Host: The Japan Society of Mechanical Engineers
Name : Dynamics and Design Conference 2019
Date : August 27, 2019 - August 30, 2019
In the development the machine structure such as cars, we predict NVH(Noise, Vibration and Harshness) performance using CAE(Computer Aided Engineering）. This has resulted in that we can identify problems, realize early solutions for them and greatly shorten development period. However, in the market, the needs for reducing the frame weight has been increasing more and more due to CO2 emission regulation and energy saving, and vibration noise likely to occur on mechanical structure. Therefore, we need an analysis method to efficiently suggest countermeasure for ensuring NVH performance. We focus on Mutual Mean Compliance (MMC) as one of the methods to support understanding for vibration phenomena. MMC is a method that can specify the part to be modified by multiplying the transfer power and the transfer function in the whole system. MMC can propose effective structural change guidelines for the frequency of interest, but this method has a problem that there is a possibility of increasing the response of the adjacent peak due to the structural change with the focused frequency.
In this research, we developed a sensitivity calculation algorithm with a bandwidth in a frequency band containing two peaks. Then, we propose a structure change guideline that can reduce responses in the focused frequency band. When the threshold value was set high, it was confirmed that the response prediction was performed with high accuracy by using the proposed algorithm. However, when the threshold value was set low and the number of elements for changing the structure was increased, the accuracy deteriorated. That reason is to increase the amount of structural change. Since the sensitivity can be assumed to be linear when the amount of design change is small, it was concluded in this study that the analysis result and the response prediction match precisely within the linear range of sensitivity. We also concluded that the proposed algorithm can reduce the response in the bandwidth.
