07. September 2021

RISC Software: Avoiding Production Errors Using Data Analysis and Machine Learning

How can an overall view and a better understanding of manufacturing operations be achieved by recognizing causal connections and identifying manufacturing errors? RISC Software GmbH explored this question together with the two companies FILL Gesellschaft m. b. H. and Nemak GmbH on a three-year research project, “Boost 4.0”.

The European project “Big Data Value Spaces for COmpetitiveness of European COnnected Smart FacTories 4.0” (BOOST 4.0) focused on developing large-scale experiments with industrial data and demonstrating data-driven connected smart factories 4.0. Together with its pilot partners, RISC Software GmbH achieves a better understanding of the machines by identifying causal connections in the stored machine data. In order to process the collected data efficiently, the design and implementation of a scalable data analytics system for big data in industry were defined as the primary objective. This system enables further specific objectives to be determined such as the selection of appropriate data analysis algorithms for very large amounts of data and the potential for parallel implementation. Another objective was the establishment of an inspection system that uses machine learning to reduce the need for manual inspection of production errors in the manufacture of cylinder heads.

The Boost 4.0 project enabled RISC Software GmbH to deepen its knowledge of big data technologies and increase its existing expertise in data science and artificial intelligence. The business processes concentrated primarily on research and development, which led to further development of the existing data analytics framework. “Although no two projects are the same and there are always new challenges in the field of R&D, we seldom begin our work ‘from scratch’,” says Wolfgang Freiseisen, Managing Director of RISC Software GmbH. “Instead we generally base it on our frameworks, our knowledge. Through our involvement in Boost 4.0, RISC Software GmbH expanded AnnaLyze, its data analytics framework, by adapting it for future big data projects. At a meta-level, the exchange of knowledge and the communications and distribution activities helped us to add to our know-how and opened the door to further cooperative research activities with companies from other branches and countries.”

The principal tasks performed by RISC Software GmbH on the pilot project with the mechanical engineering firm of FILL Gesellschaft m.b.H. concentrated on selecting methods of machine learning and data analysis suitable for very large amounts of data and with potential for parallel implementation. Integrating the results of Boost 4.0 added an architecture concept[1] to the existing research infrastructure which combines big data technologies (such as Apache Spark on Hadoop) with semantic approaches (Apache Avro and SALAD). This facilitated the exploration and analysis of large amounts of data from heterogeneous sources (machine, product, process and logging data). High frequency time series were compared to identify specific use cases and increase understanding of the data. Interactive visualizations were implemented and integrated in a dashboard that supports communication with the experts. Alois Wiesinger, CTO of FILL Gesellschaft m.B.H., explains: “Fill is an established family-owned company that focuses on the success of its customers. The new approach of using digitalization to close the information gap between customers and engineering at Fill enables customer requirements to be met more effectively. By the end of the Boost 4.0 project increased customer satisfaction was already evident. […] With this newly created process transparency and the confidence we have built up with our partners and customers we are well equipped for the factory of the future!”

The pilot project carried out in conjunction with the firm of Nemak Linz GmbH and other partners examined how data analysis could be used to improve the quality assessment of cast parts. To this end, machine learning was deployed to automatically reassess cast parts that had previously been identified as flawed by x-ray computer tomography (XCT) using XCT images. The expected outcome is therefore an optimized quality assurance process and a reduction in the number of working hours spent on back-up inspections by XCT experts at Nemak Linz GmbH. Florian Reiterer, R&D engineer at Nemak Linz GmbH, explains: “The cooperation with RISC on the BOOST project was highly professional and a real pleasure. With its expertise in the field of data analysis, RISC Software GmbH was able to show us ways of optimizing resource-intensive quality assurance processes and consequently of cutting costs.”

The project “Big Data Value Spaces for COmpetitiveness of European COnnected Smart FacTories 4.0” received grants under funding agreement no. 78073 as part of the European Union’s Horizon 2020 research and innovation programme and was successfully concluded in 2020.

Project details

  • Project name: “Big Data Value Spaces for COmpetitiveness of European COnnected Smart FacTories 4.0” (Boost 4.0)
  • Duration: 01/2018–12/2020
  • Funding scheme: H2020-EU.2.1.1.
  • Project partners: 53 partners from 16 countries
  • Grant share: approx. EUR 20 million
  • CORDIS link
  • Project website


[1] Roxana-Maria Holom, Katharina Rafetseder, Stefanie Kritzinger, Harald Sehrschön, Metadata management in a big data infrastructure, Procedia Manufacturing, volume 42, 2020, pp. 375–382, ISSN 2351-9789, https://doi.org/10.1016/j.promfg.2020.02.060.


© RISC Software GmbH, reproduction free of charge