INSIGHTS NOVEMBER / DECEMBER 2017

Can we squeeze more value from reservoirs in production - from better use of data? Truthful, professional consulting; no repeating what you already know, and no spin.

Petromall is a unique oil and gas advisory service which prides itself on technical excellence in selected fields and supplementing business management and leadership in the face of uncertainty. .

• Exploration Time to treat reservoir simulators with more scepticism? • Subsurface Using changes in gravity and subsidence on the• Projectseabed Capital Working with repeat seismic recorded on the seabed Using fibre optics in wells for measuring seismic• Field Operations Rock relative permeability from a digital rock •sample Security of Supply Improving reservoir simulation based on seismic• Decommissioning Working with well interference data A new cloud based production system for production engineers Petromall actively participates in the facilitation and conduct of seminars worldwide, and has been a proud sponsor of this event with Finding Petroleum.

Our presence is across the globe in the UK, Malta, North America, with partners in the Middle East, Time to treat Using changes Working with Using fibre optics Rock relative North and Sub- Saharan Africa. reservoir in gravity and repeat seismic in wells for permeability from simulators with subsidence on recorded on the measuring a digital rock We help you to put science and engineering more scepticism? the seabed seabed seismic sample back into every decision made. Visit us at www.petromall.org

[email protected] +44 020 3286 2556 Can we squeeze more value from reservoirs in production - from better use of data? The understanding that oil and gas companies have of reservoirs in Petromall is a unique oil and gas advisory production is a basis for multibillion dollar decisions and could hardly service which prides itself on technical excel- be more important in the overall success of the business. Is there lence in selected fields and supplementing business management and leadership; in the room for improvement in how they develop this understanding? face of uncertainty. The oil and gas industry’s understanding Between this bottom up and top down of reservoirs in production is the basis for processes is something of a black box. We offer truthful, professional opinion and multibillion dollar decisions which affect How do oil and gas companies take this advice; no playback of what you already everyone in the industry. bottom up and top down information to know, and no spin. build and improve their understanding The valuation of the company, the re- of the reservoir? Yes, we know they use Petromall was founded by 4 senior industry maining life of a well or field, sizing of and academic practitioners who consider reservoir simulators, but these are still the challenges faced by today’s oil and gas new production facilities, decisions about something of a black box, only under- environment are going to require herculean where to place infill wells, and whether stood by the people who use them. acts of leadership and technical skill in the to do workovers or enhanced oil recovery high cost provinces of the world, in order schemes, are all based on the company’s Some people we talked to think that to maintain an industry that is sustainable understanding of its reservoirs in produc- reservoir simulators, and their predic- and even recognisable compared to recent history. tion. A better understanding of reservoirs tions, should be treated with a little could also help maximise hydrocarbon more scepticism. The alternative ways Similarly nations developing an oil and gas production and minimise water. to understand reservoirs, based on ob- industry face related challenges as they served data, do not give a complete pic- seek to maximise the benefits of this new The industry has made great advances in ture either, but should perhaps be given wealth-creating opportunity - in a improving recovery over recent decades, more weight, with the raw data more responsible manner. from having a better understanding of widely shared. the reservoirs and how to extract more GLOBAL REACH from them. Understanding of oil and gas reservoirs can only become more critical in future Petromall actively participates in the facilita- But it may be possible to improve recov- years. When operating margins are tight, tion and conduct of seminars worldwide, in ery even more, if more different types making the right decision becomes more conjunction with its associate partner Find- of data could be used in better ways to ing Petroleum. important. With mature fields, predic- understand reservoirs, and more people tions need to be made about how much Our presence reaches parts of the globe, could contribute to the understanding. life is left. Often we have drilling, field such as: the UK, Malta, the USA, through This might mean having reservoir data development, operations, asset sales, partners in the Middle East, North Africa and systems which are easier for people from tie-backs to new wells, enhanced oil re- Sub-Saharan Africa. different disciplines to work with and ac- covery and decommissioning all taking cess, than the current complex reservoir place in the same field at the same time. simulators. This report explores ways it Perhaps in future we will also have CO2 could work. injection, for permanent storage and en- hanced oil recovery. This edition of Petromall Insights There is plenty of published material was written by: about new subsurface survey techniques, Karl Jeffery, editor, And there are plenty of new technologies Digital Energy Journal / Finding Petroleum how geoscientists develop their static being developed which can help under- (one point in time) geological models, stand reservoirs in production - including how reservoir engineers get a better 4D seismic techniques on the seabed and Layout by: understanding of reservoir properties in in wells, measuring changes in gravity Laura Jones, Very Vermilion Ltd specific points of the reservoir, and res- and subsidence on the seabed, under- ervoir simulation software available on standing multiphase flow from analys- the market. We can call this bottom up ing drill cuttings, studying interference Published by: information to understand a reservoir. between wells, better ways to integrate Petromall Limited and simulate data, enabling production CentralPoint There is also plenty of published material engineers to access reservoir data, and 45 Beech Street about successful wells and dry wells, preparing better geological models for London EC2Y 8AD production rates, and companies which reservoir simulators. The industry needs United Kingdom over-estimated their resources or pro- to make sure that all the information duction and let down their shareholders. generated from these technologies is Email [email protected] We could call this top down information not wasted. +44 (0)20 3286 2556 to understand a reservoir. www.petromall.org

PETROMALL | INSIGHTS 3 Integrating and analysing multi-physics - the key to improving production? We have many technologies which can acquire data on the dynamics of petroleum reservoirs - if we can integrate this multi-physics and analyse it as one, our production predictions might take a great leap forward By David Bamford Hindsight is a wonderful thing and looking back abilities quoted. Sometimes we may be sceptical data on the actual dynamics of petroleum res- one can see areas of our everyday and working but the forecasts are much more reliable now. ervoirs. lives that technology has transformed. Closer to our working lives, consider the impact But.... the trick we have not yet mastered is how One of my over-used examples is the weather the advent in the mid 1990’s of ‘regional’ or ‘ex- to integrate this multi-physics and then analyse forecasting that we see now hourly it seems on ploration’ 3D has had our exploration success it as one. TV. rates, our exploration prediction accuracy. Who would think of exploring without it nowadays? If we can do this then our production predictions Many years ago I visited the UK’s Met Office in will take a similar Great Leap Forward to the the hope of learning something about prediction For a while I have asked myself when and weather forecasts most of us see every day! success and instead discovered that their fore- how we will see the same impact on reservoir casting apparently consisted of 1st degree level management? David Bamford is a former global exploration lead physics, some IBM 360’s strapped together, and with BP, and former non executive director of data from a few meteorological stations scat- Partly this is because reservoir engineering or, Tullow Oil and Premier Oil tered up and down the UK and out in the At- more precisely, reservoir simulation has remained lantic. pretty opaque to me, an art I have described as similar to the building of Chartres Cathedral or David Bamford The resulting analyses didn’t lead to very good perhaps the Grand Mosque at Córdoba - built predictions and I left assured that the best fore- and re-designed over decades, even centuries casts were based on either “tomorrow will be and as in the case of the latter, filling now a com- pretty much like today” or “tomorrow will be pletely different purpose to that which it was pretty much like the same day last year”! designed for.

Consider what we see nowadays - satellite and But we now have, as reviewed in this Petromall airborne imagery: radar, thermal etc, even prob- Insights, several technologies which can acquire

Working with technology better understand reservoirs in production. seismic; and Exa Corporation, which can model We did not focus too much on their technol- multiphase flow through a reservoir, from a The approach of writing this report is to try to ogy itself, but focussed more on how com- piece of drill cutting or core analysed in a CT find out how the oil and gas industry actually panies work with their technology. How oil (tomography) scanner. works with reservoir data today - and explore and gas companies make a decision to use it, The second section looks at better ways to how it might be improved. and how they work with the data which is gen- erated. work with reservoir data. We look at iRes-Geo, We did not find any people who currently which has developed a workflow for preparing work at oil and gas companies willing to speak This report has two sections. The first is on high resolution geological models for reservoir on the record, but many of our interviewees better ways to gather data about a reservoir simulators; Kes Heffer of Well Dynamics, who formerly held senior roles in oil and gas in production. We look at Octio, which gathers studies flow interference data, and Paradigm, company reservoir departments. gravity, subsidence and seismic data from the with software to help production engineers seabed; Seabed Geosolutions, which gathers work with reservoir models. We also spoke to seven companies who are seismic data from the seabed; Silixa, which puts developing technology which can be used to fibre optics in wells which are used to record The reservoir simulator The reservoir simulation starts with a static picture of the reservoir, developed by geoscientists, with models generated largely from interpreting seismic data, mapping them against well data and anything else available. The reservoir simulator makes a Reservoir engineers aim to continually They can calibrate the reservoir models computational model of how fluids improve their simulations by using across the whole of the reservoir, not would flow in this reservoir model, revised geological models made by just for individual wells, looking at the with the pressures, different fluids, rock geoscientists, adding more live data, entire volume of the reservoir. properties, fractures and wells taken and making other tweaks until the into account in the computation. predicted production ‘history matches’ They want to get correct data for the the actual production. reservoir depths, surfaces, faults, net

APRIL 2017 4

to gross (total ‘pay’ footage divided by project actually has a negative net not get to learn about it. total reservoir footage), porosities and present value (NPV), when your lithologies. calculations show it as positive. We have not seen any single software system capable of managing all of the The dynamic flow simulation also Observed data various information about reservoirs takes into account that the pressure in in production – with all information the reservoir will drop as it depletes, The only alternative to reservoir at different granularity, different time leading to reduced flows, meaning simulation is to work on actual scales, and about different parts of the lower production through the wells. observation of the reservoir and there reservoir. is no perfect way to do this either – Over time, the reservoir simulator just repeat seismic surveys, gravity You might remember the big promises should show you parts of the reservoir or subsidence surveys. These are of around 2006 of the ‘digital oilfield’, which are not draining into the wells, expensive and usually only cover a seeing the oilfield as something of a and so are targets for infill drilling. small area. Or you can try to get more living being, with wells all in the perfect insights from well data or analysis of place, the flows all perfectly extracting But many of our interviewees rock samples from drilling. the maximum amount of hydrocarbon expressed private doubts about for the least investment. reservoir simulators. It isn’t possible to But observation data might be able know for sure how good they are. In to directly answer questions such as Perhaps a more useful data theory, the closer the simulator is to where untapped areas of the reservoir management concept is the ‘digital reality, the more likely it is to be able might lie. For example, if you have twin’ idea which is being developed to predict accurately. But in practise, subsidence sensors over the reservoir, by many engineering companies for if you don’t (and can’t) have a good and 90 per cent of them show that the topsides equipment. This is a digital understanding of the subsurface – the ground beneath them is compacting copy of what is happening in the real geology and fluid flows – then more slightly, the other 10 per cent of them world. Or to express in another way, investment won’t give you better show no subsidence, these 10 per cent it is all of the company’s information predictions. may be above a section of the reservoir about real world operations put which is not draining into the wells. together in an integrated format so it is Two interviewees raised the issue easy to work with. that the reservoir simulation is Data management usually based on statistics, probability A subsurface ‘digital twin’, developed and various empirical (observed) Another issue is how all of the dynamic along similar lines, would not be equations. Techniques to get a more reservoir data should be managed. a simulator at its core, but a way understanding from seismic have much At the moment, most companies use to gather all of the company’s improved in recent years, including the reservoir simulator as a kind of understanding of the reservoir, in a understanding rock physics, and the data management system. All new workable format, from one place. It is reservoir simulations do not get the data is ingested somehow into the available for everybody involved with benefits from them. simulation. It can include everything the reservoir to work with, and easy to you know about the reservoir use. Any simulations for predicting the Also, the geological model often needs including geological interpretations, future could be viewed by drilling down to be simplified (reduced in granularity) stratigraphy, sedimentology, porosity into the model, rather than being the in order to make it possible to put and permeability, core data and well core of the model itself. through computation simulation. This log data (such as saturation). can remove a lot of the detail which is Author’s required to make it accurate. But that doesn’t provide much acknowledgements confidence that the company For example well logs might show up gets maximum insights from the Thanks to Petromall for funding the small thin barriers or layers within information available. writing of this report, and David a reservoir, 1-2m thick. When the Bamford for initiating it. Thanks also reservoir model is reduced to cells with For example, someone might make to our interviewees Martha Lien and a resolution of tens of metres for a a useful observation about an Hugo Ruiz of OCTIO, Hemang Shah of simulation, they would disappear from injection well seeming to have a Seabed Geosolutions, Garth Naldrett of view. close relationship with pressure in a Silixa, David Freed of Exa Corporation, production well, indicating that water Some lithologies such as carbonates Jim Farrington of iRES Geo Technology, breakthrough to the production well is Kes Heffer of Reservoir Dynamics, Indy can be very difficult to simulate, not far away. because there can be big variations Chakrabarti of Paradigm. in the porosities and structures, with The right decision might be to choke so many complex caves or erosions down that production well. and fractures at different scales and directions. But if instead the information is just passed to a reservoir engineer, to ingest There could be information you don’t into the reservoir model in its next know which reveals why the reservoir re-run, the production engineer might

PETROMALL | INSIGHTS 5

OCTIO – measuring gravity, subsidence and seismic on the seabed OCTIO of Bergen provides services for recording gravity changes, pressure changes (an indication of seafloor subsidence) and seismic from the seabed – all useful ways to get observational data on reservoirs in production Reservoir monitoring and subsea surveillance each other nicely for interpretation, as the Hugo Ruiz, vice company OCTIO, based in Bergen, Norway, former provides good quantification of mass president G&G, provides gravity and seafloor subsidence changes while the second maps accurately Octio monitoring. The technology is currently being the extent of the area affected. applied in 7 oil and gas fields in Norway. It also makes and operate subsea passive The gravimeters used by OCTIO Gravitude seismic recording devices. are sensitive to a few microgals change, which is a billionth of the normal gravity field The gravity and subsidence monitoring on the earth’s surface, or the gravitational technology was developed internally by field between two people half a metre Statoil, and pioneered at the largest field distant. reservoir has more compaction than another, in Norway, Troll. It was then taken over by it tells you that there is a compartment of “Our clients maintain a reservoir model, OCTIO in 2013, with an aim to achieve a the reservoir that is not being depleted, and and pick a few parameters within the model wider rollout of the technology. OCTIO’s an infill well needs to be drilled there. daughter company Gravitude has been which determine mass changes,” says Hugo surveying Norway’s Ormen Lange field, the Ruiz, Vice President G&G at OCTIO. “By The subsidence measurement can be second largest gas field in Norway, since choosing the values of these parameters important for installation safety. In extreme 2014. that better fit the observed gravity data, cases, subsidence can damage the platforms they reduce significantly the space of sitting on legs on the seabed. Gravity possibilities of these parameters and hence the uncertainty in their models.” Gathering data The gravity data can be used to better understand gas production, because once gas The data can be used to monitor movement The usual workflow begins with a feasibility is produced, the space it previously occupied of a gas-water contact, quantify water influx study, modelling the expected gravity change is usually filled with water, leading to a from aquifers, map hydrocarbon depletion, and subsidence to occur. The objective is higher gravity reading, because water has identify compartmentalisation, map reservoir to determine where these measurements a higher density, says Martha Lien, CEO at properties like compressibility away from will help reducing the uncertainty in OCTIO Environmental Monitoring. wells. The economic value of information the modelling, and hence provide better arises from identification of infill well decisions during field development. This way, you can spot a section of the field planning targets, avoiding water break-in which is not communicating with the rest of to wells, or improved hydrocarbon reserve Typically 20-120 gravity / subsidence the field (the fluids are not moving, in other estimates that allow a better planning of measurement locations are selected in words), because there won’t be any gravity pipelines and resources. a survey, marked with semi-permanent change there. concrete platforms on the seafloor. The Subsidence gravity and subsidence measurement Gravity data is gathered in dedicated equipment is placed sequentially over the surveys, in which gravimeters are deployed OCTIO Gravitude’s method for platforms during a survey, with the help of sequentially at a set of locations on the measuring seafloor subsidence is based in an ROV (remote operated vehicle). This way, seafloor, situated above the producing measurements of changes of water pressure measurements at different surveys, which reservoir. at the seafloor. As gravity monitoring, it is can be one to three years apart, are gathered based on periodical surveying. at exactly the same position. By recording gravity on the seabed, the accuracy is “orders of magnitude” better than When hydrocarbons are produced, the The instrumentation used by OCTIO recording it from vessels or from the air, Ms reservoir compacts and the seafloor Gravitude includes 3 gravity meters and 3 Lien says, since the recording is much closer experiences some degree of subsidence, Ms pressure meters. The data from the sensors to the ground and at stable conditions. Lien says. Changes of seafloor depth above is sent to the ROV, and further to the vessel, the field are compared with measurements where its quality is analysed in real-time. Oil companies use the data to improve their away from the field, to provide calibration. Each reading takes about 20 minutes. reservoir simulations. “In our experience, our In this way, we obtain 2 mm accuracy in data is used directly to improve the reservoir subsidence throughout the field. There is no Stations are placed both above and around modelling, to enhance the confidence in other technology that can reach to such level the hydrocarbon field, with the stations the predictions of future production, and of accuracy. placed around it used to provide calibration. eventually to take better decisions, like The data is corrected for tides and other placement of additional producing wells” Ms Subsidence data provides a map of reservoir oceanographic effects. Lien says. compaction as it is being produced. You can also see lateral differences - if one part of the The data from gravity and subsidence meters Gravity data and seismic data complement is delivered to the clients usually as a gravity 6

change or subsidence map, and support is reservoir. This became the most producing It uses one of OCTIO’s systems to monitor provided to include this new data optimally well in the region. the injection of the fluids. The passive seismic in the reservoir simulation. system can listen continually and check if The devices were used on the Ormen the injection is creating any cracks in the Case studies Lange field to help make decisions about subsurface, which would indicate that there installing compression facilities and infill could be waste escaping from the target On the Troll field, 4D gravity data saw a rise wells, where there were uncertainties about formation. It can also detect any possible of 2 m in the gas water contact in the period compartmentalisation in the reservoir and fault re-activations. 2002 to 2009. That couldn’t be detected early water breakthrough to the production with time lapse seismic. wells. OCTIO’s systems can be used to monitor well drilling operations. In such projects, the On the Mikkel field, gravity surveys have On the Statfjord field, subsidence sensor array is used to precisely position the been performed since 2006 to monitor water monitoring was used to look for undrained drill bit with respect to the geological data of production into the reservoir, because there compartments and study aquifer properties. the area. Any small crack developed during were concerns about hydrate formation. The data was also used to calibrate the the drilling is also detected in real-time, and The monitoring showed lower water influx geomechanical model. this provides additional safety to the drilling than expected into the reservoir, which lead operation. This service has the commercial to a significant change in the estimated Passive seismic name of OCTIO DrillWatch, and is especially gas volumes in place. This, in turn, helped OCTIO provides also seismic sensor arrays suited for areas with thin overburdens, like with long term planning of pipelines and the Barents Sea. resources. – which can be used to image the reservoir using active seismic surveys, but also for The key features of OCTIO’s seismic On the Midgard field data has been passive microseismic monitoring. systems are their robustness and flexibility. gathered also since 2006, helping to Small sensor networks can be deployed for monitor the reservoir draining patterns and Passive monitoring can be used at drill cutting and waste water injection wells, to specific applications at the beginning of the aquifer support, the fault distribution and production of a field, and then expanded compartmentalisation. It was possible to make sure it is not causing any formation damage that can eventually cause waste to to larger areas and even a full reservoir see that one segment of the reservoir was monitoring system if required. underproducing, with faults acting as barriers leak to the surface. to the flow. One oil company customer injects waste The data can be collected and pre-processed water and drill cuttings back into an oil well on an offshore platform, and then sent to The reservoir model could then be updated land for further processing. to include a sealing fault, and a new well 1000m below the seafloor, which is much could be drilled in the undrained part of the less expensive than shipping it to land. Seabed Geosolutions – recording seismic on the seabed Seabed Geosolutions places seismic ‘nodes’ – small devices - on the seabed, which can be used for repeat (“4D”) seismic recording, with better resolution than conventional towed streamer seismic

Seabed Geosolutions, a company with offices be changes in the seismic response on Ocean bottom node surveys have in Houston and Massy amongst others, places repeat seismic surveys due to the changes traditionally been more costly than towed seismic ‘nodes’ – small recording devices - on in reservoir fluid and/or pressure which streamer seismic surveys depending on the seabed which can be used for repeat or could show up in amplitude attributes, or the geometry of the acquisition. Seabed “4D” seismic recording. “time-shifts” (changes in the seismic velocity Geosolutions are continually reducing costs leading to changes in the two-way time of as they focus their efforts on operational Ocean bottom nodes (OBN) are considered the reservoir). This can help indicate where innovation and efficiency enhancing a better seismic acquisition technique the hydrocarbon production is coming from technologies such as their Manta® system. than conventional towed streamer or where injected fluids are flowing in the because fuller azimuth illumination of the reservoir and provide valuable input to subsurface (looking at it from more different decisions on reservoir management such as directions) can be obtained, leading to a drilling further infill or injector wells. better understanding of the reservoir and overburden structures. This gives you better Conventional towed streamer seismic definition of the overburden and reservoir provides only 20-30m resolution of the properties, including seismic anisotropy, subsurface at typical reservoir depths depth and geological structure. The seismic which might not be enough to resolve the amplitude response of the reservoir can entire reservoir, the company says. Seabed give additional information about the rock Geosolutions’ broadband seismic technology properties, “net to gross” and possibly produces higher resolution images because of porosity. wider bandwidth acquisition and lower noise content. Seabed Geosolutions have developed a containerised After a few years of production, there may solution for efficient deployment of ocean bottom nodes either by ROV or by wire.

PETROMALL | INSIGHTS 7

Manta® ocean bottom nodes are designed to seamlessly deliver improved geophysical illumination with flexibility for dense source grid, full-azimuth and long offset surveys.

“It is a case of trying to acquire data as Companies will typically perform repeat factor of 55 to 60 percent, but the Upper efficiently and cost effectively as possible that seismic surveys every 1-2 years depending on Triassic had only produced 15 percent, gives the understanding you need to make reservoir seismic rock properties, but 5 years and therefore, was a focus for production sensible business decisions,” says Hemang is probably too long, noting oil wells typically optimisation. It is a tricky reservoir of Shah, Seabed Geosolutions’ Area have a production life of 10-20 years, he says. low net-to-gross with thin sand bodies of Geophysicist for Europe, Africa and the lateral extension from a few to hundreds of Middle East. Alwyn North Field meters. The sands have low porosity and permeability. Seabed Geosolutions’ ocean bottom nodes One example of using ocean bottom nodes run on batteries that last approximately 75 is in the North Sea Alwyn North Field where The EAGE Conference paper presented in to 120+ days. Therefore, they cannot be Total presented a paper about its experience Paris in 2017 confirmed that OBN gave kept on the seabed running permanently. It at the EAGE Conference in Vienna in June better imaging and attributes than the is possible to have ocean bottom sensors on 2016 with subsequent follow-up results at vintage towed streamer, and thanks to the cables which are powered from the platform the EAGE Conference in Paris in June 2017. wider bandwidth and improved signal-to- and left permanently on the seabed. This has noise ratio, a better description of static been done in a few places around the world, The field, which is subject to significant reservoir geometry was obtained leading to including the Norwegian field, Valhall. These faulting, is located on the Western side of the an improved understanding of the dynamic permanent systems allow repeated surveys at Northern Viking Graben, has three stacked characteristics. The results can be used as part lower cost, but with a significant initial capital reservoirs, from the Middle Jurassic, Lower of a dynamic simulation, leading to better outlay compared to OBN surveys. Jurassic to Triassic, and Upper Triassic. The correlation with well data, and improved first two reservoirs have seen a recovery infill well selection.

Results from Sparse Ocean Bottom Node on the Alwyn North Field - from Acquisition to Joint PP-PS Imaging, Brunelliere et al, EAGE Conference (2016) highlight clear improvements on both faults and lithology in the OBN data versus streamer data.

NOVEMBER / DECEMBER 2017 8 Silixa – using fibre optics in wells to measure seismic Silixa uses fibre optic cables in wells for seismic recording – so it can record data from within the reservoir and around it, getting a high resolution picture of wells in production

UK company Silixa records seismic data using geophone technology, Mr Naldrett says. into their workflows, and different companies fibre optic cables installed permanently in work in different ways. Sometimes they wells – so it can get a seismic recording from The progress has been led by advances work with independent seismic processing actually within the reservoir and near to it, so in instrumentation – measurement companies. getting a high resolution picture of wells in performance, signal to noise ratios and production. repeatabilities. One customer, an oil company in Qatar, had been operating a field for 10-15 years, and It takes advantage of the fact that the light This means that a fibre cable in oil wells finding that whenever it drilled new wells, flow through a fibre optic cable is perturbed the well performance wasn’t meeting the (changed) by changes in the sound around can “see” for miles laterally (to the side). simulation prediction. the cable, so it can act like a microphone. The system can also ‘see’ better than This led the asset manager to say, the only If there is an active seismic source, such conventional seismic where the reservoir way we can improve the reservoir model is as a vessel airgun, the seismic response at is below salt, because the seismic does not capturing more data from the field – leading the well can be recorded. It can also record have to pass through salt twice (down and to a decision to use distributed temperature passive seismic (where there is no active up) in order to see the reservoir. sensing, feeding back into reservoir and source – such as noise made by subsurface production models. It is possible to use the fibre optic cable to fracturing). record pressure, temperature and acoustics Subsea wells The same fibre optic cable can also record at the same time. So the same cable can pressure and temperature in the same be used for both production surveillance A next step might be subsea deployment of way, because the light flow is perturbed by (using pressure and temperature to monitor the technologies (installing fibre on a subsea changes in temperature and pressure. changes well, where the well head is on the seabed, not on a platform). Silixa aims to make it possible to get higher Working with the data quality seismic at lower cost – and generally Statoil did some surveys and found subsea make 4D seismic more widely available and Silixa aims to give data to oil companies fields had typically 10-15 per cent lower more frequent. Silixa’s customers typically in the most easy to use possible format. recovery than conventional wells. That could do repeat surveys every 6-12 months, rather Production data might be delivered in be attributed to having fewer wells, fewer than every 2 years, as usually seen for other PRODML format, and seismic data in SEG-Y entry points to the reservoir and a lot less types of 4D seismic, says Garth Naldrett, format. data. chief product officer of Silixa. Sometimes oil companies will ask Silixa “The value case in a subsea field Improving technology to get involved in interpretation of the development is much more extensive,” “The data – Silixa’s knowledge of the way the well costs are higher. You really want to The technology is gradually shifting from instrumentation records data can be useful optimise well and reservoir performance,” Mr being something for ‘early adopters’ to being for that. Naldrett says. accepted as a real alternative to conventional Other times oil companies want to feed it Exa and BP – get relative permeability from a digital rock sample Exa Corporation has developed software together with BP to model flows of multiple fluids through a digital image of a physical rock sample, and so find the relative permeability, a critical factor in understanding the reservoir Exa Corporation, a company based in The simulation solution was co-developed Exa claims that this is the first predictive Massachusetts, USA, has developed a way with BP, during a 3 year technology computational solver for relative to simulate fluid flow through a digital collaboration agreement. permeability for oil and gas. image of a physical rock sample without losing any resolution, working together It can be used to understand relative Relative permeability is the resistance to with BP. The technology is provided permeability – how multiple fluids flow flow for a mixture of fluids – for example as an online software product called through a reservoir, and the forces they a certain reservoir might allow water to DigitalROCK. will make on each other. flow through much more easily than oil. It is different to absolute permeability,

PETROMALL | INSIGHTS 9 which is the reservoir’s overall resistance “We plan to deploy this technology across Reservoir rocks nearly always begin to flow. our global portfolio. After a three-year with water in their pore spaces, and program of cooperative development and hydrocarbons percolate in there over time The relative permeability can be used to testing, our extensive validation studies and push the water out. understand what ultimate recovery can are drawing to a close.” be achieved from the reservoir (a function With Exa’s software, the simulation is of how much oil will be left behind in the BP uses the software as part of its “Digital made without simplifications to make pore spaces and never flow to a well). It Rocks” program, and the technology has the computer model easier to compute. can enable an understanding of how this been used across BP’s global portfolio Its simulation technique uses the full can be changed with an enhanced oil including fields in Angola, the Gulf of geometry of the pore space. recovery technique or water flood. Mexico, the North Sea, Egypt, Azerbaijan, the Middle East, India, and Trinidad and In the simulation you can see oil and The basis of the study is a 3D CT Tobago. water moving within the pore space, and (computerised tomography) scan of a see how pockets of oil are getting trapped. small piece of core or drill cutting. Clients BP’s Digital Rocks team includes experts There is a short video on exa.com website can take a scan image themselves, and in 3D imaging, fluid mechanics, numerical illustrating this. upload it to Exa’s online software, to run a modelling, computational physics, high simulation. performance computing, rock physics and The flow simulation takes into account reservoir engineering. The technology the conditions which the reservoir is Exa is a simulation software company, is implemented in its Centre for High under, and how the results will change for specialising in computational fluid Performance Computing (CHPC) in different conditions. dynamics. It also serves the automotive, Houston. aerospace and aviation industries. Using the data Replacing physical cores Exa provides purely software, provided The data about relative permeability can over the cloud. You can upload a 3D CT Until now, the only way to understand be used as part of reservoir models, used image, and start running flow analysis, how different fluids will flow through a for example to make decisions about getting results “in a relatively short time.” rock sample is to do a physical test in a where to place wells, and design enhanced laboratory, with a piece of core sample oil recovery techniques. BP agreement and test reservoir fluids, under similar Unless you understand the way different pressure and temperature to the reservoir Exa has been developing its flow fluids behave then any predictions conditions. simulation technology for a “couple of made by the simulators will not be very decades”, and realised it might be helpful Aside from the expense of setting this up, accurate. This includes simulations of how when used together with pore scale it means that you can only test a single injection water will push oil out of the imaging. core sample once, and so it is hard to pores and increase recovery. make comparisons. You can’t find out how The company met BP in 2014, who were The data can be used to work out the end the results might be different with for trying to solve the problem of relative point – how much oil you will actually be example a higher pressure water injection. permeability simulation. BP had done able to produce from the reservoir, or in Also lab results can take a year or even digital rock scanning, but not simulating other words how much oil will stay in the more. multiphase flow. pores at the point when no more oil is flowing to the oil wells. In May 2017, Exa announced it had signed Evolution of technology a multiyear “commercial agreement” with The recovery factor of reservoirs varies Oil companies have been scanning rock BP to provide its DigitalROCK relative greatly, from 20 per cent to 60 per cent, samples in tomography scans and using permeability software. and this is the major factor in the return the scan to model flow for about a decade the company gets from the investment in now. The difference with Exa’s technology BP said that the capability “will help building the oilfield. engineering teams to make more informed is that it does not simplify the rock decisions on wells, production facilities geometry at all for the modelling. The reason not all of the oil is produced and resource progression, including is because some of it is left behind in the Other companies have made a model enhanced oil recovery.” pores, trapped by capillary forces. of pores from the scanned image, which “The ability to generate reliable relative can be good for analysing porosity, or The data is also useful if the company is permeability information directly from single phase flow, but does not necessarily planning any water or CO2 flooding. digital scans on a much faster time-scale tell you how multiphase flow will travel than laboratory testing, and to gain insight through the rock, says David Freed, vice The water relative permeability also tells into the underlying pore-scale dynamics, president oil and gas at Exa Corporation. you how much water is being produced, provides substantial business value during something which operators also care Flow in real oilfields is nearly always appraisal, development, and management about, because it is expensive producing multiphase, Dr. Freed says, with oil and of our reservoirs,” said Dr. Joanne Fredrich, and handling that water – and it also gas, oil and water, water and gas, or all upstream technology senior advisor at BP, occupies topsides capacity. three. Having just one fluid is “extremely in a press release quote. rare” (except if it is water). And it is also important to be able to predict water production, so you can make

NOVEMBER / DECEMBER 2017 10

sure your topsides capacity is able to The technology could also be useful when make. This means there are no hard handle it. planning CO2 enhanced gas recovery / numbers available to justify making an enhanced oil recovery projects, such as investment decision about a carbon Oil companies are experimenting with using CO2 in coal seam gas fields. capture and storage project. surfactants (soaps) in injected water, which reduce the surface tension of the CO2 has been observed to reduce the By getting an understanding of relative fluid mixture – so changing the flow viscosity of oil, and reduce the amount of permeability using software like Exa’s, it conditions. With Exa’s technology, you gas which stays in coal seam pores, thus would be possible to get a much better can get a sense of how a surfactant will increasing oil and gas production. idea of the return of investment in such a change the hydrocarbon recovery, before project, and whether it might make sense you do it. But the precise mechanics are not very to do a pilot project. It would also help well understood, and detailed predictions compare different options. Carbon capture of what CO2 flows will lead to what improved oil and gas flows are hard to iRes – Geo: Improving understanding of reservoir behaviour using seismic UK company iRES Geo Technology believes there is room for improvement in how the oil and gas industry develops reservoir simulations from seismic data – and has developed its own software process to do it

UK company iRes-Geo Technology is no easy way to integrate the data Jim Farrington, believes that there is room for together from different disciplines, or SVP, iRes – Geo improvement in how the oil and gas even create an integrated visualisation industry develops reservoir simulations of the data, he says. from seismic data, and has developed its own workflow to do it as a one And the seismic data is underutilised, button software process. he believes. “People use it to pick the top reservoir, base reservoir, some Jim Farrington, senior geophysical important faults, that’s more or less advisor with iRes and previously a the end of the story,” he said. “Beyond geophysicist and geologist at an oil that, [they develop] the reservoir model changes in strain. major, says that one of the issues which using statistical approaches.” has “plagued the field” until now has iRES - Geo has proprietary software been that reservoir engineers’ primary Observation, simulation which can optimally generate ‘reservoir objective has been to “get a good and intermediate snapshots’ from 4D seismic data, history match,” their time trying to showing how the reservoir has changed match their models with new data as There are two basic pathways to over time due to production. it comes in (production data, well logs, achieving a greater understanding 4D seismic, fluid analysis), and they do of your reservoir’s performance If you want to build a reservoir it in a ‘patch fix’ manner and locating remaining oil – an simulation model, it needs to be as observation based approach (making geologically realistic as possible, and “It isn’t geologically realistic,” he decisions based on what you can see leverage all the available data, while says. “It gets worse and worse – the in 4D seismic) and a simulation based providing a good history match. whole thing snowballs. It is increasingly approach (making a reservoir model and difficult to get a good result.” simulating production / injection; data Reservoir models can be populated from which you then use for decision with both static parameters (from 3D Sometimes reservoir engineers making). seismic) or dynamic parameters (from introduce ‘fudge factors’ to try to get 4D). their reservoir models to work, such The 4D seismic approach is “something as saying a large area of the reservoir we are involved with very heavily,” he Reservoir modelling contains high porosity material, or says. “If you can afford it and have the workflow “parameter scaling factors”, or baffles resources and time, that’s the ideal in the model (walls obstructing approach.” You can do 4D analysis and iRes – Geo has developed cutting edge flow). interpretation on your data, and don’t reservoir modelling technology; “Single necessarily need to build a reservoir Pass Data Casting (SPDC) Closed Loop Also typically reservoir engineers, model. Reservoir Modelling,” this uniquely geologists and geophysicists work allows reservoir models to be generated on their own with different data and For example, you can use 4D seismic incorporating all the geological models, even if they are all sitting on to track the change in position of the realism contained in the seismic data. the same floor of the same office. There gas cap or water front, or understand It involves integration of seismic,

PETROMALL | INSIGHTS 11

Comparing conventional seismic inversion and well log mapping (left) with IRES-Geo’s Single Pass Data Casting (SPDC) Closed Loop Reservoir Modelling (right). Note how much better the IRES model matches the well log data (with production well in green on the left, and an injection well in blue on the right). The iRes-Geo model captures much more of the complexity of the actual subsurface, and allows a higher resolution reservoir model. petrophysical and geological data, field which has always caused trouble,” It is looking at proposing a pilot project he says. with Chevron in 2018, and working The entire workflow can be run with with Suncor. a single button press. “That’s our iRES - Geo has developed an alternative philosophy and it’s generated a lot of ‘dual-inversion’ technique, where iRes – Geo also works with Napesco interest,” Mr Farrington says. inversion for velocity is done at the in Kuwait, providing general G & G same time as the inversion for the services and more advanced, specialised Customers usually provide an initial desired reservoir parameter. 4D seismic services. reservoir grid, with accompanying petrophysical and geological data. This populates the reservoir model iRes – Geo has also entered into an at the initial grid scale with reservoir agreement with PSS-Geo, a geophysical iRes - Geo takes the seismic data, parameters located at the correct service company based in Stavanger sampled at regular time intervals such depth, with no need to upscale or and Oslo, specialising in 4D seismic as every 2, 4 or 8 milliseconds, and downscale, everything is at the same processing. maps it onto the reservoir grid with no resolution. “You lose data resolution loss of data integrity. – you introduce error every time you In August 2017, iRes - Geo announced a upscale and downscale,” he says. partnership with Stingray Geophysical, The next stage is to bring in a company that manufactures fibre petrophysical data, and define the The final stage is to take the resulting optic based permanent reservoir relationships between the seismic data reservoir models and use them to monitoring systems. and the static / dynamic reservoir generate ‘synthetic seismic cubes’, parameters, so you can integrate the working out what seismic response These fibre optic systems can record petrophysical data. The company has its you would have if the reservoir model seismic across a wide bandwidth, are own artificial intelligence (AI) software was the target of a seismic survey. cost effective, low maintenance and to derive the petro-elastic models. AI If this synthetic seismic cube closely light weight, which makes them easier can be used to make models which are resembles the recorded seismic, this is to place on the seafloor. They don’t nonlinear and interconnected, which an indication of the accuracy of your need big electrical supplies and have better reflect real life, rather than using reservoir model. a good signal to noise ratio. Their the standard calibration approach; fundamental advantage however is the curve – fitting through clouds of data Otherwise you can iteratively tweak high repeatability of the 4D monitor points. “We’ve seen a significant uplift parameters in the reservoir model until surveys. in reservoir parameter modelling results you get a closer match between the using this AI approach,” he says. synthetic and observed seismic data. In the short term however, they are expensive, so are usually only used IRES also has a different approach The output of the SPDC workflow can on the parts of the reservoir requiring to working with velocity data. The be input into any standard software, further study, such as complex areas of standard approach to reservoir such as PETREL for geological modelling the reservoir, or the crest, with reservoir parameter inversion using seismic data or Eclipse reservoir simulator. modelling and simulation covering the is to use a velocity cube to convert rest of the reservoir. the resulting reservoir model from the Clients and partnerships ‘time’ domain into the depth domain, There is work going on in the industry the physical domain we live in. iRes-Geo Technology was founded by to try to reduce the cost of 4D seismic, Dr Yi Huang, a formerly Lead Reservoir perhaps by reducing the amount of The velocity data used is generated as Geophysicist at Statoil. data being recorded, or the number of a by-product of the seismic processing geophones, Mr Farrington says. flow. “It is this selection of velocity iRes - Geo is working with Petrobras in Brazil, aiming to do a research project.

NOVEMBER / DECEMBER 2017 12 Kes Heffer - looking at well interference Kes Heffer, a former head of reservoir description at BP’s Sunbury Research Centre, is exploring how well interference – the amount that rate changes in one well affect other wells – can help better understand reservoirs Reservoir simulators and physics Prediction of production from a field requires a reservoir model of some sort. Generally, but particularly for fields at the exploration or appraisal stage, where production history is limited to a few well tests, the model used is a simulator that incorporates the geology and calculates fluid flow on a grid basis.

To account for the uncertainties in reservoir geology, reservoir engineers often run the simulations many times over with different input parameters drawn stochastically from the ranges of probability distributions. Because of the uncertainties, for a more mature reservoir you want to make use of the large amount of historical production data to update the parameters of the simulator: this is called Figure 1: The orientation of well pattern relative to permeability directionality can change oil recovery by tens of “history-matching” of the simulator. percentage points. It has long been known that directionality (anisotropy) in permeability has a large influence on areal sweep efficiency and therefore oil recovery. The above laboratory results for a five-spot well pattern (Caudle,1959) However this is not always a simple or demonstrate the sensitivity, whether at water breakthrough or at a more advanced stage with 95% watercut. quick procedure. “Workflows are speeding Kes Heffer, director of Reservoir Dynamics production well. up, but in the worst cases, by the time a Ltd and Honorary Fellow at Heriot history-match of production data up to a The data from well interference testing Watt University, has been studying how cut-off date from several months back was can be used to help calibrate reservoir fluctuations in flowrates at different wells achieved, the model no longer matched simulation models. in a field, including injection wells, can newly acquired data, and the cycle started correlate with each other. This is a form of all over again”, recalled Mr Heffer. “well interference”.

Mr Heffer previously spent 30 years working at BP, including as head of reservoir description at BP’s Sunbury Research Centre.

Well interference testing can be a good way to understand how much the reservoir geology is connected to allow fluid flow between the wells.

For example a high level of interference between a water injection well and a production well, means that an increase in injection pressure leads quickly to a measureable increase in pressure in the production well.

If such interferences are not roughly equal from an injection well to different neighbouring production wells, then injected water may be flowing preferentially in one particular direction, Figure 2: Frequencies of preferred directionalities of floods relative to local orientation of the major horizontal which might be directly towards a stress axis (Shmax): there is a strong bias towards Shmax for fields that would not be conventionally labelled as “fractured” as well as for those which are more obviously fractured.

PETROMALL | INSIGHTS 13

The unwieldy nature of simulators leads some operators to make predictions with a much simpler model, decline curve analysis, which extrapolates just historical production data semi-analytically.

An additional uncertainty in reservoir simulations that is often neglected is that of the physics of reservoir behaviour: the standard conventional model physics essentially incorporates only conservation of mass (balance of fluid flows and density change at each grid-point) assuming empirical relationships between fluid velocities, pressures, and fluid saturations. But over and above these, Mr Heffer maintains that “Geomechanics is playing a large part in fluid behaviour”.

This is particularly in terms of important preferential directionality. Figure 2 shows the plentiful field data that supports this statement: the direction that fluids move through the reservoir appears highly correlated with the orientation of the local Further analysis of flowrate correlations reveals linear features that overlay particular faults, implying that those faults are especially important to the geomechanics and fluid flow in the reservoir. stress state in the rock. directly in a “coupled” reservoir simulation The flowrates used in studies have Despite this, the only general recognition can mean a big overhead in computer time, generally been month-by-month data, of induced changes to the rock structure and also requires that reservoir engineers, as routinely reported to authorities or in most simulations is through an who build the models, are familiar with the partners; day-by-day data has also been assumed constant compressibility in the concepts of geomechanics, Mr Heffer says. used on one field, which allowed changes pore volume, which does not reflect any to the correlations through the field’s life directionality. Using well interference data to be followed: those made sense in terms of changing stress states. Incorporating more complex geomechanics The level of interference between wells can tell you something about the permeability A surprising finding from the studies was directionalities of rock. Mr Heffer analyses that flowrates can be correlated over interferences in terms of correlations large distances across a field. This ties in flowrate fluctuations because, unlike in with the concept that a reservoir is at pressure measurements, rate data are a geomechanical balance, possibly even readily available. As with waterflood before the first well is drilled: a small extra breakthrough data, flowrate correlations force, such as from a rate change at an show a strong bias towards Shmax (figure injection well, leads to changes throughout 3). the reservoir. The well interference data can be combined with whatever other In addition to overall stress-related information you have from the reservoir, directionality, Mr Heffer has found that such as fault data or measurements of the flowrate correlations can be further stress state. analysed to pick out particular major faults, implying that those faults are key Permeability directionality is often a geomechanical features and potentially consequence of fractures in the reservoir, strong flow paths across the reservoir. and Mr Heffer is currently working to develop a method of extracting both Given that flowrate correlations are related permeability and fracture density to the stress state of the rock, one can distributions across a reservoir. Figure 3: Data from several fields show that turn that around and interpolate stress correlations in the fluctuations in flowrates between pairs of wells, plotted as a function of orientation between measurements in a There are a number of other geophysical the orientation of the line between the wells, also more detailed map across a field using rate methods which would be complementary show a bias towards Shmax (green curve). In fact correlations: that has been done on one the correlations are negative on average between to these analyses, such as shear wave wellpairs that are separated along the minor large field to reveal that high watercuts in interpretation or microseismicity, which horizontal stress axis, orthogonal to Shmax: this wells are associated with locations where can also be sensitive to fractures, although is completely at odds with conventional reservoir Shmax is parallel to known faults. engineering. The bias towards Shmax persists when they are not always available, Mr Heffer the flowrate time series are detrended (filtered to says. remove low-frequency trends – red curve).

NOVEMBER / DECEMBER 2017 14 Paradigm k – a new cloud-based system for production engineers Oil and gas E&P software company Paradigm has developed a cloud-based system for production engineers that provides well surveillance data, reservoir simulations, and online collaboration capabilities.

Oil and gas E&P software company someone might post that they are planning Indy Chakrabarti, senior vice Paradigm has developed a cloud-based a workover for a certain well, and someone president software solution for production engineers else notes that the same well was worked of Product called Paradigm k, to help them perform over a few years ago and sends details of the Management at reservoir simulations and production outcome. Paradigm surveillance analysis. There are no software requirements to get “The task of production engineers is to started – you can run it from an existing maintain production targets. Historically reservoir model, or take whatever inputs you this has meant mainly surveillance, seeing have. It is hosted on Amazon Web Services. observation. However, using Paradigm what production currently looks like. Access software can help them put numbers behind to reservoir simulations will give them a Paradigm is initially offering the product their ideas, in order to compare production better understanding of why changes in to customers involved in shale oil and gas, improvements against any extra costs. production rates are happening,” says Indy where the modelling complexity can be most acute, particularly when modelling Chakrabarti, senior vice president of Product The data can be stored in different ways, fractures. Management at Paradigm. “It is a merger of including with data historians (typically used subsurface and surface workflows coming “Paradigm can offer data management as for production surveillance data), relational together.” part of the service, or oil companies can databases and models. manage the data themselves. Oil companies The system can be used by production A different kind of simulator engineers to test different plans for in-fill are increasingly paying attention to their sensor data, and historian software systems wells, and predict how much oil they might Paradigm has developed a different kind of for storing it, Mr Chakrabarti says.” “We can produce. They can investigate reasons why reservoir simulator for Paradigm k, which tap into those systems.” a well is not achieving its target production, uses the full resolution of the available or predict what might happen if you change Until now, production engineers have geological information, while running much the choke size or do an artificial lift. basically had two options if they wanted faster. to understand their reservoirs: Either The software can ‘ingest’ surveillance data Standard reservoir simulators divide the oversimplified, seeing the reservoir as a tank from well flowmeters and sensors. This data reservoir into tiny 3D boxes, and model the of hydrocarbons with no complex geology, can be used to update the reservoir model. parameters for each box individually. This or performing full-scale reservoir simulation, is a computationally intense process which The software also supports collaboration, which production engineers often find also requires simplifying the geological making it possible to share what you are challenging, Mr. Chakrabarti says. “That model into boxes. doing at each well with your colleagues, as process is onerous, and as a result, limits well as the results of those activities. who can do it”. The Paradigm k simulator, on the other hand, looks at the entire geology without Paradigm sees this evolving into a And the majority of reservoirs in the world simplification, and then uses equations to knowledge base around wells. For example, still do not have numerical simulations, Mr calculate the flows. This means that it does Chakrabarti adds. not require any reduction in resolution to Production engineer decision run, and can provide a simulation within making minutes, Paradigm says. Tests show that the outcomes of the The role of the production engineer includes simulator are very similar to those from a daily monitoring of the well (often referred more sophisticated simulator, in much less to as ‘production surveillance’) to see how computing time. individual wells are performing, whether injection systems seem to be helping, and This semi-analytical simulator has been used whether the company is on track to achieve on unconventional reservoirs, which have its production goals. fractures which are very difficult and time- consuming to numerically simulate. Production engineers also have to make more long-term decisions, such as whether “You don’t have to build a simulation to install artificial lift. Experienced deck for a production engineer,” he says. Paradigm k can predict fluid flow in complex fractured production engineers might be able to With this software, “We can represent the environments, in minutes – and without time understand well behaviour merely from consuming setup. fractures and the full complexity.”

PETROMALL | INSIGHTS