| Abstract |
We performed a simple optimization procedure and sensitivity analysis to examine designs for Enhanced Geothermal Systems (EGS) that involve horizontal wells and multiple fracturing stages. The sensitivity analysis included calculations of thermal breakthrough and the maximum flow rate that could be achieved through the system (considering pressure drop in the well and in the reservoir). Conventionally, EGS wells have been nearly vertical and stimulated in openhole sections in a single stage. We investigated a design involving two parallel horizontal wells. The first well would be drilled and completed with casing, and then stimulated sequentially in stages (using cased hole packers rated to high temperature). The second well would be drilled through the stimulated region created by the first and completed openhole. For different combinations of well spacing, fracture transmissivity, and number of stages, the optimal flow rate was determined to maximize present value (PV) of revenue (cost was not considered). The calculations showed that stimulating with multiple stages would radically improve economic performance, delaying thermal breakthrough and allowing a higher overall flow rate to be circulated through the system. At low well spacing and low number of stages, it is optimal to circulate fluid more slowly than the maximum possible rate, in order to delay thermal breakthrough. With greater well spacing and with more stages, thermal breakthrough will be relatively delayed, and it is optimal to circulate at the highest possible flow rate. It is optimal to use the lowest well spacing where NPV is maximized by circulating at the maximum possible rate. When it is optimal to circulate at the maximum possible rate, NPV is sensitive to reservoir transmissivity. When it is optimal to circulate at less than the maximum possible rate, NPV is unaffected by reservoir transmissivity. Optimizing to maximize aggregate heat extraction instead of NPV has only a limited effect on NPV. Most thermal drawdown calculations were performed assuming infinite stage spacing. Calculations assuming finite fracture spacing and a 1000 m lateral had only modestly lower NPV relative to the infinite fracture casing calculations. |