Estimation of Stress Profile with Depth from Analysis of Temperature and Fracture Orientation Logs in a 3.6 km Deep Well at Soultz, France

摘要

We presented a new method for stress measurements which isrnbased on the hypothesis that fractures which are criticallystressedrnfor frictional sliding in the current stress field tend tornpermeable and those which are not critically-stressed tend tornbe impermeable. Fracture depth, dip and strike can berndetected from borehole imaging logs such as BHTV/UBI orrnFMI/FMS. By comparing the depths of fractures and those ofrnthermal anomalies detected on temperature logs, we canrnidentify which of the fractures are permeable. Stresses onrnfracture planes can be estimated from the fracture orientationrnprovided that the state of in-situ stress is assumed. Therefore,rnknowledge of which fracture orientations tend to be permeablernand hence support critical stress states, and which fracturernorientations are not allows us to make a grid search overrnpossible states of stress to identify the stress state that bestrnsatisfies the observations and hypothesis. A simplerndescription of the stress states is necessary. Assuming thatrnone principal stress is vertical and equal to the overburden andrnpore pressure is known, the stress states are defined by threernunknowns given by the two depth gradients of the twornhorizontal stresses, and the orientation of the maximumrnhorizontal stress. The range of admissible stress states can bernfurther reduced by considering the strength of the bulk rock.rnWe applied the method to estimate stress state about a 3.6 kmrndeep well in granite at the Soultz Hot Dry Rock test site.rnFractures and thermal anomalies detected in the depth range ofrn2850~3500 m were used for the estimation. Results indicaternthat the stress regime is one of normal/strike-slip faulting,rnconsistent with the fault plane solutions of induced andrnregional earthquakes. The predicted profile of minimumrnprincipal stress magnitude agrees with that determined fromrnthe hydrofracture method. The inferred orientation of thernmaximum principal stress differs from the best availablernestimate (defined by thermally-induced tension fractures) byrn20～30°. The example demonstrates that the method can yieldrnuseful estimates of the state of stress at great depth fromrnrelatively common data of temperature, borehole imaging, andrndensity logs without requiring any additional input fromrnexpensive test. The method can also be applied in highrntemperature environments where conventional approaches tornstress estimation cannot be applied.