Modelling Preventive Maintenance based on the delay tiMe concePt in the context of a case study

Using the delay time concept and associated models, this paper presents a modelling study of optimising the preventive maintenance (PM) interval of a production plant within the context of a case study. To establish the relationship between the PM interval and expected downtime per unit time, we need the data of both failure times and the number of defects identified and removed at PM epochs. However, the available data to us was only the recorded times of failures. To overcome this problem, we obtained an estimated mean number of the defects identified at the PM epoch by the plant maintenance technicians. Based on these two types of data, we first establish a likelihood function of the observed times to failure and then a squared function of the difference between the number of defect identification estimated by the technician and the corresponding expected value from the model is mixed with the likelihood function to estimate the unknown model parameters. We test by simulation to show the validity of the above parameter estimation method. Once the parameters of the model are known, a PM model is proposed to optimize the expected downtime per unit time with respect to the PM interval. The modeling process is demonstrated by the case study presented.