A DATA-DRIVEN FRAMEWORK OF PREDICTING THE EFFECTIVE SWITCHING CYCLES OF ON-LOAD TAP-CHANGER CONSIDERING DEGRADATION

The total number of switching cycles is the most crucial performance index of the on-load tap-changer (OLTC), which determines its lifespan. Thus, this study explores a method to estimate the effective switching cycle number of OLTC based on the Kriging model considering degradation. First, the mechanical fault causes of the OLTC are analysed via establishing the fault tree, by which the principal factors of spring stress relaxation and parts wear that affect the switching cycle are determined. Then, the parameters with respect to the factors over the whole performance life are accurately identified and matched with the help of testing data. Subsequently, based on the transition time, the precise dynamic model of OLTC is constructed, and the contact separation time under the deterministic condition is obtained and verified using dynamic analysis strategy. Finally, the sampling data and corresponding contact separation time are introduced into the Kriging model, and then the effective switching cycle number is predicted according to the failure criterion. This study can provide a new data-driven framework for lifespan prediction of OLTC.