Calculation of power equipment reliability for condition-based maintenance decision-making

Abstract

Reliability is an important index for decision-making in power system. The traditional reliability calculation which is based on time cannot meet the demand of condition-based maintenance (CBM) decision-making. Now power equipment's health state can be quantified by related state evaluation guidelines, so a reliability model which is based on power equipment's health state is put forward to calculate fault rate. The main difficulty of applying the model is to get the two parameters, which are scaling parameter and curvature parameter. Someone uses statistical inversion method to get them, which needs a large scale of samples, neglects samples' health state variability, and can't reflect maintenance's impact on reliability. This paper proposes a new parameter calculation method which is based on power equipment's life cycle health state. By means of this method, we can not only overcome the shortcomings of inversion method, but can also build several kinds of state reliability models, such as state reliability model of one category power equipments, single equipment and single equipment's different life stages. After obtaining power equipment's reliability model based on health state, condition-based maintenance decision-making can be executed further. The target of decision-making is to find out the best maintenance mode such as overhaul or minor repair, which is a best balance between reliability and economy. Different maintenance modes need different maintenance costs, and have different impacts on equipment's health state. By means of analyzing each maintenance mode's ability to improve equipment's health state, we can predict reliability scientifically. Then condition-based maintenance decision-making can be carried out by calculating each maintenance mode's reliability increment per maintenance cost. The paper also gives a numerical example which shows that the decision-making method proposed here is more simple, precise and operational for condition-based maintenance management.