Reprinted from Petroleum Technology Quarterly Q1 2012 Cost estimating for turnarounds www.eptq.com

Cost estimates for refinery turnarounds can lack accuracy, but lessons learned from capital project estimating could improve matters

Gordon Lawrence Asset Performance Networks

urnarounds are major events turnaround cost estimates. Rather, it typically take place on a four-to-six for refineries and other petro- attempts to lay the groundwork for year cycle. The length of a typical T chemical facilities. They a discussion on how to improve cost turnaround execution phase (ie, the typically cost significant sums of estimating in process facility turna- period when the facility is shut operational expense (opex) and capi- rounds. It highlights the point that down and hydrocarbon free) is tal expense (capex) money to cost estimating in the turnaround usually around three-to-five weeks. execute. They cause lost opportunity world is currently not very accurate. The scope of a turnaround typi- cost through lost production while It then goes on to look at ideas on cally includes: the facility is shut down. If poorly how to improve cost estimate accu- • Inspection of equipment to managed, turnarounds offer signifi- racy, by looking at ideas that might company regulations or governmen- cant risk of accidents. If poorly be adapted from the capital project tal rules executed, they can also be the cause world. In particular, it looks at the • Inspection of pipework for corro- of significant production disruption stage gate approval system, the sion and erosion damage, both after startup, due to leaks and other development of scope and estimate internal (process weak points) and production trips. Hence, there is a bases for different levels of estimate external (corrosion under insulation, potential to save significant sums of accuracy, and at the development of or CUI) money, by ensuring turnarounds are allowances for “known unknowns” • Cleaning, repair and maintenance run correctly. and contingency for “unknown of equipment, pipework and instru- However, despite the large poten- unknowns”. Finally, it makes some mentation (pulling and cleaning heat tial for saving money, there has, suggestions for the next steps in exchanger tube bundles, repairing until recently, been very little focus developing better turnaround cost leaks in pipework or checking of in the turnaround world on making estimates. pressure relief valves) sure turnarounds are run cost effi- • Minor upgrades and modifications ciently or on finding ways to What is a process industry to the facilities (items controlled improve cost estimating and execu- turnaround? under the “management of change” tion skills. In the past seven to eight A turnaround (in the context of the [MoC] procedures) years, this attitude has begun to process industries) is defined by the • Tie-ins for capital projects. change and turnaround teams are American Petroleum Institute as “a beginning to look at how they can planned, periodic shutdown (total or Historically poor estimation and improve. One potentially rich source partial) of a... process unit or plant execution efficiency of ideas on how to improve is for to perform maintenance, overhaul Historically, there was a tendency the turnaround teams to examine and repair operations and to inspect, among operating companies to view how their capital project brethren test and replace process materials turnarounds as an inevitable and estimate and execute projects. In the and equipment”.1 Refineries and necessary evil, and to accept that capital project world, the potential other petrochemical facilities that they would “cost what they cost” for “leaving money on the table” run on a continuous rather than a and “take as long as they take”. due to poor project estimating and batch production cycle must, every However, in recent years, there has execution has meant that decades of few years, shut down operations to been a growing recognition that this time and effort have been spent, and provide access to the production attitude leaves “money on the table” continue to be spent, on developing units in order that essential mainte- in the form of the opportunity cost techniques to ensure that estimation nance, modification and inspection of lost production while the facility and execution is carried out work can be carried out that could is shut down for longer than neces- efficiently. not be done while the units are in sary; and unnecessarily excessive This article focuses on cost estima- operation. expenditure of money during the tion. It does not presume to provide Turnarounds are events that are turnaround, through running the all the answers on how to improve planned well in advance and turnaround inefficiently.

www.eptq.com PTQ Q1 2012 33 In order to improve turnaround efficiency, there is a need to examine the estimation of costs for turna- rounds, the estimation of schedule time, and the efficiency of planning and execution of the turnaround. Figure 1 Expectation of accuracy in turnaround control estimates: vox populi This article focuses on the first of those three areas, the estimation of improve cost estimating accuracy in results in a ±50% accuracy estimate, costs. turnarounds. the following conceptual design stage results in a ±30% estimate, and Accuracy of turnaround cost Comparing capital projects with the final basic design stage results in estimates turnarounds a ±10% estimate. Hence, the focus of A cost estimate needs to be accurate Cost estimating of capital projects in the stage is in developing a scope in order to provide management the process industries has gone basis to meet the estimate basis with the information needed to through a number of improvements requirements for the level of accu- decide how to proceed, to allow over the years. It may well be that racy intended cash flow planning and to aid in some of the lessons and devices • The project cannot proceed to the firm control of expenditure. used in capital projects can provide next stage without first going In a survey of conference partici- a basis from which to develop a through a stage gate review to check pants at the Turnaround Industry system for better turnaround whether it has achieved the required Networking Conference (TINC) — estimates. scope and estimate basis Europe, held in March 2011 in • Full funding of the project is Amsterdam, The Netherlands,2 83% Stage gates usually only received at the end of of respondents said that their turna- Capital projects the third, basic design phase. round control budget was intended In the capital project world, it is now The focus, therefore, in the capital to be a ±10% estimate (see Figure 1).3 generally accepted that front-end project world is on calculating “How The remaining 17% said that their definition is the key to a successful much money do we need in order to estimate was supposed to include project. In order to achieve good carry out this scope?”. sufficient contingency/reserve that front-end definition, capital project it was a “not to exceed” number. teams have largely moved toward a Turnarounds None said that their budget was stage gated approach, whereby the In the turnaround world, there has intended to be a ±30% estimate. design proceeds through three been a movement in recent years We then took a sample of 93 recent distinct stages of ever-improving towards a form of stage gated turnarounds and in each case scope definition.4 Key points are: approach, in an attempt to improve compared the total actual cost for • The length of each stage is a func- front-end definition. However, the the turnaround (including both capi- tion of the work required to be system is not yet as clearly defined tal and maintenance work) with the completed in that stage as in the capital project world. control budget. Figure 2 plots the • The culmination of each stage is a There are some key differences in over- or under-run of the actual cost estimate, with the estimates the way the approach is applied. costs, expressed as a percentage of getting more accurate with each The length of each stage is (more or the control budget. It shows the succeeding stage. Generally speak- less) fixed, based on the amount of mean (the horizontal bar) and 80% ing, the feasibility stage of a project time remaining until the turnaround confidence range (the vertical bar) of the results in our sample set and 170 compares it to the results we would 160 have expected if the cost estimates P90 truly were ±10%, or if they were 150 ±30%. 140 Based on the results shown in 130 Figure 2, we can say that turnaround 120 Mean cost budgets tend, on average, to 110 under-estimate the actual cost by 100 around 16% (the distance of the 90 P10 “mean” bar from the 100% point) N=93 and the budgets themselves show a 80 variability of closer to ±30% accu- 70 proportion of the budget, % proportion Actual turnaround costs as a Actual turnaround racy than their purported ±10% 60 accuracy (the vertical line shows the Theoretical range Theoretical range Turnaround 80% confidence range around the ±10% estimate ±30% estimate actual costs mean). Clearly, there is an opportunity to Figure 2 Accuracy of turnaround control estimates: expectation and reality

www.eptq.com PTQ Q1 2012 35 starts. Generally, by about 14–12 months before the turnaround, the ±50% ROM ±30% ROM ±10% first review of status will be held. A estimate estimate control estimate second review may be held around nine-to-six months before and the General project data last review about three-to-two Project scope General Defined Defined months before. description The stages are not focused on Facility capacity Assumed Defined Defined scope definition for cost estimates of a given accuracy. Instead of the Facility location General Specific Specific issuance of an estimate marking the end of a stage, the ends of these Ground surveys None Defined Defined stages are delineated by their prox- Project imity to the start date of the None Defined Defined turnaround. There are guidelines execution plan on what level of scope definition Contract strategy Assumed Preliminary Defined should be expected at each of these turnaround stage gates, but these Level III – Level I Level II detailed Project schedule guidelines are not as universally milestones preliminary resource loaded adopted as their counterparts are schedule on capital projects. By the nature of a turnaround, Cost estimating plan None Defined Defined refusing to let a turnaround team Engineering deliverables proceed to the next stage because it is insufficiently prepared at the time Block flow diagrams Outline Complete Complete of the stage gate review is not usually a feasible option. The budget Plot plans None Preliminary Complete tends to be fixed back at the ±30% or even ±50% estimate stage. Process flow diagrams None Complete Complete Consequently, with turnarounds, the focus tends to be on “How much Utility flow diagrams None Preliminary Complete scope can we carry out for this amount of money?” rather than the Project execution plan None Preliminary Complete capital project paradigm of “How much money do we need in order to P&IDs None Preliminary Complete carry out this scope?”. Heat and material None Preliminary Complete balances Scope basis and estimate basis Process None Preliminary Complete In this section, we make a distinc- equipment list Utility tion between the scope basis (the None Preliminary Complete level of definition of the design equipment list Electrical single line deliverables and documents) and None Preliminary Complete the estimate basis (the methodology diagram Process engineers used to calculate costs, based on None Preliminary Complete equipment datasheets those deliverables). Mechanical engineers None Preliminary Complete Capital projects equipment datasheets Equipment general None Preliminary Complete In the world of process industry arrangement capital projects, the scope basis and the estimate basis are both relatively well defined for various Table 1 levels of estimate accuracy, from ±50% down to ±10% or better. One written “basis of estimate” approach of calculating from the well-known example of where this document. bottom up. The cost of a facility is is documented is AACE International typically factored off such details as, Recommended Practice No. 18R-97.5 Early estimate: ±50% — rough order of for example, the production capacity Lawrence also lists typical require- magnitude of the new facility, with adjustments ments.4 (Table 1 and Table 2 are In capital projects, the first cost made for inflation and location. Such adapted from Lawrence, 2008.) In estimates are usually highly deter- estimates generally have little in addition, it is accepted practice to ministic, using high-level factoring terms of scope basis, often merely document the cost estimate in a methods rather than the stochastic an outline of the facility capacity

www.eptq.com PTQ Q1 2012 37 the previous turnaround scope, and ±50% ROM 30% estimate ±10% control the estimate basis uses that historical estimate estimate total cost.

Conceptual estimate: ±30% "ASIC $ETAILEDESTIMATE #OMPLETE DESIGN On turnarounds, teams appear to aim to have the overall scope frozen "UDGETESTIMATE and challenged by this point. %NGINEERING ORFACTOREDOFF &IRMQUOTE &ACTORED BASEDON EQUIPMENT However, this has not always been PREVIOUS SIMILAR achieved. In terms of deliverable FACILITIESBASEDON "UDGETQUOTESFOR &IRMQUOTESFOR documents for the scope basis, the MAJORITEMS MAJORITEMS %QUIPMENT PRODUCTIONCAPACITY mechanical workpacks are usually FACILITYSURFACEAREA DATABASECOSTS BUDGETQUOTES SIMILARMETRIC FORREMAINDER FORREMAINDER developed. The estimate basis is a "UDGETQUOTES mix of deterministic and stochastic, "ULK &ACTOREDOFF BASEDONDETAILED as the mechanical work packs are MATERIAL EQUIPMENT MATERIALTAKE OFFS costed (in-house) in some detail, but the other work (electrical, instru- "UDGETQUOTES #ONSTRUCTION &ACTOREDOFF mentation scaffolding, insulation, BASEDONDETAILED EQUIPMENT LABOUR MATERIALTAKE OFFS and so on) is usually factored off the mechanical costs, using historical factors. Table 2 Authorisation/control estimate: ±10% required. The estimate basis relies for process industry turnarounds. In Some turnaround teams do now on historical data of the total cost of the absence of such a document, we have all discipline scopes developed the entire facilities rather than have informally surveyed a number and externally costed. However, in detailed cost data. of turnarounds to look at the vari- our experience, the number of teams ous scope and estimate bases used that take their estimates to this level Conceptual estimate: ±30% by the turnaround teams to develop of detail are relatively few, despite In capital projects, a considerable estimates of nominal accuracies of the fact that most teams claim their amount of work would be completed ±50%, ±30% and ±10%. control estimates are ±10% or better. on the scope basis, coming close to The first problem we have encoun- freezing the overall scope and devel- tered is that there is some variation Work breakdown structure oping many of the design drawings in the methods used. Second, the Capital projects by the time this estimate was concept of writing down a detailed All companies have their own varia- prepared. “basis of estimate” document is tions on how they prefer to break The estimate basis would have something that still has to take root down costs within the estimate. moved from a high-level, determin- in the turnaround world; hence, Indeed, the work breakdown struc- istic estimate to a mix of the understanding how each estimate ture (WBS) may differ slightly, from stochastic and deterministic, and the was developed is difficult. (We were project to project, depending on the team moves towards Lang factors or occasionally given documents scope. Nevertheless, at a high level, other, similar methods to calculate described by the team as the “basis there is a measure of agreement the project cost (factoring off one of estimate”, but these were gener- across the industry on how to break element of the project scope, most ally extremely simple, high-level down capital project costs. The typically the equipment cost). notes on the estimate spreadsheet breakdown tends to be as shown in itself rather than a formal and Table 3. Authorisation/control estimate: ±10% detailed document.) Nevertheless, By this stage, the scope basis on a we have attempted to provide a Turnarounds capital project is generally very general overview of the methods At present, there is no clear, common detailed, with completed P&IDs, being used by turnaround teams for agreement across the process indus- order specifications for major equip- the various estimates. tries on a high-level WBS for ment and so forth. The estimate turnaround costs. Most turnaround basis is now stochastic, built bottom Early estimate: ±50% — rough order of teams split costs into direct costs — up from the material take-offs and magnitude costs that can be specifically assigned using market cost data rather than The early estimate is frequently to a cost object, such as an item of historical database information. merely the cost of the last turna- existing plant — and indirect costs round, adjusted for inflation, with — costs that cannot easily be directly Turnarounds perhaps an educated guess as to attributed to a cost object. However, There appears to be no document how the scope may have grown or the definition of what should be equivalent to AACE International shrunk since the last turnaround. considered direct or indirect cost Recommended Practice No. 18R-97 The scope basis is therefore merely varies from site to site (presumably

38 PTQ Q1 2012 www.eptq.com because of different internal account- that consumables and bulk materials allowances” allocated to each line ing practices). occasionally are not separated. item for the “known unknown” of The costs of owner supervision are that line item (for instance, wastage generally agreed to be indirect, since Allowances for uncertainty in the piping material take-off, at any hour of the day a team could Any cost estimate, by nature of it design development in equipment). be supervising work on a range of being an estimate, requires funds for The amount of design allowance existing plant areas. Contracted the uncertain elements of the scope. to include is usually determined labour on tasks that can be assigned These uncertain elements can be through expert judgment based on to specific existing plant items is divided into “known unknowns” historical data. generally categorised as direct cost. and “unknown unknowns”. The Furthermore, tracking this cost by “known unknowns” are allowances Unknown unknowns plant unit/area and by equipment for items or quantities that the esti- “Contingency” is for the “unknown class/type and by equipment mator knows, from historical unknowns” in the estimate (for number is relatively common. evidence, will be needed in a specific instance, it is discovered during However, costs such as rented line item of the estimate. The construction that the need for an equipment (cranes, heat exchanger “unknown unknowns” are the items electrical transformer was over- bundle pullers, and so on) are less or quantities that were under-esti- looked in the design). consistently assigned to either direct mated or completely overlooked in As discussed by Burroughs & or indirect categories, presumably the estimate. The estimator does not Juntima,7 the amount of contingency because such items can only some- know exactly in which line item the to include is typically calculated in times be attributed to a specific plant funds will be needed, but history one of four ways: item (for instance, if a crane is only tells the estimator that some funds • As a predetermined percentage needed to work on a specific item). will be needed to bring the estimate (of the base estimate) Theoretically, it should be possible to the P50 point. These “unknown • Through expert judgment (gut to assign bulk material (pipe, valves, unknowns” are contingency funds. feeling/experience) based on histori- for instance) to specific plant items The need for contingency is a hotly cal data and hence categorise the material as debated subject, but has been • Through a Monte Carlo-type direct. However, this is not always discussed elsewhere6 and we do not review of the cost effect of expected done. One would logically expect propose to explore this topic here. risks consumable materials such as weld- • Through use of a regression ing rods to be indirect costs because Capital projects model, based on historical data. of the difficulty of assigning them to Known unknowns specific plant items. However, the The “known unknowns” in capital Turnarounds situation can be clouded by the fact projects take the form of “design In the case of funds for uncertainty, there is a little more general agree- ment than perhaps in other areas of Allowances for estimating for turnarounds, in as ‘known unknowns’ much as it is generally accepted that

Design funds are needed for “unknown” items. Home Office Engineering Known unknowns Project In turnarounds, the “known management/supervision unknowns” are (a) “emerging work” Equipment Design Escalation (work to repair items that have allowances allowances on each on each broken/failed between the comple- Capital Direct Bulk materials line item line item tion of the estimate and the start of the turnaround); and (b) “discovery Field Construction labour work” (repair work that is discov-

Temporary facilities, ered during the turnaround once craneage, scaffold, etc equipment or pipework is opened Indirect Construction up for detailed inspection). management/supervision In common with the design allow- Allowances for ances on capital projects, estimates Contingency ‘unknown unknowns’ of likely emerging work and discov- ery work should (in theory) be Commissioning Expense calculable and assigned to specific Operations staff estimate line items. (For instance, we can expect other X valves to fail between now and the shutdown, or Table 3 based on past experience we should

40 PTQ Q1 2012 www.eptq.com expect that the trays in column Z will be discovered to be damaged 22 when we open it up.) 20 +1SD However, what we have observed is that very often such items are not 18 reported in the estimate to this level 16 of detail. Instead, they tend to be 14 listed as single line items on their Mean own. In addition, anecdotally, we 12 observe that very often the discov- 10 ery work in particular appears to be underestimated — an optimism 8 about what is likely to be found of the base estimate, % 6 −1SD N=38 seeming to prevail. allowances as proportion Total 4 The amount of emerging work Emerging work + discovery work allowance and discovery work + contingency allowance appears to be calculated on the basis of experience from previous turnarounds or from inti- Figure 3 Uncertainty allowance included in estimate mate knowledge of the state of the facility. But when it is based on included by turnaround estimators of just under 13% of the base esti- experience of previous turnarounds, and how much was actually used. mate (where base estimate = the this experience is rarely backed up total cost estimate before the three by hard numbers, showing what How much is included? uncertainty allowances are included) was spent in previous turnarounds The first problem that we encoun- as a total of all three allowance as a percentage of the base scope tered in this study was that while amounts. cost. most turnaround teams recognise That funding breaks down in turn the three main areas of unconfirmed into an average of 16% for emerging Unknown unknowns scope that need funding as being (1) work, 38% for discovery and 46% In turnarounds, just as in capital emerging work, (2) discovery work for contingency (see Figure 4). projects, a contingency for the items and (3) contingency for the unfore- that were overlooked or completely seen, few teams clearly differentiate How much is spent? unforeseen at the time of the esti- between them in their estimates. We then attempted to look at how mate issue is required. (For instance, Indeed, in a survey that we carried much in each category is typically money was included in the “discov- out, out of 216 turnarounds, only 34 consumed. From this information, we ery work allowance” to spend three (or 16%) could give a clear differen- hoped to begin developing guidelines days repairing trays in column X, tiation in their cost estimate between on calculating how much should be because the column was old and these three areas and the main included in an estimate. However, was operating at low throughput. estimate. we found that although some teams However, it took two days longer Nevertheless, from the data gath- have begun to track the consumption than expected to repair the trays in ered we are able to show (see Figure of the three elements in their actual column X. That extra two days’ cost 3) that teams typically include a total costs this practice is still immature comes out of contingency.) within the turnaround community. Whereas in capital projects contin- Without historical data on actual gency is calculated on the basis of expenditure in these three catego- one of four methods, each of which ries, it is impossible to “close the has at least some logical foundation Emerging loop” and begin developing guide- to it, in turnarounds we observe that work lines on how much should be 16% contingency often appears to be included in each category in an calculated simply as the difference Contingency estimate. between the current base estimate 38% Discovery Conclusions plus discovery and emerging work work allowances and the original, rough 38% The problem estimate quoted to management Currently, turnaround cost estimates (assuming the rough estimate to be are inaccurate in that, on average, a higher number!). turnaround costs overrun their esti- N=34 mates. They are also highly Quantification of turnaround unpredictable in the level of their allowances Figure 4 Breakdown of uncertainty inaccuracy (that is, the overrun We then looked further into how allowance into emerging, discovery and does not remain within the claimed much allowance for uncertainty was contingency ±10% accuracy of most estimates).

www.eptq.com PTQ Q1 2012 41 Furthermore, there is no clear ±50% ROM ±30% ROM ±10% agreement on the definition of estimate estimate control estimate what information a cost estimate should be based upon in order to General turnaround scope data allow it to be defined as a ±10% 95% 99% estimate. Equipment count Preliminary complete complete There are no guidelines on how to 100% 100% Turnaround premises Preliminary calculate uncertainty allowances for complete complete emerging work, discovery work and Listing – high contingency. Nor are many teams level, based Risk Based Inspection 80-90% 100% gathering the actual cost expendi- on asset (RBI) data complete complete management ture data that would allow such strategy guidelines to be developed in the Maintenance turnaround Listing – little 60-70% 99% future. backlog definition complete complete Hazard and Operability Listing – high 60-70% 100% Learning from capital projects study (HAZOP) level complete complete The capital project world historically Process Hazard Analysis Listing – high 60-70% 100% had similar problems with cost esti- (PHA) level complete complete mating. Over the years, it has Management Of Change 40-60% 100% improved through: (MOC) design changes Listing complete complete • Tying estimates to stage gates Listing – 80-90% 100% • Clearly defining the scope basis Capital projects various complete complete stages and estimate basis required for a 80-90% 100% certain level of estimate accuracy Plant projects Listing complete complete • Using similar work breakdown Plant engineering design Listing – not 60-80% 100% structures, allowing comparison changes (expense) all inclusive complete complete benchmarking of data Leak Detection And Listing – Listing – Complete • Gathering data on allowances for Repair (LDAR) preliminary preliminary listing uncertainty and developing logical Previous turnaround methodologies for estimating those Listing – 90% Listing – 90% Listing – 90% scope allowances. Engineering deliverables The turnaround world has already begun to learn from capital projects Plot plans – Baseline – Partial logistics markups no markup markup Complete by copying the stage gate approach. There seems to be little reason to Turnaround boundaries None Preliminary Complete suppose that turnaround teams Process engineering cannot continue to learn from capital design basis for plant None Partial Complete projects in order to develop turna- engineering capital round-specific standards for scope Plant engineering basis, estimate basis and calculation None Partial Complete design packages of allowances. Major capital project None Partial Complete engineering design Recommendations for immediate Heat transfer data None Preliminary Complete action If the estimation of turnarounds is Partial Original to become more accurate and turnaround P&ID markups baseline Complete markups predictable, we recommend that, as Original Equipment Baseline – a first step, a series of actions should Partial for Manufacturer (OEM) none for new Complete be adopted within the turnaround specifications & drawings capital new capital community. Startup procedures – maintenance None Partial Complete Define the scope and estimate basis (new equipment) There should be wide agreement on Utility plan None Partial Complete what the scope and estimate basis is for calling an estimate 50%, 30% Oils plan None Partial Complete or 10% accurate. The capital projects community has such agreement in Flare minimisation plan None Partial Complete documents referred to earlier. Suggested outline versions for turn- arounds are given in Tables 4 Table 4 and 5, respectively.

42 PTQ Q1 2012 www.eptq.com Standardise the WBS ±50% ROM ±30% ROM ±10% control Move towards a common standard estimate estimate estimate layout for cost estimates. The capital Quantities for mechanical work projects community has such a Mechanical Quantities for all from workpacks, work disciplines standard (see Table 3). A suggested but then priced calculated and outline version for turnarounds is Cost of the last in-house TA, adjusted for priced by shown in Table 6. inflation Other disciplines contractors, based on Other factored from workpacks Track the uncertainty allowances disciplines mechanical, based on history If the turnaround community is ever to develop even rudimentary rules of thumb for estimating Table 5 emerging work, discovery work and contingency for the unforeseen, but not only, relating to Table 4) of John Casper, 4 Lawrence G R, Stage gated approval processes it must begin to track expenditure Senior Consultant and Subject Matter Expert — a practical way to develop and filter capital in these areas. This will require two for Turnarounds and Capital Projects, at Asset investment ideas, Pharmaceutical Engineering, responses from the community: a Performance Networks. vol 20, No 2, March/April 2008. 5 AACE International Recommended Practice willingness to differentiate clearly References No. 18R-97; cost estimate classification system, the three categories in cost esti- 1 http://www.api.org/aboutoilgas/sectors/ as applied in engineering, procurement, and mates; and a willingness to refining/refinery-turnaround.cfm. construction for the process industries. accurately track actual expenditure 2 http://ap-networks.com/latest-news/tinc- 6 Lawrence G R, The use and misuse of from these three categories during europe-2011-summary.html. contingency — or why cutting contingency a turnaround. 3 We are very aware of the fact that in recent makes projects more expensive, not less. years the preference within project cost Pharmaceutical Engineering, vol 27, No 5, Sept/ Tie the estimates to the stage gates estimating circles is not to talk of a typical Oct 2007. In the world of capital projects, the ±10% or ±30% estimate, but rather to talk 7 Burroughs S E, Juntima G, Exploring Techniques stage gate culminates in a cost esti- of “classes” of estimates. This is because of for Contingency Setting, AACE International mate. Currently, the typical review the need to recognise that for the same level Transactions, Washington DC, 2004. of scope definition, two projects could have stages for preparation and planning different levels of estimate accuracy because Gordon Lawrence is a Senior Consultant with of a turnaround are not tied to a cost of the inherent risks in their characteristics. For Asset Performance Networks, Amsterdam. estimate. Tying the estimate and example, a new technology project probably He holds a bachelor’s degree in chemical stage gate together could improve carries more risk than a similar-sized project engineering from Heriot-Watt University, review of planning and preparation expanding an existing plant. However, the Edinburgh, a master’s in biochemical progress. turnaround world has not yet reached that engineering from Birmingham University, and level of sophistication. Hence, for the survey, an MBA from Strathclyde University Business Recommendations for longer-term we used terms that are readily recognised, such School, Glasgow. next steps as ±10% estimate. Email: [email protected] This article covers only the first steps in improving the accuracy and Allowances for predictability of turnaround cost ‘known unknowns’ estimates. Further steps need to be Turnaround planning Home Office taken. Once there is common agree- and preparation ment on the minimum standard of Field (before the Pre-turnaround field scope basis and estimate basis for shutdown period) work turnaround estimates, and once Direct Equipment (broken down by Bulk materials turnaround teams begin gathering unit/area, equipment actual uncertainty allowance data, type and equipment Turnaround field number labour the community will then be in a Emerging Discovery Consumables position to begin developing logical Maintenance Field work work (during methodologies for calculating more Equipment rental – allowances allowances shutdown cranes, scaffold, etc on each line on each line accurately the required level of period) item item Indirect Logistics, temporary allowance for uncertainty in emerg- facilities, etc. ing work, discovery work and Turnaround contingency for each turnaround. management/ supervision

Allowances for Acknowledgements Contingency ‘unknown unknowns’ The author would like to acknowledge the Direct encouragement offered during the preparation Capital projects of this article by Bobby Vichich, Vice President, Indirect Turnarounds, at Asset Performance Networks, as well as the invaluable advice (particularly, Table 6

www.eptq.com PTQ Q1 2012 43