Automatic Well Test Validation Empowered By Machine Learning and Natural Language Processing

Abstract

Abstract The standard operating procedure (SOP) requires a well test monthly for every active producer to assess performance behavior. With over 100 new well tests daily, a busy operation schedule can lead to delayed validation, causing high accumulated amounts throughout the month and spill over to the next month. If the well-test quality does not meet the expectation, it should be rejected and required to retest immediately to comply with SOP. The significant effort and delayed well-test validation will cause inaccuracy in production performance analysis, diagnostics, and potential issue detection. This solution aims to significantly reduce processing time from gathering enough historical information to validating with engineering models and limits human error by checking all available well tests and preparing detailed analysis for engineers to make the final decision. By having more updated accepted well tests to update well engineering models, the solution helps to improve accuracy and more confident outputs in other engineering workflows like production back allocation, well rate estimation, well and network model calibrations, and production optimization. The proposed solution leverages artificial intelligence (AI) capability learns from historical well test data with accepted and rejected flag to build a rule-based deterministic machine learning (ML) model to automatically validate and detect the possible rejected or accepted well test. The solution also considers well test comments or remarks provided by well-testing engineers which are processed via Natural Language Processing (NLP) engine. ML model can propose to accept a well test with confidence score to automate the validation and support engineer's decision. On the other hand, if the model detects a possible rejected well test, it suggests engineer to review the new well test information versus historical performance and takes actions, where early rejection triggers retesting requirement to the offshore team to prioritize the well to the test plan. Periodically, the ML model may require updates based on the most recent well test data in order to maintain its accuracy. The solution significantly reduces well test validation time from weeks to hours, improving the accuracy of other production performance analysis and optimizations. The data-driven approach can easily be adapted to different fields’ needs, offering a more flexible and efficient alternative to hard-coded rule-based well test validation.