Prognostic Reliability Prediction for Repairable System Based on Non- Parametric Model


Estimation of the reliability for repairable system after maintenance actions is usuallybased on mathematical models, which can be classified as parametric and non-parametricmodels where the parametric model is required a prior specified life time distribution whileNon-parametric model is that relaxes of the assumption of the life time distribution.Nonparametric life time models are including proportional hazard model and proportional oddmodel. In this paper we develop repairable reliability model concentrate on generalizedrepairable model that indicate the mixture of proportional hazard model and proportional oddmodel. A proportional hazard-proportional odds (PH-PO) model for the purpose of toimprove the repairable reliability to obtain accurate estimates of reliability for repairableindustrial boiler system at normal operating conditions depending on transformationparameter for reliability prediction for repairable system that represent Beji industrial boilerin power plant. The results show the odd model better than hazard model for repairablesystem after preventive maintenance depends on time to repair where transformationparameter (c) equal 0.0525094 it is closer to odds model than hazard model.In addition, reliability industrial boiler in case without temperature effect is better thanreliability with temperature effect by using exponential model where we note that thereliability at 500 it is worse state where degrade more than (400,450) .