H2020 Z-BRE4K - Interview with Dimitris Kiritsis
Interview to Dimitris Kiritsis, Professor at the Ecole polytechnique fédérale de Lausanne (EPFL) and member of the Governance Board and the Technical Oversight Board of IOF (Industrial Ontologies Foundry).
- Prof. Kiritsis, EPFL is involved in the semantic modelling of production assets, products and processes within the Z-BRE4K project. Can you please describe in detail what activities have been carried out by you within Z-BRE4K?
I coordinated the activities of the development of the semantic modelling framework for Z-BRE4K and I took care that the Z-BRE4K ontologies are developed following the state-of-the-art approach and guides recommended by the Industrial Ontologies Foundry (IOF)
- What is semantics and what is its purpose within Industry4.0?
One of the main challenges of the digitalisation of industry is the interoperability of its various and heterogeneous connected information systems and data bases. One prerequisite for this to happen is the ability of the designers of such systems and of the systems themselves to understand and capture “the Meaning of Data”, in other words the “semantics” of data. This requires the use of commonly agreed vocabularies and the fomalisation of the involved entities and associated terms and their relations in appropriate information models that we call “ontologies”. The use of commonly agreed ontologies for industry is the foundation layer of interoperability of the connected information systems: the so called “semantic interoperability”
- How can Predictive Maintenance benefit from a semantic modelling approach?
As it has been proved and demonstrated in the industrial use-cases of Z-BRE4K, a Predictive Maintenance solution is based on the trusted and standardised “data sharing” stored in heterogeneous data sources by the involved stakeholders. The semantic modelling of data provided by such heterogeneous sources using standardised ontologies is a robust foundation enabling the required trusted and standardised sharing of the data necessary to build and operate a Predictive Maintenance solution.
- In this regard, what type of information do we manage in Z-BRE4K and how do you combine and use them?
In Z-BRE4K the subject of Predictive Maintenance are industrial engineered assets such as machine tools and other types of automated machines. For a Predictive Maintenance solution we need data about both (i) the machines themselves such as BOM data, functional and technical characteristics of the components of the machines, etc., and (ii) the behaviour of the machines such as temperatures, speeds, vibrations, electrical signals from the motors etc. Also, historical data such as operational times, starts and stops, previous maintenance actions are also very useful to be considered. The correlated analysis of such data allows to assess and understand the status of “health” of an industrial asset which reflects the status of its “degradation”, which is the reason for possible defects and failures. Based on such an assessment of the health of an industrial asset, a Data Driven Predictive Maintenance solution allows to predict possible deteriorations and failures before they occur.
- EPFL is part of the Industrial Ontologies Foundry, and notably of the Maintenance WG, what is your role within it?
I have the honour to be one of the initiators of IOF (Industrial Ontologies Foundry) together with Prof. Barry Smith of the University of Buffalo and among its founders in September 2016 together with organisations such as NIST, Dassault Systems, Airbus, Autodesk, Lockheed Martin and others. I am member of the Governance Board and the Technical Oversight Board of IOF and, together with members of my group, of the WG on Maintenance and PSS (Product Service Systems), where we contribute to the development of the IOF Ontology development approach and the Maintenance and PSS ontologies and associated guides and standards which are expected to be officially published soon, as soon as IOF is becoming an international legal entity, expected to happen in the next few months.
- Which can be Z-BRE4K’s role in the development of standards?
Through our above contribution in the IOF Maintenance WG, we hope the IOF Maintenance Ontology to become one of the necessary standard for the development of interoperable Predictive Maintenance solutions.
- Lastly, EPFL is also responsible for developing strategies to improve the maintainability and life-cycle of machines and components. How did you build those strategies?
Since the beginning of the 2000, my group at EPFL was strongly active and heavily involved in the development of the pioneering concept of “Closed Loop Lifecycle Management” enabled by the so called “Product Embedded Information Devices” (PEID) such as sensors and RFIDs. It seems that the current I4.0 technologies and standards are effectively used to implement and realise Closed Loop (Circular) Lifecycle Management solution, with Predictive Maintenance being one of them, very important for the Middle-Of-Life (MOL) of industrial bust other engineered assets as well.
H2020 - Z-BRE4K