Abstract:Sulfur hexafluoride (SF6) is an artificial inert gas. It is widely used in power systems with its good arc extinguishing ability. However, once the SF6 gas leaks, it will give the safe operation of power equipment and indoor staff security poses a serious threat. Aiming at the problem of insufficient SF6 gas leak detection capabilities, an intelligent diagnosis model based on joint time-frequency analysis, particle swarm optimization (PSO) and support vector machine (SVM) is proposed. First, analyze and process the collected signal, extract the spectral centroid, spectral width, and spectral contrast of the signal to construct the feature vector; secondly, use the particle swarm algorithm to solve the problem of the SVM model in the selection of penalty parameters and kernel function; finally, establish a particle swarm Algorithm-optimized support vector machine model (PSO-SVM); through comparison with other methods, it is proved that this research method is superior to other machine learning methods in terms of recognition accuracy. Experimental results show that this method has significant advantages in the identification of SF6 gas leakage.
Keywords: Gas leaking; Joint time-frequency analysis; Particle swarm optimization; Support vector machine