The study of optimized grey model using for transformer fault prediction

Abstract

In order to ensure the safe operation and timely maintenance on mine flameproof dry-type transformer, we use MATLAB as the platform to get the simulation and to complete the edition of prediction program. After the comprehensively considering the accuracy and timeliness of the algorithm, our final optimization plan is: the GM(1,1) model must be optimized in one aspect: the accumulated order optimization, and then corrected the residuals. Compared to the traditional GM(1,1)model, the optimized and corrected GM(r,1) can improve the accuracy of the prediction result and guarantee the quick reaction of the pre-warning system before the occur of failures. Using this model enables the system to predict the potential faults in a certain period of time in the future, which can improve the actual operation efficiency of mine flameproof dry-type transformer and prolong the transformer's working life.