| Summary: | 碩士 === 國立臺灣海洋大學 === 系統工程暨造船學系 === 101 === In this study, a procedure for analysis of noise simulation of the tire tread pattern with the image-pixel technique has proposed. In this procedure, in first step, the image recognition technique is adopted to identify the two-dimensional shape of the tread pattern; and in the second step, the sound pressure parameter X1 in time domain is imposed on the two-dimensional matrix of image pixels for the tire tread pattern in the longitudinal direction of the tire, and then with the superposition approach in the width direction of the tire, the one-dimensional sound pressure in time domain is established, and in the final step, the Fast Fourier Transform method is used to calculate the narrow-band noise spectrum and the 1/3 octave-band noise spectrum.
To find the optimal sound pressure parameter X1 and weight parameter X2, in this study, the differential evolution algorithm (DE) mixed with gray relation analysis is adopted to minimize the difference between the noise simulation analysis and the measured data of the tire. To verify the rationality and feasibility of the proposed procedure of noise simulation of the tire tread pattern with the image-pixel technique through the six tread patterns of the same series. A tread pattern taken from these patterns is used to find the optimal sound pressure parameter X1 in time domain and weight parameter X2 with the above-mentioned optimization method. And then, the noise predictions of others tires, 1/3 octave-band noise spectrum, are conducted with the optimized parameters.
The noise results show that in the frequency range between 800Hz and 2000Hz, the average error of 1.1~4.6dB for the five predicted tires; in the five frequency bands, the average error of 1~1.4dB for each tire; and the error of the total sound pressure energy is 1~2.8dB. The proposed procedure, in this study, is considered as a feasible one with these noise results of the above-collected tires.
In this study, finally, the noise influences for the groove pattern of middle and shoulder of the tire are conducted, respectively. The results show that for the groove pattern of the tire’s shoulder, the noise performance in the high frequency domain, the short groove is better than the long one; for the groove pattern of the tire’s middle, the noise performance in the low frequency domain for the smaller tread area is better than others.
|
|---|
