Usage of Artificial Intelligence to Reduce Operational Disruptions of ESPs by Implementing Predictive Maintenance

Abstract

Abstract The use of electrical submersible pumps (ESPs) is a highly effective artificial lift method for boosting oil production from wells operating in both onshore and offshore fields. When an ESP is deployed, its complete pump and electrical motor assembly is positioned below the surface within the oil reservoir that the well has tapped. Once deployed, the ESPs must be carefully maintained to be highly reliable and always available to prevent costly production disruptions due to unexpected pump failures. Typically, ESPs are connected to SCADA or other distributed control systems to provide supervisory and control functions for their effective operation as well as for operational visibility. Today, many diagnostic methods are available to determine the health and status of an ESP system by making use of that functionality in its automation system. However, while these methods can provide insightful analysis of problems, they usually only provide retrospective views after failure events have occurred. But this situation is changing. With recent advances in artificial intelligence (AI) combined with the new Internet of Things (IoT) technologies, it is possible to effectively use data-driven analytics fueled by large data sets. In particular, AI technology that involves deep learning and neural networks can be extremely effective in detecting abnormal behavior of complex physical systems such as ESPs, based on the data gathered from the system, providing decision support for remediating or managing the causative issues. This paper focuses on the results of implementing this AI technology combination to detect, flag, and remediate abnormal behavior for ESPs, which can increase their availability and prevent production disruptions. The subject use case involved 30 ESPs, with pumps ranging from 200–500 kW in power, installed in a medium-depth onshore oil field. The paper discusses the architecture of the solution that was deployed and explain how it supports a predictive maintenance model that is capable of accurately identifying abnormal ESP operating behaviors in advance before an ESP can fail and disrupt production.