Cross Stockwell transform aided Random Forest based surface condition identification of Metal Oxide Surge Arrester employning Leakage current signal

This paper proposes a methodology to monitor the surface condition of Metal Oxide Surge Arrester (MOSA) in power system. Generally, the surface of polymeric housed MOSAs are affected due to various pollutants. These pollutants may result overheating of the arrester which can leads to failure of arrester. This can affect power system reliability. Therefore, a method has been proposed to estimate surface contamination of MOSA in this article based on time frequency analysis of leakage current. According to the result it can be said that proposed methodology returned very good accuracy regarding surface contamination level estimation of MOSA which in process enhances the system reliability. Therefore, the proposed methodology is useful for condition monitoring of MOSA at polluted environment. This methodology can also be used to estimate surface contamination level of insulators in service.