Research On Tuning Methods Of Speed Loop Fractional Order PI Parameters Of PMSM Servo System

Among the motors used in AC servo systems,permanent magnet synchronous motors(PMSM)are widely used in industrial and transportation fields,such as robots,CNC machine and electric vehicles,due to their high torque inertia ratio,high power density,high efficiency and compact structure.With the improvement of production technology,the control precision and response speed requirements of PMSM servo system are getting higher and higher.The traditional integer-order PI(IOPI)control method is difficult to achieve high-performance control.Fractional-order PI(FOPI)is extended as IOPI,strong and flexible,it can meet the advantages of more performance indexes than IOPI.Therefore,the FOPI controller is applied to the PMSM servo system.However,since it has more parameters,the controller parameter tuning is more difficult than the IOPI.Therefore,the following studies are carried out:Firstly,the basic knowledge of fractional calculus is introduced and the methods of fractional integral implementation are studied.The characteristics of fractional calculus and its implementation methods are analyzed.Aiming at the problem of coefficient quantization in the implementation of fractional order integral in microcontroller,a cascaded second-order section implementation method is proposed to realize FOPI control.Secondly,fractional-order PI control is applied to PMSM servo system speed loop and tuned based on relay feedback.The application of the ultimate gain and limit frequency of the relay feedback identification in the PMSM servo system is studied.The control effect of the IOPI parameter is analyzed,.The bias relay feedback is proposed to identified the first-order plus dead time(FOPDT)transfer function model of system.The model is used to tune FOPI parameters,which avoids complicated fractional calculus equations solving process.The FOPI and IOPI control effect are analyzed and compared by simulation.Then,the relay feedback setting result is used as the initial value to perform the fuzzy control real-time online tuning of the FOPI parameters.For the servo system the control quality degrading problem due to controller parameters mismatch caused by load disturbance and parameter variation,and a real-time fuzzy control online parameter tuning strategy without de-fuzzification is proposed,which uses the relay feedback tuning result as initial values and avoids complex defuzzification process.An adaptive controller saturation value algorithm adjusted according to the error percentage is proposed to reduce the overshoot.According to the mean-filter output of the speed loop PI controller,the system working condition is identified,and a gain adaptive anti-load disturbance algorithm is proposed.The proposed fuzzy control algorithm,adaptive saturation value and the gain adaptive anti-load disturbance algorithm are verified by simulation.Finally,to verify the effectiveness of the FOPI tuning strategy based on relay feedback identification method and of the fuzzy control online tuning algorithm,adaptive controller saturation value algorithm and gain adaptive anti-load disturbance algorithm,the verification is carried out on the self-developed PMSM driver test platform.The experimental results show that:(1)The proposed fractional-order integration implementation on the microcontroller based on the cascaded second-order section structure can obtain the same frequency characteristics as the Oustaloup continuous domain;(2)The proposed relay feedback identification method is feasible in the PMSM servo system,the error integral performance index and overshoot of the fractional order PI control based on relay feedback setting are less than the integer order PI.;(3)The fuzzy online self-tuning algorithm with initial value based on relay feedback identification method can obtain better control quality than the FOPI,the proposed adaptive controller saturation value and gain adaptive anti-load disturbance algorithm effectively reduce the overshoot and speed fluctuation rate of system loaded or unloaded.Under the model identification accuracy of only 18.1%,the IAE indexes and speed fluctuation rate during loading and unloading are less than the relay feedback setting in the speed range of 10～3000 rpm.