The power distribution system in coal mining areas is not only one of the infrastructure supporting coal mining production and operation, but also an indispensable key link in the entire energy industry chain. And its stable operation is crucial for the production safety and efficiency of coal mining equipment. The operating conditions of the power distribution system of coal mining equipment are complex, and the current detection scheme of the power distribution system does not extract the voiceprint features of the power distribution system, so the accuracy of fault detection is low. A multimodal fault identification method based on voiceprint features has been proposed to address this issue. This method first designs a sound signal acquisition scheme to obtain raw data, and then uses an improved noise filtering method to preprocess the collected data and extract the main features. Finally, taking the main features of the voiceprint signal as input, a fault recognition model integrating voiceprint features was constructed by combining the improved Pelican optimization algorithm and convolutional neural network. The experimental results show that the accuracy of the training and testing sets of the recognition model constructed in the study is 95.63% and 96.45%, respectively. The above data indicates that the proposed method has significant effects on improving the accuracy and robustness of fault identification in energy mining equipment and power distribution systems. It also provides effective technical support for similar power distribution systems, which is of great significance for promoting the sustainable development of the energy industry.
