Transformer fault acoustic identification model based on acoustic denoising and DBO-SVM

In order to address the impact of ambient noise on the acquisition of the transformer body acoustic pattern and completely extract the information contained in the transformer acoustic pattern, This study suggests a transformer core loosening identification technique based on the integration of wavelet threshold denoising and VMD with Identification of transformer core loosening using DBO (Dung Beetle Optimization Algorithm)-optimized SVMs (Support Vector Machine) and the difference between the kurtosis value and the mixed acoustic signal is used to get the de-noised signal with a high signal-to-noise ratio. After that, the signal is fed into an optimized support vector machine for training in order to produce the core loosening identification model. By means of the no-load tests conducted on a 500 V transformer and the examination of the acoustic signals gathered with varying levels of core looseness, the findings demonstrate that the transformer core looseness identification model with the denoised MFCC feature parameters and the dung-beetle algorithm optimized support vector machine in this work achieves an accuracy of 96.25%, hence improving the core looseness fault identification rate.