Factor analysis of power seat noise using blocked force Ryohei Usui 1 Junji Yoshida 1 Osaka Institution of Technology 5-16-1 Omiya, Asahi-ku, Osaka-city, Osaka, 535-8585, Japan Yasuo Inose TS TECH CO.,LTD. 118-1 Ota, Takanezawa-machi, Shioya-gun, Tochigi, 329-1217, Japan

ABSTRACT In recent years, operating noise generated by various devices in cabin has been focused and becomes noise reduction target due to the reduction of vehicle running noise. In this study, we then conducted the noise evaluation and investigated the noise factors of a power seat at the operating condition. At first, the operating noise at around driver’s ear position and vibration of the seat frame were meas- ured using an artificial head microphones and acceleration sensors. As a result, the noise and vibra- tion at 400-500 Hz band was found to be large and the reduction was essential to decrease the noise. Subsequently, component transfer path analysis was applied to understand the main factor of the large vibration of the frame at the target frequency. The result showed the vibration was increased largely due to the resonance between the rotational order input force from the driving motor and the natural frequency of the frame. As an instance of the countermeasure, small weight was added to the resonated part of the frame to avoid the resonance and the vibration and noise was observed to decrease at the target frequency band. 1. INTRODUCTION

As automobiles have become quieter in recent years, operating noise generated by various devices in cabin has become more noticeable. In particular, power seats are often operated while the vehicle is parked, and the operating noise has a possibility to become a problem. The purpose of this study is to propose a guideline to effectively reduce the noise deteriorating the comfortability in cabin through the noise evaluation and the vibration analysis. In addition, component transfer path analysis [1-3] (CTPA) was applied as the factor analysis method and the contributions were analyzed in each direc- tional input force from the motor and vibration transfer characteristics of the frame to identify the main cause of the noise increase. 2. NOISE MEASUREMENT AND EVALUATION OF POWER SEAT

2.1. Experiment Outline At first, in order to understand which operational condition causes large noise, radiated noise of a power seat was recorded in each movement condition (front-back, up-down, and reclining condi- tions). Figure 1 shows the experimental environment and the power seat used in this experiment.

1 junji.yoshida@oit.ac.jp

worm 2022

(a) Experimental environment (b) Power seat Figure 1: Experimental environment and power seat employed in the experiment.

The noise recording was conducted for each movement condition. For the recording, an artificial head microphone (HEAD acoustics HMS2) was placed at the occupant’s head position as shown in Figure 1 (a). In addition, a weight was placed on the seat under the artificial head microphone to set the total weight on the seat as 60 kg to prepare similar condition in which a person sits on the seat. In this condition, the radiated noise in each operating condition was recorded for 15 s and repeated five times respectively.

2.2. Loudness of the Radiated Noise Figure 2 shows the averaged specific loudness of radiated noise recorded at left and right ear for five measurements in each operational condition.

(a) Front-back condition (b) Up-down condition (c) Reclining condition Figure 2: Averaged specific loudness between left and right ear in each operational condition.

Horizontal and vertical axes in Figure 2 are critical band [Bark] and the specific loudness [sone], respectively. As shown in this figure, the radiated noise was found to be loud at low-frequency band below 5 Bark (500 Hz) in the ascending, tilt forward, and tilt backward conditions. Accordingly, we focused on the radiated noise at the ascending condition, which was particularly loud, as the target operational condition for further analysis. 3. VIBRATION AND SOUND INCREASING POWER SEAT NOISE IN OPERATION

3.1. Experiment Outline Subsequently, we analyzed the physical factors (vibration and sound) increasing the specific loudness at the ascending condition. Figure 3 shows the measurement points for the vibration and sound in ascending condition. Six microphones (PCB Piezotronics 130E20) and acceleration sensors (PCB Piezotronics 352C22) were placed at lower right, upper right, upper, upper left, lower left, and motor points.

(a) Microphones attachment points (b) Acceleration sensors attachment points Figure 3: Microphones and acceleration sensors attachment position.

In the measurement, 60 kg weight was placed on the seat as same as the previous test and the vibration acceleration and sound pressure signals were measured for 15 s in total five times.

3.2. Vibration and Sound Pressure Level of Power Seat Figure 4 shows averaged vibration acceleration and sound pressure levels at the operational condition (ascending condition).

(a) Vibration acceleration levels of power seat (b) Sound pressure levels of power seat Figure 4: Vibration acceleration and sound pressure levels of power seat at ascending condition.

Horizontal and vertical axes show the frequency [Hz] and the vibration acceleration level (VAL) [dB], respectively. And horizontal and vertical axes in Figure 4 (b) shows the frequency [Hz] and the sound pressure level (SPL) [dB], respectively. As shown in Figure 4 (a), the VAL of seat frame was observed to be high at around 400-500 Hz band especially at upper left and right points. Moreover, SPL was also high at the same frequency band at almost all points as shown in Figure 4 (b). These results suggest that vibration and sound at the frame upper left and right points were the main factors increasing the specific loudness at 5 Bark. 4. CONTRIBUTION SEPARATION OF INPUTS IN EACH DIRECTION AND TRANS- FER SYSTEM

[Tower right —Upper right 0 200400 6008001000 Frequency [Hz]

4.1. Application of Component TPA Here, we introduced component TPA [1-3] and applied the method to obtain contribution to the target (response) point. Component TPA is a relatively new method of TPA [4, 5], in which the contribution is obtained by multiplying the transfer function with the blocked force, which is the independent input force of the passive part. The blocked force ( 𝐹 Blocked ) can be obtained by obtaining the transfer func- tion ( ℎ Blocked ) of the passive part when the active part is mounted and dividing the vibration acceler- ation of the passive part in actual operation ( 𝑎 Operation ) by the transfer function, using the following formula.

Tower right —Upper right ° 200 4006008001000 Frequency {Hz}

𝐹 Blocked = 𝑎 Operation /ℎ Blocked (1)

Furthermore, the contribution ( 𝐶𝑜𝑛𝑡 Blocked ) can be obtained by multiplying the obtained blocked force with the transfer function ( ℎ Cont ) from the input to the response point as follows.

𝐶𝑜𝑛𝑡 Blocked = 𝐹 Blocked × ℎ Cont (2)

This component TPA can evaluate the contribution to the response point separately from the motor input force in each direction to the target response point vibration without decoupling the input force (motor).

i i = [Toa coniibaon m0 400 «000 10m Frequeny (Ha)

4.2. Contribution to the Response Point Vibration of the Seat Frame In the contribution separation using component TPA, the contribution from motor input force along three directions (front-back, left-right, and up-down) to the seat frame vibration at the left and right upper points (response points) were calculated by using the estimated blocked force and the transfer function. Before evaluating the calculated contribution, the calculated contribution of the three direc- tions were summarized (total contribution) and compared with the actual measured vibration accel- eration at the response point to verify whether the contribution separation can be performed accu- rately. Figure 5 shows the comparison between the total contribution and the measured vibration at the upper left point of the seat frame, where the vibration and sound were particularly large. Noting that, the noise component in the transfer function was decreased in case the component ratio was under 5 % in the calculation of the blocked force.

Figure 5: VAL comparison between total contribution and measured vibration at upper left point

of the frame.

Horizontal and vertical axes in Figure 5 are the frequency [Hz] and the VAL [dB], respectively. As shown in the figure, the total contribution (blue curve) was very similar with the measured re- sponse point vibration (black curve). Accordingly, the separation accuracy is considered to be suffi- cient.

Figure 6 shows the calculated contribution to the frame upper left vibration from the motor input force of front-back, left-right, and up-down directions, respectively.

(a) Motor front-back

(b) Motor left-right

(c) Motor up-down

contribution

contribution

contribution Figure 6: Contribution to seat frame vibration at upper left point from the motor input force

at each direction.

The results showed that the up-down contribution was the highest among all contributions at the target frequency band (400-500 Hz). We then evaluated the up-down contribution into the input force (blocked force) and transfer function for understanding the main factor increasing the VAL as shown in Figure 7.

(a) Input force (b) Transfer function (c) Contribution Figure 7: Detailed separation of motor up-down contribution.

Horizontal axes in Figure 7 are frequency [Hz]. Vertical axes in Figure 7 (a), (b), and (c) are the force level [dB], transfer function [dB], and the VAL [dB], respectively. As shown in Figure 7 (a), a large peak was observed in the input force. The frequency of this large input force was 417 Hz and this was the 10th order input force of the motor rotational speed. This input peak was considered to be made by the employed motor characteristics. Furthermore, a peak was also observed in the transfer function from the motor to the upper left point of the seat frame in the similar frequency band. This indicates that resonance occurred between the input force and the transfer system and the VAL and the SPL were increased by the resonance. 5. UNDERSTANDING THE VIBRATION CHARACTERISTICS (MODE SHAPE) OF THE SEAT FRAME

5.1. Experiment Outline For the reduction of the seat frame vibration effectively, input force reduction or transfer function reduction is necessary. In this study, we focused on the frame transfer function because the motor is used in various produces in general. Hence, we investigated the characteristics of the seat frame to understand the main factor increasing the transfer function. At first, the point inertance of the upper left point of the seat frame was obtained and the mode shape increasing the transfer function at the frequency band was analyzed.

5.2. Point Inertance and Mode Shape Figure 8 shows the point inertance at upper left point of the seat frame.

Frege Ie}

Figure 8: Point inertance at upper left point of the seat frame.

Horizontal and vertical axes in Figure 8 are the frequency [Hz] and transfer function [dB], respec- tively. As shown in this figure, large inertance was observed at around 400 Hz and this illustrates the vibration mode existence in this frequency band. Then, vibration mode analysis was carried out and the obtained vibration mode at around the target frequency (416 Hz) was expressed as follows.

Upper left

Upper right

Upper right

Upper left

Upper left

Upper right

Gere a

Figure 9: Vibration mode of the seat frame at 416 Hz.

As shown in the upper Figures 8 and 9, eigen frequency and the mode were found at 416 Hz where large vibration and noise were observed in the operational condition. And the upper part twisted in the opposite phase in the vibration mode as shown in Figure 9. This mode is considered to increase the vibration of the upper left and right points at the operational condition.

Here, we summarize the factor analysis of the large radiated noise of the power seat in operation as follows. The loudness analysis revealed that the operating noise during seat ascent was particularly loud, and vibration and sound pressure level at upper left and right of the seat frame was the main factor increasing the noise at 400-500 Hz. Through component TPA, resonance between the motor input force along vertical direction and the transfer function to the upper left and right points of the seat was found to occur and increase the vibration at the point in operation. Finally, modal analysis showed that there was a torsion mode at the upper frame, and carrying out countermeasure to this mode is found to be essential for effective noise reduction for the power seat. 6. VERIFICATION OF THE MAIN FACTOR THROUGH SIMPLE COUNTERMEAS- URE INSTANCE

6.1. Verification Test Outline In this section, countermeasure experiment was conducted to confirm whether vibration and sound could be reduced by simple countermeasure (adding mass) to the indicated main factor (vibration mode at upper left and right points) of large radiated noise at 400-500 Hz band in ascending condition. Figure 10 shows the experimental environment and mass additional points.

Figure 10: Experimental environment and mass attachment points

for noise and vibration reduction at 400-500 Hz band.

For the countermeasure to the high contributing vibration mode at 400-500 Hz band, a 260 g weight was attached to the upper left and right points, where the amplitude was large in the mode at 416 Hz, and the vibration and sound in the ascending condition at around the seat frame were meas- ured again. Each measurement point and equipment were the same as in the previous experiments.

6.2. VAL, SPL and Loudness Comparison before and after Countermeasure Figures 11, 12 and 13 show the averaged VAL, SPL and specific loudness comparison with or without additional weight.

(a) Upper right point (b) Upper left point Figure 11: VAL comparison before and after countermeasure.

(a) Upper right point (b) Upper left point Figure 12: SPL comparison before and after countermeasure.

Figure 13: Specific loudness comparison before and after countermeasure.

Horizontal and vertical axes in Figure 11 are frequency [Hz] and VAL [dB], respectively. And Horizontal and vertical axes in Figure 12 show frequency [Hz] and SPL [dB], and horizontal and vertical axes in Figure 13 show critical band [Bark] and specific loudness [sone], respectively. In addition, red and black curves in all figures show the level before (without weight) and after (with weight) countermeasure. As shown in Figure 11 and Figure 12, both VAL and SPL at the target frequency of 400-500 Hz band were decreased significantly by adding the 260 g weight. Furthermore, the specific loudness at 5 Bark, where the loudness was maximum, could be decreased about 20 % by the countermeasure as shown in Figure 13.

From these results, it was clarified that the proposed analytical procedure actual could identify effective countermeasure areas for effective power seat noise reduction.

08 Original z, — Modified Poa & Yos =02 Zou 00 ° 5 10 20 Critical band [Bark]

7. SUMMARY

In this study, we conducted experiments, evaluations, and analysis to propose a guideline for effec- tively reducing the power seat noise. At first, in order to determine which movement condition causes the large noise, we evaluated the loudness of the radiated noise in each movement condition. The result showed the loudness at around 400-500 Hz in ascending condition was especially large. Sub- sequently, in order to understand the physical factors increasing the specific loudness during the as- cending condition, the vibration and sound pressure at around the seat frame were evaluated, and it was confirmed that the vibration and sound at upper left and right were the factors. In addition, com- ponent TPA informed the resonance between the motor input force along vertical direction and the transfer function from the motor to the seat upper frame increased the noise and vibration. And vi- bration mode analysis revealed that upper frame twisted mode increased the transfer function and the mode was found to be important for the effective noise reduction. Finally, a simple countermeasure experiment was conducted for the verification by adding mass at the high amplitude points of the frame (upper left and right points), and the result showed the effective noise and vibration reduction. From these results, it was clarified that the proposed analytical procedure actual could identify effec- tive countermeasure areas for effective power seat noise reduction. 8. REFERENCES

1. D. de Klerk & DJ. Rixen, Component transfer path analysis method with compensation for test

bench dynamics, Mechanical Systems and Signal Processing , 24(6) , 1693-1710, (2010). 2. A. T. Moorhouse, A. S. Elliott & T. A. Evans, In situ measurement of the blocked force of struc-

ture-borne sound sources, Journal of Sound and Vibration , 325(4-5) , 679-685, (2009). 3. Maarten V. van der Seijs, Dennis de Klerk & Daniel J. Rixen, General framework for transfer

path analysis: History, theory and classification of techniques, Mechanical Systems and Signal Processing , 68-69 , 217-244, (2016). 4. Noumura, K. & Yoshida, J., Method of Transfer Path Analysis for Vehicle Interior Sound with

No Excitation Experiment, F2006D183 FISITA2006, pp. 1-10, Yokohama, Japan, October2006. 5. W. Bermayer, F. Brandl, R. Hoeldrich, A. Sontacci, S. Brandl & W. Fliesser, Sound Engineering

based on Source Contributions and Transfer Path Results, JSAE Paper , Vol. 318, No. 20075399, pp. 15-17, (2007).