Geomechanical Modeling and Artificial Neural Network Technique for Predicting Breakout Failure in Nasiriyah Oilfield

Wellbore instability is one of the major issues observed throughout the drilling operation. Various wellbore instability issues may occur during drilling operations, including tight holes, borehole collapse, stuck pipe, and shale caving. Rock failure criteria are important in geomechanical analysis since they predict shear and tensile failures. A suitable failure criterion must match the rock failure, which a caliper log can detect to estimate the optimal mud weight. Lack of data makes certain wells' caliper logs unavailable. This makes it difficult to validate the performance of each failure criterion. This paper proposes an approach for predicting the breakout zones in the Nasiriyah oil field using an artificial neural network. It also presents the optimal mud weight window for this field, which can be used to optimise the mud weights to minimise the wellbore instability issues. The results showed that an artificial neural network is a powerful tool for determining the breakout zones using the input data. The obtaining root mean square error and the determination coefficient were respectively 0.0082 and 0.959, by which the 1D MEM gave a high match between the predicted wellbore instabilities using the Mogi-failure criterion and the predicted breakout using the ANN model. Most borehole enlargements occur due to formation shear failures because of using low mud weights during drilling. The conclusion clarify the1.35 g/cc is the optimal mud weights for drilling new wells in this field of interest with fewer drilling issues.