HALFBACK Project: The Use of Machine Learning to Achieve High-Availability in Production

Abstract

The HALFBACK project’s goal is to achieve a high-available production, mainly by predicting failure of a manufacturing machine and machine tool to avoid downtime, associated costs, and reputation loss.
In this paper two different approaches of machine learning prediction are described. First a pattern mining approach analysing condition events and second a neural network based approach. Further, the paper discusses the need for data pre-processing and ideas how to achieve high-availabilty in production.