Complex Construction vs. Simple Deconstruction: Alternative Workflows and the Role of Ultimate Truth Models

Transcription

1

2 Complex Construction vs. Simple Deconstruction: Alternative Workflows and the Role of Ultimate Truth Models Mark Bentley, Ed Stephens AGR TRACS Tom Buckle, Rhona Hutton Heriot-Watt

3 Swanlake bay, Pembrokeshire 3

4 Complex models 50m x 50m x 3m 25m x 25m x 1m 4

5 Understand one heterogeneous bed

6 ~2.5cm Ultimate Truth ( almost) 2D cross-sectional model Typical offshore well spacing Core Plug ~5cm Cell resolution close to the scale of the input data (SCAL) 1cm Grid Cell 5cm 4m Water Injector Producer 3.2 million static cells 400m 320,000 dynamic cells (the full field equivalent would be a trillion cells)

7 Static build concept-driven Sedimentary log Conceptual sketch 4m 50m If you can sketch it, you can model it.

8 Permeability Model elements Cross-Bedded Sand (0) Channel Sand (1) Laminated Sand (2) y = e x 100 Conglomerate (3) Fine Sand (4) Homogeneous Sand (5) y = e x y = e x y = e x y = e x 10 1 Ripple Sand (6) Silt (7) Silt Nodules (8) y = e x y = e x y = e x 0.1 y = 0.082e x Porosity Petrophysically distinct 3D elements

9 Static model Model elements from multi-point statistics (MPS) Porosity (frac) Permeability (md)

10 Capillary pressure, psi Relative permeability, frac Dynamic model build FLUID Light oil (35 API), undersaturated SATURATION Mixed wet (l w 0.5) 2 1 Water / oil Pc vs Sw f = 0.20, k = 100 md S wc 0.20, S or Oil Water / oil Kr vs Sw Water Spontaneous Forced Pcw Dr Pcw Im Sw Krw Im Krw Dr Krow Im Krow Dr Sw

11 Model performance how wrong can we be? Impact of heterogeneity Impact of viscosity Need for full physics UNDERSTANDING THE PRODUCING SYSTEM effective net wettability effects of imbibition locating remaining oil and what happens when we scale back up

12 Viewing the floodfront Sw Water saturation (Sw) T = 700 d Saturation Krw Water relperm (Krw) T = 700 d Relative permeabilty

13 Heterogeneity and sweep Truth Homogeneous bodies 4 layers 2 layers 1 layer

14 Heterogeneity and sweep Permeability Homogeneous model water breakthrough late by 40% Layer models out ±10% FWCT FWCT UT1B_WI FWCT UT3B_WI FWCT UT4B_WI FWCT UT5C_WI FWCT UT6C_WI 700d Truth Bodies het. bodies 4 lyrs 2 lyrs hom. 4 Layers Water cut Layers Homogeneous 1 Layer Krw

15 Impact of viscosity 1: gas PERM, md T = 10 yrs DEPLETED PRESSURE P/Pi ~ 0.5 PRESSURE, psia Gas doesn t see the heterogeneity

16 Impact of viscosity 2: viscous oil Viscous oil, 5 cp, unfavourable mobility ratio Contrasting fast vs. slow injection (15% vs. 1.5% PV/yr) Krw T = 700d FAST T = 7000 d SLOW Krw

17 Understanding 1 - the impact of capillary forces Increased Pc shows significant recovery from upper units Delays WBT by ~ 25%, increased RF ~ 5% 700d Reference case Htr ~ md Truth reference case Pc x 3 case, Htr ~ md Truth with stronger Pc Krw

18 Capillary pressure, psi Understanding 1 capillary imbibition dry towel effects A Potential gain ~ 3 psi from Sw exchange by ca. 0.1 from unit 2 to unit 3 10 Water / oil capillary pressure curves 1,2 3 4 Pc1/2 Dr Pc3 Dr Pc4 Dr Pc1/2 Im Pc3 Im Pc4 Im B Potential gain ~ 6 psi from Sw exchange by ca. 0.1 from unit 3 to unit 4 Water saturation (Sw) Water drawn up displaces oil down Additional recovery from nominally non-net material 5 B A 0 wetter drier Water saturation, frac B A Sw

19 Understanding 2 flow-based determination of net FWCT FWCT UT1c_wi FWCT UT1c_wi-cut1 FWCT UT1c_wi-cut Water breakthrough time in Units 1 & 2 is 10% early if poorer units cut-off Water cut ,2,3 1,2 No cut off Units 1,2,3 Units 1, LOWER 2 UNITS ONLY 44 md 167 md ALL UNITS IN MODEL 0.36 md 2.5 md 44 md 167 md Krw

20 Understanding 3: the impact of wettability WW: WBT later by ~ 10%, RF higher by ~ 3% Stronger spontaneous imbibition into upper units OW: WBT earlier by ~ 20%, RF lower by ~ 10% Bypass of lower perm material within lower unit 700d Water wetting (Iw 0.8) Water Wet Oil wetting (Iw 0.2) Oil Wet Krw

21 Understanding 4 - Locating Remaining Oil Oil saturation (So) T = 700 d Saturation behind the flood front Explore sim grid cell X=5 flowing ~90% water-cut Residual oil saturation Same result in fine and sim grid models but fine model illustrates where the higher Sor is located Model Swi W/cut Krw Kro Soil Ultra fine grid 30.7% 87% % Sim grid X=5 30.6% 93% % Sim grid X = 5 Oil saturation So

22 Upscaling to normal life Permeability Ultra Fine 0.25 x 0.02 m 320,000 cells b Fine 1 x 0.04 m 40,000 cells c Logging scale 5 x 0.1 m 3,200 cells d High res simulation 20 x 0.2 m 400 cells e Normal simulation 50 x 1 m 32 cells f Kh, md

23 Upscaling to normal life Ultimate recovery similar (within 3%) 700d b Krw But breakthrough times progressively error-prone (up to 20%) c d FWCT Water cut vs. time FWCT UT1b_wi FWCT UT1c_wi FWCT UT1d_wi FWCT UT1e_wi FWCT UT1f_wi b c d e f e 0.6 Water cut f b f

24 The role of Ultimate Truth models Truth models repeatedly show that removing detail is a systematic bias Big, complex full-field models tend to over-optimism We understand more through deconstruction and can then make appropriate workflow adjustments at larger scales Sometimes, small really is beautiful

25