The variation trend of telemetry data of satellite in orbit can reflect the status and changes of the satellite direct-ly,according to which,the performance and trend of the key devices of satellite in orbit can be predicted. On the basis of the characteristics of telemetry data,the data decomposition algorithm based on X-11 is used to decompose the selected variable. The polynomial fitting,nonparametric regression,ARMA model,BP neural network and other methods are adopted to forecast the decomposed data,analyze the processes and accuracy of the methods,and evaluate the attenuation factor. The prediction ex-periment was performed for the temperature parameter of a certain satellite in orbit. The experimental results prove that the average relative error of the proposed prediction method is less than 8%,and the method can predict the performance trend of telemetry data of the satellite in orbit effectively. It provides a technical assurance for the application services such as state monitoring of the satellite in orbit,health management and fault analysis,and has an important practical value.%在轨卫星遥测数据的趋势变化能够直接体现卫星的状态和变化,根据遥测数据的变化可以对在轨卫星关键器件的性能和趋势进行预测.首先根据遥测数据特征采用基于X-11的数据分解算法,对选取变量进行分解,再利用多项式拟合、非参数回归、ARMA模型、BP神经网络等方法对分解后的数据进行预测分析,并且对方法的流程和精度进行分析,最后评估衰减因子.针对某在轨卫星温度参数的预测实验,结果证明,提出的预测方法平均相对误差小于8%,能有效地对在轨卫星遥测数据的性能趋势进行预测,为在轨卫星状态监控、健康管理与故障分析等应用服务提供技术保证,具有极其重要的实用价值.
展开▼
