Although the data base of manufacturing companies is growing rapidly and predictive analytics is no longer a foreign concept in many companies, the actual data collected is usually still insufficiently exploited. Product usage data in particular holds enormous potential but has so far hardly been incorporated into the optimization of development and production. As a result, errors and anomalies in production and the use phase are not detected at an early stage - the consequence is reactive error management. Increasingly complex value chains exacerbate the situation, since internal company data alone is not sufficient for optimal decisions in defect management.
The goal of the BMWi-funded research project "value chAIn" is to develop a value chain-spanning defect management based on intelligent data-based decision support systems. This system is intended to enable manufacturing companies, in the project group focused on the commercial vehicle industry, to improve development and production processes and to optimize the performance and availability of their products.
The partners in the "value chAIn" project are developing an integrated knowledge and information base with data from product development, production and utilization. For intelligent error management, cause-effect relationships are first identified using AI methods. On this basis, new possibilities for process optimization, predictive quality and predictive maintenance can be developed. In order to make it easier for companies to translate the insights gained into concrete and targeted measures, the Fraunhofer IPT is finally developing decision support software on this basis.
The work of the Fraunhofer IPT focuses on the use of various AI methods with which optimization potential can be derived on the basis of the existing data. The cross-value chain database, which is built up both during production and in later use, is used for process optimization in production and also for predictive fault identification. In combination with formally stored expert knowledge, this creates a decision support system for implementing the proposed measures. AI approaches hold particularly high potential for the heterogeneous database consisting of data from different phases of the product lifecycle, as they help to identify complex dependencies and anomalies.
To illustrate the possibilities of cross-value chain defect management, the project partners will develop a demonstrator at the end of the project that will show the possibilities and potential.
The research project "value chAIn" is funded by the Federal Ministry for Economic Affairs and and Climate Action in the program "New Vehicle and System Technologies" based on a resolution of the German Bundestag and is supervised by the project management organization TÜV Rheinland Consulting GmbH.