Permanent magnet synchronous motor inductance parameter online identification method based on position decoupling

Abstract

The invention discloses a permanent magnet synchronous motor inductance parameter online identification method based on position decoupling, and relates to the technical field of permanent magnet synchronous motor inductance parameter identification. According to the position decoupling online inductance parameter identification method, a virtual synchronous rotation reference system is constructed, a position error signal caused by inductance parameter deviation is extracted and compensated, firstly, the coupling effect of inductance parameter deviation and position errors in a sensorless control framework is eliminated, and the position error is corrected; and then coordinate transformation is carried out according to the compensated estimated rotor position, and finally effective identification of actual inductance is realized through a recursive total least square method. According to the position decoupling algorithm provided by the invention, the problem of failure of traditional online inductance parameter identification under a sensorless control framework is solved, additional signal injection is not needed, the inductance parameters of the system can be effectively tracked during operation, the sensorless control performance is improved, and the additional loss of the system is reduced.

Description

Technical Field

[0001] The present invention relates to the field of permanent magnet synchronous motor drive control, and is an online identification method for inductance parameters of a permanent magnet synchronous motor based on position decoupling. Background Art

[0002] Permanent magnet synchronous motors (PMSMs), with their high power density, compact size, and smooth low-speed operation, are increasingly being used in various servo systems, including electric vehicles, industrial robots, CNC machine tools, and other applications requiring high performance and precision. Compared to traditional sensored control, sensorless control eliminates the need for additional position sensors, effectively reducing the cost of PMSM control systems and improving stability. However, sensorless control schemes place high demands on PMSM model parameters, including stator resistance, stator inductance, and permanent magnet flux. While stator resistance and permanent magnet flux exhibit small temperature deviations due to temperature effects, temperature compensation can achieve relatively accurate compensation. However, stator inductance, affected by magnetic saturation, is susceptible to significant offsets due to load fluctuations during control. Furthermore, errors in the stator inductance parameters can introduce errors in rotor position estimation. Therefore, effectively identifying the accurate stator inductance parameters of PMSMs is crucial in sensorless control schemes.

[0003] Currently, inductor parameter identification is primarily divided into offline and online methods. In traditional sensored control schemes, online identification algorithms often address the lack of rank by injecting additional signals or fixing other parameters. However, in sensorless control schemes, additional signal injection introduces undesirable current loop perturbations and noise. Even with fixed stator resistance and permanent magnet flux parameters, the coupling effect between position estimation error and inductor parameter deviation persists when using fixed parameter methods, rendering traditional schemes ineffective in sensorless control. This coupling effect arises because the inductor parameter deviation introduces position estimation error within the sensorless control framework, resulting in a certain angle of deviation between the virtual synchronous coordinate system and the actual synchronous coordinate system during coordinate transformation. Consequently, traditional online inductor parameter identification schemes become ineffective. To achieve effective online inductor parameter identification in sensorless permanent magnet synchronous motor control schemes, it is necessary to decouple the coupling effect between position estimation error and inductor parameter deviation. Summary of the Invention

[0004] To address the shortcomings of the existing technology, the present invention provides a method for online identification of the inductance parameters of a permanent magnet synchronous motor based on position decoupling. This method effectively implements online identification of inductance parameters in a sensorless control framework without the need for additional signal injection, resulting in high estimation accuracy and significantly improved system stability and robustness. Initial stator resistance, stator inductance, and permanent magnet flux parameters are first obtained through offline parameter identification. The position compensation angle is then calculated based on the theoretical and actual back-electromotive force values to achieve decoupling. A recursive total least squares method is then used to identify the inductance online. The effectiveness of the present invention has been verified through simulation.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is: Step 1: Obtain the main parameters of the surface-mount permanent magnet synchronous motor, including resistance, through offline parameter identification scheme in the no-load state of the motor Permanent magnet flux and inductance Estimated rotor position based on permanent magnet synchronous motor sensorless control algorithm With estimated speed Construct a virtual synchronous rotating coordinate system γδ. Step 2: Establish the steady-state voltage equation in the virtual synchronous rotating coordinate system and extract the estimated value of the back electromotive force term through the observer Step 3: According to the back electromotive force of the δ axis in the virtual coordinate system and theoretically estimated back EMF Calculate the rotor position estimation deviation angle using the arc cosine formula And the estimated rotor position Make compensation. Step 4: Based on the estimated rotor compensation position obtained in step 3, the current and voltage in the three-phase fixed coordinate system are transformed to obtain the actual dq synchronous rotating coordinate system, and the discrete current equation is constructed based on the permanent magnet synchronous motor discrete model. Step 5: Use the recursive total least squares method to realize the online identification of the inductance parameters of the permanent magnet synchronous motor under sensorless control.

[0006] A preferred technical solution: In step 1, the actual inductance of the surface-mounted permanent magnet synchronous motor is affected by the magnetic saturation effect and will shift during operation. Therefore, there is an error between the inductance parameter obtained by offline identification and the actual inductance, and the relationship between the two can be expressed as: Among them L s is the actual inductance value. In the surface-mounted permanent magnet synchronous motor, there is L s =L d =L q , To estimate the inductance value offline, ΔL s is the inductor parameter error in the sensorless control observer.

[0007] A preferred technical solution: In step 1, the offset of the inductance parameter increases the rotor position estimation error of the sensorless control algorithm. This error can be expressed as: where θ e is the actual rotor position, Δθ e is the rotor estimation error, the estimated speed is the differential signal of the estimated position, which is not affected by the static error, so The voltage equation in the virtual synchronous rotating coordinate system γδ can be expressed as: Where u γ 、u δ are the motor virtual direct-axis and quadrature-axis voltages, i γ 、i δ are the motor virtual direct-axis and quadrature-axis currents, are the estimated values of the motor's virtual direct-axis and quadrature-axis back-electromotive force, respectively.

[0008] A better technical solution: In steps 1 and 2, it is assumed that the stator resistance and permanent magnet flux are only affected by temperature and have small fluctuations. The accurate values of the resistance and flux parameters can be obtained through offline parameter identification. Consider the magnetic saturation effect error ΔL of the inductance parameter. s and the estimated position error Δθ e When , the γδ axis back electromotive force can be further expressed as:

[0009] A better technical solution: In step three, use "i d = 0" sensorless control scheme, i γ = 0. In steady-state operation, the differential term of the inductance deviation can be ignored, and formula (4) yields: In the virtual synchronous rotating coordinate system, So we have: It can be seen that as the inductance parameter error increases, the position estimation error will also increase, and the angle error can be calculated as follows:

[0010] A better technical solution: In step 4, the obtained position estimation deviation angle is used to compensate for sensorless control. Since the arc cosine is used for calculation, Δθ eThere are two solutions, one positive and one negative. Considering the magnetic saturation effect will reduce the inductance, the estimated inductance is too large, so Δθ e Taking its positive value, after position correction, the rotor position can be expressed as: in The estimated rotor position after compensation is obtained, thereby achieving the decoupling of the position error and the inductance parameter error.

[0011] A preferred technical solution: In step 4, the estimated rotor position after compensation is used to re-transform the coordinates, and a discrete model of the surface-mounted permanent magnet synchronous motor in a virtual coordinate system is constructed. The model can be expressed as: Among them, T s is the sampling period.

[0012] A preferred technical solution: In step 5, a recursive total least squares algorithm is used to realize online identification of inductance parameters. The recursive total least squares algorithm can be expressed as: Where η is the learning rate and ε is the residual error. Substituting the discrete model into the recursive total least squares algorithm, we can obtain: According to the recursive total least squares algorithm, the inductance parameters can be quickly and effectively identified online while removing the coupling between position error and inductance parameters.

[0013] Through the above technical solution, the beneficial effects of the present invention are: Based on the traditional online inductor parameter identification method, the present invention addresses the coupling problem between sensorless control and inductor parameter identification and proposes a solution based on decoupling position correction. This solves the problem of inaccurate inductor parameter identification under the sensorless control framework and does not require the injection of additional high-frequency signals, effectively improving the accuracy of the system model. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The present invention is further described below with reference to the accompanying drawings and examples: Figure 1 This is a steady-state vector diagram when the stator inductance deviates; Figure 2 Schematic diagram of the coupling effect between parameter identification and sensorless rotor position estimation; Figure 3 This is the block diagram of the rotor position decoupling compensation algorithm; Figure 4The block diagram of sensorless control of permanent magnet synchronous motor using the proposed online inductance parameter identification algorithm; Figure 5 In the sensorless control mode, the initial value of the stator inductance is offset. Result curve of traditional online inductor parameter identification; Figure 6 In the sensorless control mode, the initial value of the stator inductance is offset. When , the online inductance parameter identification result curve after the decoupling position is corrected; Figure 7 In the sensorless control mode, the stator inductance is offset during operation, and the actual inductance is determined by L s0 =0.672mH, due to the magnetic saturation effect it is reduced to L s =0.56mH. Online inductance identification results with and without the decoupling position correction algorithm enabled. DETAILED DESCRIPTION

[0015] like Figure 1-Figure 7 As shown, this embodiment provides a method for online identification of inductance parameters of a permanent magnet synchronous motor based on position decoupling, comprising the following steps:

[0016] The main parameters of the surface-mounted permanent magnet synchronous motor, including resistance, are obtained through offline parameter identification scheme under the no-load state of the motor. Permanent magnet flux and inductance Estimated rotor position based on permanent magnet synchronous motor sensorless control algorithm With estimated speed Construct a virtual synchronous rotating coordinate system γδ as Figure 1 shown.

[0017] During operation, the actual inductance of a surface-mounted permanent magnet synchronous motor is affected by the magnetic saturation effect and will shift during operation. Therefore, there is an error between the inductance parameters obtained by offline identification and the actual inductance. The relationship between the two can be expressed as: Among them L s is the actual inductance value. In the surface-mounted permanent magnet synchronous motor, there is L s =L d =L q , To estimate the inductance value offline, ΔL s is the inductance parameter error in the sensorless control observer. When the inductance parameter is offset, the rotor position estimation of the sensorless control algorithm will have errors, such as Figure 1 The error shown in can be expressed as: where θe is the actual rotor position, Δθ e is the rotor estimation error, and the estimated speed is the differential signal of the estimated position, which is not affected by the static error. Therefore,

[0018] Establish the steady-state voltage equation in the virtual synchronous rotating coordinate system, and extract the estimated value of the back electromotive force term through the observer The voltage equation in the virtual synchronous rotating coordinate system γδ can be expressed as: Where u γ 、u δ are the motor virtual direct-axis and quadrature-axis voltages, i γ 、i δ are the motor virtual direct-axis and quadrature-axis currents, are the estimated values of the motor's virtual direct-axis and quadrature-axis back-electromotive force, respectively.

[0019] Usually, the stator resistance and permanent magnet flux are only affected by temperature and have small fluctuations. The accurate values of resistance and flux parameters can be obtained through offline parameter identification. To solve the problem of lack of rank, the stator is considered. s and the estimated position error Δθ e When , the γδ axis back electromotive force can be further expressed as:

[0020] When using "i d = 0" sensorless control scheme, i γ = 0. In steady-state operation, the differential term of the inductance deviation can be ignored, and formula (17) yields: In the virtual synchronous rotating coordinate system, So we have: It can be seen that as the inductance parameter error increases, the position estimation error will also increase, and the angle error can be calculated as follows:

[0021] The obtained position estimation deviation angle is used to compensate for sensorless control. Since the arc cosine is used for calculation, Δθ e There are two solutions, one positive and one negative. Considering the magnetic saturation effect will reduce the inductance, the estimated inductance is too large, so Δθ e Taking its positive value, after position correction, the rotor position can be expressed as: in The estimated rotor position after compensation is obtained, thereby achieving the decoupling of the position error and the inductance parameter error.

[0022] The estimated rotor position after compensation is used to re-transform the coordinates, and a discrete model of the surface-mounted permanent magnet synchronous motor in a virtual coordinate system is constructed. The model can be expressed as: Among them, T s is the sampling period.

[0023] The recursive total least squares algorithm is used to realize the online identification of inductor parameters. The recursive total least squares method can be expressed as: Where η is the learning rate and ε is the residual error. Substituting the discrete model into the recursive total least squares algorithm, we can obtain: According to the recursive total least squares algorithm, the inductance parameters can be quickly and effectively identified online while removing the coupling between position error and inductance parameters.

[0024] Table 1 Actual parameters of permanent magnet synchronous motor

[0025] Table 1 shows the basic parameters of the experimental motor

[0026] In order to verify the effectiveness of the proposed scheme based on decoupling position correction and traditional online inductance parameter identification, the inductance parameter identification scheme was tested with a sensorless control scheme under the conditions of a command speed of 1500 rpm and 50% rated load. The nominal value of the stator inductance of the motor body is Observer initial estimated inductance The initial resistance and flux linkage are both actual values. The scheme without position decoupling correction and the scheme with position decoupling correction are used to perform online parameter identification of the inductor. The learning rate parameter η in RTLS is 0.02.

[0027] like Figure 5 and Figure 6 The figures show the inductance identification results without and with the position correction scheme. Due to the coupling effect between the position error and the inductance identification results, even if the other electrical parameters are accurately fixed, the estimated inductance parameters still converge to the initial inductance. It shows that the identification of the inductance parameters is invalid at this time. When the position correction scheme is adopted, the position error caused by the inductance mismatch is first compensated, so that the estimated inductance can be effectively estimated. The estimated inductance in the steady state is 5.651mH, and the maximum error is 0.19%. Therefore, it is proved that the proposed method has high identification accuracy.

[0028] In order to verify the actual effect of the proposed online inductance parameter identification scheme based on position correction in sensorless control system, experiments were carried out at 1500rpm and 50% rated load. The initial inductance parameters were set to During the experiment, the control was switched from sensored control to sensorless control at 1s. During the steady-state process, the inductor parameters were identified online and updated at 3s. Figure 7 shown.

[0029] From the simulation results, it can be seen that in the period from 0 to 1s, inductive control is used, and the d-axis current is controlled to 0. Due to the existence of the initial error of the inductance parameter, there is an error between the estimated rotor position and the actual position. This error also causes the result of the inductance identification to converge to an erroneous value. After switching to sensorless control at 1s, the estimated position error causes a significant speed drop during the switching process. Under the modulation of the speed loop, the speed quickly recovers to 1500rpm, but the d-axis current can be clearly seen from the current waveform. Between 1 and 3s, the proposed position correction module first corrects the estimated position so that it converges to 0 after compensation. The estimated inductance is given through the online inductance parameter identification module and the cascaded mean filter. At 3s, the inductor parameters are updated, and the estimated position error after the parameter update also converges to 0. At this time, the d-axis current can also be correctly controlled to 0, reducing the loss.

[0030] The above description is only a preferred embodiment of the present invention, and therefore cannot be used to limit the scope of the present invention. In other words, equivalent changes and modifications made according to the contents of the present invention and the specification should all fall within the scope of the present invention.

Claims

1. A method for online identification of inductance parameters of a permanent magnet synchronous motor based on position decoupling, characterized in that: The following steps are involved: Step 1: Obtain the main parameters of the surface-mount permanent magnet synchronous motor, including resistance, through offline parameter identification scheme in the no-load state of the motor Permanent magnet flux and inductance Estimated rotor position based on permanent magnet synchronous motor sensorless control algorithm With estimated speed Construct a virtual synchronous rotating coordinate system γδ. Step 2: Establish the steady-state voltage equation in the virtual synchronous rotating coordinate system and extract the estimated value of the back electromotive force term through the observer Step 3: According to the back electromotive force of the δ axis in the virtual coordinate system and theoretically estimated back EMF Calculate the rotor position estimation deviation angle using the arc cosine formula And the estimated rotor position Make compensation. Step 4: Based on the estimated rotor compensation position obtained in step 3, the current and voltage in the three-phase fixed coordinate system are transformed to obtain the actual dq synchronous rotating coordinate system, and the discrete current equation is constructed based on the permanent magnet synchronous motor discrete model. Step 5: Use the recursive total least squares method to realize the online identification of the inductance parameters of the permanent magnet synchronous motor under sensorless control.

2. The method for online identification of inductance parameters of a permanent magnet synchronous motor based on position decoupling according to claim 1 is characterized in that: In step 1, the actual inductance of the surface-mounted permanent magnet synchronous motor is affected by the magnetic saturation effect and will shift during operation. Therefore, there is an error between the inductance parameter obtained by offline identification and the actual inductance. The relationship between the two can be expressed as: Among them L s is the actual inductance value. In the surface-mounted permanent magnet synchronous motor, there is L s =L d =L q , To estimate the inductance value offline, ΔL s is the inductor parameter error in the sensorless control observer.

3. The method for online identification of inductance parameters of a permanent magnet synchronous motor based on position decoupling according to claim 1 is characterized in that: In step 1, the offset of the inductance parameter increases the rotor position estimation error of the sensorless control algorithm. This error can be expressed as: where θ e is the actual rotor position, Δθ e is the rotor estimation error, the estimated speed is the differential signal of the estimated position, which is not affected by the static error, so The voltage equation in the virtual synchronous rotating coordinate system γδ can be expressed as: Where u γ 、u δ are the motor virtual direct-axis and quadrature-axis voltages, i γ 、i δ are the motor virtual direct-axis and quadrature-axis currents, are the estimated values of the motor's virtual direct-axis and quadrature-axis back-electromotive force, respectively.

4. The method for online identification of inductance parameters of a permanent magnet synchronous motor based on position decoupling according to claim 1 is characterized in that: In steps 1 and 2, it is assumed that the stator resistance and permanent magnet flux are only affected by temperature and have small fluctuations. The accurate values of the resistance and flux parameters can be obtained through offline parameter identification. Considering the magnetic saturation effect error ΔL of the inductance parameter s and the estimated position error Δθ e When , the γδ axis back electromotive force can be further expressed as:

5. The method for online identification of inductance parameters of a permanent magnet synchronous motor based on position decoupling according to claim 1 is characterized in that: In step three, use "i d = 0" sensorless control scheme, i γ = 0. In steady-state operation, the differential term of the inductance deviation can be ignored, and formula (4) yields: In the virtual synchronous rotating coordinate system, So we have: It can be seen that as the inductance parameter error increases, the position estimation error will also increase, and the angle error can be calculated as follows:

6. The method for online identification of inductance parameters of a permanent magnet synchronous motor based on position decoupling according to claim 1, characterized in that: In step 4, the obtained position estimation deviation angle is used to compensate for sensorless control. Since the arc cosine is used for calculation, Δθ e There are two solutions, one positive and one negative. Considering the magnetic saturation effect will reduce the inductance, the estimated inductance is too large, so Δθ e Taking its positive value, after position correction, the rotor position can be expressed as: in The estimated rotor position after compensation is obtained, thereby achieving the decoupling of the position error and the inductance parameter error.

7. The method for online identification of inductance parameters of a permanent magnet synchronous motor based on position decoupling according to claim 1, characterized in that: In step 4, the estimated rotor position after compensation is used to re-transform the coordinates and construct a discrete model of the surface-mounted permanent magnet synchronous motor in the virtual coordinate system. The model can be expressed as: Among them, T s is the sampling period.

8. The method for online identification of inductance parameters of a permanent magnet synchronous motor based on position decoupling according to claim 1, characterized in that: It is characterized by: In step 5, a recursive total least squares algorithm is used to realize the online identification of inductor parameters, which can be expressed as: Where η is the learning rate and ε is the residual error. Substituting the discrete model into the recursive total least squares algorithm yields: According to the recursive total least squares algorithm, the inductance parameters can be quickly and effectively identified online while decoupling the position error from the inductance parameters.

Images