Electrical Submersible Pump (ESP) lift is extensively used in offshore production systems. Although real-time pump operating parameters and the production data are commonly available according to the recently developed digital energy technologies, the analysis of the acquired data is still insufficient to monitor, diagnose, interpret, and analyze reservoir performance, wellbore integrity, and ESP operating status and efficiency in a real-time manner. Further, the traditional ammeter card diagnosis cannot identify leakage of tubing, underperforming pump lifting, and low pump working efficiency.
This paper summarizes typical characteristics of different downhole malfunctions in ESP wells, provides a novel approach to detect those problems systematically and automatically. The developed management system incorporates real-time ESP current data and interpret ammeter card by neural network analysis, which has been trained through over 900 wells. We also derived the analytical solutions for wellhead pressure buildup analysis as a supplement to neural network analysis. Consequently, this model is able to detect downhole problems that cannot be identified by ammeter card. This diagnosis model also sets thresholds for key parameters, and alarm operation team through satellite in a timely manner. This model is a reliable supplementary to the traditional ammeter card diagnosis method, and the online utility provides complement flexibility to cooperate with field operations in real-time. The early-stage identification and resolution of ESP problems can lead to a great cost-saving and less maintenance requirements owing to this intelligent system. This workflow has been successful in field trials.