Predictive maintenance using FMECA method and NHPP models

Most of predictive maintenance technologies are inaccessible to small scale and medium scale industries due to their demanding cost. This paper proposes a predictive maintenance policy using failure mode effect and criticality analysis (FMECA) and non-homogeneous Poisson process (NHPP) models which require minimal use of advanced monitoring technologies and sophisticated data acquisition systems. Most of the repairable systems show long term reliability degradation with repeated overhauls. Here, critical component of a system or machinery exhibiting sad (deteriorating) trend is used as an indicator to predict overall maintenance time of a system. Firstly, the component to be used as an indicator for predictive maintenance is chosen using FMECA method, in which the most critical component is chosen. Secondly, the failure data of the chosen component is analysed using NHPP models and based on analysis of the data, relevant NHPP model is selected and finally, the Mean Time Between Failure (MTBF) of the component is compared with the threshold mean time between failure [MTBF(Th)] of the component to decide the overall maintenance time for the system. The developed methodology is validated on an overhead crane in a steel manufacturing company.