Study on Acceleration Noise Reduction Methods by Suppressing Vibration of Cantilever-Supported Heat Shield

Yu Yonai (UD Trucks)

This study examined cases in which cantilever-supported heat shield serves as major sources of acceleration noise, resulting in the deterioration of overall noise performance. Modal analysis confirmed that the resonant frequency of the relevant component corresponded with the frequency range responsible for the worsening of acceleration noise. By introducing a sandwich structure to the steel plate component and imparting dynamic vibration absorption performance, it was possible to reduce vibration amplitude even though the resonant frequency itself remained unchanged, thereby successfully suppressing the component's impact as a noise source.

Robust Design and Mechanism Elucidation for Three-Cylinder Engine Idling Vibration Using Machine Learning in Multi-Dimensional Design Space Analysis

Kazuyuki Yamamoto･Victor Picheny･Qi Qi (Secondmind)

Using a machine learning prediction model, we derived robust design specifications that satisfy the idling vibration target of a three-cylinder engine, taking into account the manufacturing variations of key components. Furthermore, by analyzing the correlations between design variables in the visualized multi-dimensional design space, we obtained new engineering insights that contribute to establishing new design guidelines for the three-cylinder engine's idling vibration.

Efficient NVH Integration of EDUs Using Hybrid Dynamic Sub Structuring Approaches

Noriyuki Muramatsu (FEV Japan)･Christoph Steffens (FEV Europe)･Ahmed El-Mahmoudi･Christopher Lechner (FEV Vehicle)

In electric drive unit (EDU) development and vehicle integration, early decisions balancing NVH performance and cost are critical. Assessing NVH behavior in the vehicle context before physical tests is challenging. Using dynamic sub structuring, virtual EDU models can be integrated into measured benchmark vehicles for early acoustic evaluation. Experimental models from databases, including vehicle and mounting concepts, provide essential input for informed design choices. This hybrid approach enables optimization without over-engineering. The paper presents methodological foundations, challenges, and practical applications combining simulation and measurement data, supporting efficient and cost-effective NVH development for electric drives.