A novel prediction method of gas and water production in tight gas wells on the Eastern margin of Ordos Basin

Abstract

In order to address the limitations of traditional methods for predicting gas production of tight gas wells in the Linxing area, such as a single data type, simplified mathematical models, and insufficient accuracy. A novel approach is proposed that relies on comprehensive geological and engineering data from gas wells in the field. Data from 48 gas wells in the field were collected and pre-processed to obtain 129 geological and engineering influencing factors, and the correlation coefficients of each factor were calculated, ranked and chipped, and finally 39 main controlling factors for gas well production were screened out. Three models of multiple regression, support vector machine and Bayesian ridge regression were developed and evaluated, and the highest R 2 of the multiple regression model was 0.97. Relatively regression equations for daily gas and water production distribution were fitted and evaluated for their adequacy. The results demonstrated that the types of major control factors were consistent, and gas production was more sensitive than water production. Therefore, optimizing water control in gas wells with sufficient water production and adjusting the intensity of water control based on gas production adequacy is a feasible approach.