Real-time Performance Monitoring of Electrical Submersible Pumps in SAGD Process

Electrical submersible pump (ESP) is one of the preferred artificial lift systems in upstream oil production because of its wide operating range and endurance to harsh operating environments. In Alberta Canada, about two-thirds of steam assisted gravity drainage (SAGD) wells are equipped with ESPs. Keeping the ESPs long-term operational is one of the primary challenges faced by the operators. ESP failures are a common problem due to various reasons, such as, harsh operating conditions, improper installations, etc. Therefore, the real-time monitoring of ESP performance is indispensable as it can prevent unscheduled well shutdowns and enable preventative maintenance. In this work, aiming at the development of reliable ESP performance monitoring in SAGD process, a monitoring strategy based on a multivariate statistical technique is proposed. The strategy includes a novel first-principles based feature extraction approach, an online model update strategy and a new health index to evaluate ESP health conditions in real time. Test results on real ESP datasets are presented to highlight the efficacy of the proposed algorithm.