基于随机森林回归算法的感应电机驱动控制

Induction Motor Drive Control Based on Random Forest Regression Algorithm

为了提高感应电机驱动系统在不同运行条件下的控制性能,将随机森林回归(RFR)算法引入到空间矢量脉宽调制(SVPWM)中,设计了一种新型的感应电机驱动控制器。和常规SVPWM方案相比,RFR的引入为SVPWM算法提供了快速实现和预测改进的优势,从而性能得到了提高。为了进一步突出基于RFR算法的优势,还与人工神经网络(ANN)算法和自适应神经模糊系统(ANFIS)算法进行对比分析,分析结果表明RFR算法在不同工况下的稳态误差、暂态响应和鲁棒性均优于ANN算法和ANFIS算法。最后,通过仿真和试验验证了新型控制算法的实际效果。

In order to improve the control performance of induction motor drive system under different operating conditions,a random forest regression（RFR）algorithm was introduced into space vector pulse width modulation（SVPWM）,and a new induction motor drive controller was designed. Compared with the conventional SVPWM scheme,the introduction of RFR provided the advantages of fast implementation and prediction improvement for the SVPWM algorithm,which resulted in the performance improvement. In order to further highlight the advantages of RFR algorithm,it was also compared with artificial neural network（ANN）algorithm and adaptive neural fuzzy inference system（ANFIS）algorithm. The analysis results show that the steady-state error,transient response and robustness of the RFR algorithm are better than those of the ANN algorithm and the ANFIS algorithm under different operating conditions. Finally,the actual effect of the new control algorithm is verified by simulations and experiments.