Decoupling control method for improving inductance parameter identification precision of permanent magnet synchronous motor

Abstract

The invention discloses a decoupling control method for improving inductance parameter identification precision of a permanent magnet synchronous motor, and relates to the technical field of inductance parameter identification of permanent magnet synchronous motors. According to the decoupling control method for improving the inductance parameter identification precision of the permanent magnet synchronous motor, a high-frequency cosine signal is injected into a virtual shaft system, a formula required by an identification algorithm is extracted, and stray terms and coupling terms are removed through a decoupling link; and through a discrete Fourier decomposition algorithm, filtering out a responded identification signal through discrete Fourier transform to obtain a component related to a high-frequency signal, and obtaining an inductance response curve so as to calculate identification inductance. The invention provides a decoupling algorithm based on a virtual shafting, the decoupling algorithm is an improvement of an original identification algorithm, the front end of the original identification algorithm is decoupled, the influence of an interference term and a coupling term is eliminated, the problems of insufficient precision and poor tracking performance of the original method are solved, and the effect of reducing errors is achieved; the robustness of the whole system and the tracking performance of motor inductance parameters are greatly improved.

Description

Technical Field

[0001] The invention relates to the technical field of permanent magnet synchronous motor inductance parameter identification, and specifically to an optimization identification strategy for online inductance identification of a permanent magnet synchronous motor based on virtual shaft voltage injection. Background Art

[0002] Permanent magnet synchronous motors are widely used in industrial drives, transportation, household appliances, and renewable energy due to their high power density, wide speed range, and good control performance. In the control strategy of permanent magnet synchronous motor, motor control requires a more accurate motor parameter model to establish the control algorithm to achieve better control performance. At present, there are many algorithms for identifying the inductance parameters of permanent magnet synchronous motor, such as static test method, linear least squares method (RLS), fuzzy logic and neural network method and AC test method. Among them, the static test method is to apply DC voltage to the motor in a stationary state and observe the current change to infer the direct-axis and quadrature-axis inductance. However, in actual motors, the inductance of the motor will change with the running time of the motor, which leads to a large error between the inductance of the running motor and the identified inductance in the static test method, resulting in inaccurate control algorithm parameters and poor overall control performance. Fuzzy logic and neural network methods and linear least squares method require the collection of a large amount of motor operation data, analyze and learn a large amount of data to train the model, and then output the estimated inductance value through the model. This method requires the acquisition of a large amount of motor operation data, and it is necessary to obtain the operation data based on inductive control to output the estimated inductance value. In the AC test method, the traditional high-frequency signal injection method is based on the impedance model of the motor inductance at high frequency to identify the inductance by injecting a high-frequency signal into the synchronously rotating shaft system of the motor. However, when the rotor position observation error angle is large, the error of this method is large.

[0003] During the actual operation of the motor, the motor parameters are unknown and will change over time. The control algorithm for the permanent magnet synchronous motor requires a relatively accurate inductance value. In the current control algorithm, online identification of parameters is rarely involved, and most of them are offline identification cases, or a large amount of motor operation data is required to build a model, which is unrealistic under the actual working conditions of the motor. This requires an identification algorithm that can observe the inductance online to optimize the control performance of the permanent magnet synchronous motor.

[0004] The traditional high-frequency signal injection method is mainly based on the high-frequency impedance model of the motor. A high-frequency signal is injected into the dq axis of the synchronous rotating shaft system of the motor, and then the signal is separated from the main frequency signal of the motor, and then the separated signal is analyzed and calculated to achieve the purpose of identification. However, due to the error between the observed speed and rotor angle in the motor control algorithm and the actual speed and rotor angle, when the difference between the observed rotor position angle and the actual rotor position angle is large, the error between the inductance estimated by this traditional method and the actual inductance is large. The virtual shaft system rotation signal injection method is not affected by the position error due to the superiority of the principle, but the algorithm is affected by the cross-coupling effect of the motor voltage and has a certain identification error. Summary of the invention

[0005] The present invention provides a decoupling control method for improving the identification accuracy of inductance parameters of a permanent magnet synchronous motor, which overcomes the shortcomings described in the background technology.

[0006] The technical solution adopted by the present invention to solve its technical problem is: A decoupling control method for improving the inductance parameter identification accuracy of a permanent magnet synchronous motor comprises the following steps: Step 1: Perform coordinate transformation according to the equivalent model of the permanent magnet synchronous motor in the abc stationary three-phase coordinate system to obtain the equivalent model of the motor in the dq rotating coordinate system; Step 2: construct a virtual axis system rotating relative to the motor dq axis, and analyze and obtain the change matrix of the transformation from the motor dq synchronous rotating axis to the virtual axis system; Step 3: perform coordinate transformation on the equivalent model of the motor under the dq axis using the coordinate change matrix obtained in step 2, and decompose the motor equivalent model into the virtual axis system; Step 4, injecting high-frequency cosine signal into the virtual axis system; Step 5: Convert the motor equivalent model obtained in step 3 Separate them out and extract the formulas needed for the identification algorithm; Step 6: Remove the stray terms and coupling terms through the decoupling link; Step 7: Through the discrete Fourier decomposition algorithm, the response identification signal is filtered out through discrete Fourier transform to obtain the component of the high-frequency signal, and then calculated according to formula (13) to obtain the inductance response curve to calculate the identification inductance; (13) In the formula, for The amplitude of the high-frequency current component on the axis, is the high frequency signal frequency, The high-frequency voltage component obtained after discrete Fourier decomposition is then decoupled by the algorithm, and finally simplified to The relationship between the direct-axis inductance and the quadrature-axis inductance of the permanent magnet synchronous motor.

[0007] A preferred technical solution: In step 1, the equivalent model of the motor in the dq rotating coordinate system is: (1) In the formula, They are the motor stator resistance, motor direct-axis inductance and motor quadrature-axis inductance, They are the motor direct-axis voltage, motor quadrature-axis voltage, motor direct-axis current and motor quadrature-axis current, respectively. are the motor rotor electrical angular velocity and permanent magnet flux linkage respectively.

[0008] A preferred technical solution: In step 2, the transformation matrix of the motor equivalent model from the synchronous rotating shaft system to the virtual shaft system is formula (2): (2) In the formula, are the phase angle of the γ axis and the phase angle of the d axis respectively.

[0009] A preferred technical solution: In step 3, the equivalent model of the permanent magnet synchronous motor under the virtual shaft system is obtained as follows: (3) A better technical solution: The equivalent model of the permanent magnet synchronous motor in formula (3) under the virtual shaft system is expanded as follows: (4) In the formula, are the voltage and current on the γ-axis and δ-axis respectively, p is the differential operator.

[0010] A better technical solution: The high-frequency cosine signal in step 4 is: (5) In the formula, The high frequency signal is injected into , The components on the axis and the amplitude of the high-frequency voltage, are the frequency, angular velocity and initial phase of the signal respectively.

[0011] A better technical solution: Change the formula (4) Analyzing them separately yields: (6) (7) (8) (9) (10) In the formula, They are identification error, Shaft current loss, Shaft current coupling losses.

[0012] A preferred technical solution: In step 5, the formula that the identification algorithm needs to extract is: (11) A better technical solution: The discrete Fourier decomposition algorithm in step 7 adopts discrete Fourier transform, and its formula is: (12) In the formula, X amp is the specified harmonic amplitude obtained in the DFT algorithm, N is the number of points of the discrete signal in the time domain, n is the number of the discrete signal in the time domain, m is the number of the frequency domain signal, x ( n )for The nth sample value in the Fourier transform process.

[0013] A better technical solution: In step 7, simplify to get The relationship between the direct-axis inductance and the quadrature-axis inductance of the permanent magnet synchronous motor is: (14) In the formula, for The identification inductance on the axis can be obtained according to formula (14): about and The inductance curve is then calculated based on the saliency of the salient-pole motor; in the salient-pole permanent magnet synchronous motor The inductance value will be less than Inductance value, the lowest point of the curve is The highest point of the curve is The inductance value.

[0014] By adopting the above technical solution, the beneficial effects of the present invention are: The present invention proposes a decoupling algorithm based on a virtual axis system, which is an improvement on the original identification algorithm. It decouples the front end of the original identification algorithm, eliminates the influence of interference terms and coupling terms, solves the problems of insufficient accuracy and poor tracking performance of the original method, achieves the effect of reducing errors, and greatly improves the robustness of the entire system and the tracking performance of the motor inductance parameters. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The present invention will be further described below in conjunction with the accompanying drawings and embodiments: Figure 1 This is a schematic diagram of the motor coordinate system; Figure 2 This is the simulation diagram of γ-axis voltage decoupling; Figure 3 It is the inductance parameter identification curve; Figure 4 This is the block diagram of the permanent magnet synchronous motor control and identification algorithm; Figure 5 This is the inductance identification curve without adding the optimization algorithm; Figure 6 This is the inductance identification curve after adding the optimization algorithm; Figure 7 It is the inductance identification curve without adding the second set of inductance parameters of the optimization algorithm; Figure 8 This is the inductance identification curve under the second set of motor inductance parameters after adding the optimization algorithm; Fig. 9 When the optimization algorithm is not added, the inductance Inductance identification curve when the inductance suddenly changes from 12mH to 15mH at 1.5s; Fig.10 After adding the optimization algorithm, the inductor Inductance identification curve when the inductance suddenly changes from 12mH to 15mH at 1.5s; Fig.11 This is the inductance parameter identification curve when the motor load suddenly changes from 5N / m to 10N / m before adding the optimization algorithm; Fig.12 Inductance identification curve when the load changes suddenly from 5N / m to 10N / m after adding the optimization algorithm. DETAILED DESCRIPTION

[0016] like Figure 1-Figure 12 As shown, this embodiment provides a decoupling control method for improving the identification accuracy of the inductance parameters of a permanent magnet synchronous motor, comprising the following steps: (1) According to the equivalent model of the permanent magnet synchronous motor in the abc stationary three-phase coordinate system, coordinate transformation is performed to obtain the equivalent model of the motor in the dq rotating coordinate system: (1) In the formula, They are the motor stator resistance, motor direct-axis inductance and motor quadrature-axis inductance, They are the motor direct-axis voltage, motor quadrature-axis voltage, motor direct-axis current and motor quadrature-axis current, respectively. are the motor rotor electrical angular velocity and permanent magnet flux linkage respectively.

[0017] The equivalent model of the permanent magnet synchronous motor dq rotating coordinate system is transformed into the virtual axis system through coordinate transformation, as shown in Figure 1 The transformation matrix of the motor equivalent model from the synchronous rotating axis system to the virtual axis system is as follows: (2) In the formula, are the phase angle of the γ axis and the phase angle of the d axis respectively.

[0018] Through coordinate transformation (such as formula (2)), the equivalent model of the permanent magnet synchronous motor in the virtual shaft system is obtained as shown in formula (3): (3) Expanding the coordinate transformation matrix in equation (3) yields: (4) are the voltage and current on the γ-axis and δ-axis respectively, p is the differential operator.

[0019] A high-frequency cosine signal is injected into the virtual axis system, and the signal is expressed as formula (5): (5) In the formula, The high frequency signal is injected into , The components on the axis and the amplitude of the high-frequency voltage, are the frequency, angular velocity and initial phase of the signal respectively.

[0020] Will The separate analyses include: (6) (7) (8) (9) (10) The formula that the identification algorithm needs to extract is: (11) in They are identification error, Shaft current loss, Shaft current coupling loss, in which the injected high-frequency signal is only However, due to the voltage model of the motor under the virtual shaft system, it is known that cross-coupling effect will inevitably occur. The interference voltage is generated under the shaft. Since the equation for inductance identification is equation (11), it can be seen from equation (6) that these stray terms (equation (8)) and coupling terms (such as equation (9)) will affect the identification results, thereby introducing errors that affect the parameter identification accuracy. The traditional algorithm does not analyze these stray terms and coupling terms, but directly performs identification calculations on equation (6), introducing errors in these terms. This algorithm is designed as follows Figure 2 The decoupling link of the structure removes the stray terms and coupling terms, which can achieve a more accurate identification result.

[0021] (2) The discrete Fourier decomposition (DFT) algorithm uses discrete Fourier transform as shown in formula (12) to filter out the response identification signal through discrete Fourier transform to obtain the component of the high-frequency signal. Then, it is calculated according to formula (13) to obtain the inductance response curve and thus calculate the identification inductance: (12) In the formula, X amp is the specified harmonic amplitude obtained in the DFT algorithm, N is the number of points of the discrete signal in the time domain, n is the number of the discrete signal in the time domain, m is the number of the frequency domain signal, x ( n )for The nth sampling value in the Fourier transform process. The amplitude of each high-frequency signal component is decomposed by the DFT algorithm. (13) In the formula, for The amplitude of the high-frequency current component on the axis, is the high frequency signal frequency, The high-frequency voltage component obtained after DFT decomposition is then decoupled by the algorithm, and finally simplified to The relationship between the direct-axis inductance and the quadrature-axis inductance of the permanent magnet synchronous motor is: (14) In the formula, for The identification inductance on the axis can be obtained according to formula (14): about and Inductance curve, and then according to the saliency of the salient pole motor: In the salient pole permanent magnet synchronous motor The inductance value will be less than Inductance value, so the lowest point of the curve is Inductance value, similarly, the highest point of the curve is The inductance value is shown in the schematic diagram. Figure 3 .

[0022] The overall block diagram of the permanent magnet synchronous motor control based on the virtual shaft system high frequency injection optimization algorithm is as follows: Figure 4 , the control method is based on permanent magnet synchronous motor Control method: The optimization algorithm of high-frequency sinusoidal voltage injection based on virtual shaft system is established.

[0023] Table 1 Basic parameters of permanent magnet synchronous motor

[0024] Table 1 gives the basic parameters of the experimental motor based on the virtual shaft rotation high-frequency voltage injection method.

[0025] Figure 5 This is the identification curve without the coupling compensation optimization algorithm. The identification result of the experimental curve is The value of is 4.853mH and the error value of identification is 6.67%. The value is 12.17mH, the error is 1.41%, and Figure 6 The inductance identification curve after adding the optimization algorithm is given, and the experimental results is 4.974mH, The error is 4.34% and 0% respectively, and the recognition accuracy under static inductance is greatly improved.

[0026] Figure 7 , 8 is the inductance identification waveform under the second set of motor inductance parameters, where the motor inductance parameters From 5.2mH to 8mH, L q 12mH, before adding the optimization algorithm , They are 8.342mH and 11.32mH respectively, and the errors are 4.275% and 5.67% respectively. , They are 8.108mH and 11.72mH respectively, and the identification errors are 1.35% and 2.83% respectively. This shows that under the condition of changing the initial parameters of the motor, the optimization algorithm can greatly reduce the error of inductance parameter identification and improve the identification accuracy.

[0027] Fig. 9 , 10 is the inductance identification waveform under the sudden change of motor inductance parameters, where The parameter suddenly changes from 12mH to 15mH at 1.5s. The experimental results without adding the optimization algorithm are as follows: , The identification errors are 10.62% and 6.8% respectively. The experimental results after adding the optimization algorithm are as follows: , , the identification errors are 6.06% and 2.53% respectively. The experimental results show that after adding the optimization algorithm, the inductance identification curve waveform is more sensitive to the inductance mutation, and the identification of the mutation inductance is more accurate.

[0028] Fig.11 , 12 This is the inductance identification waveform under the sudden change of motor load, where the rated load of the permanent magnet synchronous motor is 5N / m. In the figure, the load suddenly changes to 10N / m at 4 seconds. By comparing the waveforms of the two figures, it can be concluded that the identification result of the inductance identification curve of the sudden change of load after adding the optimization algorithm is closer to the true value. Therefore, the inductance identification curve after adding the optimization algorithm has better stability to the sudden change of load, and the identification accuracy of the inductance is higher when the load suddenly changes.

[0029] 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. That is, equivalent changes and modifications made according to the patent scope of the present invention and the contents of the specification should still fall within the scope of the present invention.

Claims

1. A decoupling control method for improving the inductance parameter identification accuracy of a permanent magnet synchronous motor, characterized in that: The following steps are involved: Step 1: Perform coordinate transformation according to the equivalent model of the permanent magnet synchronous motor in the abc stationary three-phase coordinate system to obtain the equivalent model of the motor in the dq rotating coordinate system; Step 2: construct a virtual axis system rotating relative to the motor dq axis, and analyze and obtain the change matrix of the transformation from the motor dq synchronous rotating axis to the virtual axis system; Step 3: perform coordinate transformation on the equivalent model of the motor under the dq axis using the coordinate change matrix obtained in step 2, and decompose the motor equivalent model into the virtual axis system; Step 4, injecting high-frequency cosine signal into the virtual axis system; Step 5: Convert the motor equivalent model obtained in step 3 Separate them out and extract the formulas needed for the identification algorithm; Step 6: Remove the stray terms and coupling terms through the decoupling link; Step 7: Through the discrete Fourier decomposition algorithm, the response identification signal is filtered out through discrete Fourier transform to obtain the component of the high-frequency signal, and then calculated according to formula (13) to obtain the inductance response curve to calculate the identification inductance; (13) ， In the formula, for The amplitude of the high-frequency current component on the axis, is the high frequency signal frequency, The high-frequency voltage component obtained after discrete Fourier decomposition is then decoupled by the algorithm, and finally simplified to The relationship between the direct-axis inductance and the quadrature-axis inductance of the permanent magnet synchronous motor.

2. The decoupling control method for improving the inductance parameter identification accuracy of a permanent magnet synchronous motor according to claim 1 is characterized in that: In step 1, the equivalent model of the motor in the dq rotating coordinate system is: (1) ， In the formula, They are the motor stator resistance, motor direct-axis inductance and motor quadrature-axis inductance, They are the motor direct-axis voltage, motor quadrature-axis voltage, motor direct-axis current and motor quadrature-axis current, respectively. are the motor rotor electrical angular velocity and permanent magnet flux linkage respectively.

3. The decoupling control method for improving the inductance parameter identification accuracy of a permanent magnet synchronous motor according to claim 1, characterized in that: In step 2, the transformation matrix of the motor equivalent model from the synchronous rotating axis system to the virtual axis system is formula (2), (2) ， In the formula, are the phase angle of the γ axis and the phase angle of the d axis respectively.

4. The decoupling control method for improving the inductance parameter identification accuracy of a permanent magnet synchronous motor according to claim 1, characterized in that: In step 3, the equivalent model of the permanent magnet synchronous motor under the virtual shaft system is obtained as follows: (3) 。 5. The decoupling control method for improving the inductance parameter identification accuracy of a permanent magnet synchronous motor according to claim 4 is characterized in that: The equivalent model of the permanent magnet synchronous motor in formula (3) under the virtual shaft system is expanded as follows: (4) ， In the formula, are the voltage and current on the γ-axis and δ-axis respectively, p is the differential operator.

6. The decoupling control method for improving the inductance parameter identification accuracy of a permanent magnet synchronous motor according to claim 1, characterized in that: The high frequency cosine signal in step 4 is: (5) , In the formula The high frequency signal is injected into , The components on the axis and the amplitude of the high-frequency voltage, are the frequency, angular velocity and initial phase of the signal respectively.

7. The decoupling control method for improving the identification accuracy of inductance parameters of a permanent magnet synchronous motor according to claim 5, characterized in that: In formula (4), Analyzing them separately yields: (6) ， (7) ， (8) ， (9) ， (10) ， In the formula, They are identification error, Shaft current loss, Shaft current coupling losses.

8. The decoupling control method for improving the identification accuracy of the inductance parameters of a permanent magnet synchronous motor according to claim 7 is characterized in that: In step 5, the formula that the identification algorithm needs to extract is: (11)。 9. The decoupling control method for improving the identification accuracy of inductance parameters of a permanent magnet synchronous motor according to claim 1, characterized in that: The discrete Fourier decomposition algorithm in step 7 uses discrete Fourier transform, and its formula is: (12) ， In the formula, X amp is the specified harmonic amplitude obtained in the DFT algorithm, N is the number of points of the discrete signal in the time domain, n is the number of the discrete signal in the time domain, m is the number of the frequency domain signal, x ( n )for The nth sample value in the Fourier transform process.

10. The decoupling control method for improving the identification accuracy of inductance parameters of a permanent magnet synchronous motor according to claim 1, characterized in that: In step 7, we simplify to get The relationship between the direct-axis inductance and the quadrature-axis inductance of the permanent magnet synchronous motor is: (14) ， In the formula, for The identification inductance on the axis can be obtained according to formula (14): about and The inductance curve is then calculated based on the saliency of the salient-pole motor; in the salient-pole permanent magnet synchronous motor The inductance value is less than Inductance value, the lowest point of the curve is The highest point of the curve is The inductance value.

Images