A parameter adaptive permanent magnet synchronous motor position sensorless control method

Abstract

This invention provides a parameter-adaptive sensorless control method for permanent magnet synchronous motors. To address the uncontrollable performance degradation caused by signal errors from position sensors due to harsh working environments or sensor malfunctions, this method utilizes estimated speed and position values ​​instead of the actual speed and position detected by the position sensor. Considering that sensorless control systems do not contain any actual speed and position signals and suffer from weak low-speed load capacity and significant susceptibility to model parameter changes, this invention introduces a current estimation module to obtain the quadrature-axis and direct-axis current estimation errors. Compensated estimated speed and position are obtained through a speed estimation module and a position estimation module, both based on compensation mechanisms for the quadrature-axis current estimation error and direct-axis current estimation error, respectively. A parameter adaptive module is then used to correct the motor parameters in real time, updating the speed and position estimation modules. Compared to existing technologies, this invention significantly improves the estimation accuracy and parameter robustness of sensorless motor control systems.

Description

Technical Field

[0001] This invention relates to the field of motor drives, and specifically to a sensorless control method for a parameter-adaptive permanent magnet synchronous motor. Background Technology

[0002] Permanent magnet synchronous motors (PMSMs) are widely used in industry due to their high efficiency and control precision. Advanced torque control strategies, such as rotor field-oriented control, are common in PMSM drives. However, these require rotor position sensors, such as encoders or Hall effect sensors, to detect the rotor position. The use of sensors not only increases the cost of the control system but also introduces problems such as installation difficulties and reduced reliability. Therefore, exploring sensorless algorithms has become a key research focus for circumventing these issues.

[0003] Currently, various control methods exist for sensorless control of permanent magnet synchronous motors (PMSMs), such as model reference adaptation (MRA), sliding mode observers (SMA), and high-frequency signal injection. MRA is a common method for estimating motor speed and position. However, this method requires highly accurate motor parameters, especially when the motor is running at low speeds. When parameters change during motor operation, control stability may be affected, leading to a deterioration in the overall performance of the motor control system. While many different online parameter identification methods have been proposed, they have not fully considered the effective combination of improving the observer's own accuracy and real-time parameter updates within the algorithm.

[0004] To address the problems existing in the above-mentioned technologies, this invention provides a parameter-adaptive permanent magnet synchronous motor sensorless control method. The method obtains the current error of the direct and quadrature axes through a current estimation module, obtains the compensated estimated electric angular velocity and position using speed estimation and position modules, and uses the motor parameters obtained by the parameter adaptation module for speed and position estimation, effectively improving the estimation accuracy of the sensorless control algorithm. Summary of the Invention

[0005] To effectively combine the improvement of the observer's own accuracy with the real-time updating of parameters within the algorithm, this invention provides a parameter-adaptive permanent magnet synchronous motor sensorless control method.

[0006] This invention discloses a position-sensorless controller, including The current estimation module is used to obtain the estimated quadrature-axis current error of the permanent magnet synchronous motor. and estimation error of direct-axis current ; The velocity estimation module is used to obtain the q′ axis current required by the velocity observer. and estimated load torque And calculate the estimated value of the compensation speed; The position estimation module is used to obtain the estimated d′-axis current value required by the position observer. And calculate the estimated value of the compensation location; The parameter adaptive module is used to identify the motor's resistance. and inductor ; The speed conversion module is used to convert and obtain the estimated mechanical angular velocity. .

[0007] The present invention also provides a sensorless control method for a permanent magnet synchronous motor based on the above-mentioned adaptive parameters, comprising the following steps: Establish a virtual rotating coordinate system d′q′, rewrite the voltage equation under the virtual rotating coordinate system, and construct a current estimation error model under the virtual rotating coordinate system. Collect the phase a current i of the permanent magnet synchronous motor a b-phase current i b And reconstruct the c-phase current i c The motor α-axis current i is obtained through Clark transformation. α With motor β-axis current i β And through the Park transformation, the actual direct-axis current i of the motor is obtained. d With respect to the actual quadrature-axis current i of the motor q This allows us to obtain the current estimation error value; A speed observer is designed using mechanical equations and current estimates. When observing speed, the q′ axis current and load torque need to be estimated first. At the same time, the obtained current estimates are used to compensate for the observer's estimates and improve the observer's accuracy. A position observer is designed by integrating the observed rotational speed and estimating the current. When observing the position, the current of the d′ axis needs to be estimated first. The speed observation error will cause position integration observation error. A correction term is designed to compensate for the position estimate of the observer. At the same time, an adaptive gain coefficient is designed to improve the performance of the observer. It provides the selection criteria for the design coefficients of current, speed and position observers, and gives the pole design and determination criteria for current observers, speed observers and position observers; Taking into account the sensitivity of the observer's model parameters, the relationship between the speed and position estimation errors and the magnitude of the current speed and parameter errors is given. It is found that the main model parameters affecting the position error when the motor is running at low speed and high speed are resistance and inductance. Design an adaptive method for motor inductance and resistance based on recursive restricted total least squares method, update the resistance and inductance parameters in the above observer in real time, and give the adaptive law of the design and the suggested adaptive update period. The final speed estimate is then used to obtain a given quadrature-axis current i via a speed PI regulator. q* At the same time i d * The constant value is 0. After passing through the direct and quadrature axis current PI regulator, the quadrature axis voltage u is obtained. q Direct-axis voltage u d The estimated position value is used for coordinate transformation to obtain the α-axis voltage u. α With β-axis voltage u β The inverter control signal is obtained through space vector pulse width modulation (SVPWM) technology to control the permanent magnet synchronous motor.

[0008] Compared with the prior art, the present invention brings the following beneficial effects: (1) This invention proposes a sensorless control method for permanent magnet synchronous motor with adaptive parameters. The method obtains the compensated feedback estimated speed and position through a current estimation module, a speed estimation module, a position estimation module, and a parameter adaptive module. At the same time, the algorithm model parameters are corrected and compensated, which can improve the estimation performance of the sensorless control method and enhance the reliability of the control system.

[0009] (2) The present invention uses a parameter adaptive module to correct and compensate the parameters of the model within the algorithm, taking into account the influence of parameter error on position observation error, effectively compensating for the resistance and inductance parameter errors of the model within the algorithm, and improving the estimation accuracy of speed and position. Attached Figure Description

[0010] Figure 1 This is the control block diagram for the speed estimation module; Figure 2 Here is the control block diagram for the position estimation module; Figure 3 The control block diagram for the parameter adaptive module; Figure 4 Control block diagram of a sensorless control method for a permanent magnet synchronous motor with adaptive parameters; Figure 5 The rotational speed waveform is shown for a permanent magnet synchronous motor that has not implemented the conventional speed estimation method provided by this invention. Figure 6 The rotational speed waveform diagram for implementing the parameter adaptive permanent magnet synchronous motor speed estimation method provided by this invention; Figure 7 The position waveform diagram is for a permanent magnet synchronous motor without implementing the conventional speed estimation method provided by this invention. Figure 8 Position waveform diagram for implementing the parameter adaptive permanent magnet synchronous motor speed estimation method provided by the present invention. Detailed Implementation

[0011] To describe the present invention more specifically, the technology of the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only for explaining the present invention and are not limited to the present invention.

[0012] This invention proposes a sensorless control method for a parameter-adaptive permanent magnet synchronous motor. The controller mainly includes a current estimation module, a speed estimation module, a position estimation module, a parameter adaptation module, a speed integration module, and a speed conversion module. The specific implementation process is as follows: Establish a virtual rotating coordinate system d′q′, and rewrite the voltage equation in the virtual rotating coordinate system as follows: (1) Among them, u d′ u q′ Let i be the voltage across the d′ axis and the voltage across the q′ axis in the d′q′ coordinate system. d′ i q′ For the d′-axis current and q′-axis current in a virtual coordinate system, To estimate the angular velocity, R is the resistance of... e d′ e q′ Let be the back electromotive force in the d′q′ coordinate system, which is equivalent to the projection of the back electromotive force in the actual dq coordinate system of the motor onto the established virtual coordinate system d′q′ axis. Its expression is as follows: (2) in, For rotor position estimation Compared with the actual value of rotor position θ e The error between them What is it? What is it?

[0013] Simultaneously, a current estimation error model in a virtual rotating coordinate system is constructed, which mainly estimates the d′q′ axis current in the virtual rotating coordinate system through a rewritten voltage equation. The differential of the current estimate can be written as: (3) The phase a current i of the permanent magnet synchronous motor has now been collected. a b-phase current i b And reconstruct the c-phase current i c The motor α-axis current i is obtained through Clark transformation. α With motor β-axis current i β And through the Park transformation, the actual direct-axis current i of the motor is obtained. d With respect to the actual quadrature-axis current i of the motor q Thus, the current estimation errors are obtained as follows: (4) Where Δx(k) represents the current estimation error between the estimated and actual values. The differential expression for the current estimate is: (5) Among them, u d′ and u q′ As the voltage in the virtual d′q′ coordinate system, it can be represented by the output value u of the current loop PI regulator in the vector control system. d * and u q * replace. and The estimated value is an approximate disturbance term, which is included in the actual operation of the control system. The back electromotive force term, G d and G q These are the gain coefficients used when designing the observer.

[0014] Furthermore, combining equations (3) and (5), the dynamic error of the current estimate can be obtained as follows: (6) The current error expression is simplified using a small-signal model. Since the current estimation error is small in steady state and the control position error is close to 0, which is the control objective, the adjustment intensity of the system control loop is related to the position error. According to trigonometric functions, the above formula, equation (6), should satisfy... and Then we can obtain the following relationship for the approximate terms of the estimated value: (7) The observations of the observers used to estimate velocity and position can be compensated based on the existence relationship between the approximate terms of the current estimation error values ​​of the d′q′ shaft system, thereby improving the estimation accuracy.

[0015] A speed observer is designed based on the mechanical equations and current estimates. When observing the speed, it is necessary to first observe the q′-axis current and load torque to compensate for the observer's estimates and improve its convergence. The corresponding expression is written as: (8) Where p is the extreme logarithm. It is electromagnetic torque. It is the load torque observation value. , and These are three observer state variables. , and Gain coefficient.

[0016] A position observer is designed based on the integral of the observed rotational speed and the estimated current. When observing the position, the current along the d′ axis must be observed first. Since speed observation errors will cause position integration observation errors, a correction term can be designed to compensate for the observer's estimated value. The corresponding expression is written as: (9) in, and These are the observer state variables. and Gain coefficient, These are adaptive coefficients based on the current velocity, used to ensure the convergence of the observer. At lower speeds, the upper limit threshold can be set to... To prevent excessive values, a lower threshold can be set at higher speeds. Prevent it from being too small.

[0017] Accurate observation of rotational speed and position is related to the coefficients of the observer design; therefore, it is necessary to analyze the selection of the observer coefficients, mainly including those in the d′q′ shaft current observer. and In the speed observer and and in the position observer .

[0018] In designing a current observer and Since the goal of the control system is to accurately estimate velocity and position, i.e., its estimation error is small, the dynamic current error shown in equation (6) can be simplified as follows: (10) The poles of the current observer can be obtained. and It should meet the following requirements: (11) According to the principle of stability, in order to make the system stable, the following design is required. and The extreme point should be and There are two negative poles, meaning we only need to select... and The requirement is satisfied if the values ​​are large and the same. .

[0019] In the design of speed observers and position observers , and When the time is right, substitute equation (7) into equation (8) and subtract it from the mechanical model to obtain the dynamic error at this time: (12) There are three poles in the above equation (12). , and According to the stability principle, the observer is stable when all poles are negative real poles. In this case, its characteristic equation can be solved to obtain... , and Its expression is: (13) Directly obtaining its analytical solution is quite complex. Therefore, we can first set the desired pole values, and then numerically solve for its gain value to obtain the desired result. , and .

[0020] Furthermore, the sensitivity of the observer's velocity and position estimation errors to the model parameters is analyzed. For ease of analysis, the parameter error existing in the control system is defined as resistance error. Inductance error and flux linkage error The dynamic model of current error can be evaluated as the difference between equation (3) and equation (5). When the motor is running in steady state, the current error... and Neglecting this, it becomes a correction term in equation (5). In this case, the velocity and position estimation errors depend on the magnitude of the current velocity and parameter errors, and their expressions are as follows: (14) When the motor is running in steady state, the speed at the position integral is constant. Therefore, the position error is the main factor affecting the observer's accuracy, and it is related to the dq-axis current, speed, and parameter errors under the current operating conditions. Furthermore, when the motor is running at low speed, its position error is mainly related to the resistance error term. The position error is related to the inductance error when the motor is running at high speed. Furthermore, when the motor load is large, the large q-axis current leads to a small inductance error that causes a large position error. Therefore, this invention identifies the two key model parameters, resistance and inductance.

[0021] Specifically, this invention relies solely on existing data and uses a recursive restricted total least squares method to identify the inductance and resistance of the motor, thereby updating the resistance and inductance parameters in the aforementioned observer. The algorithm is expressed as follows: (15) in, It is a learning rate greater than zero. , , , Let $k$ represent the residual at time k-1, $input$, $output$, and $estimated$, respectively.

[0022] First, estimate the inductance, then estimate the resistance, using... , , The output signal, input signal, and estimation parameters are represented during inductance estimation. The inductance estimation expression is as follows: (16) in, The calculation period for inductance estimation is related to the computational load and convergence speed; the chosen calculation period is... The smaller the value, the faster the convergence speed, but the computational load will also increase. Therefore, we choose the system control cycle, i.e. .

[0023] Similarly, using , , The output signal, input signal, and estimation parameters are represented during resistance estimation. The resistance estimation expression is as follows: (17) in, The calculation cycle for resistance estimation is set to a relatively long value. Considering that resistance changes are temperature-dependent and relatively slowly, and also taking into account the microcontroller's processing power, a longer calculation cycle is chosen. .

[0024] Furthermore, in each algorithm loop, the inductance estimate obtained in the first run... It was updated in the second run to obtain the resistance estimate. The parameters obtained from this algorithm will be updated in real time during the operation of the aforementioned velocity and position observers.

[0025] The final speed estimate is then used to obtain a given quadrature-axis current i via a speed PI regulator. q * Meanwhile, given the direct-axis current i d * The voltage is always equal to 0. After passing through the direct and quadrature axis current regulator, the quadrature axis voltage u of the motor is obtained. q Motor direct shaft voltage u d The motor α-axis voltage u is obtained through coordinate transformation from a two-phase rotating coordinate system to a three-phase stationary coordinate system. α With motor β-axis voltage u βThe inverter control signal is obtained through space vector pulse width modulation (SVPWM) technology to control the permanent magnet synchronous motor.

[0026] The above description of the embodiments is provided to enable those skilled in the art to understand and apply the present invention, but the present invention is not limited to the above embodiments. Any improvements and modifications made to the present invention by those skilled in the art based on the disclosure thereof should be within the scope of protection of the present invention.

Claims

1. A position-sensorless controller, characterized in that, include The current estimation module is used to obtain the estimated quadrature-axis current error of the permanent magnet synchronous motor. and estimation error of direct-axis current ; The velocity estimation module is used to obtain the q′ axis current required by the velocity observer. and estimated load torque And calculate the estimated value of the compensation speed; The position estimation module is used to obtain the estimated d′-axis current value required by the position observer. And calculate the estimated value of the compensation location; The parameter adaptive module is used to identify the motor's resistance. and inductor ; The speed conversion module is used to convert and obtain the estimated mechanical angular velocity. .

2. A parameter-adaptive permanent magnet synchronous motor control method without position sensors based on the position sensorless controller described in claim 1, characterized in that, Includes the following steps: Establish a virtual rotating coordinate system d′q′, rewrite the voltage equation under the virtual rotating coordinate system, and construct a current estimation error model under the virtual rotating coordinate system. Collect the phase a current i of the permanent magnet synchronous motor a b-phase current i b And reconstruct the c-phase current i c The motor α-axis current i is obtained through Clark transformation. α With motor β-axis current i β And through Park transformation, the actual direct-axis current i of the motor is obtained. d With respect to the actual quadrature-axis current i of the motor q This allows us to obtain the current estimation error value; A speed observer is designed using mechanical equations and current estimates. When observing speed, the q′ axis current and load torque need to be estimated first. At the same time, the obtained current estimates are used to compensate for the observer's estimates and improve the observer's accuracy. A position observer is designed by integrating the observed rotational speed and estimating the current. When observing the position, the current of the d′ axis needs to be estimated first. The speed observation error will cause position integration observation error. A correction term is designed to compensate for the position estimate of the observer. At the same time, an adaptive gain coefficient is designed to improve the performance of the observer. It provides the selection criteria for the design coefficients of current, speed and position observers, and gives the pole design and determination criteria for current observers, speed observers and position observers; Taking into account the sensitivity of the observer's model parameters, the relationship between the speed and position estimation errors and the magnitude of the current speed and parameter errors is given. It is found that the main model parameters affecting the position error when the motor is running at low speed and high speed are resistance and inductance. Design an adaptive method for motor inductance and resistance based on recursive restricted total least squares method, update the resistance and inductance parameters in the above observer in real time, and give the adaptive law of the design and the suggested adaptive update period. The final speed estimate is then used to obtain a given quadrature-axis current i via a speed PI regulator. q * At the same time i d * The constant value is 0. After passing through the direct and quadrature axis current PI regulator, the quadrature axis voltage u is obtained. q Direct-axis voltage u d The estimated position value is used for coordinate transformation to obtain the α-axis voltage u. α With β-axis voltage u β The inverter control signal is obtained through space vector pulse width modulation (SVPWM) technology to control the permanent magnet synchronous motor.

3. The parameter-adaptive permanent magnet synchronous motor sensorless control method according to claim 2, characterized in that, In step (3), a speed observer is designed based on the mechanical equation and current estimate. When observing the speed, it is necessary to first observe the q′ axis current and load torque to compensate for the estimated value of the observer. Its expression is: (1) Where p is the number of pole pairs of the motor. It is electromagnetic torque. It is the load torque observation value. , and These are three observer state variables. , and Gain coefficient.

4. The parameter-adaptive permanent magnet synchronous motor sensorless control method according to claim 2, characterized in that, In step (4), a position observer is designed based on the integral of the observed rotational speed and the estimated current. When observing the position, it is necessary to first observe the current of the d′ axis. Since the speed observation error will cause the position integration observation error, a correction term can be designed to compensate for the estimated value of the observer. Its expression is: (2) in, and These are the observer state variables. and Gain coefficient, These are adaptive coefficients based on the current velocity, used to ensure the convergence of the observer. At lower speeds, the upper limit threshold can be set to... To prevent excessive values, a lower threshold can be set at higher speeds. Prevent it from being too small.

5. The parameter-adaptive permanent magnet synchronous motor sensorless control method according to claim 2, characterized in that, The pole design and determination criteria for the current observer, velocity observer, and position observer provided in step (5) include the poles of the current observer. and Should meet: (3) According to the principle of stability, in order to make the system stable, the following design is required. and The extreme point should be and There are two negative poles, meaning we only need to select... and The requirement is satisfied if the values ​​are large and the same. ; In the design of speed observers and position observers , and At that time, the three poles in the dynamic error expression can be obtained according to the stability principle. , and When all poles are negative real poles, the observer is stable, and its characteristic equation can be used to solve for the following: , and Its expression is: (4) Directly obtaining its analytical solution is quite complex. Therefore, we can first set the desired pole values, and then numerically solve for its gain value to obtain the solution that meets the requirements. , and .

6. The parameter-adaptive permanent magnet synchronous motor sensorless control method according to claim 2, characterized in that, In step (7), an adaptive law for motor inductance and resistance based on recursive restricted total least squares method and a proposed adaptive update period are designed, wherein the algorithm is expressed as: (5) in, It is a learning rate greater than zero. , , , Let K and K represent the residual, input signal, output signal, and estimated parameters at time k-1, respectively. This invention first estimates the inductance, then estimates the resistance, and uses... , , The output signal, input signal, and estimation parameters are used to represent the inductance estimation. The resulting inductance estimation expression is: (6) in, The calculation period for inductance estimation is related to the computational load and convergence speed; the chosen calculation period is... The smaller the value, the faster the convergence speed, but the computational load will also increase. Therefore, we choose the system control cycle, i.e. ; Similarly, using , , This represents the output signal, input signal, and estimation parameters during resistance estimation. The resistance estimation expression is as follows: (7) in, The calculation cycle for resistance estimation is set to a relatively long value. Considering that resistance changes are temperature-dependent and relatively slowly, and also taking into account the microcontroller's processing power, a longer calculation cycle is chosen. .

Images