A Hybrid Machine Learning Method for Pore Pressure and Wellbore Stability Prediction in a Shale Gas Field at Sichuan Basin

ABSTRACT: High pore-pressure and wellbore instability are the main challenges during drilling. The predictions of pore pressure and wellbore stability are usually executed manually by geomechanics engineers. The procedures are usually complicated and highly depend on the analyst's expertise. In this study, machine learning methods are used to perform prediction. The digital models were trained with existing well data, geology data and drilling data. The formation materials properties are trained and learnt with patterns recognition; the pore pressure, earth stresses and wellbore stability are trained with physics-based hybrid algorithm. The trained models are then used to predict pore pressure and mud weight window for planned well. This approach was applied and validated in a deep shale gas field in Sichuan basin, China. With the developed new machine learning method, the digital models were trained with eleven geology surfaces and well data from eight existing wells. The trained model was used to predict pore pressure and mud weight window for two planned horizontal wells, including formation collapse pressure, mud loss pressure and breakdown pressure. The machine learning predictions were compared against both manual results calculated by geomechanics analysts and actual drilling results, The digital results matched well with manual results and the actual drilling results confirmed its accuracy and effectiveness. 1. INTRODUCTION Wellbore instability is one of challenges for drilling. It can lead to many drilling complexities such as stuck pipe, over-pull, tight hole and loss circulation, et al. The evaluation of drilling risks with geomechanical wellbore stability analysis (WBS) is critical in well planning and drilling. High pore pressure may lead to kicks and blowouts, so it is the most crucial parameter to evaluate formation collapse pressure, mud loss pressure and formation breakdown pressure, to define a safe mud weight window. Some pore pressure prediction methods have been developed using geological data, including seismic, well logs, and drilling data (Eaton, 1975; Bowers, 1995; Dutta, 2002). In well planning, different planned well trajectories will be defined as considering reservoir target, drilling risks and operation costs. It is time consuming to evaluate and predict pore pressure and drilling risks along different planned well trajectories. Also, the prediction highly depends on the analyst's experience with local knowledge.