基于非线性直接优化方法的发电机励磁系统参数辨识

励磁模型的准确程度对于电力系统仿真的准确性有显著影响。提出了一种新的、工程中容易实现的方法，即利用非线性直接优化法辨识微分方程形式的励磁模型参数。首先利用Matlab仿真与传统的最小二乘法比较验证算法有效性。然后利用动态信号记录装置，根据模型结构获得各个环节的时域内发电机阶跃试验数据，对1台300 MW汽轮发电机组的励磁系统分别进行分环节参数辨识和整体辨识。利用电力系统分析综合程序(power system analysis software package，PSASP)的用户自定义模型功能建立励磁模型进行相同的阶跃仿真，结果验证了辨识参数在电力系统仿真中的有效性。

Parameter Identification of Generator Excitation System Based on Nonlinear Direct Optimization Algorithm

Accurate extent of generator excitation system model greatly affects the accuracy of power system simulation results. The authors propose a new method that is easy to implement in engineering practice, i.e., to identify parameters of excitation model in the form of differential equation by means of nonlinear direct optimization algorithm. Firstly by use of simulation based on Matlab and traditional least square method the effectiveness of the algorithm is compared and verified; then by means of dynamic signal recoding device and according to the structure of the model the step testing data of generator during the time domain of each link is obtained, with these data the parameter identification by links and global parameter identification for excitation system of a certain 300MW turbogeneration unit are conducted respectively. By use of the function of user-defined model in power system analysis software package (PSASP), an excitation model is built and a same step simulation with the built model is carried out. Step simulation results validate the effectiveness of parameter identification in power system simulation.