Estimation of switching surge flashover rate of point on wave switching over‐voltages along transmission line by adaptive neuro‐fuzzy inference system meta‐model

Abstract

Limiting switching overvoltages (SOVs) and performing proper insulation coordination against stresses caused by them have great importance in UHV transmission lines (TLs). The point on wave switching (PWS) is the best strategy to limit SOV without needing any arrester. Although detailed electromagnetic transient studies are carried out in the design of transmission systems, such studies are not common in day-to-day operations. However, it is important for the power utility and/or operator to ensure that the peak value of SOVs is well within the safe limits. This study presents a new PWS strategy in the EMTP/ATP environment by considering line-trapped charge to train an adaptive neuro-fuzzy inference system meta-model used to estimate the SOVs and to determine spots of critical failure risk caused by SOVs along TLs. The proposed meta-model can be used by power utilities and/or design engineers for planning the proper insulation level without consuming time to meet a desirable value of risk. Moreover, the operators can decide on the energisation of lines in sequence, which is safe and leads to successful energisation. In order to reduce the training error, an intelligent method based on two-stage data classification is introduced in which k-means clustering method is used.