Automatic Normal Mode Identification Methodology for TBIW/Powertrain

<div class="section abstract"><div class="htmlview paragraph">Mode identification, particularly Modal Map Generation, is pivotal within the NVH (Noise, Vibration, and Harshness) domain for managing the performance of complex systems like TBIW/Powertrain. This study addresses the critical task of accurately identifying Global / Local behavior of a particular system as single entity (Complete TBIW, Power train) or all the systems attached to main structure (Sub Systems i.e Seat , Fuel Tank , Pump etc), which is crucial for effective NVH post-processing.</div><div class="htmlview paragraph">Introducing a novel tool/methodology developed by the Applus IDIADA team, this paper presents an efficient approach to Global & Local mode identification across subsystems, TBIW, and Powertrain levels. Leveraging ".op2" file content, mainly Strain Energy Density[<span class="xref">1</span>] and Displacement [<span class="xref">2</span>], the tool integrates Machine Learning Techniques [<span class="xref">3</span>] to produce mode predictions along with detailed visual outputs such as graphs , pie chart , modal charts etc. Implemented as a Python-based solution compatible with major Pre and Post processors, it operates seamlessly with cloud technology [<span class="xref">4</span>], thereby reducing prediction time significantly.</div><div class="htmlview paragraph">Beyond predicting mode numbers, the tool also provides actionable insights into subsystem contributions, aiding in enhancing mode shape and continuity studies [<span class="xref">5</span>]. Validated with robust data analysis, it ensures reliability in streamlined methodology for Mode Identification for NVH applications[<span class="xref">6</span>].</div></div>