_ This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 215244, “Automated Production-Enhancement-Candidates Screening Powered by Machine Learning Unlocks Untapped Potential in Matured Oil Fields: A Case Study,” by Nusheena M. Khair, SPE, M. Farid Zaizakrani, SPE, and Nur A.I.Z. Azhar, Petronas, et al. The paper has not been peer reviewed. _ Field A offshore East Malaysia has been a productive conventional oil field for more than 3 decades. It has encountered several surface and subsurface issues, including sand production, increasing water cut, and reservoir-pressure depletion. The complete paper presents the process of identifying production-enhancement opportunities, develops a methodology to identify underperforming candidates and analyze well-integrity issues, and describes how data science and engineering analyses are integrated. Introduction Field A is undergoing redevelopment, including activities such as enhanced oil recovery, production-enhancement (PE) initiatives, idle-well reactivation (IWR), and infill drilling. However, Field A’s operations and PE initiatives face multiple challenges. These include a shortage of comprehensive data affecting reservoir management and forecasting, difficulties in managing wells on unmanned platforms during adverse weather, and high logistical costs in offshore interventions that affect the economic feasibility of certain PE and IWR jobs. Recognizing these issues, the operator team is taking a proactive approach by embarking on enhancing PE and IWR candidate generation by machine learning (ML), a strategy that aims to expedite well-by-well review (WBWR) to streamline candidate-generation processes and enhance decision-making. Production Enhancement Candidate Generation and Screening (PECGS) Methodology PECGS is a petroleum engineering advisory system that helps operators improve the efficiency of PE and IWR candidate screening and intervention planning. PECGS automatically generates a list of underperforming wells along with recommendations for remedial actions. PECGS works by first evaluating each well’s production potential and risk assessment through use of a data-driven advisory system that combines analytical and ML models with industry-standard petroleum and reservoir management logic. Once the wells have been evaluated, PECGS generates a live automated opportunity register with a ranking of candidate wells. Technical Solution The technical solution in PECGS includes four main stages: knowledge base, technical analysis, chance of success, and economic analysis. The knowledge base of PECGS is formed through data integration from various sources. Static data have been standardized and validated by subject-matter experts before being uploaded into the advisory-system database. Dynamic data, such as production and well test, are integrated daily into the advisory system database according to a predefined schedule. The technical analysis is leveraged with proven analysis methods to identify well constraints and recommend PE opportunities and present estimated production gain, success probability, and profitability. The workflow uses multicriteria ranking, or the analytical hierarchy process (AHP), to rank screened wells based on production and petrophysical key performance indicators. Once a potential PE candidate has been identified through an automated screening, the next step is to review the chance of success. Next, the engineers perform an economic analysis to evaluate the potential profitability of the PE job. The economic analysis uses a variety of factors, including the cost of the intervention, the expected increase in production, and the value of the produced oil or gas in order to calculate the net present value (NPV) of the job. If the NPV is positive, then the intervention is profitable.
Read full abstract