CN117277888A - Method and device for predicting voltage vector of stator of permanent magnet synchronous motor - Google Patents

Abstract

This application provides a method and device for predicting the voltage vector of the stator of a permanent magnet synchronous motor. This method integrates the prior estimation method and the neural network topology prediction method to obtain the optimized predicted current vector, and then constructs a vector expression. Find the minimum value of the vector expression, and determine the voltage of the d-axis of the stator and the voltage of the q-axis of the stator corresponding to the minimum value as the optimal predicted voltage vector to obtain the optimal predicted voltage vector, because it is combined with a priori estimation method and neural network topology prediction method, thereby improving the accuracy of prediction, thereby solving the problem of low accuracy of the motor model in the existing scheme of permanent magnet synchronous motor model predictive control.

Description

Prediction method and device for voltage vector of stator of permanent magnet synchronous motor

Technical field

The present application relates to the technical field of permanent magnet synchronous motors. Specifically, it relates to a method and device for predicting the voltage vector of the stator of a permanent magnet synchronous motor, a control method of a permanent magnet synchronous motor, a computer-readable storage medium and an electronic device.

Background technique

Existing solutions often use corresponding models to predict the changes in the d-axis and q-axis voltages of the stator of the permanent magnet synchronous motor in the future. However, the model used in the traditional permanent magnet synchronous motor model predictive control is a mechanism model. It often cannot truly reflect the changes of the motor in the real world with the environment, and cannot simulate the uncertainty and unknownness of the external environment.

Therefore, there is an urgent need for a prediction method of the voltage vector of the stator of a permanent magnet synchronous motor to solve the problem of low accuracy of the motor model in the existing solution of permanent magnet synchronous motor model predictive control.

Contents of the invention

The main purpose of this application is to provide a method and device for predicting the voltage vector of the stator of a permanent magnet synchronous motor, a control method of a permanent magnet synchronous motor, a computer-readable storage medium and an electronic device, so as to at least solve the existing permanent magnet synchronous solution. The problem of low accuracy of motor model in motor model predictive control.

In order to achieve the above objectives, according to one aspect of the present application, a method for predicting the voltage vector of the stator of a permanent magnet synchronous motor is provided, which method includes:

Obtain the first stator voltage, the second stator voltage and the current rotor angular speed. The first stator voltage is the voltage of the d-axis of the stator of the permanent magnet synchronous motor at the current moment. The second stator voltage is the permanent magnet The voltage of the q-axis of the stator of the synchronous motor at the current moment, and the current rotor angular velocity is the angular velocity of the rotor of the permanent magnet synchronous motor at the current moment; using a priori estimation method, according to the first stator voltage, The second stator voltage and the current rotor angular velocity are used to obtain a first predicted current vector, and a neural network topology model is used to process the first stator voltage, the second stator voltage and the current rotor angular velocity. , obtain the second predicted current vector, wherein the first predicted current vector is a vector including the first q-axis predicted current and the first d-axis predicted current, and the second predicted current vector is a vector including the second q-axis predicted current and the vector of the second d-axis predicted current. The neural network topology model is trained using multiple sets of training data. Each set of training data in the multiple sets of training data includes input data obtained within a historical time period. and output data corresponding to the input data. The input data includes the voltage of the d-axis of the stator, the voltage of the q-axis of the stator, and the angular velocity of the rotor. The output data includes the q-axis of the stator. The predicted current of the d-axis of the stator and the predicted current of the d-axis of the stator; the first predicted current vector and the second predicted current vector are weighted to obtain an optimized predicted current vector, and the optimized predicted current vector includes optimization The vector of the q-axis predicted current and the optimized d-axis predicted current; use the optimized predicted current vector and the stator voltage vector to construct a vector expression, obtain the minimum value of the vector expression, and convert the The voltage of the d-axis of the stator and the voltage of the q-axis of the stator are determined as the optimal predicted voltage vector, and the stator voltage vector is a vector including the first stator voltage and the second stator voltage.

Optionally, a priori estimation method is used to obtain the first predicted current vector according to the first stator voltage, the second stator voltage and the current rotor angular velocity, including: using the first stator voltage, The second stator voltage and the current rotor angular velocity construct a first predicted current vector expression, which includes the inductance value of the q-axis of the stator and the inductance of the d-axis of the stator. The proportional relationship between the value, the proportional relationship between the resistance value of the stator and the inductance value of the q-axis of the stator, the proportional relationship between the resistance value of the stator and the inductance value of the d-axis of the stator, the sampling period and the The first predicted current vector is determined based on the proportional relationship between the inductance value of the q-axis of the stator, the sampling period and the inductance value of the d-axis of the stator, based on the first predicted current vector expression.

Optionally, the neural network in the neural network topology model includes multiple input layers, multiple hidden layers and multiple output layers, with each input layer, each hidden layer and each output layer separated by Fully connected.

Optionally, in the process of using a neural network topology model to process the first stator voltage, the second stator voltage and the current rotor angular velocity to obtain the second predicted current vector, the method further includes :

Construct a d-axis prediction performance expression and a q-axis prediction performance expression. The d-axis prediction performance expression includes the t-th predicted d-axis current obtained by using the neural network topology model and the d-axis current using the neural network topology model. The difference between the d-axis current predicted at the t+1th time obtained, and the q-axis prediction performance expression includes the q-axis current predicted at the t-th time obtained using the neural network topology model and the q-axis current predicted using the neural network The difference in the q-axis current predicted at the t+1th time obtained by the topological model;

A prediction performance vector is determined, which is a vector including d-axis prediction performance and q-axis prediction performance.

Optionally, after determining the predicted performance vector, the method further includes:

according to ,

Determine the connection weights between the neurons of the input layer and the hidden layer and the connections between the neurons of the hidden layer and the output layer when the neural network topology model is used for the t+1th prediction. The connection weight;

in, is the connection weight between the neurons of the input layer and the hidden layer when the neural network topology model is used for the t+1th prediction,/> is the connection weight between the neurons of the hidden layer and the output layer when the neural network topology model is used for the t+1th prediction,/> is the connection weight between the neurons of the input layer and the hidden layer when the neural network topology model is used for the tth prediction,/> is the connection weight between the neurons of the hidden layer and the output layer when the neural network topology model is used for the tth prediction,/> is the preset coefficient, loss is the prediction performance vector, i is used to characterize the i-th neuron of the input layer, j is used to characterize the j-th neuron of the hidden layer, and k is used to characterize the output The kth neuron of the layer.

Optionally, the optimized predicted current vector and the stator voltage vector are used to construct a vector expression, the minimum value of the vector expression is obtained, and the voltage of the d-axis of the stator corresponding to the minimum value and the The voltage of the q-axis of the stator is determined as the optimal predicted voltage vector, including:

Use the optimized predicted current vector and the stator voltage vector to construct a vector expression, find the minimum value of the vector expression, and combine the voltage of the d-axis of the stator and the q-axis of the stator corresponding to the minimum value. The voltage of is determined as the optimal predicted voltage vector, including:

according to ,

Determine the optimal predicted voltage vector, where, is the optimal predicted voltage vector,/> The optimized prediction current vector for the predicted future Nth step,/> for/> The transposed matrix,/> To predict the difference between the optimized predicted current vector and the actual value of the corresponding current vector in step a of the future,/> for/> The transposed matrix,/> is the stator voltage vector,/> for/> The transpose matrix of current.

According to another aspect of the present application, a device for predicting the voltage vector of the stator of a permanent magnet synchronous motor is provided, which device includes:

An acquisition unit is used to acquire the first stator voltage, the second stator voltage and the current rotor angular velocity. The first stator voltage is the voltage of the d-axis of the stator of the permanent magnet synchronous motor at the current moment. The second stator voltage is the voltage of the q-axis of the stator of the permanent magnet synchronous motor at the current moment, and the current rotor angular velocity is the angular velocity of the rotor of the permanent magnet synchronous motor at the current moment;

The first processing unit is configured to use a priori estimation method to obtain the first predicted current vector based on the first stator voltage, the second stator voltage and the current rotor angular velocity, and use a neural network topology model to calculate The first stator voltage, the second stator voltage and the current rotor angular velocity are processed to obtain a second predicted current vector, wherein the first predicted current vector includes a first q-axis predicted current and a first The vector of the d-axis predicted current, the second predicted current vector is a vector including the second q-axis predicted current and the second d-axis predicted current, the neural network topology model is trained using multiple sets of training data, the Each set of training data in the multiple sets of training data includes input data obtained within a historical time period and output data corresponding to the input data. The input data includes the voltage of the d-axis of the stator, the stator The voltage of the q-axis and the angular velocity of the rotor, the output data include the predicted current of the q-axis of the stator and the predicted current of the d-axis of the stator;

A second processing unit configured to weight the first predicted current vector and the second predicted current vector to obtain an optimized predicted current vector. The optimized predicted current vector includes optimized q-axis predicted current and optimized d-axis predict the vector of current;

A third processing unit configured to construct a vector expression using the optimized predicted current vector and the stator voltage vector, obtain the minimum value of the vector expression, and obtain the voltage of the d-axis of the stator corresponding to the minimum value. and the voltage of the q-axis of the stator are determined as the optimal predicted voltage vector, and the stator voltage vector is a vector including the first stator voltage and the second stator voltage.

According to another aspect of the present application, a control method for a permanent magnet synchronous motor is provided. The method includes: obtaining a first stator voltage, a second stator voltage and a current rotor angular speed. The first stator voltage is a permanent magnet The voltage of the d-axis of the stator of the synchronous motor at the current moment, the second stator voltage is the voltage of the q-axis of the stator of the permanent magnet synchronous motor at the current moment, and the current rotor angular speed is the voltage of the permanent magnet synchronous motor. The angular velocity of the motor's rotor at the current moment; using a priori estimation method, obtaining the first predicted current vector based on the first stator voltage, the second stator voltage and the current rotor angular velocity, and using a neural network topology model , the first stator voltage, the second stator voltage and the current rotor angular velocity are processed to obtain a second predicted current vector, wherein the first predicted current vector includes the first q-axis predicted current and The vector of the first d-axis predicted current, the second predicted current vector is a vector including the second q-axis predicted current and the second d-axis predicted current, and the neural network topology model is trained using multiple sets of training data, Each set of training data in the plurality of sets of training data includes input data obtained within a historical time period and output data corresponding to the input data. The input data includes the voltage of the d-axis of the stator, the The voltage of the q-axis of the stator and the angular velocity of the rotor, the output data include the predicted current of the q-axis of the stator and the predicted current of the d-axis of the stator; for the first predicted current vector and the The second predicted current vector is weighted to obtain an optimized predicted current vector. The optimized predicted current vector is a vector including an optimized q-axis predicted current and an optimized d-axis predicted current; a vector is constructed using the optimized predicted current vector and the stator voltage vector. Expression, find the minimum value of the vector expression, and determine the voltage of the d-axis of the stator and the voltage of the q-axis of the stator corresponding to the minimum value as the optimal predicted voltage vector, the stator The voltage vector is a vector including the first stator voltage and the second stator voltage; an inverse coordinate transformation process is performed on the optimal predicted voltage vector to obtain a target three-phase voltage, and the permanent voltage is calculated based on the target three-phase voltage. Magnetic synchronous motor is controlled.

According to another aspect of the present application, a computer-readable storage medium is provided. The computer-readable storage medium includes a stored program, wherein when the program is running, the device where the computer-readable storage medium is located is controlled to execute any arbitrary A method for predicting the voltage vector of the stator of a permanent magnet synchronous motor.

According to another aspect of the present application, an electronic device is provided. The electronic device includes one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory. , and configured to be executed by the one or more processors, the one or more programs include a prediction method for executing any one of the voltage vectors of the stator of the permanent magnet synchronous motor.

Applying the technical solution of this application, by integrating the prior estimation method and the neural network topology prediction method, the optimal predicted current vector is obtained, and then a vector expression is constructed to obtain the optimal predicted voltage vector. Because the prior estimation method and the neural network topology prediction method are combined, the optimal predicted current vector is obtained. The network topology prediction method improves the accuracy of prediction and solves the problem of low accuracy of the motor model in the existing scheme of permanent magnet synchronous motor model predictive control.

Description of the drawings

The description and drawings that constitute a part of this application are used to provide a further understanding of this application. The illustrative embodiments and their descriptions of this application are used to explain this application and do not constitute an improper limitation of this application. In the attached picture:

Figure 1 shows a schematic flow chart of a method for predicting the voltage vector of the stator of a permanent magnet synchronous motor provided according to an embodiment of the present application;

Figure 2 shows the topology schematic diagram of the neural network topology model;

Figure 3 shows a schematic diagram of the process of obtaining an optimized predicted current vector based on the first stator voltage, the second stator voltage and the current rotor angular velocity;

Figure 4 shows a schematic flow chart of a control method for a permanent magnet synchronous motor;

Figure 5 shows a structural block diagram of a device for predicting the voltage vector of the stator of a permanent magnet synchronous motor provided according to an embodiment of the present application;

Figure 6 shows a schematic flow chart of another control method of a permanent magnet synchronous motor.

Detailed ways

It should be noted that, as long as there is no conflict, the embodiments and features in the embodiments of this application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and embodiments.

In order to enable those in the technical field to better understand the solutions of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only These are part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts should fall within the scope of protection of this application.

It should be noted that the terms "first", "second", etc. in the description and claims of this application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that data so used may be interchanged where appropriate for the embodiments of the application described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, e.g., a process, method, system, product, or apparatus that encompasses a series of steps or units and need not be limited to those explicitly listed. Those steps or elements may instead include other steps or elements not expressly listed or inherent to the process, method, product or apparatus.

As introduced in the background art, existing solutions often use corresponding models to predict the changes in the d-axis and q-axis voltages of the stator of the permanent magnet synchronous motor in a period of time in the future. However, the traditional permanent magnet synchronous motor model predictive control uses The model is a mechanism model. The mechanism model often cannot truly reflect the changes of the motor with the environment in the real world, and cannot simulate the uncertainty and unknownness of the external environment. In order to solve the existing solution of permanent magnet synchronous motor model predictive control, the motor model is accurate. To solve the problem of relatively low degree, the embodiments of the present application provide a method and device for predicting the voltage vector of the stator of a permanent magnet synchronous motor, a control method of the permanent magnet synchronous motor, a computer-readable storage medium, and an electronic device.

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

In this embodiment, a method for predicting the voltage vector of the stator of a permanent magnet synchronous motor is provided. It should be noted that the steps shown in the flow chart of the accompanying drawings can be implemented in a computer system such as a set of computer executable instructions. are performed, and, although a logical order is shown in the flowchart diagrams, in some cases the steps shown or described may be performed in a different order than herein.

FIG. 1 is a schematic flowchart of a method for predicting the voltage vector of the stator of a permanent magnet synchronous motor provided according to an embodiment of the present application. As shown in Figure 1, the method includes the following steps:

Step S101: Obtain the first stator voltage, the second stator voltage and the current rotor angular speed. The first stator voltage is the voltage of the d-axis of the stator of the permanent magnet synchronous motor at the current moment. The second stator voltage is the permanent magnet voltage. The voltage of the q-axis of the above-mentioned stator of the synchronous motor at the current moment, and the above-mentioned current rotor angular speed is the angular velocity of the rotor of the above-mentioned permanent magnet synchronous motor at the current moment;

Specifically, the voltage of the q-axis and d-axis of the stator of the permanent magnet synchronous motor at the current moment is helpful to predict the subsequent voltage. Therefore, it is necessary to obtain the voltage of the q-axis and d-axis of the stator at the current moment. The permanent magnet synchronous motor The same is true for the angular velocity of the rotor at the current moment, which is also helpful in predicting the subsequent voltage;

,

in this formula is a matrix including the derivative of the d-axis current and the derivative of the q-axis current of the stator, x is a matrix including the d-axis current and q-axis current of the stator, u is a matrix including the d-axis voltage and q-axis voltage of the stator, A represents System matrix, B represents the control matrix, G is the constant term;/> and/> are the d-axis and q-axis voltages of the stator respectively;/> and/> are the d-axis and q-axis currents of the stator respectively;/> and/> are the resistance value of the stator and the angular velocity of the rotor respectively;/> is the rotor flux linkage, that is, the preset constant. This preset constant is the inherent characteristic of the permanent magnet synchronous motor. If a certain motor is determined, the preset constant of the permanent magnet synchronous motor is also determined;

is the unknown disturbance current parameter related to the state,/> is the unknown disturbance current parameter of the q-axis,/> is the position disturbance current parameter of the d-axis.

Step S102, using a priori estimation method to obtain the first predicted current vector based on the above-mentioned first stator voltage, the above-mentioned second stator voltage and the above-mentioned current rotor angular velocity, and using a neural network topology model to calculate the above-mentioned first stator voltage, The above-mentioned second stator voltage and the above-mentioned current rotor angular velocity are processed to obtain a second predicted current vector, wherein the above-mentioned first predicted current vector is a vector including the first q-axis predicted current and the first d-axis predicted current. The above-mentioned second predicted current vector The current vector is a vector including the second q-axis predicted current and the second d-axis predicted current. The above-mentioned neural network topology model is trained using multiple sets of training data. Each set of training data in the above-mentioned multiple sets of training data includes history. Obtained during the time period: input data and output data corresponding to the above input data. The above input data includes the voltage of the d-axis of the above-mentioned stator, the voltage of the q-axis of the above-mentioned stator and the angular velocity of the above-mentioned rotor. The above-mentioned output data includes the voltage of the above-mentioned stator. The predicted current of the q-axis and the predicted current of the d-axis of the above stator;

Specifically, the prior estimation method is:

Using the above-mentioned first stator voltage, the above-mentioned second stator voltage and the above-mentioned current rotor angular velocity, a first predicted current vector expression is constructed. The above-mentioned first predicted current vector expression includes the inductance value of the q-axis of the above-mentioned stator and the d of the above-mentioned stator. The proportional relationship between the inductance value of the shaft, the proportional relationship between the resistance value of the above-mentioned stator and the inductance value of the q-axis of the above-mentioned stator, the proportional relationship between the resistance value of the above-mentioned stator and the inductance value of the d-axis of the above-mentioned stator, the sampling period and the The first predicted current vector is determined based on the proportional relationship between the inductance value of the q-axis, the sampling period and the inductance value of the d-axis of the stator, based on the first predicted current vector expression.

One specific embodiment (without considering unknown disturbances) is:

according to ,

Determine the first predicted current vector above, where, is the predicted current of the d-axis of the above-mentioned stator at time k+1 obtained using the above-mentioned a priori estimation method,/> is the sampling period,/> is the resistance value of the above stator,/> is the inductance value of the d-axis of the above stator,/> is the predicted current of the d-axis of the above-mentioned stator at time k obtained using the above-mentioned a priori estimation method,/> is the inductance value of the q-axis of the above stator,/> is the angular velocity of the above rotor at time k,/> is the voltage of the d-axis of the above stator at time k,/> is the voltage of the q-axis of the above-mentioned stator at time k obtained using the above-mentioned a priori estimation method, is a default constant,/> is the predicted current of the q-axis of the above-mentioned stator at time k obtained using the above-mentioned a priori estimation method, is the predicted current of the q-axis of the above stator at time k+1;

The current predicted in the previous round is used to calculate the current predicted in the current round, thereby improving the accuracy of calculating the current predicted in the current round.

In addition, regarding the neural network topology model, first, based on a determined model of permanent magnet synchronous motor, a motor open-loop control test was conducted on a motor towing platform to collect training data. When the DC bus voltage is determined, first the control input quantities of the first stator voltage, the second stator voltage and the current rotor angular velocity are discretized. Among them, the discrete interval of the first stator voltage and the second stator voltage is 5V, which is required to be voltage range across all stator d- and q-axes. The discrete interval of the current rotor angular speed of the load motor is 20 rpm, taken throughout the speed range of the motor under test. Write a script language (Python or MATLAB) on the host computer to input different first stator voltage, second stator voltage and current rotor angular speed in an automated manner. The script automatically outputs the predicted current of the d-axis and the predicted current of the q-axis. , different inputs and outputs constitute the entire data set.

Data preprocessing: Collect the data, eliminate data outliers, and normalize the data. Among them, the data normalization method uses the minimum value and maximum value method, the expression:

;

in, is the normalized data,/> is the original data in the data sequence (i.e., the matrix including the d-axis current and q-axis current of the stator), /> is the minimum value in the data sequence,/> is the maximum value in the data sequence.

As shown in Figure 2, the neural network in the above-mentioned neural network topology model includes multiple input layers, multiple hidden layers and multiple output layers. Each of the above-mentioned input layers, each of the above-mentioned hidden layers and each of the above-mentioned output layers are fully connected. The input layer needs to input three values, which are the first stator voltage, the second stator voltage and the current rotor angular speed. The output layer needs to output two values, which are the second q-axis predicted current and the second d Axis predicted current.

Wherein, in the process of using a neural network topology model to process the above-mentioned first stator voltage, the above-mentioned second stator voltage and the above-mentioned current rotor angular velocity to obtain the second predicted current vector, the above method also includes:

Construct a d-axis prediction performance expression and a q-axis prediction performance expression. The above d-axis prediction performance expression includes the t-th predicted d-axis current obtained by using the above-mentioned neural network topology model and the t-th predicted current obtained by using the above-mentioned neural network topology model. The difference between the d-axis current predicted at t+1 time. The above q-axis prediction performance expression includes the t-th predicted q-axis current obtained by using the above neural network topology model and the t-th predicted current obtained by using the above neural network topology model. +1 predicted difference in q-axis current;

Determine a prediction performance vector, which is a vector including d-axis prediction performance and q-axis prediction performance.

It is guaranteed to obtain the d-axis prediction performance and the q-axis prediction performance respectively, thereby improving the accuracy of using these two prediction performances to obtain the corresponding weights.

One of the specific embodiments is:

according to , determine the d-axis prediction performance. The above-mentioned d-axis prediction performance is the performance of the above-mentioned neural network topology model in predicting the d-axis current of the above-mentioned stator, where lossd is the above-mentioned d-axis prediction performance, is the d-axis current predicted at the tth time using the above neural network topology model,/> is the d-axis current predicted at the t+1th time using the above neural network topology model;

according to , determine the q-axis prediction performance. The above-mentioned q-axis prediction performance is the performance of the above-mentioned neural network topology model in predicting the q-axis current of the above-mentioned stator, where lossq is the above-mentioned d-axis prediction performance, is the t-th predicted q-axis current obtained using the above neural network topology model,/> is the q-axis current predicted at the t+1th time using the above neural network topology model;

Determine the prediction performance vector, which includes d-axis prediction performance and q-axis prediction performance, to improve the prediction accuracy of the neural network topology model.

In addition, after determining the prediction performance vector, the above method also includes:

according to ,

Determine the connection weights between the neurons of the above-mentioned input layer and the above-mentioned hidden layer and the connection weights between the neurons of the above-mentioned hidden layer and the above-mentioned output layer when the above-mentioned neural network topology model is used for the t+1th prediction;

Among them, first, determine the number of hidden layers, the number of nodes in each hidden layer, the threshold of hidden layer neurons, and the threshold of output layer neurons. Set training parameters, including learning rate, maximum number of iterations, and minimum tolerance error. Then, loss functions lossd and lossq are constructed to evaluate the prediction performance of the neural network. The smaller the loss, the better the prediction performance of the neural network. The activation functions of the neurons in the hidden layer and output layer of the neural network all use sigmoid ( That is, the S-shaped function) type activation function that is common in biology.

is the connection weight between the neurons of the above-mentioned input layer and the above-mentioned hidden layer when making the t+1th prediction using the above-mentioned neural network topology model,/> is the connection weight between the neurons of the above-mentioned hidden layer and the above-mentioned output layer when the above-mentioned neural network topology model is used for the t+1th prediction,/> is the connection weight between the neurons of the above-mentioned input layer and the above-mentioned hidden layer when the above-mentioned neural network topology model is used for the t-th prediction,/> is the connection weight between the neurons of the above-mentioned hidden layer and the above-mentioned output layer when the above-mentioned neural network topology model is used for the t-th prediction,/> is the preset coefficient, loss is the above-mentioned prediction performance vector, i is used to characterize the i-th neuron of the above-mentioned input layer, j is used to characterize the j-th neuron of the above-mentioned hidden layer, and k is used to characterize the k-th neuron of the above-mentioned output layer. neurons.

From this, the connection weights between the neurons of the above-mentioned hidden layer and the above-mentioned output layer and the connection weights between the above-mentioned input layer and the neurons of the above-mentioned hidden layer are obtained, so that the parameters input to the input layer can be input according to the weights, The parameters converted into the output layer improve the prediction accuracy of the neural network topology model.

Step S103, perform weighting processing on the above-mentioned first predicted current vector and the above-mentioned second predicted current vector to obtain an optimized predicted current vector. The above-mentioned optimized predicted current vector is a vector including an optimized q-axis predicted current and an optimized d-axis predicted current;

Specifically, the optimized predicted current vector is an optimized predicted current vector obtained by weighting the above-mentioned first predicted current vector and the above-mentioned second predicted current vector. By weighting the above-mentioned first predicted current vector and the above-mentioned second predicted current vector Weighting processing is performed, thereby achieving the purpose of integrating the prior estimation method and the neural network topology prediction method, thereby improving the accuracy of predicting the current vector, as shown in Figure 3. Figure 3 shows how to predict the current vector according to the first stator voltage, The above-mentioned second stator voltage and the above-mentioned current rotor angular velocity are used to optimize the process of predicting the current vector.

One specific embodiment is,

1) First write the mathematical model of the permanent magnet synchronous motor in the form of a state space expression, and the a priori predicted value (That is, the first predicted current vector corresponding to step a):

;

is the first predicted current vector of step a-1,/> is the voltage of the d-axis of the stator and the voltage of the q-axis of the stator corresponding to the minimum value of the vector expression in step a-1;

2) Calculate the covariance matrix of the a priori estimation error in step a :

;

is the covariance matrix of the posterior estimation error in step a-1, and W1 is a constant term;

A ^T is the transpose matrix of A;

3) Calculate the system gain K a in step _a :

;

H is the measurement matrix, which is the preset matrix, H ^T is the transpose matrix of H, and W2 is also a constant term;

4) Calculate the optimal estimate of step a , that is, the optimized predicted current vector of step a:

;

z _a is the observed value, which is the second predicted current vector in step a;

5) Obtain the covariance matrix of the posterior estimation error in step a :

;

I is the identity matrix.

Step S104, use the above-mentioned optimized predicted current vector and the stator voltage vector to construct a vector expression, find the minimum value of the above-mentioned vector expression, and combine the voltage of the d-axis of the above-mentioned stator and the voltage of the q-axis of the above-mentioned stator corresponding to the above-mentioned minimum value Determined as the optimal predicted voltage vector, the stator voltage vector is a vector including the first stator voltage and the second stator voltage.

The voltage of the d-axis of the stator and the voltage of the q-axis of the stator corresponding to the minimum value of the vector expression are determined as the optimal predicted voltage vector;

In the above method, by integrating the prior estimation method and the neural network topology prediction method, the optimal predicted current vector is obtained, and then a vector expression is constructed to obtain the optimal predicted voltage vector, because the prior estimation method and the neural network topology prediction are combined. This method improves the accuracy of prediction and solves the problem of low accuracy of the motor model in the existing scheme of permanent magnet synchronous motor model predictive control.

Specifically, according to ,

Determine the above optimal predicted voltage vector, where, is the above optimal predicted voltage vector,/> Predict the current vector for the above-mentioned optimization of the predicted Nth step in the future,/> for/> The transposed matrix,/> To predict the difference between the above-mentioned optimized predicted current vector and the actual value of the corresponding current vector in step a in the future,/> for/> The transposed matrix,/> is the above stator voltage vector,/> for/> The transpose matrix of , P, R, and Q are respectively preset weight matrices. The actual value of the above-mentioned current vector is the current of the d-axis of the above-mentioned stator and the current of the q-axis of the above-mentioned stator fed back in real time by the above-mentioned permanent magnet synchronous motor.

Among them, by considering the above-mentioned optimized predicted current vector predicted at the nth step in the future, the difference between the above-mentioned optimized predicted current vector predicted at the a-th step in the future and the actual value of the corresponding current vector, and the stator voltage vector, thereby improving the optimal Accuracy of predicting voltage vectors.

By fusing two models (i.e., the model corresponding to the prior estimation method and the neural network topology model) to obtain a model, and then using the fused model to perform model predictive control of the permanent magnet synchronous motor current loop, a better prediction can be obtained voltage vector (d-axis and q-axis voltage), thereby improving control accuracy.

In order to enable those skilled in the art to more clearly understand the technical solution of the present application, the implementation process of the present application's method for predicting the voltage vector of the stator of a permanent magnet synchronous motor will be described in detail in conjunction with specific embodiments.

This embodiment relates to a specific control method of a permanent magnet synchronous motor, as shown in Figure 4, including the following steps:

Step S1: Obtain the first stator voltage, the second stator voltage and the current rotor angular speed. The above-mentioned first stator voltage is the voltage of the d-axis of the stator of the permanent magnet synchronous motor at the current moment. The above-mentioned second stator voltage is the above-mentioned permanent magnet. The voltage of the q-axis of the above-mentioned stator of the synchronous motor at the current moment, and the above-mentioned current rotor angular speed is the angular velocity of the rotor of the above-mentioned permanent magnet synchronous motor at the current moment;

Step S2: Use a priori estimation method to obtain the first predicted current vector based on the above-mentioned first stator voltage, the above-mentioned second stator voltage and the above-mentioned current rotor angular velocity, and use a neural network topology model to calculate the above-mentioned first stator voltage, The above-mentioned second stator voltage and the above-mentioned current rotor angular velocity are processed to obtain a second predicted current vector, wherein the above-mentioned first predicted current vector is a vector including the first q-axis predicted current and the first d-axis predicted current. The above-mentioned second predicted current vector The current vector is a vector including the second q-axis predicted current and the second d-axis predicted current. The above-mentioned neural network topology model is trained using multiple sets of training data. Each set of training data in the above-mentioned multiple sets of training data includes history. Obtained during the time period: input data and output data corresponding to the above input data. The above input data includes the voltage of the d-axis of the above-mentioned stator, the voltage of the q-axis of the above-mentioned stator and the angular velocity of the above-mentioned rotor. The above-mentioned output data includes the voltage of the above-mentioned stator. The predicted current of the q-axis and the predicted current of the d-axis of the above stator;

Step S3: Perform weighting processing on the above-mentioned first predicted current vector and the above-mentioned second predicted current vector to obtain an optimized predicted current vector. The above-mentioned optimized predicted current vector is a vector including an optimized q-axis predicted current and an optimized d-axis predicted current;

Step S4: Use the above-mentioned optimized predicted current vector and stator voltage vector to construct a vector expression, find the minimum value of the above-mentioned vector expression, and compare the voltage of the d-axis of the above-mentioned stator and the voltage of the q-axis of the above-mentioned stator corresponding to the above-mentioned minimum value. Determined as the optimal predicted voltage vector, the above-mentioned stator voltage vector is a vector including the above-mentioned first stator voltage and the above-mentioned second stator voltage;

Step S5: Control the above-mentioned permanent magnet synchronous motor based on the above-mentioned optimal predicted voltage vector.

It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and, although a logical sequence is shown in the flowchart, in some cases, The steps shown or described may be performed in a different order than here.

The embodiment of the present application also provides a device for predicting the voltage vector of the stator of the permanent magnet synchronous motor. It should be noted that the device of predicting the voltage vector of the stator of the permanent magnet synchronous motor according to the embodiment of the present application can be used to execute the present application. The embodiment provides a prediction method for the voltage vector of the stator of a permanent magnet synchronous motor. This device is used to implement the above embodiments and preferred implementations, and what has been described will not be described again. As used below, the term "module" may be a combination of software and/or hardware that implements a predetermined function. Although the apparatus described in the following embodiments is preferably implemented in software, implementation in hardware, or a combination of software and hardware, is also possible and contemplated.

The following is an introduction to the prediction device of the voltage vector of the stator of the permanent magnet synchronous motor provided by the embodiment of the present application.

FIG. 5 is a structural block diagram of a device for predicting the voltage vector of the stator of a permanent magnet synchronous motor provided according to an embodiment of the present application. As shown in Figure 5, the device includes:

The acquisition unit 51 is used to acquire the first stator voltage, the second stator voltage and the current rotor angular velocity. The first stator voltage is the voltage of the d-axis of the stator of the permanent magnet synchronous motor at the current moment. The second stator voltage is The voltage of the q-axis of the above-mentioned stator of the above-mentioned permanent magnet synchronous motor at the current moment, and the above-mentioned current rotor angular speed is the angular velocity of the rotor of the above-mentioned permanent magnet synchronous motor at the current moment;

The first processing unit 52 is configured to use a priori estimation method to obtain the first predicted current vector based on the above-mentioned first stator voltage, the above-mentioned second stator voltage and the above-mentioned current rotor angular velocity, and use a neural network topology model to calculate the above-mentioned third current vector. The stator voltage, the above-mentioned second stator voltage and the above-mentioned current rotor angular velocity are processed to obtain a second predicted current vector, wherein the above-mentioned first predicted current vector is a vector including the first q-axis predicted current and the first d-axis predicted current. , the above-mentioned second predicted current vector is a vector including the second q-axis predicted current and the second d-axis predicted current. The above-mentioned neural network topology model is trained using multiple sets of training data. Each set of the above-mentioned multiple sets of training data The training data all include: input data obtained during the historical time period and output data corresponding to the above input data. The above input data includes the voltage of the d-axis of the above-mentioned stator, the voltage of the q-axis of the above-mentioned stator and the angular velocity of the above-mentioned rotor. The above-mentioned output data The data includes the predicted current of the q-axis of the above-mentioned stator and the predicted current of the d-axis of the above-mentioned stator;

The second processing unit 53 is configured to weight the first predicted current vector and the second predicted current vector to obtain an optimized predicted current vector. The optimized predicted current vector includes an optimized q-axis predicted current and an optimized d-axis predicted current. vector;

The third processing unit 54 is used to construct a vector expression using the above-mentioned optimized predicted current vector and the stator voltage vector, find the minimum value of the above-mentioned vector expression, and combine the voltage of the d-axis of the above-mentioned stator corresponding to the above-mentioned minimum value and the above-mentioned stator voltage The voltage of the q-axis is determined as the optimal predicted voltage vector, and the above-mentioned stator voltage vector is a vector including the above-mentioned first stator voltage and the above-mentioned second stator voltage.

In the above device, by integrating the prior estimation method and the neural network topology prediction method, the optimal predicted current vector is obtained, and then a vector expression is constructed to obtain the optimal predicted voltage vector. Because the prior estimation method and the neural network topology prediction are combined, This method improves the accuracy of prediction and solves the problem of low accuracy of the motor model in the existing scheme of permanent magnet synchronous motor model predictive control.

In one embodiment of the present application, the first processing unit includes a first processing module and a second processing module, and the first processing module is used to use the above-mentioned first stator voltage, the above-mentioned second stator voltage and the above-mentioned current rotor angular velocity, Construct a first predicted current vector expression. The first predicted current vector expression includes the proportional relationship between the inductance value of the q-axis of the above-mentioned stator and the inductance value of the d-axis of the above-mentioned stator, the resistance value of the above-mentioned stator and the q-axis value of the above-mentioned stator. The proportional relationship between the inductance value, the proportional relationship between the resistance value of the above-mentioned stator and the inductance value of the d-axis of the above-mentioned stator, the proportional relationship between the sampling period and the inductance value of the q-axis of the above-mentioned stator, the ratio between the above-mentioned sampling period and the d-axis value of the above-mentioned stator The proportional relationship of the inductance value; the second processing module is used to determine the first predicted current vector according to the first predicted current vector expression.

In one embodiment of the present application, the neural network in the above-mentioned neural network topology model includes multiple input layers, multiple hidden layers and multiple output layers. Each of the above-mentioned input layers, each of the above-mentioned hidden layers and each of the above-mentioned output layers are fully connected.

In one embodiment of the present application, the first processing unit includes a third processing module and a fourth processing module. Using a neural network topology model, the above-mentioned first stator voltage, the above-mentioned second stator voltage and the above-mentioned current rotor angular velocity are calculated. In the process of processing to obtain the second predicted current vector,

The third processing module is used to construct a d-axis prediction performance expression and a q-axis prediction performance expression. The above d-axis prediction performance expression includes the t-th predicted d-axis current obtained by using the above-mentioned neural network topology model and the t-th predicted current using the above-mentioned neural network topology model. The difference between the t+1th predicted d-axis current obtained by the network topology model. The above q-axis prediction performance expression includes the t-th predicted q-axis current obtained by using the above neural network topology model and the difference between the tth predicted q-axis current obtained by using the above neural network. The difference in the q-axis current predicted at the t+1th time obtained by the topological model;

The fourth processing module is used to determine a prediction performance vector, where the prediction performance vector is a vector including d-axis prediction performance and q-axis prediction performance.

In an embodiment of the present application, the first processing unit includes a first determination module, and after determining the prediction performance vector,

The first determination module is used according to ,

Determine the connection weights between the neurons of the above-mentioned input layer and the above-mentioned hidden layer and the connection weights between the neurons of the above-mentioned hidden layer and the above-mentioned output layer when the above-mentioned neural network topology model is used for the t+1th prediction;

is the connection weight between the neurons of the above-mentioned input layer and the above-mentioned hidden layer when making the t+1th prediction using the above-mentioned neural network topology model,/> is the connection weight between the neurons of the above-mentioned hidden layer and the above-mentioned output layer when the above-mentioned neural network topology model is used for the t+1th prediction,/> is the connection weight between the neurons of the above-mentioned input layer and the above-mentioned hidden layer when the above-mentioned neural network topology model is used for the t-th prediction,/> is the connection weight between the neurons of the above-mentioned hidden layer and the above-mentioned output layer when the above-mentioned neural network topology model is used for the t-th prediction,/> is the preset coefficient, loss is the above-mentioned prediction performance vector, i is used to characterize the i-th neuron of the above-mentioned input layer, j is used to characterize the j-th neuron of the above-mentioned hidden layer, and k is used to characterize the k-th neuron of the above-mentioned output layer. neurons.

In an embodiment of the present application, the third processing module includes a second determination module;

The second determination module is used according to ,

Determine the above optimal predicted voltage vector, where, is the above optimal predicted voltage vector,/> Predict the current vector for the above-mentioned optimization of the predicted Nth step in the future,/> for/> The transposed matrix,/> To predict the difference between the above-mentioned optimized predicted current vector and the actual value of the corresponding current vector in step a in the future,/> for/> The transposed matrix,/> is the above stator voltage vector,/> for/> The transpose matrix of , P, R, and Q are respectively preset weight matrices. The actual value of the above-mentioned current vector is the current of the d-axis of the above-mentioned stator and the current of the q-axis of the above-mentioned stator fed back in real time by the above-mentioned permanent magnet synchronous motor.

The above-mentioned device for predicting the voltage vector of the stator of a permanent magnet synchronous motor includes a processor and a memory. The above-mentioned acquisition unit, first processing unit, second processing unit, third processing unit, etc. are all stored in the memory as program units and are controlled by the processor. The above program units stored in the memory are executed to implement corresponding functions. The above-mentioned modules are all located in the same processor; or, the above-mentioned modules are located in different processors in any combination.

The processor contains a core, which retrieves the corresponding program unit from the memory. One or more kernels can be set, and the problem of low accuracy of the motor model in the existing scheme of permanent magnet synchronous motor model predictive control can be solved by adjusting the kernel parameters.

Memory may include non-permanent memory in computer-readable media, random access memory (RAM) and/or non-volatile memory, such as read-only memory (ROM) or flash memory (flash RAM). The memory includes at least one memory chip.

An embodiment of the present invention provides a computer-readable storage medium. The computer-readable storage medium includes a stored program, wherein when the program is running, the device where the computer-readable storage medium is located is controlled to execute the stator of the permanent magnet synchronous motor. Prediction methods for voltage vectors.

An embodiment of the present invention provides a processor. The processor is configured to run a program. When the program is running, the prediction method for the voltage vector of the stator of the permanent magnet synchronous motor is executed.

An embodiment of the present invention provides a device. The device includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, it implements at least the following steps: obtaining the first stator voltage, the second stator voltage, and the second stator voltage. The stator voltage and the current rotor angular speed. The above-mentioned first stator voltage is the voltage of the d-axis of the stator of the permanent magnet synchronous motor at the current moment. The above-mentioned second stator voltage is the voltage of the q-axis of the above-mentioned stator of the above-mentioned permanent magnet synchronous motor at the current moment. voltage, the above-mentioned current rotor angular velocity is the angular velocity of the rotor of the above-mentioned permanent magnet synchronous motor at the current moment; using a priori estimation method, the first predicted current is obtained based on the above-mentioned first stator voltage, the above-mentioned second stator voltage and the above-mentioned current rotor angular velocity vector, and uses a neural network topology model to process the above-mentioned first stator voltage, the above-mentioned second stator voltage and the above-mentioned current rotor angular velocity to obtain a second predicted current vector, wherein the above-mentioned first predicted current vector includes the first q axis prediction current and the vector of the first d-axis prediction current. The above-mentioned second prediction current vector is a vector including the second q-axis prediction current and the second d-axis prediction current. The above-mentioned neural network topology model is trained using multiple sets of training data. , each set of training data in the above-mentioned multiple sets of training data includes: input data obtained during the historical time period and output data corresponding to the above-mentioned input data. The above-mentioned input data includes the voltage of the d-axis of the above-mentioned stator, the voltage of the above-mentioned stator. The voltage of the q-axis and the angular velocity of the rotor, the output data include the predicted current of the q-axis of the above-mentioned stator and the predicted current of the d-axis of the above-mentioned stator; perform weighting processing on the above-mentioned first predicted current vector and the above-mentioned second predicted current vector, Obtain the optimized predicted current vector. The above optimized predicted current vector is a vector including the optimized q-axis predicted current and the optimized d-axis predicted current. Use the above optimized predicted current vector and the stator voltage vector to construct a vector expression and obtain the minimum of the above vector expression. value, and determine the d-axis voltage of the stator and the q-axis voltage of the stator corresponding to the minimum value as the optimal predicted voltage vector, and the stator voltage vector includes the first stator voltage and the second stator voltage. vector. The devices in this article can be servers, PCs, PADs, mobile phones, etc.

The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program initialized with at least the following method steps: obtaining the first stator voltage, the second stator voltage and the current rotor angular velocity, the above-mentioned third The stator voltage is the voltage of the d-axis of the stator of the permanent magnet synchronous motor at the current moment. The above-mentioned second stator voltage is the voltage of the q-axis of the above-mentioned stator of the above-mentioned permanent magnet synchronous motor at the current moment. The above-mentioned current rotor angular speed is the above-mentioned permanent magnet synchronous motor. The angular velocity of the rotor of the magnetic synchronous motor at the current moment; using a priori estimation method, the first predicted current vector is obtained based on the above-mentioned first stator voltage, the above-mentioned second stator voltage and the above-mentioned current rotor angular velocity, and a neural network topology model is used, The above-mentioned first stator voltage, the above-mentioned second stator voltage and the above-mentioned current rotor angular velocity are processed to obtain a second predicted current vector, wherein the above-mentioned first predicted current vector includes a first q-axis predicted current and a first d-axis predicted current. The vector of the current. The above-mentioned second predicted current vector is a vector including the second q-axis predicted current and the second d-axis predicted current. The above-mentioned neural network topology model is trained using multiple sets of training data. In the above-mentioned multiple sets of training data, Each set of training data includes input data obtained during the historical time period and output data corresponding to the above input data. The above input data includes the voltage of the d-axis of the above-mentioned stator, the voltage of the q-axis of the above-mentioned stator and the angular velocity of the above-mentioned rotor. , the above-mentioned output data includes the predicted current of the q-axis of the above-mentioned stator and the predicted current of the d-axis of the above-mentioned stator; the above-mentioned first predicted current vector and the above-mentioned second predicted current vector are weighted to obtain an optimized predicted current vector, the above-mentioned optimized prediction The current vector is a vector including the optimized q-axis predicted current and the optimized d-axis predicted current; use the above optimized predicted current vector and the stator voltage vector to construct a vector expression, find the minimum value of the above vector expression, and convert the above minimum value corresponding to The voltage of the d-axis of the stator and the voltage of the q-axis of the stator are determined as the optimal predicted voltage vector, and the stator voltage vector is a vector including the first stator voltage and the second stator voltage.

This application also provides a control method for a permanent magnet synchronous motor, as shown in Figure 6. The method includes:

Step S601: Obtain the first stator voltage, the second stator voltage and the current rotor angular speed. The first stator voltage is the voltage of the d-axis of the stator of the permanent magnet synchronous motor at the current moment. The second stator voltage is the permanent magnet voltage. The voltage of the q-axis of the above-mentioned stator of the synchronous motor at the current moment, and the above-mentioned current rotor angular speed is the angular velocity of the rotor of the above-mentioned permanent magnet synchronous motor at the current moment;

Step S602, using a priori estimation method to obtain the first predicted current vector based on the above-mentioned first stator voltage, the above-mentioned second stator voltage and the above-mentioned current rotor angular velocity, and using a neural network topology model to calculate the above-mentioned first stator voltage, The above-mentioned second stator voltage and the above-mentioned current rotor angular velocity are processed to obtain a second predicted current vector, wherein the above-mentioned first predicted current vector is a vector including the first q-axis predicted current and the first d-axis predicted current. The above-mentioned second predicted current vector The current vector is a vector including the second q-axis predicted current and the second d-axis predicted current. The above-mentioned neural network topology model is trained using multiple sets of training data. Each set of training data in the above-mentioned multiple sets of training data includes history. Obtained during the time period: input data and output data corresponding to the above input data. The above input data includes the voltage of the d-axis of the above-mentioned stator, the voltage of the q-axis of the above-mentioned stator and the angular velocity of the above-mentioned rotor. The above-mentioned output data includes the voltage of the above-mentioned stator. The predicted current of the q-axis and the predicted current of the d-axis of the above stator;

Step S603: Perform weighting processing on the above-mentioned first predicted current vector and the above-mentioned second predicted current vector to obtain an optimized predicted current vector. The above-mentioned optimized predicted current vector is a vector including an optimized q-axis predicted current and an optimized d-axis predicted current;

Step S604: Construct a vector expression using the optimized predicted current vector and the stator voltage vector, find the minimum value of the vector expression, and combine the voltage of the d-axis of the stator and the voltage of the q-axis of the stator corresponding to the minimum value. Determined as the optimal predicted voltage vector, the above-mentioned stator voltage vector is a vector including the above-mentioned first stator voltage and the above-mentioned second stator voltage;

Step S605: Perform coordinate inverse transformation processing on the optimal predicted voltage vector to obtain the target three-phase voltage, and control the permanent magnet synchronous motor based on the target three-phase voltage.

In the above method, by integrating the prior estimation method and the neural network topology prediction method, the optimal predicted current vector is obtained, and then a vector expression is constructed to obtain the optimal predicted voltage vector, because the prior estimation method and the neural network topology prediction are combined. This method improves the accuracy of prediction and solves the problem of low accuracy of the motor model in the existing scheme of permanent magnet synchronous motor model predictive control.

This application also provides an electronic device. The electronic device includes one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and are configured to be configured by the above One or more processors execute the above one or more programs including a method for executing any one of the above prediction methods for the voltage vector of the stator of the permanent magnet synchronous motor. By integrating the prior estimation method and the neural network topology prediction method, the optimal predicted current vector is obtained, and then a vector expression is constructed to obtain the optimal predicted voltage vector. Because the prior estimation method and the neural network topology prediction method are combined, the optimization is improved. It improves the accuracy of prediction and solves the problem of low accuracy of the motor model in the existing scheme of permanent magnet synchronous motor model predictive control.

Obviously, those skilled in the art should understand that the above-mentioned modules or steps of the present invention can be implemented using general-purpose computing devices. They can be concentrated on a single computing device, or distributed across a network composed of multiple computing devices. They may be implemented in program code executable by a computing device, such that they may be stored in a storage device for execution by the computing device, and in some cases may be executed in a sequence different from that shown herein. Or the described steps can be implemented by making them into individual integrated circuit modules respectively, or by making multiple modules or steps among them into a single integrated circuit module. As such, the invention is not limited to any specific combination of hardware and software.

Those skilled in the art will understand that embodiments of the present application may be provided as methods, systems, or computer program products. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment that combines software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each process and/or block in the flowchart illustrations and/or block diagrams, and combinations of processes and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that the instructions executed by the processor of the computer or other programmable data processing device produce a use A device for realizing the functions specified in a process or processes in a flowchart and/or a block or blocks in a block diagram.

These computer program instructions may also be stored in a computer-readable memory that causes a computer or other programmable data processing apparatus to operate in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including the instruction means, the instructions The device implements the functions specified in a process or processes in the flowchart and/or in a block or blocks in the block diagram.

These computer program instructions may also be loaded onto a computer or other programmable data processing device, causing a series of operating steps to be performed on the computer or other programmable device to produce computer-implemented processing, thereby executing on the computer or other programmable device. Instructions provide steps for implementing the functions specified in a process or processes of a flowchart diagram and/or a block or blocks of a block diagram.

In a typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.

Memory may include non-volatile memory in computer-readable media, random access memory (RAM), and/or non-volatile memory in the form of read-only memory (ROM) or flash memory (flash RAM). Memory is an example of a computer-readable medium.

Computer-readable media includes both persistent and non-volatile, removable and non-removable media that can be implemented by any method or technology for storage of information. Information may be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), and read-only memory. (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, Magnetic tape cassettes, tape magnetic disk storage or other magnetic storage devices or any other non-transmission medium can be used to store information that can be accessed by a computing device. As defined in this article, computer-readable media does not include transitory media, such as modulated data signals and carrier waves.

It should also be noted that the terms "comprises," "comprises," or any other variation thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements not only includes those elements, but also includes Other elements are not expressly listed or are inherent to the process, method, article or equipment. Without further limitation, an element qualified by the statement "comprises a..." does not exclude the presence of additional identical elements in the process, method, good, or device that includes the element.

From the above description, it can be seen that the above-mentioned embodiments of the present application achieve the following technical effects:

1) The prediction method of the voltage vector of the stator of the permanent magnet synchronous motor in this application combines the prior estimation method and the neural network topology prediction method to obtain the optimized predicted current vector, and then constructs a vector expression to obtain the optimal prediction. The voltage vector combines the a priori estimation method and the neural network topology prediction method, thereby improving the accuracy of prediction, thereby solving the problem of low accuracy of the motor model in the existing scheme of permanent magnet synchronous motor model predictive control.

2) The device for predicting the voltage vector of the stator of a permanent magnet synchronous motor in this application integrates the prior estimation method and the neural network topology prediction method to obtain the optimized predicted current vector, and then constructs a vector expression to obtain the optimal prediction. The voltage vector combines the a priori estimation method and the neural network topology prediction method, thereby improving the accuracy of prediction, thereby solving the problem of low accuracy of the motor model in the existing scheme of permanent magnet synchronous motor model predictive control.

The above descriptions are only preferred embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various modifications and changes. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of this application shall be included in the protection scope of this application.

Claims (10)

1. A method for predicting the voltage vector of the stator of a permanent magnet synchronous motor, which is characterized by including: Obtain the first stator voltage, the second stator voltage and the current rotor angular speed. The first stator voltage is the voltage of the d-axis of the stator of the permanent magnet synchronous motor at the current moment. The second stator voltage is the permanent magnet The voltage of the q-axis of the stator of the synchronous motor at the current moment, and the current rotor angular velocity is the angular velocity of the rotor of the permanent magnet synchronous motor at the current moment; Using a priori estimation method, the first predicted current vector is obtained based on the first stator voltage, the second stator voltage and the current rotor angular velocity, and a neural network topology model is used to estimate the first stator voltage , the second stator voltage and the current rotor angular velocity are processed to obtain a second predicted current vector, wherein the first predicted current vector is a vector including the first q-axis predicted current and the first d-axis predicted current, The second predicted current vector is a vector including a second q-axis predicted current and a second d-axis predicted current. The neural network topology model is trained using multiple sets of training data. Each of the multiple sets of training data A set of training data includes input data obtained within a historical time period and output data corresponding to the input data. The input data includes the voltage of the d-axis of the stator, the voltage of the q-axis of the stator, and the output data corresponding to the input data. The angular velocity of the rotor, the output data includes the predicted current of the q-axis of the stator and the predicted current of the d-axis of the stator; Perform weighting processing on the first predicted current vector and the second predicted current vector to obtain an optimized predicted current vector, where the optimized predicted current vector is a vector including an optimized q-axis predicted current and an optimized d-axis predicted current; Use the optimized predicted current vector and the stator voltage vector to construct a vector expression, find the minimum value of the vector expression, and combine the voltage of the d-axis of the stator and the q-axis of the stator corresponding to the minimum value. The voltage of is determined as the optimal predicted voltage vector, and the stator voltage vector is a vector including the first stator voltage and the second stator voltage. 2. The method according to claim 1, characterized in that, according to the first stator voltage, the second stator voltage and the current rotor angular velocity, the first predicted current vector is obtained, including: Using the first stator voltage, the second stator voltage and the current rotor angular velocity, a first predicted current vector expression is constructed, where the first predicted current vector expression includes the inductance value of the q-axis of the stator The proportional relationship between the resistance value of the stator and the inductance value of the d-axis of the stator, the resistance value of the stator and the inductance value of the q-axis of the stator, the resistance value of the stator and the inductance value of the d-axis of the stator The proportional relationship, the proportional relationship between the sampling period and the inductance value of the q-axis of the stator, the proportional relationship between the sampling period and the inductance value of the d-axis of the stator; The first predicted current vector is determined based on the first predicted current vector expression. 3. The method according to claim 1, characterized in that the neural network in the neural network topology model includes multiple input layers, multiple hidden layers and multiple output layers, each of the input layers, each of the The hidden layer and each output layer are connected in a fully connected manner. 4. The method according to claim 3, characterized in that, after using a neural network topology model to process the first stator voltage, the second stator voltage and the current rotor angular velocity, a second prediction is obtained. In the process of current vectoring, the method further includes: Construct a d-axis prediction performance expression and a q-axis prediction performance expression. The d-axis prediction performance expression includes the t-th predicted d-axis current obtained by using the neural network topology model and the d-axis current using the neural network topology model. The difference between the d-axis current predicted at the t+1th time obtained, and the q-axis prediction performance expression includes the q-axis current predicted at the t-th time obtained using the neural network topology model and the q-axis current predicted using the neural network The difference in the q-axis current predicted at the t+1th time obtained by the topological model; A prediction performance vector is determined, which is a vector including d-axis prediction performance and q-axis prediction performance. 5. The method according to claim 4, characterized in that, after determining the prediction performance vector, the method further includes: according to , Determine the connection weights between the neurons of the input layer and the hidden layer and the connections between the neurons of the hidden layer and the output layer when the neural network topology model is used for the t+1th prediction. The connection weight; in, is the connection weight between the neurons of the input layer and the hidden layer when the neural network topology model is used for the t+1th prediction,/> is the connection weight between the neurons of the hidden layer and the output layer when the neural network topology model is used for the t+1th prediction,/> is the connection weight between the neurons of the input layer and the hidden layer when the neural network topology model is used for the tth prediction,/> is the connection weight between the neurons of the hidden layer and the output layer when the neural network topology model is used for the tth prediction,/> is the preset coefficient, loss is the prediction performance vector, i is used to characterize the i-th neuron of the input layer, j is used to characterize the j-th neuron of the hidden layer, and k is used to characterize the output The kth neuron of the layer. 6. The method according to claim 1, characterized in that, the optimized predicted current vector and the stator voltage vector are used to construct a vector expression, the minimum value of the vector expression is obtained, and the corresponding value of the minimum value is The voltage of the d-axis of the stator and the voltage of the q-axis of the stator are determined as the optimal predicted voltage vector, including: according to , Determine the optimal predicted voltage vector, where, is the optimal predicted voltage vector,/> The optimized prediction current vector for the predicted future Nth step,/> for/> The transposed matrix,/> To predict the difference between the optimized predicted current vector and the actual value of the corresponding current vector in step a of the future,/> for/> The transposed matrix,/> is the stator voltage vector,/> for/> The transpose matrix of current. 7. A device for predicting the voltage vector of the stator of a permanent magnet synchronous motor, characterized in that it includes: An acquisition unit is used to acquire the first stator voltage, the second stator voltage and the current rotor angular velocity. The first stator voltage is the voltage of the d-axis of the stator of the permanent magnet synchronous motor at the current moment. The second stator voltage is the voltage of the q-axis of the stator of the permanent magnet synchronous motor at the current moment, and the current rotor angular velocity is the angular velocity of the rotor of the permanent magnet synchronous motor at the current moment; The first processing unit is configured to use a priori estimation method to obtain the first predicted current vector based on the first stator voltage, the second stator voltage and the current rotor angular velocity, and use a neural network topology model to calculate The first stator voltage, the second stator voltage and the current rotor angular velocity are processed to obtain a second predicted current vector, wherein the first predicted current vector includes a first q-axis predicted current and a first The vector of the d-axis predicted current, the second predicted current vector is a vector including the second q-axis predicted current and the second d-axis predicted current, the neural network topology model is trained using multiple sets of training data, the Each set of training data in the multiple sets of training data includes input data obtained within a historical time period and output data corresponding to the input data. The input data includes the voltage of the d-axis of the stator, the stator The voltage of the q-axis and the angular velocity of the rotor, the output data include the predicted current of the q-axis of the stator and the predicted current of the d-axis of the stator; A second processing unit configured to weight the first predicted current vector and the second predicted current vector to obtain an optimized predicted current vector. The optimized predicted current vector includes optimized q-axis predicted current and optimized d-axis predict the vector of current; A third processing unit configured to construct a vector expression using the optimized predicted current vector and the stator voltage vector, obtain the minimum value of the vector expression, and obtain the voltage of the d-axis of the stator corresponding to the minimum value. and the voltage of the q-axis of the stator are determined as the optimal predicted voltage vector, and the stator voltage vector is a vector including the first stator voltage and the second stator voltage. 8. A control method for a permanent magnet synchronous motor, characterized by comprising: Obtain the first stator voltage, the second stator voltage and the current rotor angular speed. The first stator voltage is the voltage of the d-axis of the stator of the permanent magnet synchronous motor at the current moment. The second stator voltage is the permanent magnet The voltage of the q-axis of the stator of the synchronous motor at the current moment, and the current rotor angular velocity is the angular velocity of the rotor of the permanent magnet synchronous motor at the current moment; Using a priori estimation method, the first predicted current vector is obtained based on the first stator voltage, the second stator voltage and the current rotor angular velocity, and a neural network topology model is used to estimate the first stator voltage , the second stator voltage and the current rotor angular velocity are processed to obtain a second predicted current vector, wherein the first predicted current vector is a vector including the first q-axis predicted current and the first d-axis predicted current, The second predicted current vector is a vector including a second q-axis predicted current and a second d-axis predicted current. The neural network topology model is trained using multiple sets of training data. Each of the multiple sets of training data A set of training data includes input data obtained within a historical time period and output data corresponding to the input data. The input data includes the voltage of the d-axis of the stator, the voltage of the q-axis of the stator, and the output data corresponding to the input data. The angular velocity of the rotor, the output data includes the predicted current of the q-axis of the stator and the predicted current of the d-axis of the stator; Perform weighting processing on the first predicted current vector and the second predicted current vector to obtain an optimized predicted current vector, where the optimized predicted current vector is a vector including an optimized q-axis predicted current and an optimized d-axis predicted current; Use the optimized predicted current vector and the stator voltage vector to construct a vector expression, find the minimum value of the vector expression, and combine the voltage of the d-axis of the stator and the q-axis of the stator corresponding to the minimum value. The voltage of is determined as the optimal predicted voltage vector, and the stator voltage vector is a vector including the first stator voltage and the second stator voltage; The optimal predicted voltage vector is subjected to coordinate inverse transformation processing to obtain a target three-phase voltage, and the permanent magnet synchronous motor is controlled based on the target three-phase voltage. 9. A computer-readable storage medium, characterized in that the computer-readable storage medium includes a stored program, wherein when the program is run, the device where the computer-readable storage medium is located is controlled to execute claims 1 to 6 The method for predicting the voltage vector of the stator of a permanent magnet synchronous motor according to any one of the above. 10. An electronic device, characterized by comprising: one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to Executed by the one or more processors, the one or more programs include a prediction method for executing the voltage vector of the stator of the permanent magnet synchronous motor according to any one of claims 1 to 6.