1. Netherlands Institute of Applied Geoscience TNO - National Geological Survey Using Decision support models for deep geothermal projects Jan-Diederik van Wees, Damien Bonte, Jasper Zoethout (TNO/VUA) Albert Genter (GEIE)

2. TNO ENGINE Workshop 7, Leiden 2007, november 82 Contents (1) Geothermal background – DBHE perspective DBHE: Deep borehole heat exchanger Opportunities in the Netherlands re-using E&P wells Performance calculation Taking into account uncertainties and engineering options Best practices Asset development decisions E&P Decision support system Re-using producing E&P wells:enhanced performance

3. TNO ENGINE Workshop 7, Leiden 2007, november 83 Contents (2) How about using Decision support system for other forms of deep geothermal projects Enhanced Geothermal Systems Doublet Systems

4. TNO ENGINE Workshop 7, Leiden 2007, november 84 CO- CashFlow (1) Opex + Capex Time (years) Cash flow (M $) 300 200 100 0 -100 -200 -300 12345 Production t first-oil Revenues UCF = Cash in - Cash out Pay out time  DCF DCF (DISCOUNT_RATE) NPV

5. TNO ENGINE Workshop 7, Leiden 2007, november 85 Background (1) DBHE: Deep Borehole Heat Exchanger (example 3000 m deep Prentzlau)

6. TNO ENGINE Workshop 7, Leiden 2007, november 86 Background (2) Why considering a DBHE? New interest in geothermal energy The DBHE is experimental, but there are several setups (Prentzlau, Weggis) Interesting option for re-using E&P wells Not necessary to drill new wells, use existing E&P wells Possible positive effect of the heating of the surrounding rocks during production of oil or gas Restriction LOW Enthalpy System: Short distance from the source to the consumer

7. TNO ENGINE Workshop 7, Leiden 2007, november 87 Performance Calculation (1) 1 st order temperature calculation I Fourier’s conduction law Heat flux J J=k  T/  r 2π r  Z k = conductivity  T= temp difference  r = radius difference r = radius  Z = depth difference Rock Conductivity k is very important! J Tube T0 =Temperature at outer wall of tube Difussivity, dependent on conductivity T geotherm

8. TNO ENGINE Workshop 7, Leiden 2007, november 88 Performance Calculation (2) 1 st order temperature calculation II From heat flux to temperature T idealoutlet = SUM (J)/ Q Cp + Tinlet T = Temperature J = Heat flux Q Cp = thermal capacity of the heat flow [W/K] Q = mass flow rate (kg/s) ~ 15 m3 / h Cp = specific heat at constant pressure (J/kgK)~4180 POWER = (T outlet -T inlett ) Q Cp [W or J/s] J at each depth should be such that it is consistent with temperarure of water in the tubing (heated downward/cooled upward)

9. TNO ENGINE Workshop 7, Leiden 2007, november 89 Performance Calculation (3) J consistent: Analytical solution of Kujawa and Nowak, 2000 An Isolated upper part Steel lower part Isolated upper tube Non- isolated lower tube Cross sections rw rwi rwa rwb rzo rz rw rwi rwa rwb rzo rz T1 T2 T0 T1 T2 J

10. TNO ENGINE Workshop 7, Leiden 2007, november 810 Performance Calculation (4) Heat transfer in DBHE The heat flux in the analytical model is calculated as follows: (cf. Kohl et al., 2002) outer tube: Isolated inner tube: T2 T1 T0 Thermal boundary layer

11. TNO ENGINE Workshop 7, Leiden 2007, november 811 Performance Calculation (5b: analytical) Rock conductivity = 2, Temperature =160

12. TNO ENGINE Workshop 7, Leiden 2007, november 812 Performance Calculation (5c: analytical) Rock conductivity = 1.5, Temperature =160

13. TNO ENGINE Workshop 7, Leiden 2007, november 813 Performance Calculation (6) axi-symmetric finite difference solution 1b 1a

14. TNO ENGINE Workshop 7, Leiden 2007, november 814 Performance Calculation (7) finite difference Cross-section for a specific depth Prod Water Inj Water

15. TNO ENGINE Workshop 7, Leiden 2007, november 815 Performance Calculation (8) axi-symmetric finite difference solution 1a

16. TNO ENGINE Workshop 7, Leiden 2007, november 816 Taking into account uncertainties and engineering options TNOs experience from Oil and Gas E&P “Decision and risk management” Research consortia (1997-2003) BHPBilliton Scenarios vs. Continuous probability functions (MC) AND

17. TNO ENGINE Workshop 7, Leiden 2007, november 817 Best practice: integrate techno-economics of workflow, linking technical uncertainty to economic performance 1 pdf for each KPI

18. TNO ENGINE Workshop 7, Leiden 2007, november 818 Key performance input / output (statistical distributions) Input Basin parameters Geotherm, conductivity Underground development Workover oil& gas wells Design parameters Surface development Heat exchanger transport Production parameters Volumetric rate of water Economic parameters discount rate, energy price opex, capex, tax, royalty Output (Key Performance Indicators) Technical Toutlet Economic NPV, IRR, P/I ratio, Max exposure, Pay-out time, economic life Other (not as statistical distr.) HSE Political Public opinion

19. TNO ENGINE Workshop 7, Leiden 2007, november 819 Techno-ecomonic calculation  Fast computational models, split up in modules driving philosophy is to trade-off accuracy for completeness basin properties Underground development Surface development Economics Indicators TechnicalEconomic NPV DPR, IRR Max. Exposure Payout Time Econ. Lifetime Unit Technical Cost Toutlet Well design UDSDCFBAS

20. TNO ENGINE Workshop 7, Leiden 2007, november 820 Analytical Model Decision Support System (DSS) The analytical model is written in JAVA and implemented as a module in DSS DSS is used for optimizing decisions using Monte Carlo statistics The strength of DSS Easily parameterization Fast Long term predictions

21. TNO ENGINE Workshop 7, Leiden 2007, november 821 Analytical model DSS results The results can be plotted in a graph or be stored in a table OR Easily copied to other software

22. TNO ENGINE Workshop 7, Leiden 2007, november 822 Output Value, scalar: power produced Year 10

23. TNO ENGINE Workshop 7, Leiden 2007, november 823 Output Value, timeseries :Toutlet --> power Uncertainty because of: K = 1.5..2 Geotherm = 40..50 C/km

24. TNO ENGINE Workshop 7, Leiden 2007, november 824 Sensitivity of Power to.. conductivity Geotherm [C/m] 0.75 0.59 Tornado- plot

25. TNO ENGINE Workshop 7, Leiden 2007, november 825 SCENARIO ANALYSIS: Description of DBHE case UNCERTAINTY BAS Model K=4K=1.5-2 DECISION SD model 100 kEUR eff= 0.950 kEUR Eff= 0.8 Heat exchanger DECISION UD model vacuum=0.5 cm, 50 kEUR vacuum=1 cm 100 kEUR isolation

26. TNO ENGINE Workshop 7, Leiden 2007, november 826 Decision Tree K=1.5..2 K=4 DOWNSIDE BECAUSE OF k=1.5..2 scenarios NPV [kEUR]

27. TNO ENGINE Workshop 7, Leiden 2007, november 827 Decision Tree – K=1.5..2 scenarios excluded (assesment with costs 50 kEUR prior to use) NPV [kEUR]

28. TNO ENGINE Workshop 7, Leiden 2007, november 828 Re-using a production well (1) Temperature after 10 years of operation 10 years production ~ 2000 boe/d

29. TNO ENGINE Workshop 7, Leiden 2007, november 829 Re-using a production well (2) ca 20% increase in performance

30. TNO ENGINE Workshop 7, Leiden 2007, november 830 Decision support system for other forms of deep geothermal projects  EGS EGS: necessity of fast models for technical calculation In these models flow rate and sustainable temperature are most critical Recently publication of: presentation of fast models Brief explanation of rationale behind these model Implementation of model

31. TNO ENGINE Workshop 7, Leiden 2007, november 831 In a geothermal doublet : temperature from the Temperature in rock in reservoir fractures Area of fractures (A) and flow rate (Q) and volume of fractures primarily relate to the sustainebility in time of the high temperatures. Brief explanation of rationale behind these models

32. TNO ENGINE Workshop 7, Leiden 2007, november 832 Brief explanation of rationale behind these models The flow rate can be known using the  Pres : Flow rate (Q) is assumed to increase linearly with increased pressure difference (  Pres) between injection and production well at reservoir level During a production test: a reference pressure difference at reservoir level (  Prefpres), corrected for pressure effects in the wells, results a reference flow rate (Qref) So for a different  Pres:

33. TNO ENGINE Workshop 7, Leiden 2007, november 833 Brief explanation of rationale behind these models The  Pres is calculate using the  P at the surface and have to be corrected by the pressures évolution in the boreholes.

34. TNO ENGINE Workshop 7, Leiden 2007, november 834 Brief explanation of rationale behind these models Temperatures exchanges between the fluid and the surounding rocks are double Along each borehole the fluid temperature change following In the reservoir, along fracture the fluid temperature is increasing

35. TNO ENGINE Workshop 7, Leiden 2007, november 835 Brief explanation of rationale behind these models The out temperature at surface can be calculate with : Most of the temperature is gained in the reservoir

36. TNO ENGINE Workshop 7, Leiden 2007, november 836 DSS (fast model parameters –example -1)

37. TNO ENGINE Workshop 7, Leiden 2007, november 837 DSS (fast model parameters (example)-2)

38. TNO ENGINE Workshop 7, Leiden 2007, november 838

39. TNO ENGINE Workshop 7, Leiden 2007, november 839 Conclusions (1) Usefull to adopt decision and risk management best practices from Oil and Gas industry Understanding of sensitivity of performance to uncertainties beyond control Selection of optimum design scenarios DBHE specific Geotherm, and conductivity of prime importance Insulation inner-outer tube important Preceding oil and gas production leads to pre-heating of well which can increase power by 20-30%

40. TNO ENGINE Workshop 7, Leiden 2007, november 840 Conclusions EGS (2) Fast models are available Initial results look promising Needs to be developed further in the coming months  more at final meeting