The Application of Machine Learning (Dt-Chan-Performance) in Determining Idle Well Reactivation Candidates at Pt. Pertamina Ep Regional 4 Zone 11 Cepu Field

Indonesia faces a significant challenge in achieving its goal of oil production 1 million barrels of oil per day by 2030, particularly as it relies on old fields or mature fields (brownfields) to extract remaining hydrocarbons. One of the strategies involves reactivating of idle wells in Cepu field, managed by PT. Pertamina EP Regional 4 zone 11. This study focuses on identifying suitable candidates for reactivation through combination of research, innovation and production-focus analysis. The process begins with problem definition, aiming to understand the factors influencing idle wells and review recent advancements in reactivation prediction. Data were collected from both primary and secondary sources covering period 2018-2023. The next stage is implementing Machine Learning (ML), specifically Decision Tree (DT) model, to overcome problems related to data accuracy and complexity. A web application was developed to support decision-makers in selecting wells with high reactivation potential which can provide the best solution of increasing oil recovery. The research results show a high success rate on Accuracy Under Curve and Receiver Operating Curve score of 0.99, indication strong predictive capability. Using entropy-based analysis, two potential wells were identified for reactivation for improvement. These wells were further evaluated using Chan Diagnostic and Production Performance analysis.