AI/ML based Trimmed Body NTF, VTF & Global Modes prediction & Optimization using ODYSSEE Lunar

content image

Date and Time: 25th May, 2021

Duration: 60 Minutes

Trimmed Body typically consists of 100+ parts and each part’s thickness, material, cross section, connection with neighbouring parts (bolt / weld / adhesive) makes over 400+ parameters which needs to be considered for NVH analysis. Also for Noise Transfer Function (NTFs) & Vibration Transfer Function (VTFs) typically there are 20+ excitation points (60 load cases considering unit load in X,Y,Z direction) & 4 response points (driver / passenger ear microphone) which equals 240+ response curves to post process & analyse.

With such large amount of data getting generated & with tight project delivery timelines, with regular FE methods there is a limitation on number of modification iterations which can be done for optimizing the trimmed body to meet the NVH, cost & weight targets.

For such kind of problems with large data, AI/ML techniques are very useful. ODYSSEE is such AI/ML based innovate tool developed by CADLM, a France based company, which is now acquired & part of Hexagon MSC Software. 

ODYSSEE Lunar module uses the results of existing Trimmed Body FE analysis as input & builds Reduce Order Models (ROM) using machine learning algorithms. The ROM can predict the NTF/VTF/Global Modes results within seconds without having to run the FE solver. It also predicts the sensitive parameters at each frequency & can also do optimization within a matter of few minutes, where with regular FE solver it takes multiple days for such optimization. The results predicted by ROM model are validated by comparing with regular FE solver results before using them for real-time prediction/optimization. The correlation of ROM predicted results with FE results is found to be more than 90 % for Trimmed Body NTF/VTF applications.

What you’ll learn:

Overview of AI/ML techniques for CAE applications

Use of AI/ML techniques for trimmed body -

  • To identify sensitive parameters affecting Trimmed Body NTF, VTF & Global Modes
  • To reduce the time required for predicting NTF, VTF, Global Modes from many hours to few seconds
  • To reduce the time required for optimizing NTF, VTF, Global Modes from many days to few minutes
  • To reduce the number of iterations required for NTF, VTF, Global Modes optimization
  • To take better & faster engineering decisions




Speakers

Mr. Kedar Deo, Technical Manager, MSC Software

Kedar Deo holds master’s degree in Automotive Engineering from Coventry University & has experience of over 17 years in the field of NVH / Acoustic simulations. In the past he has worked with Mahindra & Mahindra Ltd and General Motors handling full vehicle NVH design & development of SUVs, LCVs & Passenger cars. He is associated with Hexagon-MSC Software from year 2013 & works with all major Auto /Aero OEMs and Tier-1 suppliers in India & ASEAN countries providing NVH / Acoustic solutions. He has undergone advanced trainings on acoustics at ISVR (UK) & Free Field Technologies (Belgium) & has also delivered more than 35 corporate trainings on Acoustics / NVH for major Indian OEMs in last 7 years.

Mr. Debashis Karmakar, Senior Manager- Engineering Lifecycle Management, MSC Software

Debashis Karmakar has two decades of experience in software development and solution implementation in the field of CAE, CFD, Simulation Data and Process Management (SPDM), Artificial Intelligence (AI) and Machine Leaning (ML). After doing M. Tech from IIT Kanpur, he joined Bhabha Atomic Research Centre for 5 years. Afterwards he joined MSC Software where he is working on new technologies and initiatives from over 15 years. He has deployed SPDM platform SimManager at multiple auto OEMs. Currently he is driving the Machine Learning application to CAE and Simulations.



grey box
Upcoming Webinar:

Annoyed by fan or AC noise? Improve the design of Fans, Blowers and Air Conditioners to reduce noise by Simulation

Date: Thursday, 04 June, 2020

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Duis fringilla enim sit amet justo ultricies varius. Phasellus eget erat rhoncus diam scelerisque auctor nec ac lectus. Vestibulum quis orci magna. Nunc at felis laoreet, porttitor velit a, laoreet velit. Fusce elementum semper massa in tempus. Aenean at dui nunc. Sed.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Duis fringilla enim sit amet justo ultricies varius. Phasellus eget erat rhoncus diam scelerisque auctor nec ac lectus. Vestibulum quis orci magna. Nunc at felis laoreet, porttitor velit a, laoreet velit. Fusce elementum semper massa in tempus. Aenean at dui nunc. Sed.

grey box

Lorem Ipsum

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Phasellus rutrum justo eu odio finibus, in lobortis ipsum consequat.

grey box

Lorem Ipsum

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Phasellus rutrum justo eu odio finibus, in lobortis ipsum consequat.