Research On Key Technologies Of High Performance Current Loop For Permanent Magnet Synchronous Motor

Permanent magnet synchronous motor(PMSM),as a key actuator of industrial automation equipment,has been widely used in CNC machine tools,industrial robots,aerospace,general-purpose AC servo and other fields.As the innermost loop of PMSM control system,the control performance of current loop plays a decisive role in the control effect of the whole system.Compared with foreign products,the current loop control performance of the domestic system still has obvious gap,the key technical problems to be solved include automatic acquisition of physical information of controlled motor,torque or current ripple suppressio n and compensation of control delay and sampling error.In this paper,the PMSM drive system is taken as the research object,and the key technologies,such as parameter identification;inverter nonlinearity compensation;one/two times fundamental frequency torque ripple suppression;un-modeled factors compensation and modeled factors optimization of deadbeat predictive control are studied in-depth to improve the current loop control overall performance of PMSM.The specific research contents of this paper are as follows.The accurate parameter information is the basis of current controller design and subsequent algorithm optimization.According to the application status of domestic AC servo and frequency conversion products,one driver is often used to matc h multiple motors,and considering that the existing open-loop parameter identification strategy has the problem that the excitation signal cannot be accurately controlled and may lead to overcurrent of the drive system.In this paper,according to the actual industrial application requirements,a serious of parameter identification strategies based on voltage excitation signal self-learning is proposed.The proposed strategy takes the current feedback during the self-learning process as the criterion to automatically determine the required voltage amplitude and complete the different types of parameter identification,which avoids the trial-and-error of voltage reference in the traditional open-loop method,and ensures the safe operation of PMSM control system.On the basis of detailed analysis of the sources of parameter identification errors,with the aim to simplify the process of parameter identification and ensure the accuracy of identified results,the parameter identification process which can eliminate various causes of identification errors is designed.The experimental results show the proposed method can achieve an accurate parameter identification without constructing closed-loop control or using any nonlinearity compensation method.Inverter nonlinearity compensation is one of the effective methods to ensure high quality current control performance.Aimed at the problem of poor compensation effect caused by inaccurate phase voltage error calculation under small current in traditional compensation methods,an accurate calculation method of voltage error based on iterative linear interpolation is proposed.The nonlinear effects,such as dead-time,turn-on and turn-off delay,parasitic capacitance and voltage losses on the power devices are comprehensively considered by the proposed method.Accurate mapping between the voltage errors of each inverter phase and their corresponding phase current amplitudes is obtained.Aimed at the problem that the use of current feedback information for voltage error ta le look-up will lead to poor compensation effect near the zero-current zone,the phase current reference information is used to determine the voltage errors during the compensation in this paper,which effectively improves the accuracy of voltage error judgment near zero-crossing of phase current.Through the experimental comparison with the traditional inverter nonlinearity compensation methods,it is proved that the proposed compensation method has better current harmonic suppression ability under different working conditions.Low frequency torque ripple suppression is of a great significance to improve the steady state performance of PMSM.In this paper,the causes of torque ripple with frequency of one or two times fundamental frequency are analyzed theoretically.It is pointed out that the main causes of torque ripple with frequency of one or two times fundamental frequency are current sampling error and asymmetry of motor winding.Combined with closed-loop control,the influence of speed loop controller parameters on the amplitude of one/two fundamental frequencies is analyzed,and an online torque ripple suppression method based on current ripple information reconstruction is proposed.Starting from the torque ripple generation mechanism,the discrete Fourier transform is used to reconstruct the harmonics,and the current feedbacks or current references are modified according to the cause of the ripple,so that the torque ripple can be suppressed adaptively.Compared with the existing torque ripple suppression strategies,the proposed method does not need any specific information such as shaft inertia or motor parameters in advance,and has no requirement to any parameter scheduling.Experimental results show that the proposed method can effectively suppress the one/two times fundamental frequency torque ripple of different causes.Deadbeat predictive control(DBPC)has favorable dynamic response and small current ripple.However,its control performance is greatly affected by the accuracy of the model parameters,and there is an obvious current steady state error in the DBPC due to the lack of integral part.In this paper,the sensitivity of parameters,rotor angular position error,inverter nonlinear error and inherent delay of discrete control are analyzed in detail.The rotor angle position error compensation strategy based on disturbance observer is proposed,and combined with the algorithm of motor parameter identification and inverter nonlinearity compensation,the dynamic and steady state performance of deadbeat predictive current control is comprehensively improved.This paper also analyzes the current response law of DBPC under voltage boundary conditions,and simulation studies prove that higher switching frequency and lower bus voltage will lead to the saturation of the current loop under the speed or current step conditions,which affects the current dynamic performance.The experimental results show that the comprehensive solution proposed in this paper can realize the parameter robustness and zero steady state current error control of DBPC,which verifies the theoretical innovation and engineering application value of the proposed algorithm.