动态测试误差的系统建模与混合谱分析法

本文拟通过动态测试数据处理来分离测量结果、系统误差和随机误差。首先采用动态数据系统建模法(DDS)作预处理与预分析,然后主要应用逐步回归法和周期图检验法即离散谱分析法,提取非周期与周期的确定性成分,并分离出测量结果与系统误差。再对提取后的数据,应用DDS建模法作随机误差的相关分析和连续谱分析,以分离其随机性成分。必要时可借非线性最小二乘法确定两者的组合模型,面对其参数作出精确估计。该方法的分辨率与估计精确度均有所提高,适用性很广,可全部借微计算机进行实时动态数据处理及修正测试误差。

ANALYSIS OF DYNAMIC MEASUREMENT ERROR THROGH SYSTEM MODELING AND ANALYSIS WITH MIXED SPECTRA

In this paper dynamic measurement data processing is used to separate the results of measurement, systematic error and the random error.First the method of dynamic data system modeling(DDS ) is applied for preprocessing and preanalysing dynamic measurement data Then various deterministic components are extracted by emphatically applying successive regression analysis and mobified periodogram test or discrete spectral analysis, so as to separate the resutls of measurement and the systematic error. For separating random components their correlation analysis and continuous spectral analysis of the extracted data are conducted by using the DDS mobeling mtheod When necessary it is possible to use the nonlinear least square method to determine better compound models so that accurate estimation can be made for various components mentioned above.The method presented in this paper has the advantage of having detter resolution, higher estimation accuracy and broader applicability. It is expected that by using this method dynamic measurement data prcoessing and systematic error correction can be made on line.