Abnormal Detection of Pumping Unit Bearing Based on Extension Theory

Abstract

In order to solve the problem that is difficult to extract the fault data of bearing of pumping unit in operation, a negative selection algorithm based on extension theory is proposed to detect the abnormal of bearing. First, a negative selection algorithm detector set model is constructed by using matter element. Synthetic correlation degree function is used as the matching rule of detector generation stage and bearing fault anomaly detection stage. Genetic Particle Swarm Optimization (GPSO) is used to generate detector set. Aiming at the problem of large redundancy of the detector set, the correlation function is used to formulate merging rules to merge the detector set interval. Finally, bearing anomaly detection is carried out by using the bearing fault data of Case Western Reserve University. The results show that the activation rate of the inner ring fault detector is 98.89%, the activation rate of the outer ring fault detector is 98.61%, and the activation rate of the ball fault detector is 99.17%. © 2021 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC.