Validation of PERFoRM reference architecture demonstrating an application of data mining for predicting machine failure

Abstract The PERFoRM project aims to develop a reference architecture for Agile Manufacturing Control systems for plug-and-produce devices, robots and machines. The aim of the work is to improve the flexibility of the mechanical manufacturing of the housing parts involved in the production of industrial compressors and gas separators. An industrial demonstrator has been designed to implement a Data Analytics tool that provides rules beneficial for root cause analysis and a decision support system for early prediction of the failures. The tool also identifies key alarms for monitoring the machine condition.