Generating added value from data
How can machine breakdowns be predicted and avoided using production data? In the Leading-Edge Cluster it's OWL, the two partner organisation Fraunhofer IEM and BENTELER are working together on a pilot factory on predictive maintenance to optimise and streamline production and maintenance processes. The research cooperation in OstWestfalenLippe is part of the EU project BOOST 4.0 on Big Data in Industry.
Big data stands for very large amounts of data that are too complex to be evaluated using manual or conventional methods. In the Automotive Division of the global player BENTELER Group, such large amounts of machine data from a hydraulic press and a material handling system are systematically recorded, evaluated and used to derive patterns in the production process. This analysis of data in companies is not a matter of course, as Sebastian von Enzberg, Senior Expert for Industrial Data Science at the Fraunhofer Institute for Mechatronic Systems Design, explains: "A large part of the data generated in production is volatile and often remains unused. In order to create value from them, data must be contextualised, stored persistently and exchanged securely. The data analysis itself must contribute to value creation and requires structural transformations of business processes".
Optimisation thanks to machine learning methods
BENTELER and Fraunhofer IEM are therefore jointly developing a process model that can be used to predict the maintenance and inspection requirements of production machines. First, BENTELER uses sensors to capture all data generated during the manufacturing process and stores it in a data architecture. The experts of Fraunhofer IEM analyse these data using data-driven modelling procedures and machine learning procedures in order to identify patterns in the maintenance process. This enables BENTELER to detect machine faults before they happen and to forecast maintenance work in the future. However, this requires a holistic view of all components and processes. In addition to the individual further development of data analysis methods, data sources such as sensors or machine data are systematically analysed in production and enriched with expert knowledge, i.e. knowledge of the functioning and interaction of systems. These procedures are then embedded in existing manufacturing and maintenance processes in order to support employees in the maintenance and inspection of machines and thus prevent operational failures. The goal is thus a reference process as well as a tool chain for predictive maintenance, from data acquisition to integration into the maintenance process.
The potential of this solution is high: Companies can thus prevent disruptions, anticipate downtimes and make production management more efficient. Machine manufacturers can use the data about their brands to provide valuable information for product optimisation.
Training for Data Scientists
In order to enable employees to work with data science procedures and methods, Fraunhofer IEM has developed a training concept that originated in the Boost 4.0 project. The concept is based on the assignment of different professional profiles to application areas as well as competences in different forms. The training takes into account in particular the position in the company, the background knowledge and the required competencies of the employees. Both the depth of application of data science in the company and the team constellation of data science departments are taken into account as well. This way, the findings of the project are made available to a broader audience through the Leading-Edge Cluster.
BOOST 4.0 and it´s OWL
The optimal use of data, a common European Industrial Data Space, industrial Big Data solutions close to the market: In the EU project Boost 4.0, 50 partners from 16 countries are working on instruments and methods for a common data infrastructure. Here, existing German reference architectures such as RAMI 4.0 or Industrial Data Space are integrated. Industrial companies are enabled to collect, manage and comprehensively utilise data generated in production. The project is currently the largest European initiative for Big Data in industry. The project volume includes a grant of around 20 million euros by the European Commission and investments of 100 million euros by the participating companies. The BOOST 4.0 project is funded by the European Union's Horizon 2020 research and innovation programme.
In order to ensure practical relevance and feasibility of the solutions in the project, ten networked, intelligent factories are at the centre of the project. Here, solutions are to be implemented and presented which will enable the optimal use of the data generated in production. In OWL, BENTELER is developing one of these pilot factories in cooperation with the Fraunhofer IEM, which focuses in particular on the importance of data and effective data processing for predictive maintenance.
About the Leading-Edge Cluster
The project is integrated into the activities of the Leading-Edge Cluster "it´s OWL - Intelligent Technical Systems OstWestfalenLippe", in which 200 companies and research institutes cooperate. Projects open up new fields of technology and develop practical solutions for SMEs. The main topics are big data in production, machine learning, platforms and smart services as well as the working world of the future. An innovative transfer concept will make new technologies and methods available to small and medium-sized enterprises. In transfer projects, SMEs can use the approaches in cooperation with a research institute to solve concrete challenges in their operations. it´s OWL is considered to be one of the largest initiatives for Industrie 4.0 in small and medium-sized enterprises. With the support of the state of North Rhine-Westphalia, joint projects worth EUR 100 million are to be implemented by 2022. New research approaches and application examples will be presented by cluster universities and research institutes at the FMB supplier fair in Bad Salzuflen (Hall 21 A 27) from 6 to 8 November.