交通环境振动观测中本底振动去除的功率谱修正法

摘要

交通系统引起周边环境振动,其观测数据中往往含有本底成分.针对本底成分对真实振动的干扰问题,提出对观测功率谱修正的方法以去除本底振动.假设真实振动与本底振动为互不相关的随机过程,推导了功率谱修正法计算公式,并通过算例考查了方法的有效性.首先,取一条振动记录设定为真实振动曲线,与本底振动记录叠加合成振动数据以模拟现场观测记录.然后假设真实振动未知,分别采用振动级修正法、谱幅值修正法及功率谱修正法估计真实振动曲线.比较计算曲线与设定曲线,判断方法的可行性与准确性.算例显示,功率谱修正法计算的时程、功率谱与设定曲线符合良好,误差低于谱幅值修正法;功率谱修正法、振动级修正法计算的振动级与设定值基本一致.谱幅值修正法计算值略低于设定值.结果说明功率谱修正法克服了振级修正法不能计算时程和功率谱的不足,并且计算结果优于谱幅值修正法.%Background vibrations typically reside in field observation data recording traffic-induced environmental vibration in surrounding areas, and may disturb the true vibration signals excited by traffic transportations. A method of PSD revision was suggested here to remove the background vibration from field records observed. Based on the assumption that the true vibration and the background vibration were mutually uncorrelated stochastic processes, calculation formulae of the method were deduced. A group of data was then taken as an example to demonstrate the effectiveness of the method. First, a vibration curve was set as the true vibration data, and was superimposed on a background one to synthesize a filed observation record. Then, the true vibration was supposed to be unknown, and was estimated with the vibration level revising approach, the Fourier amplitude revising approach and the proposed PSD revising approach,respectively. The feasibility and accuracy of these approaches were examined by comparing the true vibration data with the estimated ones. The results showed that the time history and the PSD obtained with the PSD revising approach agree well with those of the true curves, and the errors are smaller than that of Fourier amplitude revising approach. The vibration level calculated with the PSD revising approach or the vibration level revising approach is almost the same as that of the true curres, but the calculated value with the Fourier amplitude revising approach is relatively smaller. It was concluded that the proposed approach behaves more accurately than the Fourier amplitude revising approach, and produces more types of true vibration data, i. e. , time histories and PSDs, than those obtained with the vibration level revising approach.