Voiceprint Fault Diagnosis of Converter Transformer under Load Influence Based on Multi-Strategy Improved Mel-Frequency Spectrum Coefficient and Temporal Convolutional Network

Hui Li,Qi Yao,Xin Li
DOI: https://doi.org/10.3390/s24030757
IF: 3.9
2024-01-24
Sensors
Abstract:In order to address the challenges of low recognition accuracy and the difficulty in effective diagnosis in traditional converter transformer voiceprint fault diagnosis, a novel method is proposed in this article. This approach takes account of the impact of load factors, utilizes a multi-strategy improved Mel-Frequency Spectrum Coefficient (MFCC) for voiceprint signal feature extraction, and combines it with a temporal convolutional network for fault diagnosis. Firstly, it improves the hunter–prey optimizer (HPO) as a parameter optimization algorithm and adopts IHPO combined with variational mode decomposition (VMD) to achieve denoising of voiceprint signals. Secondly, the preprocessed voiceprint signal is combined with Mel filters through the Stockwell transform. To adapt to the stationary characteristics of the voiceprint signal, the processed features undergo further mid-temporal processing, ultimately resulting in the implementation of a multi-strategy improved MFCC for voiceprint signal feature extraction. Simultaneously, load signal segmentation is introduced for the diagnostic intervals, forming a joint feature vector. Finally, by using the Mish activation function to improve the temporal convolutional network, the IHPO-ITCN is proposed to adaptively optimize the size of convolutional kernels and the number of hidden layers and construct a transformer fault diagnosis model. By constructing multiple sets of comparison tests through specific examples and comparing them with the traditional voiceprint diagnostic model, our results show that the model proposed in this paper has a fault recognition accuracy as high as 99%. The recognition accuracy was significantly improved and the training speed also shows superior performance, which can be effectively used in the field of multiple fault diagnosis of converter transformers.
engineering, electrical & electronic,chemistry, analytical,instruments & instrumentation
What problem does this paper attempt to address?
### Problems the Paper Aims to Solve This paper aims to address the issues of low recognition accuracy and difficulty in effective diagnosis in traditional voiceprint fault diagnosis of converter transformers. Specifically: 1. **Improving Diagnostic Accuracy**: - A new method is proposed that considers the impact of load factors and utilizes improved multi-strategy Mel Frequency Cepstral Coefficients (MFCC) for voiceprint signal feature extraction. - Fault diagnosis is performed in combination with a temporal convolutional network. 2. **Optimizing Parameter Selection**: - The Hunter Prey Optimization (HPO) algorithm is improved, using IHPO combined with Variational Mode Decomposition (VMD) to denoise the voiceprint signal. - The Stockwell transform is used to combine the preprocessed voiceprint signal with the Mel filter, further processing it to adapt to the stationary characteristics of the voiceprint signal. - Ultimately, multi-strategy improved MFCC is used for voiceprint signal feature extraction. 3. **Introducing Load Signal Segmentation**: - Load signals are introduced to segment the diagnostic interval, forming a joint feature vector. - An improved temporal convolutional network (ITCN) is proposed using the Mish activation function, and the IHPO-ITCN model is introduced to adaptively optimize the convolution kernel size and the number of hidden layers, constructing a fault diagnosis model for converter transformers. Through specific instances, multiple sets of comparative experiments are constructed and compared with traditional voiceprint diagnosis models. The results show that this model has a fault recognition accuracy of up to 99%, and the training speed also demonstrates superior performance, making it effectively applicable in various fault diagnoses.