LAUSR

Abstract The V-I characteristic of ZnO varistor which has been obtained from experimental test, is an important parameter to model surge arrester electrical behavior. Generally, in grid voltage, this characteristic depends on temperature variations. In this paper, an improved metal oxide surge arrester electrical model has been proposed under grid voltage. Temperature effects on surge arrester performance have been considered in the proposed model. Artificial neural network has been used to determine V-I characteristic at temperature variations. Therefore, in proposed model, the V-I characteristic of each varistor has been estimated based on applied voltage and operating temperature. To validate proposed model, laboratory setup and infrared thermal camera have been arranged and experimental tests results have been compared with simulation ones. Results revealed that there has been a good agreement between simulation and experimental results. Moreover, electrical behavior of surge arrester is predictable with high precision through the proposed artificial model.