A PdM framework Through the Event-based Genomics of Machine Breakdown

A novel event-based predictive maintenance framework based on sensor signal measurements and regressive predictions to minimise machine breakdown and component failure is proposed. Such capabilities will be complemented by Event-Clustering technique to cluster and remove less impact sensor signals and also build breakdown genomics from the root of a failure in order to predict the upcoming machine breakdowns and components failures. The creation of machine breakdown genomics requires the knowledge of systems state observed as well as the state change at specified time intervals (discretization). The proposed framework is applied to a real application case study. An industrial case study of a continuous compression moulding machine that manufactures the plastic bottle closure (caps) in the beverage industry has been considered as an experiment. The machine breakdown genomics theory is tested in this case to build the sequence of events or the genomics of breakdown, where sequences of contiguous events lead to failure or healthy machine status. This is complemented by the Regression Event-Tracker method to estimates the condition monitoring of the components and provide components real-time remaining useful life estimation. The Weibull failure-rate analysis is carried out on the remaining useful life estimates for each element to understand and estimate the mean time to failure for the manufacturing machine.

Download paper here

Recommended citation: M. Danishvar, V. C. Angadi and A. Mousavi (2020), "A PdM framework Through the Event-based Genomics of Machine Breakdown", In The 9th Asia-Pacific International Symposium on Advanced Reliability and Maintenance Modelling, pp. 1-6, Vancouver, Canada.