1 Primary funding is provided by The SPE Foundation through member donations and a contribution from Offshore Europe The Society is grateful to those companies that allow their professionals to serve as lecturers Additional support provided by AIME Society of Petroleum Engineers Distinguished Lecturer Program www.spe.org/dl

2 Anne Valentine Principal Instructor, Production Engineering Schlumberger (retired) Integrated Historical Data Workflow: Maximizing the Value of a Mature Asset

3 Outline Workflow – Data required Case Study – History Initial underperformer identification Water and formation damage indicators Waterflood success – History Update Results of interventions Summary and Conclusions 3

4 Opportunity identification more important than ever Low cost, quick techniques to identify opportunities, for example: – Well interventions: acid jobs, squeezes, recompletions, refracturing jobs – Wells to shut in or reactivate – Improved waterflood management Can be completed within a few days 4

5 Goal: Improve production / recovery at low cost 5 Impact of technology and workflows Source: The Digital Oil Field – Oil & Gas Investor – April 2004 Cash Flow Explore Delineate DevelopProduce to Maturity + _

6 What data do you need? Historical Dynamic Data – Monthly Production – Monthly Injection (if applicable) – Pressures – Well Events Static Data – Petrophysical: Permeability, Porosity, Net Pay, Initial Water Saturation – PVT Properties 6

7 The workflow approach 7 Define performance and reservoir KPIs Measure & compare well KPIs and select underperformers Analyze reasons for underperformance Recommend actions for improvement Monitor and learn from results Buell, Turnipseed, “Application of Lean Six Sigma in Oilfield Operations”, 84434-PA SPE Journal Paper – 2004 D M A IC

8 Case study: A large waterflood 8 Ferrier field, Alberta, Canada Upper Cretaceous Cardium sandstone Low permeability Main waterflood area (303 wells) Original oil in place (OOIP) ~ 30 million m 3 Recovery factor (RF) ~19% An outsider’s “look-back” Jones, McCord, Cummer, “Reservoir Simulation Pays Big Dividends”, SPE 2428 Geological Atlas of the Western Canadian Sedimentary Basin – Chapter 23 Oil producer Shut-in producer Water injector Shut-in injector

9 Historical production Date: May 2010 Base 10-year forecast Expected ultimate recovery (EUR)  6.33 million m 3 ~ 21% recovery factor Goal: optimize production at a low cost 9 Liquid Rates – m 3 /d Oil Rate Water Rate Water Injection Rate Forecast Alberta Energy Regulator (AER) – public data Date

10 DEFINE KPIS D

11 Fundamental assumption Performance should be a function of reservoir quality How to define “reservoir quality”? – Flow capacity (kh) = Permeability x Net pay – Original oil in place (OOIP): proportional to hydrocarbon column (per well) = Net pay x Porosity x (1 – Initial water saturation) 11

12 Definition of “performance” Oldest well: 48 years of production Newest: 2 years of production An old well, even a poor one, normally has higher cumulative oil than a new well. For this field, cumulative oil is not a good indicator of “good performer” versus “bad performer”. 12 Number of Wells Producing Other options: Current rate (if same age) Lifetime average rate Peak rate Cum prod at x years Cum prod / Cum prod days EUR (uncertain) Combination of above

13 Selected indicator of “performance” Smooth (moving average) oil rate and select best value Data quality control – removes noise and anomalous points 13 Oil Rate – m 3 /d Actual Oil Rate Smoothed Oil Rate

14 MEASURE KPIS, SELECT UNDERPERFORMERS MD

15 Plot performance indicator vs. reservoir indicator Categorize wells and view on map 15 Initial underperformance identification Map Reservoir Indicator Performance Indicator Best Smoothed Oil Rate – m 3 /d

16 ANALYZE REASONS FOR UNDERPERFORMANCE A

17 Individual wells: Water production (overall water cut = 18%) Formation damage Wellbore or completion problems * – Perforations inadequate – Artificial lift restrictions – Surface constraints Overall: Waterflood management 17 Possible under- performance reasons

18 Water distribution Water production not generally a big problem Some individual wells - increasing water cuts Cumulative Oil Prod (Mm 3 ) Cumulative Water Prod (Mm 3 ) 18

19 Identify wells with above average water production Heterogeneity index (HI) compares each individual well with the group average HI = 0 for a well that behaves like the average Above average > 0, below average < 0 Calculate a running sum to see long-term trends Plot of two HI values shows trends 19 SPE 36604: Completion Ranking Using Production Heterogeneity Index SPE 138229: Performance Model Analysis for Candidate Recognition

20 Underperformers with higher water production 20 Cumulative HI Water Cumulative HI Oil High liquid High oil, low water High water, low oil Weak wells Date: May 2010

21 Water control diagnostics Technique to diagnose water production behavior “Chan plot”: – WOR (water-oil ratio) – WOR 1 (first derivative of water-oil ratio) – Versus cumulative days on production – Log-log scales Widely used Also applicable for WGR or GOR SPE 30775: Water Control Diagnostic Plots Plus many later papers based on this 21

22 Possible insight into water problems Log WOR Log WOR 1 Log Cum Days Possible Breakthrough Possible Water Coning Log Cum Days Water control diagnostics - Theory 22

23 Diagnosis of possible water source 23 Log WOR Log WOR 1 Possible breakthrough, but not conclusive Further investigation required

24 Avoid misleading conclusions 24 Log WOR Log WOR 1 Possible breakthrough? Not likely, highest water cut = 4% Log WOR Log WOR 1 Chan plots can be inconclusive

25 Signs of formation damage For damage during drilling or completion – Formation damage index (FDI) may be low FDI = Q / kh = oil rate / flow capacity For damage anytime – Gas/oil ratio (GOR) may be high due to pressure drop across damaged zone – Gas comes out of solution in the wellbore 25

26 Formation damage indicators Previously identified over- and under- performing wells indicated Potential acid candidates marked 26 Flow Capacity (kh) – md.m Gas/Oil Ratio (GOR) – MSm 3 /m 3 Oil Rate (Q) - m 3 /d

27 Locations of acid candidates 27 Best Smoothed Oil Rate – m 3 /d Map Hydrocarbon Column - m

28 Other reasons for underperformance Waterflood management is crucial Voidage replacement ratio (VRR) = injected volume / produced volume – Volumes include oil, water and gas and are expressed at reservoir conditions – Target VRR = 1.0 28

29 Voidage replacement ratio Waterflood as a whole is quite well balanced May 2010: – VRR = 1.17 – Cum VRR = 1.03 Monthly VRR Cumulative VRR VRR – m 3 / m 3 Target = 1.0 29

30 Pattern voidage replacement ratios 30 Cumulative VRR 0.2 – 0.8 0.8 – 1.2 1.2 – 2.0 2.0 – 8.0 Well balanced Decrease injection rate Increase injection rate Possibly stop injection

31 Injector analysis – Hall plot Skin analysis technique for injection wells Y-axis = Hall coefficient =  (  pressure x  days) 31 Hall, H.N. "How to Analyze Waterflood Injection Well Performance", World Oil (Oct. 1963) 128-130 1 = Damaged well 2 = Gradual plugging in well 3 = No change, no plugging, no damage 4 = Stimulated well or sudden channeling Hall CoefficientCumulative Water Injection

32 Example injectors 32 All types are commonly found, particularly type 2 Type 3 No damage Type 4 Stimulated Cumulative Water Injection - Mm 3 Type 2 Severe plugging

33 Hall plot slopes All injectors are shown Steeper slopes mean more resistance to injection 33 Low resistance to injection High resistance to injection Cumulative Water Injection - Mm 3 Hall Coefficient

34 Injectivity relationship to permeability We expect injectivity to be related to permeability Overlay resistance to injection on permeability grid map Not always related – investigate further 34

35 TAKE ACTIONS, MONITOR, LEARN FROM RESULTS I C

36 Three years later: May 2013 36 Interventions done in 2010 – 2011 in 40 wells identified here as underperforming Impact on field total to date  gain of ~ 60,000 m 3 compared to original forecast = rate increase of 57 m 3 /d Increase in EUR  ~ 220,000 m 3 Projected recovery factor ~21%  ~22% Oil Rate – m 3 /d Updated Forecast Original Forecast Date

37 Formation damage candidates Example well with intervention: gain ~1150 m 3 37 Oil Rate – m 3 /d Forecast workover

38 High water candidates No squeezes were carried out Higher cost intervention Problem was not severe 38

39 Injection changes – Balance waterflood Injector interventions unknown, field VRR changed little Some injection rates were decreased – 2010: 1082 m 3 /d – 2013: 768 m 3 /d 39 Liquid Rates – m 3 /d Oil Rate Water Rate Water Injection Rate Forecast

40 Results with vs without interventions Producers with interventions (40) Gain ~ 40,000 m 3 No interventions (95) Gain (maybe due to waterflood balancing) ~ 20,000 m 3 40 Oil Rate – m 3 /d Forecast Oil Rate – m 3 /d Forecast

41 Results Underperforming wells identified in a short time (2 to 3 days) – Producers – potential acid candidates – Injectors – plugging and/or resistant to injection – Patterns – where to increase / decrease injection? Action taken on 40 underperforming wells Some injection rate adjustments Gain in reserves ~ 220,000 m 3 (1.4 million bbl) Cash flow improved, life cycle extended 41

42 Conclusions This workflow is – Simple and effective – Flexible, can be adapted to multiple reservoir / field types – Able to handle huge amounts of data Key is to determine appropriate performance indicators with built-in quality control Demonstrates value of historical data Can result in production gains 42

43 Society of Petroleum Engineers Distinguished Lecturer Program www.spe.org/dl 43 Your Feedback is Important Enter your section in the DL Evaluation Contest by completing the evaluation form for this presentation Visit SPE.org/dl

44 Backup 44

45 Selection of performance KPIs 45 Poor correlation, poor choice Good correlation, good choices Best correlation, best choices

46 Possible Water Coning For a high rate well (e.g. 20 Mb/d) in good reservoir, the cone could reach > 200 ft high with width > 200 ft! 46 Water control diagnostics - Theory Further evidence of coning: When liquid production rate drops, WOR also drops.

47 47 Water control diagnostics - Theory Possible Breakthrough

48 48 Gas field example Strong wells – drop line pressure Weak wells, possibly stimulate Good wells Weak wells – drop line pressure &/or stimulate Cum HI Gas Cum HI Pressure

49 Additional workovers done 49 Best Smoothed Oil Rate – m 3 /d Hydrocarbon Column - m Six workovers done among the Expected group Results: rate increases with fastèr declines

50 Example well 50 Oil Rate – m 3 /d Forecast workover Multiple workovers – very short term results

51 Data gathering KPI selection ID Under- performers Water problems* Formation damage Completion problems* Waterflood analysis* Recommended actions 51 Workflow – Define, Measure & Analyse stages If applicable If data available

52 Selection of completion KPIs Two wells, same reservoir quality – we expect better performance from “better completion” Completion indicator depends on data available Vertical / deviated / horizontal Meters perforated / open Frac job data (e.g. fluid volume) A combination (e.g. fluid volume / completion length) NOTE: not analyzed for case study due to lack of data 52

53 CI: Completion Length PI: Best 6 Mo Gas Rate Shale gas example – completion KPI 46 Completed poorly and poor producers Completed OK but poor producers Good Candidates Good completion and production Horiz