In automotive engineering, numerical simulations in computers have replaced complex and expensive real tests for many years. Crash tests are one example. The development process results in enormous amounts of simulation data, the evaluation of which poses a big challenge. The ViPrIA project – virtual product development using intelligent assistance systems – therefore aims to improve the complex data analysis. In the design process, engineers should be able to fall back on intelligent assistance systems that, for example, point out inconsistencies in the simulation data. This would simplify the work flow and accelerate the entire development process.

"Virtual product development often involves changes in design," says Prof. Dr. Jochen Garcke, head of the "Numerical Data-Based Prediction" department at the Fraunhofer Institute for Algorithms and Scientific Computing SCAI. "We are developing machine learning (ML) methods that can automatically detect anomalies in simulation results." Garcke and his team would like to prepare the simulation data in such a way that common ML methods can be used for evaluation. In addition to existing open source packages, Fraunhofer SCAI uses specially developed analysis algorithms that are adapted for the application domain.

The ViPrIA research project is coordinated by SCALE GmbH in Ingolstadt. Further project partners are SIDACT GmbH in Sankt Augustin, a spin-off company of Fraunhofer SCAI, and Fraunhofer SCAI. Associated industrial partners are the automobile companies AUDI, Porsche and Volkswagen.

The Federal Ministry of Education and Research (BMBF) is funding the project for three years, from 1 October 2019 to 30 September 2022.

Contact

Prof. Dr. Jochen Garcke
Head of Business Area Numerical Data-Driven Prediction
Fraunhofer Institute for Algorithms and Scientific Computing SCAI
Schloss Birlinghoven
53754 Sankt Augustin, Germany
Phone +49 2241 14-2286
jochen.garcke@scai.fraunhofer.de