| IPMSM(Interior Permanent Magnet Synchronous Machine)has been gradually applied to urban railway traction system due to its nature of small volume,high mechanical strength and high output torque at low speed.Generally,Mechanical sensors are required to detect the speed and the rotor position in order to achieve high-performance transmission control.However,it increases the cost and volume,and when machine operates under the harsh conditions,the accuracy of the sensor is affected,which reduces the performance of the control system.To improve the operation performance and reliability of the whole system,the sensorless control technology of IPMSM has become a hot topic in traction control field.Hence,this paper focuses on the sensorless approach and are as follows:Firstly,three mathematical models of IPMSM built in different coordinate system are given,and the basic principles of vector control are introduced.The mathematical model of IPMSM in discrete time domain has been established to solve the problem,that the delay of the digital control system will have a great impact on the system,and the digital control delay is also compensated.Furthermore,since the coupling effect of the stator current is increased at low carrier ratio condition,a complex vector current regulator is introduced to enhance the decoupling effect of the stator current regulator in vector control,and the control performance of the system at low switching frequency is improved.Secondly,the principle of sliding mode variable structure control is elaborated combined with the special asymmetric structure of IPMSM,the extended back EMF is introduced to reconstruct the voltage model,an extended back EMF sliding mode observer for the IPMSM has been established,and the sliding mode gain is designed by using the Lyapunov stability criterion.The phase-locked loop is used to extract the speed and position.The super-twisting algorithm is introduced to optimize the sliding mode chattering phenomenon caused by traditional sliding mode observer,which will deteriorate the steady-state observation performance.Considering the special structure of the second order sliding mode observer,the active flux is used to reconstruct the voltage model,and the back-EMF second order sliding mode observer based on the super-twisting algorithm is established.The sliding mode gain is also designed by using the Lyapunov stability criterion.The simulation results verify the effectiveness of the second order sliding mode observer to eliminate the chattering phenomenon,which can improve the steady-state performance.Besides,Conventional phase-locked loop based speed and rotor position observer produce finite speed observation errors when the motor is working at the acceleration and deceleration process.Aiming at solving the above-mentioned problem,a novel frequency-locked loop is designed by using the characteristics of pure sinusoidal wave,which can achieve the non-error observation when the input is ideal sinusoidal.Since the actual back EMF input are sinusoidal waves containing the harmonic contents,and the sinusoidal waves amplitude changes with time,the frequency estimation algorithm under non-ideal input is introduced.Based on this,the small-signal model of the frequency-locked loop is used to analyze the proposed scheme,which can prove the no-error tracking performance when inputting the frequency ramp signal.In order to further improve the control performance of the speed estimation scheme,the relationship between control parameters and control performance is analyzed,and the control parameter design strategy is given.Finally,The controlled object modeling and the hardware-in-the-loop tests was completed on the hardware-in-the-loop test platform based on RT-LAB and TMS320F28335 controller.The test results verify the correctness and effectiveness of the control algorithm. |
PDF Full Text Request