Mechanical Fault Diagnosis of Reactor Core Winding Looseness Based on Multi-feature Fusion and Improved KELM

Abstract

To solve the problems of frequent occurrence of loose cores of high-voltage shunt reactor (HVSR) windings and low operating efficiency, this paper proposes a fault diagnosis method based on multi-feature fusion and an improved bat optimization-based kernel extreme learning machine (IBA- KELM). This method is mainly based on the time-frequency feature fusion of the vibration signal of the reactor. Firstly, this paper extracts the time-frequency domain feature quantities of the original vibration signals of multiple sensors, and performs parallel superposition and fusion of the feature levels to obtain a fusion data set. Secondly, using the fusion data sets, this paper establishes an IBA-KELM-based reactor winding core looseness fault diagnosis and identification model. Finally, the experimental data of the 20kVA reactor experimental platform is adopted to verify the effectiveness and superiority of the proposed method. Experimental results demonstrate that the proposed method has higher recognition accuracy and diagnosis accuracy compared with similar algorithms.