Method and apparatus for commissioning PMSM position sensorless control parameters

Abstract

The present application relates to a method for commissioning position sensorless control parameters of a control system of a position sensorless permanent magnet synchronous motor (PMSM), comprising: configuring the control system of the PMSM using the position sensorless control parameters; acquiring an estimated position of a rotor of the PMSM generated by a control system of the PMSM under the position sensorless control parameter; acquiring an actual position of the rotor detected by a position sensor mounted on the PMSM; determining a position estimation error based on the estimated position and the actual position of the rotor; and taking the position estimation error and the estimated position as input of a neural network (NN) model to obtain updated position sensorless control parameters output by the NN model.

Description

Technical Field

[0001] The present application relates generally to permanent magnet synchronous motors, and more particularly to methods and apparatus for tuning position sensorless control parameters of a control system of a position sensorless permanent magnet synchronous motor (PMSM). Background Art

[0002] PMSM is widely used in industrial control, servo, automobile, aerospace and medical equipment. The PMSM control system needs to obtain the position information of the motor rotor in real time. The position sensor installed on the rotor can get the position of the motor rotor. However, the existence of the position sensor makes the structure of PMSM more complicated. The position sensor is greatly affected by the external environment. It is possible that the position information detected in a harsh environment is inaccurate, which reduces the reliability of PMSM.

[0003] A position sensorless method has been developed to solve the problems that may be caused by the above position sensors. The position sensorless method enables the control system of the PMSM to estimate the position of the rotor without installing a position sensor on the motor. The PMSM using the position sensorless method can be called a position sensorless PMSM.

[0004] Commonly used position sensorless methods include observer-based control methods such as back-electromotive force (EMF) estimation and Kalman filtering methods. These methods estimate the rotor position by estimating the back-EMF. Since the amplitude of the back-EMF is proportional to the motor speed, this method works well in the high-speed range, but it is difficult to accurately estimate the position information in the zero-speed or low-speed area.

[0005] The high frequency injection (HFI) method is a typical position sensorless method that can estimate the rotor position in the zero speed or low speed range. The principle of the HFI method is to inject a high frequency voltage or current signal superimposed on the basic voltage control signal, process the high frequency current or voltage response signal of the PMSM, and obtain the estimated rotor position through the error tracking algorithm. Since the response of the injected high frequency signal is the response of the external signal superimposed on the high frequency signal of the motor basic control current, the estimated position signal is independent of the intensity of the motor basic voltage and current signal, and is based on the high frequency response of the external high frequency injection signal. Therefore, the HFI method can accurately estimate the motor position in the zero speed or low speed stage of the motor, which has a very obvious advantage. However, the HFI method is very sensitive to the position sensorless control parameters, so it is necessary to find suitable position sensorless control parameters for the PMSM control system during the debugging stage. When debugging the position sensorless control parameters of the PMSM control system, the debugger needs to measure, track errors, and adjust parameters for a long time to find the best combination of position sensorless control parameters. Therefore, it would be beneficial if the debugging efficiency of the position sensorless control parameters of the control system of the position sensorless PMSM can be improved. Summary of the invention

[0006] The following summary is provided to introduce selected concepts in a simplified form that are further described in the detailed description that follows. This summary is not intended to highlight key features or essential features of the claimed subject matter, nor is it intended to limit the scope of the claimed subject matter.

[0007] According to one aspect of the present application, a method for debugging position sensorless control parameters of a control system of a position sensorless PMSM is provided, comprising: configuring the control system of the PMSM using the position sensorless control parameters; obtaining the estimated position of the rotor of the PMSM generated by the control system of the PMSM under the position sensorless control parameters; obtaining the actual position of the rotor detected by a position sensor installed on the PMSM; determining a position estimation error based on the estimated position and the actual position of the rotor; and using the position estimation error and the estimated position as inputs of a neural network (NN) model to obtain updated position sensorless control parameters output by the NN model.

[0008] According to one aspect of the present application, a system for debugging position sensorless control parameters of a control system of a position sensorless PMSM is provided, comprising: the PMSM; a position sensor installed on the PMSM; a control device configured to perform the following operations: configure the control system of the PMSM using the position sensorless control parameters; obtain the estimated position of the rotor of the PMSM generated by the control system of the PMSM under the position sensorless control parameters; obtain the actual position of the rotor detected by the position sensor installed on the PMSM; determine a position estimation error based on the estimated position and the actual position of the rotor; and use the position estimation error and the estimated position as inputs of a NN model to obtain updated position sensorless control parameters output by the NN model.

[0009] According to one aspect of the present application, a device for debugging position sensorless control parameters of a control system of a position sensorless PMSM is provided, comprising: a parameter configuration module, used to configure the control system of the PMSM using the position sensorless control parameters; a position information acquisition module, used to obtain the estimated position of the rotor of the PMSM generated by the control system of the PMSM under the position sensorless control parameters, and to obtain the actual position of the rotor detected by a position sensor installed on the PMSM; a position estimation error determination module, used to determine a position estimation error based on the estimated position and the actual position of the rotor; and a parameter updating module, used to use the position estimation error and the estimated position as inputs of a NN model to obtain updated position sensorless control parameters output by the NN model.

[0010] According to one aspect of the present application, a device for debugging position sensorless control parameters of a control system of a position sensorless PMSM is provided, comprising: one or more processors and a memory, the memory storing computer executable instructions, which, when executed, enable the one or more processors to perform methods according to various embodiments of the present disclosure.

[0011] According to one aspect of the present application, a machine-readable storage medium is provided, which stores executable instructions. When the instructions are executed, one or more processors execute the method according to various embodiments of the present disclosure.

[0012] By using the PMSM position sensorless control parameter debugging method according to the present disclosure, an automatic adjustment method based on a neural network is used to find the optimal PMSM position sensorless control parameter that gives the minimum position estimation error. By adopting the debugging system according to the present disclosure, the control system of the test device is automatically configured by the control device, the online test of the test device is controlled, the test results are collected, and the position sensorless control parameters are efficiently adjusted using a neural network model, thereby continuously improving the position sensorless control parameters. As a result, it is possible to save the operator's long-term tracking and debugging, use the neural network to more efficiently predict the optimal combination of PMSM position sensorless control parameters, and automatically complete the debugging work, thereby completing the PMSM position sensorless control parameter debugging work more quickly and efficiently in various PMSM application projects. Other advantages of the embodiments of the present application will be described below. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] A further understanding of the nature and advantages of the present application may be achieved by referring to the following drawings, in which similar components or features may have the same reference numerals.

[0014] Figure 1 A block diagram of a system for tuning position sensorless control parameters of a control system of a position sensorless PMSM is shown according to one embodiment.

[0015] Figure 2A and 2B Schematic diagrams of a neural network model for updating position sensorless control parameters of a control system of a PMSM according to an embodiment are respectively shown.

[0016] Figure 3 A flow chart of a method for tuning position sensorless control parameters of a control system of a position sensorless PMSM according to one embodiment is shown.

[0017] Figure 4 A flow chart of a method for tuning position sensorless control parameters of a control system of a position sensorless PMSM according to one embodiment is shown.

[0018] Figure 5 A block diagram of an apparatus for tuning position sensorless control parameters of a control system of a position sensorless PMSM according to one embodiment is shown. DETAILED DESCRIPTION

[0019] The subject matter described herein will now be discussed with reference to example embodiments. It should be understood that the discussion of these embodiments is only to enable those skilled in the art to better understand and implement the subject matter described herein, and is not a limitation of the scope of protection, applicability or examples set forth in the claims. The functions and arrangements of the elements discussed can be changed without departing from the scope of protection of the present application. Various examples can omit, replace or add various processes or components as needed. For example, the described method can be performed in an order different from the described order, and various steps can be added, omitted or combined. In addition, the features described relative to some examples can also be combined in other examples.

[0020] Figure 1 A block diagram of a system 100 for tuning sensorless control parameters of a control system of a sensorless PMSM is shown according to one embodiment.

[0021] The system 100 includes a position sensorless PMSM 110, a control device 120, and a position sensor 130. The control device 120 includes a debugging system 1210 and a control system 1220 of the position sensorless PMSM 110. The position sensor 130 is mounted on the PMSM 110 and is used to detect the position of the rotor of the PMSM 110. For ease of description, the position of the rotor detected by the position sensor 130 is referred to as the actual position, and the position of the rotor estimated by the position estimation module 12230 of the control system 1220 of the PMSM 110 is referred to as the estimated position.

[0022] The control system 1220 of the PMSM 110 includes a high frequency injection (HFI) module 12210, a high frequency response signal acquisition module 12220, and a position estimation module 12230. The HFI module 12210 injects a high frequency voltage or current signal into the basic control voltage signal of the motor itself. For example, the HFI module 12210 injects or superimposes a high frequency pulsating voltage signal having a specific amplitude into the basic control voltage signal of the D axis of the PMSM 110. The high frequency response signal acquisition module 12220 acquires a high frequency response signal generated based on the high frequency injection signal. For example, the high frequency response signal acquisition module 12220 acquires a D axis high frequency current response signal generated based on the injected high frequency pulsating voltage signal. The D axis high frequency current response signal may be a high frequency response current change value, for example, the difference between the current high frequency response current and the high frequency response current of the previous control cycle. The position estimation module 12230 estimates the position of the rotor of the PMSM 110 based on the acquired high frequency response signal. For example, the position estimation module 12230 estimates the position of the rotor of the PMSM 110 based on the acquired high-frequency response current change value. The position estimation module 12230 can be implemented as a position observer, which is configured with a proportional gain (P gain) coefficient, an integral gain (I gain) coefficient, and a differential gain (D gain) coefficient as position sensorless control parameters. The position observer obtains the estimated rotor position by processing the high-frequency response current change value based on the P gain coefficient, the I gain coefficient, and the D gain coefficient. The performance of the PMSM 110 is sensitive to the amplitude of the high-frequency injection signal injected by the HFI module 12210, the P gain coefficient, the I gain coefficient, and the D gain coefficient of the position estimation module 12230, which are position sensorless control parameters. The goal of the system 100 is to find the optimized above-mentioned position sensorless control parameters for the PMSM 110. Those skilled in the art will appreciate that the PMSM 110 and its control system 1220 may include more components. Figure 1 Only components related to the technical solution of the present disclosure are shown, and other components are not shown. The technical solution of the present disclosure is not limited to the specific implementation details of the PMSM and its control system, and the technical solution of the present disclosure can be used to debug the position sensorless control parameters of the control system of the position sensorless PMSM known in the art and known in the future.

[0023] The debugging system 1210 includes a parameter configuration module 12110, a position information acquisition module 12120, a position estimation error determination module 12130, a position estimation error discrimination module 12140, and a parameter update module 12150. The parameter configuration module 12110 is used to configure the position sensorless control parameters of the control system 1220 of the PMSM 110. For example, Figure 1As shown, the parameter configuration module 12110 can configure the high frequency injection signal amplitude of the HFI module 12210, the P gain coefficient, the I gain coefficient and the D gain coefficient of the position estimation module 12230. Figure 1Although not shown in the figure, the debugging system 1210 may further include a PMSM startup module. After the parameter configuration module 12110 configures the position sensorless control parameters of the control system 1220 of the PMSM 110, the PMSM startup module may start the operation of the PMSM 110 and the control system 1220. The position estimation module 12230 of the control system 1220 of the PMSM 110 estimates and outputs the position of the rotor, such as the position of the rotor salient pole. The position sensor 130 measures the position of the rotor. The position information acquisition module 12120 acquires the estimated position of the rotor from the position estimation module 12230 and the measured position or actual position from the position sensor 130. The position estimation error determination module 12130 determines the position estimation error between the estimated position and the actual position of the rotor acquired by the position information acquisition module 12120. The position estimation error determination module 12140 determines whether the position estimation error is less than the error threshold. In one embodiment, if the position estimation error is less than the error threshold, it is determined that the optimal position sensorless control parameter is obtained, and the debugging process is stopped accordingly; if the position estimation error is greater than the error threshold, it is determined that the optimal position sensorless control parameter is not obtained, and the debugging process continues. In one embodiment, if the position estimation error determination module 12140 determines that the position estimation error is less than the error threshold for multiple consecutive times, it is determined that the optimized position sensorless control parameter is obtained, and the debugging is stopped accordingly. Specifically, in this embodiment, when the position estimation error determination module 12140 determines that the position estimation error is less than the error threshold for the first time, the next round of debugging is still performed, and the optimized position sensorless control parameter is determined to be obtained only when the position estimation error is determined to be less than the error threshold in multiple consecutive rounds of debugging, and the debugging is stopped accordingly. The number of times that the position estimation error is determined to be less than the error threshold for multiple consecutive times can be recorded by a counter. When the recorded number of times reaches the number threshold, the position estimation error determination module 12140 determines that the optimized position sensorless control parameter is obtained, and the debugging is stopped accordingly. The parameter updating module 12150 may include a NN model, which updates the position sensorless control parameters based on the position estimation error determined by the position estimation error determining module 12130 and the estimated position obtained by the position information acquiring module 12120. For example, the weight coefficient of the NN model is updated based on the position estimation error, the position estimation error and the estimated position are used as inputs of the NN model, and the updated position sensorless control parameters output by the NN model are obtained based on the weight coefficients of the updated NN model. Next, the parameter configuration module 12110 uses the updated position sensorless control parameters output by the parameter updating module 12150 to configure the position sensorless control parameters of the control system 1220 of the PMSM 110, and repeats the above debugging process.

[0024] Understandable, although Figure 1In the embodiment shown in FIG. 1 , the control system 1220 and the debugging system 1210 are implemented on one control device 120 . However, in other embodiments, the control system 1220 and the debugging system 1210 may also be implemented on separate control devices.

[0025] Figure 2A A schematic diagram of a NN model 200A for updating position sensorless control parameters of a control system of a position sensorless PMSM according to one embodiment is shown.

[0026] The parameter updating module 12150 may include the NN model 200A, or may be implemented as the NN model 200A. The NN model 200A includes an input layer, a hidden layer, and an output layer. The input layer includes two neurons 211 and 212, the hidden layer includes four neurons 221, 222, 223, and 224, and the output layer includes four neurons 231, 232, 233, and 234. The neurons in two adjacent layers are connected in a fully connected manner, that is, each neuron is connected to all neurons in the previous layer and all neurons in the next layer, so Figure 2A The NN model 200A shown is implemented as a fully connected NN model. Each connection line between neurons has a corresponding weight coefficient, for example, the connection line between neurons 211 and 221 has a weight coefficient w1121, the connection line between neurons 212 and 224 has a weight coefficient w1224, and so on. Figure 2A Only the weight coefficients of some connecting lines are shown in FIG. 1 . It can be understood that each connecting line has a corresponding weight coefficient.

[0027] The neurons 211 and 212 of the input layer receive the estimated position p obtained by the position information acquisition module 12120 and the position estimation error e determined by the position estimation error determination module 12130 as the input of the NN model 200A. Taking the neuron 221 as an example, the output value of the neuron 221 can be obtained at the neuron 221 of the hidden layer based on the output values ​​of the neurons 211 and 212 in the previous layer and the corresponding weight coefficients w1121 and w1221. Specifically, the output value of the neuron 221 is obtained at the neuron 221 based on the weight coefficients w1121 and w1221 and the estimated position p and the position estimation error e. For example, a linear function can be used at the neuron 221 to obtain the output value of the neuron 221 based on the weight coefficients w1121 and w1221 and the estimated position p and the position estimation error e. The linear function can be, for example, the weighted sum of the estimated position p and the position estimation error e, or other linear functions. For another example, a nonlinear function may be used at neuron 221 to obtain the output value of neuron 221 based on weight coefficients w1121 and w1221, the estimated position p, and the position estimation error e. The output values ​​of other neurons in the hidden layer may be obtained in a similar manner. Similarly, taking neuron 231 as an example, the output value KP of neuron 231 may be obtained at neuron 231 in the output layer based on the output values ​​of each neuron 221, 222, 223, and 224 in the previous layer and the corresponding weight coefficients w2131, w2231, w2331, and w2431. For example, a linear function may be used at neuron 231 to obtain the output value KP, which may be, for example, the weighted sum of the output values ​​of each neuron 221, 222, 223, and 224 in the previous layer, or other linear functions. For example, a nonlinear function may be used at neuron 221 to obtain the output value KP. The output values ​​KI, KD, and U of other neurons in the output layer may be obtained in a similar manner. The output values ​​KP, KI, KD, and U of each neuron in the output layer are respectively the updated P gain coefficient, I gain coefficient, D gain coefficient, and HFI signal amplitude in this iteration or debugging cycle.

[0028] Before the updated position sensorless control parameters KP, KI, KD, and U are obtained based on the input estimated position p and the position estimation error e by the NN model 200A, the weight coefficient w of the NN model 200A is updated. In one embodiment, the weight coefficient w of the NN model 200A is updated by adopting an optimal estimation method, and the weight coefficient w of the NN model 200A can be updated by an optimal estimation method such as the least squares method, the gradient descent method, the Gauss-Newton method, the Levenberg-Marquardt method, and the Powell method. An optimal estimation method is to train the NN model by numerically approximating back propagation when the NN model does not have an ideal output. Taking the Levenberg-Marquardt method as an example, the Jacobian matrix (derivative matrix) of the weight coefficient of the NN model can be numerically approximated at each time step using deviations relative to small changes in each weight coefficient, and then the control error is reduced by adjusting each weight coefficient. The control error in the technical solution disclosed in the present invention can be the position estimation error e. The above-mentioned optimal estimation methods for optimizing the NN model weight coefficients are all methods known in the art. Those skilled in the art will understand that the NN weight coefficient optimization methods known in the art or to be known in the future can be used to optimize the weight coefficients of the NN model in each debugging cycle so that the optimized weight coefficients reduce the position estimation error e.

[0029] Figure 2B FIG. 2 is a schematic diagram of a NN model 200B for updating position sensorless control parameters of a PMSM according to an embodiment. Figure 2A The NN model 200A shown is different only in that it includes five neurons 221 to 225 in the hidden layer. The operation of each neuron in each layer of the NN model 200B is the same or similar to the operation of each neuron in each layer of the NN model 200A. It will be appreciated by those skilled in the art that although Figure 2A and 2B A fully connected NN model including an input layer, an output layer and one hidden layer is shown in the figure, but the technical solution of the present disclosure may also adopt other NN model structures. For example, the NN model may include an input layer, multiple hidden layers and an output layer. The input layer includes two neurons corresponding to the estimated position p and the position estimation error e. The output layer includes four neurons corresponding to the P gain coefficient, the I gain coefficient, the D gain coefficient and the HFI signal amplitude. The number of neurons included in the hidden layer is not limited to a specific number. For example, each hidden layer may include four neurons, or five neurons, or other numbers of neurons.

[0030] Figure 3 A flow chart of a method 300 for tuning sensorless control parameters of a control system of a sensorless PMSM is shown according to one embodiment.

[0031] The method 300 may be performed by the control device 120, specifically by the debugging system 1210. In step 310, the control device 120 performs parameter configuration of the control system of the PMSM 110, for example, configuring the amplitude of the high frequency injection signal of the HFI module 12210, the P gain coefficient, the I gain coefficient and the D gain coefficient of the position estimation module 12230. In step 320, after the parameter configuration, the control device 120 starts the operation of the PMSM 110 and the control system 1220. In step 330, the control device 120 obtains the estimated position of the rotor output by the position estimation module 12230 of the control system 1220 of the PMSM 110, and obtains the actual position of the rotor measured by the position sensor 130. In step 340, the control device 120 calculates the position estimation error based on the obtained estimated position and actual position of the rotor. In step 350, the control device 120 determines whether the calculated position estimation error is less than the error threshold. If yes, then the debugging process ends at step 370, and the current PMSM position sensorless control parameters are stored or reported. If no, then the PMSM position sensorless control parameters are updated at step 360, for example, by updating the P gain coefficient, I gain coefficient, D gain coefficient and the amplitude of the high-frequency injection signal based on the estimated position and the position estimation error through the NN model 200A or 200B. Then, step 360 is returned to step 310, and the debugging process 300 is repeated until the debugging is completed at step 370.

[0032] Figure 4 A flow chart of a method 400 for tuning sensorless control parameters of a control system of a sensorless PMSM is shown according to one embodiment.

[0033] At step 410 , the control system of the PMSM is configured using position sensorless control parameters.

[0034] In step 420 , an estimated position of the rotor generated by the control system of the PMSM under the position sensorless control parameters is obtained.

[0035] In step 430 , the actual position of the rotor detected by a position sensor mounted on the PMSM is acquired.

[0036] At step 440, a position estimation error is determined based on the estimated position and the actual position of the rotor.

[0037] In step 450, the position estimation error and the estimated position are used as inputs of the NN model to obtain updated position sensorless control parameters output by the NN model.

[0038] In one embodiment, the position sensorless control parameters of the PMSM control system include: the amplitude of the HFI signal, the gain coefficient of the position observer of the control system, wherein the gain coefficient of the position observer includes the P gain coefficient, the I gain coefficient and the D gain coefficient of the position observer.

[0039] In one embodiment, method 400 further comprises: repeating the operations of steps 410 to 450 using the updated position sensorless control parameters output by the NN model until a predetermined condition is satisfied. In one embodiment, the predetermined condition comprises: the position estimation error is less than a position estimation error threshold.

[0040] In one embodiment, the NN model includes a fully connected NN model. In one embodiment, the NN model includes an input layer, one or more hidden layers and an output layer, the input layer includes a first neuron and a second neuron, which respectively receive the position estimation error and the estimated position, and the output layer includes a first neuron, a second neuron, a third neuron and a fourth neuron, which respectively output the P gain coefficient, the I gain coefficient, the D gain coefficient and the amplitude of the HFI signal. In one embodiment, the NN model includes a hidden layer, the hidden layer includes a first neuron, a second neuron, a third neuron and a fourth neuron, or the hidden layer includes a first neuron, a second neuron, a third neuron, a fourth neuron and a fifth neuron.

[0041] Figure 5 A block diagram of an apparatus 500 for tuning position sensorless control parameters of a position sensorless PMSM according to one embodiment is shown.

[0042] According to one embodiment, the apparatus 500 may include one or more control systems or processing units 510 that execute one or more machine-readable instructions stored or encoded in a machine-readable storage medium (ie, memory 520). Figure 5 However, those skilled in the art will appreciate that the device 500 may include various other components, such as various communication modules, bus modules, and possible user interface modules. The device 500 may be used to implement Figure 1 The control device 120 shown in FIG. 1 , wherein the processing unit 510 is configured to execute the above-mentioned combined Figure 1-4 Various operations and functions of the control device 120 are described.

[0043] According to one embodiment, a program product such as a non-transitory machine-readable medium is provided. The non-transitory machine-readable medium may have instructions, which, when executed by the processing unit 510, cause the apparatus 500, the control device 120, or the system 100 to perform the above combined operation in various embodiments of the present application. Figure 1-4 Describes the various operations and functions.

[0044] The specific embodiments described above in conjunction with the accompanying drawings describe exemplary embodiments, but do not represent all embodiments that can be implemented or fall within the scope of protection of the claims. The term "example" used throughout this specification may mean "used as an example, instance or illustration" and does not mean "preferred" or "having advantages" over other embodiments. For the purpose of providing an understanding of the described technology, the specific embodiments include specific details. However, these technologies can be implemented without these specific details. In some instances, in order to avoid making the concepts of the described embodiments difficult to understand, well-known structures and devices are shown in block diagram form.

[0045] The above description of the present application is provided to enable any person of ordinary skill in the art to implement or use the present application. It is obvious to a person of ordinary skill in the art that various modifications are made to the present application, and the general principles defined herein may be applied to other variations without departing from the scope of protection of the present application. Therefore, the present application is not limited to the examples and designs described herein, but is consistent with the widest range of principles and novel features disclosed herein.

Claims

1. A method for debugging a position sensorless control parameter of a control system of a position sensorless permanent magnet synchronous motor (PMSM), comprising: configuring the control system of the PMSM using the position sensorless control parameters; obtaining an estimated position of a rotor of the PMSM generated by the control system of the PMSM under the position sensorless control parameters; Acquiring an actual position of the rotor detected by a position sensor mounted on the PMSM; determining a position estimation error based on the estimated position and the actual position of the rotor; The position estimation error and the estimated position are used as inputs of a neural network (NN) model to obtain updated position sensorless control parameters output by the NN model.

2. The method according to claim 1, wherein: The position sensorless control parameters of the control system of the PMSM include: the amplitude of a high frequency injection (HFI) signal and a gain coefficient of a position observer, wherein the gain coefficient of the position observer includes a P gain coefficient, an I gain coefficient and a D gain coefficient of the position observer.

3. The method according to claim 1 or 2, further comprising: The operation of claim 1 is repeated using the updated sensorless control parameters output by the NN model until a predetermined condition is met.

4. The method according to claim 3, wherein: The predetermined condition includes: the position estimation error is less than a position estimation error threshold.

5. The method according to claim 1 or 2, wherein: The NN model includes a fully connected NN model.

6. The method according to claim 5, wherein: The NN model includes an input layer, one or more hidden layers and an output layer, the input layer includes a first neuron and a second neuron, which respectively receive the position estimation error and the estimated position, and the output layer includes a first neuron, a second neuron, a third neuron and a fourth neuron, which respectively output the P gain coefficient, the I gain coefficient, the D gain coefficient and the amplitude of the HFI signal.

7. The method according to claim 6, wherein: The NN model includes only one hidden layer, and the one hidden layer includes a first neuron, a second neuron, a third neuron, and a fourth neuron, or the one hidden layer includes a first neuron, a second neuron, a third neuron, a fourth neuron, and a fifth neuron.

8. A system for debugging a position sensorless control parameter of a control system of a position sensorless permanent magnet synchronous motor (PMSM), comprising: the PMSM; a position sensor mounted on the PMSM; A control device configured to: configuring the control system of the PMSM using the position sensorless control parameters; obtaining an estimated position of a rotor of the PMSM generated by the control system of the PMSM under the position sensorless control parameters; Acquiring an actual position of the rotor detected by a position sensor mounted on the PMSM; determining a position estimation error based on the estimated position and the actual position of the rotor; The position estimation error and the estimated position are used as inputs of a neural network (NN) model to obtain updated position sensorless control parameters output by the NN model.

9. The system according to claim 8, wherein: The position sensorless control parameters of the control system of the PMSM include: the amplitude of a high frequency injection (HFI) signal and a gain coefficient of a position observer, wherein the gain coefficient of the position observer includes a P gain coefficient, an I gain coefficient and a D gain coefficient of the position observer.

10. The system according to claim 8 or 9, wherein: The control device is configured to perform the following operations: The operation of claim 8 is repeated using the updated sensorless control parameters output by the NN model until a predetermined condition is met.

11. The system according to claim 10, wherein the predetermined condition comprises: The position estimation error is less than a position estimation error threshold.

12. The system according to claim 8 or 9, wherein: The NN model includes a fully connected NN model.

13. The system according to claim 12, wherein: The NN model includes an input layer, one or more hidden layers and an output layer, the input layer includes a first neuron and a second neuron, which respectively receive the position estimation error and the estimated position, and the output layer includes a first neuron, a second neuron, a third neuron and a fourth neuron, which respectively output the P gain coefficient, the I gain coefficient, the D gain coefficient and the amplitude of the HFI signal.

14. The system according to claim 13, wherein: The NN model includes only one hidden layer, and the one hidden layer includes a first neuron, a second neuron, a third neuron, and a fourth neuron, or the one hidden layer includes a first neuron, a second neuron, a third neuron, a fourth neuron, and a fifth neuron.

15. A device for debugging position sensorless control parameters of a control system of a position sensorless permanent magnet synchronous motor (PMSM), comprising: a parameter configuration module, configured to configure the control system of the PMSM using the position sensorless control parameters; a position information acquisition module, used to acquire an estimated position of a rotor of the PMSM generated by the control system of the PMSM under the position sensorless control parameters, and to acquire an actual position of the rotor detected by a position sensor installed on the PMSM; a position estimation error determination module, configured to determine a position estimation error based on the estimated position and the actual position of the rotor; as well as A parameter updating module is used to use the position estimation error and the estimated position as inputs of a neural network (NN) model to obtain updated position sensorless control parameters output by the NN model.

16. A device for debugging position sensorless control parameters of a control system of a position sensorless permanent magnet synchronous motor (PMSM), comprising: one or more processors; as well as A memory storing computer executable instructions which, when executed, cause the one or more processors to perform the method of any one of claims 1 to 7.

17. A machine-readable storage medium storing executable instructions which, when executed, cause one or more processors to perform the method of any one of claims 1 to 7.

Images