Vibration Fundamental Frequency Amplitude Prediction of Electric Transformer by Proposed Hybrid Artificial Bee Colony Method

Abstract

Existing research express by processing about the vibration signal of electric transformer cover, the corresponding fundamental frequency amplitude can be obtained, which can be used to dig the specific operation state of the electric transformer. For accurately predicting the amplitude at the fundamental frequency about cover vibration signal of electric transformer, a proposed method based on artificial bee colony is designed. But the basic artificial colony method has some defects, such as slow tracking speed and poor feature description ability. Therefore, in view of the shortcomings of the existing ABC based method, the proposed artificial bee colony method is designed, where an adversarial learning mechanism and optimizing each specific search formula is employed. The proposed ABC method improves the prediction accuracy of electric transformer cover vibration signal’s fundamental frequency amplitude. The sepcific results of experimental introduce the effectiveness of the proposed ABC method.