Virtual reality of interior noises of vehicles based on boundary element and neural networks

摘要: Interior noises of vehicles would be caused when the vibration body panels excites indoor air. In paper, load engine was obtained firstly through experiments. Secondly, applied in finite element model white to compute velocity and realize virtual reality, indicating that front support had large frequency lower than 60 Hz. The boundary then adopted interior noise extract sound pressure at a point near driver’s head. Two obvious peaks were shown curves, 270 310 Hz, respectively. impact on peak determined panel contribution analysis, remarkably reduced after applying absorption materials these panels. Nevertheless, many more additional not always better. If multilayer needed, an optimal value existed thickness material. And great played toward cabin by reasonable selection different their thicknesses. Finally, neutral network (NN) also used predict noises, which compared with result element. maximum difference between prediction values NN within 5 dB, neural feasible noise. Subsequently, method conduct optimization analysis for