Model-based Design Of Permanent Magnet Synchronous Traction Motor Controller For Electric Vehicles

As the power source of driveline in electric vehicle, permanent magnetsynchronous motor (PMSM) maintains the operation of the vehicle. Motor control isessential to the vehicle’s safety, reliability, operability, dynamic performance, fueleconomy and comfort. To improve the reliability of motor control system, faulttolerance of speed sensor is added to the motor control system.At first, a mathematical model of PMSM and sensor models are built and acontrol software prototype is established based on the vector control algorithm andspace vector pulse width modulation. These modules constitute the motor controlsystem model in simulation. In order to acquire more accurate motor model, thispaper expounds the factors that result in the change of parameters in PMSM, anddiscusses the influence on the control algorithm and speed sensorless control causedby time-varying characteristics of parameters. Recursive least squares algorithm withforgetting factor is introduced to identify the parameters in PMSM. Based on thePMSM model with revised parameters, several typical sensorless control algorithmsof PMSM are studied. To eliminate the error introduced by open-loop estimation andparameter variation, the method of state observer is adopted to observe the status ofelectromotive force, which can derive the position and rotary speed. Then, based onthe Popov stability theory, model reference adaptive method is applied to theestimation and reconstruction of position and speed in motor. To overcome the noiseand disturbance in the control and measurement system extended Kalman filtermethod is proposed to estimate the position and speed. By several methods forposition and speed estimations, fault tolerance is realized on the condition of failurein speed/position sensor.The development platform is established on MATLAB/Simulink， whichcan develop the software system of PMSM controller. The platform has threemain functions，such as model-based code auto-generation, modular graphicalprogramming and step-by-step verification. Model-based code auto-generation can effectively avoid the error in manual code. Modulargraphical programming can improve the readability，the scalability and theportability of the program. Step-by-step verification can help developers findthe mistake and correct it timely.At last，the design of hardware system is introduced. The construction ofmotor controller and allocation of power sources are discussed. The schematicdiagrams of control circuit and driver circuit are detailed described.