【摘要】 针对国际全球卫星导航系统服务组织(IGS)站高程非线性速度场建模时周期初值人为假定存在不稳定的问题,提出以功率谱分析为手段来探测高程时间序列主要周期值的方法,并将探测得到的主要周期值作为初值建立非线性速度场模型:以国内外8个IGS站20余年的高程时间序列数据为对象,利用功率谱分析的方法得到每个IGS站高程时间序列数据的主周期值;然后作为非线性速度场模型的周期初值进行迭代求解;最后与非线性模型线性化人为给定初值的方法进行对比。结果表明,当迭代求解得到的模型拟合残差精度相等时,功率谱分析给定周期初值的方法比人为给定固定周期初值的方法在迭代次数、迭代时间上均可得到降低,且能实现IGS站高程时间序列数据的有效拟合。更多还原

【Abstract】 Aiming at the problem of artificial instability in the determination of initial periodic values of nonlinear velocity field modeling for IGS station elevation, the paper proposed a method to detect the main periodic values of elevation time series by means of power spectrum analysis, and established the nonlinear velocity field model by taking the main periodic values obtained from the detection as the initial values: the elevation time series data of more than 20 years from 8 IGS stations at home and abroad were taken as the objects, and the main period values of the elevation time series data of each IGS station were obtained by using the power spectrum analysis method; then the iterative solution was conducted with the initial periodic values of the nonlinear velocity field model; finally the method of linearizing artificial initial value of nonlinear model was compared with the proposed method. Results showed that: the method of power spectrum analysis with given initial value of period could be reduced both in number of iteration and in iteration time compared with the method with fixed initial value when the accuracy of model fitting residuals obtained by iterative solution was equal; moreover, the proposed method could implement the effective fitting of time series data for the IGS station elevations.更多还原