Refracturing Candidate Selection in Tight Oil Reservoirs Using Hybrid Analysis of Data and Physics Based Models
Die Hu,Zhengdong Lei,Stephen Cartwright,Steven Samoil,Siqi Xie,Zhangxin Chen
DOI: https://doi.org/10.2118/208883-ms
2022-01-01
Abstract:Refracturing candidate selection problems can be solved via production statistics, virtual intelligence and type-curve matching, and these methods are mostly developed using data-based models. They unleash great power of data but have not considered the influence of geological distributions in physics-based models. This paper combines the strengths of data and physics based models and proposes a hybrid analysis method to improve and strengthen the current methods. Three criteria, production performance, a completion index and a geological distribution around an offset well, and their sub-criteria are selected to build an evaluation system for refracturing candidate wells. Field data is collected and processed to calculate a completion index and production performance. To quantify a geological distribution around a well, a history-matched reservoir simulation model is required. Besides, a graph theory algorithm, Dijkstra’s shortest path, is used to quantify the influence of geological distributions in 3D reservoir models on wells. An analytic hierarchy process and grey correlation analysis are then used to establish a multi-level evaluation system and determine and rank each individual strategic factor. Finally, datapoints are shown in a 3D coordinate system, and custom defined weights are used to calculate the final ranking of potential refracturing wells. In addition, the hybrid analysis is presented on our self-developed visualization platform. A history-matched reservoir simulation model from the Y284 tight oil reservoir is used as a study case. Eight refractured wells’ data is collected and analyzed. As a grey correlation analysis result, a sub-criteron of productivity performance, relative productivity, ranks the first, followed by cumulative liquid production. Completion and resistance rank third and fourth with a small gap. Based on the analysis results, an evaluation system is built up. 14 refracturing candidate wells are analyzed and ranked using the evaluation system. These wells are displayed in a 3D coordinate system, where x, y and z directions represent three criteria separately. Wells distributed in the first quadrant are regarded as optimum candidates to apply refracturing treatments. Correlations of evaluation factors and increased oil production after refracturing treatment are plotted to validate the method. This study explores how to conduct hybrid analysis in a selection workflow of refracturing candidate wells. Combing visualization, interpretability, robust foundation and understanding of reservoir models with accuracy and efficiency, data-driven artificial intelligence algorithms, the experiences distilled, and insights gained from this project show great potential to apply hybrid analysis as well as modelling in oil and gas industry.