A prediction techniques in resistive current measurement of metal oxide arrester

It is impossible for quantitatively forecasting the deterioration trend and service life through periodic test. The current of MOA under the operating voltage reflects the insulation performance or the nonlinear properties of its resistor. In this paper, a new predicting method is introduced to combine gray prediction using equal dimensional innovation with BP Neural Network through the optimal weighting algorithm and the method can be used to predict leakage current of MOA by on-line test. Through the experiments of current predicting of MOA, it is proved that the proposed method is of availability and practicality in the measurement. The method also provides a new idea for the data analysis in other experiment items in condition-based maintenance.