Dynamic discriminant analysis for predictive maintenance of electrical components subjected to stochastic wear
ISSN: 0332-1649
Article publication date: 1 March 2002
Abstract
An optimal maintenance program for electrical power system components should be based on their reliability. Since, for components characterized by high reliability and cost such as HV circuit breakers, available statistical data are in limited number, a physical model for their ageing is opportune. In the paper a Predictive Maintenance Program (PMP), for determining when a HV circuit‐breaker should be rebuilt, is formalized; it is based upon an adequate stochastic model of electrical wear associated with breaking operations due to system faults. In the model, both fault times and amplitudes are described by means of random variables, in order to deduce a reliability function used as input data for a Bayesian discriminant analysis which dynamically estimates, also in the presence of observation errors, the state of the component, determining the optimal times to perform a maintenance action.
Keywords
Citation
Allella, F., Chiodo, E. and Pagano, M. (2002), "Dynamic discriminant analysis for predictive maintenance of electrical components subjected to stochastic wear", COMPEL - The international journal for computation and mathematics in electrical and electronic engineering, Vol. 21 No. 1, pp. 98-115. https://doi.org/10.1108/03321640210410779
Publisher
:MCB UP Ltd
Copyright © 2002, MCB UP Limited