Research on Fault Diagnosis Method of Rod Pumping Wells Based on CNN_APRCSO_SVM

Abstract

The traditional work status recognition methods based on indicator diagrams require manual selection of indicator diagram features, and the recognition accuracy is low. In response to this problem, this paper proposes an intelligent fault diagnosis method combined convolutional neural network (CNN) with support vector machine (SVM). The CNN is used to automatically extract the features of the indicator diagrams, SVM is used to make diagnosis, and the improved chicken swarm optimization is used to solve the problem of difficult determination of the SVM parameters. The improved chicken swarm optimization avoids the problem that chicken swarm optimization (CSO) is easy to fall into local optimum, and it is better than particle swarm optimization (PSO) and the traditional CSO in accuracy. Compared with the traditional CNN model fault diagnosis method, the fault diagnosis method proposed in this paper has better recognition performance.