Initial situation

This project addresses the gap in systematically linking knowledge and insights gained during the usage and implementation phases of technical systems with the planning and development phases of new product generations. It aims to replace the currently experience-based process with a data-driven methodology by analyzing sensor and usage data using artificial intelligence (AI) and connecting them with graph-based models.

To achieve this, an innovative prototype will be developed to enable seamless integration between requirements management and sensor data. An interdisciplinary team combines the extensive expertise of the project partners. The consortium includes ConSenses GmbH, specializing in data processing and AI integration, and Conweaver GmbH, a leader in graph databases and data modeling. The team is further strengthened by the Department of Product Development and Machine Elements (pmd) at TU Darmstadt, which contributes expertise in Natural Language Processing (NLP) and AI-based data analysis. The team is supported by Werner Schmid GmbH (providing use cases and practical testing) and the associated partner Andritz Kaiser GmbH (offering expertise in mechanical engineering and production systems).

Project goals

1. Closing Digital Process Gaps:

• Enhancing requirements gathering and verification through hybrid models that combine measured data and physical models
• Reducing development time and costs through early adjustments to requirements

2. Sustainable Integration:

• Applying digital methods among project partners to optimize production processes and predict maintenance and wear conditions

3. Digital Transformation:

• Developing models and processes that enable more efficient and resource-saving approaches to product development

Procedure

The project is divided into several key phases. During the infrastructure setup phase, measurement and analysis systems will be established to collect initial data. This is followed by method development, where a methodology is created to link structured data with AI algorithms to interpret relationships between requirements and measurement results. In the implementation and validation phase, the developed methods will be tested in practice, and algorithms will be adjusted based on the results. Finally, practical recommendations will be derived in the conclusion and transfer phase, and the tools will be adapted for broader applications.

Funding

This project is funded by the Hessian State Chancellery – Minister for Digitalisation and Innovation.