Keynote presentations

Keynote1

J. Antoni, INSA Lyon.

Signal processing in mechanical engineering: a historical perspective and current challenges

Abstract:

If research in mechanical engineering has largely borrowed from signal processing, the reverse is also true. This keynote lecture aims at highlighting these interactions from historical and technical points of view. After a brief overview of the role of signal processing in modal analysis, SHM, condition monitoring, and NDT, the presentation will cover the main breakthroughs that signal processing has made possible in these fields, and eventually identify the remaining challenges. These will be recognized as being both scientific and educational in nature. A few suggestions, based on simple ideas, will be discussed to push the identified limits. Next, the presentation will dwell into the fascinating connection between signal processing and machine learning when it comes to applications in mechanical engineering. It will be shown that many signal processing techniques have structural similarities with CNNs, an observation that opens the way to explainable AI, but also to the design of more versatile signal processing tools. The keynote will extensively rely on the feedback of renown scholars in the domain.

Keynote2

C. Dindorf, Robert Bosch GmbH.

Shaping vibration engineering for the automotive transformation

Abstract:

In this presentation, we delve deep into the pivotal role of vibration engineering in the automotive industry's transformation, with a special focus on electrified powertrains, automated driving, and the virtualization of vibration engineering, including two EU-funded projects.

With the advent of fully electrified powertrains we are facing new challenges in automotive vibration. The internal combustion engine is no longer masking other sources within the vehicle that suddenly have to fulfill much stricter noise requirements. Also, automotive E/E equipment now has to reliably deal with new sources of vibration.

Further into the future, once we are being driven as passengers in fully automated shuttles the disturbance of the passengers by noise as well as vibration-induced failure of electric powertrains is of high importance. Testament to this, the EU-funded project “GAP-Noise” is dealing with these topics and is showcased here.

One keystone in this automotive transformation is the development of methods for the virtualization of vibration loads. This enables at a much earlier time in the development cycle of a component the need for a change of design at the fraction of a physical test which would be available at a much later stage. Approaches and solutions developed in the EU-funded project “VOLTCAR” and also beyond this project are being presented.