Reliability modelling for electricity transmission networks using maintenance records

Abstract

Maintenance decisions for transmission network assets (TNAs) require accurate reliability prediction. However, there are a large number of operating, design and environmental variables that potentially influence their reliability. This paper presents a new reliability prediction method for TNAs. Failure times were identified by extracting significant unplanned maintenance events for critical failure modes. A regression tree-based grouping analysis was utilized to analyse the influences by variety of factors on future unplanned maintenance. These results were then used to build the reliability prediction model allowing a decision maker to have an estimate of future unplanned maintenance requirements. A case study using real industry data was conducted to test the proposed reliability prediction model. The results demonstrate the feasibility of using this approach for TNA maintenance decision support.