DAWOPT research project - from data flow to knowledge-based production optimization.

In collaboration with the Fraunhofer Institute for Manufacturing Engineering and Automation (IPA) in Stuttgart, we have driven forward a joint project that is being funded by the InvestBW innovation program on behalf of the Baden-Württemberg Ministry of Economics, Labor and Tourism. The project is entitled DAWOPT, which stands for "From data flow to knowledge-based production optimization". Together, we want to redefine production optimization through a combination of data- and knowledge-based approaches. 

The overarching goal of our research cooperation is an automatic diagnosis of production deviations (anomaly detection) through machine learning, a recommendation for action provided by artificial intelligence for the user through root cause analysis and optimization measures with trained (externalized) expert knowledge. The task of our research project is therefore to develop a new approach for future applications of the value stream method, which provides the user with knowledge-based suggestions for targeted measures to optimize production on the basis of an automatically recorded data stream. Based on the established value stream method, a real-time-oriented value stream analysis is available. The further development is based on data connections to IIoT sources which, together with the implemented, newly developed key performance indicator system, make it possible to evaluate existing value creation processes of a value stream at different points in time. The users of such a digital application are faced with the operational task of recognizing where in the production system the next meaningful lever for improvement should be applied. Without proven lean expertise, most users lack the knowledge of which KPI deviation makes which improvement measure necessary. The lack of expert knowledge combined with the difficulty of simply deriving these correlations from the available data means that production employees often draw the wrong conclusions. We want to close this gap in a multi-stage hybrid approach (data- and knowledge-driven) using machine learning (ML), knowledge graphs and artificial intelligence (AI) methods. 

We have pooled our expertise since August 2023: In addition to iFAKT Research & Technology, the Factory Planning and Production Management departments at the Fraunhofer Institute for Manufacturing Engineering and Automation (IPA) in Stuttgart with the competencies "Autonomous Production Optimization", " WertstromDigital - die digitale Wertstromanalyse ", "Value Stream Design and Lean Production", as well as Cyber Cognitive Intelligence (CCI) with the competence "Dependable AI" are working together and want to draw as many insights as possible from the research project at the beginning of 2025, which we want to put into practice. 

The project is funded by the Invest BW program of the Ministry of Economics, Labor, and Tourism. 

Project partner:

"For me, we are taking the next step towards a digital value stream method together in the DAWOPT research project. It's no longer just about digitally capturing the value stream, but analyzing it with the help of AI. In this way, we can combine the identification of improvement opportunities with learned expert knowledge and make it more widely available." 

- Tim Teriete

M.Sc. M.Sc. Tim Teriete
Research Associate
Fraunhofer Institute for Manufacturing Engineering and Automation IPA
Factory Planning and Production Management