Production Data Analysis and Pressure Prediction of Shale Gas Well in Fuling Jiaoshiba Area

摘要

The production pressure of shale gas well is a key parameter for guiding the yield allocation, thus the prediction of production pressure based on historical data can significantly impair the optimization of the subsequent production. However, the strong nonlinear correlation among the historical production data weakens the effect of traditional linear prediction method. Hence, a novel prediction and data analysis method for the shale gas well production pressure based on Elman neural network is proposed in this paper. First, the adaptive segmentation algorithm is used to piecewise eliminate the incomplete and abnormal data to guarantee the accuracy of the prediction. Second, the correlation analysis of production data is developed by Spearman method. Finally, Elman neural network is applied in predicting the production pressure of shale gas well. Compared with BP neural network and curve fitting prediction, the experimental results of the shale gas in Fuling Jiaoshiba area indicate that the proposed method can effectively improve the prediction accuracy of production press.