基于分块Hankel矩阵的抽水蓄能电机扩展卡尔曼滤波模型子空间循环辨识

Sub Space Cycle Identification of Pumped Storage Motor Extended Calman Filter Model Based on Block Hankel Matrix

为提高抽水蓄能电机模型辨识结果的精度,提出一种基于分块Hankel矩阵的抽水蓄能电机扩展卡尔曼滤波(extended Calman filter,EKF)模型子空间循环辨识方法。首先,给出抽水蓄能电机磁场方程,建立基于EKF滤波器抽水蓄能电机模型,并利用电阻和电感的两组状态向量构建两组辨识模型,两组模型协同工作构建循环辨识算法;然后,引入子空间辨识策略,利用分块Hankel矩阵建立相关数据估算方程,利用零空间投影进行输入项删除,降低辨识复杂度,采用最小二乘实现对系统参数矩阵的有效辨识;通过在Matlab平台和硬件环境下的仿真实验,验证了所提算法在抽水蓄能电机转速、转子电阻、励磁电感等参数辨识上的有效性。

In order to improve the accuracy of the model identification results of pumped storage motor models, a new method based on block Hankel matrix is proposed to identify the extended Calman filter（EKF） model of pumped storage motor. Firstly, the pumped storage motor field equation was presented, an extended Calman filter based pumped storage motor model was also presented, and the state vector of resistance and inductance were used to construct the two sets of identification model, which realized the collaborative identification for the model. Secondly, based on the subspace identification strategy, the block Hankel matrix was used to construct the estimating equation of related data, and the null space projection was used to delete the input items, which could reduce the recognition complexity, then the least squares was used to achieve effective identification of system parameter matrix. Finally, the effectiveness of the proposed algorithm is verified by the simulation experiments on the Matlab platform and hardware environment.