Analysis of SF6 contact based on QPSO-SVR

Abstract

In the condition evaluation of the high-voltage SF6 circuit breaker, contact resistance and mass loss have a significant impact on the arc contact. To that end, this paper proposes a method based on quantum particle swarm optimization and support vector regression (QPSO-SVR), the implementation of which can effectively predict the contact resistance increment and mass loss of the circuit breaker arc contacts under different arc current conditions, and the best support vector regression (SVR) algorithm training parameters are obtained through experimental data. To validate the proposed method’s accuracy, it is compared to other prediction methods, and the results show that the QPSO-SVR method has good predictive ability for experimental data under different discharge parameters. The relative error of prediction for contact resistance increment is 3.023%, and the relative error of prediction for mass loss is 4.61%, indicating good accuracy and robustness. It can serve as a reference for the maintenance of high-voltage SF6 circuit breakers, which is useful. It is of great significance to the maintenance of SF6 circuit breaker.