基于补偿二乘法的VSG惯性辨识问题及解决方法

虚拟同步发电机(VSG)的惯性是微网逆变器的重要参数,现有文献在对VSG惯性参数辨识过程中,为避免算法在递推过程中不收敛,需对输出数据进行滤波预处理。然而实验发现,输出滤波器截止频率会影响递推算法的最终辨识结果。针对上述问题,文章首先分析了滤波方式影响辨识结果的机理,然后提出对输入和输出数据等效滤波之后再进行递推运算的辨识策略,最后通过实验和仿真验证了理论分析的正确性及所提方法的有效性。

Problems and solutions of VSG inertial identification based on compensation least squares

The inertia of the virtual synchronous generator (VSG) is an important parameter of the microgrid inverter. In the process of identifying VSG inertial parameters of existing literature, in order to avoid the data divergence during the recursive process, the output data needs to be filtered preprocessing. However, experiments have found that the cut-off frequency of the output filter will affect the final identification result of the recursive algorithm. Aiming at the above problem, firstly this paper analyzes the mechanism of the filtering method affecting the identification results, and then proposes an identification strategy which performs recursive calculation after equivalent filtering of input and output data. Finally, the correctness of the theoretical analysis and the effectiveness of the proposed strategy are verified by the simulations and experiments.