The Optimal Speed-Torque Control of Asynchronous Motors for Electric Cars in the Field-Weakening Region Based on the RFR

For the optimal speed–torque control of asynchronous motors for the electric cars in the field-weakening region, a random forest regression (RFR) algorithm is introduced in this article, to solve the problem of stator d -axis ( ${\boldsymbol{i}_{\boldsymbol{sd}}}$ ) and the q -axis current ( ${\boldsymbol{i}_{\boldsymbol{sq}}}$ ) matching. First, a vector control system for the asynchronous motors is built on the basis of the current distribution output model. Second, according to the limiting conditions of maximum voltage ( ${\boldsymbol{U}_{\boldsymbol{smax}}}$ ) and current ( ${\boldsymbol{I}_{\boldsymbol{smax}}}$ ), an analytical model of the maximum torque output is established, and the current distribution law is analyzed. Third, based on the variation laws of ${\boldsymbol{i}_{\boldsymbol{sd}}}$ and ${\boldsymbol{i}_{\boldsymbol{sq}}}$ , a closed-loop voltage vector analytical model is designed and embedded in the vector control system to analyze the simulation results. Fourth, an AVL dynamometer experimental platform is built to collect the measured sample data under a constant temperature of 85 °C and maximum power output. The working condition parameters serve as the input of the RFR model, whereas ${\boldsymbol{i}_{\boldsymbol{sd}}}$ and ${\boldsymbol{i}_{\boldsymbol{sq}}}$ are the outputs. The validity of the current distribution is verified. Finally, the regression model is embedded in the vector control system to determine ${\boldsymbol{i}_{\boldsymbol{sd}}}\,$ and ${\boldsymbol{i}_{\boldsymbol{sq}}}$ under different working conditions. Results verify the correctness and effectiveness of the proposed algorithm.