Application of Time Series Analytics to Assess Performance of Artificial Lift Systems Deployed in Coal Seam Gas (CSG) Wells

Abstract

Artificial Lift Systems (ALS) play a crucial role in producing natural gas from Coal Seam Gas (CSG) wells in Australia. These systems are employed in over five thousand wells located in the Bowen and Surat Basins of Queensland. Operators face significant challenges in managing and maintaining ALS-supported production due to regular failures caused by factors like coal fines. Failure of ALS can have a detrimental impact on meeting both local and international gas export commitments; hence, effective management and maintenance of ALS-supported production are paramount. The thesis highlights the importance of utilizing real-time data and time series analytics to evaluate ALS performance. Real-time data can help manage CSG wells with artificial lift proactively and with greater insight. Petroleum and well surveillance engineers' expertise is combined to enhance the analysis of time series data. The research presents an innovative approach that involves transforming time series data into images through Symbolic Aggregate Approximation (SAX). SAX serves as a feature extraction technique that converts time series data into a symbolic representation, which is then translated into performance heatmap images. Petroleum and well surveillance engineers label these SAX-generated performance heatmap images with expert precision. By incorporating domain-specific insights and utilizing novel time series analytics techniques, operators can detect abnormal ALS behavior, proactively address performance issues, and improve overall production efficiency. This research enabled the creation of a tailor-made ALS analytics application that helps monitor an extensive network of CSG wells, detect abnormal ALS behavior early, and provide insights for proactively managing performance issues, thereby imparting a significant economic impact on CSG operations in Australia.