Today’s technology means that maintaining equipment often involves analysis of a large amount of data. Keeping an eye on every element of a machine, knowing when something has gone wrong and what the root cause is becomes increasingly difficult as machines become more complex.
It can be hard to find the right computer software to analyse all the information required to maintain a physical system, be it an energy source like a wind turbine, an industrial machine or a vehicle.
Now, the EU, ECSEL Participating States and industry-funded MANTIS project is working to provide companies with everything they need to keep track of their machines and keep them running – from the sensors to the data analysis – using artificial intelligence. The team is developing what it calls its own maintenance service platform.
The project will help businesses save time and money by enabling them to anticipate large problems before they occur or even to change the working conditions of assets, allowing maintenance tasks to be scheduled at an appropriate time for the business.
“We are now working on getting feedback from experts and stakeholders in order to make the last iteration [of the platform] before publishing it,” says project coordinator Urko Zurutuza of Mondragon University in Spain.
Because the project involves developing something suited to a huge variety of uses, the team went to the type of businesses that would be using the system first, to ask what they would need from a maintenance system.
They took into account everything from the sensors they would use, the purpose for the intelligent algorithms and the human-machine interface (HMI) scenarios, to the way they would like to provide maintenance services. From these results, they began to develop a reference that could be tailored to suit each individual requirement.
The prototype platform is nearly finished. When complete, it will consist of a network of smart sensors that keep an eye on the machine and the environment around it, and robust communication mechanisms for transferring that information in harsh environments.
It will have machine-learning capabilities to make the platform smarter and cloud-based data-processing and analysis. The platform will also involve an HMI that will provide information to the people involved when it is needed.
The platform takes into account two kinds of data analysis: the data gathered by an individual company and the data shared in the cloud by stakeholders, which can be used for machine learning. By putting the two together, it can help build up a catalogue of recommended techniques.
Building on success
“They can serve for early failure diagnosis and prediction, root-cause analysis of asset failures, for estimating the remaining useful life of assets, for collaborative decision-making, and for optimisation of maintenance plans,” says Zurutuza.
The team has high hopes for the MANTIS platform. Two of the partners involved in the project have already started to implement maintenance platforms based on MANTIS, specific to their own sectors. A follow-up project has also started, known as Prophesy, which hopes to build on the success of MANTIS.
The project team aims to build on its academic and industry-based research partners, with seven PhD students already contributing to the research. Industrial partners are expected to showcase what MANTIS can do during Hannover Messe, one of the world’s largest trade fairs, scheduled for April 2018.
The project was funded through the ECSEL Joint Undertaking, a public-private partnership keeping Europe at the forefront of technology development in the electronic components and systems sector. ECSEL JU is running an ambitious programme with a financial contribution from the EU, industry and research organisations and 29 participating European States approaching EUR 5 billion for the period 2014-2020.