Methods, devices, computer equipment, and storage media for multi-parameter identification of permanent magnet synchronous motors

Abstract

This invention discloses a method, apparatus, computer device, and storage medium for multi-parameter identification of permanent magnet synchronous motors, relating to the field of motor control technology, and aimed at solving the problem of severe parameter coupling in traditional online motor parameter identification methods. The method includes: inputting the acquired dataset into an initial parameter identification model; in the initial parameter identification model, determining predicted motor parameters based on motor operating parameters, and determining the predicted voltages of the direct and quadrature axes corresponding to the predicted motor parameters; adjusting the parameters in the parameter identification model based on the loss value of the total loss function until preset conditions are met, thus obtaining a target parameter identification model; and inputting the acquired motor operating parameters of the target motor into the target parameter identification model to obtain the target motor parameters.

Description

Technical Field

[0001] This invention relates to the field of electric motor control technology, and in particular to a method, apparatus, computer equipment, and storage medium for multi-parameter identification of permanent magnet synchronous motors. Background Technology

[0002] In the field of electric drives, especially in high-dynamic applications such as electric vehicles, permanent-magnet synchronous motors (PMSMs) typically need to frequently switch between complex operating conditions such as stall start, rapid acceleration, high-speed cruising, and regenerative braking energy recovery. These operating conditions cause the motor's current and winding temperature to change drastically on a timescale of seconds or even milliseconds, thus affecting key motor parameters, including stator resistance. Direct-axis and quadrature-axis inductors ) and permanent magnet flux ( It exhibits significant time-varying and coupling characteristics.

[0003] To achieve high-performance control of permanent magnet synchronous motors, such as vector control or direct torque control, accurate motor parameters, including stator resistance, must be obtained. Direct-axis and quadrature-axis inductors ) and permanent magnet flux ( If the above motor parameters are misidentified, it will directly lead to torque control errors, reduced system efficiency, and even system oscillations, seriously affecting the accuracy and stability of control.

[0004] Common online motor parameter identification methods, such as model reference adaptation, extended Kalman filtering, and particle swarm optimization, often rely on locally linearized models or gradient-based optimization searches. When the motor parameters change rapidly and synchronously, these identification methods struggle to effectively decouple them, easily outputting incorrect parameter combinations. They cannot provide accurate and reliable parameter estimates under extreme dynamic conditions, thus limiting further improvements in the performance of motor control systems. Summary of the Invention

[0005] Based on this, a method, device, computer equipment, and storage medium for multi-parameter identification of permanent magnet synchronous motors are provided to solve the problem of severe parameter coupling in traditional online motor parameter identification methods.

[0006] In a first aspect, the present invention provides a multi-parameter identification method for a permanent magnet synchronous motor, the method comprising: The acquired dataset is input into the initial parameter identification model; wherein, the dataset includes motor operating parameters and labels corresponding to the motor operating parameters, and the labels include the measured voltages of the direct axis and quadrature axis when the motor is running; In the initial parameter identification model, based on the motor operating parameters, predicted motor parameters are determined, as well as predicted voltages for the direct and quadrature axes corresponding to the predicted motor parameters; wherein, the predicted motor parameters include predicted stator resistance, predicted inductances for the direct and quadrature axes, and predicted permanent magnet flux linkage; Based on the loss value of the total loss function, the parameters in the initial parameter identification model are adjusted until a preset condition is met to obtain the target parameter identification model; wherein, the loss value includes the first loss value of the physical loss function, and the first loss value characterizes the difference between the measured voltage and the predicted voltage; The obtained motor operating parameters of the target motor are input into the target parameter identification model to obtain the target motor parameters, wherein the target motor parameters include the target stator resistance, the target inductance of the direct axis and the quadrature axis, and the target permanent magnet flux linkage.

[0007] Optionally, determining the predicted voltage corresponding to the predicted motor parameters includes: generating the predicted voltages for the direct axis and quadrature axis based on the motor state equation in the total loss function and using the predicted motor parameters.

[0008] Optionally, the first loss value of the physical loss function is determined by: obtaining the measured voltage from the tags corresponding to the motor operating parameters; and determining the error between the predicted voltage and the measured voltage using the physical loss function to obtain the first loss value.

[0009] Optionally, the loss value of the total loss function further includes a second loss value of the data loss function; wherein, the label further includes reference motor parameters corresponding to the motor operating parameters, the reference motor parameters including reference stator resistance, reference inductance of the direct axis and quadrature axis and reference permanent magnet flux linkage, and the second loss value characterizes the difference between the predicted motor parameters and the reference motor parameters.

[0010] Optionally, the loss value of the total loss function is obtained by a weighted sum of the first loss value and the second loss value.

[0011] Optionally, the second loss value of the data loss function is determined by: obtaining the reference motor parameters from the labels corresponding to the motor operating parameters; and determining the error between the predicted motor parameters and the reference motor parameters using the data loss function to obtain the second loss value.

[0012] Optionally, the motor operating parameters include the current feedback values ​​of the direct axis and quadrature axis during motor operation; the step of generating the predicted voltages of the direct axis and quadrature axis based on the motor state equation in the total loss function and using the predicted motor parameters includes: using the automatic differentiation function of the initial parameter identification model to calculate the current feedback values ​​of the direct axis and quadrature axis to obtain the differential terms of the direct axis current and quadrature axis current; substituting the differential terms of the direct axis current, the quadrature axis current, and the predicted motor parameters into the motor state equation to obtain the predicted voltages of the direct axis and quadrature axis.

[0013] Secondly, the present invention provides a multi-parameter identification device for a permanent magnet synchronous motor, the device comprising: The input module is used to input the acquired dataset into the initial parameter identification model; wherein, the dataset includes motor operating parameters and labels corresponding to the motor operating parameters, and the labels include the measured voltages of the direct axis and quadrature axis when the motor is running; The determination module is used to determine the predicted motor parameters and the predicted voltages of the direct axis and quadrature axis corresponding to the predicted motor parameters in the initial parameter identification model based on the motor operating parameters; wherein the predicted motor parameters include the predicted stator resistance, the predicted inductance of the direct axis and quadrature axis, and the predicted permanent magnet flux linkage; An adjustment module is used to adjust the parameters in the initial parameter identification model based on the loss value of the total loss function until a preset condition is met to obtain the target parameter identification model; wherein, the loss value of the total loss function includes a first loss value of the physical loss function, and the first loss value characterizes the difference between the measured voltage and the predicted voltage; The output module is used to input the obtained motor operating parameters of the target motor into the target parameter identification model to obtain the target motor parameters of the target motor, wherein the target motor parameters include the target stator resistance, the target inductance of the direct axis and the quadrature axis, and the target permanent magnet flux linkage.

[0014] Optionally, the determining module is further configured to generate the predicted voltages of the direct axis and quadrature axis based on the motor state equation in the total loss function and using the predicted motor parameters.

[0015] Optionally, the device is further configured to obtain the measured voltage from tags corresponding to the motor operating parameters; and determine the error between the predicted voltage and the measured voltage using the physical loss function to obtain the first loss value.

[0016] Optionally, the loss value of the total loss function further includes a second loss value of the data loss function; wherein, the label further includes reference motor parameters corresponding to the motor operating parameters, the reference motor parameters including reference stator resistance, reference inductance of the direct axis and quadrature axis and reference permanent magnet flux linkage, and the second loss value characterizes the difference between the predicted motor parameters and the reference motor parameters.

[0017] Optionally, the loss value of the total loss function is obtained by weighted summation of the first loss value and the second loss value.

[0018] Optionally, the device is further configured to obtain the reference motor parameters from the tags corresponding to the motor operating parameters; and to determine the error between the predicted motor parameters and the reference motor parameters using the data loss function to obtain the second loss value.

[0019] Optionally, the motor operating parameters include the current feedback values ​​of the direct axis and quadrature axis during motor operation; the determining module is further configured to use the automatic differentiation function of the initial parameter identification model to calculate the current feedback values ​​of the direct axis and quadrature axis to obtain the direct axis current differential term and the quadrature axis current differential term; and to substitute the direct axis current differential term, the quadrature axis current differential term, and the predicted motor parameters into the motor state equation to obtain the predicted voltages of the direct axis and quadrature axis.

[0020] Thirdly, the present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the multi-parameter identification method for permanent magnet synchronous motors described in the first aspect.

[0021] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the multi-parameter identification method for permanent magnet synchronous motors described in the first aspect.

[0022] The aforementioned method, apparatus, computer equipment, and storage medium for multi-parameter identification of permanent magnet synchronous motors introduce a physical loss function calculated based on the motor state equation into the total loss function used to optimize the parameter identification model. This forces the output of the parameter identification model to conform to the physical laws of motor operation, enabling the model to efficiently complete training under the constraints of the physical loss function even without a large amount of offline calibration data. It can also learn autonomously from the input data to obtain a robust target parameter identification model. This achieves high-precision and strong generalization parameter identification under data-scarce conditions, solving the problem that existing methods, which train parameter identification models using only data loss functions calculated from limited measured data, face difficulties in parameter decoupling under multiple operating conditions. Attached Figure Description

[0023] Figure 1 This is a flowchart illustrating a multi-parameter identification method for a permanent magnet synchronous motor in one embodiment; Figure 2 This is a structural block diagram of a multi-parameter identification method device for permanent magnet synchronous motors in one embodiment; Figure 3 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings. The specific operational methods in the method embodiments can also be applied to the device embodiments or system embodiments. It should be noted that in the description of this invention, "multiple" is understood as "at least two". "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing together, or B existing alone. A connected to B can represent: A and B directly connected, or A and B connected through C. Furthermore, in the description of this invention, terms such as "first" and "second" are used only for descriptive purposes and should not be construed as indicating or implying relative importance or order.

[0025] In this invention, the acquisition, transmission, storage, and use of data all comply with the requirements of relevant national laws and regulations.

[0026] The technical solutions provided in the embodiments of this application will be described in detail below with reference to the accompanying drawings and specific implementation methods.

[0027] Figure 1 This is a flowchart illustrating a multi-parameter identification method for a permanent magnet synchronous motor in one embodiment. This process can be executed by a multi-parameter identification device for the permanent magnet synchronous motor. This device can be implemented through software, hardware, or a combination of both. Figure 1 As shown, the process includes the following steps: S101, input the obtained dataset into the initial parameters identification model.

[0028] In this embodiment, the dataset includes a sample set and a label set. The samples in the sample set are motor operating parameters, which include, but are not limited to, the current feedback values ​​of the direct axis and quadrature axis (d / q axis) during motor operation. electric angular velocity Each tag in the tag set includes the measured voltages of the direct axis and quadrature axis during motor operation. and reference motor parameters The reference motor parameters are measured motor parameters collected at specified operating points (such as rated operating point, no-load point, etc.). These parameters include the reference stator resistance (i.e., measured stator resistance), the reference inductance for the direct and quadrature axes (i.e., measured inductance), and the reference permanent magnet flux linkage (i.e., measured permanent magnet flux linkage). Each sample corresponds to a tag, meaning each motor operating parameter corresponds to a measured voltage for the direct and quadrature axes, and a reference motor parameter.

[0029] Specifically, the dataset is obtained as follows: During motor operation, the measured d / q axis voltages, d / q axis current feedback values, and electrical angular velocities are collected in real time. The measured d / q axis voltages can be either voltage command values ​​or measured values. Through finite element analysis and variable frequency no-load experiments, high-precision reference motor parameters are obtained under various operating conditions. The collected d / q axis current feedback values ​​and electrical angular velocities are used as samples, and the obtained reference motor parameters and measured d / q axis voltages are used as labels. A dataset is constructed based on all collected samples and all labels. This dataset is then input into the initial parameter identification model, which is a pre-built parameter identification model to be trained.

[0030] S102, in the initial parameter identification model, based on the motor operating parameters, the predicted motor parameters are determined, as well as the predicted voltages of the direct axis and quadrature axis corresponding to the predicted motor parameters are determined.

[0031] In one embodiment, the constructed parameter identification model includes a neural network that feeds back the d / q-axis current values ​​from the motor operating parameters. and electric angular velocity The neural network is input and outputs predicted motor parameters. These parameters include the predicted stator resistance. Predicted inductance along the direct axis and quadrature axis And predicting permanent magnet flux linkage The predicted motor parameters are expressed as follows: .

[0032] In one embodiment, the neural network in the parameter identification model employs a multilayer feedforward perceptron with a five-layer fully connected structure. Its input layer is designed to receive and standardize three key real-time motor operating parameters: d-axis current... q-axis current With electric angular velocity Following the input layer, three consecutive hidden layers are configured, with 64, 32, and 16 neurons respectively, all employing the ReLU activation function to introduce necessary nonlinear mapping capabilities. The final output layer contains four linear neurons, corresponding to the four target motor parameters to be identified: stator resistance, d-axis inductance, q-axis inductance, and permanent magnet flux linkage. All layers are fully connected via trainable weight matrices and bias vectors.

[0033] In one embodiment, based on the motor state equation in the total loss function, the predicted voltages for the direct axis and quadrature axis are generated using the predicted motor parameters. The specific steps are as follows: S1021, using the automatic differentiation function of the initial parameter identification model, calculates the d / q axis current feedback value in the motor operating parameters to obtain the direct axis current differential term and the quadrature axis current differential term.

[0034] The automatic differentiation function of the neural network in the parameter identification model is used to calculate the current differential term. Since the automatic differentiation is an operation of analyzing the differentiation of internal variables in the model calculation diagram, rather than the numerical processing of the measured current signal with noise by the traditional numerical differentiation method, it avoids the amplification effect of high-frequency noise inherent in the numerical differentiation method, significantly improves the signal-to-noise ratio, and thus improves the identification accuracy of the model.

[0035] S1022, substitute the direct-axis current differential term, quadrature-axis current differential term, and predicted motor parameters into the motor state equation to obtain the predicted voltages of the direct and quadrature axes.

[0036] The motor state equations are as follows: ; ; in, Indicates the predicted voltage along the direct axis. Indicates the predicted voltage of the quadrature axis. This indicates the predicted stator resistance. This indicates the predicted direct-axis inductance. Indicates the predicted quadrature axis inductance. This indicates the predicted flux linkage of the permanent magnet. This represents the current feedback value along the direct axis. This indicates the current feedback value of the quadrature axis. Represents the current differential term along the direct axis. This represents the differential term of the current along the quadrature axis.

[0037] S103. Based on the loss value of the total loss function, adjust the parameters in the initial parameter identification model until the preset conditions are met to obtain the target parameter identification model.

[0038] In one embodiment, the loss value of the total loss function includes a first loss value of the physical loss function, which characterizes the difference between the measured d / q-axis voltage and the predicted d / q-axis voltage. The first loss value of the physical loss function can be determined as follows: S1031, obtain the measured voltages of the direct axis and quadrature axis from the labels corresponding to the motor operating parameters.

[0039] S1032, through the physical loss function, determines the error between the measured voltage and the predicted voltage of the direct axis and the quadrature axis, and obtains the first loss value.

[0040] In one embodiment, the error is chosen as mean square error, then the physical loss function is expressed as: ; in, Represents the predicted voltages along the direct axis and quadrature axis. This represents the measured voltage along the direct axis and the quadrature axis.

[0041] By embedding the physical laws of the motor (i.e., the motor state equation) as hard constraints into the total loss function of the parameter identification model, the parameter identification model is guided to learn parameter decoupling mappings that conform to physical laws based on the total loss function, thereby improving the parameter decoupling capability of the model.

[0042] In one embodiment, the loss value of the total loss function further includes a second loss value of the data loss function, which characterizes the difference between the predicted motor parameters and the corresponding reference motor parameters. The second loss value of the data loss function can be determined in the following manner: S1033, obtain the reference motor parameters from the labels corresponding to the motor operating parameters.

[0043] Since the acquired dataset contains multiple labels, it is necessary to select the labels corresponding to the motor operating parameters and perform error analysis on the baseline motor parameters and predicted motor parameters in the corresponding labels.

[0044] In one embodiment, the method for selecting a tag corresponding to the motor operating parameters can be as follows: Determine the operating condition to which the motor operating parameters belong. For example, determine the operating condition based on the magnitude of the d / q-axis current feedback value and the magnitude of the electrical angular velocity in the motor operating parameters. The aforementioned reference motor parameters carry operating condition information. Then, select a target tag corresponding to the operating condition from multiple tags, and obtain the reference motor parameters from the target tag. The aforementioned operating condition could be, for example, the motor being at its rated point or no-load point.

[0045] S1034, through the data loss function, determines the error between the predicted motor parameters and the reference motor parameters, and obtains the second loss value.

[0046] In one embodiment, the error is chosen as mean square error, then the data loss function is expressed as: ; in, To predict motor parameters, To predict stator resistance, To predict direct-axis inductance, To predict quadrature axis inductance, To predict the magnetic flux linkage of permanent magnets, Based on the motor parameters, As the reference stator resistance, As a reference direct-axis inductor, As a reference quadrature axis inductance, The reference permanent magnet flux linkage.

[0047] S1035, the total loss function is obtained by weighted summing of the first loss value of the physical loss function and the second loss value of the data loss function. The total loss function is expressed as: ; in, To preset hyperparameters, Represents the data loss function. This represents the physical loss function. In one embodiment, The model can be adjusted as needed during training. For example, in the initial training phase, it can be... Setting it to a small value allows the data loss function to dominate the parameter identification model, accelerating model convergence and ensuring the model's parameters quickly converge to a reasonable range, avoiding getting trapped in local optima. In the mid-to-late stages of training, this value can be increased. The value of makes the physical loss function dominant. In some implementations, 0 < <1. It should be noted that the embodiments of this application... There are no restrictions on the adjustment methods.

[0048] S1036, based on the input sample set and the loss value of the total loss function, iteratively train the initial parameter identification model, and adjust the parameters in the parameter identification model through the optimization algorithm until the preset conditions are met, thus obtaining the trained target parameter identification model.

[0049] In some implementations, the preset condition may be that the loss value of the total loss function is less than a set value, or that the change in the loss value of the total loss function in N consecutive iterations is less than a preset threshold, or that the number of training rounds of the model reaches a set maximum number. In the embodiments of this application, the preset condition is not specifically limited.

[0050] In this embodiment, the training process of the parameter identification model is a hybrid supervised learning approach. Its core is defining a composite loss function that integrates data-driven and physical law constraints. Specifically, the total loss function consists of two weighted parts: one is the mean square error loss between the neural network output motor parameters and the labeled reference motor parameters; the other is the mean square error loss between the predicted voltage calculated by substituting the motor operating parameters into the voltage state equation of the permanent magnet synchronous motor and the actual system measured voltage—that is, the physical loss. For example, the model training uses the Adam optimizer, with an initial learning rate set to 0.001, and the learning rate is multiplied by a decay factor γ=0.5 every 200 training epochs. Using the aforementioned total loss function as the target, the backpropagation algorithm iteratively updates all weights and biases of the network until the model converges. This ultimately enables the neural network not only to fit limited offline data but also to strictly adhere to the inherent physical laws of the motor, thereby achieving high-precision and robust online decoupled identification of motor parameters under all operating conditions and time-varying parameters.

[0051] S104. Input the obtained motor operating parameters of the target motor into the target parameter identification model to obtain the target motor parameters.

[0052] After obtaining the target parameter identification model, the real-time motor operating parameters of the target motor are input into the target parameter identification model, and the output parameters are the target motor parameters, which are also the identified motor parameters of the target motor. The target motor parameters include the target stator resistance, the target inductance of the direct axis and quadrature axis, and the target permanent magnet flux linkage.

[0053] The multi-parameter identification method for permanent magnet synchronous motors provided in this application integrates prior data with physical laws through a composite loss function that incorporates both data and physical constraints. This composite loss function consists of two parts: 1) a data loss term, calculated based on limited measured data, providing accurate initial guidance for model training and ensuring that the model output conforms to the measured data; and 2) a physical loss term, calculated based on the motor's state equations, forcibly constraining the model output to always satisfy the fundamental laws of motor operation. By optimizing the parameter identification model through this composite loss function, a target parameter identification model is ultimately obtained that fully utilizes limited data while strictly adhering to physical laws. Since the core training of the model relies on the physical loss, it can learn autonomously from sample data even without a large amount of offline calibration data, thus achieving high-precision, strong-generalization parameter identification under data-scarce conditions.

[0054] It should be understood that, although Figure 1 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 1 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0055] In one embodiment, such as Figure 2 As shown, a multi-parameter identification device for a permanent magnet synchronous motor is provided, comprising: The input module 201 is used to input the acquired dataset into the initial parameter identification model; wherein, the dataset includes motor operating parameters and labels corresponding to the motor operating parameters, and the labels include the measured voltages of the direct axis and quadrature axis when the motor is running; The determination module 202 is used to determine the predicted motor parameters and the predicted voltages of the direct axis and quadrature axis corresponding to the predicted motor parameters in the initial parameter identification model based on the motor operating parameters; wherein the predicted motor parameters include the predicted stator resistance, the predicted inductance of the direct axis and quadrature axis, and the predicted permanent magnet flux linkage. The adjustment module 203 is used to adjust the parameters in the initial parameter identification model based on the loss value of the total loss function until a preset condition is met to obtain the target parameter identification model; wherein, the loss value includes the first loss value of the physical loss function, and the first loss value characterizes the difference between the measured voltage and the predicted voltage; The output module 204 is used to input the obtained motor operating parameters of the target motor into the target parameter identification model to obtain the target motor parameters of the target motor, wherein the target motor parameters include the target stator resistance, the target inductance of the direct axis and the quadrature axis, and the target permanent magnet flux linkage.

[0056] In this embodiment, the specific limitations of the multi-parameter identification device for permanent magnet synchronous motors can be found in the above-described limitations of the multi-parameter identification method for permanent magnet synchronous motors, and will not be repeated here. Each module in the aforementioned multi-parameter identification device for permanent magnet synchronous motors can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0057] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 3As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores multi-parameter identification data for permanent magnet synchronous motors. The network interface communicates with external terminals via a network. When the processor executes the computer program, it implements a multi-parameter identification method for permanent magnet synchronous motors. The display screen can be an LCD screen or an e-ink display screen. The input device can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device casing, or an external keyboard, touchpad, or mouse.

[0058] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the computer device to which the present invention is applied. A specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0059] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any step of the multi-parameter identification method for permanent magnet synchronous motors described in the first aspect.

[0060] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements any step of the multi-parameter identification method for permanent magnet synchronous motors described in the first aspect.

[0061] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0062] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0063] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A method for multi-parameter identification of a permanent magnet synchronous motor, characterized in that, The method includes: The acquired dataset is input into the initial parameter identification model; wherein, the dataset includes motor operating parameters and labels corresponding to the motor operating parameters, and the labels include the measured voltages of the direct axis and quadrature axis when the motor is running; In the initial parameter identification model, based on the motor operating parameters, predicted motor parameters are determined, as well as predicted voltages for the direct and quadrature axes corresponding to the predicted motor parameters; wherein, the predicted motor parameters include predicted stator resistance, predicted inductances for the direct and quadrature axes, and predicted permanent magnet flux linkage; Based on the loss value of the total loss function, the parameters in the initial parameter identification model are adjusted until a preset condition is met to obtain the target parameter identification model; wherein, the loss value includes the first loss value of the physical loss function, and the first loss value characterizes the difference between the measured voltage and the predicted voltage; The obtained motor operating parameters of the target motor are input into the target parameter identification model to obtain the target motor parameters, wherein the target motor parameters include the target stator resistance, the target inductance of the direct axis and the quadrature axis, and the target permanent magnet flux linkage.

2. The method according to claim 1, characterized in that, Determining the predicted voltage corresponding to the predicted motor parameters includes: Based on the motor state equation in the total loss function, the predicted voltages of the direct axis and quadrature axis are generated using the predicted motor parameters.

3. The method according to claim 1, characterized in that, The first loss value of the physical loss function is determined in the following way: Obtain the measured voltage from the tags corresponding to the motor operating parameters; The error between the predicted voltage and the measured voltage is determined using the physical loss function, thus obtaining the first loss value.

4. The method according to claim 1, characterized in that, The total loss function also includes a second loss value from the data loss function; The label also includes reference motor parameters corresponding to the motor operating parameters. The reference motor parameters include reference stator resistance, reference inductance of the direct axis and quadrature axis, and reference permanent magnet flux linkage. The second loss value characterizes the difference between the predicted motor parameters and the reference motor parameters.

5. The method according to claim 4, characterized in that, The total loss function is obtained by weighted summation of the first loss value and the second loss value.

6. The method according to claim 4, characterized in that, The second loss value of the data loss function is determined in the following way: Obtain the reference motor parameters from the labels corresponding to the motor operating parameters; The error between the predicted motor parameters and the reference motor parameters is determined using the data loss function, and the second loss value is obtained.

7. The method according to claim 2, characterized in that, The motor operating parameters include the current feedback values ​​of the direct axis and quadrature axis when the motor is running; The step of generating the predicted voltages for the direct axis and quadrature axis based on the motor state equation in the total loss function and using the predicted motor parameters includes: Using the automatic differentiation function of the initial parameter identification model, the current feedback values ​​of the direct axis and quadrature axis are calculated to obtain the direct axis current differential term and the quadrature axis current differential term; Substituting the direct-axis current differential term, the quadrature-axis current differential term, and the predicted motor parameters into the motor state equation yields the predicted voltages for the direct and quadrature axes.

8. A multi-parameter identification device for a permanent magnet synchronous motor, characterized in that, The device includes: The input module is used to input the acquired dataset into the initial parameter identification model; wherein, the dataset includes motor operating parameters and labels corresponding to the motor operating parameters, and the labels include the measured voltages of the direct axis and quadrature axis when the motor is running; The determination module is used to determine the predicted motor parameters and the predicted voltages of the direct axis and quadrature axis corresponding to the predicted motor parameters in the initial parameter identification model based on the motor operating parameters; wherein the predicted motor parameters include the predicted stator resistance, the predicted inductance of the direct axis and quadrature axis, and the predicted permanent magnet flux linkage; An adjustment module is used to adjust the parameters in the initial parameter identification model based on the loss value of the total loss function until a preset condition is met to obtain the target parameter identification model; wherein, the loss value includes a first loss value of the physical loss function, and the first loss value characterizes the difference between the measured voltage and the predicted voltage; The output module is used to input the obtained motor operating parameters of the target motor into the target parameter identification model to obtain the target motor parameters of the target motor, wherein the target motor parameters include the target stator resistance, the target inductance of the direct axis and the quadrature axis, and the target permanent magnet flux linkage.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 7.

Images