This paper presents the analysis of in-situ stress magnitudes from observed borehole failures. Knowledge about the state of in-situ stress is important for many applications in the petroleum industry such as predicting sanding potential and reservoir behavior during depletion, and defining the range of safe mud weights to prevent lost circulation or wellbore collapse while drilling.
Borehole breakouts and drilling-induced tensile fractures revealed by wellbore image logs are often used to determine the orientation of principal stresses and to constrain stress magnitudes. One difficulty encountered in doing this is a lack of knowledge of the specific wellbore conditions, including mud pressure and temperature, at which each feature occurred. While-drilling images are acquired a very short time after the drill bit has passed the formation. It is however possible to use while-drilling images to pinpoint the time interval during which failures initiate and to monitor failure growth overtime, since continuous downhole annular pressure measurements provide constraints on conditions under which the failure has occurred.
Using wireline acoustic image data from one well, and LWD high-resolution resistivity images in a second well, we demonstrate the approach and evaluate the resulting uncertainties. Although use of real-time data does reduce uncertainties, the uncertainty in the magnitude of maximum horizontal stress remains large due to uncertainties in the formation strength. Therefore, the results presented are preliminary, and will be improved in later research when better data are available.
BETA (the Baker Hughes Experimental Test Area) is located south of the city of Tulsa, in northeastern Oklahoma. It is a unique field research facility where downhole tools can be tested and various researches can be performed. Over time, more than one hundred wells were drilled at BETA and therefore an abundance of well logging data is available. The geology at BETA has been carefully studied based on core analysis and the lithology is well known . Bedding is flat-lying and laterally continuous, and therefore measurements made in one well can be applied with little depth correction to other wells.