Integration of Neural Networks and Wellbore Stability, a Modern Approach to Recognize Drilling Problems Through Computer Vision

Cavings are a valuable source of information when drilling operations are being performed, and multiple parameters can contribute to producing cavings which indicate that failure has occurred or is about to occur downhole. This study will describe a project which is an integrated study of Machine Learning, Computer Vision, Geology, and Photography so that the recognition of cavings in the shaker is possible and how to link the cavings morphology with causal mechanisms related to wellbore instability problems. This study aims to develop a model which can extract caving features such as Shape, Edge Definition, Color, and Size. One of the core aspects of this study was to develop a structured image database of cavings from the Norwegian Continental Shelf which contains important feature information and the application of different algorithms used for automation enabled several opportunities to analyze and identify causal mechanism related to wellbore instability problems in real-time. As a result of that, the drilling operations would experience an improvement in terms of a faster decision-making process to solve operative problems related to wellbore stability which will lead to optimization not only in time and resources but also in safer drilling operations. Different algorithms and artificial intelligence tools were used to investigate the best approach to correctly detect and derive meaningful information about the shape, color size and edge from cavings like supervised learning, unsupervised learning, neural networks and computer vision. A key part of this study was image augmentation which plays a significant role for the detection of the cavings and their features. Multiple data sets can be created, and by using data augmentation, this will enable recognition of more complex patterns that will have on-rig applicability. Also, this new approach can deliver multiple outcomes besides failure mechanism identification such as volume of rocks being drilled, transport of cutting, type of formation being drilled.