Research On Phase Current Reconstruction Technique For Pmsm Based On Single Branch Sampling

Thanks to its simple structure,high reliability,high power density,and high efficiency,the permanent magnet synchronous motor(PMSM)is widely used in science and technology,industry,civil and other fields.The usage of multiple current or postion sensors in PMSM drive system will increase the driver’s volume,weight,cost,and the risk of failure of the control system,decreasing the operating reliability.Recently,single current sensor techniques have been widely concerned.However,the current control quality of DC-link sampling methods is adversely influenced by its inherent measurement dead zones(MDZs).The application range of multiple-branch methods is limited due to its complex circuit structure and the necessity of Hall-effect current sensors.Comprehensively considering the shortcomings of the current techniques,this paper proposes a new phase current reconstruction strategy and conducts research on its related problems.Focusing on the inherent MDZs of DC-link sampling method,a novel phase current reconstruction strategy based on single-branch sampling is proposed by changing the installing position of the SCS.Based on detailed analysis of the relationship between the sampling value of the SCS and the phase current under each voltage vector,a hybrid SVPWM(HSVPWM)technique,serving as an auxiliary,is put forward by modifying conventional SVPWM.The proposed method can guarantee the symmetry of three-phase PWM waveforms with simpler control algorithm and less impact on current when compared with traditional MDZ-elimination scheme.The single-branch sampling method(SBSM)has a simpler circuit structure,which can be achieved with a sampling resistor,expanding the application range and increasing the reliability.Considering the MDZs of the single-branch sampling method at high/over modulation region,an improved measurement vector insertion method is proposed by changing the position of the inserted vectors in each switching period.Influence of the maximum active vector ratio on the MDZs of SBSM is theoretically analyzed,and a comparative analysis with the traditional DC-link sampling method is also carried out.In theory,the improved strategy simplies the control algorithm,with reduced switching times and switching loss of the power device.The synthesized voltage vector can achieve the maximum output amplitude by avoiding the MDZs at high/over modulation region,which is confirmed by experimental results.Aiming at the problem of inaccurate sampling and large current reconstruction errors in single-branch sampling method during the switching process of different sectors,a current prediction solution with partially updating is proposed.Reasons for the appearance of inaccuracy and distortion in reconstructed currents at sector switching process under HSVPWM-based techniques are theoretically analyzed.The predicted dq-axis current information in the switching period is obtained in the control period before sector switching,and on this basis,the αβ-axis current is partially updated by utilizing the measurable current.Eventually,current reconstruction errors at sector switching process are eliminated.The experimental results show that current reconstruction errors and harmonics in reconstructed currents are both reduced to a great extent under the proposed method.Especially when the motor is in high-speed operating conditions,the effect is more apparent.The inherent sampling-delay error(SDE)existing in single-branch sampling method is studied,with an error compensation strategy based on synchronous rotation coordinate axis system proposed.The mechanism of SDE is theoretically analyzed and the error expression is also derived.Through analyzing the error composition of δγ-axis error component caused by the SDE,the compensation algorithm is implemented by sequentially eliminating the ac and dc error components.The proposed error compensation scheme which is independent of motor parameters,owns higher parameter robustness when compared with traditional methods.Theoretically,this method is also suitable for partial DC-link sampling methods and multi-branch sampling methods.Experimental results show that the current reconstruction error can be effectively reduced when the proposed error compensation scheme is used.Research on PMSM position sensorless control based on current reconstruction technique is carried out.Firstly,a nonlinear flux observer is adopted for the position estimation.Then,an iterative-step-size optimization method based on golden section search is proposed to furtherly reduce rotor position estimation error.At last,based on the content of the previous chapter,the overall control scheme of PMSM position sensorless control based on sampling-delay error compensation is established.Experimental results reveal that under the proposed position sensorless control scheme,the position estimation accuracy can reach the level of traditional multi-current sensor scheme at steady state and dynamic conditions.