An Ica-Based Audio Feature Fault Detection Method For Transformer Equipments

Experienced engineers in transformer substation can judge the equipment condition via just listening to the working sounds of electrical equipments. Use audio signal processing applied in engines and other mechanical equipments for reference. A scheme to monitor the working condition of electrical equipments is proposed. Firstly, the basic principles and system structure of this scheme is outlined. It introduces the method of colleting electrical equipments' working sounds by Microphone array, because Microphone array form a beam to target the source sound, which can reduce the noise and reverberation. When substation is working, the environmental background interference sounds exist and are independent from electrical working sound. So we can use FastICA algorithm that is based on the largest negentropy to separate the collected sound to several independent source signals. It has the advantage of fast convergence and robust. The simulation result shows this algorithm can effectively separate the multiple independent source signals. The separation accuracy is above 95% for typical sample mixed sounds and the reliability of electrical equipment fault detection system based on audio signal processing is ensured.