【正文】 data of the experiment and the calculation is identified in this domain, the power plant modeling and the exciting force can be considered reasonable.However, around 12 Hz of 1storders, data is not much identified. In this frequency domain, the vibration of the engine and the vehicle body are mutually coupled through the engine mount. Therefore, the accuracy of the vehicle body model has a damaging effect.IMPROVEMENT OF THE MEASURING ACCURACY IN LOWFREQUENCY VIBRATIONThe engine exciting force was determined using Souma’s method, and the vibration in each part of the engine was calculated by adding the exciting force. So far, however, the calculated data has not been much identified with the actual measurement. Therefore, the accuracy of the actual measurement is improved. In the surface vibration of the engine, the lowfrequency vibration, which causes the idling vibration, and the highfrequency vibration, which causes noise, are mixed. When the mixed vibration is measured with a piezo element acceleration pickup, the highfrequency order is emphasized and the target lowfrequency order bees relatively small. For example, the measured acceleration to time waveform for the vertical vibration in the right engine foot is shown in Figure 4. In this paper, a strain gage acceleration pickup, which measures force acting on the inner weight by strain, is used. This device, which is larger than a piezo element acceleration pickup, is more sensitive to the acceleration. Besides, silicon oil is filled inside to protect the detecting parts in this device, which mechanically blocks off the highfrequency order. The measured acceleration to time waveform for the vertical vibration with the device is shown in Figure 5. Compared with Figure 4, Figure 5 shows only the lowfrequency order although the same area was measured. In this way, the highfrequency order is blocked off, which results in the higher sensitivity with the device. This time, the device, which measures the acceleration ranging from 0 to 20m/s2,was used. This device is easily calibrated using Gforces because it has the higher sensitivity. When a piezo element acceleration pickup was used, the differences between the calculation and the experiment were 2040% in the main order of the vibration, and a few times in other orders. Therefore, the principle of Souma’s method using a piezo element acceleration pickup has been in doubt. However, the data of the experiment and the calculation has been identified as shown in Figure 2 and 3 since a strain gage acceleration pickup, which has been used in the experiment of movement performance, was used for an engine.Fig. 1 Seat rail vertical vibration Fig. 2 Head cover lateral vibrationFig. 3 Right engine foot vertical vibration Measurement with piezo element acceleration pickupENTIRE VEHICLE MODELFigure 6 shows the body model. Interior and exterior equipments such as doors and seat are added in the form of 85 mass points to the main structure modeling detailed with sheet metal finite elements. The grid points are 61,912. Figure 7 shows the model where a frame, a suspension, and an engine are bined, and a fuel tank and a bumper is added in the form of concentrated mass. The grid points are 39,262.Combining the models shown in Figure 6 and 7 using cabmount makes the entire vehicle model. Total grid points mounts to 101,174. The calculation time is 3,293 seconds using IBMSP2, MSC/NASTRAN Version . The calculating method is package calculation. If the model bees on larger scale, the model must be calculated by the block structure.Figure 8 shows the frequency response function, indicating the responses of the frame with the right back engine mount after exciting the driver’s seat rail. In the frequency ranging from 20 to 30 Hz, which is required for the analysis, the data of the experiment is qualitatively identified with that of the calculation.Fig. 5 Measurement with strain gage acceleration pickup Fig. 6 Body mode Frame,power plant and suspension model Frequency response functionCORRELATION ANALYSIS OF THE MODESFrom the viewpoint of vibration characteristics, it can be considered that an entire vehicle is insulated by the engine mount and the cabmount, which have relatively small spring constants, although the insulation is not plete. When the entire vehicle is divided into block structures by each insulating mount and suspension, the body has 4 block structures:(1) Block where interior equipment is added in the form of concentrated mass to the body as shown in Figure 6, which is described as “body”, hereafter.(2) Block where the fuel tank and the bumper are added in the from of concentrated mass to the frame as shown in Figure 7, which is described as “frame,” hereafter. (3) Power plant (4) SuspensionAmong the above block structures, (1) body and (2) frame have the natural frequency around 24 Hz in the idling vibration. The vibration characteristics for the body, the frame and the entire vehicle model are pared and investigated.COMPARISON OF NATURAL FREQUENCYFigure 9 shows the distribution of the natural vibration frequency in each block structure and in the vehicle condition. The frame has 17 natural modes below 50Hz. In Figure 7, the model mounting a power plant and a suspension on the frame, is called Y chassis, which has 35 natural modes below 50 Hz. Y chassis makes the entire vehicle model by mounting the body, which has 94 natural modes below 50 Hz.When the number of natural modes of Y chassis is added to 61 natural modes of the body, total number of the modes amounts to 96. The number of the natural modes of the entire vehicle model (94) is less than the above total number by 2 modes. This is because 2 natural modes became above 50 Hz by bining Y chassis and the body, as the result of analyzing the mode correlation described later.Fig. 9 Natural modes in frequency domain附錄B具有車架結(jié)構(gòu)車