Reducing uncertainty in fracture modeling using dynamic data: McElroy field case

Transcription

1 Reducing uncertainty in fracture modeling using dynamic data: McElroy field case Carlos Collantes DWEP Appraisal Earth Modeler Chevron North America Exploration and Production Feb 10 th 2015 Gocad-SKUA User Meeting 2015

2 Agenda McElroy field basic introduction McElroy depositional model Static fracture characterization Streamline simulation and feedback Towards a static solution using dynamic data Deliverables Lessons learned

3 McElroy field basic introduction Discovered in mi south of Midland Carbonate reservoir (dolomite) wells drilled

4 McElroy depositional model Carbonate Ramp Model Evaporites & Mudstones Wackestones - Packstones Grainstones / Ooid Shoals Wackestones - Packstones Low Perm Area High Quality Area Low Press. Area East Flank Area

5 McElroy depositional model Prograding carbonate ramp 2018 wells w/ formation tops

6 Reservoir Characterization Pre-simulation static porosity 2011 Pre-simulation static permeability 2011 Initial characterization handles OOIP effectively

7 Static fracture characterization Static fracture characterization Matrix and fracture permeability Fracture permeability is just an overprint at 400 md

8 Simulation Results at field level via GTTI Oil Production Rate (STB/day) OPR Water Cut (fraction) WWCT FIELD OPR OPR-FSIM_31_INIT_PIMULT OPR-FSIM_31_FINAL_PIMULT FIELD 1 WWCT WWCT-FSIM_31_INIT_PIMULT WWCT-FSIM_31_FINAL_PIMULT Oil Prod Rate (STB/DAY) WWCT Date Date mc_hq_lp_variable_sort.hist 19 Nov 2012 mc_hq_lp_variable_sort.hist 19 Nov 2012 : Production Data : Before WCT matching : After WCT matching Conclusions: Production can be history matched using heavy hitter wells (80+ wells) We get the energy of the reservoir right, but input data needs to be refined to match individual well performance

9 Calibrate Geologic Model via History Matching Permeability Before and After Water Cut Matching via GTTI Before water cut matching K=1 K=5 After water cut matching K=10 K=15 K=20 K=25

10 Calibrate Geologic Model via History Matching Permeability Change Made via Water Cut Matching K=1 K=5 K=10 K=15 K=20 K=25 Feedback: Diagnose the geologic model Need more permeability in LPA Need less fractures and vugs in HQA and LKA

11 Towards a static solution using dynamic data Did we get lucky the first time around? Two different characterizations to be tested Short fracture, high heterogeneity Long fracture, low heterogeneity

12 Simulation Results Water Cut Match -Field Oil Production Rate (STB/day) : Production History : Before Water Cut Match : After Water Cut Match Water Cut (Fraction) LONG LONG SHORT SHORT

13 Permeability Change required for History Matching LONG case - Layer 5 (E-D5) Before History Matching After History Matching Change of Permeability required for History Matching Mostly increased permeability in LPA (Low Pressure Area)

14 Permeability Change required for History Matching SHORT case - Layer 5 (E-D5) Before History Matching After History Matching Relatively small changes in permeability Change of Permeability required for History Matching

15 Flood Efficiency Map Flux Map Colored by Flux Colored by Producer Colored by Injector

16 Deliverables Flood Efficiency Map Filtered Leave the connection showing flux movement from HQA to LPA Filtered out - Producer/injector paring with negligible amount of flux - Well connections supporting producers nearby (well-patterned flooding)

17 Deliverables Production forecasting using different pattern alternatives Expected ultimate recovery for drilling campaign CM107 Allocation factors CM517W 2% CM516W 11% CM518W 23% CM483W 12% CM315C 25% CM482W 12% CM458W 10% CM453W 1% CM457W 2% CM515W 1% CM462W 1% CM463W 0%

18 Lessons learned TheupdateoftheMcElroyearthmodellowerstheriskassociatedtothefuture oil production and strengthens the quality of its production forecast. The Central Permian AD team selected the SHORT model for pattern scenario testing. The forecast from the streamline simulation displayed that recovery associated with the one scenario was larger and the cost of drilling was smaller. Carbonate fields are extremely heterogeneous. Well based variograms cannot honor field heterogeneity. In order to account for it, heterogeneity needs to be forced into the model. The different permeability simulations in the earth model controls the well allocation factors in mature fields with dense well population, hence looking for common well allocation factors is not a value adding proposition.

19 Benefits of streamline simulation Validation of the reservoir energy Update of permeability field Testing of production scenarios/patterns Forecasting production Characterization of high swept areas Generation of development strategies per quarter section