FailureProcessing Scheme Based on Kalman Predicti

Authors: S K Yang

Publish Date: 2011/03/26

Volume: 11, Issue: 4, Pages: 417-431

PDF Link

Abstract

A 25 kVA generator is one of the most important components in the power system of an indigenous defensive fighter IDF However the generators have failed frequently since being equipped on IDFs The poor reliability and long repair time of a 25 kVA generator have seriously impacted the availability of IDFs This study aims to build a failureprocessing mechanism for the 25 kVA generators based on the stateestimation function of the Kalman filter By way of this mechanism the left or rightside load of the fighter can be switched from the supplied generator that is going to fail to the other generator before the abnormal generator actually fails Therefore this mechanism not only mitigates the impact resulting from the irregular power on systems of the fighter but also obtains reaction time to shutdown the generator that is going to fail so as to perform a preventive maintenance or other effective actions on the generator such that the further failures can be avoided In the simulations of this study the failure times predicted by the Kalman filter are compared with the results obtained by Monte Carlo simulation for judging the prediction accuracy In order to simulate more accurately an exponential attenuator is placed at each end of the outputs of both the generator and the Kalman filter so as to conform to the actual aging model The prediction errors are acceptable Besides reliabilities of the generators in four different configurations are compared to know the effect of redundancy design

Keywords: