Estimating Probability of Gas Breakdown Using Grey-Fuzzy Logic

Designing the protection system of high voltage power systems involves elucidating the properties of gas breakdown. This work presents a simplified mathematical approach based on grey and fuzzy methods to estimate the probability of gas breakdown, which differs from other grey-fuzzy methods. With the help of the data normalization process and a linear fuzzy membership function in determining the proper value of the distinguishing coefficient, the proposed approach outperforms other grey-fuzzy-based methods, the curve fitting method, and the neural network method in terms of estimation accuracy. The average relative error of test cases 1 and 2 with the normalization processing is 6.43%, which is 2.53 lower than that of the neural network method, 4.35% lower than that of the curve fitting method, 17.64 lower than that of the 5-interval max-max fuzzy-grey method, 1.51% lower than that of the 5-interval min-min fuzzy-grey method, 13.87% lower than that of the 3-interval max-max method, and 2.72% lower than that of the 3-interval min-min method