Method and system for estimating stator and rotor resistances of an induction motor based on a hybrid model

Abstract

The present application belongs to the technical field of motor drive control, and particularly relates to an induction motor stator and rotor resistance estimation method and system based on a hybrid model. By real-time acquisition of stator voltage and stator current, a stator voltage equation is constructed, a corrected stator flux linkage is obtained, and a rotor flux linkage reference value is calculated, thereby providing a reference for subsequent model optimization. By constructing a hybrid model architecture of a rotor flux linkage adjustable model and a stator current adjustable model, and introducing a multi-modal attention mechanism to dynamically adjust model parameters, the adaptability and estimation accuracy of the model are improved, thereby enabling high-precision back-propagation estimation of rotor resistance and stator resistance, effectively improving the performance and reliability of the induction motor control system, and providing support for stable operation and accurate control of the induction motor.

Description

Technical Field

[0001] This invention belongs to the field of motor drive control technology, specifically relating to a method and system for estimating the stator and rotor resistance of an induction motor based on a hybrid model. Background Technology

[0002] In sensorless induction motor drive systems, the accuracy of motor parameters (especially rotor and stator resistances) is crucial for ensuring control performance. During actual operation, rotor and stator resistances can drift significantly due to dynamic factors such as internal motor temperature changes and magnetic circuit saturation. Any deviation in parameter estimation will directly increase the speed estimation error in the sensorless control loop, and in severe cases, may even lead to unstable operation of the entire system, posing a serious threat to its reliability and stability. Therefore, online parameter estimation technology has become a core research area and key technological direction for improving the robustness and control accuracy of sensorless induction motor drive systems.

[0003] At present, the existing technical means in the field of rotor and stator resistance estimation of induction motors mainly cover the following categories: (1) Analytical method based on motor model, which relies on the mathematical model of motor and combines the real-time measured voltage and current signals to calculate parameters. However, this method is easily affected by the simplification of the mathematical model of motor itself, the unavoidable noise interference during the measurement process, and the influence of different operating conditions, resulting in a large error in the estimation result of this method, which is difficult to meet the requirements of high-precision control; (2) Adaptive observer and Kalman filter method, which adjusts the parameter estimation value in real time through dynamic correction mechanism to improve the estimation accuracy. However, in practical applications, there is a problem of poor robustness. That is, when facing complex and changeable operating conditions, the estimation performance is easily affected. At the same time, the calculation process is relatively complex and requires high hardware computing resources. Moreover, this method is extremely sensitive to the initial parameter setting. Small deviations in the initial parameters may lead to large deviations in the estimation result; (3) Intelligent algorithms based on non-neural networks, such as fuzzy logic and genetic algorithms. These algorithms can achieve parameter estimation through preset empirical rules or optimization iteration processes. However, fuzzy logic algorithms rely heavily on expert experience to build rule bases and lack broad adaptability to different working conditions; genetic algorithms find the optimal solution through iterative optimization, but the iterative process takes a long time, resulting in poor real-time performance of the algorithm, making it difficult to meet the requirements of real-time control of induction motors, and its generalization ability is weak, with unstable performance under different motors or operating conditions. Summary of the Invention

[0004] This invention provides a method and system for estimating the stator and rotor resistance of an induction motor based on a hybrid model. By real-time acquisition of stator voltage and stator current, a stator voltage equation is constructed to obtain a corrected stator flux linkage and calculate a reference value for the rotor flux linkage, thus providing a reference for subsequent model optimization. By constructing a hybrid model architecture combining an adjustable rotor flux linkage model and an adjustable stator current model, and introducing a multimodal attention mechanism to dynamically adjust model parameters, the adaptability and estimation accuracy of the model are improved. This enables high-precision back-calculation of rotor and stator resistance, effectively improving the performance and reliability of the induction motor control system and providing support for the stable operation and precise control of the induction motor.

[0005] A method for estimating the stator and rotor resistance of an induction motor based on a hybrid model includes:

[0006] The stator voltage and stator current are acquired in real time and transformed into a stationary coordinate system to obtain voltage and current characteristics;

[0007] Based on voltage and current characteristics, a stator voltage equation is constructed. Combined with multi-stage low-pass filter correction, the corrected stator flux linkage is obtained. The stator flux linkage and stator current are substituted into the preset voltage model to calculate the rotor flux linkage reference value.

[0008] The current model of the induction motor is discretized to construct an adjustable rotor flux model, and a stator current observer is introduced to construct an adjustable stator current model. Based on the adjustable rotor flux model and the adjustable stator current model, the estimated values ​​of rotor flux and stator current are calculated respectively.

[0009] A multimodal attention mechanism is introduced, which combines the error values ​​between the rotor flux estimation value and the rotor flux reference value, as well as the error values ​​between the stator current estimation value and the stator current acquisition value, to update the network parameters of the multimodal attention mechanism and dynamically adjust the model parameters of the rotor flux adjustable model and the stator current adjustable model.

[0010] Based on the updated rotor flux adjustable model and stator current adjustable model model parameters, the rotor resistance and stator resistance are estimated by reverse calculation.

[0011] By acquiring stator voltage and stator current in real time, the stator voltage equation is constructed to obtain the corrected stator flux linkage and calculate the rotor flux linkage reference value, thus providing a reference for subsequent model optimization. By constructing a hybrid model architecture of an adjustable rotor flux linkage model and an adjustable stator current model, and introducing a multimodal attention mechanism to dynamically adjust the model parameters, the model's adaptability and estimation accuracy are improved. This enables high-precision back-calculation of rotor resistance and stator resistance, effectively improving the performance and reliability of the induction motor control system and providing support for the stable operation and precise control of the induction motor.

[0012] Furthermore, the real-time acquisition of stator voltage and stator current, and their transformation to a stationary coordinate system to obtain voltage and current characteristics, includes:

[0013] Based on voltage and current sensors, the stator voltage and stator current of the induction motor are acquired in real time; the stator voltage includes three-phase stator voltage; the stator current includes three-phase stator current.

[0014] Based on the stator voltage, the Clark transformation is used to convert it to a stationary coordinate system to obtain the voltage characteristics; the voltage characteristics include those in the stationary coordinate system. shaft Axis voltage components, and Under the axis Axis voltage components;

[0015] Based on the stator current, the Clark transformation is used to convert it to a stationary coordinate system to obtain the current characteristics; the current characteristics include those in the stationary coordinate system. shaft Axis current components, and Under the axis Axial current component.

[0016] By collecting the three-phase stator voltage and three-phase stator current of the induction motor and converting them to two-phase stator voltage and two-phase stator current in a stationary coordinate system, the system structure is simplified, the data complexity is reduced, and a data foundation is provided for subsequent model construction and analysis.

[0017] Furthermore, based on voltage and current characteristics, a stator voltage equation is constructed, and combined with multi-stage low-pass filter correction, a corrected stator flux linkage is obtained. The stator flux linkage and stator current are then substituted into a preset voltage model to calculate a rotor flux linkage reference value, including:

[0018] Based on voltage and current characteristics, the stator voltage equation in the stationary coordinate system is constructed. The stator flux linkage is initially estimated by back electromotive force integration, and the initially estimated stator flux linkage is corrected by a multi-stage low-pass filter to obtain the corrected stator flux linkage.

[0019] The corrected stator flux linkage and the collected stator current are substituted into the preset voltage model to construct the rotor flux linkage state equation based on the voltage model, and the rotor flux linkage is calculated and used as the rotor flux linkage reference value.

[0020] By using a multi-stage low-pass filter to correct the initially estimated stator flux linkage, DC drift and initial value errors during integration can be effectively suppressed, improving the accuracy of the stator flux linkage and providing a reliable basis for calculating the rotor flux linkage reference value. At the same time, by setting the rotor flux linkage reference value, a target reference is provided for subsequent network parameter optimization.

[0021] Furthermore, the discretization of the current model of the induction motor to construct an adjustable rotor flux linkage model and the introduction of a stator current observer to construct an adjustable stator current model, and the calculation of rotor flux linkage estimates and stator current estimates based on the adjustable rotor flux linkage model and the adjustable stator current model, respectively, include:

[0022] The current model of the induction motor in the rotating coordinate system is discretized, and an adjustable rotor flux model in the stationary coordinate system is constructed. The estimated value of the rotor flux is then calculated.

[0023] A stator current observer based on stator voltage and rotor flux linkage estimates is introduced to construct an adjustable stator current model in a stationary coordinate system and calculate the stator current estimate.

[0024] By constructing an adjustable rotor flux linkage model and an adjustable stator current model, the rotor flux linkage and stator current are estimated in parallel. The dual-adjustable model architecture provides the necessary state variables for the adaptive adjustment of subsequent model parameters, and at the same time provides a data foundation for subsequent network parameter optimization.

[0025] Furthermore, the introduction of a multimodal attention mechanism, combining the error values ​​between the rotor flux estimation and the rotor flux reference value, and the error values ​​between the stator current estimation and the stator current acquisition value, updates the network parameters of the multimodal attention mechanism and dynamically adjusts the model parameters of the adjustable rotor flux and the adjustable stator current models, including:

[0026] Based on the voltage and current features in the stationary coordinate system, the feature dimension and channel number are expanded respectively to obtain high-dimensional voltage features and high-dimensional current features. The high-dimensional voltage features and high-dimensional current features are then spliced ​​along the feature dimension to form a combined feature block.

[0027] Based on the combined feature blocks, a convolution operation is performed and an activation function is used to process the data to generate an attention mask. The attention mask and the combined block features are then multiplied element-wise to obtain weighted features.

[0028] Self-attention and channel attention mechanisms are introduced to dynamically assign weights to weighted features, generating high-dimensional spliced ​​features, which are then compressed to generate low-dimensional feature vectors. These low-dimensional feature vectors correspond to the model parameters of the rotor flux adjustable model and the stator current adjustable model. The model parameters of the rotor flux adjustable model are model parameters associated with the rotor resistance, and the model parameters of the stator current adjustable model are model parameters associated with the stator resistance.

[0029] Based on the error values ​​between the rotor flux estimate and the rotor flux reference value, and the error values ​​between the stator current estimate and the stator current acquisition value, the model parameters of the rotor flux adjustable model and the stator current adjustable model are dynamically updated.

[0030] Substitute the updated rotor flux adjustable model parameters into the rotor flux adjustable model and the updated stator current adjustable model parameters into the stator current adjustable model, and calculate the rotor flux estimated value and stator current estimated value respectively.

[0031] By introducing a multimodal attention mechanism, the operating characteristics of the induction motor are adaptively extracted, and the model parameters are adaptively adjusted according to the operating conditions. This allows the model to be continuously optimized based on the actual error situation, thereby improving the accuracy of the rotor flux estimation and stator current estimation.

[0032] Furthermore, the dynamic updating of the model parameters of the adjustable rotor flux linkage model and the adjustable stator current model based on the error values ​​between the rotor flux linkage estimate and the rotor flux linkage reference value, and the error values ​​between the stator current estimate and the stator current acquisition value, includes:

[0033] Based on the rotor flux estimate and the rotor flux reference value, a rotor flux error function is constructed.

[0034] Based on the stator current estimate and the stator current acquisition value, a stator current error function is constructed;

[0035] With the minimization of rotor flux error function and stator current error function as the optimization objective, the gradient backpropagation algorithm is used to adaptively update the model parameters of the rotor flux adjustable model and the stator current adjustable model.

[0036] By constructing a dual error function and using the minimization of the error function as the optimization objective to adaptively update the network parameters of the multimodal attention mechanism, the network weights can be adaptively adjusted according to the actual error, thereby improving system performance and accuracy.

[0037] Furthermore, the method of back-estimating rotor resistance and stator resistance based on the model parameters of the updated rotor flux adjustable model and stator current adjustable model includes:

[0038] Based on the updated rotor flux adjustable model parameters and their physical relationship with the rotor time constant, the rotor resistance is calculated in reverse.

[0039] Based on the updated stator current adjustable model parameters and the calculated rotor resistance, the stator resistance is estimated by reverse calculation.

[0040] A system for estimating the stator and rotor resistance of an induction motor based on a hybrid model includes:

[0041] The data acquisition and processing module is used to acquire stator voltage and stator current in real time and convert them to a stationary coordinate system to obtain voltage and current characteristics.

[0042] The model building and calculation module is used to construct the stator voltage equation based on voltage and current characteristics, combine it with multi-stage low-pass filter correction to obtain the corrected stator flux linkage, and substitute the stator flux linkage and stator current into the preset voltage model to calculate the rotor flux linkage reference value; it also discretizes the current model of the induction motor to construct an adjustable rotor flux linkage model, and introduces a stator current observer to construct an adjustable stator current model, and calculates the rotor flux linkage estimate and stator current estimate respectively based on the adjustable rotor flux linkage model and the adjustable stator current model;

[0043] The model training module is used to introduce a multimodal attention mechanism, which combines the error values ​​between the rotor flux estimation value and the rotor flux reference value, as well as the error values ​​between the stator current estimation value and the stator current acquisition value, to update the network parameters of the multimodal attention mechanism and dynamically adjust the model parameters of the rotor flux adjustable model and the stator current adjustable model.

[0044] The resistance calculation module is used to back-estimate the rotor resistance and stator resistance based on the model parameters of the updated rotor flux adjustable model and stator current adjustable model.

[0045] An electronic device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method described above.

[0046] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described above.

[0047] The beneficial effects of this invention are as follows:

[0048] This invention acquires stator voltage and stator current in real time, constructs stator voltage equations, obtains corrected stator flux linkage, and calculates rotor flux linkage reference values, thus providing a reference for subsequent model optimization. By constructing a hybrid model architecture of adjustable rotor flux linkage and adjustable stator current, and introducing a multimodal attention mechanism to dynamically adjust model parameters, the adaptability and estimation accuracy of the model are improved. This enables high-precision back-calculation of rotor and stator resistance, effectively improving the performance and reliability of the induction motor control system, and providing support for the stable operation and precise control of the induction motor. Attached Figure Description

[0049] Figure 1 This is a flowchart of the present invention;

[0050] Figure 2 This is a schematic diagram of the structure of the self-attention mechanism;

[0051] Figure 3 A flowchart of the self-attention mechanism;

[0052] Figure 4 This is a schematic diagram of the neural network model structure;

[0053] Figure 5 This is a schematic diagram of the network structure of the present invention;

[0054] Figure 6 This is a schematic diagram of the system structure of the present invention;

[0055] Figure 7 This is a schematic diagram of the structure of a computer device. Detailed Implementation

[0056] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0057] It should be noted that various aspects of embodiments within the scope of the appended claims are described below. It will be apparent that the aspects described herein can be embodied in a wide variety of forms, and any particular structure and / or function described herein is merely illustrative. Based on this disclosure, those skilled in the art will understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number of aspects set forth herein can be used to implement the device and / or practice the method. Additionally, this device and / or method can be implemented using other structures and / or functionalities besides one or more of the aspects set forth herein.

[0058] In addition, specific details are provided in the following description to facilitate a thorough understanding of the examples, and those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0059] Example 1

[0060] Figure 1This paper presents a hybrid model-based method for estimating the stator and rotor resistance of an induction motor. By real-time acquisition of stator voltage and current, a stator voltage equation is constructed to obtain the corrected stator flux linkage and calculate the rotor flux linkage reference value, thus providing a reference for subsequent model optimization. A hybrid model architecture combining an adjustable rotor flux linkage model and an adjustable stator current model is constructed, and a multimodal attention mechanism is introduced to dynamically adjust model parameters, improving model adaptability and estimation accuracy. This enables high-precision back-calculation of rotor and stator resistance, effectively enhancing the performance and reliability of the induction motor control system and supporting stable operation and precise control of the induction motor. The specific steps include:

[0061] S1: Real-time acquisition of stator voltage and stator current, and conversion to a stationary coordinate system to obtain voltage and current characteristics;

[0062] S11: Based on voltage and current sensors, the stator voltage and stator current of the induction motor are collected in real time;

[0063] In this embodiment, the stator voltage includes the three-phase stator voltage; the stator current includes the three-phase stator current;

[0064] S12: Based on the stator voltage, the Clark transformation is used to convert it to a stationary coordinate system to obtain the voltage characteristics;

[0065] In this embodiment, voltage characteristics Including in the stationary coordinate system shaft Axis voltage components ,as well as Under the axis Axis voltage components ;

[0066] S13: Based on the stator current, the Clark transformation is used to convert it to the stationary coordinate system to obtain the current characteristics;

[0067] In this embodiment, current characteristics Including in the stationary coordinate system shaft Axis current components ,as well as Under the axis Axis current components .

[0068] S2: Based on voltage and current characteristics, construct the stator voltage equation, combine it with multi-stage low-pass filter correction to obtain the corrected stator flux linkage, and substitute the stator flux linkage and stator current into the preset voltage model to calculate the rotor flux linkage reference value.

[0069] S21: Based on voltage and current characteristics, the stator voltage equation in the stationary coordinate system is constructed. The stator flux linkage is initially estimated by back electromotive force integration. The initially estimated stator flux linkage is then corrected by a multi-stage low-pass filter to obtain the corrected stator flux linkage.

[0070] The expression for the stator voltage equation in the stationary coordinate system is as follows:

[0071] ;

[0072] In the formula, Indicates stator flux linkage The components of the axis; Indicates stator flux linkage The components of the axis; Indicates the stator voltage at The components of the axis; Indicates the stator voltage at The components of the axis; Indicates the stator current at The components of the axis; Indicates the stator current at The components of the axis; Indicates stator resistance; Indicates time;

[0073] Because the DC voltage applied to the inverter fluctuates, a multi-stage low-pass filter is used to correct the initially estimated stator flux linkage, which can effectively suppress DC drift and initial value error during integration and improve the accuracy of the stator flux linkage.

[0074] S22: Substitute the corrected stator flux linkage and the collected stator current into the preset voltage model, construct the rotor flux linkage state equation based on the voltage model, calculate the rotor flux linkage, and use it as the rotor flux linkage reference value.

[0075] The expression for the preset voltage model is as follows:

[0076] ;

[0077] In the formula, This indicates that the rotor flux linkage based on a preset voltage model is... The components of the axis; This indicates that the rotor flux linkage is based on a preset voltage equation. The components of the axis; Indicates rotor self-inductance; Indicates mutual intuition; Indicates the self-inductance of the stator; The leakage flux coefficient is expressed as follows: ;

[0078] The expression for the rotor flux linkage state equation based on the voltage model is as follows:

[0079] ;

[0080] In the formula, express Rotor flux based on voltage model at all times The components of the axis; express Stator flux linkage at time The components of the axis; express Stator current at time The components of the axis; express Rotor flux based on voltage model at all times The components of the axis; Indicates stator flux linkage The components of the axis; express Stator current at time The components of the axis;

[0081] S3: Discretize the current model of the induction motor to construct an adjustable rotor flux linkage model, and introduce a stator current observer to construct an adjustable stator current model. Based on the adjustable rotor flux linkage model and the adjustable stator current model, calculate the estimated values ​​of rotor flux linkage and stator current respectively.

[0082] S31: Discretize the current model of the induction motor in the rotating coordinate system, construct an adjustable rotor flux linkage model in the stationary coordinate system, and calculate the estimated value of the rotor flux linkage.

[0083] The expression for the adjustable rotor flux linkage model in the stationary coordinate system is as follows:

[0084] ;

[0085] In the formula, express Rotor flux based on current model at all times The components of the axis; express Rotor flux based on current model at all times The components of the axis; express Rotor flux based on current model at all times The components of the axis; express Rotor flux based on current model at all times The components of the axis; , , The model parameters of the rotor flux adjustable model are expressed as follows: , , ; Indicates the sampling time; The rotor time constant is expressed as follows: ;

[0086] According to the expression of the model parameters of the rotor flux adjustable model, the model parameters are... Independent of rotor time constant, model parameters and model parameters It is related to the rotor time constant, and from the expression for the rotor time constant, it can be seen that the rotor time constant is inversely related to the rotor resistance, that is, when the rotor inductance... When kept constant, the model parameters can be changed by altering the rotor resistance. and model parameters This makes it easier to obtain higher accuracy.

[0087] S32: Introduce a stator current observer based on the stator voltage and rotor flux estimates, construct an adjustable stator current model in a stationary coordinate system, and calculate the stator current estimate.

[0088] As shown in the preset voltage model, stator resistance significantly affects the calculation of rotor flux linkage, especially in the low-speed region where the back electromotive force is extremely small. This can easily lead to the rotor flux linkage calculated using the adjustable rotor flux linkage model deviating from the true value. To reduce the error in rotor flux linkage estimation caused by stator resistance, a stator current observer based on the stator voltage and rotor flux linkage estimates is introduced to construct an adjustable stator current model in a stationary coordinate system. Its expression is:

[0089] ;

[0090] In the formula, express Stator current at time The components of the axis; express Stator current at time The components of the axis; express Stator current at time The components of the axis; express Stator current at time The components of the axis; express Stator voltage at time The components of the axis; express Stator voltage at time The components of the axis; , , , The model parameters are represented by the following expressions: , , , , Indicates the sampling period. Indicates the speed of the induction motor;

[0091] According to the expression for the model parameters of the adjustable stator current model, the model parameters are... Model parameters Model parameters Related to induction motor parameters and induction motor speed Sampling period Related to, and related to, stator resistance Irrelevant, but model parameters With stator resistance This relates to the ability to change model parameters by altering the stator resistance before updating network parameters. This makes it easier to obtain higher accuracy.

[0092] In this embodiment, the stator current observer is a Luneburg observer.

[0093] S4: Introduce a multimodal attention mechanism, combine the error values ​​between the rotor flux estimation value and the rotor flux reference value, and the error values ​​between the stator current estimation value and the stator current acquisition value, update the network parameters of the multimodal attention mechanism, and dynamically adjust the model parameters of the rotor flux adjustable model and the stator current adjustable model.

[0094] S41: Based on the voltage and current features in the stationary coordinate system, the feature dimension and channel number are expanded respectively to obtain high-dimensional voltage features and high-dimensional current features. The high-dimensional voltage features and high-dimensional current features are then spliced ​​along the feature dimension to form a combined feature block.

[0095] S411: Voltage characteristics based on stationary coordinates Multiple one-dimensional convolutional layers, one-dimensional inverted convolutional layers, and activation functions are used to expand the feature dimension and number of channels to obtain high-dimensional voltage features.

[0096] In this embodiment, voltage characteristics The number of samples is The initial channel size is 1, and the initial feature size is 2, i.e. .

[0097] The expression for the voltage feature after feature dimension expansion is as follows:

[0098] ;

[0099] In the formula, This represents the voltage feature after feature dimension expansion, which expands the feature size from 2 to 16; This represents a one-dimensional inverse convolution function; This represents the sigmoid activation function;

[0100] The expression for the voltage characteristic after channel expansion is as follows:

[0101] ;

[0102] In the formula, This represents the voltage characteristic after channel number expansion, which expands the channel number from 1 to... ; Represents a one-dimensional convolution function;

[0103] The voltage features after feature dimension expansion are fused with the voltage features after channel number expansion to obtain the expression for the high-dimensional voltage features:

[0104] ;

[0105] In the formula, This represents high-dimensional voltage characteristics, and ; Indicates feature concatenation operation;

[0106] S412: Current characteristics based on stationary coordinate system Multiple one-dimensional convolutional layers, one-dimensional inverted convolutional layers, and activation functions are used to expand the feature dimension and number of channels to obtain high-dimensional current features;

[0107] In this embodiment, voltage characteristics The number of samples is The initial channel size is 1, and the initial feature size is 2, i.e. .

[0108] The expression for the current characteristic after feature dimension expansion is as follows:

[0109] ;

[0110] In the formula, This represents the current feature after the feature dimension is expanded, which increases the feature size from 2 to 16;

[0111] The expression for the current characteristic after channel expansion is as follows:

[0112] ;

[0113] In the formula, This represents the current characteristic after the channel number is expanded, which increases the channel number from 1 to... ;

[0114] The current features after feature dimension expansion are fused with the current features after channel number expansion to obtain the expression for the high-dimensional current features:

[0115] ;

[0116] In the formula, Represents high-dimensional current characteristics, and ;

[0117] It should be noted that the number of channels, as a hyperparameter, is generally expanded before model training.

[0118] S413: Combine high-dimensional voltage features and high-dimensional current features along the feature dimensions to form a combined feature block;

[0119] The expression for the combined feature block is:

[0120] ;

[0121] In the formula, Represents a combined feature block, and ;

[0122] S42: Based on the combined feature blocks, perform convolution operations and use activation functions to generate an attention mask. Then, multiply the attention mask and the combined block features element-wise to obtain weighted features.

[0123] The expression for the attention mask is:

[0124] ;

[0125] In the formula, This indicates an attention mask, and ;

[0126] The expression for the weighted feature is:

[0127] ;

[0128] In the formula, Indicates weighted features, and ;

[0129] S43: Introduce self-attention and channel attention mechanisms to dynamically assign weights to weighted features, generate high-dimensional concatenated features, and compress the feature dimensions to generate low-dimensional feature vectors.

[0130] By introducing self-attention and channel attention mechanisms, average pooling and max pooling operations are performed on the weighted features to obtain the corresponding global feature vectors. These vectors are then input into a shared fully connected layer, and combined with an activation function to generate high-dimensional concatenated features, the expression of which is:

[0131] ;

[0132] In the formula, Represents high-dimensional splicing features; Indicates a fully connected operation; This indicates an average pooling operation; This represents the max pooling operation;

[0133] The high-dimensional concatenated features are compressed to generate a low-dimensional feature vector, expressed as follows:

[0134] ;

[0135] In the formula, It is a low-dimensional feature vector, and ;

[0136] In this embodiment, the low-dimensional feature vectors correspond to the model parameters of the rotor flux adjustable model and the stator current adjustable model; that is, the model parameters of the rotor flux adjustable model are model parameters related to the rotor resistance. and model parameters The model parameters of the adjustable stator current model are model parameters related to the stator resistance. .

[0137] Figure 2 The diagram shown is a structural schematic of the self-attention mechanism; Figure 3 The diagram shown is the algorithm flowchart for the self-attention mechanism.

[0138] S45: Based on the error values ​​between the rotor flux estimate and the rotor flux reference value, and the error values ​​between the stator current estimate and the stator current acquisition value, dynamically update the model parameters of the rotor flux adjustable model and the stator current adjustable model.

[0139] S451: Construct a rotor flux error function based on the rotor flux estimate and the rotor flux reference value;

[0140] The square function of the difference between the estimated rotor flux linkage and the reference rotor flux linkage is used as the rotor flux linkage error function, and its expression is:

[0141] ;

[0142] In the formula, This represents the rotor flux linkage error function value; express The square of the difference between the estimated rotor flux linkage and the reference rotor flux linkage at any given time.

[0143] S452: Construct a stator current error function based on the stator current estimate and the stator current acquisition value;

[0144] The square function of the difference between the estimated stator current and the acquired stator current is used as the stator current error function, and its expression is:

[0145] ;

[0146] In the formula, This represents the value of the stator current error function; express The square of the difference between the estimated stator current and the acquired stator current at any given time.

[0147] S453: With minimizing the rotor flux error function and stator current error function as the optimization objective, the gradient backpropagation algorithm is used to adaptively update the model parameters of the rotor flux adjustable model and the stator current adjustable model.

[0148] Model parameters based on the adjustable rotor flux model By minimizing the rotor flux error function, the gradient backpropagation algorithm is used to update it, resulting in the updated model parameters of the rotor flux adjustable model. Its expression is:

[0149] ;

[0150] ;

[0151] In the formula, This represents the model parameters of the updated rotor flux adjustable model. ; This represents the model parameters of the unupdated rotor flux adjustable model. ; Indicates the learning rate;

[0152] Similarly, the updated model parameters of the rotor flux adjustable model can be obtained. Model parameters of the adjustable stator current model .

[0153] It should be noted that in practical applications, it is necessary to obtain operating data of voltage and current characteristics of the same type of induction motor under different stator and rotor resistances. Multiple sets of operating data are then input into a neural network model based on a multimodal attention mechanism for training. This allows the neural network to load the trained weights in subsequent applications, eliminating the need for further updates, ensuring real-time computation, improving computational efficiency, and ultimately quickly outputting the updated model parameters of the adjustable rotor flux linkage model. Model parameters of the rotor flux adjustable model Model parameters of the adjustable stator current model .

[0154] Figure 4 The diagram shows the structure of the neural network model.

[0155] S46: Substitute the updated rotor flux adjustable model parameters into the rotor flux adjustable model and the updated stator current adjustable model parameters into the stator current adjustable model, and calculate the rotor flux estimated value and stator current estimated value respectively.

[0156] S5: Based on the model parameters of the updated rotor flux adjustable model and stator current adjustable model, the rotor resistance and stator resistance are estimated by reverse calculation.

[0157] S51: Based on the model parameters of the updated rotor flux adjustable model, and combined with their physical relationship with the rotor time constant, the rotor resistance is calculated in reverse.

[0158] In this embodiment, the model parameters of the rotor flux adjustable model include model parameters. and model parameters There are two options; in actual calculations, either one can be chosen to calculate the rotor resistance.

[0159] Among them, based on model parameters The expression for calculating the rotor resistance is:

[0160] ;

[0161] In the formula, Indicates based on model parameters Calculated rotor resistance;

[0162] Among them, based on model parameters The expression for calculating the rotor resistance is:

[0163] ;

[0164] In the formula, Indicates based on model parameters Calculated rotor resistance;

[0165] In this embodiment, the model parameters will be used as the basis. Calculated rotor resistance and based on model parameters Calculated rotor resistance Defined as rotor resistance .

[0166] S52: Based on the model parameters of the updated adjustable stator current model and the calculated rotor resistance, the stator resistance is estimated by reverse calculation.

[0167] Among them, the model parameters based on the adjustable stator current model The calculated expression for the stator resistance is:

[0168] ;

[0169] In the formula, This indicates the stator resistance.

[0170] Figure 5 The diagram shown is a schematic of the network structure. IM represents the induction motor; the indirect magnetic field control is a combined representation of the speed loop and current loop, forming the overall control architecture; the Luneburg observer acts as a speed observer, used to estimate the induction motor speed in real time.

[0171] In this control system structure, the indirect magnetic field control receives the reference rotation speed. And the magnetic flux obtained from the calibration of the induction motor Information, the current component of the output stator current of the induction motor in the stationary coordinate system. and and the voltage components of the stator voltage in the stationary coordinate system. and The Luneburger observer estimates the induction motor speed based on the acquired current and voltage components. This information is then fed back to the indirect magnetic field control to form a closed-loop control. After the system is running, the predicted value of the current component at the next moment can be obtained. Predicted values ​​of current components Voltage component prediction value Voltage component prediction value This is then used as the system input for the next iteration. An error function is constructed based on the parameter difference between the current and previous time steps, and the gradient backpropagation algorithm is used to update the weights of the neural network model, thereby updating the model parameters of the rotor flux adjustable model. Model parameters And the model parameters of the stator current adjustable model. Meanwhile, the rotor resistance will be calculated and obtained. Feedback to indirect magnetic field control and stator resistance Feedback is sent to the Luneburg observer to enable dynamic adjustment and optimization of system parameters, thereby improving system performance.

[0172] Example 2

[0173] like Figure 6 As shown in the figure, this embodiment provides a system for estimating the stator and rotor resistance of an induction motor based on a hybrid model, including a data acquisition and processing module, a model building and calculation module, a model training module, and a resistance calculation module.

[0174] Specifically, the data acquisition and processing module is used to acquire stator voltage and stator current in real time and convert them to a stationary coordinate system to obtain voltage and current characteristics;

[0175] Specifically, the model building and calculation module is used to construct the stator voltage equation based on voltage and current characteristics, combine it with multi-stage low-pass filter correction to obtain the corrected stator flux linkage, and substitute the stator flux linkage and stator current into the preset voltage model to calculate the rotor flux linkage reference value; and to discretize the current model of the induction motor to construct an adjustable rotor flux linkage model, and to introduce a stator current observer to construct an adjustable stator current model, and to calculate the rotor flux linkage estimate and stator current estimate respectively based on the adjustable rotor flux linkage model and the adjustable stator current model;

[0176] Specifically, the model training module is used to introduce a multimodal attention mechanism, which combines the error values ​​between the rotor flux estimate and the rotor flux reference value, as well as the error values ​​between the stator current estimate and the stator current acquisition value, to update the network parameters of the multimodal attention mechanism and dynamically adjust the model parameters of the rotor flux adjustable model and the stator current adjustable model.

[0177] Specifically, the resistance calculation module is used to back-estimate the rotor resistance and stator resistance based on the model parameters of the updated rotor flux adjustable model and stator current adjustable model.

[0178] Example 3

[0179] Based on the same technical concept, embodiments of this application also provide a computer device, including a memory 1 and a processor 2, such as... Figure 7 As shown, memory 1 stores a computer program, and processor 2 executes the computer program to implement any of the methods described above.

[0180] The memory 1 includes at least one type of readable storage medium, including flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 1 can be an internal storage unit of the hybrid model-based induction motor stator and rotor resistance estimation system, such as a hard disk. In other embodiments, the memory 1 can also be an external storage device of the hybrid model-based induction motor stator and rotor resistance estimation system, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. Furthermore, the memory 1 can include both an internal storage unit and an external storage device of the hybrid model-based induction motor stator and rotor resistance estimation system. The memory 1 can be used not only to store application software and various types of data installed in the hybrid model-based induction motor stator and rotor resistance estimation system, such as the code of the hybrid model-based induction motor stator and rotor resistance estimation system program, but also to temporarily store data that has been output or will be output.

[0181] In some embodiments, processor 2 may be a central processing unit (CPU), controller, microcontroller, microprocessor or other data processing chip, used to run program code stored in memory 1 or process data, such as executing a program for estimating the stator and rotor resistance of an induction motor based on a hybrid model.

[0182] The present invention also discloses a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of the method described in the above-described method embodiments. The storage medium may be a volatile or non-volatile computer-readable storage medium.

[0183] The computer program product of the application page content refresh method provided in the embodiments of the present invention includes a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the steps of the method in the above method embodiments. For details, please refer to the above method embodiments, which will not be repeated here.

[0184] The present invention also discloses a computer program that, when executed by a processor, implements any of the methods described in the foregoing embodiments. This computer program product can be implemented specifically through hardware, software, or a combination thereof. In one optional embodiment, the computer program product is specifically embodied in a computer storage medium; in another optional embodiment, the computer program product is specifically embodied in a software product, such as a software development kit (SDK), etc.

[0185] It is understood that the same or similar parts in the above embodiments can be referred to each other, and the contents not described in detail in some embodiments can be referred to the same or similar contents in other embodiments.

[0186] It should be noted that in the description of this invention, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance. Furthermore, in the description of this invention, unless otherwise stated, "a plurality of" means at least two.

[0187] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process, and the scope of the preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of the invention pertain.

[0188] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0189] Those skilled in the art will understand that all or part of the steps of the methods described in the above embodiments can be implemented by a program instructing related hardware, and the program can be stored in a computer-readable storage medium. When executed, the program includes one or a combination of the steps of the method embodiments.

[0190] Furthermore, the functional units in the various embodiments of the present invention can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0191] The storage media mentioned above can be read-only memory, disk, or optical disk, etc.

[0192] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0193] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for estimating the stator and rotor resistance of an induction motor based on a hybrid model, characterized in that, include: The stator voltage and stator current are acquired in real time and transformed into a stationary coordinate system to obtain voltage and current characteristics; Based on voltage and current characteristics, a stator voltage equation is constructed. Combined with multi-stage low-pass filter correction, the corrected stator flux linkage is obtained. The stator flux linkage and stator current are substituted into the preset voltage model to calculate the rotor flux linkage reference value. The current model of the induction motor is discretized to construct an adjustable rotor flux model, and a stator current observer is introduced to construct an adjustable stator current model. Based on the adjustable rotor flux model and the adjustable stator current model, the estimated values ​​of rotor flux and stator current are calculated respectively. A multimodal attention mechanism is introduced, which combines the error values ​​between the rotor flux estimation value and the rotor flux reference value, as well as the error values ​​between the stator current estimation value and the stator current acquisition value, to update the network parameters of the multimodal attention mechanism and dynamically adjust the model parameters of the rotor flux adjustable model and the stator current adjustable model. Based on the updated rotor flux adjustable model and stator current adjustable model model parameters, the rotor resistance and stator resistance are estimated by reverse calculation.

2. The method for estimating the stator and rotor resistance of an induction motor based on a hybrid model according to claim 1, characterized in that, The real-time acquisition of stator voltage and stator current, and their transformation to a stationary coordinate system, yields voltage and current characteristics, including: Based on voltage and current sensors, the stator voltage and stator current of the induction motor are acquired in real time; the stator voltage includes three-phase stator voltage; the stator current includes three-phase stator current. Based on the stator voltage, the Clark transformation is used to convert it to a stationary coordinate system to obtain the voltage characteristics; the voltage characteristics include those in the stationary coordinate system. shaft Axis voltage components, and Under the axis Axis voltage components; Based on the stator current, the Clark transformation is used to convert it to a stationary coordinate system to obtain the current characteristics; the current characteristics include those in the stationary coordinate system. shaft Axis current components, and Under the axis Axial current component.

3. The method for estimating the stator and rotor resistance of an induction motor based on a hybrid model according to claim 1, characterized in that, Based on voltage and current characteristics, a stator voltage equation is constructed. Combined with multi-stage low-pass filter correction, the corrected stator flux linkage is obtained. The stator flux linkage and stator current are then substituted into a preset voltage model to calculate the rotor flux linkage reference value, including: Based on voltage and current characteristics, the stator voltage equation in the stationary coordinate system is constructed. The stator flux linkage is initially estimated by back electromotive force integration, and the initially estimated stator flux linkage is corrected by a multi-stage low-pass filter to obtain the corrected stator flux linkage. The corrected stator flux linkage and the collected stator current are substituted into the preset voltage model to construct the rotor flux linkage state equation based on the voltage model, and the rotor flux linkage is calculated and used as the rotor flux linkage reference value.

4. The method for estimating the stator and rotor resistance of an induction motor based on a hybrid model according to claim 1, characterized in that, The process involves discretizing the current model of the induction motor to construct an adjustable rotor flux linkage model, and introducing a stator current observer to construct an adjustable stator current model. Based on the adjustable rotor flux linkage model and the adjustable stator current model, the estimated rotor flux linkage and estimated stator current are calculated, respectively. The current model of the induction motor in the rotating coordinate system is discretized, and an adjustable rotor flux model in the stationary coordinate system is constructed. The estimated value of the rotor flux is then calculated. A stator current observer based on stator voltage and rotor flux linkage estimates is introduced to construct an adjustable stator current model in a stationary coordinate system and calculate the stator current estimate.

5. The method for estimating the stator and rotor resistance of an induction motor based on a hybrid model according to claim 1, characterized in that, The introduction of a multimodal attention mechanism, combining the error values ​​between the rotor flux estimation and the rotor flux reference value, and the error values ​​between the stator current estimation and the stator current acquisition value, updates the network parameters of the multimodal attention mechanism and dynamically adjusts the model parameters of the adjustable rotor flux and the adjustable stator current models, including: Based on the voltage and current features in the stationary coordinate system, the feature dimension and channel number are expanded respectively to obtain high-dimensional voltage features and high-dimensional current features. The high-dimensional voltage features and high-dimensional current features are then spliced ​​along the feature dimension to form a combined feature block. Based on the combined feature blocks, a convolution operation is performed and an activation function is used to process the data to generate an attention mask. The attention mask and the combined block features are then multiplied element-wise to obtain weighted features. Self-attention and channel attention mechanisms are introduced to dynamically assign weights to weighted features, generating high-dimensional spliced ​​features, which are then compressed to generate low-dimensional feature vectors. These low-dimensional feature vectors correspond to the model parameters of the rotor flux adjustable model and the stator current adjustable model. The model parameters of the rotor flux adjustable model are model parameters associated with the rotor resistance, and the model parameters of the stator current adjustable model are model parameters associated with the stator resistance. Based on the error values ​​between the rotor flux estimate and the rotor flux reference value, and the error values ​​between the stator current estimate and the stator current acquisition value, the model parameters of the rotor flux adjustable model and the stator current adjustable model are dynamically updated. Substitute the updated rotor flux adjustable model parameters into the rotor flux adjustable model and the updated stator current adjustable model parameters into the stator current adjustable model, and calculate the rotor flux estimated value and stator current estimated value respectively.

6. The method for estimating the stator and rotor resistance of an induction motor based on a hybrid model according to claim 5, characterized in that, The model parameters of the adjustable rotor flux linkage model and the adjustable stator current model are dynamically updated based on the error values ​​between the rotor flux linkage estimate and the rotor flux linkage reference value, and the error values ​​between the stator current estimate and the stator current acquisition value. This includes: Based on the rotor flux estimate and the rotor flux reference value, a rotor flux error function is constructed. Based on the stator current estimate and the stator current acquisition value, a stator current error function is constructed; With the minimization of rotor flux error function and stator current error function as the optimization objective, the gradient backpropagation algorithm is used to adaptively update the model parameters of the rotor flux adjustable model and the stator current adjustable model.

7. The method for estimating the stator and rotor resistance of an induction motor based on a hybrid model according to claim 1, characterized in that, The method of back-estimating rotor resistance and stator resistance based on the model parameters of the updated rotor flux adjustable model and stator current adjustable model includes: Based on the updated rotor flux adjustable model parameters and their physical relationship with the rotor time constant, the rotor resistance is calculated in reverse. Based on the updated stator current adjustable model parameters and the calculated rotor resistance, the stator resistance is estimated by reverse calculation.

8. A system for implementing the induction motor stator and rotor resistance estimation method based on a hybrid model as described in claim 1, characterized in that, include: The data acquisition and processing module is used to acquire stator voltage and stator current in real time and convert them to a stationary coordinate system to obtain voltage and current characteristics. The model building and calculation module is used to construct the stator voltage equation based on voltage and current characteristics, combine it with multi-stage low-pass filter correction to obtain the corrected stator flux linkage, and substitute the stator flux linkage and stator current into the preset voltage model to calculate the rotor flux linkage reference value; it also discretizes the current model of the induction motor to construct an adjustable rotor flux linkage model, and introduces a stator current observer to construct an adjustable stator current model, and calculates the rotor flux linkage estimate and stator current estimate respectively based on the adjustable rotor flux linkage model and the adjustable stator current model; The model training module is used to introduce a multimodal attention mechanism, which combines the error values ​​between the rotor flux estimation value and the rotor flux reference value, as well as the error values ​​between the stator current estimation value and the stator current acquisition value, to update the network parameters of the multimodal attention mechanism and dynamically adjust the model parameters of the rotor flux adjustable model and the stator current adjustable model. The resistance calculation module is used to back-estimate the rotor resistance and stator resistance based on the model parameters of the updated rotor flux adjustable model and stator current adjustable model.

9. An electronic device, characterized in that, The electronic device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, 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