Intelligent compaction harmonic ratio indexes are more variable and their detection accuracy is not ideal. This paper summarizes the optimal wheel vibration signal processing method by comparing and analyzing a variety of noise-reduction methods such as Fourier transform denoising, wavelet thresholding method, and empirical modal decomposition (EMD). On this basis, by truncating the whole period of the signal after noise reduction, the error caused by Fourier transform leakage and fence effect is overcome, and the stability of the harmonic ratio index is improved. There are several key findings. The vibration signal of the roller wheel is approximately smooth, and the noise frequency and the effective signal frequency are obviously differentiated, and the Fourier transform denoising method can achieve excellent results. Wavelet thresholding does not have any obvious advantage over the Fourier transform method when dealing with the vibration signal of the roller wheel, and it is not flexible in choosing the frequency threshold of the retained signal. EMD is susceptible to the influence of the mode aliasing phenomenon, and there are certain defects in the stability of the waveform of the decomposed signal; in addition, the actual excitation frequency of the roller is not stable. Therefore, the Fourier transform of the intercepted signal can easily cause the leakage of the spectrum and the fence effect, which leads to the lack of calculation accuracy of the harmonic ratio index, and there is a high degree of variability. By truncating the whole period of the intercepted signal, the errors of the Fourier transform can be effectively overcome, the amplitude of the fundamental wave of the signal spectrum after truncation is significantly improved, and the frequency is closer to the real value. The introduction of the whole period of the truncation method in the vibration signal processing can reduce the volatility of the harmonic ratio index, and effectively alleviate the interference of its judgment of the quality of the compaction of the roadbed.
Tian et al. (Mon,) studied this question.