The productivity of oil and gas wells plays a decisive role in the selection of reservoir completion methods and related opera-tions,which is one of the key indicators for reservoir development.At present,the prediction of oil and gas well productivity based on machine learning algorithm is obviously influenced by sample data.Aiming at the characteristics and advantages of the support vector machine method and the gray wolf algorithm in dealing with small data samples,combining the support vector machine method with the gray wolf algorithm to form the gray wolf algorithm-the support vector machine SVM algorithm(GWO-SVM algorithm).The algorithm before and after optimization and the commonly used machine learning algorithm were compared and tested by using the actual well data of an oilfield.The results show that the optimized GWO-SVM algorithm exhibits significant advantages in terms of computational speed and accuracy,and can more accurately determine the production capacity of oil and gas wells.The research results have some guiding significance for the productivity prediction of oil and gas wells.
关键词
油气井/产能预测/支持向量机算法/灰狼算法/GWO-SVM算法
Key words
oil and gas well/productivity prediction/support vector machine algorithm/gray wolf algorithm/GWO-SVM algorithm
维普
万方数据