论文部分内容阅读
上述实例及其它算例表明,试验数据与理论值偏差1％时,所得估值参数与理论参数的最大相对误差不大于4％,因此,Levy提出的系统辨识方法是有实用价值的。 二、试验频率点数和频率范围对辨识精度影响 Levy法不是无偏估计,增加试验点数,不一定能提高估值精度,还会使系数阵X行数增加,易于导致病态矩阵和试验数据的弱相关,使辨识精度下降,甚至估值参数不可信。
The above examples and other examples show that the maximum relative error between the estimated parameters and the theoretical parameters is less than 4% when the experimental data deviates from the theoretical value by 1%. Therefore, the system identification method proposed by Levy is of practical value. Second, the test frequency points and frequency range of the recognition accuracy Levy method is not unbiased estimation, increasing the number of test points, may not be able to improve the valuation accuracy, but also make the coefficient matrix X lines increased, easily lead to the weakness of the pathological matrix and experimental data Related, so that the identification accuracy decreased, and even the valuation parameters are not credible.