Development of a Ball Seat Event Recognition Algorithm with Convolutional Neural Network for the Real-Time Hydraulic Fracturing Analytics System

This paper provides the technical details of our latest progress of the hydraulic fracturing ball seat event recognition model. In our previous work (Shen et al., 2020), the U-Net architecture was used to build models to recognize the ball seat event. In this paper, we train the Convolutional Neural Network (CNN) to solve the same problem. Comparing with U-Net approach, the CNN model requires less computation resource during training phase while it has slight better performance.