Insights and feedbacks
UPTIME 1st Live Webinar
Lessons Learned and Best Practices
in White Goods Industry
The UPTIME 1st live webinar “Lessons Learned and Best Practices in White Goods Industry”, held on 19 March 2020, attended by 30 stakeholders, was focused on UPTIME white goods business case, which deals with Whirlpool’s complex automatic production line that produces drums for clothes dryers.
UPTIME partners Pierluigi Petrali (Head of Manufacturing R&D Whirlpool Corporation), Manuele Barbieri (Corporate R&D Project Technical Coordination – RINA Consulting S.p.A.) and Dr. Fenareti Lampathaki (Technical Director Suite5 – Data Intelligence Solutions Limited) presented:
1) Differences between classic preventive maintenance and predictive maintenance using historical data and real time data 2) Benefits and some lessons learned of predictive maintenance by a concrete implementation in the White Goods Industry 3) Live demonstration of the implementation in the UPTIME Platform.
Moreover, during the webinar, we gained useful insights and feedback from the audience. In this article, we present 3 main insights from the discussion on the lessons learned and the main feedbacks received from the audience.
Insights from the lessons learned
1) FMECA is a powerful tool. It is an important tool for implementation of predictive maintenance means, especially on the components of high value.
2) Correlate data is the key. Data alone do not mean anything. We need also a standard semantic model in order to understand the meaning of data without wasting time.
3) Prediction must gain the trust of the process experts. The predictions need to be validated and the algorithms need to be fine-tuned and challenged with the real world.
Equally important, but it is underestimated, is to share the knowledge between process experts and data experts, as well as to increase the knowledge of each one in the field of the other.
“We are not just talking about technology, we are talking about the way people are using some suggestions generated by a machine. We could have a perfect prediction by the machine, but if the people in the factory, who should take this output are not able to read it, it is useless.” Pierluigi Petrali, Whirlpool Corporation
Audience questions and feedbacks
Main questions from the audience are related to how UPTIME deals with anomaly detection and false alerts as well as which kind of algorithms are used for failure prediction. What are the number and types of sensors? Is the prediction tool able to predict a failure in the next 2-3 days? What is the ROI of such a solution? What is UPTIME experience using standards in predictive maintenance?
ISO 13374 and the compliant on MIMOSA schema are the basis of the UPTIME data model. The UPTIME platform is developed in compliance with the RAMI 4.0 architecture. More details on the discussion is available on the webinar recording.
At the end of the webinar, some participants have taken part in the provided survey. The answers to the questionnaire confirm that there is a high level of interest of the participants to such event that shares real-life predictive maintenance experience. These answers help us to refine the future roadmap of the UPTIME platform, such as consideration of new features e.g. integration with the spare parts management system, and also to better identify obstacles, as for instance provision cost/benefit analysis to convince decision makers.
Did you miss this webinar? You can find the webinar recording and the presentation here.
Lastly, we would like to invite you also to join the UPTIME community partner program (free of charge) and benefit from new advances in predictive maintenance. If you have any questions or comments, please contact us.