Fault diagnosis study of pumping well schematic based on SCN-integrated learning

Abstract

With the bad oil exploitation environment, the safe operation of pumping wells is also affected to a certain extent. Once a fault occurs, it will bring great losses. Therefore, the realization of rapid and accurate diagnosis of pumping wells is of great significance for reducing losses, avoiding safety accidents and ensuring oil field production. In this paper, a new model of SCN- integrated learning is proposed to train and classify the data. The indicator diagram data is standardized by Z-score. The Gray Level Co-occurrence Matrix (GLCM) is used to extract the feature vector, and the feature vector is input into the SCN- integrated learning model for training. Through the comparison with SCN, the accuracy of this method is improved by 7.3%, and the final accuracy reaches 97.93%, which verifies the effectiveness and accuracy of this method.