Applying proportional hazards modelling in reliability

Rotating machinery prognostics: State of the art, challenges and opportunities

“…PHMs assume that hazard changes proportionately with covariates (asset condition in this case) and that the proportionality constant is the same at all time. The use of PHM for incorporating CM information in reliability calculation started in the 1980s [56][57][58][59]. Jardine et al used Weibull distribution to model the baseline hazard function in PHM.…”

Proportional Intensity Models Robustness with Overhaul Intervals

“…Dale 18 applied the PWP approach to simulated failure times data on 30 equipment units for a reliability growth program with design improvements implemented after each of the five stages, resulting in a decreasing rate of occurrence of failures (DROCOF). Wightman and Bendell 19 and Bendell et al 20 also cited the PWP model and advised that engineering applications should be exploratory in nature and performed by investigators familiar with the problem domain. Hansen and Ascher 21 examined a case of automotive intermittent failures, which often lead to a series of unsuccessful repair attempts, and reported that repair times for intermittent failures cannot be assumed negligible and the model must be designed to account for machine downtimes.…”

Optimizing the performance of a repairable system under a maintenance and repair contract

“…A review of the theory of the PIM/PHM in the context of maintenance and reliability can be found in Bendell 16 , Bendell et al 17 , Kumar and Klefsjo 18 , Ansell and Phillips (chapter 3) 19 and Kumar 20 . By a recent review of case studies of the PIM, the authors have found more than 50 applications (see e.g.…”

“…However, the key to getting satisfactory results that are of practical and predictive value is the appropriate formulation of the model itself and explanatory variables. In order to accomplish this, technical knowledge of the system and how time and explanatory variables act on the system must be taken into consideration a priori (Bendell 16 , Ansell and Phillips 21 , Bendell et al 17 ). At the same time, this knowledge is necessary to interpret the results and to verify the practical usefulness of the models.…”