基于最小二乘法的热试验温度数据预测  

Prediction of temperature in thermal test based on least square method

作　　者：李西园[1] 龚喆[1] 马昆[1] 曹志松[1] 张景川[1] 

机构地区：[1]北京卫星环境工程研究所,北京100094

出　　处：《航天器环境工程》2013年第2期146-150,共5页Spacecraft Environment Engineering

摘　　要：航天器真空热试验中,温度数据的正确判读及预测十分困难,又十分重要。文章针对典型的太阳电池板试件建立了热模型,使用热分析中常见的有限差分法建立离散方程,通过最小二乘法拟合其系数,以预测下一个测控周期各个测量点的温度,并提出了滑动窗口修正的方法。基于某型号电池板试验数据对热模型的预测精度进行了验证和误差分析,最后对通过测点温度逆向预测加热功率的应用进行了讨论。The data validation and prediction are diffcult yet important issues in the thermal vacuum test. In this paper, a typical solar panel thermal model is build, and the discrete equation is derived by the finite difference method. The coefficients are fitted by the least square method with ＂slip window＂ correction, to predict the temperatures of measuring points in the next measurement and control period. Correction slip window. Based on the test data of a certain solar panel, the prediction capability of this model and the related error analysis are made. In the end. the method to nrediet tha hoatincr nnxxr^r thrrmrvh ＋h~ ＋ h ~ ：-.- ：_ J：

关 键 词：真空热试验 最小二乘法 温度测量 数据预测 

分 类 号：V416.6[航空宇航科学与技术—航空宇航推进理论与工程] TP206.1[自动化与计算机技术—检测技术与自动化装置]