Method, system, device and medium for identification of wireless power transfer coil parameters

Abstract

The present application relates to a kind of wireless power transmission coil parameter identification method, system, device and medium.The method includes: the operating frequency of the wireless power transmission coil to be measured is obtained, and the image of the wireless power transmission coil under the operating frequency;The image of the wireless power transmission coil and the operating frequency are input into parameter identification model, the structural features in the image are extracted, and the operating frequency feature is extracted, and the quality factor and inductance value of the wireless power transmission coil are obtained based on the structural features and the operating frequency feature.The present application improves the accuracy and speed of coil parameter identification.

Description

Technical Field

[0001] This invention relates to the field of wireless power transmission technology, and in particular to a method, system, device, and medium for identifying parameters of a wireless power transmission coil. Background Technology

[0002] In recent years, wireless power transfer has gradually become an important charging technology due to its ability to overcome the limitations of physical connections. In wireless power transfer systems, the transmitting and receiving coils are crucial components, and they are often wound with Litz (Litzendraht, stranded wire) coils. The inductance and quality factor of the Litz coil have a significant impact on the operation of the system; however, accurately identifying the coil parameters is a challenging task.

[0003] Currently, electromagnetic simulation software or impedance analyzers are commonly used to identify coil parameters, but both methods have certain drawbacks: Electromagnetic simulation software (such as Ansys Q3D) suffers from long simulation times for Litz coils. Furthermore, at high frequencies, the skin effect and approach effect often lead to non-convergence or poor accuracy in simulation results. In addition, while impedance analyzer methods offer acceptable accuracy, the equipment is expensive and the operation is cumbersome, making rapid and accurate coil parameter identification impossible. Therefore, there is a need for a method, system, device, and medium for identifying the parameters of wireless power transmission coils. Summary of the Invention

[0004] In view of the shortcomings of the prior art described above, the purpose of this invention is to provide a method, system, device and medium for identifying wireless power transmission coil parameters, thereby improving the problem that the speed and accuracy of coil parameter identification in the prior art cannot be achieved simultaneously.

[0005] To achieve the above and other related objectives, the present invention provides a method for identifying parameters of a wireless power transmission coil, comprising: acquiring the operating frequency of the wireless power transmission coil under test, and an image of the wireless power transmission coil at the operating frequency; inputting the image of the wireless power transmission coil and the operating frequency into a parameter identification model, extracting structural features from the image and the operating frequency features, and obtaining the quality factor and inductance value of the wireless power transmission coil based on the structural features and the operating frequency features.

[0006] In one embodiment of the present invention, the step of inputting an image of the wireless power transmission coil and the operating frequency into a parameter recognition model, extracting structural features from the image and the operating frequency features, and obtaining the quality factor and inductance value of the wireless power transmission coil based on the structural features and the operating frequency features includes: inputting the image into an image encoder module of the parameter recognition model to extract the structural features of the wireless power transmission coil from the image to obtain image embedding features; wherein the image encoder module is a convolutional neural network; inputting the operating frequency into a text encoder module of the parameter recognition model to extract the operating frequency features to obtain text embedding features; wherein the text encoder module is a convolutional neural network; inputting the image embedding features and the text embedding features into a fusion module of the parameter recognition model to fuse the features to obtain multimodal hybrid features; and inputting the multimodal hybrid features into a decoder module of the parameter recognition model to decode the multimodal hybrid features to obtain the quality factor and inductance value of the wireless power transmission coil.

[0007] In one embodiment of the present invention, the step of inputting the image into the image encoder module of the parameter recognition model and extracting the structural features of the wireless power transmission coil from the image to obtain image embedding features includes: performing size normalization processing on the image; inputting the normalized image into the image encoder module, extracting multiple structural features of the wireless power transmission coil from the image, and performing feature compression on the multiple structural features to generate image embedding features.

[0008] In one embodiment of the present invention, the step of inputting the normalized image into the image encoder module, extracting multiple structural features of the wireless power transmission coil from the image, and performing feature compression on the multiple structural features to generate image embedding features includes: inputting the normalized image into a feature extraction unit group of the image encoder module, performing feature extraction step by step through each feature extraction unit, and generating multiple initial structural features of the wireless power transmission coil after each extraction, and using each initial structural feature generated by the last feature extraction unit as the corresponding structural feature; wherein, the feature extraction unit group includes multiple cascaded feature extraction units; and inputting the multiple structural features into a fully connected unit of the image encoder module for dimensionality reduction and merging processing to generate image embedding features.

[0009] In one embodiment of the present invention, the step of inputting the running frequency into the text encoder module of the parameter recognition model and extracting running frequency features to obtain text embedding features includes: extending the running frequency to obtain a one-dimensional tensor; inputting the one-dimensional tensor into the text encoder module, extracting multiple running frequency features of the one-dimensional tensor, and compressing the multiple running frequency features to obtain text embedding features.

[0010] In one embodiment of the present invention, the step of inputting the image embedding features and the text embedding features into the fusion module of the parameter recognition model and fusing the two features to obtain multimodal hybrid features includes: weighting and summing the image embedding features and the text embedding features to obtain multimodal hybrid features.

[0011] In one embodiment of the present invention, the parameter recognition model is trained. The training process of the parameter recognition model includes: inputting a pre-acquired image of a wireless power transmission coil and its corresponding operating frequency into the parameter recognition model to obtain a predicted quality factor value and a predicted inductance value; calculating the mean square error of the quality factor between the pre-acquired true quality factor value and the predicted quality factor value; calculating the mean square error of the inductance value between the pre-acquired true inductance value and the predicted inductance value; weighting the mean square error of the quality factor value and the mean square error of the inductance value to obtain a comprehensive error; and updating the parameters of the parameter recognition model based on the comprehensive error using the gradient descent method to obtain a trained parameter recognition model.

[0012] In one embodiment of the present invention, a wireless power transmission coil parameter identification system is also provided. The system includes: a data acquisition module for acquiring the operating frequency of the wireless power transmission coil under test and an image of the wireless power transmission coil at the operating frequency; and a parameter prediction module for inputting the image of the wireless power transmission coil and the operating frequency into a parameter identification model, extracting structural features from the image and the operating frequency features, and obtaining the quality factor and inductance value of the wireless power transmission coil based on the structural features and the operating frequency features.

[0013] In one embodiment of the present invention, an electronic device is also provided, comprising: one or more processors; and a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the electronic device enables the wireless power transmission coil parameter identification method described in any of the preceding claims.

[0014] In one embodiment of the present invention, a computer-readable storage medium is also provided, on which a computer program is stored, which, when executed by a computer processor, causes the computer to perform the wireless power transmission coil parameter identification method described in any of the preceding claims.

[0015] As described above, the method, system, device, and medium for identifying wireless power transmission coil parameters of the present invention have the following beneficial effects: By employing a multi-modal model approach, simultaneously considering both the coil's image and operating frequency, the structural features in the image and the operating frequency features in the input operating frequency are extracted after inputting the image and operating frequency into the parameter identification model. Based on these two features, the quality factor and inductance value of the coil can be predicted. The present invention, by simply photographing the coil and inputting the operating frequency, can quickly and accurately complete the parameter identification task without disassembling the original equipment, greatly improving the speed and accuracy of coil parameter identification. Attached Figure Description

[0016] Figure 1 A flowchart illustrating a method for identifying wireless power transmission coil parameters provided in an embodiment of the present invention;

[0017] Figure 2 This is an architecture diagram of the parameter recognition model of the present invention;

[0018] Figure 3 This is a schematic diagram of the decoder structure of the parameter recognition model of the present invention;

[0019] Figure 4 This refers to the image data of the coils in the training set of this invention;

[0020] Figure 5 This is a schematic diagram of the loss function curve during the training of the parameter recognition model of the present invention;

[0021] Figure 6 This refers to image data of the coils in the test section of this invention;

[0022] Figure 7 This is a schematic diagram illustrating the loss of the test results of this invention;

[0023] Figure 8 The diagram shown is a structural block diagram of a wireless power transmission coil parameter identification system provided in an embodiment of the present invention;

[0024] Figure 9 The diagram shown is a structural schematic of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0025] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, unless otherwise specified, the following embodiments and features described therein can be combined with each other.

[0026] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0027] In the following description, numerous details are explored to provide a more thorough explanation of embodiments of the invention. However, it will be apparent to those skilled in the art that embodiments of the invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring embodiments of the invention.

[0028] This invention provides a method for identifying parameters of a wireless power transmission coil. It employs a multimodal model approach, considering that the coil's inductance and quality factor are not only related to the coil's own parameters but also affected by its operating frequency. Therefore, this invention simultaneously considers the coil's image and operating frequency. After inputting the image and frequency into the parameter identification model, corresponding image embedding features and text embedding features are obtained. These two features are then fused to obtain a multimodal hybrid feature. Decoding this hybrid feature allows for the prediction of the coil's quality factor and inductance. This invention, by simply photographing the coil and inputting the operating frequency, can accurately complete the parameter identification task in milliseconds without disassembling the original equipment, significantly improving the speed and accuracy of coil parameter identification.

[0029] Please see Figure 1 The method for identifying the parameters of a wireless power transfer coil includes the following steps:

[0030] S1. Obtain the operating frequency of the wireless power transfer coil under test, and an image of the wireless power transfer coil at that operating frequency.

[0031] When identifying the quality factor and inductance value of a wireless power transmission coil, it is first necessary to photograph the coil and obtain its operating frequency. The operating frequency refers to the electromagnetic wave frequency of the wireless power transmission coil under preset operating conditions, which includes, but is not limited to, common frequencies such as 85kHz, 100kHz, 200kHz, 400kHz, and 6.78MHz. Furthermore, to more clearly demonstrate the coil's structural characteristics so that the subsequent parameter identification model can comprehensively and accurately extract these characteristics, a top view of the coil is preferably used as the coil image. It should be noted that the image of the wireless power transmission coil does not change with the operating frequency; for the same wireless power transmission coil, it has the same image at different operating frequencies.

[0032] S2. Input the image of the wireless power transmission coil and the operating frequency into the parameter recognition model, extract the structural features and the operating frequency features from the image, and obtain the quality factor and inductance value of the wireless power transmission coil based on the structural features and the operating frequency features.

[0033] An image of the wireless power transmission coil and its corresponding operating frequency are input into a trained parameter recognition model. A convolutional neural network (CNN) or other image processing network is used to extract the coil's structural features from the input image. These structural features include, but are not limited to, the coil's geometry, winding method, wire diameter, and number of turns. Similarly, a CNN or other processing network is used to extract the operating frequency features. After fusing the frequency and structural features, a comprehensive multimodal hybrid feature is generated, which can then be used to predict the coil's quality factor Q and inductance L.

[0034] In one embodiment of the present invention, the step of inputting an image of the wireless power transmission coil and the operating frequency into a parameter recognition model, extracting structural features from the image and the operating frequency features, and obtaining the quality factor and inductance value of the wireless power transmission coil based on the structural features and the operating frequency features includes:

[0035] The image is input into the image encoder module of the parameter recognition model to extract the structural features of the wireless power transmission coil from the image, thereby obtaining image embedding features; wherein, the image encoder module is a convolutional neural network;

[0036] The running frequency is input into the text encoder module of the parameter recognition model to extract the running frequency features and obtain the text embedding features; wherein, the text encoder module is a convolutional neural network;

[0037] The image embedding features and the text embedding features are input into the fusion module of the parameter recognition model, and the two are fused to obtain multimodal hybrid features;

[0038] The multimodal hybrid features are input into the decoder module of the parameter recognition model to decode the multimodal hybrid features and obtain the quality factor and inductance value of the wireless power transmission coil.

[0039] like Figure 2 As shown, the parameter recognition model in this embodiment includes an encoder module, a fusion module (pack), and a decoder module (Decoder). The encoder module includes an image encoder module and a text encoder module, used to extract the structural features and operating frequency features of the coil, respectively. Specifically, both the image encoder module and the text encoder module use a Convolutional Neural Network (CNN) architecture. The image of the coil is input to the image encoder module, where multiple convolutional layers of the CNN progressively extract the coil's structural features. These extracted structural features are then further combined and compressed to generate a fixed-dimensional image embedding feature. Similarly, the operating frequency of the wireless power coil is input to the text encoder module, where multiple convolutional layers of the CNN progressively extract the coil's operating frequency features, ultimately generating a fixed-dimensional text embedding feature. Text embedding features and image embedding features are fused using a fusion module to obtain a high-dimensional multimodal hybrid feature. This multimodal hybrid feature integrates the structural information and operating frequency information of the image, enabling a more comprehensive description of the coil's electrical performance characteristics, thus facilitating the subsequent prediction of the coil's inductance value L and quality factor Q. The multimodal hybrid feature is then input into the decoder module, where it is replaced with the quality factor Q and inductance value L of the wireless power transmission coil. Specifically, as shown... Figure 3 As shown, the decoder module includes multiple fully connected layers and one output layer. These fully connected layers serve as hidden layers, progressively mapping high-dimensional multimodal mixed features to the target output space through layer-by-layer transformation. The final output layer includes two neurons, each predicting a parameter (i.e., inductance value and quality factor). Preferably, the decoder module has two fully connected layers with sizes of 8192 and 128 respectively, and includes an output layer, enabling the model to perform parameter identification efficiently and accurately.

[0040] In one embodiment of the present invention, the step of inputting the image into the image encoder module of the parameter recognition model and extracting the structural features of the wireless power transmission coil from the image to obtain image embedding features includes:

[0041] The image is then normalized in size.

[0042] The normalized image is input into the image encoder module, which extracts multiple structural features of the wireless power transmission coil from the image and performs feature compression on the multiple structural features to generate image embedding features.

[0043] To ensure consistent image size during both training and inference phases and to reduce computational complexity caused by varying image sizes, the coil images are normalized by scaling or cropping them to a preset size (e.g., 64*64 pixels), ensuring all input images have the same size and resolution. The normalized images are then fed into an image encoder module, which decomposes them into multiple pixel blocks and extracts low- to high-level features layer by layer through multiple neural network layers, ultimately yielding multiple structural features of the coil image. These extracted structural features are then compressed and converted into fixed-dimensional image embedding features.

[0044] In one embodiment of the present invention, the step of inputting the normalized image to the image encoder module, extracting multiple structural features of the wireless power transmission coil from the image, and performing feature compression on the multiple structural features to generate image embedding features includes:

[0045] The normalized image is input to the feature extraction unit group of the image encoder module. Feature extraction is performed step by step by each feature extraction unit. After each extraction is completed, multiple initial structural features of the wireless power transmission coil are generated. The initial structural features generated by the last feature extraction unit are used as the corresponding structural features. The feature extraction unit group includes multiple cascaded feature extraction units.

[0046] Multiple structural features are input into the fully connected unit of the image encoder module for dimensionality reduction and merging processing to generate image embedding features.

[0047] like Figure 2 As shown, the normalized image is input into the image encoder module. The specific structure of the image encoder module is as follows: Figure 2As shown in the CNN Model, a convolutional neural network captures detailed features such as the coil diameter and coil bends in the coil image, thereby improving the recognition accuracy of subsequent models. Specifically, the image encoder module includes a feature extraction unit group composed of multiple cascaded feature extraction units. Each feature extraction unit includes a convolutional layer, an activation layer, and a pooling layer. Each feature extraction unit extracts features based on the input image / initial structural features, generating the initial structural features of the current layer. Specifically, for the first feature extraction unit, the image is input to the first feature extraction unit. Through multiple convolutional kernels in the convolutional layer, low-level features such as edges, corners, and texture information of the image are extracted from the image. Then, these features are nonlinearly transformed through an activation layer (such as the ReLU function) to enhance the model's ability to express complex features. The transformed features are then downsampled through a pooling layer to reduce the computational complexity of the model and to impart a certain degree of invariance to the features (such as translation invariance). The output of the pooling layer is used as the initial structural features generated by the current feature extraction unit. Starting from the second feature extraction unit, the system receives the initial structural features generated by the previous feature extraction unit. These features are then combined and analyzed through convolutional layers, activation layers, and pooling layers in the current feature extraction unit to extract deeper features. This process is repeated until the last feature extraction unit, at which point the extracted higher-level features are output as the final structural features. These structural features are then input to a fully connected unit, where multiple fully connected layers perform layer-by-layer dimensionality reduction. Finally, a weighted summation method is used to merge the input structural features, generating fixed-dimensional image embedding features.

[0048] In one embodiment of the present invention, the step of inputting the running frequency into the text encoder module of the parameter recognition model and extracting the running frequency features to obtain text embedding features includes:

[0049] The operating frequency is feature-extended to obtain a one-dimensional tensor.

[0050] The one-dimensional tensor is input into the text encoder module, multiple operating frequency features of the one-dimensional tensor are extracted, and feature compression is performed on the multiple operating frequency features to obtain text embedding features.

[0051] By normalizing or standardizing the original execution frequencies, the numerical values ​​of the execution frequencies are scaled to a preset range, and this data is repeatedly copied to form a one-dimensional tensor with multiple identical elements. Furthermore, a transformation function can be used to convert the execution frequencies into a multi-dimensional representation, forming a one-dimensional tensor. This one-dimensional tensor is input into the text encoder module, where it extracts multiple execution frequency features layer by layer through multiple convolutional layers, activation layers, and pooling layers. These execution frequency features are then linearly transformed through the fully connected layer of the text encoder module to generate a fixed-dimensional text embedding feature.

[0052] In one embodiment of the present invention, the feature fusion of the image embedding features and the text embedding features to obtain multimodal hybrid features includes: weighted summing of the image embedding features and the text embedding features to obtain multimodal hybrid features. By assigning different weights to the image embedding features and the text embedding features respectively, and then weighted summing the two, multimodal hybrid features are obtained. It is understood that the weights are obtained through training a parametric recognition model.

[0053] In one embodiment of the present invention, the parameter recognition model is trained, and the training process of the parameter recognition model includes:

[0054] The image of the wireless power transmission coil and its corresponding operating frequency are pre-acquired and input into the parameter identification model to obtain the predicted value of the quality factor and the predicted value of the inductance.

[0055] Calculate the mean square error of the quality factor between the pre-obtained true value and the predicted value of the quality factor;

[0056] Calculate the mean square error of the inductance value between the actual value and the predicted value of the inductance value obtained in advance;

[0057] The mean square error of the quality factor and the mean square error of the inductance value are weighted and processed to obtain the comprehensive error.

[0058] Based on the comprehensive error, the parameters of the parameter recognition model are updated according to the gradient descent method to obtain the trained parameter recognition model.

[0059] When training the parameter recognition model, pre-acquired images of the wireless power transmission coil and their corresponding operating frequencies are input into the model. An image encoder module extracts image embedding features from the images, and a text encoder module extracts text embedding features from the operating frequencies. These two embedding features are fused to obtain multimodal hybrid features, which are then input into a decoder module to obtain predicted quality factor and inductance values ​​corresponding to the input images and operating frequencies. For each sample in the current batch: the mean squared error between the predicted and true quality factor values ​​for that sample is calculated, and the mean squared errors of all values ​​in the current batch are averaged to obtain the mean squared error of the quality factor for that batch. Similarly, the mean squared error of the inductance value for the current batch can also be obtained. These two mean squared errors are weighted and summed to obtain a combined error. Based on the combined error, gradient descent is used to update the parameters in the model layer by layer. This process is repeated until the combined error converges or the training reaches a preset number of iterations, resulting in a trained parameter recognition model.

[0060] like Figure 4 As shown, this illustrates a portion of the training dataset. The training set includes coils with and without ferromagnetic cores, excitation wires of varying thicknesses, and coils of different shapes. This comprehensive dataset enables the model to adapt to a wide range of recognition tasks. To improve the model's generalization performance and prediction accuracy, a larger dataset could be collected. However, considering the difficulty of data collection, the dataset in this invention includes 20 different types of coils, each with five different frequencies (85kHz, 100kHz, 200kHz, 400kHz, 6.78MHz), totaling 100 data sets. During training, to normalize the original images, these training images were uniformly adjusted to 64*64 pixels. To evaluate output quality, two different evaluation metrics, mean squared error and error, were used, such as... Figure 5 As shown, the loss steadily decreases with increasing training epochs, demonstrating the effectiveness of the invention. Figure 6 The test set used in this invention is shown below. Figure 7 The following is an example of using Figure 6 The results of testing on the test set. From Figure 7 It can be seen that at an operating frequency of 85kHz, both the mean squared error loss (MSELoss) and the error reach their lowest values, at 0.011% and 21.6%, respectively. Furthermore, as the frequency increases, both the mean squared error loss and the error remain at a low level. Understandably, since only 20 sets of coil images are currently used as the training dataset, increasing the number of datasets will further improve the accuracy of the parameter recognition model.

[0061] Please see Figure 8The wireless power transfer coil parameter identification system 100 includes a data acquisition module 110 and a parameter prediction module 120. The data acquisition module 110 acquires the operating frequency of the wireless power transfer coil under test and an image of the coil at that operating frequency. The parameter prediction module 120 inputs the image of the wireless power transfer coil and the operating frequency into a parameter identification model, extracts structural features from the image and the operating frequency features, and obtains the quality factor and inductance value of the wireless power transfer coil based on the structural features and the operating frequency features.

[0062] Specific limitations regarding the wireless power transfer coil parameter identification system can be found in the limitations of the wireless power transfer coil parameter identification method described above, and will not be repeated here. Each module in the aforementioned wireless power transfer coil parameter identification system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in hardware format within or independently of the processor in a computer device, or stored in software format in the memory of a computer device, so that the processor can call the corresponding operations of each module.

[0063] It should be noted that, in order to highlight the innovative aspects of this invention, this embodiment does not include modules that are not closely related to solving the technical problems proposed by this invention, but this does not mean that there are no other modules in this embodiment.

[0064] Please see Figure 9 The electronic device 1 may include a memory 12, a processor 13 and a bus, and may also include a computer program stored in the memory 12 and executable on the processor 13, such as a wireless power transfer coil parameter identification program.

[0065] The memory 12 includes at least one type of readable storage medium, such as flash memory, portable hard drive, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 12 can be an internal storage unit of the electronic device 1, such as a portable hard drive. In other embodiments, the memory 12 can be an external storage device of the electronic device 1, such as a plug-in portable hard drive, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the electronic device 1. Furthermore, the memory 12 can include both internal and external storage units of the electronic device 1. The memory 12 can be used not only to store application software and various types of data installed on the electronic device 1, such as codes for identifying wireless power transmission coil parameters, but also to temporarily store data that has been output or will be output.

[0066] In some embodiments, the processor 13 may be composed of integrated circuits, such as a single packaged integrated circuit or multiple integrated circuits with the same or different functions, including combinations of one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips. The processor 13 is the control unit of the electronic device 1, connecting various components of the electronic device 1 via various interfaces and lines. It executes programs or modules stored in the memory 12 (e.g., a wireless power transmission coil parameter identification program) and calls data stored in the memory 12 to perform various functions and process data of the electronic device 1.

[0067] The processor 13 executes the operating system of the electronic device 1 and various installed applications. The processor 13 executes the applications to implement the steps in the above-described method for predicting the power battery capacity trajectory.

[0068] For example, the computer program may be divided into one or more modules, which are stored in the memory 12 and executed by the processor 13 to complete this application. The one or more modules may be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program in the electronic device 1. For example, the computer program may be divided into a data acquisition module 110 and a parameter prediction module 120.

[0069] The integrated unit implemented as a software functional module described above can be stored in a computer-readable storage medium, which can be non-volatile or volatile. The software functional module, stored in the storage medium, includes several instructions to cause a computer device (which may be a personal computer, computer equipment, or network device, etc.) or processor to execute some functions of the power battery capacity trajectory prediction method described in the various embodiments of this application.

[0070] In summary, the present invention discloses a method, system, device, and medium for identifying parameters of a wireless power transmission coil. It employs a multimodal model approach, considering that the coil's inductance and quality factor are not only related to the coil's own parameters but also affected by its operating frequency. Therefore, the present invention simultaneously considers the coil's image and operating frequency. After inputting the image and frequency into the parameter identification model, corresponding image embedding features and text embedding features are obtained. These two features are then fused to obtain a multimodal hybrid feature. Decoding this hybrid feature allows for the prediction of the coil's quality factor and inductance. This invention, by simply photographing the coil and inputting the operating frequency, can accurately complete the parameter identification task in milliseconds without disassembling the original equipment, significantly improving the speed and accuracy of coil parameter identification. Therefore, the present invention effectively overcomes the various shortcomings of existing technologies and has high industrial application value.

[0071] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A method for identifying parameters of a wireless power transmission coil, characterized in that, The method includes: Acquire the operating frequency of the wireless power transfer coil under test, and an image of the wireless power transfer coil at that operating frequency; The image is then normalized in size. The normalized image is input into the feature extraction unit group of the image encoder module of the parameter recognition model. Feature extraction is performed step by step by each feature extraction unit. After each extraction is completed, multiple initial structural features of the wireless power transmission coil are generated accordingly. The initial structural features generated by the last feature extraction unit are used as the corresponding structural features. The image encoder module is a convolutional neural network, and the feature extraction unit group includes multiple cascaded feature extraction units. Multiple structural features are input into the fully connected unit of the image encoder module for dimensionality reduction and merging processing to generate image embedding features; The running frequency is input into the text encoder module of the parameter recognition model to extract the running frequency features and obtain text embedding features; wherein, the text encoder module is a convolutional neural network; The image embedding features and the text embedding features are input into the fusion module of the parameter recognition model, and the two are fused to obtain multimodal hybrid features; The multimodal hybrid features are input into the decoder module of the parameter recognition model to decode the multimodal hybrid features and obtain the quality factor and inductance value of the wireless power transmission coil.

2. The method for identifying wireless power transmission coil parameters according to claim 1, characterized in that, The text encoder module that inputs the running frequency into the parameter recognition model, extracts the running frequency features, and obtains text embedding features includes: The operating frequency is feature-extended to obtain a one-dimensional tensor. The one-dimensional tensor is input into the text encoder module, multiple operating frequency features of the one-dimensional tensor are extracted, and feature compression is performed on the multiple operating frequency features to obtain text embedding features.

3. The method for identifying wireless power transmission coil parameters according to claim 1, characterized in that, The step of inputting the image embedding features and the text embedding features into the fusion module of the parameter recognition model and fusing the two features to obtain multimodal hybrid features includes: weighting and summing the image embedding features and the text embedding features to obtain multimodal hybrid features.

4. The method for identifying wireless power transmission coil parameters according to claim 1, characterized in that, The parameter recognition model is trained, and the training process of the parameter recognition model includes: The image of the wireless power transmission coil and its corresponding operating frequency are pre-acquired and input into the parameter identification model to obtain the predicted value of the quality factor and the predicted value of the inductance. Calculate the mean square error of the quality factor between the pre-obtained true value and the predicted value of the quality factor; Calculate the mean square error of the inductance value between the actual value and the predicted value of the inductance value obtained in advance; The mean square error of the quality factor and the mean square error of the inductance value are weighted and processed to obtain the comprehensive error. Based on the comprehensive error, the parameters of the parameter recognition model are updated according to the gradient descent method to obtain the trained parameter recognition model.

5. A system for identifying parameters of a wireless power transmission coil, characterized in that, The system includes: The data acquisition module is used to acquire the operating frequency of the wireless power transmission coil under test, and the image of the wireless power transmission coil at that operating frequency; The parameter prediction module is used to perform size normalization processing on the image; The normalized image is input into the feature extraction unit group of the image encoder module of the parameter recognition model. Feature extraction is performed step by step by each feature extraction unit. After each extraction is completed, multiple initial structural features of the wireless power transmission coil are generated accordingly. The initial structural features generated by the last feature extraction unit are used as the corresponding structural features. The image encoder module is a convolutional neural network, and the feature extraction unit group includes multiple cascaded feature extraction units. Multiple structural features are input into the fully connected unit of the image encoder module for dimensionality reduction and merging processing to generate image embedding features; The running frequency is input into the text encoder module of the parameter recognition model to extract the running frequency features and obtain text embedding features; wherein, the text encoder module is a convolutional neural network; The image embedding features and the text embedding features are input into the fusion module of the parameter recognition model, and the two are fused to obtain multimodal hybrid features; The multimodal hybrid features are input into the decoder module of the parameter recognition model to decode the multimodal hybrid features and obtain the quality factor and inductance value of the wireless power transmission coil.

6. An electronic device, characterized in that, The electronic device includes: One or more processors; A storage device for storing one or more programs, which, when executed by the one or more processors, cause the electronic device to implement the method for identifying wireless power transmission coil parameters as described in any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that, It stores a computer program that, when executed by the computer's processor, causes the computer to perform the method for identifying the parameters of the wireless power transmission coil as described in any one of claims 1 to 4.

Images