DatProForge data-driven process modelling of drop forging processes to increase productivity using adaptive tool design methodology

• Investigation of measurement data changes in the continuous forging process
• Development of an innovative data-driven approach: combination of traditional sensor technology and high-resolution radar sensor technology in the forging process
• Data fusion of heterogeneous sensor data, modelling, feature generation and selection: Detection and extraction of process-relevant sensor data to detect process instabilities (e.g. tool wear)
• Derivation of design guidelines to generate more process-resilient mould surfaces

The aim of the research project is to develop a basic understanding of the interaction between changes in measurement data (pattern recognition, as a result of specifically altered process conditions) and the active surface design in die forging (e.g. helix angle or die pitch) on product quality in continuous operation. With the aid of numerical process simulation, relevant sensors are selected and integrated locally in the forming system in a targeted manner. Furthermore, the existing domain knowledge is used to develop specific features of the process measurement data with which the type of process fault can be recognised. This includes in-depth analyses of the press kinematics as well as the elastic and thermal deformations that occur during forging, and measures are proposed to improve the quality and accuracy of the forgings.

The expected result of the research project is a data-driven methodology that makes it possible to make the active surface design of forging tools more process-resilient using process measurement data and AI methods. For the first time, high-resolution radar sensors are being integrated with other process measurement technology in a smart sensor network. Feature-based detection of process disturbance variables and the AI-driven derivation of design guidelines for active tool surfaces will increase process resilience during forging.