Fracture Parameters Inversion of Hydraulic Fracturing Based on Deep Learning Neural Network

In the design of backflow scheme after hydraulic fracturing, the fracture size of length, width and height are needed. However, these parameters are difficult to obtain. Therefore, a fracture parameter inversion method is proposed. For this method, the deep learning neural network is used. Firstly, the basic principle of fracture parameter inversion is introduced. Then the structure of deep learning neural network is discussed. It is based on the Alexnet network, which has 5 convolutional layers, 3 pooling layers, and 3 fully connected layers. In this paper, a method is proposed to convert one-dimensional data into two-dimensional image for neural network training. And the testing precision of neural network is defined. Finally, the simulation results show that the training loss is 0.001, and the testing precision is 0.95. The research of this paper is the application of the intelligent technology to the oil and gas industry.