| PMSM have the characteristics of simple structure,high energy density and good adjustment performance.So it has received widespread attention.In high performance motor speed regulation system.Many scholars use double closed-loop system to control the motor,such systems generally have high accuracy.Sensors are designed in traditional control systems to improve control accuracy.The position sensor is relatively expensive,it has high requirements for the environment and it is particularly vulnerable to damage.This limits the use of PMSM.In order to overcome the weak point of the use position sensor,research on position detection without position sensor for PMSM.Design a motor control system,it can control the motor in the widest possible speed range.A large number of experiments have been done in this system,and the results show that this system is feasible.According to the structural characteristics of PMSM,this paper derives the mathematical formula of PMSM.We deeply analyze SVPWM and vector control.Sliding mode observer is used at high speed.However,this algorithm also has the problems of chattering and phase lag.Our focus is on the second-order sliding mode observer.Add adaptive coefficients to the integral term in the algorithm,use saturation function instead of switching function,use a filter with a variable cutoff frequency.Two models are built in the simulation software.The result indicated that the improved super spiral has a faster response speed,and its position estimation is accurate.In the low-speed stage,the back electromotive force of the PMSM is relatively small,so its observer accuracy is reduced.To solve the problem,during low speed rotation,the high frequency injection is designed and used in the motor control process.The simulation model is made in the software.The simulation waveform shows that its estimation performance is better when the motor runs at low speed and it can meet the control requirements.In order to verify the rationality of the algorithm,a control system based on TMS320F28335 was made by us.Firstly,the running experiment with sensor is carried out on this controller.Then,a position sensorless control algorithm was introduced.Finally we did a comparative experiment.Experiments show that the improved sliding mode observer designed in this paper reduces the error of speed estimation by 5r/min compared with the traditional algorithm at high speed and steady state.The rotor position estimation accuracy is 0.09 rad higher than the traditional algorithm.This system has high estimation accuracy,able to meet the control requirements of traction system. |