Uncertainty calculation applied to different regression methods in the quasi-dynamic collector test
Uncertainty calculation applied to different regression methods in the quasi-dynamic collector test
复制标题
DOI：
10.1016/j.solener.2006.03.010
复制
发表时间：
2006-11
期刊：
影响因子：
6.7
通讯作者：
M. Kratzenberg;H. Beyer;S. Colle
中科院分区：
文献类型：
--
作者：
M. Kratzenberg;H. Beyer;S. Colle
鉴于测试获得的集热器模型参数的不相关回归不确定性，本文根据EN12975讨论和评估了太阳能集热器的准动态测试程序。在本讨论中，分析了回归系数识别的两个程序的结果，即最小二乘（LS）和加权最小二乘（WLS）回归方法。计算并验证了 LS 和 WLS 结果的不确定性。两种方法的相互比较显示了可比较的归一化效率曲线，但 WLS 方法在不确定性方面表现更佳。
This paper discusses and evaluates the quasi-dynamic test procedure for solar thermal collectors according to EN12975 in view of the uncorrelated regression uncertainty of the collector model parameters obtained from the test. In this discussion the results of two procedures for the regression coefficients identification, the least square (LS) and the weighted least square (WLS) regression methods are analyzed. The uncertainties of the both, the LS and WLS results are calculated and validated. The inter-comparison of the two methods shows comparable normalized efficiency curves, but superior performance of the WLS method with respect to the uncertainties.