest quality, tested beyond required standards, and ﬁeld proven around the world in enhancing efﬁciency, minimizing costs, and reducing risks.

From start to end, we partner with clients to deliver completion systems tailored to the unique needs of speciﬁc wells. With Weatherford, you get a complete and deeply reliable solution for any offshore well.

weatherford.com/offshore

DRILLING & FORMATION EVALUATION WELL CONSTRUCTION COMPLETION & STIMULATION

OFFICERS SOUTH AMERICA AND CARIBBEAN Anelise Quintao Lara, Petrobras 2017 President Janeen Judah, Chevron SOUTH, CENTRAL, AND EAST EUROPE Matthias Meister, Baker Hughes 2016 President Nathan Meehan, Baker Hughes SOUTH ASIA AND THE PACIFIC Salis Aprilian, PT Badak NGL 2018 President SOUTHWESTERN NORTH AMERICA Darcy Spady, Broadview Energy Libby Einhorn, Concho Oil & Gas Vice President Finance WESTERN NORTH AMERICA Roland Moreau, ExxonMobil Annuitant Andrei Popa, Chevron REGIONAL DIRECTORS TECHNICAL DIRECTORS AFRICA DRILLING Adeyemi Akinlawon, Adeb Konsult Jeff Moss, ExxonMobil CANADIAN Cam Matthews, C-FER Technologies HEALTH, SAFETY, SECURITY, ENVIRONMENT, AND SOCIAL RESPONSIBILITY EASTERN NORTH AMERICA Trey Shaffer, ERM Joe Frantz, Range Resources MANAGEMENT AND INFORMATION GULF COAST NORTH AMERICA J.C. Cunha J. Roger Hite, Inwood Solutions COMPLETIONS MID-CONTINENT NORTH AMERICA Jennifer Miskimins, Colorado School of Mines Chris Jenkins, Independent Energy Standards

MIDDLE EAST PRODUCTION AND FACILITIES Khalid Zainalabedin, Saudi Aramco Hisham Saadawi, Ringstone Petroleum Consultants

NORTH SEA RESERVOIR DESCRIPTION AND DYNAMICS Karl Ludvig Heskestad, Det Norske Oljeselskap Tom Blasingame, Texas A&M University

NORTHERN ASIA PACIFIC DIRECTOR FOR ACADEMIA Phongsthorn Thavisin, PTTEP Dan Hill, Texas A&M University ROCKY MOUNTAIN NORTH AMERICA Erin McEvers, Clearbrook Consulting AT-LARGE DIRECTORS

RUSSIA AND THE CASPIAN Khaled Al-Buraik, Saudi Aramco Anton Ablaev, Schlumberger Helena Wu, Santos Ltd.

JPT The Journal of Petroleum Technology® magazine is a STAFF registered trademark of SPE.

Glenda Smith, Publisher SPE PUBLICATIONS: SPE is not responsible for any statement made or opinions expressed in its publications. John Donnelly, Editor EDITORIAL POLICY: SPE encourages open and objective Alex Asfar, Senior Manager Publishing Services discussion of technical and professional subjects per- tinent to the interests of the Society in its publications. Pam Boschee, Senior Manager Magazines Society publications shall contain no judgmental remarks or opinions as to the technical competence, personal Chris Carpenter, Technology Editor character, or motivations of any individual, company, or group. Any material which, in the publisher’s opinion, Trent Jacobs, Senior Technology Writer does not meet the standards for objectivity, pertinence, Anjana Sankara Narayanan, Editorial Manager and professional tone will be returned to the contribu- tor with a request for revision before publication. SPE Joel Parshall, Features Editor accepts advertising (print and electronic) for goods and services that, in the publisher’s judgment, address the Stephen Rassenfoss, Emerging Technology Senior Editor technical or professional interests of its readers. SPE reserves the right to refuse to publish any advertising it Stephen Whitfield, Staff Writer considers to be unacceptable. Adam Wilson, Special Publications Editor COPYRIGHT AND USE: SPE grants permission to make Craig Moritz, Assistant Director Americas Sales & Exhibits up to five copies of any article in this journal for personal use. This permission is in addition to copying rights grant- Mary Jane Touchstone, Print Publishing Manager ed by law as fair use or library use. For copying beyond that or the above permission: (1) libraries and other users David Grant, Electronic Publishing Manager dealing with the Copyright Clearance Center (CCC) must pay a base fee of USD 5 per article plus USD 0.50 per Laurie Sailsbury, Composition Specialist Supervisor page to CCC, 29 Congress St., Salem, Mass. 01970, USA Dennis Scharnberg, Proofreader (ISSN0149-2136) or (2) otherwise, contact SPE Librarian at SPE Americas Office in Richardson, Texas, USA, or e-mail [email protected] to obtain permission to make more than five copies or for any other special use of copyrighted material in this journal. The above permission notwithstanding, SPE does not waive its right as copyright holder under the US Copyright Act. Canada Publications Agreement #40612608. PERFORMANCE INDICES

WORLD CRUDE OIL PRODUCTION+‡ HENRY HUB GULF COAST NATURAL GAS SPOT PRICE‡

THOUSAND BOPD 6 OPEC FEB MAR APR MAY 5 USD/million Btu Algeria 1320 1320 1320 1320 4 Angola 1840 1845 1840 1865

Ecuador 540 552 555 556 3 Indonesia 837 847 844 841 2 Iran 3550 3700 4000 4100 Iraq 4225 4225 4475 4325 1 Kuwait1 2550 2550 2320 2550 JUL SEP FEB JAN JUN DEC APR OCT AUG MAY 2016 NOV Libya 360 320 330 285 MAR Nigeria 2200 2120 2100 1850 Qatar 1517 1537 1537 1537 Saudi Arabia1 9990 10040 10240 10340 WORLD CRUDE OIL PRICES (USD/bbl)‡ UAE 2745 2595 2595 2670 Venezuela 2400 2400 2400 2300 JAN FEB MAR APR MAY JUN JUL AUG TOTAL 34074 34051 34556 34539 Brent 30.70 32.18 38.21 41.58 46.74 48.25 44.95 45.84

WTI 31.68 30.32 37.55 40.75 46.71 48.76 44.65 44.72 THOUSAND BOPD NON-OPEC FEB MAR APR MAY Canada 3797 3767 3429 2811 China 4133 4091 4036 3973 WORLD ROTARY RIG COUNT† Egypt 491 494 493

Mexico 2247 2249 2210 2180 REGION FEB MAR APR MAY JUN JUL AUG Norway 1675 1632 1666 1607 US 532 478 437 408 417 449 481 Russia 10485 10522 10450 10440 Canada 211 88 41 42 63 94 129 UK 1014 986 985 979 USA 9168 8947 8894 Latin America 237 218 203 188 178 186 187 2 12803 12640 12453 12751 Other Europe 107 96 90 95 91 94 96 TOTAL 45802 45546 44670 44128 Middle East 404 397 384 391 389 390 379 Total World 79876 79597 79226 78667 Africa 88 91 90 91 87 82 81 Asia Pacific 182 183 179 190 182 186 194 INDICES KEY + Figures do not include natural gas plant liquids. TOTAL 1761 1551 1424 1405 1407 1481 1547 1 Includes approximately one-half of Neutral Zone production. 2 From this issue of JPT, the “Other” line item also includes Argentina, Australia, Azerbaijan, Brazil, Colombia, Denmark, Equatorial Guinea, Gabon, India, Kazakhstan, Malaysia, Oman, Sudan, Syria, Vietnam, and Yemen. Monthly production from these countries was listed individually WORLD OIL SUPPLY AND DEMAND3‡ in previous JPT issues. Ongoing work on the US Energy Information Administration (EIA) website is disrupting the regular updating of these countries’ production numbers. Additional annual and monthly MILLION BOPD 2016 international crude oil production statistics are available at: http://www.eia.gov/beta/international/. Quarter 3rd 4th 1st 2nd 3 Supply includes crude oil, lease condensates, natural gas plant liquids, biofuels, other liquids, and refinery processing gains. SUPPLY 96.38 96.47 95.59 95.79 † Source: Baker Hughes. ‡ Source: EIA’s Monthly Energy Review. DEMAND 95.06 94.15 94.14 95.04 Numbers revised by EIA are given in italics.

66 JPT • OCTOBER 2016 CAMShale FRACTURING FLUID DELIVERY AND FLOWBACK SERVICE

Monoline* ﬂ anged-connection fracturing ﬂ uid delivery technology reduces installation time by more than 60%, Eagle Ford Shale.

More stages. Less cost. Integrated reliability. The CAMShale* fracturing fl uid delivery and fl owback service provides seamless delivery of hydraulic fracturing fl uid from the pumping service provider’s missile trailer to the wellbore, integrating the entire high-pressure surface completion system through fl owback and well testing. An operator in the Eagle Ford Shale reduced hydraulic fracturing costs by approximately USD 2.7 million per month and achieved 100% valve reliability with this integrated service, which uses only Cameron equipment, maintenance programs, work procedures, and a multiskilled crew.

Find out more at cameron.slb.com/CAMShale

AFRICA holds a 13.5% interest in the block, which is 1.096 million bbl followed a nearly identical operated by OGDC. net total of 1.101 million bbl produced during Z Tullow Oil announced in mid-August the first quarter. Production in both periods that first oil has flowed from the Tweneboa, AUSTRALIA/OCEANIA exceeded a net output of 997,904 bbl during Enyenra, and Ntomme (TEN) fields the fourth quarter of last year. Tethys is one offshore Ghana, which was on time and Z Beach Energy completed logging at of the largest onshore oil concession holders on budget for the project’s development the Callawonga-12 oil development well in in the Sultanate of Oman with a current net plan approved by the government in May Petroleum Production License 220 on the production of about 12,000 B/D. 2013. The company expects oil output to western flank of South Australia’s Cooper ramp up gradually through the rest of the Basin. The well was being cased and Z Gas Plus Khalakan (GPK) reported that year toward a production facility capacity suspended after reaching a total depth of the No. 1 sidetrack well in the Shewashan of 80,000 B/D. Tullow is the operator 4,754 ft in the Westbourne formation. The field of the Kurdistan Region of Iraq was of the TEN fields with a 47.18% stake. Its presence of a gross oil column of 13 ft, or successfully drilled and recompleted recently joint-venture partners are Anadarko (17%), a net column of 6.6 ft, was interpreted in as a horizontal producer in the Qamchuga Kosmos Energy (17%), Ghana National the primary target, the McKinlay member formation. The well is producing from Petroleum Corporation (15%), and and the Namur sandstone. The company 500 B/D to 700 B/D of oil, although it may PetroSA (3.82%). operates and holds a 75% interest in the require further stimulation to meet predrill license, with the remaining stake held by production estimates that were based upon ASIA Cooper Energy. the original Shewashan-1 vertical well. The company expects total production at the Z Rosneft has made a discovery at the EUROPE Shewashan development to hit 10,000 B/D PLDD well in the Wild Orchid gas condensate early next year. GPK is the operator with an field in Block 06.1 of the Nam Con Son Basin Z Premier Oil has discovered oil at the 80% interest in the development. offshore Vietnam. The discovery is being Bagpuss 13/25-1 well on the Halibut Horst evaluated for the volume of reserves and in the Outer Moray Firth of the United NORTH AMERICA commercial attractiveness, and there is a Kingdom North Sea. The well encountered potential synergy with the nearby Rosneft- 41 ft of hydrocarbon-bearing sands within a Z Anadarko has drilled a successful operated Lan Tay production platform, the 68-ft hydrocarbon column, which was in line appraisal well at the Shenandoah field in company said. Rosneft is the operator and with predrill estimates, and reached a total Walker Ridge Block 51 of the US Gulf of holds a 35% interest in Block 06.1, with the depth of 1,532 ft. Hydrocarbon and reservoir Mexico. Drilled in 5,900 ft of water to a total remaining interests held by Petrovietnam analysis are under way to determine depth of 31,100 ft, the Shenandoah No. 5 (20%) and ONGC (45%). whether the discovery is commercial. The well encountered more than 1,000 net ft of joint-venture partners in the Bagpuss high-quality oil pay in the Lower Tertiary Z Oil & Gas Development Company prospect are operator Premier, with a Wilcox sands. The well’s results indicate an (OGDC) has successfully tested and 37.5% interest, Maersk (25%), EnCounter extension of the field reservoir boundaries completed the X-6 and X-7 development Oil (15%), North Sea Energy (15%), and to the east. Anadarko, the operator, holds wells in the Nashpa field in the Karak and Groliffe (7.5%). a 33% interest in the field, with other Kohat district of Pakistan’s KPK Province. interests held by ConocoPhillips (30%), The Nashpa X-7 well, drilled to a 15,079-ft Z Statoil has started oil and gas production Cobalt International Energy (20%), and depth, produced 2,700 B/D of oil and from the Fram C East well offshore Norway. Venari Resources (17%). 7.4 MMscf/D of gas through a 36/64-in. The well was drilled from the Fram subsea choke at wellhead flowing pressure of template, and production is being tied SOUTH AMERICA 1,886 psig. The Nashpa X-6 well, drilled back to the Troll C hub in the North Sea. to a 16,486-ft depth, produced 2,550 B/D The company said that project capital cost Z Andes Energia announced discoveries of oil and 16.3 MMscf/D of gas through a had been reduced to USD 73 million from in the Chachahuen Block in Mendoza 36/64-in. choke at wellhead flowing pressure an originally estimated USD 97 million as a Province, Argentina. About 20 ft of net of 2,667 psig. OGDC is the operator with result of “a simple, smart well concept and oil pay was discovered by exploration a 56.45% interest in the field. Pakistan significantly increased drilling efficiency.” well Cerro Redondo x-1 in the Rayoso Petroleum (28.55%) and Government Statoil, the operator, holds a 45% interest formation sandstone, and gas was found by Holdings (15%) have the remaining interests. in the well, with ExxonMobil (25%), Engie exploration well La Orilla x-1 in the deeper (15%), and Idemitsu (15%) holding the horizon of the Lotena formation. Gas was Z Jura Energy reported a gas discovery remaining interests. also found by exploration well Remanso at the Khamiso-1 exploration well in del Colorado x-1 in the deeper horizon of Pakistan’s Guddu Block. The well was drilled MIDDLE EAST the Cuyo Group. The discoveries reaffirm to a total depth of 2,470 ft in an Eocene “the large-scale potential of Chachahuen,” limestone of the Pirkoh formation. During a Z Tethys Oil is maintaining its increased said company Chairman Nicolas Mallo short-duration prestimulation test conducted level of oil production in Oman, where Huergo. Andes Energia is developing on a 32/64-in. choke, the well encountered it produces from onshore blocks 3 the block in a joint venture with YPF and an average gas flow of 2.95 MMcf/D. Jura and 4. A second-quarter net output of Energia Mendocina. JPT

8 JPT • OCTOBER 2016 You can’t aord to use the wrong additive.

Saving money using products not suitable to your application costs you a lot more in the long run. Our chemical expertise and decades of experience serving the oilield industry ensures you’ll get the right specialty additive for your application for the right price.

Our technical expertise is just an email or call away. [email protected] 1 800 458 4034

Ingevity_JPT_April 2016.indd 1 3/7/16 2:01 PM RISK AND REWARD

Mitigating Risks To Improve Project Results Janeen Judah, 2016 SPE President

I am honored to serve the Society of economic hurdles. Even with today’s sophisticated probabilis- Petroleum Engineers as president this tic economic analysis, results still fall short of expectations and year. Each president chooses a theme, promises to stakeholders. and for me, my theme was obvious: risk and reward. Too Many Projects Fail Ours is a risky business, and that is The stakes are incredibly high. Independent analysis of large what makes it fun and challenging. Risk capital projects has shown that as an industry, we are not very also drives the big rewards when what we good at delivering on our promises. In its spotlight article on do pays off. Mitigating risk is the very core of what we do on oil and gas megaprojects, Ernst & Young estimates that 64% of a daily basis: drilling risk, subsurface risk, cost/schedule risk, major capital projects overrun on cost, and 74% face schedule performance risk, safety risk, geopolitical risk, and price risk. delays. Cost and schedules are overrun in every region of the Projects are a constant juggle to find the sweet spot where we world. Independent Project Analysis (IPA) comes to very similar produce oil and gas safely, responsibly, and profitably. But as an conclusions in its subscription benchmarking service, estimat- industry, we make decisions to eliminate risk through analysis ing that 60% of projects overrun costs by more than 20% while rather than manage the risk that will always be there. The pe- only 18% of projects meet both cost and schedule targets. IPA troleum industry is fundamentally a commodity business. Com- also estimates that projects deliver 35% less of their planned modity prices change, but combine that with a capital-intensive overall net present value from financial investment decision to business with big upfront investments and long payouts, and lookback. Some estimate that more than 75% of projects deliver commodity price swings hit us hard. more than 25% less value than promised. Engineers are paid to predict the future, especially produc- On the production and reserves side, the story is not any bet- tion, reserves, cost, and schedule. We cannot control price risk, ter. Ernst & Young’s 2015 US Oil and Gas Reserves study shows but we can improve performance. I have always been interest- that integrated companies failed to replace US production for ed in the intersection of engineering and risk evaluation with 2010–2014, despite more than USD 2 billion in capital invest- oil and gas economics. I wrote my master’s thesis on what we ment. Independents fared much better at production and re- would now call “decision analysis.” I was intrigued by deter- serve replacement during this time period due to high activity ministic project and economic evaluation and how data are in- in the US shale plays, but shale production will decline rapidly terpreted within the bands of possibility to get a project over with diminished development programs.

Capex and production for five supermajors* (indexed to 2005) 300 250 200 Development cost Production 150 100 50

Percentage Increase Percentage 0 2005 2006 2007 2008 2009 2010 2011 2012 2013

*Includes BP, Chevron, ExxonMobil, Shell, and Total. Source: Annual reports; Strategy& analysis

Oil and gas production has not kept up with development costs. Source: PricewaterhouseCoopers

To contact the SPE President, email [email protected].

10 JPT • OCTOBER 2016 Those new barrels came dearly. During the period 2005– ranges that are too narrow, confirmation bias, anchoring, and 2013, PricewaterhouseCoopers’ (PwC) analysis showed that general overconfidence. With all those negative results, can en- global production for the five super majors (BP, Chevron, gineers challenge ourselves to reduce risk and improve reward? ExxonMobil, Shell, and Total) has remained flat, while develop- Possible solutions include: ment cost almost tripled (Tideman et al. 2014). 1. Technology. Our past performance has shown that advances in technology can help reduce risk. We can We Overestimate and Under-Deliver increase the predictability of outcomes through better All published studies agree that, as an industry, we consistent- analysis. We can reduce failures and increase safety ly overestimate production attainment and underestimate cost and operational performance by better monitoring of and schedule. IPA’s analysis of production attainment shows the our operations. And we can create new technologies to too-narrow range of expectations and failure of the results to increase our opportunities. History has shown that our fall within the P10/P90 range. The average production attain- industry tends to make technology leaps in hard times ment is only 78%, so production facilities are also oversized. because we are forced by economics to innovate. I will How could we be so wrong? discuss this issue more in future columns. Fundamentally, our industry usually manages the business to 2. Better engineering. We need a rebirth of technical short-term metrics. For publicly held companies, that usually oversight for our initial project designs. I believe that at means analysts’ and shareholders’ expectations, annual perfor- least part of this will come through better institutional mance promises, and executive performance goals. For national memory contained in design systems. We also need to companies, the stakeholders are different, but the pressure to reduce complexity and standardize more designs. Too meet annual metrics, such as cash delivered to the finance min- many projects are completely custom-designed, creating istry, are very similar. openings for design errors and reducing opportunities for The oil business is a long-term game. We invest billions in cost savings through standardization. projects destined to produce for 30 years or more. Future un- 3. Better business analysis. Are we doing the right discovered technology advancements will produce margin- development, in the right place, at the right time? Large al barrels over those 30+ years, but we do not know the final capital projects have skewed project risks, not normal estimated ultimate recovery until the last barrel is produced. distributions, with a lot of ‘tail risk’ (low probability but Short-term metrics in a long-term business is not a formula for often high consequence). Do we assess those risks well? sustainable success. Do we include learnings from prior projects to improve our economic analysis? Do we monitor changes in risk So, What Can We Do Differently? profiles as projects move forward and act on those Ours is a risky business and not for the fainthearted. But we can changes? mitigate some of the unnecessary risks. Many of our gaps are in: 4. Make tough decisions. Finally, management should 1. Engineering oversight. Most people would agree that “just say no” to some projects. Reward systems should engineering oversight has declined in the last 20 years as reinforce right decisions, not encourage making promises operators have outsourced almost all design to service and that we cannot keep. We need to find new ways to monitor engineering, procurement, and construction contractors. and reward long-term actual results of projects, not short- In the last 10 years or so, these contractors have been term metrics such as final investment decision and first stretched thin to meet the very high demand for their oil promises. services. Operators were also people-constrained as Finally, I think we need a new level of cooperation between their activities ramped up, along with the first wave of operators and the service sector. Service companies are essen- retirements from the big crew change. Many misses on tial to implement projects, yet when times get tough, operators cost and schedule have been due to poor interpretation stop the music too quickly. In addition, innovations will be more and design mistakes that could have been avoided with of a cooperative venture between the asset owner and the ser- more technical review. vice sector. We need a healthy service sector to effectively devel- 2. Executive overconfidence. The PwC article summarized op projects, and we need to find a better way to work together. the problem very well: Management sets optimistic goals Risk and reward drive all of our decisions. We balance many, for a first oil date, and then creates a very tight project often unmeasurable, risks to develop resources to provide ener- to meet that deadline. Once projects are underway, it gy for the world. We can do a better job for ourselves, our com- becomes apparent that the deadline cannot be met within panies and our stakeholders. JPT the given budget. Project teams try to speed the schedule by adding resources, and then often blow both cost and For Further Reading schedule. Tideman, D., Tuinstra, H.T., and Campbell, B.J. 2014. Large The real question is: Would we make the same investments Capital Projects in the Oil and Gas Sector: Keys to Successful if we were honest with ourselves around investment metrics, Project Delivery, http://www.strategyand.pwc.com/media/file/ especially cost and schedule? Analysis of project results con- Large-capital-projects-in-the-oil-and-gas-sector.pdf (accessed sistently reveals the same issues across all projects: outcome 7 September 2016).

JPT • OCTOBER 2016 11 COMMENTS EDITORIAL COMMITTEE Bernt Aadnøy, University of Stavanger

Syed Ali—Chairperson, Consultant

Tayfun Babadagli, University of Alberta Resilient Production William Bailey, Schlumberger Mike Berry, Mike Berry Consulting John Donnelly, JPT Editor Maria Capello, Kuwait Oil Company Frank Chang, Saudi Aramco

Simon Chipperfield, Santos This issue of JPT offers two suggestions as to why crude oil Alex Crabtree, Hess Corporation prices may not rise significantly in the near future. The role of Gunnar DeBruijn, Schlumberger unconventional production becoming the industry’s “swing pro- Mark Egan, Retired

ducer” in reaction to oil prices is laid out in the Management Mark Elkins, ConocoPhillips column beginning on page 42 (see also Comments, March 2015 Alexandre Emerick, JPT). And the continued strength of Permian Basin output is Petrobras Research Center described in an article beginning on page 24. Niall Fleming, Statoil Although global crude oil production has declined in some regions since the collapse in oil prices began, and there have been massive cuts in Ted Frankiewicz, SPEC Services company spending, world output has remained stubbornly resilient. Global petroleum Stephen Goodyear, Shell production rose in 2015 and 2016, and is projected to rise slightly in 2017, according to Omer M. Gurpinar, Schlumberger the US Department of Energy’s Energy Information Administration (EIA) Short-Term A.G. Guzman-Garcia, Retired Energy Outlook, published in September. Greg Horton, Consultant US crude oil production averaged 9.4 million BOPD in 2015 and is forecast to aver- John Hudson, Shell age 8.8 million BOPD this year and 8.5 million BOPD in 2017. The US production forecast is 0.2 million BOPD higher than one the EIA made in August as it cited an uptick Morten Iversen, Karachaganak Petroleum in drilling activity, rig efficiency, and well productivity. Leonard Kalfayan, Hess Corporation The steep decline in North American rig activity over the past year came primarily Thomas Knode, Statoil from the onshore light tight oil sector, which has short development cycles that allow Sunil Kokal, Saudi Aramco

producers to react quickly to price signals. Production from the US Gulf of Mexico is Marc Kuck, Eni US Operating expected to reach record highs next year due to previous big deepwater discoveries. Jesse C. Lee, Schlumberger Changes to Performance Indices Page Douglas Lehr, Baker Hughes Beginning with this issue, the World Crude Oil Production table in the Performance Silviu Livescu, Baker Hughes Indices page will include fewer countries. The table lists the monthly crude oil pro- Shouxiang (Mark) Ma, Saudi Aramco duction numbers from the EIA. Ongoing work on the EIA’s website is disrupting the John Macpherson, Baker Hughes

regular updating of some countries’ production numbers. This has been going on for Graham Mensa-Wilmot, Chevron several months, which is causing considerable delays in our reporting of this infor- Stéphane Menand, DrillScan mation. EIA recently launched the redesign of its International Energy Portal at http://www.eia.gov/beta/international and is currently beta testing it. Badrul H. Mohamed Jan, University of Malaya Over the years, our readers have indicated that they use the Performance Indices Michael L. Payne, BP plc to gather data for their research and analyses. Therefore, we decided not to source Zillur Rahim, Saudi Aramco production data from another provider in order to preserve the continuity of data Eric Ringle, FMC Technologies

for our readers. Other reporting services use different methodologies to collect Martin Rylance, BP GWO Completions production information. Engineering In this issue of JPT and going forward, crude oil production data (crude oil including Robello Samuel, Halliburton lease condensate) will be sourced from the EIA’s Monthly Energy Review. The Review Otto L. Santos, Petrobras includes statistics on total energy production, consumption, and trade; energy prices; overviews of petroleum, natural gas, coal, electricity, nuclear energy, renewable Luigi A. Saputelli, Frontender Corporation energy, and international petroleum; carbon dioxide emissions; and data unit conver- Sally A. Thomas, ConocoPhillips sion values. It is available at http://www.eia.gov/totalenergy/data/monthly/, including Win Thornton, BP plc historical data back to October 1974. JPT Xiuli Wang, Baker Hughes

Mike Weatherl, Well Integrity, LLC

Rodney Wetzel, Chevron ETC

Scott Wilson, Ryder Scott Company

Jonathan Wylde, Clariant Oil Services

To contact JPT’s editor, email [email protected].

12 JPT • OCTOBER 2016 JPT Magazine - October | 8.125” x 10.875” | FINAL

peloton.com/cloud

An overhaul for Land Software was desperately needed. So we did it. We designed our new land data management system from the ground up and ended in the Cloud. Because our design is based on real user experiences, LandView™ captures and visualizes your land information like no other before.

LandView™ is a web-based management solution in the MasterView® suite of products.

1660-24680_JPT_Magazine_Oct_Typewriter_8_125x10_875_FINAL.indd 1 2016-08-29 10:32 AM GUEST EDITORIAL

Risk Management at NASA and Its Applicability to the Oil and Gas Industry

David Kaplan, Safety and Mission Assurance, NASA Johnson Space Center

On initial consideration, one might an option” because of the consequences (both of hardware failures and human reasonably ask: What can the Nation- to life and the environment should a cat- errors) that led to the high-consequence al Aeronautics and Space Administra- astrophic mishap occur. near-meltdown of the nuclear core. As a tion (NASA) contribute to the oil and At NASA, we use qualitative tech- result, the US Nuclear Regulatory Com- gas industry? niques—such as fault trees, failure mission has required a facility-specific About 3 years ago, a senior principal modes and effects analyses, hazard PRA for every nuclear power plant in the at Deloitte Advisory’s Energy & Resourc- assessments, etc.—to understand risk United States. es Operational Risk Group reached out based on statistics, experience, or pos- In February 2003, Space Shuttle to NASA to better understand the safe- sibilities that our engineers can antici- Columbia was lost on re-entry when a ty culture at NASA with the intent of pate. Similarly, upstream oil and gas piece of insulation foam broke off from understanding how that culture might exploration and production uses qual- the external tank and struck the wing translate to oil and gas operations. Very itative techniques—such as process leading edge of the space shuttle. Recog- quickly, the conversation expanded to safety methods, barrier analyses, bow- nizing that the cause of this accident was the realm of risk management. tie charts, hazard identification, haz- a low-probability, high-consequence Working with Deloitte, NASA came to ard and operability studies, etc.—to event, NASA committed to strengthen appreciate the remarkable similarities assess risk. At NASA, these qualitative its safety and mission assurance capa- between an offshore deepwater facil- approaches are augmented by a quanti- bilities. PRA was adopted and embraced ity and the International Space Station. tative risk-assessment technique called by the Space Shuttle and International Both exist in extremely hostile environ- probabilistic risk assessment (PRA) to Space Station programs. ments. Both function in remote loca- uncover and mitigate low-probability A PRA creates a rigorous logic flow for tions where movement of crew and sup- sequences of events that can lead to a complex system. Every safety-related plies must be carefully choreographed. high-consequence outcomes. hardware component is captured as a Both are extremely complex engineer- node and quantitative reliability per- ing structures where human reliability Why PRA? formance numbers are assigned to each plays a critical role in mission success, The technique of PRA was developed by possible outcome. For example, a pump and both have a deep commitment to the nuclear power industry and initially can function as commanded, remain off personal and process safety. published in mid-1975, though not wide- when commanded on, remain on when It also should be noted that both have ly publicized. However, the investigation commanded off, or operate at only a par- dedicated teams—the onboard crew of the Three Mile Island incident in 1979 tial level of capability. Human actions and the onshore support experts—that revealed that the PRA had documented also are captured as logic nodes that live by the mentality that “failure is not the sequence of low-probability events can have quantitative reliability information assigned to them. For example, a person can push the correct button David Kaplan is a leader at the National Aeronautics and Space within the assigned timeframe, push Administration (NASA) Johnson Space Center with more than the wrong button, push the correct but- 30 years of experience in aerospace engineering and management. ton outside the assigned timeframe, or He has been a project manager for Mars hardware, a space shuttle do nothing. flight controller, and managed the crew health-care equipment on A rigorous PRA also can account for the International Space Station. Most recently, Kaplan served as chief of the Quality Division at the space center. In that position, he common cause failures in both hard- managed the NASA Failure Analysis Laboratory, which is ware and software. For example, if a instrumental in detecting counterfeit parts and assisting projects to reduce their risks pump fails in one system, then all simi- associated with fabrication and operations. Currently, he is involved in assessing the lar pumps from the same lot/vendor that applicability of NASA’s quantitative risk-management techniques to the oil and gas may exist in entirely separate systems industry. He may be contacted at [email protected]. are now suspect.

14 JPT • OCTOBER 2016 Given a high-consequence undesir- Projects With Anadarko presented to Anadarko management on able event (such as loss of hydrocar- In addition to working with other gov- 28 July 2016. A final report was deliv- bon containment), every single path ernment agencies, NASA has a special ered at the end of August. through the logic model that could lead mechanism for working with commer- While it is not my place to discuss to that event can be assessed. Should a cial organizations. In situations where any facet of the work that NASA did in low-probability action occur (perhaps NASA has unique facilities, technolo- partnership with Anadarko, I am able to a highly trained individual is distract- gies, techniques, or experiences, it state that Anadarko followed up the BOP ed and fails to observe a change in the may enter into a reimbursable agree- work by asking NASA to perform a PRA mud flow rate in vs. the mud flow rate ment (referred to as a Space Act Agree- of the dynamic positioning system being out), then every other subsequent low- ment) to perform work for the mutu- considered for the Shenandoah devel- probability action(s) can be identified to al benefit of the Space Act partner opment. The PRA for that began in June mitigate the undesirable event. and NASA. and is ongoing. Anadarko Petroleum is working with NASA is just beginning to work Why BSEE? suppliers to develop various subsea with BSEE and the oil and gas indus- In April 2015, I attended a conference equipment with working pressures of try. Our hope is that the benefits of that explored crossover technologies more than 15,000 psi for their Shenan- a quantitative assessment of risk will that might have applications to the doah field in the Gulf of Mexico. The both complement the industry’s cur- space and energy sectors. Brian Salerno, director of Engineering and Technol- rent approach to risk management as director of the Bureau of Safety and ogy Global for Anadarko, Jim Raney, well as help with risk-informed deci- Environmental Enforcement (BSEE), wanted to have a set of eyes from out- sion making. It has worked for NASA in gave a presentation that included an side the industry look over the approach the exploration of space. Could it also acknowledgement that BSEE would to risk management being used by work for offshore deepwater drilling need better tools to assess risk as oper- his team for this activity. Anadarko and production operations? JPT ators moved to deeper drilling; higher entered into a Space Act Agreement temperatures and pressures; less well with NASA in November 2014, enabling understood environments; and intro- NASA to engage and participate in duced new, emerging technologies. He the project. suggested the need for a quantitative Anadarko introduced NASA to the approach to risk management. unique layout of bowtie charts (an inte- The outcome of several meetings was gration of fault trees and event trees), to a US Government Interagency Agree- the barrier analysis approach, etc. Our ment between BSEE and NASA signed eventual assessment back to Anadarko in January 2016, formalizing a partner- was that all their risk-management tech- ship between the two organizations for niques were qualitative and, while excel- 5 years. Under this agreement, NASA lently executed, might not capture low- will work with BSEE to develop a pro- probability, high-consequence events. cess for preparing PRAs for offshore NASA explained its use of quantitative deepwater drilling and production oper- PRA modeling to capture these types ations. Together with the oil and gas of events. industry, we will evaluate whether the Anadarko was open-minded to the additional insights of a PRA provide possibility that PRA might provide meaningful information for the opera- insights not otherwise available through tors and contractors as well as for the their more traditional qualitative risk- regulator, BSEE. management techniques. Since the NASA has a document to guide in project would require a blowout pre- the preparation and execution of a PRA venter (BOP) with a rated working pres- referred to as the “Probabilistic Risk sure up to 20,000 psi, Anadarko asked Assessment Procedures Guide for NASA NASA to prepare a PRA for a gener- Managers and Practitioners” (NASA ic 20,000-psi BOP. The work began in document number SP-2011-3421). The October 2015. first task that BSEE has given NASA is to The development of the BOP PRA was rewrite the PRA Guide to be relevant to a true partnership; Anadarko provided the oil and gas industry. NASA is sched- world-class expertise on the design and uled to deliver the initial version of the operations of BOPs, and NASA provided document to BSEE by the end of the world-class modelers and data analysts. 2016 calendar year. The results of the BOP PRA model were

JPT • OCTOBER 2016 15 TECHNOLOGY APPLICATIONS

Chris Carpenter, JPT Technology Editor

Shear-History Simulator additional stainless-steel pump is able ments and basic sediment and water key The rheology of fracturing fluids used to deliver 20 mL/min of crosslinker. The performance indicators, and enables a by the oil and gas industry is affected by simulator features a maximum operating more-controlled water/oil interface. This a range of factors and conditions, and pressure of 2,000 psi at ambient temper- allows steam-assisted-gravity-drainage shear-history simulators have proven use- atures; features also include multiple tub- operators to reduce equipment fouling ful for determining how well these fluids ing configurations, four shear-history- and upsets as well as decrease slop oil perform in transporting proppants. The tube assemblies, and a panel-mounted generation and disposal, which results Chandler Engineering Model 5600-Auto computer with touch-screen interface in less recycling and more production shear-history simulator simplifies the and automatic loop-flushing features. throughput. For conventional producers, process of preparing and loading water- ◗ For additional information, visit the products can provide oil and water based fracturing fluids into rotational vis- www.chandlereng.com. separation to meet water-discharge- cometers (Fig. 1). The process begins regulation requirements. with a gel-based fluid that is placed into Fluids-Separation Technologies ◗ For additional information, visit a pressurized supply reservoir that deliv- High-quality oil and water separation is www.bakerhughes.com/SNAP. ers the fluid to an injection pump. The critical to the success of most upstream fluid is pumped through a combination oil production facilities but often results Hydraulic-Fracturing System of tubes at various rates and durations in increased operational challenges The ability to use multiple fuel types, to simulate pumping conditions experi- and costs associated with the process. including field gas, offers the potential enced during fracture treatments. Pres- Baker Hughes’ TRETOLITE SNAP fluids- for fracturing-cost-reduction. Evolu- sure is applied to the reservoir through separation technologies help operators tion Well Services’ hydraulic-fracturing an air regulator mounted onto the face overcome these challenges by achieving system uses 100% electrically powered panel. The gel pump is constructed of dry oil and clean water (Fig. 2). This gen- process equipment (Fig. 3). The electric high-pressure, corrosion-resistant stain- eration of water clarifiers and demul- power is generated by burning natural less steel and is driven by a servo motor sifiers exceeds current performance gas through a customized-for-purpose that can deliver 0 to 140 mL/min. An standards, including oil-in-water require- turbine system. For primary equipment,

Fig. 2—Baker Hughes’ TRETOLITE SNAP fluids-separation technologies Fig. 1—The Chandler Engineering Model 5600-Auto shear-history simulator. achieve dry oil and clean water.

16 JPT • OCTOBER 2016 BroadShear OFF-CENTER TOOL JOINT SHEAR RAMS

Successful shear test of an of -center tool joint pin and box, unshearable using conventional technologies.

Distinguished by shear ability.

BroadShear rams now make it possible to shear any part of the drillstring above the BHA, including tool joint hardbanding and of -center tubulars—components long considered unshearable.

Designed to stringent Bureau of Safety and Environmental Enforcement (BSEE) regulations, BroadShear rams bring enhanced security and dependability to marine drilling, so you can maximize safety in of shore operations.

Watch the BroadShear rams in action at cameron.slb.com/BroadShear

BroadShear is a mark of Schlumberger. © 2016 Schlumberger. 16-DRL-177836 the system uses a blender with ambi- comfort, using real-time data-acquisition Slim-Profile Formation Sampler dextrous suction/discharge capability software. Remote-control cameras moni- The Compact formation sampler is a slim, with rate capacity of up to 240 bpm. The tor equipment and allow the removal of lightweight tool that provides efficient chemical additive system features low- all personnel from high-pressure areas pressure/volume/temperature (PVT) and high-rate pumps and flowmeters. and exposure to chemical hazards and sampling in diverse boreholes (Fig. 4). The fracturing pumps are mounted with silica dust. The sampler can capture up to three dual pumps on each trailer. The data ◗ For additional information, visit true 700-cc PVT samples in a range of van is designed for safety, control, and www.evolutionws.com. wellbore sizes from less than 3 in. to 14 in. The self-centering design reduces formation-sticking risks, enables a more-efficient and faster connection with the target zones, and makes it possible to deploy the tool on traditional wireline or through drillpipe. The sampler can be deployed on monocable or heptacable, which is particularly advan- tageous in offshore environments. Real- time and recorded pressure data acquired by the sampler can be used to calculate formation-pressure gradients, ascer- tain fluid-contact levels, determine fluid mobility, and define formation permeability. The combined capabilities of the sampler enable operators to obtain reliable, representative formation fluid samples by wireline, resulting in operations that are safer and simpler than those using traditional formation-testing tools. ◗ For additional information, visit www.weatherford.com.

Pressure-Reducing System Tech-Flo’s Surface Jet is a Venturi-based pressure-reducing system. This system operates by channeling a high-pressure source (high-pressure wells, high-pressure Fig. 3—An Evolution Well Services hydraulic-fracturing facility, featuring gas from a process system, or a high- electrically powered equipment. pressure water or oil line) through the surface jet to create a pressure drop, allowing a lower-pressure system to enter the higher-pressure system (Fig. 5). Typi- cal applications include reducing line- operating pressure, reducing vessel pressure, high-pressure well boosting of low-pressure wells, and artificial-lift support. The system functions by using the Venturi effect, a special case of Bernoulli’s principle stating that as fluid flows through Fig. 4—The Compact formation sampler from Weatherford provides efficient PVT sampling in diverse boreholes. a pipe with a constriction in it (nozzle), the fluid must speed up in the restriction, reducing its pressure and producing a vac- uum through the Bernoulli effect. This vac- uum is what allows a low-pressure system to enter a higher-pressure system. JPT ◗ For additional information, visit Fig. 5—Schematic of Tech-Flo’s Surface Jet pressure-reducing system. www.tech-flo.net.

18 JPT • OCTOBER 2016 Our focus: Reducing your cost.

FMC Technologies works closer and earlier with you to deliver Surface Integrated Services that reduce your costs during drilling, completion and production. With a strong focus on improving our customers’ efciencies, we simplify the way equipment and services work at your well-site, cutting installation time and minimizing the number of people needed onsite. Finally a process that delivers streamlined and safer surface operations. Rethink. Reinvent. Reimagine.

www.fmctechnologies.com

#RethinkReinventReimagine TECHNOLOGY UPDATE

A Holistic Approach To Controlling Torsional Dynamics in the Drillstring

Liam Lines, SPE, Weatherford

Understanding drilling dynamics, the potential for damage, effect on drill- FSS is characterized by distinct peri- downhole shock and vibration that ing efficiency and—importantly—most ods of zero or low drill-bit RPM (the occur while drilling, is a crucial step to effective means of mitigation. “stick”) followed by an abrupt release improving drilling efficiency and reduc- (the “slip”) where the string unwinds ing nonproductive time (NPT). Differing Dynamics at an extremely high rate that in some Industry statistics suggest that 20% In addition to full stick/slip (FSS), the cases has exceeded the average drill- to 30% of all downhole failures are modes include low-frequency torsional string rotation rate by more than a fac- related to adverse drilling dynamics oscillation (LFTO) and high-frequency tor of six. The high torque (sticking) (SPE 127413 and SPE 98150), and that torsional oscillation (HFTO) (Fig. 1). required to create FSS can be generated drilling tool providers spend roughly LFTO and FSS both typically occur at by multiple points in the drillstring, not USD 750 million per year on repair and less than 2 Hz; HFTO occurs at between just the drill bit. maintenance. This amount pales com- 50 and 250 Hz. HFTO, the most damaging form of pared with the untold costs borne by LFTO, probably the most common of torsional dynamics, is the excitation of exploration and production companies all three modes, is the excitation of a a local torsional resonance in the bot- as the result of drilling and completion fundamental torsional natural frequency tomhole assembly (BHA) and is preva- inefficiencies, NPT, and tools lost in of the drillstring. The excitation causes lent when drilling through hard carbon- the hole. the drillstring to act as a long torsion- ate formations (SPE 167968). Because Preventing adverse dynamics requires al spring, which results in sinusoidal of the high frequency, the resultant a strong understanding of the fundamen- oscillations in bit speed and torque. The torque and RPM fluctuations are atten- tals, particularly in addressing torsion- oscillation frequency is a function of the uated as they travel up the drillstring al dynamics: time-based variations in drillstring length and stiffness, and is and are not measureable at the surface. downhole rotational speed and torque. typically less than 0.5 Hz for all drill- Without downhole sensors providing a A common misconception is that tor- strings (SPE 21945). LFTO is character- real-time indicator, HFTO can quickly sional dynamics is limited to stick/slip. ized by a maximum revolutions per min- cause extensive and premature failure However, stick/slip is only one of three ute (RPM) of up to twice the average of downhole drilling tools and polycrys- distinct modes of torsional dynamics, drillstring RPM and a bit that may stop, talline diamond compact (PDC) drill bits each with its own unique indicators, but only instantaneously. (SPE 49204).

500 A B C D 400

300

200

100

0 Angular Velocity (RPM) Velocity Angular 0 5 10 15 20 25 30 35 40 –100 Time (seconds)

Fig. 1—There are three distinct modes of torsional dynamics (A) HFTO is marked by rapid downhole speed changes. (B) LFTO is indicated by sinusoidal variations in RPM (C) FSS has distinct periods of low or zero RPM during a stick, followed by a ramp-up in RPM during a slip. (D) The modes may also occur simultaneously; the example shows FSS with periods of HFTO during the slip phase (arrowed). Source: Weatherford.

20 JPT • OCTOBER 2016

or the addition of novel technologies offer significant benefits. In some cases, a simple increase in the drillpipe outside diameter can significantly reduce FSS and LFTO (SPE 151133). The addition of a positive displacement motor (PDM) into the BHA can significantly reduce the severity of both LFTO and HFTO. Because the amplitude of LFTO is based on string speed, slow- ing the rotation and providing a second RPM source near the drill bit reduces the amplitude of drill-bit RPM oscillation. PDMs also act as a low-pass filter, which prevents HFTO generated at the drill bit from travelling across into the Fig. 2—MWD data showed that the highest shock and vibration levels in a PDM BHA occurred while circulating off-bottom with no rotation. measurement-while-drilling (MWD)/ Source: Weatherford. LWD sensors above. It should be noted that PDMs may cause a dramatic increase in lateral A Multistage Solution unscrew downhole, and bit wear to the and axial shock and vibration. During Understanding the exact nature of tor- face and gauge are indications of FSS. full-scale flow-loop testing, Weather- sional dynamics is essential for deter- Rapid fatigue failure of alignment pins, ford found that some PDMs generated mining when and how to mitigate it. The such as those in RSSs, axial cracks in more than 50 times the peak axial and most effective strategies involve four tubulars, dislodgement of large compo- 20 times the peak lateral accelera- aspects: measurement, modeling, miti- nents from printed circuit boards, and tions, compared with the baseline. Field gation, and monitoring. spalling of PDC cutters signal HFTO. trials also showed, surprisingly, that the LFTO rarely causes significant damage. highest levels of shock and vibration Measurement. The availability of qual- occurred while circulating off-bottom, ity downhole dynamic measurements Modeling. LFTO and HFTO result from waiting for a wellbore survey, when the has increased rapidly over the past the excitation of structural resonanc- drillstring was not rotating (Fig. 2). decade (SPE 170586). Some sensors are es in the drillstring and BHA, respec- Novel torque-limiting downhole tools integrated with rotary-steerable sys- tively. Industry-standard critical speed can be very effective at combating the tems (RSSs) and logging-while-drilling analysis—the method of calculating the severity of bit-induced FSS and HFTO. (LWD) tools. Others are available as RPM at which these resonances occur— Weatherford’s latest torque-limiting standalone devices that can be installed requires the user to tune the mod- tool is a purely mechanical device that throughout the drillstring, even in the els. This method has significant lim- is placed above the BHA, which extends bit itself, and connected by fiber-optic itations and relies on the availability and contracts in response to drill-bit cables to surface (SPE 170770). of a significant amount of offset data. torque. The extension and contraction Transparency and standardization of Without these data, predictions are moderates bit engagement with the rock measurement is very important, espe- almost worthless. face, contracting as the drill bit begins cially with the drive to feed these data To remedy this problem, there has to torque up and preventing a full stick, directly into drilling automation sys- been a recent focus on time-domain then extending during the slip (or high- tems (SPE 174874). One invaluable simulations. These simulate the com- RPM) phase to allow the bit to remain real-time metric for distinguishing FSS plex interaction between the wellbore, engaged as the rate of rock removal from LFTO is the stick/slip index (SSI) stabilizers, under-reamers, drillstring increases. These tools, like PDMs, also (SPE163428): SSI <1 indicates LFTO members and—especially important for act as low-pass filters, and can be used and SSI >1 indicates FSS. HFTO detec- the prediction of FSS and HFTO—the to isolate sensitive BHA components tion requires measurements from tan- drill bit and formation (SPE 178866 and from HFTO. gential and radial accelerometers. When SPE 178874). A technology that has seen a big the ratio of tangential to radial accelera- resurgence is intelligent topdrive sys- tion is greater than 1, HFTO exists. Mitigation. FSS can respond well to tems (TDSs). TDSs provide torque to Post-run damage assessment is surface-based parameter optimization, rotate the drillstring. Intelligent TDSs also important for understanding and namely a reduction in weight on bit and use sophisticated control algorithms to identifying torsional dynamics. Over- an increase in RPM. LFTO and HFTO do reduce torsional fluctuations and are torqued connections, connections that not. In these cases, changes to the BHA extremely effective at combating LFTO,

22 JPT • OCTOBER 2016 if tuned correctly (SPE170925). New bit SPE 163428 Torsional Resonance, technologies also help to reduce tor- An Understanding Based on sional dynamics. Drill-bit design plays Field and Laboratory Tests with a significant role in FSS and HFTO, Latest Generation Point-the- especially the bit aggressiveness. One Bit Rotary Steerable System We give recent innovation is PDC bits that self- by L.A. Lines, D.R.H. Stroud, adjust bit aggressiveness while drilling and V.A. Coveney. you the (SPE 178815). SPE 167968 High frequency Torsional Dynamics of Drilling Systems: superpowers Monitoring. To reduce the demands An Analysis of the Bit-System on drillers, there is a movement toward Interaction by J. Jain, H. Oueslati, automation techniques that moni- and A. Hohl et al. you’ve tor parameters and provide warnings SPE 170586 Advanced Drilling and instruction to the drilling team Dynamics Sensor Allows Real- always (SPE 178773). The ultimate goal is to Time Drilling Optimization, close the feedback loop using rig sys- Damage Prevention, and Condition dreamed of. tems that automatically optimize drill- Monitoring of RSS and LWD BHAs ing parameters through the control of by L.A. Lines, C.L. Mauldin, and Introducing the world’s the drawworks and topdrive. J.W. Hill et al. first X-Ray technology SPE 170770 Field Premiere of Along- Conclusions String Dynamic Measurements for for oil wells. Drilling dynamics and optimization Automated Drilling Optimization VISURAY’s revolutionary VR90 ® have received a growing share of atten- Using Downhole Information not only finds downhole blockages tion in the past few years, particularly by D.M. Veeningen, A.P. Pink, faster, it lets you see 2D and 3D with the advances in drilling automation and M. McRay et al. reconstructions of the obstruction. technology and the move into deeper, SPE 170925 Surface Control Software We’ll illuminate the problem, you’ll more complex formations. The ability to Dramatically Mitigates Downhole detect, categorize, and mitigate torsion- Torsional Vibration in the Eagle eliminate the problem. Better yet, al dynamics downhole is critical if the Ford Shale Play as Validated by you’ll eliminate downtime and driller is to reach target depth quickly, High-Speed Downhole Dynamics increase profitability. safely, and cost-effectively. JPT Data by F. Efteland, A. Coit, and K.I. Mckenna. Contact us for a References SPE 174874 A Framework for demonstration SPE 21945 Detection and Monitoring Transparency in Drilling Mechanics visuray.com of the Slip-Stick Motion: Field and Dynamics Measurements Experiments by M-P. Dufeyte and by J.D. Macpherson, P. Pastusek, Visit us at Ofshore Energy H. Henneuse. and M. Behounek et al. in Amsterdam at stand 2.055. SPE 49204 Torsional Resonance SPE 178773 Vibration Monitoring of Drill Collars with PDC Bits in and Mitigation—An Integrated Hard Rock by T.M. Warren and Measurement System J.H. Oster. by J.A. Greenwood. SPE 98150 Risk-Based Reliability SPE 178815 A Step Change in Drill Engineering Enables Improved Bit Technology with Self-Adjusting Rotary-Steerable-System PDC Bits by J.R. Jain, G. Ricks, and Performance and Defines New B. Baxter et al. Industry Performance Metric SPE 178866 Integration of Time- by P.A. Wand, M. Bible, and Based Dynamics Simulation with I. Silvester. Milling and Underreaming System SPE 127413 MWD Failure Rates Due to to Implement Reliable Single Trip Drilling Dynamics by H. Reckmann, Solution in Plug and Abandonment VISURAY P. Jogi, and F. Kpetehoto et al. Operation by Y. Wang, M. Vos, and X-RAY VISION SPE 151133 Eliminating Stick-Slip A. Karimpour et al. by Managing Bit Depth of Cut and SPE 178874 Prediction and Mitigation Minimizing Variable Torque in the of Torsional Vibration in Drilling Drillstring by J. Davis, G.F. Smyth, System by A. Hohl, M. Tergeist, and N. Bolivar et al. and H. Oueslati et al.

JPT • OCTOBER 2016 Shift Toward Permian Drives Rising Rig Count Joel Parshall, JPT Features Editor

A drilling rig in the sunset works Apache’s Alpine High field in the southern Delaware Basin of west Texas, where the company recently made a major discovery. Source: Apache

on’t look now, but the United The number of active oil rigs nation- and STACK (Sooner Trend, Anadarko DStates rig count has inched up in wide rose to 407 for the week end- [Basin], Canadian [County], and King- recent months, and the driver has been ing on 2 September, according to Baker fisher [County]) plays. the old reliable of onshore oil produc- Hughes, up from 316 for the week of On the US natural gas side, the 88 rigs tion, the Permian Basin of west Texas 27 May. Of the additional 91 rigs, 65 operating at the beginning of Septem- and New Mexico. were activated in the Permian Basin. ber reflected little change since spring. While some observers might see the Beginning from a low point of 132 rigs However, a shift in deployment had scenario as a simple case of drillers operating in the week of 29 April, the added six rigs to the Marcellus since responding to the recovery of oil prices basin has added 70 rigs, and it current- 5 August and four rigs to the Utica since over the spring and summer, the fact that ly has more than 50% of the horizon- 20 May. The total number of rigs oper- rig growth occurred chiefly in the Perm- tal oil rigs operating in the US, up from ating in the two basins were 27 and ian indicates there is more to the story. 15% in 2011. 14, respectively. And along with a similar, smaller shift of A smaller growth trend has emerged gas rig activity toward the Marcellus and in Oklahoma, where 10 rigs were added High-Grading Portfolios Utica basins in Pennsylvania and Ohio, between 13 May and 2 September to “Companies are high-grading with- the statistics suggest what the geograph- bring the statewide total to 66. The in their portfolios, which leads to an ic footprint of an eventual, long-term increase has centered on the SCOOP overall high-grading within the domes- recovery may look like in the US. (South Central Oklahoma Oil Province) tic system in the Lower 48,” said Reed

24 JPT • OCTOBER 2016 Olmstead, director of commercial plays 1.6 Permian Basin Tight 160 and basins at IHS Markit. “So that’s why Oil Production we see operators dropping rigs out of 1.4 140 the worst parts of their acreage, and that leads to a natural contraction of where 1.2 120 activity is happening. Operators are looking to focus their money on areas 1.0 100 that can generate economic returns at depressed oil prices, and the Permian 0.8 Oil Price 80 keeps rising to the top when you look at 0.6 60 cost and performance.” The attractions of the Permian Basin 0.4 40 are high-quality rock with signifi- Daily Oil Production (Million BOPD) cant stacked play opportunities, well- 0.2 20 developed infrastructure in many areas, Horizontal Drilling Begins West Texas Intermediate Oil Price (USD/bbl) and the fact that the industry has a – 0 wealth of experience in the basin. '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 '10 '11 '12 '13 '14 '15 '16 The most active driller is Pioneer Nat- Countering the trend of falling crude oil prices, Permian Basin tight oil ural Resources, which is adding a 17th production has continued to rise strongly. Source: Pioneer Natural Resources rig to its Permian operations. Others (US Energy Information Administration data). with high activity levels include Concho Resources, ExxonMobil’s XTO unit, and Parsley Energy, while Anadarko, Chev- Long-Term Production Growth leum for USD 2.5 billion. The addition- ron, Cimarex Energy, Apache, EOG, and Scott Sheffield, chairman and chief al Permian acreage would complement Occidental are among the companies executive officer (CEO) of Pioneer, told EOG’s current acreage there as the com- with substantial current activity and/or a plenary session audience at the SPE- pany shifts emphasis away from the high production. cosponsored Unconventional Resources more expensive Eagle Ford Shale. With Technology Conference in August that the acquisition, EOG Chairman and CEO Permian Resurgence the Permian will drive long-term US oil Bill Thomas said, “We’ll be able to grow For a basin that has produced oil since production growth. He forecast that the oil [production] with less capital and 1920, the Permian has seen a remark- basin’s production will increase by an more efficiently than we do now.” able resurgence. In its core Central average of 300,000 B/D annually over The news followed Concho’s mid- Basin platform (CBP), the location of the next 10 years and reach 5 million August announcement that it would buy most of its conventional fields, produc- B/D by 2025. 40,000 core Midland Basin acres from tion in 2007 stood at a nearly 6-decade The Permian Basin is now in direct Reliance Energy for USD 1.6 billion. low of less than 800,000 B/D. Since competition with Saudi Arabia, Shef- then, production has grown to 2 mil- field said, noting the basin has produced Land Deals lion B/D, a level not seen since the early 35 billion BOE over its commercial life Recent land deals have transacted at 1970s, with a shift toward horizon- and has an additional 150 billion BOE per-acre prices of USD 58,000 in the tal drilling since 2011 driving most of of recoverable reserves that can be pro- Midland Basin and USD 27,000 in the the increase. duced as the Midland and Delaware more remote Delaware Basin, although In the Midland Basin, adjacent to basin plays expand to their potential. one independent reportedly paid the CBP, the Spraberry and Wolfcamp Sheffield said that Pioneer expects USD 150,000 per acre to acquire some plays have generated an increase of to achieve a 30% annual production Delaware Basin holdings. approximately 850,000 BOE/D since growth in the Permian and that the Acreage swaps have also become com- 2009, according to data from IHS. Ini- company’s Midland Basin horizon- monplace, as land is a currency available tial growth primarily reflected vertical tal drilling prospects can break even regardless of commodity prices. well activity, but horizontal drilling has at an average price of USD 25/bbl— Bruce Palfreyman, general manager taken over in the last 4 years. which he noted was the price needed to of the Permian asset at Shell, discussed The surging production has made the earn a 10% return on the present value swaps in the Delaware Basin at a recent Permian the only major US oil shale play of investment. presentation on the company’s uncon- to experience growth since crude prices The opportunities present in the ventional resources business. “There began to fall, with US Energy Informa- Permian have spurred a rash of acqui- is going to be continued consolidation tion Administration data showing an sitions and land deals, including EOG’s across the basin in terms of people put- almost 20% rise in basin output since September announcement that it would ting acreage together,” he said. “Every- the start of last year. purchase privately held Yates Petro- body knows the value enhancement of

JPT • OCTOBER 2016 25 long laterals, so you’re going to see lots SCOOP and STACK Plays Similar to Continental’s refocus from more trading and swaps as we go for- Interest has also been building in Okla- the Bakken to the SCOOP play, South- ward from smaller to larger units, where homa. Continental Resources, a leading western Energy curtailed gas drilling they can lay out their development pro- operator and the largest acreage holder in the Fayetteville Shale to concentrate grams with as long a lateral as they in the Bakken Shale of North Dakota and “on the better part of their portfolio, can execute.” Montana, has shifted significant atten- the Marcellus acreage,” Olmstead said. Shell holds a net 300,000 acres in the tion to the SCOOP play, which repre- The best near-term opportunities, he Delaware Basin through a 50/50 joint sented 29% of company production at believes, lie in the northeastern Penn- venture to develop an area of mutual the end of last year. sylvania dry gas window. interest with Anadarko. Assets include “When you look at their presenta- “We’ve seen some very strong wells 400 Shell-operated wells and more than tions and have conversations with them, out of the Utica,” Olmstead continued. 5,000 possible well locations. “In the they’re much more focused on Oklaho- “EQT and a couple of other companies current environment, we’re ready to ma than they had been, say, 2 or 3 years have had some amazing results. The move a few things into development, ago,” said IHS’ Olmstead. “When you question is really just the repeatability.” and that’s going to be our focus going look at where their growth story is and forward,” Palfreyman said. where they feel their best returns are, The Shape of Things To Come Overall, Delaware Basin operators it’s the SCOOP.” Current trends suggest that the geo- have been shifting from delineation to Continental had significantly delin- graphic footprint of the US shale oil optimization, which is reflected in lon- eated various SCOOP discoveries before industry will be more centered on the ger laterals and higher proppant loading. oil prices fell and thus has been able to Permian Basin when drilling activity According to Jeanie Oudin, a senior move quickly into development drill- recovers than it was before oil prices research manager at Wood Mackenzie, ing. Marathon Oil and Newfield Explo- collapsed. The case is plausible, but “The Midland and Delaware basins hold ration are also major operators in the there are some caveats when projecting the largest number of undrilled, low- SCOOP. Some smaller companies active to a full recovery environment. cost tight oil locations in the Lower 48. in the play hope to be able to prove up Analyst Richard Zeits, who covers oil, No other region comes close.” acreage to sell to the larger companies, gas, and commodities on the website Olmstead said. Seeking Alpha, believes that Pioneer’s Apache’s Big Discovery In the STACK play, Devon Energy has projection for 2025 industry produc- One company that has been patient- a strong legacy position that it enhanced tion in the Permian is achievable and ly drilling the Delaware is Apache, early this year with the USD 1.9-billion possibly conservative. Nonetheless, he which announced a major discovery on acquisition of privately held Felix notes some hurdles to be scaled as a 7 September. Energy, which held key prospects in the recovery builds. The Alpine High discovery, primarily Anadarko Basin. ◗ Additions to processing, storage, in Reeves County, Texas, in the south- Similarly, Marathon is bolstering and takeaway capacity and ern part of the basin, holds an esti- its presence in the Anadarko with the particularly the time and expense mated resource in place of 75 Tcf of gas USD 888-million purchase of privately to develop processing capability for and 3 billion bbl of oil in the Barnett held PayRock Energy Holdings, which liquid-rich natural gas. and Woodford formations. Apache has was due to close by the end of the third ◗ The need to spread drilling to less- also confirmed oil-bearing potential in quarter. Newfield, Husky Ventures, productive acreage. the Pennsylvanian, Wolfcamp, and Bone Cimarex, and Chaparral Energy are ◗ Eventual cost inflation in the Springs formations, company CEO and other major STACK play operators. supply chain. President John J. Christmann IV said. At least some operators have turned Zeits emphasizes shale oil’s broadly He called Alpine High “a world-class toward the SCOOP and STACK plays, distributed US production base that he resource play.” Olmstead said, because “they’re at a believes can again come to the fore in The discovery followed a meticu- point in their portfolio where they need a full recovery. “The industry will like- lous 18-month process of assembling another asset—whether it’s because ly continue to rely on several prolific a position of 307,000 net contiguous they are running out of inventory in basins, most of which will contribute acres among 352,000 gross acres and other assets (or) they don’t like the eco- significantly to production growth in an drilling 19 wells, nine of which were nomics and think that despite the risks upcycle,” he said. on production when the discovery the economics here are worth it.” However, it appears safe to say that was announced. To accelerate delinea- while the Permian Basin did not lead tion and development at Alpine High, Marcellus and Utica the shale oil revolution, it will lead Apache boosted current capital spend- In natural gas activity, the Marcellus and the US shale oil industry into the next ing by USD 200 million. The new play Utica plays have lately attracted addi- recovery when it happens. And the represents more than 25% of the com- tional rigs. “We’re very bullish on those same can be said of the Marcellus for pany’s 2016 capital budget. plays,” Olmstead said. shale gas. JPT

26 JPT • OCTOBER 2016 TAM-189_Corporate_Ad_051016_outlines.indd 1 5/10/16 11:58 AM Devon Energy Rises to the Top as a Data-Driven Producer Trent Jacobs, JPT Senior Technology Writer Source: Getty

fter becoming one to work a small arsenal of sophisticated sors, cloud-based analytics services to of the first shale algorithms designed to find the other- crunch the data in seconds, and effective A wise hidden insights and engineering procedures on how to take action on that producers to adopt tweaks that are driving more value out of information—depends largely on solid advanced data analytics, the shale fields it operates. teamwork and sound business decisions. This realm of computer science has Part of this requires getting analyt- Devon Energy is now also made routine tasks that once took ics experts, information technology staff, one of its most ardent engineers hours to complete possible to and engineering leaders from different do in a few minutes, or less. And that departments to climb out of their respec- champions. The results some of these programs can literally tive silos so they can share data and speak to why that is. think for themselves, although in a limit- develop new solutions together. ed mathematical sense, is not hyperbole. On the economic justification, one sim- Today’s offerings in artificial intelli- ple way to assess the value of advanced Since 2012, when Devon began investing gence and machine learning technology analytics was put to Ball by a completions in big data technology, its 90-day pro- can be trained to look for defined out- engineer, who she quoted as saying, “If it duction rates from horizontal wells have comes, or they can be set free to decide doesn’t help us extract oil and gas from increased by 250%—an improvement on their own what to look for and will the ground, then so what?” that according to IHS Markit data has lift- rewrite their underlying algorithms in Ball has come to call this the “so what” ed the Oklahoma City-based company’s order to do so. test. She applies it to herself, the engi- early-time production from new wells to neers who come to her team wanting a the highest level in the US onshore mar- Humans Still in the Loop shiny new algorithm, and to the software ket. Additionally, Devon’s costs to drill, As intriguing as this emerging world of vendors touting their product as the lat- complete, and operate its well have fallen intelligent software is, when discussing est and greatest breakthrough in artifi- by as much as 40%, the company said. how it has changed her company, Kathy cial intelligence. It is fair to attribute some of these Ball, manager of advanced analytics and This approach filters out the ideas not significant gains to the quality of the data science at Devon, begins by stressing worth pursuing, while allowing engineers company’s positions inside North Amer- the importance of the human element. and data experts to focus resources on ica’s most prolific shale plays, and also “It is about culture and collaboration,” ones with the most potential to generate the pressure that oilfield contractors are she insisted, adding that those two ingre- revenue. So far, more than a dozen ana- under to make price concessions in a dients constitute the “secret sauce that lytics programs have made the grade and down market. makes this all work.” are now integrated into the company’s However, that does not complete the In other words, implementing a sophis- day-to-day operations. story. ticated data-driven approach to field They address several key aspects of From exploration and drilling to com- development—one that relies on real- shale development: pilot testing, well pletions and production, Devon has put time streaming from remote field sen- spacing, expected drilling time, geosteer-

28 JPT • OCTOBER 2016 600 ing, geohazard prediction, bottomhole assembly selection, real-time hydrau- 450 IMPROVEMENT lic fracturing risk analysis, and artificial IN DEVON WELLS lift optimization. (Since 2012) 300 Finding the Bit Geosteering is one area where Devon 150 believes analytics is clearly making a difference. This drilling technology has become a favorite among shale produc- 0 Top U.S. Producers ers because it gives them a degree of precision when tapping into relatively thin Based on its own data and that of IHS Markit, Devon Energy reports to have pay zones. achieved the best initial production rates in the US onshore. Source: Devon Energy The downside is that geosteering tools sit some distance behind the bit and sur- hydraulic fracturing that happens when “We are now able to do that through veys are only taken every 90 ft, leaving sand or proppant stops flowing into open a big data solution and come up with a more than 100 ft of unknown wellbore fractures and begins to plug up the well- calculation for the Lower 48 in probably trajectory each time drilling moves for- bore instead. Screenouts can be expen- 10 minutes,” she said. ward. To overcome this data gap, Devon sive to remediate and have thus encour- The mapping program includes data developed a real-time drill bit tracker aged the development of algorithms to from more than 100,000 horizontal that leans on artificial intelligence and detect their telltale signs. wells drilled by Devon and its competi- machine learning to work the math and The analysis behind Devon’s old warn- tors in the past decade. Tools such as this figure out a well’s true path. ing system was capable of assigning risk are allowing producers to peer over the Engineers at the company’s drilling levels between 1–4 to indicate the proba- lease line and compare their completion center keep a watchful eye on this drilling bility that a screenout was near. That sys- designs to offset operators. data as it is streamed in and visualized, tem has since evolved into a much more For Devon, the mapping program has enabling them to call for adjustments if deterministic model that factors in the identified refracturing candidates and is the bit is veering off the planned course. type of fracturing fluid being used, for- accelerating pilot programs that seek to “That is just priceless,” Ball said of mation characteristics, and field history. prove new techniques. This form of pre- the bit tracker. “It keeps us in a produc- Characterizing the role of artificial dictive analytics is also used to assess the ing zone, and helps us assess the land- intelligence in making models like this company’s entire fleet of artificial lift units. ing curve,” the critical juncture where the more accurate, Ball said, “It has gone The program determines the mean time to vertical well section gives way to the hor- from an engineer saying, ‘Here is the failure and flags lift units if the potentiality izontal section. On its website, the com- pattern that you need to look for,’ to ‘All for a rod or pump failure is detected. pany credits this program with reducing right, we have built that pattern, now let As a result, Devon’s field maintenance the number of unplanned sidetracks it the machine make it even smarter.’” strategy has transformed from one based has had to drill to get wells back on their In this case, the smarter model does on a cyclical schedule to one that is largely intended trajectory. away with the risk levels and instead sends prescriptive. Instead of field technicians Similar analytic programs are used to engineers a 15-second warning only when inspecting a set list of wells in sequential model which bottomhole assembly kits a screenout is imminent. The next step is order, the analytics program now sends will perform best in a particular well and to develop a 30-second warning: plenty of them two reports a day that target only will chart their actual performance as time to tone down the pumping while also those units truly in need of attention. they rotate down the hole. Another one ensuring that maximum pressures inside “It is really important to go beyond predicts which sections of a well are most the wellbore are maintained for as long as just preventative,” Ball explained. “That likely to give drillers trouble (i.e., non- possible. Ideally, that would lead to a bet- will tell you that I need to replace some- productive time) and which parameters ter stimulation treatment. thing because it might fail—the analyt- are contributing most to the rate of pen- ics is going to tell you when it is going etration (e.g., weight-on-bit). Taken as Production and Reservoir to fail.” She added that this program has a whole, these drilling-focused analytics Management turned out so well that it now represents programs are preventing costly errors One example of how analytics is giving one of the company’s largest areas of and helping rig crews deliver better wells valuable time back to engineers is an cost savings. in less time. interactive map that Devon uses to calcu- In terms of uplift, the company has late how much oil or gas is left in any given noted that this combination of analyt- Screening Out Screenouts shale field in the US. Ball explained that ics and automation has netted single- Devon is also using real-time analysis to an engineer tasked with doing this for 50 digit percentage increases in production avoid screenouts—a common problem in wells might have needed a week to do so. across its portfolio.

JPT • OCTOBER 2016 29 Four Answers to the Question: What Can I Learn From Analytics?

Stephen Rassenfoss, JPT Emerging Technology Senior Editor

he future of data- ◗ What are the changes in a fracturing For example, one of the programs used Tdriven analysis design that will offer the biggest to analyze fracturing used facial recog- production payoff? nition to classify pressure changes dur- in exploration and ◗ Why has the drilling slowdown ing each stage to sort them into differ- not depressed production from ent classes. production (E&P) will unconventional gas plays? Production data revealed that one of depend on whether it ◗ What is the half-life of my field, and the classes of fractures, characterized why should I care? by a pressure bump near the end of the can add value in the field. The stories below bring together stage that could mean trouble, were often advanced statistical analysis methods more productive than those that went with multiple names: analytics, big data, exactly according to plan. But it was up to Four examples of what is possible were machine learning, and even a “physio sta- the completion team to figure out how to presented recently at the 2016 Uncon- tistical engine for automatic stochastic apply that observation. ventional Resources Technology Con- production forecasting.” The process is “brutally empirical,” said ference (URTEC) in San Antonio with All this new E&P math is aimed at iden- Roger Anderson, president of AKW Analyt- the authors of papers posing questions tifying patterns and relationships that ics, who led the study for Range Resourc- such as: otherwise would be missed. But auto- es. He said the method is good at identify- ◗ Does it matter if a lateral is drilled matic and self-learning does not mean ing what is happening, but it is unable to toe-up or toe-down? all-knowing. determine why or how it happened.

Toe-Up or Toe-Down: Does It Really Matter?

For years, an unresolved question for tive rock, which can mean large chang- its response based on the performance of those drilling horizontal wells has been: es in elevation from the heel—the curve more than 300 similar wells drilled in the Does it matter if it is toe-up or toe-down? from vertical to horizontal—to the toe. Cana-Woodford Shale in Oklahoma. Horizontal wells generally follow an up Results from modeling have not settled The comparison was possible because or down slope following the most produc- the matter. Recently, Devon Energy offered the company had mass-produced simi-

Toe-Up Toe-Down

Example of the difference between a toe-up and a toe-down well. Source: URTEC 2461175

30 JPT • OCTOBER 2016 lar wells, giving it a large sample of wells The results were compared based on in the future, which will address ques- with 4,800-ft laterals where it fractured the depth of the wellbore—shallow, tions about whether changes that come 10 stages with 40 perforation clusters middle, and deep—because of differ- with age, such as increased water pro- using 3.5 million lb of proppant located ences in the zones. For example, the duction requiring pumping, will alter in a compact area with similar geology in condensate-rich production in the shal- the conclusion. the three depths studied. lower zone is more likely to be affect- Devon is using this study when plan- The analysis was done in a way that ed by liquid holdup than the gas-prone ning new wells in other areas with simi- ensured, as much as is statistically possi- deep zone. lar conditions, such as projects to drill ble, that “only toe-up or toe-down could The widest variation between toe-up infill wells in spots with relatively low be affecting the production performance and toe-down wells was seen in the mid- pressure to overcome liquid buildup. of the wells analyzed,” said Sam Brown- dle zone where the elevation changes The location of the well, though, may ing, a reservoir engineering for Devon were greatest. In those toe-down wells, require drilling a well toe-down to stay who delivered the paper at URTEC. the elevation changes were nearly all in within the most productive rock. It concluded that toe-up wells pro- the 100–200 ft range, while the dips at And the sample size in this case is too duce more. And the greater the elevation other depths were less than 100 ft. small to generalize about all horizon- change between the heel and the toe, the The paper predicted that an extreme tal wells. But based on the Devon study, greater the impact. toe-down well in the lowest-pressure the potential reward for knowing if toe- “The more toe-up they are, they [wells] area would produce 30% less over up is better or worse can be big enough appear to be better, and the more toe- its life. to justify the cost of a study to answer down is worse,” Browning said, adding The paper suggested toe-up wells the question. “All the best wells were toe-up.” may perform better by offering a grav- Devon found that toe-down wells pro- ity boost for the liquids-rich flow, and URTEC 2461175 Effects of Toe-Up versus duced 25% less based on a year’s worth of toe-down can allow liquid to build up in Toe-Down Wellbore Trajectories production, he said. The difference was the production zone, hindering the flow. on Production Performance in the narrow in the early days of production Browning is planning to come back Cana Woodford by S. Browning and and widened over time. to these wells to see how they are doing R. Jayakumar, Devon Energy.

Is the Struggle Really Worth It?

When it comes to pumping a proppant But the production from the fractures per foot nearly tripled, with coarse sand into a fracture, a job that goes according that went smoothly was often not as used and more stages in longer laterals to plan may not be the best option. good as those that required a struggle. drilled deeper in the formation. A statistical analysis of fractur- “Perfect means you will not do as well The paper points out that “determining using machine learning found that if you do not struggle to get in that ing which (change) led to the great- the smoothest fracturing stages often last sand,” said Roger Anderson, presi- est improvement is difficult because” underperformed those where the oper- dent of AKW Analytics, who delivered there were so many big ones. The bot- ators struggled to get all the sand into the paper on a 4-year study for Range. tom line, though, was positive, with the the formation. “Perfect may mean the sand (volume) new approach yielding a “remarkable The study presented by AKW Analyt- was short.” increase in condensate and gas produc- ics at URTEC used a system comprising The difference among fracture classes, tion per well in 2013,” the paper said. multiple software programs to seek out which was only noticed in 69 of the wells, And Anderson pointed out the changes telling patterns in several streams of data was one of many variables studied to see were generally in line with the study’s from 1,800 fracture stages in 156 wells which factors have the biggest impact conclusions about which factors mat- completed by Range Resources in the on production. ter most. Marcellus Shale. The goal was to learn how to pre- The most influential ones were not sur- The study focused on two groups of dict future performance when comple- prising—the percentage of the wellbore classes of fractures: ones where the pres- tion teams are considering which vari- in the targeted reservoir rock, the total sure rose before it completed pumping ables will have the most effect during the volume of proppant placed, and the num- the planned amount of proppant, rais- design stage. ber of stages were among them—but it ing concerns that an excessive amount of The multivariable analysis showed an also quantified the potential impact and sand might block the fracture by causing oil field is no place for a statistician look- predicted how different combinations of a screenout, and another class where the ing for clean, consistent data sets over attributes will perform. pressure hardly changed because “every- time. Prior to the last year of the study, By using 34 attributes, for example, thing went as scheduled,” according to Range Resources revamped its comple- Anderson said the model was 78% accu- the paper. tion method. The pounds of sand placed rate, with a wide error range. With more

JPT • OCTOBER 2016 31 350

300

250

200

150

100 % Change in 2013 Wells % Change in 2013

Stage Length 0 30/50 Mesh 30/50 Mesh # Stages Total Sand Deeper Landing Lateral Length lb/ft per Stage per Well per ft Zone –50

This quantifies changes in the completion methods used by Range Resources, including a reduction in the space between stages, to produce significantly more gas in the Marcellus. Source: URTEC 2430481

data to consider and time to learn, the URTEC 2430481 Using Machine Learning to URTEC 2426612 Petroleum Analytics system could improve its accuracy. “It Identify the Highest Wet Gas Producing Learning Machine To Forecast Production is not something you would want to Mix of Hydraulic Fracturing Classes in the West Gas Marcellus Shale by report to the SEC; it is not good enough and Technology Improvements in the R. Anderson, B. Xie, L. Wu, AKW Analytics yet,” he said, adding, “but it is better Marcellus Shale by R. Anderson, B. Xie, et al. than guessing.” L. Wu, AKW Analytics et al.

Why Does Drilling Drop, Not Kill, Gas Production?

Shale wells are known for their rapid pro- 3,000 wells producing 1 Bcf/D of gas. increases the EUR by about 20% over a duction declines. So when drilling activ- The results showed that the impact of 3-year period, a simulation concluded the ity dropped in 2008 and again in 2014, the rapid early decline rate of new wells increased productivity would be possible steep declines were widely predicted. can be muted by the slow, steady produc- to maintain the production plateau for But that is not how it has played tion from thousands of older wells drilled 10 years with only two rigs. out in the three biggest US shale gas over many years. The simulator was also used to pre- plays—the Barnett, Haynesville, and While the paper said it takes a fleet of dict what would happen in the three big- Marcellus—where production has held drilling rigs to reach that peak—16 rigs gest US gas shale plays if drilling stopped. up even though only a few rigs have working for years—once at that peak It found that the Barnett was the most been running. level, only four are required to main- resilient based on the expected rate of The reason why the expectation and tain that level for many years. In prac- decline, the Haynesville the least, and the the realization have been different is tice, operators are likely to add rigs when Marcellus in between. due to the industry’s increasing pro- gas prices are high enough to make it an “The Barnett is very resilient. There is ductivity and the fact that thousands attractive investment. “When the price is a large well portfolio,” drilled over many of unconventional wells perform differ- relatively high, I drill a lot to feed my well years, Charlez said. ently than a single well, according to portfolio. When prices are low, I live on If drilling stopped for 10 years, the Philippe Charlez, a senior technical advi- existing wells,” Charlez said. Barnett production would be down sor to Total, who delivered a paper on his There is a significant reward for pro- 50% while the Haynesville would have research on resilience at URTEC. ductivity gains increasing the estimated dropped to zero by 2025, according to Total’s Unconventional Factory Devel- ultimate recovery (EUR). “If it is high- the study that said Marcellus would drop opment Simulator was used to study the er, I need less rigs to maintain the pla- nearly 75% if there was no drilling dur- decline rate of a large shale play with teau rate,” he said. And if an operator ing that period.

32 JPT • OCTOBER 2016 Fully automated closed-loop drilling

The future of well construction built on a history of success. NOV Automation oers simple steps to unleash the full potential of your drilling system. We’ve merged downhole and surface drilling environments with rig control systems to deliver consistent and impactful results across your fleet.

Join the drilling revolution at nov.com/automation.

© 2016 National Oilwell Varco | All Rights Reserved 27 400 As the oldest shale play, the Barnett has the edge because about 15,000 wells have been drilled there vs. 3,700 in the Haynesville, and the Barnett has more, older wells whose steady output reduces 23 300 the impact of the sharp declines seen in new wells. The net effect of a mix of prolific fast-declining young wells and low-

19 200 Rig Count producing steady old wells is roughly analogous to a diversified investment Production (Bcf/D) Production portfolio with volatile high-performing stocks balanced by the modest, steady cash flow from bonds. 15 100 April 2011 April 2012 April 2013 April 2014 URTEC 2439429 Resilience of the US Shale While the number of rigs drilling in the Barnett, Haynesville, and Marcellus Production to the Collapse of Oil & Gas formations dropped sharply (blue line), production has risen in the three Prices by P. Charlez, Total, and P. Delfiner, largest US gas-producing shale formations (red line). Source: URTEC 2439429 PetroDecisions.

What is the Half-Life of That Shale Play?

Half-life is a commonly used way to substantial production from uncon- To make her case, Heidi Kuzma, measure the life of radioactive materials, ventional oil fields, because it can be founder and chief technical officer of but not oil fields. used to estimate when a company will BetaZi, applied its statistical analy- The founder of a data-driven forecast- need to spend money to sustain pro- sis method to forecast production from ing company, though, argues that it is a duction at a level that satisfies invest- more than 13,000 wells in North Dakota useful benchmark for companies with or expectations. and concluded it would take 3 years and 4 months for the output of all those wells to decline 50%. Monthly Volumes 2000–2025 Peak June 2015 “Somehow, if production is to be brought up to its peak 2015 levels, half of 40 13,000 that production must be replaced by the 37 12,150 end of 2018. Worst case scenario, that 35 11,300 means another 6,000 or 7,000 wells,” 32 10,400 29 9,550 need to be drilled and completed, said 27 8,700 Kuzma, a geophysicist by training. “That May 2012 October 2018 is a huge amount of capital.” 24 Half peak 7,800 21 6,940 The actual statewide decline rate will 19 6,100 not be that fast, because while the calculation assumed no drilling, there were 16 5,200 Count Well 13 4,300 27 drilling rigs working in North Dako- 11 3,500 ta in late August, according to the Baker Hughes rig count. Those wells are for BBL/month boe p50 (million) 8 2,600 5 1,700 companies focused on speeding the con- 3 900 struction of significantly more produc- 0 0 tive wells. For individual operators, however, the half-life of their assets is shorter. The Feb Feb Feb Feb Aug Aug Aug Aug pressure is greatest for those who drilled 2000 May 2001 2002 2002 Nov 2003 2004 May 2005 2005 Nov 2006 2007 May 2008 2008 Nov 2009 May 2010 Feb 2011 Nov 2011 Aug 2012 May 2013 Feb 2014 Nov 2014 Aug 2015 May 2016 Feb 2017 Nov 2017 Aug 2018 May 2019 2020 2020 Nov most of their wells at the tail end of the boom, which ended during the second If all drilling stopped when oil production peaked in the Bakken, it would drop to half that level in 3 years and 4 months, according to a study half of 2014. based on a statistical analysis of the production reports from 14,000 wells. “There are already a significant num- Source: URTEC 2461672 ber of operators whose Bakken produc-

34 JPT • OCTOBER 2016 tion is very close to half of what it was that are difficult to test. They constantly dicted future monthly production rates. at its peak in 2015,” she said. Many of change,” Kuzma said. And both said their predictions were in those cannot afford to drill and complete Production decline rates have always line with actual production results. wells. “In this case, depletion might be a been tracked, but are not featured in Charlez concluded that if “no new function not of the asset, but of the capi- financial reporting. Predicting long- development activity occurred, it would tal needed to exploit it,” Kuzma said. term decline rates from unconventional take 10 years for Bakken production to BetaZi’s statistical approach estimated wells is a challenge. The proper meth- drop 50%,” which is three times lon- the decline rate by inputting monthly oil, od for predicting the long-term decline ger than Kuzma’s prediction. His esti- gas, and water production figures report- rate for unconventional wells is not set- mate was mentioned at the end of a ed to the state of North Dakota for each tled among petroleum engineers. And study focused on major gas plays, where well, and also the number of days the the skills and methods of those drill- he said the longer histories allow more well was flowing. The company’s predic- ing and completing wells varies widely dependable statistical analysis. tions were created using its proprietary among operators, which also have vary- Given the many variables, Kuzma said “physio statistical engine for automatic ing approaches, and budgets, for main- it is hard to prove who is right. “For the stochastic production forecasting.” taining production. moment, any purely data-driven fore- Estimates of reserves and ultimate Based on the conclusions of Kuzma’s cast that claims to be accurate beyond 6 recoveries do not figure into the calcu- paper, and one presented by Phillippe or 7 years for unconventional production lation. In shale, it appears the amount Charlez, a senior technical advisor for is a bit suspect, since we just don’t have of oil produced is a small percentage of Total, long-term production predictions that much experience,” she said, adding, what is in the ground. Actual recoveries are tricky. “let’s get back together in 10 years.” will depend on the future price of oil, and Both estimated production on a well whether improved techniques are devel- by well basis and combined it using their URTEC 2461672 Blinking Out: North Dakota oped and applied. company’s proprietary software engine. Without Capital for Replacing Production “Reserves and EUR [estimated ulti- He drew his information from decline by H. Kuzma, C. Eklund, T. Rapp et al., mate recovery] are estimated numbers curve analysis, while she statistically pre- BetaZi.

ENVENTURE. PUNCHING THROUGH.

When you’re pushing the limits of exploration and development, you need extreme technology. Enventure’s proven ESET® solid expandable liner technology is engineered for toughness and reliability to help you reach Total Depth while minimizing NPT. Here’s how: s Rotate across ledges and through tortuous wellbores s Rotate and reciprocate to improve cementation of expanded liner Add more punch to your process with ESET® technology from Enventure. To ﬁnd out more, visit us at: www.EnventureGT.com/ESET Analytics Firms Explore Oil and Gas Market Trent Jacobs, JPT Senior Technology Writer

he oil and gas Some are selling analytics as a way up from Silicon Valley called Maana. The Tindustry is facing an to combat the growing threat of cyber list of financiers includes the venture attacks. Other programs interpret human arms of Saudi Aramco, Chevron, Shell, invasion of data analytics semantics to extract valuable informa- and GE. tion out of the entirety of a company’s This may be a leading indicator of startups who saw a wide- document library. things to come because rather than open gap in the market a Below is a closer look at what several focusing on any particular niche solu- of these analytics startups are offering tion, Maana is a big-picture analyt- few years ago when talk to the industry—only there will be no ics platform. Able to reach between dif- of big data first began. deep dive into the layer cake of terminol- ferent silos within an organization, the ogy that dominates this emerging soft- program analyzes seemingly disparate ware arena, e.g., artificial intelligence, sets of operational data and ties them Many of these young companies vying machine learning, edge computing, etc. all together. for attention from producers are focused Donald Thompson, cofounder and on alleviating the headaches associ- Power of Integration president of Maana, said such data inte- ated with artificial lift systems. There A sign of the industry’s accelerating gration is able to generate “new knowl- are also a number of software products uptake of analytics came in May when edge” that oil and gas producers can designed to handle difficult computa- a number of the world’s largest produc- leverage to drive performance in the field. tions such as production forecasting and ers made a USD-26 million investment “Most projects involve going and tack- reserve estimation. in an industrial-focused analytics start- ling a single isolated problem, but everything in a business tends to be highly interrelated,” he explained. “So the failure of an individual pump is very interesting to understand, but its impact on production schedules is also interesting … if one pump fails, how should you adjust all the other pumps in order to still meet your production requirements?” Thompson, who pioneered knowledge- and semantic-based search engine tools during his 15 years at Microsoft, gave another example of an oil and gas project involving malfunctioning field sensors. The sensors, managed by a remote monitoring team, were sending faulty telemetry alerts to the operator’s support center. Unbeknownst to the latter group, the sensors had received a software update which was responsible for the glitch. By being in between the monitoring and support team, Maana connected the dots and determined the root cause. A worker installs a small sensor on a pump jack which turns the machine into an internet-of-things device that can be tracked, monitored, and controlled “Individual analytics projects with- remotely or set to optimize itself autonomously. Source: Ambyint in each one of those groups,” Thomp-

36 JPT • OCTOBER 2016 son said, “would never have detected a correlation between the two groups and those are the type of things we’re enabling now.” Pumping Up the Data One of the biggest targets for analytics vendors is artificial lift because rods, pumps, and motors are relatively easy for the latest generation of smart software to interpret and predict. Calgary-based Ambyint is one of those firms. The company feels it has an advantage over its peers because it is leveraging a decade of Canadian oilfield pump data gathered by its holding company and predecessor, PumpWell. Ambyint was formed to market an improved predictive maintenance program that analyzes the symptoms that precede a pump failure. Such programs alert producers to the precursors of a pump failure days before a total failure occurs. This allows field maintenance A random forest modeling program allows users to quickly run queries programs to be scheduled with precision over large swaths of producing areas to find new trends or well completion and efficiency. strategies. Source: Ruth.ai The firm’s software can also automate pumping operations based on produc- into an autopilot where it will start mak- is an internal search engine that mim- tion targets or to stretch out the mean ing decisions for you,” Dhunay said. ics many of the features used by Google time to failure. Though clearly logical, or Amazon. this pairing of analytics and automation Reading Between the Lines The company is currently fine-tuning is in its infancy. Anyone with a smartphone has by now its language recognition platform in a Nav Dhunay, president and chief exec- probably tried speaking to it to get direc- partnership with the SPE, the gatekeep- utive officer of Ambyint, said producers tions, send a text, or run an internet er to hundreds of thousands of techni- who buy into predictive analytics must search. Led by former oil and gas profes- cal papers that date back to the 1800s. first become confident of its core fea- sionals who specialized in computer sci- Smith said: “It’s the best available con- tures before taking that next leap. ence, i2k Connect is trying to bring this tent that could be imagined to train “This industry is a bit conservative natural language recognition technology our system.” when it comes to technology adoption, to oil and gas companies. The goal is to create an enhanced so the last thing somebody wants to do is But instead of spoken seman- search capability across all SPE online trust that a computer can make a better tics, this company is teaching its soft- content including the technical paper decision than they can—even though if ware to understand written seman- database, OnePetro. Content and papers you look at the majority of accidents that tics. Moreover, it is trying to teach it will be categorized based on the i2k happen worldwide, they happen because oil and gas jargon using 15 different program’s understanding of their of the human element,” he said. industry taxonomies. true context as opposed to only rely- Ultimately, he added, software will “Companies run on data, and much of ing on keywords. Smith described it as win the day as it becomes better than that data is in the text documents they “enriching the documents with subject humans at managing oil and gas wells create and store every day,” said Reid matter expertise.” running on pumps This transition will Smith, one of i2k’s founders. “Unfortu- That means, for example, techni- begin through a semiautomated sys- nately, those documents are hard to find.” cal papers from related topics will be tem Ambyint calls a “recommendation To solve the problem, this startup’s tagged more accurately. A paper relat- engine,” which requests that prescriptive intelligent software will pore over a ing to production logging but that does actions be taken. company’s unstructured text files and not use that exact term within the doc- As production engineers accept more add structure by autotagging each PDF ument will still be tagged “production and more of those recommendations, file, email, slideshow, or Word docu- logging,” making it more visible in One- “what we’re going to do is turn this thing ment with a keyword. The end result Petro searches on this topic.

JPT • OCTOBER 2016 37 Real-Time Processing same thing now for big, fast data, except He added that tools like this also allow In August, real-time analytics firm SQL- I think this is even easier to use than operators to find answers to important stream received a major endorsement a spreadsheet.” questions. Why are some wells doing of its platform when Amazon adopted worse than others? Is it a completion it to become the engine behind a new Disruptive Pricing problem, or a problem with the reser- web-based service for businesses. The Troy Ruths founded his namesake start- voir? Are there any clear trends that San Francisco-based company is hoping up Ruths.ai in 2013 after earning his show why one part of the field is produc- to make a similar impression on oil and PhD in computational biology from Rice ing better than another? gas companies. University in Houston. While working Named after the world’s most com- as a self-described “data monkey” for Cyber Analytics mon computer language, SQLstream Chevron before graduating, he realized SparkCognition is another company says its platform is able to process more the potential of analytics to become a specializing in predictive maintenance than a gigabyte of data per second with central component of the oil and gas and has deployed its system with an a delay of only 5 milliseconds. business going forward. airline manufacturer and the two larg- Such processing capability is at the But something was missing—a est operators of wind turbines in North upper limit of what today’s systems can marketplace where producers could America. It recently inked an agree- handle and more than an hour fast- buy these applications without going ment with Flowserve, the world’s larger than processing giant Hadoop, said through the cumbersome process of est maker of oilfield valves, which is Damian Black, the company’s chief exec- vendor selection. marketing the system as a way to avoid utive officer. He described the platform So Ruths.ai created an exchange complete failures. as a “universal mechanism for trans- where it sells analytics apps it developed The Austin, Texas-based firm has also forming any form of input, in any for- along with those created by other soft- developed a cyber security analytics mat, arriving at any time, into any out- ware developers. Today the exchange application that is being used by cyber put format.” hosts more than 80 programs designed antiterrorism authorities and as a fraud This streaming data technology can for geoscience, play-wide analysis, pro- detection system for banks. This may be applied to just about anything using duction operations, and other areas. be an interesting area to watch since SCADA—supervisory control and data Many of these apps can be download- the oil and gas industry is one of the acquisition—systems. In modern oil ed for a few hundred dollars and only most frequent victims of cyber attacks fields it is hard to find anything not run- a handful cost more than USD 1,000. and espionage. ning on SCADA. Ruths said his intent is to disrupt the Philippe Herve, vice president of solu- The SQLstream platform’s flexibility current pricing model for these prod- tions at SparkCognition, said the cyber includes a user-friendly interface that ucts to make his exchange stand out security product can be used as a stand- Black said requires no specialized train- amid an increasingly crowded field. alone service, or it can be married with ing to master. Guided by the program’s “These prices are an order of magni- the firm’s predictive maintenance soft- recommendations, users can build cus- tude less than the competing counter- ware to add an extra level of assurance. tom apps in minutes and establish rules parts out there, and what we are trying “When an asset is starting to misbe- that could be used to control automated to do is lower the bar so more people can have, is it because we have a failure on it field equipment. have access to these types of solutions,” or is it because someone has attacked it,” With the back-end programming he explained. he asked. “If something goes wrong, you doing the heavy-lifting, users can change One of the apps enables operators need to know why.” their minds and modify how data are to quickly analyze workover candidates Herve added that unlike convention- analyzed or visualized on the fly. All and generate economic models for each al antivirus software, which base their modifications take effect immediately potential job. This is an example of a defenses around past attack meth- and without so much as a simple reboot. task that might take hours of research ods, analytics-based security platforms Black stressed that up until just a few for a single engineer to complete for an actively monitor a company’s networks years ago, such app development for entire field. for anomalies that may indicate when real-time data was the domain of very When using this program, as the cur- something or someone has invaded a expensive software consultants that sor is dragged across a field map, the cal- secure system. needed weeks to deliver. culations are completed almost instan- The SparkCognition software also “I like to draw the analogy with taneously. “You just give it the three runs constant internet queries to find spreadsheets,” he said. “Spreadsheets pieces of data that it needs—workover new threats. When it discovers one, or changed the way that people analyzed history, production, and well location— detects an ongoing attack, the program data, utterly. They are so easy to use and and then it will go through and calcu- automatically sends alerts and written you can do complex formulations and late the uplift factor-ratios for each job,” reports to security experts to brief them build financial models. We’re doing the Ruths said. on the situation.

38 JPT • OCTOBER 2016 JPT Magazine - October - Mermaid | 8.125” x 10.875” | FINAL

peloton.com/mermaid

It’s no myth. Finally, you can get production software even PA’s will love.

1660-24626_JPT_Magazine_Oct_Mermaid_8_125x10_875_FINAL.indd 1 2016-08-29 2:09 PM Accelerating the Uptake Cycle Through Collaboration, Outsourcing Trent Jacobs, JPT Senior Technology Writer

s a newcomer in Founded in 2015, Darcy is trying a partner at Darcy, adding, “which means to accomplish this by melding its in- outsourcing elements of the work that is Athe arena of oilfield depth research studies with a collabor- required for technology development to market research, Houston- ative approach to consulting. The lat- a group like us, where we do the research, ter part of this strategy involves placing scouting, screening, and evaluation.” based Darcy Partners has a group of promising young companies The culmination of all that ground- and their potential suitors into the work is quarterly technology forums set an ambitiously high same room where the free flow of ideas that Darcy hosts and moderates. The bar for itself: to speed up is encouraged. company has held two of these day-long The firm’s ultimate goal is to prove to events this year: one on advanced data the oil and gas industry’s oil and gas companies that technology analytics and another on new water man- validation becomes a much faster process agement technologies for unconvention- widely acknowledged when done collectively and openly com- al developments. Its upcoming forums and painfully slow rate of pared with when it is all done internally. will focus on unmanned aerial vehicles, “We are saying let’s do innovation completion optimization, cyber security, technology adoption. more efficiently,” said Hossein Rokhsari, and another on advanced data analytics.

Identify internal business opportunities that need technological solutions

Deploy de-risked external Identify technology solution internally with pilots and at enterprise scale Research the technology space to understand trends, Deploy Research established vendors and innovation opportunities

Validate Scout

Validate technology solution Scout for new vendors, startups, through due diligence and or R&D efforts in the ecosystem peer network that address the business opportunity

A new look at the innovation cycle proposes operators identify a challenge and then hand off many of the subsequent vetting duties to an outside firm before deploying the technology in a test environment. Source: Darcy

40 JPT • OCTOBER 2016 Host In attendance are key decision makers from oil and gas companies whose biggest pain points are used to steer the direction of the forum. Presenting are the entrepreneurs that Darcy has high-graded based on their ability to cure those issues. For the up-and-coming technology vendors, the forums are STRATEGIES FOR THE NEW ENERGY LANDSCAPE a chance to get invaluable face time with their target audience. Keynote Speaker Keynote & Opening Darcy takes a hands-on role with those entrepreneurs by pro- Ceremony Speaker viding them with corporate coaching to make their message as Rex W. Tillerson Dr. Sultan Ahmed Al Jaber ChairmanC & Chief Executive Ofﬁcer UAE Minister of State & Chief Executive Ofﬁcerr effective as possible. Once the pitch is made, operators have Exxon Mobil Corporation ADNOC GROUP OF COMPANIES the opportunity to put questions to the entrepreneurs about the feasibility of their innovations. Another of the firm’s partners, Jeremy Sweek, emphasized WORLD’S LARGEST TECHNICAL OIL & GAS CONFERENCE that this sort of transparency is rare in the oil field but necessary to address the fact that the industry has a tendency to run too many pilot tests in isolation. He explained that Patrick Pouyanné Bob Dudley Alexander Medvedev Vicki A. Hollub when operators test-drive startups, the results are typically Chairman & CEO CEO Deputy Chairman, President and CEO TOTAL BP Management Committee OCCIDENTAL PETROLEUM not shared publicly but end up percolating through word-of- GAZPROM CORPORATION mouth between professional colleagues and field workers. This practice of informal information sharing is less than desirable from a startup’s point of view because it makes Dr. Daniel Yergin Sir Richard Shirreff Mark Garrett Jeff Miller it hard for the company to appear on another operator’s Pulitzer Prize-Winning Deputy Supreme Allied Commander CEO President radar screen and grow its business. According to Sweek, it Author NATO Europe 2011-2014 BOREALIS HALLIBURTON also means disinformation about those startups often goes unchecked and leaves operators in a position where they are unable to fully leverage the sum of their aggregate experience. Toshiaki Kitamura Paal Kibsgaard Lars Christian Bacher President and CEO Chairman and CEO Executive Vice President “Most of the operators we have talked to have tried a pilot INPEX/JODCO SCHLUMBERGER Development and Production International with one, two, or three of these companies but it is a different STATOIL combination for each of them,” he said. “So why as an industry 4 EASY WAYS TO REGISTER AS A DELEGATE 1. Online: www.adipec.com/confreg 3. Tel: +971 2 6970 500 do we need 100 pilot projects to determine that there is a clear 2. Email: [email protected] 4. Fax: +971 2 4444 383 winner, or two or three winners?” OPENING CEREMONY & 1 106 HSE Another problem Darcy is trying to address with its KEYNOTE SPEECH TECHNICAL SESSIONS approach is vendor deluge. Using data analytics as an exam- E&P GEOSCIENCES GLOBAL BUSINESS 3 LEADERS SESSIONS OFFSHORE & MARINE OPERATIONAL ple, in this space alone there are hundreds of companies try- 8 CONFERENCE SESSIONS ing to break into the oil and gas business. Darcy has nar- EXCELLENCE MINISTERIAL 2 SESSIONS DRILLING COMPLETION rowed the list for its clients by distinguishing the startups SECURITY IN ENERGY 12 TECHNOLOGY selling what Sweek referred to as “fluff” vs. those selling SESSIONS PANEL GAS TECHNOLOGY actual breakthroughs that can lower operating costs or en- 8 SESSIONS MIDDLE EAST 8 PROJECTS ENGINEERING hance reserves. PETROLEUM CLUB AND MANAGEMENT WOMAN IN ENERGY VIP SESSIONS “When you say big data, we ask what do you really mean, 8 SESSIONS FIELD DEVELOPMENT and when you say analytics, what do you really mean,” he said. INDUSTRY TOPICAL LUNCHEONS UNCONVENTIONAL 3 3 BREAKFAST “You have to pull it back to see whether or not it’s really just a SESSIONS RESOURCES SESSIONS business intelligence tool that has a different user interface.” Rokhsari, who worked at an energy-focused venture capi- WHAT TO EXPECT DURING THE 4 DAYS OF ADIPEC? tal firm before starting Darcy, insists that the point of all this is not only to encourage operators to become investors in the 95,000 + 2,000 + 130,000 + sqm startups—it is to also encourage them to consider becoming Attendees Exhibiting Gross Exhibition Companies Space paying customers. The initial results of Darcy’s research efforts and technol- 25 14 15 ogy forums appear to be positive. The company reports that Inte rnational Country National International Pavilions Exhibiting Oil Companies Oil Companies every operator has discovered at least two or three startups that they plan to work with, and in turn, each startup has 8,500 + 700+ 120 walked away from the forums with three to five operators as Delegates Expert Countries Speakers potential clients. JPT REGISTRATION FOR THE EXHIBITION IS NOW OPEN www.adipec.com/visreg Supported By

Offcial Airline Offcial Travel Host City Offcial Hotel Offcial Media Venue Conference ADIPEC Agent Partner Organiser Organised By JPT • OCTOBER 2016 MANAGEMENT

The Reason to Expect Prolonged USD 30–60/bbl Oil

Rodney Schulz, President, Schulz Financial

Since 1982, the year oil prices were The Organization of Petroleum Export- TABLE 1—WORLD’S TOP TEN OIL deregulated in the US, the import ing Countries (OPEC) produces approxi- PRODUCERS BY COUNTRY oil price to the US has averaged mately 30 million BOPD. Although OPEC USD 52/bbl in 2015 dollars. It is vola- may attempt to collude and control oil Production, January–March 2016 tile, having a monthly standard devi- prices, world history has repeatedly Country (million BOPD) ation of 9%, although volatility ebbs proved that global cartels generally can- and flows. The global oil price, like not control markets for any sustained Russia 10.5 Saudi Arabia 10.0 the price of any commodity or securi- period of time. United States 9.2 ty in a free and open market, incorpo- This article is not about geopoliti- Iraq 4.3 rates all available information almost cal forces, but one can quickly see that China 4.1 instantaneously and follows a random OPEC members have distinct cultures, Canada 3.8 walk pattern. governments, customers, and priori- Iran 3.5 UAE 2.7 With the above factors in mind, what ties, as well as different production and Kuwait 2.5 can today’s oil and gas professional reserve replacement costs. Hence, it is Venezuela 2.4 expect for the near- to mid-term oil unlikely, and perhaps even unrealistic, price? Although the answer may not be for OPEC members to uniformly agree on Total 53.0 welcome, a fairly stable set of condi- production quotas and then expect each tions coalesce to make a strong reason member country to behave accordingly. understand this one needs to look at the to expect the oil price to generally range Therefore, it is probably a good assump- domestic oil production of the US vs. between USD 30/bbl and USD 60/bbl tion that global oil prices will submit to time (Fig. 1). for the foreseeable future. global market forces; i.e., oil prices will Although oil prices spiked from the likely continue to depend on the supply late 1970s into the early 1980s (Fig. 2), The Global Macro Situation and demand equation of competing pro- domestic production for the US moved To understand the situation, it is helpful ducers and consumers. minimally. However, during the strong to first look at two global macro factors. oil price period from 2006 to 2014, US First, the world consumes approximate- The Reasons for oil production increased from 5 million ly 90 million bbl of oil every day. And a USD 30–60/bbl Range BOPD to 10 million BOPD due to shale oil. second, three countries, Saudi Arabia, Why can one expect prices to fall in the Moreover, it is now widely known that Russia, and the US, are the largest sup- range of USD 30–60/bbl as previously the US has 78 billion bbl of technically pliers, with each producing between 9 suggested? The biggest reason is the sup- recoverable shale oil. and 11 million BOPD (Table 1). Beyond ply trade-off between conventional and The investment economics of US shale these three countries, production shale oil production, combined with fun- production is easy to describe in terms drops precipitously. damental investment economics. To first of option pricing. Simply put, when the break-even, or option “strike” price, moves above where it is economically Rodney Schulz, SPE, is president of Schulz Financial, a retail advisory firm, and is the viable, the acreage owner can exercise its founder and exclusive instructor for Oil & Gas Economics and Uncertainty, a 2-day option to make a profit by drilling. This seminar taught both privately and through SPE. His oil and gas experience includes a wide range of engineering, finance, and operational responsibilities throughout situation begs a few simple questions that North America for both major and independent producers as well as gas processing can be easily answered. companies. Schulz’s expert witness experience includes work for a major oil company, First, what strike price is necessary to the Federal Bankruptcy Court in Corpus Christi, and an international law firm. He has add shale oil production in the US? From also served as the financial director/chief financial officer for an organization with 150 a wide variety of sources and depen- employees in six states. He earned a BS degree in petroleum engineering from the dent on acreage location as well as other University of Kansas and an MBA from Duke University. factors, the strike price of US shale oil

42 JPT • OCTOBER 2016

12,000 to drill and complete a shale well USD 7 million, and the production cost 10,000 USD 8/bbl after drilling and completing a well, it takes: 8,000 (1−3/16)×(USD 40−USD 8)×

6,000 (average BOPD for an 18 month payout)×365×1.5=USD 7 million

4,000 Thus, with the above cost and price structure, the average daily produc- 2,000 tion rate for 18 months required for one well to satisfy the above economic 0 equation is 491 BOPD. To simplify the Average US Crude Oil Production (thousand BOPD) US Crude Oil Production Average equation, assume one drilling rig can add 500 BOPD per well, with the initial 1974 2010 2012 2014 1976 1960 1962 1964 1966 1968 1970 1972 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 Year production rate being higher and the 18-month end rate being lower. Fig. 1—US crude oil production by year. Source: US Energy Information Administration. If one next assumes it takes 6 weeks to drill and complete a shale oil well, and if 120 fracturing resources are available upon Average real price since 1982=USD 52 per Barrel demand, one rig can drill approximately

100 8 wells per year. Hence, one rig is capable of adding 8×500=4,000 BOPD in a year. Therefore, it takes approximate- 80 Real ly 250 drilling rigs to add 1 million bbl of oil per day to the production capacity 60 of the US. With 1,600 domestic rigs available and 40 approximately 400 rigs currently running, it is easy to see that the US has hun-

20 dreds of rigs that could be put to work,

Average Imported Oil Price (USD/bbl) Imported Oil Price Average Nominal as conditions warrant, and quickly add substantial production, thus putting the 0 resulting downward pressure on oil prices. 2011 1974 2010 2012 2013 2014 2015 1976 2001 1968 1969 1970 1971 1972 1973 1975 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2002 2003 2004 2006 2007 2008 2009 As prices move higher, more produc- Year ers will pick up drilling rigs to add pro- Fig. 2—US imported oil price vs. time. Source: US Energy Information duction, make a profit, and supply mar- Administration. kets with oil. And as prices move lower, producers will release rigs, allowing pro- ranges from USD 30/bbl to USD 50/bbl. answer is a simple “yes.” Specifically, the duction to fall with natural decline, until Therefore, when the oil price moves US has a total inventory of 1,600 drilling global production falls to a point where above the particular producer’s break- rigs, a sizeable percentage of which can demand again drives prices higher. even point/strike price, that producer can drill shale wells, and the current number One may ask: “Why is the expected range be expected to drill. of active domestic rigs is approximate- USD 30–60/bbl when the common strike Second, how much production can US ly 400. Therefore, the US has ample rig price for shale oil is USD 30–50/bbl?” In producers add as the oil price increases? capacity to seize drilling opportunities as short, the USD 50–60/bbl zone is a rea- Although all 78 billion bbl of technically conditions warrant. sonable buffer range. Should one also add recoverable shale oil may not be finan- This raises the question: “How many a lower buffer range of USD 20–30/bbl? cially viable at USD 30–50/bbl, it takes drilling rigs does it take to add, for exam- Probably not, because global reserve only 0.365 billion bbl, or 0.5% of the ple, production of 1 million BOPD?” replacement becomes exceedingly diffi- 78 billion bbl, to add 1 million BOPD of The answer to this is straightforward cult below USD 35/bbl. Again, one should production in a year. and simple. look at the USD-30/bbl and USD-60/bbl Third, are an adequate number of drill- For example, if the oil price is price marks not as absolute borders, but ing rigs and fracturing crews available USD 40/bbl, the required payout time as the range that contains most of the bell to exploit the shale oil in the US? The 18 months, the royalty 3/16, the cost curve for sustainable oil prices.

44 JPT • OCTOBER 2016 The latest In summary, production capacity, natural decline, prices, and drilling will move together to supply market demand. The stator elastomer world, in totality, generally operates on a capitalistic market supply and demand equation that has significantly influenced all aspects of global life and economics for centuries, if technology from not millennia. Of course, low-cost producers will always have an advantage. PV Fluid Products! Other Considerations One may argue that oil is different because a small number of producers “control” prices. However, if one compares the oil and gas industry with other industries, it is not unusual for two suppliers, in this case Saudi Arabia and Russia, to have a 25% market share. Another argument is that the production economics in the Middle East are far more favorable than elsewhere in the world. This is true, but it is nothing unusual for a small number of suppliers to have an economic equation that is better than their competitors. For example, a farmer can raise wheat in West Texas, but the farmer with land in western Kansas will have a significantly better economic equation for wheat production than the farmer in West Texas. One may also wish to bring political stability into the discussion. But again, political stability, or lack thereof, has always been a factor in economic equations. From the government Drive down non-productive time with side, the possibility of economic gain will always be an incen- drilling solutions that work!! tive to promote political stability. And from the investment side, companies will pursue business in politically unstable regions if vanced fi t philosophy provides the reward is worth the risk. UUÊÊ "#"$>%ÊVÊV&i$"V>%ÊÃÜiÊÃÜi%% UUÊÊ`vanced fi t philosophy provides er temperature range Technological advances have made shale oil produc- UUÊ Improved$«À(ved stator life in LÀ(>èr temperature range tion viable, given an adequate oil price. And it is safe to aggressive applications UUÊ,i`ÕVi`ÊV Reduced chunking&Õ#*"#}Êv> failures"%ÕÀiÃ assume that technology will continue to advance. However, UUÊÊL%e to maintain higher UUÊÊ"}&iÀÊ,"* a 1,500-millidarcy, 7,000-ft deep, 40-°API oil reservoir will differential pressures UUÊÊ#VÀi>Ãi`ÊL(ÌÌ($ÊÊ%"#i always have a competitive advantage over a 1-microdarcy, 7,000-ft deep, 40-°API oil reservoir. Why? It simply takes more horsepower and resources to extract oil from the 1-microdarcy reservoir than from the 1,500-millidarcy reservoir. But this is not the whole equation, as it is evident the world does PV Fluid Products not have enough 1,500-millidarcy oil production to fully supply demand. Conclusion Summarizing, it is likely that oil prices will range from USD 30–60/bbl for the foreseeable future. Short-term excur- sions outside of this range should not be surprising, but will probably not last for a sustained period of time. The reason for the likely USD 30–60/bbl range is the global interaction of producers and consumers, with a large potential shale oil supply that can be rapidly tapped as conditions warrant. A diminutive 0.5% of technically viable US shale oil can add 1 million BOPD of production for a year. The global oil supplier makeup is similar to that of other industries. Technology and political stability will factor into the equation, but this is nothing new. As with other industries, different suppliers have different economic The PROGRESSIVE rotor and stator company equations, but this is common with the constantly shifting global economic order. JPT

www.pvﬂ uid.com JPT • OCTOBER 2016 SPE HONOREES

Four Honored as Legends of Artificial Lift

Shauna Noonan, Occidental Petroleum, and Jeanne Perdue, Occidental Petroleum

The 2016 SPE Artificial Lift Conference and Exhibition for North America, to be held 25–27 October 2016 in The Woodlands, Texas, will feature a special Legends of Artificial Lift Luncheon honoring four people who have made outstanding contributions to the field of artificial lift technology. The 2016 SPE Legends of Artificial Lift Award recipients are John Bearden, Mark Mahoney, James McCoy, and Sid Smith Sr.

John Bearden has published 19 papers Mark Mahoney has received industry and holds 16 patents on the topic of recognition for his work in developing electric submersible pumps (ESPs). improvements to rod pumping equip- He has worked on the SPE task groups ment to address problems such as sand that developed Recommended Prac- production and high gas/oil ratios. tices on ESPs, authored the ESP chap- Mahoney was involved in the pioneer ter in the new edition of the Petroleum work of placing rod pumps in the devi- Engineering Handbook, and served on ated section of horizontal wells, and he the committee that developed the ISO Standard for ESPs. helped develop new designs and strategies for rod pumping in After earning BS and MS degrees in mechanical engineering deviated wellbores. from Texas A&M University, Bearden joined the Borg-Warner Author/coauthor of numerous SPE papers and coauthor of Ingersoll Research Center, which was Centrilift’s corporate three patents, Mahoney served on the API Committee for Stan- research facility. While there, he worked on various projects dards and Recommended Practices for Sucker Rod Pumps for studying the effects of gas on downhole pumps, which result- 10 years. He serves on the steering committees for the bian- ed in the development of Centrilift’s Rotary Gas Separator nual Middle East Artificial Lift Forum, an annual international prototype. He transferred to Centrilift’s Byron-Jackson Pump beam-pumping workshop held in Texas, and the biannual SPE Department in 1976 as a project engineer and, after Baker Enhanced Oil Recovery/Improved Oil Recovery Conference in Hughes acquired Centrilift, he worked his way up to the posi- Oman. He is a peer review editor for the SPE Production & Oper- tion of director of R&D systems engineering for Baker Hughes’ ations journal and the SPE Canadian Journal of Petroleum Tech- artificial lift product line. He was presented the Baker Hughes nology, and is a master class instructor for SPE. Lifetime Technology Achievement Award in 2007, and retired Mahoney worked at Harbison-Fischer for more than 26 years in 2015. in research and development of new products, reverse engi- His master’s thesis was on the subject of turbomachinery neering and redesign of sucker rod pumps (SRP), SRP system and fluid dynamics, and his whole career involved the design, designs, failure analysis, and training on SRP systems. He is an development, and application of ESP components and sys- ISO/API Q1 trained internal auditor, and he developed new qual- tems. “A career in artificial lift was good to me,” he said. “I ity processes, work instructions, and procedures for both inter- learned new technologies and new application solutions on a nal and external customers. consistent basis. That learning came from others in the com- Mahoney also worked as the senior artificial lift consultant pany, customer technologists, and even competitors.” for the Production Optimization Group at Lufkin Industries, Bearden credits his mentors for giving him the advice and and held the positions of senior artificial lift projects manag- support he needed to be successful. er and Middle East operations manager for Lufkin in Oman. “Find a mentor or mentors as you start pursuing your pro- Mahoney recently retired from Occidental Petroleum, where he fessional career,” he recommends. “They can help guide you, worked as the artificial lift advisor for 3 years. but it is up to you to determine your course. I was fortunate “My advice is to go and get as much field experience as you can, to have several mentors during my career, and the one that especially visiting the teardown of pumps/equipment and wit- helped guide me the most was John Tuzson, my manager at nessing installation and pulls of pumps, gas lift, and plunger lift the Borg-Warner Research Center,” who wrote the book Cen- equipment,” he said. “Visit manufacturers as well. Learn about trifugal Pump Design. how the equipment is made and the materials that are used so you better understand failures. Last but not least, find someone that is good at failure analysis and make good friends with him or her.”

46 JPT • OCTOBER 2016 James McCoy graduated from the Uni- Sid Smith Sr. earned a BS degree in versity of Oklahoma with a BS degree in petroleum engineering from the Uni- petroleum engineering and from Penn versity of Louisiana in Lafayette in 1958 State University with an MS degree in and went to work for Camco, where he petroleum and natural gas engineer- coauthored Camco’s Gas Lift Manual ing. He worked for a consulting firm with H.W. Winkler. In 1968, he joined and a major oil company, purchasing Conoco as a gas lift expert and held and operating oil and gas properties. In positions in production engineering, 1962 he acquired Echometer Company and expanded the acous- operations, and management in Houston, Dubai, and Indonesia. tic liquid level instrument’s capabilities into a full-fledged well After taking early retirement from Conoco in 1993, Smith analyzer system that is used worldwide to measure and improve chaired the SPE Gulf Coast Section’s Community Services Com- the efficiency of producing oil wells. He has received industry rec- mittee while consulting part-time for Camco and Shell. With ognition for his dedication and passion for teaching others how Shell, he taught gas lift schools in The Hague, Gabon, Nigeria, to use downhole pressure data and inflow performance to maxi- Brunei, and Oman while assisting with the deployment and use mize well production. of WinGLUE, Shell’s patented gas lift surveillance software. He McCoy is active in SPE, having authored or coauthored more is currently working as a part-time consultant for Appsmiths, than 30 SPE papers, the first being SPE 337 titled “Analyzing assisting with gas lift software development and services in Well Performance.” He holds numerous artificial lift patents Houston. For the past 10 years, he has been executive director and was instrumental in establishing the McCoy School of Engi- of Urban Outreach Inc., a faith-based nonprofit serving inner- neering at Midwestern State University in Wichita Falls, Texas. city families in Houston. “The artificial lift industry has been a great way to support Smith said engineers can remain in production engineering myself and my family,” he said. “The job has been both in the on the technical ladder throughout their career, rather than office and in the field, which I like. The industry has a lot of climb the management ladder. “The key thing about being con- challenges, and I have really enjoyed working on those challeng- nected in artificial lift is, artificial lift can be and usually is uti- es—the varied challenges have been fun. I recommend the arti- lized throughout the reservoir life cycle. So the opportunity for ficial lift industry to young people looking for a profession that gainful employment is there even in times of low oil and gas is interesting and exciting. I have been privileged to work with prices. I am still as busy as I was before oil prices dropped below such a great group of people.” USD 100 a barrel.” JPT Survival of the Simplest

A new species of toe valve >> OmniFrac Pro Toe™ For decades, engineers have had to trust the future of their wells to the lesser of two evils: exceptionally expensive TCP or unreliable ﬁrst generation toe valves. Superior Completion Services has advanced toe valve technology with the new OmniFrac Pro Toe™, a mechanically simple design that simply works.

superiorenergy.com/completionservices TECHNOLOGY FOCUS

Field Development Projects Maria A. Capello, SPE, Executive Adviser, Kuwait Oil Company

The low price of oil has had an immediate es; and inclusiveness of breakthrough shareholders, consideration of other effect in the planning departments of oil technology. Additionally, a new trend projects’ supply chains in nearby fields, companies. They were forced to shift the in field development that has proved to contract strategies, and even weather. focus and carefully rank and select only be beneficial and that assesses profit- Remarkably, the academic, theoretical, those developments that would ensure ability of production strategies is the modeling, and research studies have also profitability in the production of oil and incorporation of pilot or early- or initial- shifted toward assessing the optimal or gas. Hence, the field-development proj- production units to establish real project economic development of shale oil and ects need to include and consider not economics and boost opportunities for gas, offshore fields, and environmentally only a static or dynamic subsurface char- understanding the cost-reduction oppor- protected areas, marking an alignment acterization but also the production- tunities. This is a strategy incremental- within the industry to solve the current systems and facilities options, to trigger ly applied globally by international and concerns and incorporating in the studies profitability and establish clear break- national oil companies. The complexities a very practical engineering or economic even thresholds. More than ever, the con- of the development plans have propelled approach not frequently present before. sideration of deep water, tight reservoirs, an integration among professionals of That all the optimizations, synergies, shale oil, remote locations, or environ- subsurface studies, production systems, and successes presented in the papers mentally critical plays is placed under the construction engineering, transporta- you will read are flourishing now is microscope. Increasingly difficult project tion, and even marketing teams, to an not coincidental but clearly is a result economics has delayed or stopped invest- increased collaboration not seen before. imposed by the hard times derived from ments that were estimated to be safe and The three papers selected for this the low price of oil and the accelerat- profitable before the price drop. feature typify this necessary integrated pace with which we want to produce In this extended-low-oil-price frame- ed approach in field developments. In our resources. An extended-low-price work, a positive economic return is these papers, as well as in those sug- era seems to have settled in, and, even if actively sought by ever-incremental inte- gested for additional reading, the main in the future 3 to 4 years the price of oil gration of subsurface and surface engi- factor is the integration of all elements rises, the new field-development-opti- neering, to select strategies that lower involved in ensuring economic profit- mization strategies we are testing and the production cost. I would like to men- ability, including subsurface studies, applying now will serve our industry well, tion that some of the key elements that drilling and completion, construction optimizing and lowering the break-even enabled better economics in the most engineering, production systems, tech- thresholds for the huge resources we still successful projects were new proj- nology, environmental considerations, need to produce worldwide. JPT ect designs that ensure flexibility; syn- global supply and construction integra- ergies with contractors, especially for tion, compliance with government reg- supply-chain improvements; allianc- ulations, synergy with contractors and Recommended additional reading at OnePetro: www.onepetro.org.

Maria A. Capello, SPE, is an executive adviser with the Kuwait Oil SPE 178191 Cluster Wells Applied in Wetland Environment in North Azadegan Company (KOC) for the North Kuwait Asset, advancing strategic by Bingshan Liu, CNPC Drilling Research initiatives in reservoir-management best practices for all assets Institute, et al. of KOC and diversity for all companies upstream and down- stream of the Kuwait Petroleum Corporation holding. She is an SPE 177204 Defining the Optimum Exploitation Strategy Combining Water experienced consultant for the oil and gas industry and an Injection, Field Development, and Artificial- expert in field-development and -monitoring strategies. Capello Lift Analysis to a Mature Field Through has worked in Latin America, the United States, and the Middle Surface and Subsurface Coupled Models East. She holds a licentiate degree in physics from Simon Bolivar University and an by O. Espinola, Schlumberger, et al. MS degree in geophysics from the Colorado School of Mines. Capello holds an honor- SPE 175531 Effect of Well Spacing on ary lifetime membership from the Society of Exploration Geophysicists and has Productivity of Liquid-Rich Shale Reservoirs received its Distinguished Member and Regional Service awards. She serves on the With Multiphase Flow: A Simulation Study JPT Editorial Committee and can be reached at [email protected]. by A. Khanal, University of Houston, et al.

48 JPT • OCTOBER 2016 Sapinhoá Field, Santos Basin Presalt: From Design to Execution and Results

his paper presents the development Tof Sapinhoá field, covering the fast- track transition and decision-making process, from appraisal to conceptual and basic engineering of the Sapinhoá pilot project and on to its subsequent execution, highlighting the challenges, lessons learned, and results. Even though, at the time, several uncertainties about developing presalt areas were present, a fast-track strategy was chosen. A pilot project was seen as a means to provide valuable information for the remaining development of Sapinhoá and other fields in the presalt cluster. Fig. 1—Sapinhoá Field and Santos Basin presalt cluster. Introduction The Sapinhoá field is in Block BM-S-9 at ◗ Estimate vertical and horizontal ◗ Number of permanent downhole the central portion of the Santos Basin permeability gauges (PDGs) per zone, to manage (Fig. 1). The water depth is approximate- the reservoir better ly 2140 m, and reservoir depth lies be- Project Engineering: ◗ Remote selectivity of the producing/ tween 5000 and 6000 m, with salt-layer Concept Overview injection zone, achieved through the thickness up to 2 m. Production-development plans consisted selection of hydraulically operated The main objectives of the Sapinhoá of the installation of a chartered float- control valves pilot were to evaluate the production and ing production, storage, and offloading ◗ Lithology uncertainties, mitigated injection behaviors in a carbonate reser- (FPSO) vessel in the field’s south area; by the use of large-bore wellheads, voir of microbial origin and obtain infor- a well-construction campaign; a sub- allowing an extra casing string mation and provide technology devel- sea gathering system; and gas-export- without affecting production-casing opment so that those could be used for pipeline installation. diameter the remaining development of the pre- The project final scope consisted of salt cluster. This information would be Well System. Well design for the Sapin- 11 wells—six producers, one gas injec- u s e f u l t o hoá pilot had to take into consideration tor, one water-alternating-gas (WAG) ◗ Define the best recovery method to requirements from other disciplines, injector, and three water injectors that be applied in future projects which affected final configuration, in- later could be converted to WAG injectors ◗ Define the best exploitation strategy cluding the following: if necessary. ◗ Optimize location, geometry, and ◗ Scaling precipitation risk, which As for completion design, intelligent number of wells was mitigated through inclusion completion was the primary choice, fol- ◗ Evaluate reservoir hydraulic of a downhole chemical-injection lowed by single-zone completion, de- communication mandrel pending on the number of intervals found. All wells had a 6⅝- and 5½-in. combined production string, in order to This article, written by Special Publications Editor Adam Wilson, contains highlights cope with the expected high flow rates. of paper OTC 26320, “Sapinhoá Field, Santos Basin Presalt: From Conceptual Design Acid stimulation was considered for all to Project Execution and Results,” by J. Turazzi Naveiro, SPE, and D. Haimson, wells, not only to achieve the maximum Petrobras, prepared for the 2015 Offshore Technology Conference Brasil, Rio de flow rate but also to guarantee that the Janeiro, 27–29 October. The paper has not been peer reviewed. production or injection profile would be close to the simulated distribution. Smart Copyright 2015 Offshore Technology Conference. Reproduced by permission. wells included PDGs, chemical-injection

The complete paper is available for purchase at OnePetro: www.onepetro.org.

JPT • OCTOBER 2016 49 mandrels, and hydraulic valves below the tegrated at Brasfels shipyard in Brazil. Despite this environment, the Sapin- feedthrough packer. The process plant has overall capacities hoá pilot well-construction-campaign of 120,000 B/D of oil, 150,000 B/D of costs and schedule performance were Subsea Production System. The devel- liquid, 150,000 B/D of water injection, reduced significantly during its execu- opment of a subsea production system 5 000 000 m3/d of gas production, and tion because of an accelerated experience for the Sapinhoá pilot was challenging. 3 250 000 m3/d of gas export. curve. At the basic-engineering phase, The carbon dioxide and possible hy- the campaign duration was estimated to drogen sulfide presence affected mate- Project Strategies: Dealing last 2,368 days. It is now projected to rial selection for equipment and pipes With a Fast-Track Execution take 1,997 days, a 15% reduction. in multiple ways because special alloys Because of the fast-track guidelines ad- have longer lead times to delivery and re- opted for this project, the project scope Subsea System. Because of the fast-track quired qualification of material and weld- began being defined at an early stage. implementation and the innovative as- ing process. The 2000-m water depth, Strategies were needed that could guar- pects related to the presalt subsea equip- combined with metocean data that in- antee some flexibility for the project, ment, the first deliveries suffered delays, dicated a harsher environment in San- allowing for changes in the future. mainly with regard to the trees. To deal tos Basin than in Campos Basin, posed a with that, the project team had to review problem to riser design and installation. Well-Construction Campaign. The its well-campaign schedule continuously, Finally, reservoir pressures and the re- strategy adopted for well construction postponing some well completions and quirement to inject gas demanded a high- was to charter three rigs to be dedicat- anticipating the drilling of new wells. ly resistant gas-injection line. ed to all projects of the block, including At the end, those delays had no direct A buoy-supporting-riser (BSR) concept the Sapinhoá pilot. With dedicated rigs effect on the project ramp up because was considered the alternative with the for the project, it was possible to special- the wells campaign was executed faster best economics after technical-viability ize each of them in certain activities and, than planned. verification. The decoupled BSR gath- consequently, optimize the wells cam- ering system was a new riser concept paign, reducing time for each activity. Conclusion featuring a tethered buoy, measuring In spite of all the challenges and risks 52×40×10 m, which is submerged ap- Subsea Equipment. Because of the scope envisioned at the conceptual phase, the proximately 250 m and has 22 to 28 risers uncertainties and the long-lead-time sup- Sapinhoá pilot managed to become a suc- connected to it, including steel catenary ply for subsea equipment, the presalt tree cessful project, both technically and eco- risers (SCRs) for production and injec- standard was defined as a unique model nomically. Approximately 1 year after the tion purposes, flexible lines for gas lift, that could be used for producers, water field’s declaration of commerciality, the and electric/hydraulic umbilicals. injectors, gas injectors, and WAG wells. production unit was already operational, The remaining architecture of the sub- That way, the order could be made even and, 2½ years after startup, oil produc- sea system included satellite production if the scope or sequence of well construction is at peak. wells connected individually from each tion had not been fully defined. In order to deal with the complexity pipeline end termination to its respective and uncertainties of the reservoir, com- tree through flexible flowlines and a ver- FPSO-Vessel Construction. The Cidade bined with the fast-track implementa- tical connection module. de São Paulo process plant was designed tion, the project conception and devel- so that it would be able to receive differ- opment strategies needed to be based Flow-Assurance System. The main con- ent kinds of fluids and to deal with differ- on two pillars: flexibility and robustness. cern regarding flow assurance for the ent recovery methods in the future. Sapinhoá pilot was the possibility of Flexibility. This was applied, for exam- scaling precipitation in the production Project Execution: ple, in different possible recovery meth- string. To mitigate this risk, it was rec- Results and Lessons Learned ods, number of wells, well-construction ommended that all wells be stimulated to Reservoir Results. The drilling results configurations, contingency plans for full capacity because smaller drawdowns confirmed the expectations for the Sapin- subsea activity, and different FPSO opera- reduced precipitation. Downhole scaling- hoá field, presenting excellent productiv- tion methods. inhibitor injection also was necessary be- ity and injection. The initial-production cause the precipitation normally occurs potential flow rates are ranging between Robustness. This was applied, for exam- at the largest-differential-pressure point, 29,000 and 36,000 B/D, surpassing ple, with tree specifications, metallurgy which is the transition between the reser- conceptual-design estimates of 25,000– selection, and wellhead and casing design. voir and the well. 30,000 B/D. The continuous search for and implementation of flexible and robust solu- FPSO Vessel. The FPSO vessel Cidade de Wells Campaign. The Sapinhoá pilot tions were key success factors of the proj- São Paulo was converted from a double- well-engineering program had to deal ect, enabling the management team to sided single-bottom vessel in China, with with uncertainties and challenges from deal with design and execution chang- the process modules being construct- a technical point of view and from the es as they appeared, keeping the proj- ed at different sites worldwide and in- imposed fast-track approach. ect optimized. JPT

50 JPT • OCTOBER 2016 KBC Adv_ SPE Journal of Petro Tech_ Oct2016.pdf 1 9/7/2016 8:34:07 AM

World-class technology and consulting services for designing, operating and managing hydrocarbon production and processing facilities.

Production System Optimisation Flow Assurance Production Network Design Fluids and Oil Property Analysis Reliability & Maintenance Energy Efficiency C A Yokogawa Company Organisational Effectiveness M

CM MY Our Engineering CY CMY Software Suite K

KBC will be at ADIPEC in Abu Dhabi, 7-10 November. Visit us at Booth 8412.

KBC Advanced Technologies [email protected] • www.kbcat.com • blog.kbcat.com AMER: +1 281 293 8200 APAC: +65 6735 5488 EMEIA: +44 1932 242424 Improving Megaproject Economics Through the Use of Early Production

key aspect of the project discussed the goal of capturing the benefits of the with the drilling program and construc- Ain this paper is the use of minimal initial-production strategy. tion phase of the permanent production initial-production facilities to achieve facilities. Piping headers, as well as basic significant early production from Benefits and Challenges lighting, instrumentation, and controls each of four preconstructed artificial The initial production strategy evolved to allow sustainably safe operation, were islands. The initial-production facilities and was optimized from conceptual de- all designed to fit within a small space in leverage the use of spare processing sign through front-end engineering and such a way that the drilling rig could ma- capacity on existing adjacent satellite detailed design, resulting in the fit-for- neuver over the facilities to reposition to platforms to allow for production from purpose initial-production facilities for new drilling locations free of clashes with each of the islands during the 4-year the first three islands. Given the short op- the initial-production facilities. This also period between completion of the first erating duration of the initial-production will allow for easy dismantling of the facil- wells and completion and startup of the facilities, cost efficiency was the key driver ities without production interruption fol- permanent production facilities. in decision making. To ensure implemen- lowing startup of the permanent produc- tation of only the most cost-effective fit- tion facilities. It was a significant design Introduction for-purpose facilities, the early operating challenge to arrange all components of The project is approximately 80 km off- experience from the first island was used the facilities within the defined envelope, shore in the Persian Gulf and consists to simplify further the design of the initial- but achieving this allowed for increased of four artificial islands. The objective production facilities on the fourth island. production and, ultimately, approval of a of the initial production phase is to pro- Achieving the execution schedule for the more economically feasible project. vide early and sustained production from initial-production facilities to allow start- Coordination with the drilling group was wells drilled on the islands as soon as they up to be achieved as early as possible in integral during the design of the facility, to are available, thereby achieving advanced the drilling program, to maximize the du- ensure that emphasis on safety in design revenue generation and improved project ration of the initial-production phase, was was applied while also helping to ensure economics while the permanent produc- a key driver and challenge. Without effec- minimum downtime of the facilities. As tion facilities are under construction. tively managing the execution schedule to part of the simultaneous operations on the Initial startup was achieved from the achieve startup on each island as planned, islands, the initial-production-facilities first island in November 2014, followed the economics of the initial-production scope is designed, installed, and commis- by initial startup of the second island in phase becomes considerably challenged. sioned in such a way that the drilling se- April 2015. This is the first phase of the The project management team took a quence of the wells requires no shutdown plan to increase production from the field structured approach and implemented for the hookup of a new well to com- by approximately 40%. The project is seg- several key initiatives to allow efficient mence production. Continuous coordina- mented into an initial-production phase execution of the work while maintaining tion meetings were held with the drilling and permanent production buildup phas- excellent safety performance. This struc- group along with a structured approach es. The initial-production phase will con- tured approach was instrumental in man- to design freezes to the initial-production- tinue until completion of the first perma- aging the key project interfaces associ- facilities engineering to allow the drilling nent production buildup phase in 2018. ated with the ongoing drilling program, group the maximum flexibility in planning The project achieved startup of initial space limitations inherent to an island, and initial drilling on each island while en- production within 4 years of conceptual multiple contractor interfaces, logistics suring that fabrication and construction design in 2010 and a minimum of 4 years challenges, and simultaneous operations. of the initial-production facilities could before completion of the permanent The initial-production facilities on each commence without affecting the schedule. production facilities, thereby achieving island were designed to avoid interference The initial-production facilities follow a low-cost, keep-it-simple philosophy. Given that the facilities are intended only This article, written by Special Publications Editor Adam Wilson, contains highlights for short-term operation, they require of paper SPE 177729, “Improving Megaproject Economics Through the Use of Early mainly attended operation, using manu- Production,” by James A. Volker and Adam J. Bond, Zakum Development Company, al valves, with the exception of the shut- prepared for the 2015 Abu Dhabi International Petroleum Exhibition and Conference, down valves. On the first three islands, the Abu Dhabi, 9–12 November. The paper has not been peer reviewed. facilities also share test headers and mul-

For a limited time, the complete paper is free to SPE members at www.spe.org/jpt.

52 JPT • OCTOBER 2016 tiphase flowmeters. The fourth island was production facilities. This proved to be development scheme has been imple- added after the initial three islands were a substantial challenge considering the mented successfully in a cost-effective designed, and experience learned from close, 6-m well spacing on the islands and manner by leveraging the spare process- the previous islands was used to achieve numerous drilling locations per island. ing capacity on the existing satellite plat- greater efficiencies in the design and op- This challenge is managed and controlled forms in conjunction with installation erations by removing the test header and by implementing detailed simultaneous- of minimal new initial-production facili- temporary-equipment room altogether. operations guidelines and procedures ties on each of four preconstructed ar- In addition to generating revenue containing rigorous work-management tificial islands along with short inter- with minimized investment, the initial- and permitting requirements. Finally, connecting pipelines. Tightly controlled, production facilities also provided a valu- strict and ever-changing area ownership cost-driven design and operating crite- able mitigation for any schedule delays and work-management principles have ria ensured that the completed initial- that might have been incurred regard- been applied, allowing for project, drill- production facilities were fit for purpose ing the completion of the permanent ing, or operations ownership, depending for the short 4-year operating life. The production facilities because all drilled on the phase of the project in any given structured approach to managing the si- wells are still able to produce through area. Hard barriers are erected before multaneous operations and key inter- the initial-production facilities up until startup, and areas adjacent to the live faces helped deliver the execution cer- the planned startup of the permanent production facilities are under controlled tainty required to achieve completion production facilities. access by the operations group as the of the initial-production facilities on Startup of production from the first and area authority. Work performed outside the first three islands safely within the second islands occurred soon after the first this controlled-access area remains under same time frame as completion of the two wells were completed on each island, the stewardship and permitting control initial wells required for startup, thereby allowing for the minimum design flow re- of either the drilling group or the project maximizing the duration of the initial- quirements for the initial-production fa- group. This allows the party best suited to production phase and maintaining the cilities. Subsequent wells are progressively the environment to control the work and attractive project economics. Successful hooked up to the initial-production facili- safe-working procedures. implementation of the initial-production ties immediately following completion of phase also provides a valuable mitigation drilling. Project teams have minimized the Conclusion for any schedule delays that may be in- duration from completion of drilling to The strategy to include an initial- curred with regard to completion of the production flow by preinstalling all major production phase in the overall project- permanent production facilities. JPT components, except for the closing spool to the wellhead. All commissioning activities are completed before final hookup. Well-operations and project-construction teams complete their activities in parallel, to the extent possible, to reduce hookup duration following completion of drill-'RZQORDGIURP ing. This approach has been successful at ZZZVHUDILPOWGFRP quickly producing wells as they become available. Wells continue to be hooked up in an expeditious manner, and the third island will start up immediately following the completion of the first well. One of the key challenges during the )8785( initial-production phase is the high degree 5HSODFHVWKUHH of simultaneous operations during construction, and even more so after startup, 6WRSZDVWLQJ RI\RXUROG on each island. Before initial startup on H[SHQVLYH each island, the detailed coordination of work activities requires construction ac- PRQH\ V\VWHPVRU tivities to progress efficiently in conjunction with drilling operations while mini- VSUHDGVKHHWV mizing the risk to personnel. Following startup, construction and commissioning activities associated with future well tie-ins occur in conjunction with ongoing 6(5$),0 )XWXUH drilling operations adjacent to producing facilities and alongside ongoing well op- Productionforecastingandreservesmanagementsystem erations and construction of permanent

JPT • OCTOBER 2016 53 Completion and Well-Spacing Optimization for Horizontal Wells in Pad Development

n organic shales, hydraulic fracturing on the well spacing. The following are Complex Reservoir Description Iis important for optimizing the some of the methods that are applied to Without Explicit Hydraulic-Fracture production of horizontal wells. For define possible well spacing. Description. In the previous two meth- a standalone lateral, the propped Analogs. Using analog methods re- ods, the simple reservoir description al- surface should be maximized to quires an older field already in develop- lows for an analytical model to be run increase production. In the case of ment with similar rock and fluid quality, in seconds or a numerical model in min- a pad, well spacing is an additional structural complexity, well architecture, utes. However, describing each reservoir factor. Competition for production and hydraulic-fracture stimulation. The property by a single value does not allow between laterals of a given pad should analog approach leads to large uncer- consideration of spatial variation of the be minimized and is the result of both tainties if any of these parameters are not reservoir. Once a larger set of static data well spacing and hydraulic-fracture within the same range or if the older ana- becomes available, a more comprehen- design. A numerical model coupling an log field does not have sufficient statisti- sive model can be set up to assess the explicit description of the hydraulic- cal representation. presence of, for example, petrophysical fracture geometry and reservoir Simple Reservoir Description With- variations within the reservoir and their simulation is proposed to evaluate out Explicit Hydraulic-Fracture De- effect on the well spacing. This approach production interference. scription. The first step in simulation requires a complex numerical reservoir is to start with a simple reservoir de- simulator. The fracture geometry is kept Introduction scription in which each property is uni- simple, building each fracture as a sur- Well Spacing. One of the more critical form (porosity, permeability, and fluid face of constant properties in a single- questions that any asset team must con- saturations). The reservoir model is porosity model or as creating a zone of sider concerns well spacing when de- then coupled with an analytical or nu- enhanced porosity. signing the development plan for a new merical production simulator. This kind Complex Reservoir Description In- field. Well spacing defines the total num- of approach is appropriate for a field cluding Explicit Hydraulic-Fracture ber of wells, the drilling and comple- with minimal information. The hydrau- Description. The last type of model tion schedule, and the field-production lic fractures are represented as paral- features the explicit description of the curve. The ultimate decision relies on lel planes that intersect the wellbore. hydraulic-fracture geometry by use of economic analysis, balancing the expect- Hydraulic-fracture properties are kept a fracture simulator along with a de- ed ultimate recovery against capital and constant over their whole surface and tailed description of the reservoir. Both operational expenditures. along the lateral. hydraulic-fracture geometry and reser- This work focuses on methods to eval- Simple Reservoir Description In- voir description are then merged into a uate different well spacing and on how to cluding Explicit Hydraulic-Fracture simulation grid. reduce the range of uncertainty in the de- Description. The next level of complex- velopment of the Vaca Muerta play. ity with a simple reservoir- description Production-Interference Character- model is to include variation in propped- ization. Interference between wells in Well-Spacing Evaluation. Well spac- fracture geometry. Some commer- unconventional developments is more ing depends on geomechanics, petro- cial software can handle this type of common than expected and can occur at physics, stimulation design, hydraulic- model, where each fracture can be de- different stages of well life. The interfer- fracture geometry, and reservoir fluids. fined individually in terms of length, ence could be These factors are combined by use of nu- height, and the angle with respect to ◗ Between a well being drilled and merical simulation to study their effects the wellbore. an adjacent well being stimulated ◗ Between two wells being stimulated ◗ Between a well in production and This article, written by Special Publications Editor Adam Wilson, contains highlights an adjacent well being stimulated of paper SPE 180956, “Completion and Well-Spacing Optimization for Horizontal ◗ Between two wells in production Wells in Pad Development in the Vaca Muerta Shale,” by M. Suarez, SPE, YPF, and In this work, focus is placed on the S. Pichon, SPE, Schlumberger, prepared for the 2016 SPE Argentina Exploration and latter two types of interference, given Production of Unconventional Resources Symposium, Buenos Aires, 1–3 June. The their possible reduction of estimated paper has not been peer reviewed. ultimate recovery.

For a limited time, the complete paper is free to SPE members at www.spe.org/jpt.

54 JPT • OCTOBER 2016 Petrophysics

Geomechanics Static Model Hydraulic-Fracture No Simulation

Fracture Natural Match? Fractures

Yes

Forecast No Gridding

Production Simulation Production Yes Completion Match? Optimization

Fig. 1—Reservoir-centric fracturing-to-production work flow.

Methodology from the formation from the properties of the reservoir to the Reservoir-Centric Fracturing-to- petrophysical inputs and the production data. Production Work Flow. To model hor- properties of the hydraulic fractures izontal wells in organic-rich shale opti- as per the hydraulic-fracture From Single Well to Multiple-Well mally, emphasis is made on maximizing simulation. Pad. Well spacing being the final ob- the integration of reservoir description ◗ A production history match is jective of the study, and knowing that with the resulting propped hydraulic performed, adjusting dynamic production interference is because of fracture that can be designed. This is achieved by coupling a hydraulic-fracture simulator with a reservoir simulator. The reservoir-centric fracturing-to- production work flow (Fig. 1) allows for a high level of integration along with the possibility of separately varying each input to evaluate its effect on overall production. The basic steps of the fracturing-to-production work flow are the following: ◗ Integrate petrophysics, geomechanics, and natural-fracture characterization along with reservoir structure in a static model. ◗ Hydraulic-fracture simulation is performed using observed pressure data from fracturing operations as calibration points. It must be noted that the hydraulic-fracturing STAR-CCM+: Discover better model solves explicitly for the designs, faster. interactions between geomechanics and natural fractures to define Virtual Wave Basin: Explore designs digitally to reduce cost without compromising safety. the resulting hydraulic-fracture geometry. ◗ For flow/production simulation, siemens.com/mdx hydraulic-fracture geometry is explicitly gridded in an unstructured manner, populating properties

JPT • OCTOBER 2016 55 the overlap of hydraulic fractures from tion is to increase the treatment size lateral length is another design param- different wells, the reservoir-centric and proppant mass. The higher produc- eter. A sensitivity analysis is performed fracturing-to-production work flow is tion rate then can justify the increased keeping the number of clusters per stage the best-suited methodology. cost. However, when considering com- constant but increasing the number of To mimic the behavior of laterals in peting production between neighboring hydraulic-fracturing stages from 15 to field development, a three-well-pad wells within a pad, a new constraint to 20 within the lateral. Increasing the approach is proposed. Then, to evalu- stimulation-treatment size is introduced. number of fracturing stages increases ate production interference, two cases To evaluate the case, a sensitivity the total number of perforation clus- are proposed. analysis is performed by varying the size ters and reduces the spacing between Single-Well Simulation. This con- of the fracturing treatments while main- perforation clusters. siders a central standalone lateral. The taining constant the number of stages When comparing simulated pro- single-well case is used as a baseline in and fluid and proppant-type fractions duction in the pad configuration, the which no production interference im- and comparing scenarios of a stand- 15-stage scenario is shown to outper- pairs well final recovery. alone lateral and a lateral surrounded form the 20-stage scenario for long- Pad Simulation. After emulating by hydraulically fractured and produc- term hydrocarbon recovery, regard- the hydraulic fracture in each of the ing neighboring wells. When compar- less of the size of the hydraulic-fracture wells, the pad is produced in a multiwell ing standalone-well and well-in-pad treatment. This result is counterintuitive reservoir-simulation mode. This allows scenarios, the relative cumulative hy- but must be considered in the scope of the pad case to consider explicitly the po- drocarbon gain when increasing the hydraulic-fracture geometry and the re- tential reduction in the rate of the central total amount of proppant per well is resulting propped surface. In a horizontal well caused by overlap of the drainage duced by half when the well is compet- well, competition for growth from the areas of the three wells. ing for production with neighbors. Sim- stress shadow within and between stag- ply increasing the hydraulic-fracture es is an important factor to consider. By Discussion size, even if effective for a single well, reducing the spacing between perfora- Hydraulic-Fracturing-Treatment Size. might not be the optimal solution for a tion clusters, each independent hydraulic When designing a stimulation treatment pad development. fracture suffers a stronger stress shadow. for a well, a strong correlation has been This stronger stress shadow results in seen between the propped area and final Stage Number and Cluster Spacing. a higher fracture-propagation pressure hydrocarbon recovery. The simplest op- The number of hydraulic fractures per developing within each hydraulic fracture. This higher fracture-propagation pressure strengthens the disequilibrium of fluid distribution among perforation Motivate clusters that limited entry might not be able to overcome. An uneven fluid distri- Inspire bution finally leads to a smaller created propped surface and then affects hydro- Educate carbon recovery. Well Spacing. The first option to re- Nominate duce the competition between neighboring wells is to increase well spacing. Do you have colleagues who are authorities in their ﬁelds and Two cases are simulated by keeping the experienced public speakers? If you do, consider nominating well completion constant but consid- one or more of them for the Society of Petroleum Engineers ering two different well spacings. In- Distinguished Lecturer Program. creasing well spacing leads to higher production and reduces the relative dif- Learn more about the program at www.spe.org/go/NomDL. ference between a well in a pad and a Nominations are accepted until 15 March. standalone well. However, by increasing well spacing, even if the performance of an individual well is improved, the total number of wells to develop a given area would be reduced, which might reduce total field production. The method presented allows for the The SPE Distinguished Lecturer Program is funded by the SPE Foundation, Ofshore Europe, AIME, and companies evaluation of the relative effect of each that allow their professionals to serve as lecturers. scenario and its economic outcome over the global field development. JPT

56 JPT • OCTOBER 2016 TECHNOLOGY FOCUS

New-Frontier Reservoirs II Leonard Kalfayan, SPE, Global Production Engineering Adviser, Hess Corporation

With crude-oil prices continuing to lan- [E]nhancing production improved containment and fracture con- guish, margins in tight-reservoir-asset ductivity, even formation permeability developments have continued to tighten from multizone, propped- enhancement? Are there learnings from and new drilling and completion activity, fracture completions in tight long-term injection operations in tight of course, is substantially reduced. Look- reservoirs, for example, is reservoirs that can be applied to well- ing back 2 years, the focus in onshore- stimulation operations? asset development has essentially shift- not straightforward. There Also, for existing well completions, ed entirely from fast-paced growth of are questions to address, can reactive fluid (e.g., acid or other tight hydrocarbon reservoirs to produc- chemicals) be used effectively, and, if so, tion enhancement from existing (but especially with respect to how? Can well production be enhanced still profitable) wells, as well as to max- the understanding of the by injection of fresh water to remove salt imizing productivity from the smaller potentially residing in existing natural- number of new well completions. contribution of natural fracture systems? However, enhancing production from fractures and induced The papers featured this month provide multizone, propped-fracture comple- unpropped fractures. assessments and discussions concerning tions in tight reservoirs, for example, these and other related questions. Each is not straightforward. There are ques- paper is unique, but they all share the tions to address, especially with respect wells that are candidates for refractur- intent of advancing efforts in enhancing to the understanding of the contribu- ing? Is there an opportunity with the production from existing and new well tion of natural fractures and induced use of smaller proppants? What are the completions in tight reservoirs. JPT unpropped (IU) fractures. How can nat- implications of fracture and well spac- ural and IU fractures be accessed or ing? Can the pumping and rate steps enhanced in new wells or in existing of fracture stages be redesigned for Recommended additional reading at OnePetro: www.onepetro.org. Leonard Kalfayan, SPE, is a global production-engineering SPE 180274 Study of the Rock/Fluid adviser with the Hess Corporation in Houston. He has 35 years of Interactions of Sodium and Calcium Brines experience in the industry, working with a major operator and a With Ultratight Rock Surfaces and Their major pressure pumping company and as an independent con- Effect on Improving Oil Recovery by sultant before joining Hess in 2009. Kalfayan’s background is in Spontaneous Imbibition by M.K. Valluri, conventional and unconventional oil and gas, geothermal pro- Texas A&M University, et al. duction enhancement and stimulation, new-technology devel- SPE 169843 Estimating Long-Term Well opment and deployment, and business development. He was an Performance in the Montney Shale Gas SPE Distinguished Lecturer in 2005 and has served on numerous SPE program and Reservoir by Vu P. Dinh, Murphy Oil, et al. technical committees. Kalfayan is author or coauthor of more than 30 SPE and other IPTC 17739 A Comparison of North society publications, serves as a technical reviewer for SPE Production & Operations, American and International Risks and is coeditor of the SPE monograph Acidizing Fundamentals. He is a member of the in Unconventional Resource Plays JPT Editorial Committee. by D. Nathan Meehan, Baker Hughes

JPT • OCTOBER 2016 57 The Role of Induced Unpropped Fractures in Unconventional Oil and Gas Wells

he term induced unpropped (IU) 8,250 Tfractures refers to fractures created 2 million psi around the main propped fracture that are 8,200 3 million psi too small to accommodate any proppant. 4 million psi 5 million psi These could include natural fractures

-Direction (psi) -Direction 8,150

and microfractures induced along X bedding planes or along other planes of weakness. On the basis of production 8,100 data, diagnostic methods, and field observations, it is becoming increasingly 8,050 clear that IU fractures created during the hydraulic-fracturing operation play a critical role in determining the success of 8,000 fracture treatments. Magnitude in the Stress 7,950 Introduction 0 10 20 30 40 50 60 70 80 90 There has been a debate about whether Distance Along the X-Direction (m) certain shales are naturally fractured. Fig. 1—Effect of Young’s modulus on the stresses in the direction perpendicular The classic example of this is the Barnett to a penny-shaped fracture. The in-situ minimum horizontal stress used here is Shale. Many authors have suggested that 8,000 psi, Poisson’s ratio is 0.2, the desired maximum width of the fracture is the Barnett Shale is highly naturally frac- 1 cm, and fracture height is 100 m. tured, whereas others have argued that field observations in the Barnett can be Shale. In some instances, the boundary stresses can result in the tensile or shear explained on the basis of geologic litho- between the lithologically distinct layers failure of planes of weakness that may be facies and heterogeneity. Observations is sharp and well-defined, but, in other in- present in the rock. These stress-induced made on cores clearly indicate the pres- stances, it is diffuse and displays its own cracks can be referred to as IU fractures. ence of fractures. Mineralization on the characteristic lithologic gradation. Bound- For the purposes of this paper, IU frac- faces of these fractures indicates that aries between lithologically distinct layers tures are broadly defined as fractures that they are not drilling-induced fractures can often act as planes of weakness along have widths that are so small that they are but are instead native to the Barnett. Sim- which fractures can develop under stress. unable to accommodate any proppant, ilar questions and discussions have aris- and as such will close over time as the en about the role of natural and induced What Are IU Fractures? fluid pressure within them is decreased. fractures in other shale plays. Microfractures can also be created in The presence of these induced microfrac- One reason that it is often difficult to the rock by the mechanical stresses and tures is difficult to detect and prove be- resolve questions about natural fractures strains imposed on the rock fabric dur- cause it cannot be observed in cores or on in shales is the extreme level of heteroge- ing the process of creating the main hy- logs. Instead, one must infer the existence neity commonly observed in many types draulic fracture. Fig. 1 shows the magni- of these fractures indirectly by observa- of shale. Differences in lithology and the tude of the stress induced by the creation tions made through microseismic moni- highly laminated nature of many types of of the main fracture as a function of dis- toring, production history matching, and shale are apparent when viewing cores tance from the fracture for several dif- interference between wells. The follow- from source rocks such as the Eagle Ford ferent values of Young’s modulus. These ing subsections present five forms of evidence for the presence of these fractures.

This article, written by JPT Technology Editor Chris Carpenter, contains highlights The Fate of Injected Fracturing Water. of paper SPE 174946, “The Role of Induced Unpropped Fractures in Unconventional Thousands of gallons of water are pumped Oil and Gas Wells,” by M.M. Sharma and R. Manchanda, The University of Texas into each fracture stage during a fractur- at Austin, prepared for the 2015 SPE Annual Technical Conference and Exhibition, ing treatment. The fluid injected goes from Houston, 28–30 September. The paper has not been peer reviewed. the wellbore to the propped fracture and

For a limited time, the complete paper is free to SPE members at www.spe.org/jpt.

58 JPT • OCTOBER 2016 FACULTY & INSTRUCTOR POSITION leaks off into the reservoir. It is illustrative ing the process of fracturing. Not all of – Petroleum Engineering – to perform a simple volume balance on the these shear-failure events result in the injected fluid to estimate the amount of creation of fractures or even microfrac- fluid that is expected to go into each of the tures. However, slippage along planes of hydraulically connected regions. weakness does result in rupture of the On the basis of the simple material bal- existing rock fabric and the possibility of ance of the fracture fluid presented in the shear-induced dilation. Several authors complete paper, the created area of these have calculated the stresses in the vicini- The Dave C. Swalm School of Chemical IU fractures is expected to be at least an ty of hydraulic fractures. These measure- Engineering at Mississippi State University is order of magnitude larger than that of ments can help to identify regions that seeking individuals to fill one (1) tenure-track the main propped hydraulic fracture. are more probable to have shear failure position, at the rank of Assistant/Associate/ Another observation that supports this and, thus, estimate the regions in which Professor of Petroleum Engineering and hypothesis of a large area associated with microseismic events may occur. one (1) Instructor of Petroleum Engineering the network of IU fractures is the con- The density of microseismic events has position, with the appointments to begin sistent observation from field data that, sometimes been used as a surrogate for August 16, 2017. The School is especially during flowback, most of the water that is the density of microfractures that are interested in candidates with proven expertise in Reservoir Engineering, Reservoir injected is trapped in the formation and being created. Whether these microfrac- Simulation, and/or Production Engineering. is not produced back immediately. tures enhance permeability depends entirely on rock properties such as min- Applicants for the tenure-track position must have B.S. in Petroleum Engineering and a Ph.D. Production Response of Fractured eralogy and the in-situ stress state. In degree in Petroleum Engineering or related Wells. When reservoir simulations are reservoirs with a high stress contrast, field. The successful tenure-track candidate conducted to match the production re- more-planar fractures are observed, will develop an externally funded research sponse of horizontal wells completed while, in rocks under low stress contrast, program and demonstrate excellence in in ultralow-permeability rocks such as the microseismic pattern indicates more- teaching. Specific activities include teaching, shales, the production predicted from complex fracture patterns. research, service, and advising students in simulations is significantly lower than In most horizontal wells, microseismic the Petroleum Engineering program. that of the field-measured production events from one perforation cluster or Applicants for the Instructor position must when using the matrix permeability. This stage overlap with those from an adjacent have B.S. and M.S. degrees in Petroleum is not the case with higher-permeability stage. This implies that shear-induced Engineering or a related field. Petroleum rocks. Several authors have discussed the microfractures from one stage may hy- industry experience is particularly valuable concept of having a stimulated reservoir draulically interact with another stage or for Instructor applicants. Specific activities include teaching, supervising and advising volume (SRV) in the vicinity of the hy- another well. This is the case in many in- undergraduate students, and providing service draulic fractures and have analyzed the stances where pressure communication to the Petroleum Engineering program. effect of SRV on the observed production. between stages and between wells is com- The Dave C. Swalm School of Chemical This region of high permeability is a monly observed. This would not be possi- Engineering is housed in a state-of-the-art volume-averaged representation of a re- ble without high-permeability channels building dedicated to Chemical and Petroleum gion where the presence of induced micro- between wells that allow the pressure Engineering. The School currently enrolls over fractures has increased the effective per- pulses to propagate such large distances. 550 undergraduate and graduate students. meability of the rock. The precise density More information about the Dave C. Swalm of microcracks and their orientation will Tracer Data. Tracer data also strongly School of Chemical Engineering can be found depend on the rock fabric and the distri- suggest that pathways for fluid migra- at www.che.msstate.edu. Mississippi State bution of planes of weakness in the rock. tion (IU fractures) exist. The results of University is located in Starkville, Mississippi, A description of the rock at this level of de- previous studies that support this con- a vibrant college town located in Northeast Mississippi (visit.starkville.org). tail is usually not available when describ- clusion are presented in detail in the ing the region around the main fracture. complete paper. Applicants should apply via the MSU Clearly, the process of history matching employment website (www.msujobs.msstate. edu) and include a detailed vita, a statement production data is not unique and there Pressure Communication Between of research (tenure-track applicants only) may be other sets of parameters that may Wells. Several authors have presented and teaching interests, and the names and provide a comparable history match. The data showing the pressure response in a addresses of three references. Review of characteristic of high initial production well when an adjacent well is being frac- applications will begin on November 1, 2016, and a large initial decline rate is fairly tured. In some instances, pressure com- and continue until the position is filled. ABD typical of most shale production and can munication is observed not only in adja- may apply. Separate applications are required be matched best by postulating a region cent wells, but also in a well that is one for each position. of high permeability around the fracture. well removed from the well being frac- MSU is an equal opportunity employer, and all qualified applicants will receive consideration for employment tured and is more than 1,200 ft away. without regard to race, color, religion, ethnicity, sex Microseismic Data. There is a large body Clearly, the fracturing process is creating (including pregnancy and gender identity), national origin, disability status, age, sexual orientation, genetic of microseismic data that clearly shows information, protected veteran status, or any other the creation of shear-failure events dur- (Continued on page 65) characteristic protected by law. We always welcome nominations and applications from women, members of any minority group, and others who share our passion for building a diverse community that reflects the diversity in JPT • OCTOBER 2016 our student population. Injection in Shale: Experience on the Norwegian Continental Shelf

aste injection in shale, with ically favorable than shipping to shore rier for the vertical propagation of the Wmatrix permeability in the and onshore treating and disposal. hydraulic fractures created during injec- nanodarcy range and without the A survey of all CRI wells on the NCS tion. One certainty is that the injection of presence of any permeable layers, has noted that 110 had been or were operat- such high volumes cannot be performed been performed on the Norwegian ing, with approximately 30 of them still through a single discrete fracture, be- Continental Shelf (NCS) for more than 15 active at the time of the survey. This con- cause it would necessarily extend all the years. To avoid leakages to the seafloor firms the attractiveness of CRI wells as a way to the seafloor. using this method, techniques have been means of waste disposal, at least as far as The main opportunities, though, involve developed that allow wells to dispose of the NCS is concerned. Unfortunately, the the quantity and quality of monitoring sys- several million barrels into individual same survey also revealed that 14 of these tems used in the Ekofisk field, which are shale domains, with vertical propagation 110 wells had suffered from a leakage almost unique in the world. The systems of the disposal domain less than 1,000 ft to the seabed, thus illustrating the risks include frequent seafloor bathymetry, reg- above the injection point. Recently, use associated with CRI operations. ular logging of all wells, permanent down- of frequent 4D interpretations of seismic An outcome of the investigation was hole gauges, use of fiber optics along the surveys shot over a permanent sensor the discovery that a vast majority of the tubings, and permanent seismic arrays on array allowed detailed domain mapping leakage-to-seafloor cases occurred in wells the seafloor, allowing both passive seismic and independent dynamic monitoring. where shallow sand bodies were lacking or monitoring and active seismic shooting where their quality was very poor. This every 6 months. A side benefit of this ex- Historical Perspective justifies the general strategy traditional- ceptional array of techniques is the possi- Changes of North Sea regulations in the ly adopted by Norwegian operators (i.e., bility of detecting any anomaly caused by mid- to late 1990s made the seabed dis- avoiding the upward migration of hydrau- waste reinjection and the possible migra- posal of oily cuttings and other waste lic fractures with high-permeability sand tion of fractures toward the seafloor. from drilling and production impossi- bodies); this emphasizes the risks associ- The decision by the operator to inject ble without the use of significant top- ated with the lack of good-quality “blan- the waste into the Ekofisk overburden re- side cleaning systems. Meanwhile, the de- ket” sands above the injection point. sulted from a balanced approach using the velopment of fields required increasingly opportunities offered by the monitoring complex wells, resulting in the almost sys- Challenges and Opportunities systems to detect any abnormal vertical tematic use of oil-based mud for efficient for Waste Injection propagation of the waste-disposal domain drilling. This led to a dual challenge for in the Ekofisk Field Region and possible leakage that could have been the industry, with an increased amount of The Ekofisk field is situated in the south- triggered by the lack of a good-quality waste generated and the difficulty of dis- ern region of the Norwegian North Sea, sand layer in the Ekofisk overburden for- charging it. Options were then reviewed close to the British and Danish sectors in mation. The environmental and econom- between shipping onshore and treatment an area where the Utsira sand is not pres- ic benefit of this balanced approach has or reinjection into the subsurface with ent. The main challenges are the amount been demonstrated clearly by the more- cuttings-reinjection (CRI) wells (a typical of waste generated every year at a regular than-15-year history of cuttings injection design of a CRI well in the Ekofisk Field rate (i.e., approximately 850,000 bbl/yr into the overburden in six wells in the region is outlined in the complete paper). to be injected) and the absence of any greater Ekofisk area with no abnormal up- Most operators concluded that the CRI significant permeable layer in the over- ward migration of the waste-injection do- option was much safer and more econom- burden of the field that can act as a bar- main, with no leakage to the seafloor, and with the capacity of some wells to accommodate up to 5 million bbl of waste fluids. This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 170851, “Injection in Shale: Review of 15 Years of Experience on the Round-the-Clock Monitoring Norwegian Continental Shelf and Implications for the Stimulation of Unconventional and Real-Time Analysis Reservoirs,” by F.J. Santarelli and F. Sanfilippo, Geomec, and R.W. James, The analysis of past leakage cases shows H.H. Nielsen, M. Fidan, and G. Aamodt, ConocoPhillips, prepared for the 2014 SPE that data acquisition alone is insufficient Annual Technical Conference and Exhibition, Amsterdam, 27–29 October. The paper to ensure injection confinement and the has not been peer reviewed. avoidance of surface leakages. Conse-

For a limited time, the complete paper is free to SPE members at www.spe.org/jpt.

60 JPT • OCTOBER 2016 Sources:

Operator Integrated Control Historian and Safety Alarm Systems Visualizer and Interpreter Process-Control Operator Collector Network Historian Reciproctor of Operators’ Transformer Fiscal Metering Network Historians Platform Special Collaborator Monitoring A

Sensors Store Platform B

Fig. 1—Architecture of the system developed during the JIP on injection safety to ensure round-the-clock monitoring and real-time analysis of injection data. quently, a joint-industry project (JIP) on tical extension of the disposal domain. tic downhole conditions was measured to injection safety was undertaken to devel- The systematic extraction of information be between 10–9 and 10–8 darcies. op a software system capable of perform- from all injection cycles has allowed quan- 3. Another important observation is ing automatic round-the-clock monitor- tification of three important features. that the pressure inside the fracture net- ing and real-time analysis of injection data 1. Injection of volumes of significant work connected to the well is far in ex- for any given well. The architecture of this magnitude has been achieved by creating cess of the overburden stress and that the system is shown in Fig. 1. a multitude of fractures in the disposal fracture network does not show any sign The system has two key functions. One domain instead of having a discrete num- of extension after the well has been shut is to send an alarm or “flag” in case of ber of fractures accommodating the in- in, despite this very large net pressure. anomalies in the injection process, and the jected fluid. From a physical perspective, two mech- other is to extract as much information as 2. The secondary permeability in the anisms acting simultaneously can be iden- possible from the injection records. Both disposal domain was repeatedly measured tified: One impairs the propagation of the functions are performed simultaneously to be in the 0.1-darcy range, while the ma- and benefit greatly from each other (e.g., trix permeability of the shale under realis- (Continued on page 65) the real-time extraction of the reservoir pressure in the injection domain allows round-the-clock calculation of the injec- tivity index and the detection of abnormal increases, which could be indicative of upward migration and eventual leakage). The round-the-clock monitoring-and- automatic-alarm system fulfills three main roles, corresponding to three levels of alarms. The first level simply ensures that all SPE INTERNATIONAL HEAVY OIL sensors are active when the injection pro- CONFERENCES & EXHIBITIONS cess occurs and that adequate measurements are recovered effectively. The second level ensures that the procedures Mark Your Calendar! Join E&P leaders in diferent parts of the world to learn carefully established by the operator are about regional and global best practices in heavy oil. followed effectively by the personnel performing the job. The third level detects 19–20 October 2016 | Lima, Peru and reports anomalies in the performance SPE Latin America and Caribbean Heavy of the CRI well, which could correspond to and Extra Heavy Oil Conference a loss of injection efficiency or could mean www.spe.org/go/HeavyOilLatin an upward migration of the disposal do- 6–8 December 2016 | Kuwait City, Kuwait main, leading to leakage. SPE International Heavy Oil Conference & Exhibition www.spe.org/go/HeavyOilSPE Mechanisms It has been possible to dispose of millions 15–16 February 2017 | Calgary, Canada of barrels of waste fluids in a limited vol- SPE Canada Heavy Oil Technical Conference ume of shale through a 10-ft-long perfo- www.spe.org/go/HeavyOilCanada ration interval without any undue ver-

JPT • OCTOBER 2016 61 Model for a Shale-Gas Formation With Salt-Sealed Natural Fractures

ost multiple-transverse-fracture to the material-balance model that the secondary fracture system acts as a Mhorizontal wells in shale-gas authors have developed. These phenom- seal and as a barrier to flow. formations remain in transient bilinear ena include the imbibition and entrap- ◗ Where sodium chloride is the or linear flow for very long periods. ment of injected fracturing fluid, the in- mineral in the secondary fracture However, there are often reported crease in salinity of the fluid in contact system, hereafter it will be referred cases of shale wells that exhibit with the shale, and the deterioration of to as salt, with a fixed molar boundary-dominated flow in a very shale samples. density of 1,049 mol/ft3. short period, which implies a stimulated Several researchers have studied the ◗ As the main fracture opens under rock volume (SRV) much smaller than relationship between the low load recov- tensile failure, secondary fracture would be expected. This paper offers ery of hydraulic-fracturing fluid and the systems activate. an alternative explanation for the early elevated salinity of the fluid that does re- ◗ Apart from imbibition, any leakoff boundary-dominated flow related turn. Many authors have proposed rela- during injection is only into the to dissolution of salt-sealed natural tionships between the surface area of the natural-fracture system. fractures in the shale. formation exposed to fracturing fluid, ◗ Injected water in the water-based the characteristics of the fluid, the wet- fracturing fluid dissolves the salt Introduction tability of the rock, and the distribution encountered through leakoff into Operators producing gas from the or characteristics of the natural-fracture sheared mineralized fractures Haynesville, Marcellus, and Horn River system. These findings are examined in (equations for determination of shale formations have observed that pro- detail in the complete paper. compact dissolution are provided duced water is more saline than the in- in the complete paper). jected fracturing fluid. Additionally, the Material-Balance Model ◗ Water in shear cracks of sufficient fraction of injected water that flows back for Injection, Flowback, volume to dissolve the salt creates when the well is put on production— and Production a productive secondary fracture termed load recovery—is low. The authors base a material-balance system approximated as a dual- This study investigates the possibil- model on assumptions chosen so that the porosity continuum, described ity that salts removed by flowing back model applies to any formation that has as fractures surrounding matrix injected low-salinity water-based frac- a dissolvable mineral volume, measur- cubes. turing fluid may be a mechanism for in- able salt concentration in the flowback, ◗ The authors model the creasing the effective permeability in inferred distribution of mineral in the heterogeneous shale volume as shale-gas wells within a limited volume formation, and inferred geometry of hy- homogeneous, with averaged surrounding the created hydraulic frac- draulic fracture or SRV. Several of these inputs. tures. The study offers strong evidence assumptions are listed as follows: ◗ Water in shear cracks of that salt may be a key factor in productiv- ◗ During a single-stage hydraulic- insufficient volume to dissolve the ity and ultimate gas recovery from shale- fracture treatment, the main salt fails to mobilize gas. gas wells. fracture encounters a mineralized ◗ Water salinity increases in the secondary fracture system. The main hydraulic fracture through Interactions Between secondary fracture system is diffusion from the secondary Fluids, Salt, and Shale represented as fracture porosity. fracture system. Several fracturing-fluid/formation in- ◗ Before the hydraulic-fracture ◗ Leakoff volume is equal to the teractive phenomena relate strongly treatment, the mineral in the volume remaining in the productive secondary fracture system plus the volume remaining in mineralized This article, written by JPT Technology Editor Chris Carpenter, contains highlights cracks plus the volume imbibed of paper SPE 175061, “Model for a Shale-Gas Formation With Salt-Sealed Natural into the shale matrix. Fractures,” by Hoagie Merry and C.A. Ehlig-Economides, University of Houston, ◗ Initial flowback water is produced and Pang Wei, Sinopec Research Institute of Petroleum Engineering, prepared for the from the primary and secondary 2015 SPE Annual Technical Conference and Exhibition, Houston, 28–30 September. fracture systems by diffusion The paper has not been peer reviewed. and displacement in response

For a limited time, the complete paper is free to SPE members at www.spe.org/jpt.

62 JPT • OCTOBER 2016 EN_AFENav_ad_JPT_Marionette_2015_09_c01New.pdf 1 2016-02-11 10:50 AM

CONTROL COSTS AND SPENDING. BEFORE THEY CONTROL YOU.

Now is the time to optimize Authorization for Expenditure processes in order to drive efficiency, manage spending, and protect your margins.

Budgets today are tight with extra scrutiny on

C every expenditure. You need a process that can

M withstand that scrutiny, justify your spending and provide visibility. AFE Navigator® is that process. Y Eliminate lost AFEs. Drastically speed up approvals. CM Cut waste so you can capitalize on more MY opportunities. Take control with a secure, all-in-one, CY mobile-friendly AFE workflow and capital tracking CMY system used by more than 100 upstream and K midstream oil and gas companies.

Protect your margins with AFE Nav.

Visit energynavigator.com/optimize or call Lynn Babec at 1-866-856-9544. to a pressure gradient and gas trix into which it imbibes. This gas can versatility of the model are provided in production, respectively. contribute to elevated gas rate in early the complete paper. ◗ Imbibed water and residual water time, in part because of gravity segrega- saturation in the propped fracture tion and the very high mobility ratio of Discussion and in productive secondary gas to water. The hypothetical considerations in fractures produce slowly in Because the gas is replaced in the ma- this work are consistent with many di- solution with gas. trix by water that imbibes, this volume verse investigations related to salt. An ◗ Newly productive secondary contributes to the displacement of water important implication is that fractur- fractures are distributed evenly that is in the DSRV. ing with a fluid that would not dissolve along the fracture plane. When the fracture has closed before the mineralized salt would not enable ◗ Dissolution of salt in shear cracks the onset of well production, if no im- connection between previously min- is slow compared with the time bibition occurs, the volume that was eralized fractures and the matrix and taken to flow into the cracks, and occupied by salt and is exposed to sur- would not enhance the effective perme- fast compared with the shut-in time faces open to imbibition has a maxi- ability of the shale. Despite its poten- before the onset of production to mum of 0.2 times the leakoff volume tial scarcity, in this scenario, essentially the well. from the main fracture. The saturated fresh water is the best fracturing fluid. ◗ The time required for water to volume leaked off into shear fractures This work also suggests that refractur- imbibe into shale is fast compared outside the DSRV does not contribute ing even without proppant may serve with the time before the onset of to dissolution of salt and cannot imbibe to restore lost permeability and extend production to the well. because it is not in contact with an im- the DSRV. ◗ The bulk volume of rock containing bibing surface. The production mecha- newly productive secondary nism for the water outside the DSRV is Conclusions fractures has an effective only pressure diffusion and isothermal The objective of this research was permeability much larger than the expansion. Saturated water flowback to investigate the possibility that shale-matrix permeability. from outside the DSRV can contribute freshwater-based fracturing fluid inject- These assumptions allow adjustable no more than approximately 4% of the ed into a shale formation could increase user-defined inputs in order to match leakoff volume. the permeability in a limited volume by field data. Gas will displace leakoff fluid from dissolving salt in an existing natural- the main fracture and secondary frac- fracture system. Shear-Fracture-Network Dissolution. tures down to the residual water satura- The authors developed a material- Complete dissolution of salt in a sheared tions in the main hydraulic fracture and balance model that estimates the DSRV fracture enables the fracturing fluid to in the secondary fracture system inside defined by the volume of shale gas ren- imbibe into the shale matrix and releas- the DSRV. dered productive by injected water es gas trapped in the shale matrix to flow Remaining residual water will then be dissolving salt found in mineralized into the newly developed effective per- vaporized with water/gas-ratio plot de- fractures sheared during hydraulic frac- meability. However, complete salt dis- cline as the reciprocal of the cumulative turing. The model presumes that pro- solution requires an excess of leakoff produced gas. If imbibition occurs, the duced gas will initially displace brine fluid that will flow into sheared fractures ultimate load recovery will be higher be- from propped and secondary fractures but will not provide connection to gas in cause less of the leakoff fluid will flow but that continued gas production will the shale matrix. A microseismic survey past the DSRV. vaporize water from the remaining could give an impression of an apparent Regardless of whether imbibition oc- salt-saturated residual brine, deposit- SRV larger than the bulk shale volume, curs, the initial load recovery is a dis- ing salt that could explain observed loss with enhanced effective permeability placement process, and the ultimate load of shale effective permeability previ- caused by salt dissolution. The authors recovery through water vaporization is ously modeled as pressure-dependent apply the term dissolved stimulated rock slow and will be produced in solution permeability. The authors expect that, volume (DSRV) to the latter volume. with the gas at much lower rates. As the eventually, flowback-water salinity will Subsections of the complete paper pro- residual water is produced from the main drop to near zero. vide expressions for the determination fracture and secondary fracture system, A field application performed on the of DSRV, fracture porosity, and imbibi- inferences can be made as to the change basis of data from Haynesville Shale tion volume. in permeability that may take place. wells shows that the developed material- Conversely, as water is produced in balance model estimated a bulk DSRV Flowback and Production. The free- solution with the gas, the dissolved that could be consistent with observed gas flow into hydraulic fractures before salt that was suspended in the pore gas-production behavior exhibiting production can be described as coun- water is left behind; this salt may act boundary-dominated flow after a tercurrent imbibition. By this mecha- to reduce the porosity and, hence, the short period on production and that nism, the volume of water imbibed from absolute permeability. the modeled water-trapping mecha- the main hydraulic fracture displaces The results of an example field case nisms are consistent with observed low an equivalent gas volume from the ma- to demonstrate the applicability and the load recovery. JPT

64 JPT • OCTOBER 2016 The Role of Induced . . . Injection in Shale . . . (Continued from page 59) (Continued from page 61) a network of fractures that extends primary fracture, and the other favors ƕƏƏķƏƏƏ[ through the failure of planes of weakness the opening of secondary fractures. of Sand Screens and can propagate a long distance. The In the case of very-low-permeability proppant cannot take this tortuous path- shale such as that above the Ekofisk chalk installed globally way and is only placed in the main frac- hydrocarbon reservoir, the important ture, leading to the creation of rather long issue is what is happening inside the frac- and complex patterns of IU fractures that ture ahead of the fluid front. The matrix temporarily provide hydraulic conductiv- permeability value makes it obvious that With 100% ity over long distances. no fluid from the rock matrix can be dis- An equally important observation that placed to fill the volume created by the track record, is made in this and other studies is that newly formed fracture ahead of the inject- be part of the this pressure communication does not ap- ed fluid. Consequently, this fracture vol- pear to last much beyond the time when ume ahead of the fluid becomes filled with ELITE the fracture is being pumped. For exam- vapor. This phenomenon is known as cavi- ple, the production response of these wells tation. It corresponds to the partial dehy- with Tendeka’s does not indicate any production interfer- dration of the shale and leads to mechani- ence for the first few years of production. cal swelling when the dehydrated shale is new premium This suggests that the microfractures that contacted by water. were created close over a period of time Here, it must be emphasized that screen range because they did not have any proppant mechanical-swelling stresses can often to keep them open permanently. IU frac- depend on the level of dehydration of the tures are the only reasonable explanation shale and are mainly independent of shale for the pressure communication between mineralogy. In turn, this swelling stress wells that then disappears over time. applies not only at the tip of the fracture SND $!L but also over its entire length from pre- Implications of the Presence of IU vious sudden propagation events and can Fractures. Use of Smaller-Diameter be viewed as a stress barrier added to the Proppant. As mentioned previously, the minimum in-situ stress. This added swell- surface area of IU fractures can be at least ing stress explains why the pressure re- an order of magnitude greater than that quired to propagate a hydraulic fracture in $;m7;h-ruo%b7;v-u-m];o=voѴ'ঞomv of the main propped fracture. This im- shale can be much higher than the mini- =ouv-m7=-1;1olrѴ;ঞomv)b|_uo0'v| v-m7v1u;;morঞomv*$_;m;)|o plies that, if it is possible to keep these mum in-situ stress and why no fracture l-uh;|Eliteu-m];bv0-v;7om|_; fractures open by some means, one may propagation was ever observed during _b]_Ѵ+v'11;vv='Ѵ&Ѵ|u-ru;lb'l v1u;;mu-m];=ou;,|u;l;v;u%b1; be able to achieve a significant increase in the shut-in periods of the CRI wells. The -rrѴb1-ঞomv*&vbm]|_;v-l;C%; well productivity. One way to achieve this swelling stresses developed during the Ѵ-+;u1omv|u'1ঞom=ouv'r;ubou rѴ']]bm]u;vbv|-m1;|_; Ѵb|;_-v0;;m is to use smaller-sized proppant so that fracture-propagation process contribute ;m]bm;;u;7bmu;vromv;|o|o7-+Ľv it may have a better opportunity to enter to the mechanical equilibrium of the sys- Cm-m1b-Ѵ1omv|u-bm|v|oruo%b7;-1ov| the microfractures that would otherwise tem, and the absence of propagation after ;@;1ঞ%;-m7uo0'v|v1u;;mvoѴ'ঞom v'b|;7|olov|C;Ѵ7-rrѴb1-ঞomv* close during flowback. well shut-in is further proof that the sys- Use of Acid or Other Chemicals To tem is at a mechanical equilibrium during $;m7;h-Ľv1olrѴ;ঞom7;vb]mv)bѴѴ ruo%b7;-lou;;@;1ঞ%;-m7;L1b;m| Etch the Surface of the IU Fractures. shut-in periods, despite the presence of a )-+|o0o|_orঞlb/;bmYo)-m7 Another possible strategy would be to use high net pressure in the fracture network. lomb|ou+o'uu;v;u%obuķ|o7;Ѵb%;u dilute mineral acid to etch the surface of Another mechanism governing the cre- ;m_-m1;7);ѴѴr;u=oul-m1;* the IU fractures to provide some degree ation of the multiple-fracture network ob- of surface roughness to the IU fractures. served around the CRI wells in the Ekofisk Choosing Fracture Spacing and Well field is the thermal effect. The fluid inject- Spacing. Both well spacing and produced during the CRI operations in Ekofisk is )))*|;m7;h-*1ol tion have an important effect on the frac- colder than the formations where the in- tures that can be created in the infill well jection takes place. Under typical condi- between two wells. Fracture spacing, well tions, the fluid is injected at 70 to 85°F for spacing, and production time of the outer a virgin formation temperature of 175°F, wells could be optimized with the help of corresponding to a cooling of approxi- geomechanical simulations to create fa- mately 100°F. This means that the forma- vorable conditions for creating networks tions in contact with the injected fluid will of IU fractures. JPT be cooled. JPT ! "&$"!""+&!! " !(!

JPT • OCTOBER 2016 TECHNOLOGY FOCUS

Petroleum Data Analytics Luigi Saputelli, SPE, Senior Reservoir Engineering Adviser, ADNOC, and Frontender Corporation

One of the many challenges we face today sensor availability, and engineering mod- tics (regression, time series, and factor in the petroleum industry is the manage- els has promoted the exponential growth analysis); pattern recognition (Markov ment of data and information. In some of data types and volumes. Data-driven models, principal components, ensemble instances, we are overwhelmed by the techniques also have diversified and averaging, classification, and regression); amount and diversity of formats, and, improved to address such incremental business intelligence (key-performance- in other cases, we are blinded from the complexities. We are now referring to the indicator dashboards, multidimension- right information to understand a pro- professionals who manage and find value al visualization); artificial intelligence for cess (What has happened?), to predict the from data as “data scientists,” and we planning, creativity, perception, and social immediate future (What could happen?), are calling the management of large and intelligence (knowledge representation, or to make proper decisions (What should complex data volumes “big data.” neural networks, support vector machines, we do?). The answer to these questions Data analytics, either big or small, is Bayesian inference, decision tree, natural- is data analytics supporting appropriate the collection of tools that leverages data language programming); machine learn- engineering and management judgment collection, aggregation, processing, and ing (inductive logic programming, rule and the modeling of actual energy sce- analysis for describing insights into the learning, and clustering); and manage- narios. Data analytics for strategic deci- past, predicting future performance, ment of large data sets, distributed and sion making is being constantly devel- and prescribing actions from the opti- parallel computing, cloud computing ser- oped to mitigate low-oil-price scenarios. mization of possible outcomes. Current vices, and data cleansing and profiling. For many decades, our technical and trends of data analytics differ from tradi- A graduate degree may be required to business processes have benefited from tional statistics in the sense that the new master some of the techniques around the wide use of data statistics for deci- data-driven predictive and prescriptive data analytics, and decades may be sion making. In many instances, pre- models go beyond data averaging, out- required to adopt them across the indus- dicting and prescribing have relied more lier detection, correlations, and multiple- try, but it is also true that many of these on data evidences and trends than on parameter regression fitting. techniques are evolving at such a fast first-principle simulation models. The Data-analytics tools may include one pace that they become obsolete by the advancement of computational power, or more of the following groups: statis- time we plan to roll out a trial pilot. We need to learn how to experiment with, implement, and capture results from Luigi Saputelli, SPE, is a senior reservoir-engineering adviser data analytics faster than ever. We either with ADNOC. During the past 25 years, he has held various posi- evolve quickly or disappear. JPT tions as reservoir engineer, drilling engineer, and production engineer. Saputelli previously worked for 3 years with Hess Corporation, for 5 years with Halliburton, and for 11 years with Petróleos de Venezuela. He is a founding member of the SPE Recommended additional reading Petroleum Data-Driven Analytics technical section and recipient at OnePetro: www.onepetro.org. of the 2015 SPE International Production and Operations Award. SPE 176791 Use of Regression and Saputelli has authored or coauthored more than 70 technical publications in the Bootstrapping in Drilling Inference areas of digital oil field, reservoir management, reservoir engineering, real-time opti- and Prediction by Chiranth M. Hegde, mization, and production operations. He holds a BS degree in electronic engineering The University of Texas at Austin, et al. from Universidad Simon Bolívar, an MS degree in petroleum engineering from SPE 174985 Topological Data Analysis Imperial College London, and a PhD degree in chemical engineering from the To Solve Big-Data Problem in Reservoir University of Houston. Saputelli serves on the JPT Editorial Committee, the SPE Engineering: Application to Inverted Production and Operations Advisory Committee, and the Reservoir Description and 4D-Seismic Data by Abdulhamed Alfaleh, Dynamics Digital Oil Field subcommittee. He has served as reviewer for SPE Journal Saudi Aramco, et al. and SPE Reservoir Evaluation & Engineering and as an associate editor for SPE SPE 179958 Detecting and Removing Economics & Management. Saputelli also serves as managing partner at Frontender, Outliers in Production Data To Enhance a petroleum engineering services firm based in Houston. He can be reached at Production Forecasting by Nitinkumar L. [email protected]. Chaudhary, University of Houston, et al.

66 JPT • OCTOBER 2016 Functional Approach to Data Mining, Forecasting, and Uncertainty Quantification

he difficulty in applying ries. This is often challenging because or concentration of oxygen or plankton Ttraditional reservoir-simulation production time series come as noisy, as a function of ocean depth. FDA starts and -modeling techniques for discrete observations of production from discrete measurements, regularly unconventional-reservoir forecasting rates over time. Conventional approach- or irregularly sampled over the time do- makes the use of statistical and modern es to this problem rely on parameteriz- main by assuming they came from some machine-learning techniques a relevant ing the system with decline curves and smooth process corrupted by noise. proposition. However, the most current work in the parameter space of the as- The first objective of FDA is to trans- applications of these techniques often sumed decline model, or, even simpler, form discrete measurements into a ignore the systematic time variations work with the raw data. This paper takes continuous approximation of the true in production-decline rates. This paper an alternative nonparametric approach underlying function. This is routinely proposes a nonparametric statistical wherein the data are used to find the accomplished with a basis expansion. approach, using a modern technique most-appropriate smooth and continu- In this basis expansion, raw measure- termed functional data analysis ous representation of declining produc- ments of functional data are approxi- (FDA). In FDA, production data are tion time series. The approach for this mated with a scaled sum of smooth an- modeled as a time series composed of nonparametric form relies on the statis- alytical basis functions that span the a sum of weighted smooth analytical tical discipline termed FDA. same time domain as the original data. basis functions. FDA allows for the exploration of sto- Once the basis system is selected, it is chastic variation in functional data and necessary to find appropriate values for Introduction construction of low-dimensional rep- the scaling coefficients. This is accom- Many companies have adopted a so- resentations of time series. In the con- plished with an objective to minimize called “data-centric process” for un- text of shale production, these low- the mean squared error of the fit. derstanding and forecasting in dimensional representations will enable In general, estimation of underlying unconventional reservoirs. This data- a better understanding of relationships functions has two competing objectives: centric process comes as a consequence between the wells and, in conjunction to match observations as accurately as of the shortcomings of conventional with a distance-based generalized sen- possible and to avoid overfitting, which reservoir-data-analysis and -modeling sitivity analysis (DGSA), identification results in fits that are too wiggly. In approaches, which mostly belong to the of the most-influential completion and order to avoid overfitting and secure preshale era. Either the huge quanti- reservoir parameters. Additionally, smooth variation of the fitted function ty of data collected in shales cannot FDA enables formulation of a forecast- across neighboring observations, an ad- be used fully with conventional mod- ing framework with high-dimensional ditional roughness penalty is often in- eling techniques or the rapid nature Kriging-based regression, to produce troduced into the fitting criteria. The of shale development simply does not best-guess estimates of the entire pro- simplest measure of a function’s rough- allow for lengthy reservoir-modeling duction profiles for new well locations. ness is given by its integrated squared and - simulation studies. second derivative. Decision variables in shales are rates FDA or volumes of produced hydrocarbons. FDA is a statistical discipline that focus- Functional-Principal- Therefore, understanding shales and es on data that can be considered to be Component Analysis (FPCA) identifying value-creating practices by infinitely dimensional, such as oil and After basis-expansion fitting, FDA al- use of data-driven techniques require gas production curves over time, daily lows for exploration of variations in proper handling of production-time se- temperature measurements over time, production profiles and construction of low-dimensional spaces for data visualization. The main tool for this type This article, written by Special Publications Editor Adam Wilson, contains highlights of analysis is FPCA. FPCA is a tech- of paper SPE 174849, “Functional Approach to Data Mining, Forecasting, and nique for data transformation and di- Uncertainty Quantification in Unconventional Reservoirs,” by Ognjen Grujic, mensionality reduction that is similar to Stanford University; Carla Da Silva, Anadarko Petroleum; and Jef Caers, Stanford the well-known multivariate principal- University, prepared for the 2015 SPE Annual Technical Conference and Exhibition, component analysis (PCA). The main Houston, 28–30 September. The paper has not been peer reviewed. difference between functional and

For a limited time, the complete paper is free to SPE members at www.spe.org/jpt.

JPT • OCTOBER 2016 67 multivariate PCA lies in the way prin- is sensitive to the analyzed parameter. from a difficult rate-vs.-time forecast- cipal components are represented. In DGSA was used to calculate the numeri- ing to a simple problem of forecasting FPCA, principal components are func- cal values (magnitudes) of relative pa- functional principal-component scores, tions, which span the same time do- rameter importance. which are scalars. main as the original functional data. The The main idea behind DGSA is that authors observed that more than 95% influential input parameters to some Kriging of variance, within a given functional- transfer function would separate result- An alternative approach that is capa- data set, is fully described by the first ing responses into discrete categories ble of exactly reproducing training data two principal components. This also or clusters. In the original development, (up to measurement error) and is less means that, in all subsequent analyses, DGSA was focused on computing param- sensitive to outliers is Kriging, a tech- the first two principal components can eter sensitivities in Earth modeling with nique extensively used in geostatistics. be worked with alone without much loss flow simulations where flow responses Kriging falls into a bin of interpola- of information. were transformed with multidimension- tion methods that is local in nature, al scaling, producing low-dimensional given that it weights training data on Functional Sensitivity Analysis plots. In such low-dimensional space, the basis of their distances from an es- Because similar production profiles clustering was performed and cumula- timated point in the input space. Even plot close to one another in a low- tive distribution functions (CDFs) of pa- though it was originally developed for dimensional score plot, a visualiza- rameters associated with models in each low-dimensional problems, modern tion of what parameters (geological or cluster were constructed. If a parameter adaptations in metamodeling for com- completion) affect production is now were important, its cluster-based CDFs puter experiments efficiently general- possible. This is accomplished by col- would be well-separated (meaning de- ize Kriging to high-dimensional spaces. oring points in the score plot by pa- viated from the overall CDF), and, if it The idea here is to perform linear Krig- rameter values and observing whether were not important, then the converse ing in the space constructed from the that coloring varies systematically with- would occur. input parameters (geological and com- in the plot. The authors argue that such In order to rank parameters from the pletion parameters) and to interpolate/ systematic variation, or presence of a most to the least sensitive, the L1 norm Krige production decline for any new trend, indicates whether production is computed between class-specific set of geological and completion param- CDFs and the preceding CDF. The re- eters—for example, for new well loca- sult is averaged and standardized with tions, where geological parameters are the P95 bootstrapped L1 norm from the estimated or measured and completion same distribution. The averaging and parameters are optimized. One Stop for standardization enable construction of JPT tornado charts with parameter sensitiv- Conclusions Everything ities and parameter ranking to aid bet- This paper introduces a functional Get all your online JPT ter interpretation of the data. approach for analysis and forecast- content in one place at ing of unconventional-reservoir data. www.spe.org/jpt Reconstructing Production It shows how functional basis expan- Profiles With FPCA sion can be used to represent produc- Responsive Design FPCA also allows for production-profile tion decline in nonparametric fashion SPE members can access reconstruction and dimensionality re- and without assumptions about the un- the latest issue of JPT duction. To reconstruct one production derlying physical mechanisms or rely- from any of their devices. profile, one has to sum products of prin- ing on previously formulated decline Optimized for desktop, cipal components and associated scores models. It also demonstrates how to tablet, and phone, JPT is and add that sum to the mean function transform smoothed production pro- easy to read and browse computed on the entire ensemble. files into a low-dimensional functional anytime you are online. principal-component score plot to aid Forecasting Framework data mining and identification of the Assuming that the mean function and most-influential parameters. functional principal components are This work deals with single variate Offline Access reasonably estimated from existing wells reservoir-flow responses (oil rates), (training set), all that is needed to fore- whereas wells in the used data set pro- Download PDF versions cast production at some new well location duced oil, water, and gas. Real busi- of 180+ issues dating are the functional principal-component ness and engineering applications re- back to 1997 for reading scores. In essence, this means that, by quire joint analysis and forecasting of online or when an performing basis expansion and func- these functional quantities, and such Internet connection is tional principal-component analysis complex multivariate data-mining and not available. on the training data, the production- forecasting problems will be the focus of forecasting problem is transformed future work. JPT

68 JPT • OCTOBER 2016 Mitigating Drilling Dysfunction With a Drilling Advisory System

he ExxonMobil Drilling Advisory using MSE trends to maximize drilling Often, MSE data are noisy and TSystem (DAS) is a rig-based performance were an effective means their presentation on a doghouse drilling-surveillance and -optimization of optimizing the drilling process. computer can be limited to strip platform that encourages regular The basic idea was to use MSE trends charts displayed over a fixed drilloff tests, carefully monitors to identify and respond to drilling- interval of time. drilling performance, and provides system-performance limiters such as ◗ Correctly identifying formation recommendations for controllable bottomhole-assembly (BHA) whirl, changes is key to a successful drilling parameters to help improve bit balling, and stick/slip, with the ul- interpretation of MSE trends. the overall drilling process. Key timate goal of producing consistently ◗ Human resources on the rig technical components of the DAS better and longer bit runs. The initial need to be prioritized, and other include high-quality adaptive program was deemed highly success- activities often take precedence, filtering algorithms for computing ful and since has become a key compo- resulting in the need for an various drilling-system-performance nent of the rate-of-penetration (ROP) automated drilling-performance- measures, data-encapsulation/ - optimization process. monitoring system. dimensionality-reduction techniques, Although MSE-surveillance activities These challenges can become com- and reduced-order-modeling hold great potential for identifying and pounded when a real-time MSE- techniques that are ultimately used avoiding drilling-performance limiters, surveillance program is combined with to generate recommendations for the following challenges often arise in additional drilling-surveillance work improved performance. the implementation of a rig-based MSE- flows that leverage downhole or sur- surveillance work flow: face drilling data to further assess the Introduction ◗ Frequent drilloff tests are needed dynamical behavior of the drilling as- Approximately a decade ago, the com- to maximize the benefit of the sembly. An example of such a work pany initiated an internal program to method. MSE surveillance requires flow that has been used by the opera- pursue real-time rigsite mechanical- almost constant interrogation tor is real-time bit stick/slip surveil- specific-energy (MSE) surveillance. of the weight-on-bit (WOB) lance monitoring, which enables virtual MSE, a quantity initially developed and and revolutions/minute (RPM) real-time observation of stick/slip se- applied in the mining industry and sub- parameter space to collect the MSE verity on the basis of surface torque and sequently applied to oilwell drilling, pro- data required for trending analysis. RPM and a drilling-mechanics model of vides a measure of the energy required For statistically meaningful the drilling assembly. Reacting to sev- for a drilling assembly to drill through trending analysis, it is important eral interdependent (and sometimes an interval of subterranean formation; to conduct drilloff tests over conflicting) performance metrics (MSE, and underefficient drilling conditions sufficiently long depth intervals for ROP, stick/slip severity) in real time is a correlate with the compressive strength a given WOB/RPM set point. significant challenge. of the rock being drilled. It was quickly ◗ MSE data must be interpreted To address many of the challenges demonstrated that frequent drilloff tests correctly to identify trends. and to ensure a more-consistent application of MSE-surveillance (and other This article, written by Special Publications Editor Adam Wilson, contains highlights surveillance) work flows, the operator of paper IPTC 18333, “Mitigating Drilling Dysfunction With a Drilling Advisory System: has developed a real-time DAS. DAS is a Results From Recent Field Applications,” by Gregory S. Payette, SPE, Darren Pais, real-time rig-centric software platform SPE, Benjamin Spivey, SPE, and Lei Wang, SPE, ExxonMobil Upstream Research that is designed to interact with existing infrastructure available at a typical Company; Jeffrey R. Bailey, SPE, and Paul Pastusek, SPE, ExxonMobil Development rigsite, such as the wellsite information Company; and Michael Owens, XTO Energy, prepared for the 2015 International transfer specification (WITS). The DAS Petroleum Technology Conference, Doha, Qatar, 7–9 December. The paper has not been system collects real-time drilling data, peer reviewed. performs various levels of data processing, computes measures of drilling per- Copyright 2015 International Petroleum Technology Conference. Reproduced by formance, and provides operational rec- permission. ommendations designed to assess the

For a limited time, the complete paper is free to SPE members at www.spe.org/jpt.

JPT • OCTOBER 2016 69 ◗ Optimization Aggregating input and drilling- 90 100 performance-metric data over 80 Zoom 95 substantial periods of time and 70 90 depth into response points for the 60 85 50 purpose of obtaining characteristic

RPM 80 40 75 response over average formation 30 70 intervals. 20 65 ◗ Upscaling the overall system 18 20 22 24 26 28 30 32 WOB (1,000 lbf) response using collections of

Default response points, fitted response V2 DAS 35.8 ft surfaces, and optimization 38 31 algorithms such as the simplex, 36 Vr 30 34 29 steepest-descent, and Newton 32 HOLE SECTION 30 28 8.75 in. *6: AutoTrak - LS methods, ultimately to generate 28 27 WOB (1,000 lbf) WOB MSE (1,000 psi) 160 200 the DAS recommendations for 150 Ts 150 operating parameters such as 140 100 System 130 50 WOB and RPM. RPM Play ROP (ft/hr) ROP 120 0 Back A key constraint in developing data- –5 –4.5 –4 –3.5 –3 –2.5 –2 –1.5 –1 –0.5 0 –5 –4.5 –4 –3.5 –3 –2.5 –2 –1.5 –1 –0.5 0 processing algorithms for real-time sys- Time Relative to Present (minutes) Time Relative to Present (minutes) tems is guaranteeing efficient and reli- Fig. 1—A screenshot of one of the views of the DAS user interface. able computational performance over long continuous run times (possibly sev- drilling-performance landscape contin- ability to perform stick/slip surveillance eral days per hole section and weeks or ually and ultimately maximize real-time and generally reduce the effectiveness months per well). As a result, particular drilling performance. of DAS. DAS operational recommenda- care must be taken to ensure that run- tions for WOB and RPM, along with time algorithms are simple and effi- Technical Overview of DAS other computed performance metrics, cient. This issue is particularly relevant The DAS system runs on a rig-based are sent to the vendor computer using when run-time hardware has a relative- computer with limited computing a customized WITS record. These DAS ly small footprint, as is the case for DAS power. Several versions have been de- outputs then can be distributed among (i.e., running on a conventional desktop veloped as research has progressed, but, vendor computers at the rig and shown or laptop in real time vs. off-site calcula- generally speaking, the DAS computer as part of standard strip charts in front tions on a server farm). has a standard serial-cable WITS data of the driller and rig supervisor. Fi- feed from the mud logger and a serial nally, the DAS custom WITS records Conclusions output to provide the DAS recommen- may be communicated to a WITS com- The DAS platform was developed to dations back to the rig data-processing puter and streamed in real-time to an serve as a real-time MSE and stick/slip equipment. The DAS computer screen information-management-center serv- surveillance-based digital assistant, is configured to display state-of-health er. This setup allows engineers at office with the ultimate goal of helping the metrics as well as optimization charts sites to view real-time drilling data and driller achieve consistently better and that show the drilling parameters and DAS recommendations by means of a longer bit runs. The system encourages DAS calculation results (Fig. 1). Web-based application. Furthermore, it regular (guided) drilloff tests and care- A generic drilling-data architecture enables information-center staff to log fully monitors drilling performance to has been established to enable on-site onto the DAS computer remotely for provide recommendations for improved DAS at a rig and remote monitoring by configuration and troubleshooting. performance. The system has now dem- use of a Web-based application. The At a high level, the main steps in the onstrated value in both offshore and DAS computer may be at any number DAS data-processing/optimization algo- onshore environments. of locations at the rig, including the rithms are Future development efforts for DAS doghouse, provided that a WITS feed ◗ Receiving raw WITS Record 1 will focus on software engineering and can be established between the mud- surface data characterizing the user’s experience in particular. A fit- logging system and the DAS computer. the state of the drilling assembly for-driller implementation will trans- Communication is established between as measured at the surface. form DAS from research software into the DAS and vendor computers through ◗ Performing preprocessing an interactive, easy-to-use platform WITS records sent by means of serial and cleanup of the raw WITS that continuously assists in the drilling- connections. The minimal data-rate re- Record 1 data. optimization process in a manner that quirement for WITS Record 1 inputs to ◗ Computing drilling-performance removes the most tedious components DAS is 1 second. Slower data rates (e.g., metrics such as MSE, ROP, stick/ of surveillance and trending analysis 5 seconds or more) limit the system’s slip severity, and depth of cut. from the driller’s work flow. JPT

70 JPT • OCTOBER 2016 Big-Data Analytics for Predictive Maintenance Modeling: Challenges and Opportunities

ig-data analytics can allow a better Bunderstanding of a production system’s abnormal behavior. This CM and D Application CM and D System knowledge is essential for the adoption Design Phase Use Phase of a proactive maintenance approach, leading to a shift toward condition- based maintenance (CBM). CBM focuses Machine/process Risk assessment on performing interventions on the Components breakdown Criticality/prognosis basis of actual and future states of a Faults/degradations Diagnosis system determined by monitoring Symptoms modeling Recognition underlying deterioration processes. One of the building blocks of CBM design Descriptors Processing and implementation is the prognostic Measurements approach, which aims to detect, classify, and predict critical failures. This paper Fig. 1—Condition monitoring (CM) and diagnostics (D) cycle. presents approaches for constructing a prognostic system. Traditional approaches to reliabil- ures in a floating production, storage, ity estimations are based on the dis- and offloading (FPSO) vessel. Introduction tribution of event records of a popula- Diagnostics can be described as a pro- Optimization of maintenance costs is tion of identical units. Many parametric cedure of reasoning to interpret the among operators’ main concerns in the failure modes, such as Poisson, expo- health condition of machinery by use search for operational efficiency, safety, nential, Weibull, and log-normal distri- of data acquired during its operation. and asset availability. The ability to pre- butions, have been used to model ma- It has a vital role in decision making dict critical failures emerges as a key fac- chine reliability. This project attempts to for both operation and maintenance. In tor, especially when reducing logistics create an integrated solution using big addition, diagnostic procedures should costs is mandatory. data and analytics techniques to imple- be adjusted according to potential fail- Experience has shown that significant ment a CBM standard procedure for the ures (on the basis of their likelihood and benefits can be achieved when major target problem. severity) that could occur in a machine. maintenance interventions (overhauls, The principle is shown in Fig. 1. The usually performed periodically) can be The V-Shaped Method V-shaped array represents the high-level postponed on the basis of conclusions The International Organization for Stan- and low-level concerns. from the use of degradation models. Such dardization’s standards for condition Condition monitoring for offshore in- an approach can be complex, but its re- monitoring and diagnostics of machines stallations is certainly a challenge, espe- sults may reduce maintenance and logis- offer good guidance to establish a CBM cially when it comes to data quality and tics costs while keeping availability with- standard procedure, especially for the analysis. Having identified the critical in required levels. target problem of turbogenerator fail- functions, it would be possible to identify the critical components, failure modes, This article, written by Special Publications Editor Adam Wilson, contains highlights and degradation mechanisms. of paper OTC 26275, “Big Data Analytics for Predictive Maintenance Modeling: Challenges and Opportunities,” by I.H.F. Santos, M.M. Machado, E.E. Russo, Machine/Process. The system under study belongs to the main power- D.M. Manguinho, V.T. Almeida, R.C. Wo, M. Bahia, and D.J.S. Constantino, generation system of an FPSO unit op- Petrobras; D. Salomone, M.L. Pesce, C. Souza, and A.C. Oliveira, EMC—Brazil erating in the Campos Basin. It has Research Center; and A. Lima, J. Gois, L.G. Tavares, T. Prego, S. Netto, and E. Silva, four turbogenerators, each consisting PEE-COPPE/UFRJ, prepared for the 2015 Offshore Technology Conference Brasil, of an aero- derivative gas turbine driv- Rio de Janeiro, 27–29 October. The paper has not been peer reviewed. ing an electric generator. The main emphasis of the complete paper is the Copyright 2015 Offshore Technology Conference. Reproduced by permission. gas-turbine engines.

The complete paper is available for purchase at OnePetro: www.onepetro.org.

JPT • OCTOBER 2016 71 Components. From a maintenance per- the human factor and those caused by occurrences (in less than a given time spective, it is of interest for the operator faulty or malfunctioning sensors. interval) is also performed. For all vali- to have a functional tree representing The following step is data analysis. dated NS or MF situations, a 24-hour the machine. This tree should be com- Several models, algorithms, and methods interval is identified during which the posed mostly of the maintainable parts, are available for data analysis, depending machine operated without interruption and, from the component breakdown, on the type of data collected. before the stop that originated the asso- one should list all possible failure modes ciated event. and their respective causes and degrada- Diagnosis and Prognosis. The final step tion mechanisms. in all CBM approaches is making deci- Feature Extraction. For all NS and MF Then, the criticality of each of the fail- sions. The diagnostics of machine fail- events, the machine operation within the ure modes should be assessed through ures is basically a procedure of mapping 24-hour interval identified in the pre- expert judgment or from historical data the information obtained in the mea- ceding step is characterized by mean- on the basis of significance and probabil- surement space or features in the feature ingful features that should act as the ity of occurrence. space to machine failures in the failure- classifier input. mode space. Symptoms Modeling, Descriptors, Prognostics is a complex task. In gen- Classifier Training. Using the features and Measurements. In the modeling of eral, it is divided into two main types. extracted in the preceding stage, the symptoms, the operator must rely on the The first includes a prediction of time chosen classifier is trained, following the expertise within the organization with until machine or component failure and event labels defined in the second step. respect to a particular asset. is called “remaining useful life.” The sec- The result of this four-step procedure Descriptors can be obtained from ond is used to predict the time that a ma- is a smart algorithm capable of identify- condition-monitoring systems, either chine could operate without failure. ing a faulty operation the next time the directly or after processing of the mea- machine is started, on the basis of the surements. Descriptors have one big ad- Research Hypothesis: features extracted during the 24 hours vantage over measurements: Their selec- The Challenge before the machine stop. tivity helps to increase the accuracy of One can observe that most failures are the diagnostics significantly. related to machine startups. This kind Final Considerations Data from sensors were stored, proof event is a hidden failure and is dif- and Future Work cessed, and analyzed in order to identi- ficult to predict. For the main failure Starting from a challenging research fy correlations between parameters that event considered in the complete paper, proposition, such as to model and pre- explain the events best (e.g., principal- one critical component is a valve from dict hidden failures, this study discusses component analysis); patterns of behav- which precise behavior is demanded dur- the modeling of machine failures by the ior related to major occurrences; if there ing the startup. use of a big-data analytics approach with were variables that should be included in Considering that the action (counter- different classifiers. the monitoring set; and what more can measure) for those events is the replace- In the search for a smart algorithm be considered in the correlation of vari- ment of the valve, a question was raised capable of identifying a faulty operation ables with their failure modes or critical for the research team: “If the abnormal the next time the machine is started, on components and subsystems. valve’s behavior could be detected during the basis of features extracted before the In general, all collected data can be the run, could a predictive model be de- most recent machine stop, some classifi- subdivided into two groups: (1) events, veloped to assign a probability of failure ers were tested in a series of experiments. data that include information on what ac- at the next startup?” The results of those experiments are pre- tually happened, what caused the event, Working with this notion, the team sented in the complete paper. and what was done, and (2) condition extracted sensor data from the indus- Among the problems encountered monitoring (CM), measurements related trial repository in order to train off- in this research were data-collection to the health state of the machine. line classifiers. difficulties in terms of database Typically, the event data collection re- standardization. quires manual data entry while CM data, Classification Processes From the use of predictive models, nowadays, is collected automatically with Database Preprocessing. This step in- once useful models are constructed in the help of sensors. cludes the removal of all major outliers the near future, another problem that and adjustment of the sampling frequen- arises relates to the decision making, Processing and Recognition. Data pro- cy to a unique value (1 sample/min). in which the process must include the cessing should be started with data filtra- model’s predictions. In that sense, fu- tion and cleaning because the collected Event Annotation. In this stage, the ture work will consider more than one data (especially those entered manually) time stamps of all normal stops (NSs) model, resulting in a voting system that may contain errors. The most common and machine failures (MFs) are deter- would be able to provide reasoning types of errors include those caused by mined. Removal of repeated NS and MF for decisions. JPT

72 JPT • OCTOBER 2016 TECHNOLOGY FOCUS

Sand Management and Sand Control R.J. Wetzel, SPE, Drilling and Completions Senior Adviser and Team Lead, Chevron Energy Technology Company

I suppose that many of us are taking a [C]ould the development us to a specific course for improvement deep breath just now. Many of us could because the variation from well to well be revisiting how we have been complet- of consistent practices might simply be explained away by the ing wells and what we might be able to be a critical first step differences in planning and execution. improve. These improvement areas often on our journey toward Given the preceding idea, could the involve some sort of trade-off between development of consistent practices be well deliverability and well/completion achieving improvements in a critical first step on our journey toward costs (in terms of equipment and rig time completion quality? achieving improvements in completion to deploy these various alternatives). I quality? The list of those practices that suspect that we all have been involved we should carry out in a consistent man- with completions where these two areas I can think of two primary reasons: ner is quite long. For sand-control appli- are debated. (a) We lack the metrics to support our cations, we could start with those activi- In my experience, it seems that much decisions and (b) we do not have con- ties that occur early in the design process. of our discussion revolves around what sistent practices (e.g., laboratory and After reviewing the many high-quality the various participants “feel” is the best design work, deployment process- technical papers written over the past approach. Much of the decision even- es) across our wells to allow us to com- year, I have found a few that I think offer tually hinges on what we will do in the pare our results. I suspect that you can a good place for you to start your journey short term (deployment) rather than in think of others. Both of these areas offer toward consistency. The three summa- the long term (deliverability). The rea- improvement opportunities. rized papers are all related to the selec- son is, in my opinion, that we are fairly For those of us who have a robust set tion of proppant and screens in your sure about the near-term items (related of metrics to evaluate our overall sand- sand-control completions. I am not pro- to cost) but often very uncertain about control planning and deployment pro- moting any one of these papers over the the longer-term items (deliverability as a cess, if the preparation work (e.g., core others. However, I am suggesting that result of how well we deployed the lower testing, compatibility testing, equipment whatever your organization does in this completion). Why is it that many of our selection) is not carried out in a consis- area, your organization should do it con- completion quality decisions are focused tent fashion, the variation in results as sistently. You may find your organiza- on cost and not deliverability? depicted in our metrics would not lead tion’s new, preferred approach to proppant and screen selection in one or more of the presented articles. JPT R.J. Wetzel, SPE, is a drilling and completions senior adviser and the team lead for the SandFace Completions Team at Chevron Energy Technology Company. He has 36 years of experience, in all geographic areas, in various aspects of drilling, completions, Recommended additional reading and workover. Wetzel has performed technical, operations, and at OnePetro: www.onepetro.org. management roles. In addition to supplying technical support SPE 178966 Sand-Retention Testing: for Chevron’s business units through the SandFace Completions Reservoir Sand or Simulated Sand—Does Team, he also participates in numerous design reviews of It Matter? by Tracey Ballard, Weatherford, Chevron’s major capital projects, with specific focus on drilling and completions et al. design selection and deployment plans. Wetzel also manages the SandFace SPE 179036 Sand-Screen Design and Completions technology-development program focused on improving Chevron’s Optimization for Horizontal Wells Using lower-completion reliability and performance. He holds a BS degree in mechanical Reservoir Grain-Size-Distribution Mapping engineering from the University of Louisiana at Lafayette and serves on the JPT by Mahdi Mahmoudi, University of Alberta, Editorial Committee. Wetzel can be reached at [email protected]. et al.

JPT • OCTOBER 2016 73 Re-Evaluation of Gravel-Pack-Sizing Criteria

n this paper, gravel-pack pore size Gravel-Pack Permeability vs. Stress Iis evaluated further by use of the 500 1 permeability of the gravel pack and Permeability other methods. A new sizing method is Length proposed that is based on the effective 400 0.8 formation size and the gravel-pack pore size. In this manner, the gravel pack is effectively treated like a screen and the 300 0.6 selection of gravel-pack size becomes similar to the selection of screen size. Introduction 200 0.4 Pack Length (in.) Pack Use of a coarser-grained material to stop Permeability (darcies) Permeability the production of a finer-sized formation sand in oil wells has been practiced 100 0.2 for decades. The selection of the coarser- grained sand has usually been performed on the basis of a multiple of one or more 0 0 of the smaller particle sizes in the forma- 0 200 400 600 800 1,000 1,200 tion material, with the goal being that Stress (psi) the formation material will bridge on the Fig. 1—Gravel-pack permeability for 20/40 gravel with increasing uniaxial stress. larger gravel-pack particles without reducing flow capacity or allowing excessive solids production. horizontal wellbores. Testing methods sizing ratios based on particle-size mea- This paper presents an alternative ap- have been modified so that gravel-pack surements, or pore size based on pack- proach to testing and selecting gravel tests can be performed by both constant- ing theory. sizes in which the size of the formation drawdown and constant-rate methods. Permeability depends on the sizes and material is addressed by using an effec- A summary of previous impor- shapes of interconnections between adja- tive size of the formation grains defined tant studies investigating gravel-size- cent pores, which, in turn, are influenced as the median grain size (d50) divided selection methods is provided in the by the entire grain-size distribution. Par- by the uniformity coefficient (d40/d90) complete paper. ticle shape has an important effect on and the gravel pack is described by the permeability because it influences the apparent pore size of the gravel. This ap- Gravel-Pack Pore Size From Permea- size and shape of interconnections be- proach allows laboratory performance bility Data. The packing arrangements tween particles. The more angular the testing for solids production, size of sol- of gravel, along with the size distribu- grains are, the smaller the voids and the ids, and retained permeability to be com- tion and shape, ultimately determine the more tortuous the flow paths. pared directly for gravel-pack comple- pore size between gravel particles. Sev- A discussion of methods of pore-size tions and for screen-only completion. eral methods exist for estimating the calculation and estimation is included in These data are especially useful in the pore size in porous media. These meth- the complete paper. selection of the completion method for ods tend to be based on permeability, Pore Size From Packing Arrange- ments. Pore size can also be obtained This article, written by JPT Technology Editor Chris Carpenter, contains highlights by assuming single-size particles in dif- of paper SPE 179023, “Gravel-Pack-Sizing Criteria: It’s Time To Re-Evaluate,” by ferent packing arrangements. The pore Christine Fischer, Vernon Constien, and Carla Vining, Constien and Associates, size changes from a maximum for a cubic prepared for the 2016 SPE International Conference and Exhibition on Formation pack of 0.414 times the gravel size to Damage Control, Lafayette, Louisiana, USA, 24–26 February. The paper has not been a minimum of 0.1547 times the gravel peer reviewed. size for hexagonal packing. While gravel

For a limited time, the complete paper is free to SPE members at www.spe.org/jpt.

74 JPT • OCTOBER 2016 packs certainly have a range of sizes The most-common methods for an- two methods are summarized in Table 4 and roundness, these very simple as- alyzing formation-particle sizes in- of the complete paper. sumptions offer advantages for compar- clude dry sieving, occasionally wet sieving data from large laboratory-test sets. ing, and laser particle-size analysis. In Conclusions Changes in packing order most frequent- some cases, dry-sieve analysis may yield Approaches to determine a representa- ly occur from differences in sedimenta- larger-than-actual particle-size values tive gravel-pack pore size have been eval- tion or stress. if fine particulates are present in large uated, and a simple relationship based quantities. The electromagnetic forces on cubic packing has been selected to Comparison of Methods for Estimat- between grains tend to promote aggrega- keep data sets constant. For special proj- ing Gravel-Pack Pore Size. Gravel packs tion, which shifts the percentage of fines ects, other relationships such as perme- are never single-sized particles, so there present in the formation sand. ability could also be used. Two testing are variable pore sizes inside the pack. For this work, all formation-particle- methods that can use constant draw- For laboratory-testing purposes, the size data were determined by laser analy- down or constant flow rate have been gravel pore size can be adjusted on the sis and all gravel data were determined used in screen-only and gravel-pack test- basis of gravel distribution, shape, and by sieve analysis. Before beginning ing, and the data can be compared di- stress. However, for comparative pur- particle-size analysis, the disaggregat- rectly for performance of produced sol- poses across large data sets, it is helpful ed samples must be cleaned of reser- ids, retained permeability, and size of to maintain a single method for estimat- voir fluids. It may also be necessary to produced solids. Selection criteria can ing gravel pore size. re- establish wettability after drying the be established by choosing ranges on For the tests in this work, the gravel- material if the particles are dispersed in the performance master curves that pack thickness was approximately 0.5 in. water for laser analysis. meet expectations for gravel-pack per- and the maximum net uniaxial confining The parameters calculated from formance. The ability to directly com- stress was 1,000 psi. Fig. 1 illustrates the particle-size analyses—besides the cu- pare screen-only- and gravel-pack-test change in gravel-pack permeability for mulative plots by weight or particle vol- data should aid in completion designs in a 20/40 gravel across the uniaxial-net- ume—include the median, which is the soft- formation reservoirs. JPT stress range of 200 to 1,000 psi in the oil- 50th percentile on the cumulative curve; flow-test equipment. Very little change in the mean, which is considered the aver- permeability occurred, which indicates age grain size; sorting, which measures that the packing arrangement has prob- the grain-size variation of a sample by ably not shifted to a smaller pore size. encompassing the largest parts of the To maintain a constant method for es- size distribution as measured from a cu- timating gravel-pack pore size in evaluat- mulative curve; skewness, which mea- ing performance master curves, the me- sures the degree to which a cumulative Changing Your dian gravel size multiplied by 0.414 was curve approaches symmetry; and the uni- Address? selected. This expression has been found formity coefficient, which is the ratio of to compare well with results from screen- the sizes at d40 to the size at d90 on the Let SPE know. only tests on the same formations. cumulative curve. For this work, the criti- +1.972.952.9393 cal formation sizes are the d50 and the Formation- and Gravel-Particle-Size uniformity (d40/d90). Analysis. Regardless of the method used to select the gravel-pack sand, the first Gravel-Pack- and Screen-Test Methods. step is to obtain representative sam- Several laboratory- evaluation meth- Update Your ples of the formation and determine the ods for testing sand-control screens particle-size distributions. As wellbore have been documented in the litera- Member Profile construction has shifted to horizontal ture, but test methods that include grav- wellbores, the number of formation- el packs are rarer. The two methods http://www.spe.org/ particle-size distributions in the lateral that have been used to develop data members/update can be quite large. Particle-size analysis presented in the complete paper are should be performed for each interval known as the constant-drawdown oil- sample. Because only one size of gravel flow (CDOF) test and the constant- flow- pack is normally placed into a wellbore, rate brine (CFRB) test. The CDOF meth- SPE Benefits methods for analyzing the large amount od has been used in many screen-only of formation-particle-size data and re- and gravel-pack-plus-screen tests. The Discover the possibilities. ducing it to a small number for testing CFRB test has been developed for evalu- must be implemented. Frequency plots ating screens or gravel packs by a meth- http://www.spe.org/ of all the distributions present in a well- od that more closely simulates depo- members/benefits bore are helpful in reducing the data to sition on a screen or gravel pack from critical distributions. erosion of a sandface environment. The

JPT • OCTOBER 2016 75 Evaluating Sand-Screen Performance With Sand-Retention Tests and Numerical Modeling

his paper presents a combined how sand-retention testing should be Development Texperimental and numerical- performed or results interpreted. Sand- of a Sand-Retention-Test modeling study on sand-screen screen performance is often evaluated Apparatus and Test Procedure performance. The objective is to on the basis of two criteria: screen plug- A new sand-retention-test apparatus has develop an improved methodology ging and sand retention. been developed aiming to overcome or for optimal sand-screen-aperture reduce some of the problems and limita- selection by addressing some of Limitations of the Existing Approach. tions identified with the existing sand- the limitations present in existing While the acceptance criteria for sand slurry-retention tests. The apparatus sand-retention tests. A new sand- production and produced-particle size comprises a clean-fluid (water) -injection retention-test facility has been are well-defined and these factors are system, a sand-slurry-injection system, developed that incorporates a number relatively straightforward to measure, a sand-retention-test cell, and a fluid- of improvements into the design and screen plugging is defined differently by and solid-collection-and-separation sys- experimental procedure. different investigators. There is current- tem and is described in detail in the ly a lack of measurements that can be complete paper. Existing Approach to Sand- used to accurately quantify screen plug- Screen-Performance Evaluation ging and screen permeability. Experimental Procedure. Once the Empirical criteria have traditionally A recent review on sand-slurry ex- sand-retention cell was set up with the been applied for selecting screen size perimental procedures and consequent test screen coupon of the required slot on the basis of one or two parameters interpretation methodologies identi- size, all the pressure transducers for derived from the particle-size distribu- fied a number of limitations of exist- screen, cell, slurry, and clean-fluid pres- tion (PSD). Because of the problems as- ing sand-slurry testing, one of which sure measurements were installed. The sociated with the empirical criteria, the was fluid-flow velocity, which is ad- procedure for conducting a sand-screen- industry standard practice is to con- opted in some of the laboratories as retention test was as follows: duct laboratory sand-retention tests on being one to two orders of magnitude ◗ Fill the sand-retention cell and all screen coupons. This is particularly rel- higher than the flow velocity typical- the hydraulic tubes to pressure evant for premium sand screens, the fil- ly expected in the field. A much higher transducers with clean fluid (water). ter media of which are fundamentally fluid-flow velocity could exaggerate the ◗ Pump clean fluid at a flow rate of different from those of a slotted liner difference in screen performances in the approximately 400 mL/min using or wire-wrap screen. The same mecha- laboratory that might not exist under the triplex pump. nisms for forming stable, multiparticle field-fluid-flow conditions. Others have ◗ Inject dense sand slurry at a rate bridges may or may not be applicable. recognized that simulating reservoir- of 5 mL/min while the mixer cell In most cases, the cross-sectional path sand-production situations in a sand- oscillates at 2 Hz. through these filter media is tortuous retention test may be problematic, be- ◗ Collect the fluid and any sand and 3D. Laboratory sand-retention test- cause the reservoir sand-production passing through the screen coupon ing is typically required to determine situation would be highly variable and in a plastic container emptied the effective screen opening size. There a much slower fluid-flow velocity could approximately once per liter. are no universal industry standards on cause other ambiguities. ◗ Inject the clean fluid and dense sand slurry continually at the same time This article, written by JPT Technology Editor Chris Carpenter, contains highlights until the height of the sandpack of paper OTC 26434, “Evaluating Sand-Screen Performance With Improved Sand- is 10 to 20 mm or the maximum pressure capacity of the sand- Retention Test and Numerical Modeling,” by B. Wu, S.K. Choi, Y. Feng, R. Denke, retention cell is reached. T. Barton, C.Y. Wong, J. Boulanger, W. Yang, and S. Lim, CSIRO, and M. Zamberi, ◗ Inject clean fluid at a constant rate S. Shaffee, M.B. Jadid, Z. Johar, and B.B. Madon, Petronas, prepared for the 2016 to evaluate sand-screen and sand- Offshore Technology Conference Asia, Kuala Lumpur, 22–25 March. The paper has not pack permeability. been peer reviewed. The photographs in Fig. 1 were taken during a typical sand-retention Copyright 2016 Offshore Technology Conference. Reproduced by permission. test and show the process of sand

The complete paper is available for purchase at OnePetro: www.onepetro.org.

76 JPT • OCTOBER 2016 Proppant pack consolidation technology Protect your integrity

Conventional proppant FUSION technology Unconsolidated pack Consolidated pack

Now you can create a bonded, high integrity proppant pack without closure stress.

FUSION® proppant pack consolidation technology prevents proppant washout from the non-compressive annulus and near-wellbore areas. Utilizing the technology creates a high-integrity, high-permeability proppant pack that sets up even in low temperature environments. FUSION technology has been successfully deployed in deepwater injection wells in the Gulf of Mexico, and is ideal for critical applications on high rate production wells incorporating frac and pack technology for sand control. Proppant pack integrity and placement can be evaluated for the life of the well using the built-in CARBONRT® inert tracer technology and our proprietary evaluation process. FUSION technology provides you with critical well integrity to inject and produce at the ultra-high rates required to improve well economics and increase EUR. carboceramics.com/fusion Fig. 1—Screen, sandpack formation, final sandpack height, and screen-pressure-measurement port (left to right).

retention on the screen, the sandpack Simulation Conditions. To generate the no clear particle-passing phenomenon, height measurement, and the screen- particle population required for the sim- while the height of accumulated particles pressure measurement. ulation, the PSD curve was divided into increased continually. equally sized bins, within which the par- Experimental Results ticle size was assumed to be the same and Parametric Studies. Parametric stud- and Analyses equal to the bin’s average particle size. ies were carried out for three key pa- The apparatus developed has been used The computation domain was to be kept rameters (i.e., liquid velocity, screen slot to conduct sand-slurry-retention tests, small to reduce computation time while size, and particle concentration for two which are described in detail in the com- remaining representative of the fluid- sands). The effect of liquid velocities was plete paper. The experimental program and particle-flow conditions in a sand- investigated with Geometry 1, and the ef- included eight types of sand and wire- retention test. Thus, the sand screen was fect of slot width and solid concentration wrapped-sand-screen coupons with five represented by a slice containing three was investigated with Geometry 2. A total different apertures ranging from 0.152 to slots. Two geometries were used for the of 19 cases were studied. 0.254 mm. simulation. Geometry 1 had a width, thick- Effect of Liquid Velocity on Amount ness, and height of 7.5, 2.5, and 30 mm, re- of Sand Produced. Sand production Numerical Modeling spectively, while Geometry 2 had a height was enhanced by increasing the liquid- of Sand-Retention Tests of 15 mm and its width and thickness re- flow velocity. With a Fully Coupled mained the same as those of Geometry 1. Effect of Slot Size on Amount of Sand Discrete-Element-Method Five slot sizes were considered in the sim- Produced. The volume of particles pro- (DEM) and Computational- ulation study. The sand particles were as- duced from the slot increased as the slot Fluid-Dynamics (CFD) Model signed with properties of glass beads; liq- size increased. Also, the time to reach A fully coupled numerical model, by in- uid properties were those of water. a stable retention increased as the slot corporating CFD in an in-house DEM size increased. code, has been developed and used to Sand-Retention-Test Simulation. In Effect of Particle Concentration on simulate sand-slurry flow and the sand- this simulation, for Rocla90 sand when Amount of Sand Produced. The amount retention process in the sand-retention the slot width was 0.2286 mm, solid vol- of particles produced reduced as the solid tests described in the complete paper. ume fraction was 1%, and the liquid-flow concentration increased. Also, the time The objective of the numerical simulation velocity was 0.003 m/s, the solids and to reach a stable sand retention reduced was to carry out parametric studies to un- liquid continually flew into the simula- as the solid concentration increased. derstand the effect of some key parame- tion domain at the specified rate from ters on sand-screen performance. the top inlet. The solids were treated Conclusions In the numerical simulation of the as spherical particles, with their sizes ◗ The amount of sand produced from sand-retention process, the DEM was generated randomly according to the a sand screen is correlated mainly used to model the solid phase and CFD PSD of the test sand. Once the parti- with the ratio of screen aperture modeled the liquid phase. The move- cles hit the screen, some of the particles to particle size. ments of discrete sand particles with dif- with a diameter less than the slot width ◗ Uniformity coefficient plays a ferent PSDs, the retention process of sand passed through the screen. After the par- dominant role in sand-screen particles by the screen, and the degree of ticles settled down to the screen, par- permeability. plugging/blockage of the screen aperture ticle numbers passing through the slots ◗ The correlations for sand amount can be illustrated clearly. The interac- were quickly reduced, with few parti- produced and the retained sand- tions between individual particles, par- cles passing through the slots at times screen permeability can be improved ticles and screen, and fluid and particles, along with the steady and continuous by incorporating sand-particle and their effects on retention behavior, flow of the liquid phase. With more and geometry into the correlations. can be studied at the sand-particle scale. more particles retained above the screen, ◗ A parametric study using the DEM- Most importantly, different PSDs of the the particles with diameters larger than CFD model has demonstrated the sand particles can be introduced read- the slot width accumulated above the major effect of liquid-phase velocity, ily in the simulation, and the underlying slot. These particles blocked the slot. A solid volume ratio, and slot size on mechanisms can be demonstrated. steady state was quickly reached, with the amount of sand produced. JPT

78 JPT • OCTOBER 2016 Factors Governing the Performance of Multilayered Metal-Mesh Screens

ultilayered metal-mesh screens Results merely induced by the deformation of the M(MMSs) are widely used as Three types of multilayered MMSs— filter layer but is induced by the screen standalone screens for sand control in 175-µm PSM, 250-µm PSM, and 115-µm assembly as a whole. unconsolidated formations. The nominal plain Dutch weave (PDW)—are studied. With regard to the effect of overlap on rating of such screens is usually based The name of each type corresponds to model-predicted sand production of a on the specifications of the filter layer. the nominal rating and the design of the 175-µm PSM, a universal decrease in sand It is often found that screens with the filter layer. All of the MMSs have the same production with an increase in overlap same filter-layer nominal rating perform three-layered structure, with one protec- was observed for all tested PSDs. differently. It is shown in this study tion layer above and one support layer 250-µm PSM. In the current case, in- that the primary reason for this is that below the filter layer. The protection and creasing the overlap not only decreases the sand-retention performance of support layers are PSMs with nominal the percentage of large pores but also de- multilayered MMSs is a strong function ratings of 542 and 864 µm, respectively. creases the size of the large pores. This of not only the filter layer but also the phenomenon was not seen in the 175-µm protection layer and the support layer. Effect of Layer Overlap. In this study, the PSM where the peak of all size distribu- compression/sintering-induced thick- tions was at 180 µm. The reason that this Introduction ness change of a multilayered MMS is different phenomenon is observed is that The sand-retention ability of a multilay- characterized by the average overlap. It the variation of PSD of a PSM screen with ered MMS is evaluated by examining the is a combined factor that accounts for respect to layer overlap is also influenced pore-size distribution (PSD) of the screen the reduction in screen thickness caused by the alignment of mesh layers and the as well as its slurry-test sand production by the inter- and intralayer wire defor- relative pore size of adjacent layers. [the sand-production evaluation is for mation during screen manufacturing. 115-µm PDW. When the overlap was plain-square-mesh (PSM) screens only]. The thickness of a screen could be changed on PSD of a 115-µm PDW, the re- Three major steps are involved in the obtained either from a direct measure- sults show that, as the overlap increases, evaluation process: ment of a sample or from computed- the PSD shifts toward smaller pore sizes. 1. Generate a multilayered virtual tomography-scan images. Once the The dominant pore size could be as small mesh assembly (an example is thickness is known, the overlap can be as 90 µm when there is 40% overlap for provided in Fig. 1) based on the calculated accordingly. the screen. Moreover, a comparison be- mesh specification and different 175-µm PSM. In investigating the ef- tween the PSD of a multilayered screen layer-overlap, layer-alignment, and fect of overlap on the 175-µm PSM screen, and the PSD of a single-layered PDW filter relative pore-size-ratio conditions. it is found that, for a multilayered screen, reveals that the PSDs at different overlaps 2. Evaluate the PSD of the virtual screen having larger overlap or, equivalently, are mainly controlled by the filter layer. with a specialized filtration model. smaller screen thickness decreases the The layer overlap changes the PSD of a 3. Enter the PSD information into percentage of large pores. This is because PDW screen and a PSM screen in differ- the analytical slurry-test model the lateral movement of sand particles ent ways. When varying the overlap, the to quantify the sand-retention between mesh layers is more limited with PSD of a multilayered PDW screen is pri- performance of the target screen. increased percentage of overlap. The size marily governed by the change in the fil- The model incorporated in this of the smaller pores becomes even small- ter layer. In contrast, for a multilayered study uses the entire PSD of a PSM er as the overlap increases. The change in PSM screen, the effect of varying layer screen to predict sand production. screen PSD with different overlap is not overlap on PSD is obvious only when the protection and support layers coexist with the filter layer. This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 178955, “Factors Governing the Performance of Multilayered Metal- Effect of Layer Alignment. The effect Mesh Screens,” by Chu-Hsiang Wu, SPE, and Mukul M. Sharma, SPE, The University of layer alignment is studied by shifting of Texas at Austin, and Rajesh Chanpura, SPE, Mehmet Parlar, SPE, and Joseph the position of the filter layer with re- Ayoub, SPE, Schlumberger, prepared for the 2016 SPE International Conference and spect to the protection and support lay- Exhibition on Formation Damage Control, Lafayette, Louisiana, USA, 24–26 February. ers. This is performed during the pro- The paper has not been peer reviewed. cess of stacking individual mesh layers

For a limited time, the complete paper is free to SPE members at www.spe.org/jpt.

JPT • OCTOBER 2016 79 (a) Layer 1 (b) Layer 2 (c) Layer 3 (d) Layers 1–3 (Top Protection Layer) (Filter Layer) (Bottom Support Layer) (Stacked) Fig. 1—(a through c) Individual layers are first generated according to mesh specifications. (d) All layers are then stacked together to simulate a multilayered virtual mesh screen.

together when building a multilayered interplays could be diminished or even causes a negligible change in its PSD. The screen. Note that all simulations in this eliminated. Here, the issue is addressed screens were all built with 30% layer section use the same mesh designs. by studying the relative-pore-size ratio, overlap and 0% filter-layer-alignment 250-µm PSM. In this case, the results defined as the ratio of the pore size of a shift. The result agrees well with the pre- show that the PSD is a strong function protection/support layer to that of the vious observation that the PSD of a PDW- of mesh alignment. It is shown that the filter layer. screen assembly is mainly governed by filter-layer pores, which have a nominal 250-µm PSM. The filter layer of a stan- the filter layer and is less affected by rating of 250 µm, may not be fully open dard 250-µm PSM screen in this study is adjacent layers. to the incoming sands in a small sample sandwiched between a protection layer Summarizing all of the study’s observa- because of the pore blockage by wires and a support layer. To isolate the influ- tions leads to the following conclusions: of adjacent layers. Also, there is no ob- ence of each layer, the simulation starts ◗ Having a protection layer can vious trend regarding how PSD chang- with modifying only the protection layer. affect the PSD of a multilayered es with the filter-layer-alignment shift. The results show that, as the pore size of screen. The PSD of a PDW screen This often-uninspected alignment varia- the protection layer increases, the per- is less sensitive to the existence tion between mesh layers plants some centage of small pores decreases. This of a protection layer when the degree of uncertainty into the results of allows a larger area of the filter layer to protection-layer pore size is laboratory slurry tests conducted with have unhindered exposure to the incom- larger than the filter-layer pore coupon samples. ing sand. size. The PSD of a PSM screen is 115-µm PDW. Here, almost no differ- When both the protection layer and the very sensitive to the addition of ence is found between the different align- support layer are modified, it is shown protection and support layers. ment cases. This can be attributed to the that the interplays between the filter ◗ Layer overlap changes the PSD fact that the PSD of a multilayered PDW layer and its adjacent layers diminish of a PSM mainly by imposing screen is mostly dominated by the PDW as the pore-size ratio of the protection constraints on the lateral movement filter; the filter-layer pore size is mainly layer and that of the support layer both of particles between mesh layers. governed by the mesh count, wire diam- get larger. At high pore-size-ratio values Interplays between mesh layers eter, and the layer overlap. for both the protection layer and the sup- should thus be considered when The study also determined that a gen- port layer, the PSD of a multilayered PSM designing a screen. eral decrease in sand production with an screen will be very similar to that of a increase in overlap is observed for both single-layered filter. Conclusions alignment scenarios. The amount of sand One can significantly overestimate For a PSM screen, the PSD and the cor- being produced from a PSM screen is af- sand production when the nominal rat- responding sand production are strong fected by its mesh alignment; this can ing is used to model the multilayered- functions of layer overlap, layer align- cause the variations in results common- PSM-screen performance. The difference ment, and the relative-pore-size ratio. ly seen in slurry tests performed with between predictions made by the nomi- These factors govern the PSM-screen per- screen coupons. nal rating and by the true PSD of a screen formance mainly by affecting the inter- can be remarkable and is highly unpre- play between individual mesh layers. The Effect of Relative-Pore-Size Ratio. Cer- dictable because of variations in mesh presence of protection and support layers tain degrees of interplay between the alignment and layer overlap. The situa- can downsize the pore throats in a mul- filter layer and its adjacent layers, es- tion is improved as the pore-size ratios tilayered screen, which results in a lower pecially for PSM screens, significantly of the protection and support layers be- sand production than that produced from affect the PSD and the sand-retention come larger. a single filter layer with the same nomi- performance of a screen. It is important 115-µm PDW. Increasing the pore- nal rating. Designing protection/support to review the design of a multilayered size ratios of the protection and sup- layers with large pore-size ratios helps to screen to determine whether these layer port layers of a multilayered PDW screen eliminate this complexity. JPT

80 JPT • OCTOBER 2016 PEOPLE

ZUHAIR A. AL-HUSSAIN, SPE, retired from Saudi Aramco as CHUCK MATULA, SPE, vice president of southern area oil operations. He joined Saudi SHASHI RAJAGOPALAN, Aramco as a petroleum engineer in 1978 and held positions of SPE, and LYLE LEHMAN, increasing responsibilities over his 38 years of service with the SPE, founded Shale 2.0, a company, including vice president of drilling and workover. data analytics company Under Al- Hussain, the drilling program expanded from 25 rigs focused on the improve- in 1996 to more than 200. The company’s drilling and workover ment of completion Matula Rajagopalan division has the highest percentage of Saudis manning service strategies for oil and gas companies’ and drilling contractors’ rigs. He pioneered the use operators. Matula was previously an officer of local sand and seawater for gas well fracturing operations. at ThruBit, cofounder and owner of Quanti- co Energy Solutions, and held executive po- DARRELL HOLLEK, SPE, sitions at Weatherford and Halliburton. He was appointed executive is a graduate of Texas A&M University and vice president, opera- holds a BA in engineering and an MBA in tions, at Anadarko, with finance. Rajagopalan is experienced in the Lehman responsibilities for US on- commercialization of business and technol- shore exploration, pro- ogy startups in energy and worked at Halliburton,WellDynamics, duction, and midstream Quanta Services, Genel Energy and other energy-related compa- Hollek Leyendecker activities, along with Gulf nies. He holds a bachelor of engineering in mechanical engi- of Mexico and international operations. He was previously neering from Queen Mary and Westfield College, University of executive vice president, US onshore exploration and produc- London, and an MBA from the Fuqua School of Business, Duke tion. He will also remain a director of Western Gas Holdings, a University. Lehman was previously managing principal consul- subsidiary of Anadarko, and Western Gas Equity Holdings, the tant at StrataGen and Pinnacle Technologies. He has 40 years of general partner of Western Gas Equity Partners. ERNIE experience in stimulation and has written more than 24 techni- LEYENDECKER, SPE, was appointed executive vice president, cal papers and 10 business white papers, and holds 7 working international and deepwater exploration. He was previously patents. He is a member of the SPE Distinguished Lecturers senior vice president, international exploration, and the other Committee. Lehman holds a BS degree in chemistry from the positions he held at the company include general manager for University of Oklahoma. worldwide exploration engineering, planning, and international negotiations; vice president of corporate planning; and TERRY EARL SWIFT, SPE, is retiring from Swift Energy as chief senior vice president, Gulf of Mexico exploration. executive officer (CEO). Swift served as chairman of the board of directors from June 2006 to April 2016 and will remain a C. SUSAN HOWES, SPE, was appointed vice member of the board until he exits the CEO role. He joined the president of engineering at Subsurface Consul- company in 1981 as manager of engineering and served in a tants and Associates. She will be responsible variety of roles over the years, including executive vice presi- for maintaining the technical quality stan- dent and chief operating officer, and president. Swift holds a dards for the company’s recruitment, consul- BS degree in chemical engineering from the University of Hous- tancy, and training services. Howes began her ton and an MBA under the president/key executive program of career at Anadarko and held a variety of engi- Howes Pepperdine University. neering positions in reservoir engineering, operations, mergers and acquisitions, technology and planning, and human resources, including senior staff engineer, engineering training and recruiting manager, employment manager, and learn- In Memoriam ing and organizational development manager. In 2007, she joined The In Memoriam section lists with regret SPE members who Chevron as Horizons Program manager and later moved into its recently passed away. If you would like to report the passing reservoir management function. Howes has coauthored papers of a family member who was an SPE member, please write to and articles on uncertainty management, risk management, and [email protected]. talent management for SPE conferences and publications. She chairs the SPE Soft Skills Committee and previously served as the Ben Abrahams, Aberdeen, UK director of the SPE Gulf Coast North America region. She is a Dis- Fakhry M. Abu Shakra, Amman, Jordan tinguished Member of SPE and is a recipient of the SPE DeGolyer S. Hugh Christianson, Midland, Texas, USA Distinguished Service Medal. Howes holds a BS degree in petro- Joseph C. Isaac III, Magnolia, Texas, USA leum engineering from the University of Texas and is a certified Franklin R Midkiff, North Newton, Kansas, USA Professional Engineer and Professional in Human Resources. Michael A. Watts, Anchorage, Alaska, USA

JPT • OCTOBER 2016 81 PROFESSIONAL SERVICES

[email protected] ™ www.AvasthiConsulting.com Changing Your Address? PLATTFTI SPARKS AVASTHI • Consulting Petroleum Engineers • Reservoir Engineering Let SPE know & ASSOCIATES, INC. • Reservoir Simulation Worldwide Energy Consulting, • Reservoir Characterization +1.972.952.9393 Engineering, Training, Shale and • Oil and Gas Reserves Evaluation EOR/IOR Project Operations • Fair Market Value and Acquisition Valuation • Enhanced Oil Recovery Since 1990 • Economic Evaluation Field Development Planning (FDP) of Onshore and Offshore • Oil and Gas Production lenandersen.com (Shallow and Deep water) Oil, Gas, Gas-condensate, and • Gas Storage Design and Screening Heavy Oil Fields, and Unconventional/ Shale Resources. • Regulatory Filings and Database Acquisition Reservoir Characterization, Engineering and Simulation • Expert Petroleum Engineering Testimony (Static and Dynamic Modeling), Economic Evaluation, IAM, Len Andersen EOR/IOR, CCUS, Geomechanics, and Fracturing/Stimulation FTI Platt Sparks www.FTIConsulting.com 925-A Capital of Texas Highway S. www.PlattSparks.com Head Office: 800 Rockmead Drive, Suite 212 Austin, Texas 78746 U.S.A. [email protected] Houston, Texas 77339, U.S.A. • Phone: +1-281-359-2674 (512) 327-6930 800-428-4801 H.J. Gruy and LONQUIST & CO. LLC SPE Benefits Associates, Inc. Petroleum Engineers • Energy Advisors www.lonquist.com Discover the possibilities. Oil and Natural Gas Reserve Advisors Reserve Determinations Market Valuations • Reservoir Engineering • Merger & Acquisition Support http://www.spe.org/members/benefits Geologic Studies Expert Testimony • Reserve Determinations • Mineral and Royalty Field Studies Arbitration Petrophysical Analysis Commercial Models • Economic Evaluations Management Reservoir Simulation Risk Analysis • Underground Storage • Regulatory Filings and Stochastic Evaluations Acquisition Screening Engineering Testimony • Salt Cavern Engineering • Facilities Engineering 6575 West Loop South, Suite 550 Bellaire, Texas 77401 • Supply Studies • Disposal Well Design & TEL: (713) 739-1000 • Graphical Information Systems FAX: (713) 739-6112 • Mining Engineering [email protected] CGCAWLEY, GILLESPIE & ASSOCIATES,A INC. www.hjgruy.com Austin — Houston — Wichita — Calgary PETROLEUM CONSULTANTS SINCE 1960 Worldwide Petroleum Consulting M.J. ENGLAND, P.E. CONSULTING PETROLEUM ENGINEER HOT Reserve Reports Estate Appraisals Fair Market Value Expert Witness Fort Worth Houston Austin (817) 336-2461 (713) 651-9944 (512) 249-7000 Engineering 215 Union Blvd., Suite 350 Lakewood, CO 80228-1840 Exploration / Field Development / Training www.cgaus.com – [email protected] Telephone: 303/298-0860 Integrated Reservoir Studies • Lead & Prospect Facsimile: 303/298-0861 COUTRET AND ASSOCIATES, INC. Generation • Reservoir Characterisation • Field Development Planning • Enhanced Oil Recovery • SERVING THE OIL AND GAS INDUSTRY FOR Petroleum Reservoir Engineers Underground Gas Storage • Reserves Audits • Training OVER 60 YEARS Property Evaluation, Reservoir Engineering & HR Development Fluid Injection, Property Management www.hoteng.com MILLER AND LENTS, LTD. Parkstrasse 6, 8700 Leoben, Austria 401 Edwards Street, Suite 810 Phone (318) 221-0482 INTERNATIONAL OIL AND GAS CONSULTANTS Shreveport, LA 71101 Fax (318) 221-3202 Phone: +43 3842 430530 / Fax: +43 3842 430531 www.coutret.com [email protected], [email protected] Specializing in All Phases of Reserves Evaluations, Including Petroleum Economics, FORREST A. GARB Huddleston & Co., Inc. Reservoir Engineering, Geology, and Petrophysics Domestic and International & ASSOCIATES, INC. Petroleum & Geological Engineers 1221 McKinney, Suite 3700 Two Houston Center Phone: (713) 651-9455 International Petroleum Consultants 909 Fannin St., Ste. 1300 Fax: (713) 654-9914 Houston, TX 77010 Houston, TX 77010 e-mail: [email protected] Reservoir Engineering Expert Witness Ph: (713) 209-1100 Fax: (713) 209-1104 Web pages: http://www.millerandlents.com Economic Evaluation Reservoir Simulation e-mail: [email protected] Geologic Studies Due Diligence Forensic Engineering Technical Staffing International Reservoir Technologies, Inc. 5310 Harvest Hill Road, Suite 275 NITEC LLC Dallas, Texas 75230-5805 INTEGRATED RESERVOIR STUDIES International Petroleum Consultants Tel: 972-788-1110 Fax: 972-991-3160 Email: [email protected] Seismic Interpretation & Modeling Fractured Reservoir Characterization/Modeling

Web pages: www.forgarb.com Stratigraphy & Petrophysics Gas Storage • Unconventional • EOR • CO2 CCS Black Reservoir Simulation Oil/Compositional/Thermal Reservoir Simulation Frank Givens, CPA Enhanced Oil Recovery Studies Well Test Design & Analysis Provider of LYNX®, MatchingPro®, Structured approach to Investment Decisions ® ® Well Completion Optimization PlanningPro and ForecastingPro Software 662-404-3798 [email protected] 300 Union Blvd., Suite 400 Denver, Colorado Lakewood, CO 80228 475 17th Street, Suite 1400 LPL FINANCIAL PH: (303) 279-0877 Fax: (303) 279-0936 Ph. (303) 292-9595 Member FINRA/SIPC www.irt-inc.com [email protected] www.NITECLLC.com

82 JPT • OCTOBER 2016 ROBERTO AGUILERA, Ph.D. Tarek Ahmed & Associates Ltd. PERA SERVIPETROL LTD. Taking Petroleum Engineering Training Curtis H. Whitson International Petroleum Consultants to a New Level & Associates Independent Oil and Gas Producers For dates & descriptions Naturally Fractured Reservoirs of courses held worldwide, please visit us at EOS Fluid Characterization Log Interpretation Well Test Analysis www.TarekAhmedAssociates.com Design & Analysis of PVT Data Performance Forecasts Gas Condensate Specialists Economics Compositional Simulation Expertise Numerical Simulation WILLIAM M. COBB Petroleum Short Courses Pipe-It Integrated-Model Optimization & ASSOCIATES, INC. E-mail: [email protected] — WORLDWIDE PETROLEUM CONSULTANTS — 903-19th Avenue SW, Suite 502 Skonnertveien 7, 1st floor • 7053 Ranheim Norway Calgary, Canada T2P 3T7 Waterflood & EOR Studies Phone 47 7384 8080 / Fax 47 7384 8081 Phone: (403) 266-2535 Fax: (403) 264-8297 Geological & Petrophysical Analysis [email protected] / www.pera.no http://www.servipetrol.com Reservoir Simulation Unconventional Resource Evaluation Reserves & Property Valuation PetroTel Gas Storage & CO2 Sequestration Analysis SiteLark Expert Witness • Technical Training Leaders in Oil and Gas Technology A Flotek Company FIELD DEVELOPMENT PLANS | RESERVOIR 12770 Coit Road, Suite 907, Dallas, TX 75251 CHARACTERIZATION & SIMULATION | Reservoir Modeling and Phone (972) 385-0354 www.wmcobb.com WATERFLOODING | ENHANCED OIL FAX (972) 788-5165 [email protected] RECOVERY | SEISMIC INTERPRETATION, Engineering Services ® MAPPING & ANALYSIS | WELL TESTING | ZAETRIC DRILLING | RESERVES AUDITS | TRAINING Please Contact: Providing technical document development, busi- Worldwide Offices: ness process support and printing/binding services USA | UAE | Malaysia | Oman | Russia www.sitelark.com to the oil & gas industry since 2000. Telephone: 469-222-5436 Ph: +1-972-473-2767 • DOCUMENTATION — Drilling & Completion, [email protected] | www.petrotel.com E-mail: [email protected] Rig Operations, QA/HSE, Equipment, Reports, Instructions & Procedures PRA JAMES E. SMITH & ASSOCIATES, INC. • BUSINESS PROCESS — Technical Contracts, Petrotechnical Resources of Alaska, LLC SPARTAN OPERATING CO., INC. RFQs, Process Evaluation, Project & Vendor Alaska’s Oil and Gas Consultants Management 310 South Vine Avenue, Tyler, TX 75702 Geology, Geophysics, and Engineering 903-593-9660 • 903-593-5527 (FAX) • 800-587-9660 • PRINTING/BINDING — Turnkey, In-House, www.Petroak.com Customizable, Quick Turnaround 3601 C Street (907) 272-1232 voice [email protected] • http://www.jes-engineer.com Suite 1424 (907) 272-1344 fax www.zaetric.com • The Woodlands, Texas James E. Smith, P.E., Registered Professional Engineer Anchorage, AK 99503 281-298-1878 • [email protected] ADVERTISERS IN THIS ISSUE JPT ADVERTISING SALES ADIPEC 2016 NCS Multistage, LLC Page 41 Page 21 Archer OneSubsea, a Schlumberger company Dana Griffin Page 3 Cover 4 Advertising Sales Manager (Americas, Asia Pacific, and South Asia) C&J Energy Services Peloton LandView Tel: +1.713.457.6857 Page 2 Page 13 [email protected] Cameron, a Schlumberger company Peloton, Inc. Jane Bailey Pages 7, 17 Page 39 Advertising Sales Manager CARBO PETRONAS (Europe, Middle East, Russia, and Africa) Page 77 Page 43 Tel: +44 (0) 1227.266.605 [email protected] CD-adapco Production Tool Solution Page 55 Page 5 Craig W. Moritz Energy Navigator Inc. PV Fluid Products, Ltd. Assistant Director Americas Sales & Exhibits Page 63 Page 45 Tel: +1.713.457.6888 [email protected] Enventure Rock Flow Dynamics Page 35 Cover 3 FMC Technologies Schlumberger Page 19 Cover 2 Ingevity formerly MWV Specialty Seraﬁm Ltd. Chemicals Page 53 Page 9 Superior Energy Services, Inc. ADDRESS CHANGE: Contact Customer Services at 1.972.952.9393 to notify of address change or make KBC Advanced Technologies Plc Page 47 changes online at www.spe.org. Subscriptions are Page 51 TAM International USD 15 per year (members). JPT JOURNAL OF LEUTERT Page 27 PETROLEUM TECHNOLOGY (ISSN 0149-2136) is pub lished monthly by the Society of Petroleum Page 15 Tendeka Engineers, 222 Palisades Creek Drive, Richardson, Mississippi State University Page 65 TX 75080 USA. Periodicals postage paid at Page 59 Visuray Richardson, TX, and additional offices. National Oilwell Varco Page 23 Send address changes to JPT, P.O. Page 33 POSTMASTER: Weatherford Box 833836, Richardson, TX 75083-3836 USA. Page 4

JPT • OCTOBER 2016 83 SPE EVENTS

WORKSHOPS 26–27 October ◗ Houston—SPE Diagnostic Price Environment Through Cost-Optimised Fracture Injection Test (DFIT)—Processes Well Integrity Activities 10–12 October ◗ Amsterdam—SPE Gas Well and Applications 5–6 December ◗ Kuala Lumpur—SPE Deliquification 2–3 November ◗ The Hague—SPE GIWS: Challenges of CO2—Creating Value Through 10–13 October ◗ Kuala Lumpur—SPE Where Authorities Meet the Industry To Innovative Technologies for Development, Reservoir Surveillance & Production Define Well Integrity Separation and Utilisation Enhancement Through Cost-Effective 7–9 November ◗ Houston—AIChE/SPE: Technology Integration and Operation Next Generation Deepwater Facilities CONFERENCES Efficiency 9–10 November ◗ The Woodlands—SPE 11–12 October ◗ Houston—SPE /IMarEST Strategies for Identifying and Exploiting the 19–20 October ◗ Lima—SPE Latin America Loop Current Eddy: Observations, Impacts, Sweet Spot in Unconventional Reservoirs and Caribbean Heavy and Extra Heavy Oil and Forecasts Conference 13–16 November ◗ Cairo—SPE Deep Wells 11–12 October ◗ Doha—SPE Reservoir Challenges in a Challenging Condition 24–26 October ◗ St. John’s, Newfoundland and Labrador—Arctic Technology Characterisation ◗ 16–17 November Abu Dhabi—AAPG/ Conference 17–18 October ◗ Houston—SPE Deepwater EAGE/SEG/SPE Knowledge Management Asset Integrity and Field Life Extension Challenges 24–26 October ◗ Moscow—SPE Russian Petroleum Technology Conference and 17–20 October ◗ Kuala Lumpur—SPE Asset ◗ 21–23 November Dubai—SPE/AAPG/ Exhibition Abandonment—Emerging Reality EAGE Unconventional Plays: Achieving 25–27 October ◗ The Woodlands—SPE 18–19 October ◗ Calgary—SPE Montney Play Efficiency and Effectiveness Through Integration Artificial Lift Conference and Exhibition- 19–20 October ◗ Bergen—SPE Workshops North America in conjunction with Offshore Technology 21–23 November ◗ Kuala Lumpur—SPE 25–27 October ◗ Perth—SPE Asia Pacific Oil Days (OTC) Flow Assurance—Survival During Industry Downturn and Gas Conference and Exhibition 24–26 October ◗ Abu Dhabi—EAGE/SPE 1–3 November ◗ Astana—SPE Annual Integrated Geomechanics in E&P 28–29 November ◗ Kuala Lumpur—SPE GIWS: Maximising Asset Value in a Low Oil Caspian Technical Conference and Exhibition 7–10 November ◗ Abu Dhabi—Abu Dhabi International Petroleum Exhibition and STATEMENT OF OWNERSHIP, MANAGEMENT AND CIRCULATION (Required by 39 U.S.C. 3685). 1. Title of publication, Journal of Petroleum Technology. 2. Publication No. 0028-1960. 3. Date of filing, September 2016. 4. Frequency Conference (ADIPEC) of issue, monthly. 5. No. of issues published annually, 12. 6. Annual subscription price, $4.50. 7. Complete mailing 14–16 November ◗ Bangkok—International address of known office of publication, SPE, 222 Palisades Creek Drive, Richardson, TX 75080-2040, Dallas County. 8. Complete mailing address of the headquarters or general business offices of the publishers, SPE, 222 Palisades Petroleum Technology Conference Creek Drive, Richardson, TX 75080-2040. 9. Name and address of publisher, Glenda Smith, 222 Palisades Creek 30 November–1 December ◗ Manama—SPE Drive, Richardson, TX 75080-2040. Name and address of editor, John Donnelly, 10777 Westheimer, Suite 1075, Middle East Artificial Lift Conference and Houston, TX 77042-3455. 10. Owner, Society of Petroleum Engineers (SPE), 222 Palisades Creek Drive, Richardson, TX 75080-2040. 11. Known bondholders, mortgagees, and other security holders owning or holding 1 percent or Exhibition more of total amount of bonds, mortgages, or other securities (none). 12. The purpose, function, and nonprofit status 5–7 December ◗ Nairobi City—SPE/AAPG of this organization and the exempt status for Federal income tax purposes have not changed during preceding 12 months. 13. Publication name: Journal of Petroleum Technology. 14. Issue date for circulation data below: September Africa Energy and Technology Conference 2016. 15. Extent and nature of circulation: 6–8 December ◗ Mangaf—SPE International Average Number Number Copies Heavy Oil Conference & Exhibition Copies Each Issue of Single Issue During Preceding Published Nearest 12 months to Filing Date SYMPOSIUMS A. Total number of copies (net press run) 79,573 75,404 29 November–1 December ◗ Banff—SPE B. Paid circulation (by mail and outside the mail) 1. Mailed outside-county paid subscriptions 33,253 30,369 Thermal Well Integrity and Design 2. Mailed in-county paid subscriptions stated on Form 3541 45,250 43,960 3. Paid distribution outside the mails including sales through dealers and carriers, street vendors, counter sales, and other 0 0 CALL FOR PAPERS paid distribution outside USPS SPE Reservoir Characterisation and 4. Paid distribution by other classes of mail through the USPS 0 0 Simulation Conference and Exhibition ◗ C. Total paid distribution 78,503 74,329 Abu Dhabi D. Free or nominal rate distribution (by mail and outside the mail) Deadline: 26 October 1. Free or nominal rate outside-county copies included on Form 3541 0 0 2. Free or nominal rate in-county copies included on Form 3541 0 0 SPE/IATMI Asia Pacific Oil & Gas 3. Free or nominal rate copies mailed at other classes through the USPS 0 0 Conference and Exhibition ◗ Bali 4. Free or nominal rate distribution outside the mail 409 500 Deadline: 9 December E. Total free or nominal rate distribution 409 500 F. Total distribution 78,912 74,829 OTC Brasil ◗ Rio de Janeiro G. Copies not distributed 661 575 Deadline: 8 March 2017 H. Total 79,573 75,404 I. Percent paid and/or requested circulation 99.48% 99.33% Find complete listings of upcoming I certify that the statements made by me above are correct and complete. Alex Asfar, Senior Manager Publishing SPE workshops, conferences, Services. symposiums, and forums at www.spe.org/events.

84 JPT • OCTOBER 2016 Your reservoirs ARE AS OLD asas thethe DINOSAURS.

Your SIMULATION TECHNOLOGY doesn’tdoesn’t hhaveave ttoo be.

Reservoirs are ancient, but your simulation technology shouldn’t be. Today, tNavigator is the apex predator, with the ability to close the loop on model simulation cycles. With tNavigator, you can quickly create and convert geological models to HD models for a complete loop solution. So don’t settle for prehistoric software—switch to tNavigator.

www.trytnavigator.com

RFD-014_Dinosaur_tNavigator_Campaign_080816.indd 1 8/9/16 9:28 AM Multiphase Compressor

Introducing the world’s fi rst subsea wet gas compressor. Developed in collaboration with Statoil, the multiphase compressor is an industry fi rst, designed as a contrarotating machine specifi cally for pressure boosting unprocessed well streams with no requirements for an antisurge system or upstream gas treatment. Based on our multiphase pump technology—which has accumulated more than 2.6 million hours of run time—this compact and robust system enables longer step-outs, continuous operation for lower capex, and increased production from subsea gas fi elds.

Find out more at onesubsea.slb.com/multiphasecompressor