AI and ROM in Crash, Aero and Thermal Modeling

  • The agenda clearly puts the main emphasis on extremely interesting topics in CAE-assisted vehicle development so don’t miss out!
  • Listen to field reports on specific industrial applications and gain a deeper understanding of the subject matter. Altogether, lectures are at a high scientific level and will add tremendous value to your business.
  • The conference is your chance to network on an international scale, sharing views with leading simulation-experts from Europe, Asia and North America.
Automotive CAE | VDI Conference

The current debate is about unresolved issues in the field of Computer Aided Engineering, which will be addressed at the conference. Learn how the use of Artificial Intelligence (AI) and Reduced Order Modeling (ROM) can really push vehicle development in the automotive sector.

Get your simulation results even faster

Automobile manufacturing is characterised by diminishing development cycles and rising demands. In order to shorten development schedules and this way reduce costs, car manufacturers have long been relying on CAE methods. Now simulation methods like the Reduced Order Modeling plus the use of Artificial Intelligence in CAE are promising major time savings for developers. Then VDI Conference will discuss the use of AI and ROM, specifically in crash, aero and thermal simulation.

The conference will highlight current research findings and developments and show examples of industrial applications on the use of AI and ROM in computer-based vehicle development. It provides a comprehensive overview on the whole topic backed up by evidence from  enterprises and research facilities across Europe, North America and Asia.  

Top Topics

  • Real-time interactive aerodynamic and crash simulation
  • Integration of AI in CAE solvers
  • New approaches of AI and ROM in the area of vehicle safety
  • Applications for battery cooling

Download Program PDF

1st Conference Day, November 19, 2019

Registrierung / Registration


Eröffnung durch den fachlichen Leiter / Opening by the Chairman

Reduced Order Modeling (ROM) & Computer Aided Engineering (CAE)


Model Reduction and Engineered Artificial Intelligence (AI) for Augmenting CAE

  • Main MOR techniques for real-time simulation
  • Main AI-applications: visualizing, classifying, modeling, certifying, explaining, consuming
  • Application based on virtual, digital and hybrid twins

Prof. Dr. Francisco Chinesta, ESI Chair, ENSAM ParisTech, France, Co-Authors: Jean Louis Duval, ESI Group, Paris, France, Dr. André Berger, ESI Engineering System International GmbH, Berlin


From 3D Field Solution to Reduced Order Models

  • Classification of ROMs (linear/nonlinear, steady/transient, blackbox/system matrix, physics, scalar-output/field-output) and ROM-portfolio of ANSYS
  • Details of nonlinear transient dynamic ROM
  • Application of ROMs to automotive cooling problems

Dr. Valery Morgenthaler, Principal Software Developer, ANSYS, Lyon, France


Kaffeepause und Fachausstellung/Coffee Break and Exhibition

ROM & Crash Modeling


PGD for Crash Simulation

  • Reduced Order Modeling
  • Proper Generalized Decomposition
  • Optimization

Dr. Michael Andres, Forschung Crashmethoden, Co-Author: Dr. Lars Greve, Teamleiter Crashmethoden, both Volkswagen AG, Wolfsburg


Machine Learning (ML) based ROM for Real-Time Optimization – Applications for Crash and Safety

  • Introduction and short review of ROM (SVD, ML)
  • ML versus SVD (pros and cons)
  • Applications in safety and crash
  • Optimization and application in reliability and robustness

Dr. Kambiz Kayvantash, CTO, CAdLM, Antony, France

ROM & Aero / Thermal Modeling


Interactive Aerodynamic Design via Reduced Order Modeling and GPU-Computing

  • Integration of vehicle styling and aerodynamics
  • Model reduction via Proper Orthogonal Decomposition (POD) and interpolation
  • Leveraging Lattice-Boltzmann solver technology on GPUs for the computation of aerodynamic training data

Markus Mrosek, M.Sc., Project General Manager, Innovation Center Europe, Volkswagen AG, Wolfsburg, Co-Authors: Dr. Bastian Schnepf, Altair Engineering GmbH, Unterschleißheim, Dr. Carsten Othmer, Volkswagen AG, Wolfsburg, Prof. Dr.-Ing. Rolf Radespiel, TU Braunschweig


Mittagspause und Fachausstellung / Lunch and Exhibition


Development of Reduced Model for Aerodynamic Drag and Lift

  • Application of a reduced model based on POD to Computational Fluid Dynamics (CFD) simulation
  • Prediction of aerodynamic drag and lift of scale models of vehicle and the predicted results of using the reduced model correlated well to that of CFD
  • 3 groups of flow pattern
  • Improving accuracy: Application of POD by referencing standard deviation of pressure in wake of the scale models

Dr.-Ing. Tsuyoshi Yasuki, Project General Manager, Advanced CAE Division, Advanced Vehicle Engineering Development Div., Co-Authors: Mashio Taniguchi, Junichi Inokuchi, Yasuo Yamamae, Hiroshi Tanaka, all Toyota Motor Corporation, Toyota, Japan


The Proper Generalized Decomposition for Turbulent Flow-Control Problems

  • Model Order Reduction
  • Proper Generalized Decomposition
  • Parametrized flow-control

Vasileios Tsiolakis, M.Sc., CAE Methods, Innovation Center Europe, Volkswagen AG, Wolfsburg, Co-Authors: Prof. Dr. Ruben Sevilla, Swansea University, U.K., Dr. Carsten Othmer, Volkswagen AG, Wolfsburg, Prof. Dr. Antonio Huerta, Dr. Matteo Giacomini, both Universitat Politècnica de Catalunya, Barcelona, Spain


Kaffeepause und Fachausstellung / Coffee Break and Exhibition

Podiumsdiskussion / Panel Discussion: Die Zukunft von ROM in der Fahrzeugentwicklung / Future of MOR in Vehicle Development



Application of  Machine Learning Techniques through Simulation Data Management (SDM)

  • CAE-processes integrated with SDM
  • Intelligent support of simulation engineers with ML technologies
  • Data mining based on large scale structured simulation data
  • Introduction of examples from the automotive industry

Marko Thiele, Niederlassungsleiter Ingolstadt, Produktmanager SDM-System LoCo, SCALE GmbH, Ingolstadt


Ende des 1. Veranstaltungstages / End of the 1st Conference Day



Wir laden Sie herzlich ein, in entspannter Atmosphäre tiefergehende Gespräche mit anderen Teilnehmern und Referenten zu führen. / We kindly invite you to use the relaxed atmosphere for in-depth conversations with other participants and speakers.

2nd Conference Day, November 20, 2019



Applications of Artificial Intelligence in Automotive Engineering for Crashworthiness Design

  • AI
  • CAE
  • FEM
  • Long short term memory
  • Crashworthiness

Professor Dr.-Ing. Kaan Inal, Dept. of Mechanical and Mechatronics Engineering, University of Waterloo, Canada


Machine Learning unterstützte Identifikation von konstitutiven Parametern in der Werkstoffcharakterisierung

  • Neuronale Netze zur Unterstützung der Parameteridentifikation
  • Erzeugung von adäquaten Daten in großer Menge
  • Training eines künstlichen neuronalen Netzes
  • Evaluierung der Methode

Dr. David Koch, Materialcharakterisierung, Co-Autor: Dr. André Haufe, Leiter Prozess-Simulation, beide DYNAmore GmbH, Stuttgart

AI & Crash Modeling


Upscale: AI Enhanced Methods for Virtual Prediction of Short Circuit in Full Vehicle Crash Scenarios

  • Introduction to the UPSCALE/Crash project
  • AI for battery pack characterization
  • Virtual prediction of battery cell internal short circuit
  • Safety prediction for electric vehicle

Alexandre Dumon, Lightweight Assemblies, ESI Group, Paris, France, Co-Authors: Cristian Jiménez Castells, Applus IDIADA, Santa Oliva, Spain, Michael Andres, Volkswagen AG, Wolfsburg, Stefano Menegazzi, CRF, Trento, Italia, Prof. Dr. Francisco Chinesta, ENSAM ParisTech, France, Fatima Daim, ESI Group, Paris, France, Dr. techn., Dipl-Ing. Christoph Breitfuss, ViF, Graz, Austria


Machine Learning based Constitutive Modeling

  • Neural network based models describing metal plasticity
  • Neural network model: temperature and rate-dependent hardening of polymers
  • Neural network model predicting history effect on FLC
  • Development of robot-assisted uniaxial and multi-axial testing systems for sheet metal

Prof. Dr. Dirk Mohr, Dept. of Mechanical and Process Engineering, ETH Zürich, Switzerland


Kaffeepause und Fachausstellung / Coffee Break and Exhibition


Maschinelles Lernen von Crash-Ereignissen mit virtuellen Fahrzeugmodellen

  • Analyse von Beschleunigungs-Signalen mit Methoden des Maschinellen Lernens
  • Big Data mit Simulations-Modellen aus dem virtuellen Benchmark
  • Automatische Unterscheidung normaler Fahrmanöver von kritischen Ereignissen
  • Identifikation von Low-Speed-Crash-Ereignissen mit nicht-sichtbaren Struktur-Schäden
  • Entwicklung eines IoT-Devices mit Künstlicher Intelligenz

Dr. André Backes, Techn. Manager, TECOSIM GmbH, Rüsselsheim


Machine Learning for Pre- and Postprocessing in the Automotive Development Process

  • Representation of CAE meshes for ML
  • Subjective evaluation of crash simulation results
  • AI methods for engineering

Dr. Joachim Sprave, AI Research, Daimler AG, Sindelfingen

Necking Prediction using Neural Networks

  • Comprehensive virtual testing program using imperfection based model
  • Training of computationally efficient feed-forward neural network for prediction of localized instability
  • Validation on virtual and physical tests

Dr. Lars Greve, Teamleiter Crashmethoden, Co-Author: Bram P. van de Weg, Masterstudent Universität Twente, Enschede, Niederlande und Volkswagen AG, Wolfsburg


Mittagspause und Fachausstellung / Lunch and Exhibition


Combining AI and CAE in order to save Lives

  • Real-time trauma analysis system
  • CAE for training deep-learning algorithm which translates real-time crash pulse data into an organ-specific medical report for EMS and trauma physicians
  • Real-life feedback to calibrate dummy and human body models

Gilad Avrashi, Co-Founder and CTO, Co-Author: Dr. Itay Bengad, both MDGo, Haifa, Israel

AI & Aero / Thermal Modeling


A New Framework for 3D Deep-Learning based Surrogate Modeling and Optimization

  • Neural Concept Shape (NCS): only CAE-oriented Deep-Learning code able to process industrial scale raw unstructured 3D data without any preprocessing
  • Performance of NCS surrogate models on several benchmark tasks and real industrial problems, including fluid dynamics and heat exchanges
  • Optimization of a fixed wing drone, the eBee Classic: Improvement of the L/D ratio with physical constraints

Dr. Pierre Baque, CEO, Neural Concept, Switzerland, Co-Authors: Luca Zampierri, EPFL/Neural Concept, Lausanne, Switzerland, Prof. Pascal Fua, EPFL, Switzerland


UPSCALE: Machine Learning for Accelerated Aero-Thermal Design in the Age of Electromobility

  • Introduction to the UPSCALE/CFD project
  • The data-physics continuum: Looking at intersection of CAE and ML with a wide focus lens
  • Making data-driven methods more physical: Overview of methods for improving Machine Learning training efficiency
  • Using data to augment physics-based methods
  • Bringing it all togehter: an outline and status update on on the systems and procedures for utilizing the various methods in the context of a vehicle design framework.
  • Encoding expertise without knowledge: where is this all going and is it a good thing?

Dr. Eugene de Villiers, Managing Director, ENGYS LTD, London, U.K.


Schlussworte des fachlichen Leiters / Closing by the Chairman


Ende der Veranstaltung / End of the Conference

Book your ticket now

Would you like to share your views on the most fascinating topics in CAE with simulation-experts from all over the world? Would you like to know how AI and ROM have proven successful in industry applications at various enterprises and what are the developments that can be expected in future? Then it’s time to book your ticket for the VDI conference "Automotive CAE".

Eventnumber: 01KO509

The event is currently not available for booking.

We will inform you by email as soon as it is possible to book the event again.

By sending through your contact details, you won’t miss any further updates about our events.

In case of queries, you can reach us here

You can reach us at:


Become an exhibitor or sponsor

Request information


  • MSC Software GmbH
  • SCALE GmbH

Are you missing something?

Do you need more information about this event? Then please feel free to send us a message.

Write a message