Parameter Identification And Sensorless Control Strategy Of Permanent Magnet In-wheel Motor

Due to the in-wheel motor drive system including motor,controller,sensors and so on are installed in the narrow space,traditional hall position sensors are vulnerable to road impact,vibration,motor temperature rise,inverter dead zone,resulting in poor sensitivity,temperature drift and other problems.Some problems such as structural complications,reduced availability of space and increased production and maintenance costs are leaded by the installation of sensors.It is an effective measure to solve the above problems to study a technology that does not need sensors to complete the estimation of electrical angle and angular velocity.In this paper,the on-line parameter identification technology and sensorless control technology of in-wheel motor are studied.Firstly,the technical background and research status are expounded,the existing problems are analyzed,and the research objectives are drawn forth.The mathematical model of ideal interior PMSM is built,and the vector transformation of different coordinate systems is derived.The vector control technology is analyzed.Secondly,an on-line parameter identification method of motor based on improved teaching and learning optimization algorithm(ITLBO)is proposed considering the temperature rise environment of in-wheel motor.The temperature of the in-wheel motor will increases under the condition of long time and overload operation,which causes the motor parameter drift and reduces the motor performance.The bench is built and the off-line parameter identification method of in-wheel motor is designed,and the influence of temperature change on stator resistance,inductance and flux is analyzed according to the results.Gauss perturbation function is used to improve the initialization process,and group learning and last elimination mechanism are established to improve the identification speed and accuracy.A parameter identification algorithm model is established by Matlab/Simulink for verification.The results show that compared with the identification algorithms being compared,the improved algorithm has faster convergence rate and exacter convergence results.Thirdly,the driving condition of the vehicle is analyzed,and the model reference adaptive system(MRAS)method is selected to estimate the rotor speed of the in-wheel motor.The nonlinear variable gain PID technique in ADRC is used to replace the fixed gain PID adaptive law in the traditional model reference adaptation,and simulation model of model reference adaptive estimation is built.The experimental data show that the improved sensorless control technology has high robustness,which is more suitable for wide speed and variable load condition.Finally,in order to verify the actual operation effect of the control algorithm proposed in this paper,a motor test bench is built,and the hardware circuit of the control system is built based on the DSP28335.Based on the load motor,and the China automotive test cycle,by bench experiments the stability and anti-interference ability of the proposed sensorless control method are verified.