Key technology in vehicle construction: SCALE researches on Artificial Intelligence

SCALE presents its activities on artificial intelligence and machine learning.

Artificial intelligence (AI) is becoming increasingly important as a key technology of the 21st century. With the German automotive industry, a large, economically strong AI scope is anchored in this country, which is on the verge of a major digital transformation. SCALE is also active in the field of Artificial Intelligence and Machine Learning (ML) and deals with questions concerning the application of AI methods in the automotive sector

Research projects


ViPriA: Since October 2019, SCALE has been coordinating the BMBF-funded research project on "Virtual Product Development using Intelligent Assistance Systems" (ViPriA for short). In cooperation with SIDACT and the Fraunhofer Institute SCAI, AI methods are being researched to make them useful in the CAE field. The aim of the project is to support calculation engineers in vehicle development with intelligent assistance systems. When setting up numerical simulations, AI approaches are to be used to provide information on meaningful modifications and result forecasts. Automated analyses of simulation results should, among other things, enable anomalies to be presented more efficiently.


ML@KaroProd: Since December 2018, SCALE has been involved in another BMBF project in the field of machine learning as a sub-area of AI. Together with Fraunhofer IWU and the Chair for Artificial Intelligence at Chemnitz University of Technology, research is being conducted into how ML approaches can be used to predict process parameters and component quality in car body production.


Master Theses

Besides participating in research projects on Artificial Intelligence, SCALE has supervised master theses in the field of Machine Learning on the following topics.


Erkennung von FEM-Geometrien mit Mehtoden des Maschinellen Lernens: The thesis presents a conceptual approach for the efficient processing of FEM data and an evaluative comparison of two deep learning approaches for geometry recognition of existing constructions.


Parameter estimation for spot weld design in automotive construction: The thesis is dedicated to the parameter estimation for spot welded constructions in automotive engineering.

ML-supported Design Methods for Determining Permissible Input Variables for Crash Simulation using the Example of FEM

Parameter prediction and automated generation of Car-body-connector-technologies in CAX-applications: The thesis is dedicated to basic research in order to automatically place welding spots on car body components and implement them in CAX applications.

Further publications


For the utilization of ML-based technologies for our SDM solution LoCo further publications have been created:



contact Heiner Müllerschön
  • Projekte
  • Software Sales