Project Update: Real-time data-driven maintenance logistics

February 2020 – Q&A with Assistent Professor Willem Jaarsveld

When did the project officially start?

The official start of the research project was in the summer of 2018.

How many people are working on the project?

At the time of the interview there are 14 researchers, including three PhD's, a postdoc, professors and university lecturers working on the project. The research side is about 20 people and then master students and who do part of the execution, at the companies a total of 10 people are directly involved, and another 10 people indirectly. Recently the last vacancy has been filled.

How did the project start?

It started before the call for proposal became public. Partering companies felt that research had to be done on the subject and to collect data and valorize it. The Call for proposal accelerated the project and made it scale up. This has led to an actionable agenda between all parties.

What is the biggest challenge so far?

Finding out and setting up the legal part of the collaboration was a lengthy process in which various lawyers had to work together to find out how everything had to be laid down contractually.

Is there a lot of contact with the private partners and what does the collaboration look like?

There was already contact with the private partners and they already had a clear ambition. This corresponded well with the ambitions of the university. A plenary meeting with the companies takes place three times a year and more regularly there are small meetings in which the agendas for the research are drawn up. This is where agreements are made about obtaining concrete results. PhDs visit the companies and talk directly to contact persons. Master's students are also deployed at the companies under the supervision of a PhD.

Were the research questions still adjusted in the course of the research?

No, but as a result of progressive insight, the way in which we formulate these research questions has changed. Projects now start with formulating ambitions and not with transferring data.

In which phase of the research are you now?

The project has entered the phase in which we can make a number of predictions based on collected data and even data that is not necessarily from the asset in question (transfer learning). This can be compared to a person who has played soccer for years, stopped playing soccer and is now playing hockey. Not all the knowledge and skills of football are useful, but the generic elements are, and that probably makes the game easier to understand.

We are now concretely in the phase in which we are working on the actual deployment of the collected data and knowledge. The logistical follow-up of given alerts, the failure behaviour is being predicted better and better, now the step has to be made that action can be taken before an asset fails.

What is your ultimate goal?

To make the chain work in a number of situations for the companies within the consortium. A predictive algorithm is just not worth anything, an agile maintenance department is just not worth anything, and a stand-alone asset monitoring is also not worth anything. The goal is to bring all these things together so that the whole thing can create real value.

26 February 2020


Sign up for our newsletter and stay informed of the latest news Commit2data