振动试验中削波信号功率谱密度补偿  

Power Spectral Density Compensation Algorithm for Signal Clipping in Vibration Test

作　　者：严鲁涛[1] 杨志鹏[1] 高飞[1] 刘洁[1] 

机构地区：[1]北京强度环境研究所,北京100076

出　　处：《振动．测试与诊断》2015年第4期690-696,798,共7页Journal of Vibration,Measurement & Diagnosis

基　　金：民用航天资助项目

摘　　要：振动试验过程中削波会引起信号的功率谱密度下降,局部频率点对应的功率谱密度误差过大将导致试验结果的可信度降低,尤其对于模态试验,输入信号削波可能导致信号谱型局部"下凹",甚至产生错误的试验结果。从偏斜度、峭度、概率分布和功率谱密度等方面分析了高斯信号及非高斯信号的特性,介绍了功率谱密度补偿的频谱均衡及比例-积分-微分(proportion-integration-differentiation,简称PID)两种算法,并对比了这两种算法的补偿效果。结果表明:对于高斯信号,两种算法从迭代次数及最小误差对比区别不大,且均能满足均衡要求;对于非高斯信号,PID算法实现较小误差的同时,迭代次数少,具有一定优势。In environmental shaker testing applications,sigma clipping of the shaker drive signal is used to protect the test system.However,the clipped signal spectrum will no longer correspond exactly to the given power spectral density(PSD).This may cause reduced vibration test reliability and even the wrong results,especially for modal tests.Both the power spectrum equalization control algorithm and PID control strategy are presented in order to compensate for the difference between certain spectra and the clipped specification.The results show that,in the case of Gaussian random signals,the two methods show almost the same compensation effect in terms of minimum error and iterative steps.For non-Gaussian random signals,however,the PID control strategy obtained fewer iterative steps and minor errors.

关 键 词：削波 高斯信号 非高斯信号 频谱均衡 比例-微分-积分控制 

分 类 号：TB535[理学—物理] O324[理学—声学]