Optimization Method of Hydraulic Fracturing Design for Horizontal Shale Gas Well Based on Artificial Neural Network and Genetic Algorithm

Multistage hydraulic fracturing in horizontal well is the key means to realize efficient development of shale reservoir. Traditional fracturing optimization mainly applies single factor analysis or orthogonal test based on numerical simulation, which is time consuming and is difficult to obtain the global optimal solution. In this paper, an optimization method of hydraulic fracturing design for horizontal shale gas well based on artificial neural network (ANN) and genetic algorithm (GA) is proposed. On the basis of collecting geology, engineering, and production data of fractured horizontal wells, the main production influencing factors are obtained with recursive feature elimination and cross validation method, then a data-driven production prediction model is constructed based on ANN, and the GA is used to search the global optimal combination of fracturing parameters. With the data set of 162 fractured horizontal wells from a shale gas field in Sichuan Basin, an ANN production prediction model is trained and tested. The calculation efficiency is high, and the determination coefficient exceeds 0.8. Furthermore, an optimal fracturing parameters combination strategy is proposed for a new completed well in this area. This method can provide some guidance for production evaluation and hydraulic fracturing design of horizontal shale gas wells.