Health Condition Assessment of Pole-mounted Switch Assemblies Based on Hybrid Algorithm

With the continuous construction of distribution automation, the reliability of pole-mounted switch assemblies had been paid more and more attention. This paper presents a health condition assessment model based on multi-source data, using support vector regression (SVM), Back Propagation Neural Network (BPNN), Extreme Learning Machine (ELM) and Random Forest (RF). Firstly, the four single evaluation models are established. Then a hybrid algorithm evaluation model of four intelligent algorithms based on the four single evaluation models is established. And in order to optimize the results simulation a hybrid algorithm evaluation model of three intelligent algorithms which eliminating RF algorithm is built. According to the simulation results, the health condition assessment model synthesizing three intelligent algorithms is the best one. The results can be used in engineering practice to arrange the maintenance of the pole-mounted switch assemblies reasonably and improve the reliability of distribution system.