基于混合算法和PMU数据的发电机励磁参数辨识方法研究

Excitation systems' parameters identification based on hybrid algorithm with PMU's data

励磁参数辨识是电力系统稳定计算校核的基础。研究利用WAMS数据在线辨识励磁系统参数的方法。首先,将励磁系统的在线参数辨识归结为一个输入输出系统的最优化问题。即输入PMU数据是发电机的PMU出口电压,输出PMU数据是发电机的励磁电压和励磁电流。目标函数是在同一输入下,模型的输出和实际测量输出的差异最小。在实现上,采用数据点的差异平方和作为目标函数。上述最优化问题为包含微分方程的非线性问题,所以采用基于GA方法和最优梯度搜索相结合的方法求解,从而获得辨识参数。最后,基于实测PMU数据和BPA-FV型的仿真,验证了该方法的有效性。

Parameter identification for excitation systems（ES） is the foundation of the power system stability analysis.The paper on the identifying the parameters of the excitation system with the field data obtained with PMU/WAMS.Firstly,the on-line parameters identification of ES is formulated as an optimization problem of an input-output system.In detail,the input is the generator＇s terminal voltage and the output is generator＇s excitation voltage/current,both are obtained with PMU.The objective function is to minimize the different of the output and virtual output errors,responding to the same input.In computation,the objective function is the square sum of the differences at sampling time.The above optimization problem is nonlinear as it holds the differential equations,therefore the hybrid algorithm,which is the combination of the genetic algorithm（GA） and gradient-based searching,is taken to obtain the parameters.Finally,simulation results in a BPA-FV type ES,with the field PMU data,show the effectiveness of the proposed approach.