Intelligent Prediction of Stuck Pipe Remediation Using Machine Learning Algorithms

Abstract

Abstract Stuck pipe is still a major operational challenge that imposes a significant amount of downtime and associated costs to petroleum and gas exploration operations. The possibility of freeing stuck pipe depends on response time and subsequent surface action taken by the driller during and after the sticking is experienced. A late and improper reaction not only causes a loss of time in trying to release stuck pipe but also results in the loss of an important portion of expensive tubular, downhole equipment and tools. Therefore, a fast and effective response should be made to release the stuck pipe. Investigating previous successful responses that have solved stuck pipe issues makes it possible to predict and adopt the proper treatments. This paper presents a study on the application of machine learning methodologies to develop an expert system that can be used as a reference guide for the drilling engineer to make intelligent decisions and reduce the lost time for each stuck pipe event. Field datasets, including the drilling operation parameters, formation type, and fluid mud characteristics, were collected from 385 wells drilled in Southern Iraq from different fields. The new models were developed to predict the stuck pipe solution for vertical and deviated wells using artificial neural networks (ANNs) and a support vector machine (SVM). The results of the analysis have revealed that both ANNs and SVM approaches can be of great use, with the SVM results being more promising. These machine learning methods offer insights that could improve response time and strategies for treating stuck pipe.