Prediction of Powerplant Vibration Using FRF Data of FE Model

Recently, for the purposes of shortening the development period, the estimation of powerplant vibration has become more important in the early design stage, and eigenvalue analysis by FEM is commonly used to solve this problem. Eigenvalue analysis cannot directly predict vibration levels that affect the durability of each component and the vibration of a car body, however it is necessary to predict powerplant vibration in order to estimate exciting force under running conditions. Another factor adding to the difficulty of prediction is the instability of exciting force and various other non-linear characteristics. This paper presents a new approach using FRF data from FE models for accurate prediction of engine vibration under running conditions. By applying this approach to an in-line four cylinder engine, the predicted vibration is reasonably comparable with experimental results.