Permanent magnet synchronous motor off-line inductance surface identification method based on sampling error correction

Abstract

The application discloses a permanent magnet synchronous motor offline inductance surface identification method based on sampling error correction and belongs to the technical field of permanent magnet synchronous motor offline inductance identification. The application is aimed at the problems that the existing offline inductance identification method does not fully consider the real-time change of inductance with working conditions and is easily affected by sampling errors in a small current interval. The application comprises the following steps: after the motor rotor is positioned at zero based on the d-axis direct current signal injection, square wave voltage signals with the same frequency and synchronous phase are synchronously injected into the motor dq axis; during the square wave voltage signal injection process, the discretized voltage and induced current are sampled corresponding to each sampling period, and the three-phase zero current points are removed; according to the voltage and induced current sampling results, the discretized flux linkage is solved through the discretized integral operation; then, the inductance identification results of the motor under different saturation working conditions are calculated; finally, the polynomial fitting algorithm is adopted for fitting to obtain the offline inductance surface identification result. The application is used for offline inductance surface identification.

Description

Technical Field

[0001] This invention relates to a method for identifying the offline inductance surface of a permanent magnet synchronous motor based on sampling error correction, and belongs to the field of offline inductance identification technology for permanent magnet synchronous motors. Background Technology

[0002] Permanent magnet synchronous motors (PMSMs) have wide applicability and significant application value in various industrial fields due to their high power and torque densities. Currently, the diverse and complex control methods for PMSMs often rely heavily on accurate motor inductance information and its real-time changing characteristics; therefore, PMSM inductance identification methods are of great importance.

[0003] Offline inductor identification via signal injection allows for flexible selection of signal injection methods, enabling effective information extraction and state decoupling for different motors under various operating conditions, resulting in high efficiency and accuracy. However, current inductor identification methods do not address sampling errors in the low-current range, and existing methods struggle to acquire inductor surfaces under different saturation conditions. To further improve the accuracy of inductor identification and expand the algorithm's general applicability, further research is needed on offline inductor surface identification methods for permanent magnet synchronous motors that take into account sampling errors.

[0004] Due to the electromagnetic properties of motor materials, complex cross-saturation and magnetic saturation phenomena occur under different operating conditions, causing significant variations in motor inductance. Traditional offline inductance identification algorithms typically only identify inductance under no-load conditions, failing to adequately consider real-time changes in inductance under different saturation conditions. Furthermore, traditional offline inductance identification methods rarely address sampling errors in the low-current range of the motor, leading to errors in the identification results. Therefore, to improve the accuracy of offline inductance identification and further expand its application scope, a method that considers sampling errors is needed. Summary of the Invention

[0005] To address the issues that existing offline inductance identification methods do not fully consider the real-time changes in inductance with operating conditions and are susceptible to sampling errors in the low current range, this invention provides an offline inductance surface identification method for permanent magnet synchronous motors based on sampling error correction.

[0006] The present invention provides an offline inductance surface identification method for permanent magnet synchronous motors based on sampling error correction, comprising:

[0007] After positioning the motor rotor at zero position based on the d-axis DC signal injection, a square wave voltage signal with the same frequency and phase is synchronously injected into the dq axis of the motor. The amplitude of the square wave voltage signal is adaptively adjusted according to a preset number of cycles.

[0008] The sampling period is determined according to a preset number of periods; during the square wave voltage signal injection process, the discretized voltage and induced current are sampled for each sampling period; the sampling results of the discretized voltage and induced current containing the three-phase zero current point are removed; based on the retained sampling results of the discretized voltage and induced current for each sampling period, the discretized flux linkage is solved by discretization integration.

[0009] Then, based on the discretized flux linkage, the inductance identification results of the motor under different saturation conditions are calculated;

[0010] Based on all inductor identification results, a polynomial fitting algorithm is used to fit the data to obtain offline inductor surface identification results.

[0011] According to the offline inductor surface identification method for permanent magnet synchronous motors based on sampling error correction of the present invention, the square wave voltage signal includes the d-axis square wave voltage signal u. d and q-axis square wave voltage signal u q :

[0012]

[0013]

[0014] In the formula U d U is the amplitude of the square wave voltage along the d-axis. q Let T be the amplitude of the square wave voltage along the q-axis, t be time, and k be the count of the injected period of the square wave voltage signal. d T is the frequency of the d-axis square wave voltage. q The frequency of the q-axis square wave voltage is given.

[0015] According to the offline inductance surface identification method for permanent magnet synchronous motors based on sampling error correction of the present invention, T d and T q All are 0.1-0.2 times the PWM frequency.

[0016] According to the offline inductance surface identification method for permanent magnet synchronous motors based on sampling error correction of the present invention, U d and U q The adaptive adjustment method is as follows:

[0017]

[0018] In the formula, l represents the number of amplitude combinations of the square wave voltage signal injected along the dq axis, l0 represents the total number of amplitude combinations of the square wave voltage signal injected along the dq axis, and U max This is the voltage amplitude limit value;

[0019] Each value of l corresponds to one sampling period.

[0020] The method for determining the three-phase zero-current point according to the offline inductance surface identification method for permanent magnet synchronous motors based on sampling error correction of the present invention includes:

[0021]

[0022] In the formula i a Let i be the phase A current. b Let i be the phase B current. c For phase C current, i d For the discretized induced current along the d-axis, i q To collect the discretized induced current along the q-axis, θ e To identify the initial rotor position of the motor offline, θ e =0.

[0023] According to the offline inductance surface identification method for permanent magnet synchronous motors based on sampling error correction of the present invention, the discretized flux linkage calculation method includes:

[0024]

[0025] In the formula ψ dn Let ψ be the discretized flux linkage along the d-axis for the nth sampling period. qn Let u be the q-axis discretized flux linkage for the nth sampling period. dn Let u be the d-axis square wave voltage signal of the nth sampling period. qn R is the q-axis square wave voltage signal of the nth sampling period. s For stator resistance, i dn For the discretized induced current along the d-axis acquired in the nth sampling period, i qn The discretized induced current along the q-axis is collected during the nth sampling period; n = 0, 1, ..., l0.

[0026] According to the offline inductance surface identification method for permanent magnet synchronous motors based on sampling error correction of the present invention, the sampling period is 2-5 PWM cycles.

[0027] According to the offline inductance surface identification method for permanent magnet synchronous motors based on sampling error correction of the present invention, the method for obtaining the inductance identification results of the motor under different saturation conditions is as follows:

[0028]

[0029] L d For the d-axis inductance identification results, L q For the q-axis inductance identification results, i da i is the average value of the discretized induced current along the d-axis between the current sampling period and the adjacent previous sampling period. qa It represents the average value of the discretized induced current along the q-axis between the current sampling period and the adjacent previous sampling period.

[0030] According to the offline inductance surface identification method for permanent magnet synchronous motors based on sampling error correction of the present invention, within the sampling period corresponding to the amplitude combination of square wave voltage signals injected into each pair of dq axes, i d and i q This forms a trajectory line on the dq-axis current plane;

[0031] Methods for obtaining offline inductor surface identification results include:

[0032]

[0033] In the formula a PQ b is the fitting parameter for the d-axis inductance surface. PQ The parameters for fitting the q-axis inductance surface are: P = 0, 1, 2, ..., P0, Q = 0, 1, 2, ..., Q0; P is the order of the d-axis polynomial, Q is the order of the q-axis polynomial, P0 is the maximum order of the d-axis polynomial, and Q0 is the maximum order of the q-axis polynomial.

[0034] According to the offline inductance surface identification method for permanent magnet synchronous motors based on sampling error correction of the present invention, a PQ and b PQ Solve using the recursive least squares method:

[0035] e = y - λ T ρ

[0036] y T =[L d (i d i q )](or [L q (i d i q )])

[0037]

[0038] or

[0039] In the formula, e is the least squares algorithm error matrix, y is the least squares algorithm output matrix, λ is the least squares algorithm feedback matrix, and ρ is the least squares algorithm parameter matrix.

[0040] The beneficial effects of the present invention are as follows: The method of the present invention achieves inductor identification based on amplitude adaptive square wave voltage injection, and realizes the correction of sampling error in the small current range and the fitting of the inductor surface through data sampling and fitting strategies.

[0041] This invention addresses the problems of existing offline inductance identification methods for permanent magnet synchronous motors, which fail to adequately consider real-time changes in inductance under varying operating conditions and are susceptible to sampling errors in the low-current range. The method combines discrete sampling with real-time inductance calculation during parameter identification. Furthermore, by comparing the injected signal with the three-phase zero-current range, it eliminates sampling errors in the low-current range. Simultaneously, it utilizes polynomial fitting to obtain the inductance surface under different saturation conditions, thus broadening the applicable operating range of offline inductance identification and improving its accuracy. Attached Figure Description

[0042] Figure 1 This is a schematic diagram of amplitude adaptive square wave voltage injection for the offline inductor surface identification method for permanent magnet synchronous motors based on sampling error correction as described in this invention;

[0043] Figure 2 This is a schematic diagram of the sampling of voltage, current, and magnetic flux during the inductor identification process in the method of this invention; ψ in the diagram d and ψ q Let ψ be the flux linkage along the dq axis. d.q Let i be the ordinate variable of the flux linkage along the dq axis. d.q Let u be the ordinate variable of the discretized induced current along the dq axis. d . q T represents the ordinate variable of the dq-axis square wave voltage signal. s The sampling period for signal sampling;

[0044] Figure 3 This is a block diagram of the inductor identification method in this invention; the diagram includes a voltage injection stage, an information sampling stage, and an inductor surface fitting stage; where U dc This refers to the bus voltage.

[0045] Figure 4 This refers to the positional relationship between the injected signal sampling point and the three-phase zero current in the method of this invention;

[0046] Figure 5 This is a schematic diagram of the inductor surface generation method based on polynomial fitting algorithm in the present invention.

[0047] Figure 6 This is the result of offline inductance surface identification along the d-axis;

[0048] Figure 7 This is the result of offline inductor surface identification along the q-axis. Detailed Implementation

[0049] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0050] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0051] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but this is not intended to limit the scope of the invention.

[0052] Specific Implementation Method 1: Combination Figures 1 to 7 As shown, this invention provides an offline inductance surface identification method for permanent magnet synchronous motors based on sampling error correction, including:

[0053] Based on the injection of DC signal along the d-axis, the motor rotor is positioned at zero. After determining the position of the motor's dq axis, in the open-loop case, based on PWM modulation, square wave voltage signals with the same frequency and phase are synchronously injected into the motor's dq axis. The amplitude of the square wave voltage signal is adaptively adjusted at preset cycle intervals to achieve scanning of the voltage amplitude value under the dq axis system.

[0054] The sampling period is determined according to a preset number of cycles; during the square wave voltage signal injection process, the discretized voltage and induced current are sampled for each sampling period; the sampling results of the discretized voltage and induced current that contain three-phase zero current points are removed; based on the retained sampling results of the discretized voltage and induced current for each sampling period, the discretized flux linkage is solved by discretized integration; during the signal injection process, it is determined in real time whether the data sampling points required for inductance solution contain three-phase zero current regions, in order to distinguish and eliminate inductance identification data affected by sampling errors.

[0055] Then, based on the discretized flux linkage, the inductance identification results of the motor under different saturation conditions are calculated;

[0056] Based on all inductor identification results, a polynomial fitting algorithm is used to fit the data to obtain offline inductor surface identification results.

[0057] This implementation method achieves inductor identification based on amplitude-adaptive square wave voltage injection. It also incorporates data sampling and fitting strategies to correct sampling errors in the low-current range, thereby obtaining the inductor surface identification result. During the inductor identification process, the RLS algorithm (Recursive Least Squares) is used to solve for the fitting polynomial parameters, based on real-time data sampling values, ultimately obtaining the complete inductor surface.

[0058] Furthermore, combined with Figure 1 As shown, the square wave voltage signal includes the d-axis square wave voltage signal u. d and q-axis square wave voltage signal u q :

[0059]

[0060]

[0061] In the formula U d U is the amplitude of the square wave voltage along the d-axis. q Let T be the amplitude of the square wave voltage along the q-axis, t be time, and k be the count of the injected period of the square wave voltage signal. d T is the frequency of the d-axis square wave voltage. q T is the frequency of the q-axis square wave voltage. d =T q .

[0062] As an example, T d and T q All are 0.1-0.2 times the PWM frequency.

[0063] By the given U d and U q Injection enables inductor identification under a set of states. To allow signal injection to simulate different motor operating conditions, i.e., to achieve different inductance values ​​under offline conditions... d and i q Motor inductance identification under combined conditions. An amplitude-adaptive voltage injection strategy is employed, U d and U q The adaptive adjustment method is as follows:

[0064]

[0065] In the formula, l represents the number of amplitude combinations of the square wave voltage signal injected along the dq axis, l0 represents the total number of amplitude combinations of the square wave voltage signal injected along the dq axis, and U max This is the voltage amplitude limit value, which must be lower than the voltage limits for the inverter and motor. A signal injection diagram is shown below. Figure 1 As shown.

[0066] Each value of l corresponds to one sampling period.

[0067] The injected voltage amplitude must be less than the voltage limits of the inverter and motor. During signal injection, selecting a value of 8 to 15 for l0 can keep the overall signal injection time within 20 seconds. As l increases, U... d and U q The i induced on the dq axis of the motor is generated by the combination of equally spaced distribution. d and i qSimilarly, the signals are evenly distributed in the dq axis coordinate system, which allows the injected signal range of the method of the present invention to cover different motor operating conditions.

[0068] The following section explains the method for correcting sampling errors in the low current range. Figure 4 The positional relationship between the sampling points of the injected dq-axis signal and the three-phase zero current is given, providing guidance for the correction of sampling errors in the low current range.

[0069] Combination Figure 4 As shown, in the inductor identification method using square wave signal injection, the parameter identification in the low current range is significantly affected by sampling errors due to the inverter's nonlinearity. Since the inverter's nonlinear effect is most pronounced near the three-phase zero current region, to compensate for sampling errors during offline inductor identification, it is necessary to determine the location corresponding to the zero current during parameter identification. Methods for determining the three-phase zero current point include:

[0070]

[0071] In the formula i a Let i be the phase A current. b Let i be the phase B current. c For phase C current, i d For the discretized induced current along the d-axis, i q To collect the discretized induced current along the q-axis, θ e To identify the initial rotor position of the motor offline, since the initial position of the motor is determined by DC injection along the d-axis, θ e =0. At this time, the distribution of the three-phase zero current positions on the dq-axis current plane is as follows: Figure 4 As shown, the angle between the zero current lines of each phase is 60°.

[0072] By combining the sampling method of motor voltage and current during the dq-axis square wave voltage injection process, it can be seen that the flux linkage can be obtained by solving the sampling points of two adjacent states, with a sampling interval of T. s By continuously sampling the dq-axis current, it is relatively easy to determine whether there is a three-phase zero-current point between two sampling points. It can be considered that when the three-phase zero current is directly located within the dq-axis current sampling interval, the inductance identification is significantly affected by sampling errors.

[0073] To eliminate the impact of sampling errors in the small current range on inductor identification, the current corresponding to all sampling points during the inductor identification process should be detected to determine whether the continuous sampling values ​​of the dq-axis current within the sampling interval include the three-phase zero current point. If so, the identification result for that current interval can be directly discarded, and the remaining sampling data can be combined to further fit the inductor surface. Since the sampling error only affects the calculation result within one sampling period, the error data accounts for a low proportion of the entire inductor data.

[0074] Combination Figure 3 As shown, the discretized flux linkage calculation method includes: based on the injection of a square wave voltage signal, an instantaneous induced current is generated in the dq axis of the motor. To achieve offline inductance calculation, the induced flux linkage of the motor dq axis is first calculated using the motor voltage and current information. In the continuous domain, the formula for calculating the dq axis flux linkage is as follows:

[0075]

[0076] In the formula, the flux linkage along the dq axis can be considered as a function of the current along the dq axis, expressed as ψ. d (i d i q ) and ψ q (i d i q ).

[0077] Since the actual motor control process is discrete, a discrete flux linkage identification strategy is required in the actual identification process. The solution formula is as follows:

[0078]

[0079] In the formula ψ dn Let ψ be the discretized flux linkage along the d-axis for the nth sampling period. qn Let u be the q-axis discretized flux linkage for the nth sampling period. dn Let u be the d-axis square wave voltage signal of the nth sampling period. qn R is the q-axis square wave voltage signal of the nth sampling period. s For stator resistance, i dn For the discretized induced current along the d-axis acquired in the nth sampling period, i qn The q-axis discretized induced current is collected in the nth sampling period; n is the discrete sampling count variable, n = 0, 1, ..., l0.

[0080] By adaptively adjusting the injection voltage and injection frequency, the value of the induced current in the motor can be controlled. Furthermore, different induced currents can simulate the magnetic circuit saturation state of the motor under different operating conditions. Therefore, the method of this invention can simultaneously inject a square wave with adaptive amplitude through the dq axis under offline operating conditions to obtain the inductance parameters of the motor under different saturation states.

[0081] To achieve dq axis inductance L d and L q The identification of u during signal injection. d u q i d i q ψ d and ψ qBy a fixed sampling period T s Real-time sampling. For example... Figure 2 As shown, to ensure sampling accuracy, the sampling period T can be set to... s It takes 2-5 PWM cycles. Considering the actual control characteristics of the motor, only the motor i... d ≤0 and i q Samples are taken from the ≥0 range.

[0082] L under discretized control d and L q The sampled values ​​i from the two sampling states can be directly obtained. d i q ψ d and ψ q The inductance was calculated. To eliminate pulsation errors in actual calculations, the average value of the center of the sampling period was taken as the final inductance solution. Simultaneously, the operating current corresponding to this inductance was also obtained from the average value of the two sampled state currents.

[0083] Combination Figures 2 to 7 As shown,

[0084] The method for obtaining the inductance identification results of the motor under different saturation conditions is as follows:

[0085]

[0086] L d For the d-axis inductance identification results, L q For the q-axis inductance identification results, i da i is the average value of the discretized induced current along the d-axis between the current sampling period and the adjacent previous sampling period. qa It represents the average value of the discretized induced current along the q-axis between the current sampling period and the adjacent previous sampling period.

[0087] During signal injection, to ensure the safety of the parameter identification method, it is necessary to ensure that the actual injected motor's i... d and i q The current should not exceed the maximum allowable current for the motor and controller. The overall offline inductance identification method is as follows: Figure 3 As shown.

[0088] The following section explains the method for obtaining inductor surfaces based on polynomial fitting:

[0089] Combination Figures 5 to 7 As shown, when U is selected d and U q During combined injection, the induced current i d and i q Real-time changes, during continuous sampling, can be achieved within a U d U q Obtain a series of i under combinationd i q The corresponding inductance value corresponds to a trajectory line on the dq-axis current plane. For different U... d U q Combined sampling can obtain a series of radioactive sampling trajectories, such as Figure 5 As shown. The inductance surface of the motor at any i d i q It is continuously integrable within the range, and the inductor surface can be regarded as a polynomial function with current as its domain.

[0090] Methods for obtaining offline inductor surface identification results include:

[0091]

[0092] In the formula a PQ b is the fitting parameter for the d-axis inductance surface. PQ The parameters for fitting the q-axis inductance surface are: P = 0, 1, 2, ..., P0, Q = 0, 1, 2, ..., Q0; P is the order of the d-axis polynomial, Q is the order of the q-axis polynomial, P0 is the maximum order of the d-axis polynomial, and Q0 is the maximum order of the q-axis polynomial. The dq-axis inductance L d and L q It can be viewed as a function of the dq-axis current, expressed as L in the formula. d (i d i q ) and L q (i d i q The specific values ​​of P0 and Q0 are determined manually before the fitting process, and the selected range in the method of this invention can be 2 to 4.

[0093] Furthermore, a PQ and b PQ By solving using the recursive least squares method, the optimal fitting characteristics of the fitting parameters to the actual inductor surface are achieved:

[0094] e = y - λ T ρ

[0095] y T =[L d (i d i q )](or [L q (i d i q )])

[0096]

[0097] or

[0098] In the formula, e is the least squares algorithm error matrix, y is the least squares algorithm output matrix, λ is the least squares algorithm feedback matrix, and ρ is the least squares algorithm parameter matrix.

[0099] In summary, the inductor identification method based on amplitude adaptive square wave voltage injection in this embodiment can fully consider the variation law of inductance under different operating conditions; at the same time, based on the proposed data sampling and fitting strategy, the inductor surface after correcting the sampling error in the small current range is obtained. Figure 6 and Figure 7 The offline inductor surface identification results based on the method of the present invention show that the method of the present invention can achieve stable and accurate inductor identification under different operating conditions, proving the effectiveness and feasibility of the method of the present invention.

[0100] While the invention has been described herein with reference to specific embodiments, it should be understood that these embodiments are merely examples of the principles and applications of the invention. Therefore, it should be understood that many modifications can be made to the exemplary embodiments, and other arrangements can be designed without departing from the spirit and scope of the invention as defined by the appended claims. It should be understood that different dependent claims and features described herein can be combined in ways different from those described in the original claims. It is also understood that features described in conjunction with individual embodiments can be used in other described embodiments.

Claims

1. A method for offline inductance surface identification of a permanent magnet synchronous motor based on sampling error correction, characterized in that... include, After positioning the motor rotor at zero position based on the d-axis DC signal injection, a square wave voltage signal with the same frequency and phase is synchronously injected into the dq axis of the motor. The amplitude of the square wave voltage signal is adaptively adjusted according to a preset number of cycles. The sampling period is determined according to a preset number of periods; during the square wave voltage signal injection process, the discretized voltage and induced current are sampled for each sampling period; the sampling results of the discretized voltage and induced current containing the three-phase zero current point are removed; based on the retained sampling results of the discretized voltage and induced current for each sampling period, the discretized flux linkage is solved by discretization integration. Then, based on the discretized flux linkage, the inductance identification results of the motor under different saturation conditions are calculated; Based on all inductor identification results, a multinomial fitting algorithm is used to fit the results to obtain offline inductor surface identification results. and The adaptive adjustment method is as follows: ， In the formula The number of combinations of amplitude values ​​for injecting square wave voltage signals into the dq axis. The total number of combinations of square wave voltage signal amplitude values ​​injected into the dq axis. This is the voltage amplitude limit value; The amplitude of the square wave voltage along the d-axis. This represents the amplitude of the square wave voltage along the q-axis. Each Each value corresponds to one sampling period.

2. The offline inductance surface identification method for permanent magnet synchronous motors based on sampling error correction according to claim 1, characterized in that, Square wave voltage signals include d-axis square wave voltage signals. and q-axis square wave voltage signal : In the formula For time, Injecting period counting into the square wave voltage signal, The frequency of the square wave voltage along the d-axis. The frequency of the q-axis square wave voltage is given.

3. The offline inductance surface identification method for permanent magnet synchronous motors based on sampling error correction according to claim 2, characterized in that, and All are 0.1-0.2 times the PWM frequency.

4. The offline inductance surface identification method for permanent magnet synchronous motors based on sampling error correction according to claim 3, characterized in that, Methods for determining the three-phase zero-current point include: ， In the formula This is the current in phase A. This is the B-phase current. For phase C current, The induced current is discretized along the d-axis. The induced current is discretized along the q-axis. To identify the initial rotor position of the motor offline, .

5. The offline inductance surface identification method for permanent magnet synchronous motors based on sampling error correction according to claim 4, characterized in that, Discretized flux linkage calculation methods include: ， In the formula For the d-axis discretized flux linkage in the nth sampling period, For the q-axis discretized flux linkage in the nth sampling period, This is the d-axis square wave voltage signal of the nth sampling period. This is the q-axis square wave voltage signal in the nth sampling period. For stator resistance, The discretized induced current along the d-axis is collected during the nth sampling period. The discretized induced current along the q-axis is collected during the nth sampling period; .

6. The offline inductance surface identification method for permanent magnet synchronous motors based on sampling error correction according to claim 5, characterized in that, The sampling period is 2-5 PWM cycles.

7. The offline inductance surface identification method for permanent magnet synchronous motors based on sampling error correction according to claim 6, characterized in that, The method for obtaining the inductance identification results of the motor under different saturation conditions is as follows: ， The result of d-axis inductance identification. The results of q-axis inductance identification. This represents the average d-axis discretized induced current of the current sampling period and the adjacent previous sampling period. It represents the average value of the discretized induced current along the q-axis between the current sampling period and the adjacent previous sampling period.

8. The offline inductance surface identification method for permanent magnet synchronous motors based on sampling error correction according to claim 7, characterized in that, Within the sampling period corresponding to the amplitude combination of each pair of square wave voltage signals injected along the dq axes and This forms a trajectory line on the dq-axis current plane; Methods for obtaining offline inductor surface identification results include: ， In the formula The parameters are for fitting the d-axis inductance surface. These are the parameters for fitting the q-axis inductance surface. , P is the order of the polynomial along the d-axis, Q is the order of the polynomial along the q-axis, P0 is the maximum order of the polynomial along the d-axis, and Q0 is the maximum order of the polynomial along the q-axis.

9. The offline inductance surface identification method for permanent magnet synchronous motors based on sampling error correction according to claim 8, characterized in that, and Solve using the recursive least squares method: In the formula This is the error matrix of the least squares algorithm. The output matrix of the least squares algorithm. This is the feedback matrix for the least squares algorithm. This is the parameter matrix for the least squares algorithm.

Images