背景参数的不确定性对参数识别的影响

Influence of Background Parameter Uncertainty on Parameter Identification

识别电力系统参数时,通常假定除极为有限的几个目标参数外,全部非目标参数(下称背景参数)都为准确值。但是,事实却是系统模型和工况等都存在很强的随机性和不确定性,许多背景参数用的都是典型值或经验值,甚至连合理的误差区间都难以确定。背景参数的误差必然会影响目标参数的识别误差,甚至使问题不可识别。采用基于系统稳定机理的受扰轨迹差异度,通过灵敏度分析识别目标参数,考察背景参数不确定性对参数识别的影响。仿真表明,背景参数的误差可能导致严重的识别误差,甚至无解,从而警示了参数识别中的巨大风险。

Except for two or three target parameters,all the remaining parameters（hereinafter referred to as background parameters） are usually assumed to be accurate in parameter identification.However,both models and operating conditions are highly random and uncertain in nature,typical or experiential values have to be used for the background parameters.It is also difficult to assign reasonable value-intervals for some background parameters.Errors of the background parameters will affect the identification errors of the target parameters,or even invalidate the identification.Based on stability mechanisms,the influence of background parameter uncertainty on parameter identification by using the difference index and its sensitivity analysis are investigated.Simulation results show that background parameter error may lead to serious error for identification or even no solution.This reveals the potentially very high risks for parameter identification.