On the 8th of December we organized an event on “Condition monitoring & forecasting” at ASML. All in all, we look back on an energizing session with both high-level content presentations and a fantastic tour through the experience center of ASML. We have shared first results, best practices and lessons learned, and had an exploratory view on the possible benefits for service supply chain. Condition monitoring & forecasting seems to be a hot topic for many companies, or is now on the top of the research agenda of companies still exploring the benefits of it!
First results are promising!
Robert Lemmens (Marel) presented a case study in which they demonstrated the value add of condition monitoring techniques. By monitoring chain length differences, they are able to accurately predict failures. Nice side effect from this study, is that they found significant quality differences between their suppliers of the chain parts. Marel now has a company wide challenge how to valorize the findings from this case study, as this requires significant changes of their business model.
Multiple approaches seem successful
Margot Peters (Dutch Railways) presented their research on condition monitoring. Big data is analyzed in multiple ways, and their conclusion: “It takes two to tango!”. That is, you need to combine the knowledge of both data scientists and physicians in order to create value add of big data. Their first results are promising: the time period to predict failures in advance ranges from 1 minute to 2 days and in one case even two weeks!
ASML decided to start dancing the tango alone. They used datamining techniques to find anomalies in big data sets. Their data driven model outperforms their physical model in predicting failures (82% vs. 55%). Now they revert the insights from this data driven model to their physicians in order to get a better understanding of their (physical) machine functioning.
A promising future is ahead of us!
Simon Jagers (Semiotic Labs) highlighted the need for condition monitoring and forecasting. Trends (skilled labor shortage, servitization) and opportunities (less uplanned asset downtime, costs savings) form the basis for his dream: “maintenance on (mechanical) parts of assets is 100% predictive by 2020!”. Multiple cases he researched truly gave him the belief that this dream is in fact realistic.
Maarten Driessen (Gordian) presented an exploratory view on the impact of condition monitoring and forecasting on service (parts) supply chains, and how service organizations could benefit from condition monitoring information in terms of asset operational availability, working capital and operational costs.
Interested in the presentations? Please send us an email: firstname.lastname@example.org.