Numerical analysis plays a significant role in the development of passenger comfort in modern vehicles. Customer perception of quality and comfort is largely influenced by its vibro-acoustic properties.

    To develop well engineered and robust vehicles that meet such customer requirements, many vibro-acoustic aspects and loading conditions need to be accounted for. This is only feasible when numerical analysis techniques, like the Finite Element Method, are applied over all design stages including the early concept phase.

    As vibro-acoustic behaviour is typically of global nature, meaning vibrations of different components mutually interact throughout the whole system, complex full-vehicle analyses accounting for all components as well as the interior air cavity are required. Moreover, with the rise of electric vehicle development, the rotational motor speeds and hence the excitation frequencies of interest increase dramatically compared to conventional combustion engine based vehicles. Such large-scale simulations challenge even today’s high-performance computing hardware and software to its limits. For this reason, efficient numerical reduction schemes of the underlying systems of equations are essential to avoid excessive evaluation times and costs, which becomes especially relevant when optimization approaches are considered.

    The presented paper illustrates the structure-borne acoustic analysis and optimization procedure of an electric full vehicle model subject to high frequency motor load cases. To reduce simulation times, the Frequency Response Function-Substructure reduction technique is applied to model components that are not modified in the successive optimization procedure enabling a very efficient reuse of precomputed data in each iteration cycle. As the focus in this work is on modifications in the powertrain, sub-systems like trimmed body and chassis are viable to be reduced.

    A new machine learning based optimization approach is applied that combines the Nastran based fluid-structure coupled analyses with self- learning algorithms. A detailed/high-level description of this new easy-to-use and highly automated approach is provided along with a discussion of the obtained results. Besides the optimized acoustic response and design state, these results consist of relevant design parameter rankings and tolerance analysis to reveal the robustness of the optimized design state.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Interior Noise Optimization of Powertrain Induced Vibrations for an Electric Vehicle Using Machine Learning Methods


    Additional title:

    Proceedings


    Contributors:

    Conference:

    International ATZ Conference on Vehicle Acoustics ; 2023 ; Zürich, Schweiz July 11, 2023 - July 12, 2023


    Published in:

    Publication date :

    2025-04-29


    Size :

    17 pages





    Type of media :

    Article/Chapter (Book)


    Type of material :

    Electronic Resource


    Language :

    German




    Optimization of Vehicle Interior Noise Performance Based on Real Powertrain Excitation

    Zhou, Jianwen / Yao, Lingyun / Zhou, Zhou | SAE Technical Papers | 2013


    Optimization of Vehicle Interior Noise Performance Based on Real Powertrain Excitation

    Zhou, J. / Yao, L. / Zhou, Z. et al. | British Library Conference Proceedings | 2013


    POWERTRAIN, NOISE SUPPRESSION METHOD, AND ELECTRIC VEHICLE

    TANG ZHENGYI / MEI XUEYU / EERHEMU BAYAER et al. | European Patent Office | 2024

    Free access

    POWERTRAIN, NOISE SUPPRESSION METHOD AND ELECTRIC VEHICLE

    TANG ZHENGYI / MEI XUEYU / EERHEMU BAYAER et al. | European Patent Office | 2023

    Free access

    Machine-Learning-Based Modelling of Electric Powertrain Noise Control Treatments

    Kocsis, Attila / Jacqmot, Jonathan | SAE Technical Papers | 2023