采用改进递归最小二乘法的电动负载模拟器参数在线辨识

For the purpose of achieving accurate on-line identification of all parameters for the electric load simulator, an improved recursive least squares (IFRLS) online parameter identification method is proposed by introducing a low-pass filter and covariance matrix readjustment to the traditional least squares algorithm. Firstly, the mathematical model of the electric load simulator was established, and the filtering method in identification process was designed. Then, based on the traditional least-squares algorithm with forgetting factor, covariance resetting was introduced to deal with the non-convergence problem. Finally, on-line parameter identification and verification experiments were conducted. The experimental results show that the non-convergence problem in the identification process was solved effectively by using the proposed improved recursive least squares algorithm, and online identification for all parameters of the electric load simulator was realized. The obtained identification model has a fit of 93.8% with the real model.