Research on productivity prediction and identification of main control factors of fractured vuggy reservoirs based on machine learning

Abstract

For SHB oil and gas reservoir, natural fracture development using support vector machine (SVM) and support vector regression method to SHB oil and gas field with the main fault zone of 18 flowing Wells for the single well production forecast, drilling and well completion by input data, production data and dynamic data, bottom hole flowing pressure, adjoining well production data as the input variables such as time, The predicted output value is used as the output variable for yield prediction. The results show that SVM is not only more efficient than the traditional DCA method, but also avoids geological modeling and a large amount of historical fitting work. It has a certain reference value for the formulation of rational production system in SHB block.