1. ¹Bruce Power LP, Tiverton, Ont., Canada
2. ²University of Toronto, Toronto, Ont., Canada
3. ³Brock University, St. Catherines, Ont., Canada

Correspondence: EK Doyle, Bruce Power LP, Box 4000B12, Tiverton, Ont., Canada N0G2T0. E-mail: kevin.doyle@brucepower.com

Received February 2007; Accepted October 2007; Published online 20 February 2008.

Abstract

Variant forms of reliability centred maintenance (RCM) have been the maintenance improving tools of choice for the last 20 years. In this case study paper, justification is made for implementing, and a path is laid out to implement, Operations Research in the form of statistical modelling as the next step forward after RCM. The lack of failure data issue has been addressed by using elicitation protocols to provide component lifetime distributions. Graphical analysis and Crowe/AMSAA (Army Materials Systems Analysis Activity) methodologies are developed as a basis for justifying expenditures on maintenance improvement initiatives. A review of historical empirical—inferential techniques dating back to WWII is presented as well as discussions of the current applicability of the same. The deregulation of the Electrical Generation Industry has produced severe restraints on the publishing of failure data. Owing to a fortuitous set of circumstances, a limited amount of data became releasable, which allowed the promise of the method to be demonstrated.