Assessment of repairable-system reliability using proportional intensity models: a review

This paper provides an overview of methods, and surveys the literature on engineering applications of proportional intensity (PI) models with explanatory variables (covariates), for repairable systems reliability assessment. The semi-parametric PI method relaxes the assumption of an underlying distribution, and is potentially useful in engineering practice, where the underlying information for a failure process is usually not available. PI semi-parametric models initially proposed for clinical studies in medical applications include PWP (Prentice, Williams, and Peterson), AG (Andersen, and Gill), and WLW (Wei, Lin, and Weissfeld). Abundant funding received in medical research has advanced PI models to become well developed, and widely referenced in the biostatistics field. This paper reviews both the available methods for repairable-system reliability assessment, and the published engineering application case studies. An engineering application example that applies PI model to a maintainability process used in US Army M1A2 Arams Main Battle Tank is presented.