Data-driven models for advanced fault detection


A data-driven model of the dynamic manufacturing system is a method to provide insights on the stochastic nature of the operation and interdependence of equipment. Currently no closed form equation exist to model large manufacturing systems. Using plant floor data it is possible to study performance and the effects of machine interaction on throughput and quality. Using machine learning and time series analysis, it is possible to identify patterns and trends from the data. Moreover, different machine operating conditions can be studied and tested using discrete event simulation. The data-driven models and simulations can help plant floor operations by suggesting optimal operating parameters to improve different aspects of manufacturing.
Researchers: Shuai Xiong, Miguel Saez