In situ stress estimation using borehole failures — Even for inclined stress tensor

This paper derives a method to estimate the in situ stress using pressure and/or image data from borehole failures. It is based on linear elasticity and common failure criteria for brittle rock. Current methods for estimating in situ stress are only applicable to one type of failures and they either put restrictions on the in situ stress tensor, or need many data sets, or are numerically unstable. Even though they may yield satisfactory results in some situations, more accurate and less restrictive methods are needed to yield reliable in situ stress estimates in general cases. The proposed method overcomes these challenges. It can combine fractures and breakouts data from several locations, and it does not restrict the in situ stress to have vertical as a principal direction. If, on the other hand, vertical is a principal direction, the method only needs failure data from one location to uniquely estimate the in situ stress tensor. The method gives accurate results when applied to synthetic data, and is shown to be fairly stable against noise in the failure input data. The method is demonstrated on field data from the Matsukawa geothermal field in Japan, and from the Gjøa field in the North Sea. In both cases the estimated in situ stress is highly inclined and predicts observed data with good accuracy. And in the latter case, where structural data are available, the estimated in situ stress tensor follows the formation layers.

► A new method to estimate the in situ stress using borehole failures is developed.
► It can estimate in situ stress with arbitrarily oriented principal directions.
► It is demonstrated on two field cases where the in situ stress tensor is found to be highly inclined.