Welding parameter determination method and device and nonvolatile storage medium

Abstract

The invention discloses a welding parameter determining method and device and a nonvolatile storage medium. The method comprises the steps that material data of a to-be-welded battery cell are obtained, and the material data represent the structure and performance of the to-be-welded battery cell; the material data are input into a pre-trained neural network model, welding parameters of the to-be-welded battery cell are output through the neural network model, the neural network model is obtained by training an original neural network model through a training sample, and the training sample comprises sample material data and sample welding parameters of the sample battery cell. The technical problem that the efficiency of a method for determining the welding parameters used for welding the tabs of the battery cell is low is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of battery manufacturing, and in particular to a method and device for determining welding parameters and a non-volatile storage medium. Background Art

[0002] With the development of new-generation information technologies such as 5G, cloud computing, and big data, as well as the widespread application of new devices such as electric vehicles, smart terminals, and drones, the battery industry has ushered in explosive growth. The development of battery technology is driving the innovation of cells and batteries. Cells are the basic components of batteries, and the energy transfer of cells is crucial.

[0003] At present, the energy density of batteries used in new energy vehicles is generally less than 300Wh / kg, which is far lower than the target of more than 500Wh / kg for the energy density of the new generation of batteries. In order to meet the needs of the rapid development of new energy vehicles, the energy density of the new generation of batteries needs to be further improved. The energy density and power density of the new generation of batteries are much higher than those of traditional batteries, and the welding process of the battery cell has an important impact on the energy density and power density of the battery.

[0004] In the welding process, the selection of welding parameters is proportional to the welding quality. Whether the welding parameters are selected reasonably directly affects the welding quality, welding efficiency and welding cost. However, the selection of welding parameters depends on the experience of the welding operator. Different welding operators have different choices of welding parameters and different welding quality. Ultrasonic welding uses the high-frequency vibration energy of ultrasound to weld the battery cell tabs, but the selection of welding parameters needs to be based on the material data of the battery cell and obtained through experiments. Therefore, it requires a lot of manpower and material resources, which reduces welding efficiency and increases welding costs.

[0005] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0006] The embodiments of the present invention provide a method and device for determining welding parameters and a non-volatile storage medium, so as to at least solve the technical problem that the method for determining welding parameters used in welding the tabs of a battery cell is inefficient.

[0007] According to one aspect of an embodiment of the present invention, a method for determining welding parameters is provided, comprising: obtaining material data of a battery cell to be welded, wherein the material data characterizes the structure and performance of the battery cell to be welded; inputting the material data into a pre-trained neural network model, and the neural network model outputs the welding parameters of the battery cell to be welded, wherein the neural network model is obtained by training an original neural network model through training samples, and the training samples include sample material data and sample welding parameters of the sample battery cell. This embodiment automatically obtains welding parameters by directly inputting material data into the neural network model, thereby achieving real-time prediction and adjustment of welding process parameters. This greatly reduces the reliance on manual experience, improves the automation level and response speed of the battery cell production line, and makes the battery cell manufacturing process more efficient and flexible.

[0008] Optionally, the material data is input into a pre-trained neural network model, and the neural network model outputs the welding parameters of the battery cells to be welded, including: inputting the material data into a convolutional network unit included in the neural network model, and the convolutional network unit extracts material features from the material data, wherein the material features characterize the material properties of the battery cells to be welded; inputting the material features into a long short-term memory network unit included in the neural network model, and the long short-term memory network unit outputs the welding parameters. The neural network model obtained by the combination of the convolutional neural network and the long short-term memory network in this optional embodiment can output the best welding parameter combination to meet the welding requirements of specific battery cell materials. This parameter optimization is based on autonomous learning and adjustment of deep learning algorithms, which can automatically adapt to subtle differences in materials, ensure the consistency and high quality of the welding process, and reduce welding defects and production delays caused by improper parameter settings.

[0009] Optionally, the neural network model is trained in the following manner: the training sample is input into the original neural network model, and the original neural network model adjusts the parameters of the original neural network model according to a predetermined loss function; the neural network model is determined according to the parameters that make the loss function meet the predetermined conditions. This optional embodiment trains the neural network model to optimize the welding parameter prediction, which not only improves the prediction accuracy and stability of the model, but also significantly reduces the production cost and improves the production efficiency, providing strong technical support for the intelligent and efficient production of the battery manufacturing industry.

[0010] Optionally, the training samples are input into the original neural network model, and the original neural network model adjusts the parameters of the original neural network model according to a predetermined loss function, including: inputting the sample material data into the original neural network model, and the original neural network model determines the predicted welding parameters; the original neural network model determines the value of the loss function according to the predicted welding parameters and the sample welding parameters; and the back propagation algorithm is used to adjust the parameters of the original neural network model according to the value of the loss function. This optional embodiment inputs the training samples into the original neural network model, and adjusts the model parameters through the loss function and the back propagation algorithm, which can significantly improve the accuracy and reliability of welding parameter prediction, reduce production costs, and improve the controllability and consistency of the welding process, thereby promoting the intelligent transformation and efficiency improvement of the battery manufacturing industry.

[0011] Optionally, a back propagation algorithm is used to adjust the parameters of the original neural network model according to the value of the loss function, including: determining the gradient of the value of the loss function; and using an adaptive moment estimation algorithm to adjust the parameters of the original neural network model according to the gradient of the value of the loss function. This optional embodiment uses a back propagation algorithm and an adaptive moment estimation algorithm to adjust the parameters of the neural network model, which can significantly improve the training efficiency and prediction accuracy of the model, reduce the risk of overfitting, accelerate model convergence, and provide a smarter, more efficient, and more reliable welding parameter prediction solution for the battery manufacturing industry.

[0012] Optionally, the material data includes at least one of the following: the number of tab layers, the foil thickness and the weld mark depth, wherein the number of tab layers is the number of layers of metal foil included in the tab of the battery cell to be welded, the foil thickness is the thickness of the metal foil constituting the tab of the battery cell to be welded, and the weld mark depth is the depth of the weld mark that needs to be formed on the tab of the battery cell to be welded. This optional embodiment uses the number of tab layers, foil thickness and weld mark depth as material data to provide specific structure and performance information of the battery cell tab. The neural network model can accurately output welding parameters that match the characteristics of the battery cell material by learning the relationship between these data and welding parameters. This means that tabs with different numbers of layers, thicknesses and weld mark depth requirements can obtain customized welding parameters to ensure the consistency of welding quality and battery cell performance.

[0013] According to another aspect of an embodiment of the present invention, a device for determining welding parameters is also provided, including: an acquisition module, used to acquire material data of a battery cell to be welded, wherein the material data characterizes the structure and performance of the battery cell to be welded; a determination module, used to input the material data into a pre-trained neural network model, and the neural network model outputs the welding parameters of the battery cell to be welded, wherein the neural network model is obtained by training an original neural network model through training samples, and the training samples include sample material data and sample welding parameters of the sample battery cell.

[0014] According to another aspect of an embodiment of the present invention, a non-volatile storage medium is provided, the non-volatile storage medium including a stored program, wherein when the program is running, the device where the non-volatile storage medium is located is controlled to execute any one of the above-mentioned methods for determining welding parameters.

[0015] According to another aspect of the embodiments of the present invention, a computer device is provided. The computer device includes a processor, and the processor is used to run a program. When the program is run, any one of the above-mentioned methods for determining welding parameters is executed.

[0016] According to another aspect of the embodiments of the present invention, there is also provided a computer program product, including a computer program, which implements any one of the above-mentioned methods for determining welding parameters when the computer program is executed by a processor.

[0017] In an embodiment of the present invention, material data of a battery cell to be welded is obtained, wherein the material data characterizes the structure and performance of the battery cell to be welded; the material data is input into a pre-trained neural network model, and the neural network model outputs welding parameters of the battery cell to be welded, wherein the neural network model is obtained by training an original neural network model through training samples, and the training samples include sample material data and sample welding parameters of the sample battery cell, thereby achieving the purpose of accurately predicting the optimal welding parameters based on the material data of the battery cell, greatly improving the efficiency and welding quality of the welding process, and further solving the technical problem of low efficiency of the method for determining the welding parameters used in welding the tabs of the battery cell. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0019] Figure 1 A hardware structure block diagram of a computer terminal for implementing a method for determining welding parameters is shown;

[0020] Figure 2 is a flow chart of a method for determining welding parameters according to an embodiment of the present invention;

[0021] Figure 3 is a schematic diagram of the structure of a neural network model provided according to an optional embodiment of the present invention;

[0022] Figure 4 is a schematic diagram of a welding parameter prediction process according to an optional embodiment of the present invention;

[0023] Figure 5 4 is a structural block diagram of a device for determining welding parameters provided according to an embodiment of the present invention. DETAILED DESCRIPTION

[0024] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0025] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0026] According to an embodiment of the present invention, a method embodiment for determining welding parameters is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0027] The method embodiment provided in the first embodiment of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Figure 1 FIG. 1 shows a hardware structure block diagram of a computer terminal for implementing a method for determining welding parameters. Figure 1 As shown, the computer terminal 10 may include one or more (102a, 102b, ..., 102n are used to illustrate) processors (the processor may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 104 for storing data. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the BUS bus), a network interface, a power supply and / or a camera. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 1More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0028] It should be noted that the one or more processors and / or other data processing circuits described above may generally be referred to herein as "data processing circuits". The data processing circuits may be embodied in whole or in part as software, hardware, firmware, or any other combination thereof. In addition, the data processing circuit may be a single independent processing module, or may be incorporated in whole or in part into any of the other components in the computer terminal 10. As described in the embodiments of the present application, the data processing circuit acts as a processor control (e.g., selection of a variable resistor terminal path connected to an interface).

[0029] The memory 104 can be used to store software programs and modules of application software, such as program instructions / data storage devices corresponding to the method for determining welding parameters in the embodiment of the present invention. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, the method for determining welding parameters of the above-mentioned application program is realized. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the computer terminal 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0030] The display may be, for example, a touch screen liquid crystal display (LCD) that enables a user to interact with a user interface of the computer terminal 10 .

[0031] The tabs of a battery cell refer to the metal sheets or wires that connect the positive and negative electrode materials in batteries such as lithium batteries. Tabs are usually located at both ends of the battery and are used to connect the positive and negative electrodes of the battery to the external circuit. The design and connection method of the tabs have a great impact on the performance and safety of the battery. In order to connect the positive and negative electrodes of the battery to form a complete circuit, the tabs of the battery cell need to be welded.

[0032] Ultrasonic welding of tabs is one of the key links in the production process of battery cells. The quality of welding affects the performance, safety and production cost of the battery cells. Good welding quality can ensure the stability of battery cell performance, while improving the vibration resistance and impact resistance of welding points and enhancing the safety of battery cells. On the contrary, poor welding quality may cause battery cells to malfunction during use, reduce product safety, and increase the cost of subsequent repairs or recalls.

[0033] The influence of ultrasonic welding parameter setting on the welding of the tab is crucial. The welding parameters determine the heating, pressing and vibration of the material during the welding process, thus affecting the welding quality and efficiency. The increase in energy parameters will accelerate the softening and flow of the material during the welding process and improve the welding strength, but too high energy may cause overheating of the welding area, so the energy needs to be controlled. Appropriate pressure parameters can ensure that the welding contact surface fits tightly and improves the welding strength, while too high or too low pressure may affect the welding quality and efficiency. The amplitude parameters affect the vibration intensity and frequency of the welding area. The appropriate amplitude can improve the welding strength and sealing, but too large or too small amplitude may result in poor welding results. Therefore, the selection of ultrasonic welding parameters needs to be based on comprehensive consideration of factors such as welding materials, shapes, and process requirements, and appropriate adjustments and controls should be made to ensure that the welding process is stable and efficient, and obtain good welding quality.

[0034] Traditional ultrasonic welding parameter selection requires multiple welding experiments to find the best value. Its disadvantage is that multiple welding experiments require a lot of time and resources. In addition, traditional methods can usually only find the best parameters among the preset parameter values. Manual parameter preset has certain limitations and cannot guarantee that the preset parameters contain the best parameters.

[0035] In order to solve the above problems, the present invention provides a method for determining welding parameters. Figure 2 is a flow chart of a method for determining welding parameters according to an embodiment of the present invention, such as Figure 2 As shown, the method comprises the following steps:

[0036] Step S202, obtaining material data of the battery cells to be welded, wherein the material data represents the structure and performance of the battery cells to be welded.

[0037] In this step, the battery cell to be welded is the battery cell for which welding parameters need to be determined this time. The material data can be data that can characterize the characteristics of the tab material of the battery cell, and can reflect the actual situation of the tab material, so as to determine more appropriate welding parameters.

[0038] As an optional embodiment, the material data includes at least one of the following: the number of tab layers, the foil thickness and the weld mark depth, wherein the number of tab layers is the number of layers of metal foil included in the tab of the battery cell to be welded, the foil thickness is the thickness of the metal foil constituting the tab of the battery cell to be welded, and the weld mark depth is the depth of the weld mark that needs to be formed on the tab of the battery cell to be welded.

[0039] Optionally, the material data may include the number of tab layers, foil thickness and weld depth, and correspondingly, the welding parameters may include welding energy, pressure and amplitude. By adjusting the energy, pressure and amplitude of the welding machine, it is possible to ensure that the tabs of the battery cells with different tab layers and foil thickness can obtain an ideal weld depth, thereby ensuring the firmness of the welding and the performance of the battery cells.

[0040] The number of tab layers refers to the number of layers of the battery cell tab (usually the conductive material of the positive or negative electrode). In battery cell manufacturing, the tab is usually made of multiple layers of thin metal foil to increase the conductive area and mechanical strength. The more tab layers there are, the more welding energy is required to ensure that all layers are fully fused. An increase in the number of layers will result in an increase in the thickness of the tab, so greater pressure is required to ensure the firmness of the weld. An increase in the number of layers may affect the transmission of vibration, so the amplitude may need to be adjusted to ensure the welding effect.

[0041] Foil thickness refers to the thickness of the metal foil that makes up the tab. Foil thickness directly affects the mechanical strength and electrical conductivity of the tab. The thicker the foil, the more welding energy is required to penetrate the thicker metal layer and achieve good fusion. Increased thickness will cause uneven pressure distribution in the welding area, so more pressure is required to ensure uniform welding. Increased foil thickness may affect the transmission of vibration, so the amplitude also needs to be adjusted to ensure the welding effect.

[0042] Weld depth refers to the depth of the weld mark formed on the lug after welding. Weld depth is an important indicator to measure welding quality. It is usually required that the weld depth reaches a certain standard to ensure the firmness of the weld. The higher the weld depth requirement, the greater the welding energy required to ensure that the weld depth reaches the expected standard. The higher the weld depth requirement, the greater the welding pressure required to ensure that the metal in the welding area is fully fused. The higher the weld depth requirement, the amplitude may need to be adjusted to ensure sufficient metal fusion and weld mark formation during welding.

[0043] This optional embodiment uses the number of tab layers, foil thickness, and weld depth as material data to provide specific structural and performance information of the cell tabs. The neural network model can accurately output welding parameters that match the cell material characteristics by learning the relationship between these data and welding parameters. This means that tabs with different number of layers, thickness, and weld depth requirements can all obtain customized welding parameters to ensure consistency in welding quality and cell performance.

[0044] Step S204, inputting material data into a pre-trained neural network model, and having the neural network model output welding parameters of the battery cells to be welded, wherein the neural network model is obtained by training an original neural network model using training samples, and the training samples include sample material data and sample welding parameters of the sample battery cells.

[0045] In this step, the material data can be directly input into the pre-trained neural network model to automatically obtain the welding parameters and realize the real-time prediction and adjustment of the welding process parameters. The neural network model can be trained based on a large number of training samples, which include sample material data of a large number of sample cells and the optimal sample welding parameters determined when welding the sample cells. The use of the neural network model to predict the welding parameters greatly reduces the dependence on manual experience, improves the automation level and response speed of the cell production line, and makes the cell manufacturing process more efficient and flexible.

[0046] As an optional embodiment, material data is input into a pre-trained neural network model, and the neural network model outputs welding parameters of the battery cells to be welded, including: inputting material data into a convolutional network unit included in the neural network model, and the convolutional network unit extracts material features from the material data, wherein the material features characterize the material properties of the battery cells to be welded; inputting material features into a long short-term memory network unit included in the neural network model, and the long short-term memory network unit outputs welding parameters.

[0047] Optionally, the neural network model provided by this optional embodiment can be a convolution-long short-term memory network model (CNN-LSTM), that is, the neural network model is a combination of a convolutional neural network and a long short-term memory network. Specifically, the model takes the number of tab layers, foil thickness and weld depth of the battery cell as input data, and first extracts the spatial features of the input data through a convolutional neural network (CNN), wherein the convolution layer and pooling layer of the convolutional neural network (CNN) can identify and extract these important local features, and the extracted spatial features reflect the geometric properties of the battery cell tab material. The extracted spatial features can then be input into a long short-term memory network (LSTM) for processing. The long short-term memory network (LSTM) is good at processing sequence data, can capture temporal dependencies and sequence information, and further model the dynamic changes of these features in the time dimension. In this way, the model can more comprehensively analyze and understand the input data, thereby more accurately predicting the parameters required for welding, including energy, pressure and amplitude.

[0048] This optional embodiment uses a neural network model obtained by combining a convolutional neural network with a long short-term memory network to output the best welding parameter combination to meet the welding requirements of specific battery core materials. This parameter optimization is based on autonomous learning and adjustment of deep learning algorithms, which can automatically adapt to subtle differences in materials, ensure the consistency and high quality of the welding process, and reduce welding defects and production delays caused by improper parameter settings.

[0049] As an optional embodiment, the neural network model is trained in the following manner: the training samples are input into the original neural network model, and the parameters of the original neural network model are adjusted according to a predetermined loss function; the neural network model is determined according to the parameters that make the loss function meet predetermined conditions.

[0050] Optionally, during the training process of the model, the parameters within the model can be adjusted according to the loss function so that the model learns the knowledge in the training samples. This optional embodiment uses mean square error (MSE) as a loss function to measure the difference between the model's predicted value and the actual value. Mean square error is a loss function commonly used in regression problems, and is suitable for situations where the model needs to predict continuous values ​​rather than classification labels. Specifically, mean square error calculates the average of the squares of the difference between the predicted value and the actual value, and in this way quantifies the size of the prediction error, helping to optimize the model to minimize the prediction error and improve the precision and accuracy of the prediction.

[0051]

[0052] Where n is the number of samples, y i is the predicted value of the i-th sample, is the actual value of the i-th sample.

[0053] In addition, you can set R 2 As the convergence criterion of the model. The mean square error is used to measure the difference between the model prediction value and the actual value, while R 2 The model's ability to explain the data variance is evaluated. In regression tasks, R 2 It is used as an indicator to measure the goodness of fit of the model. Its value range is between 0 and 1. The closer it is to 1, the better the model fits the data. 2 As the convergence criterion of the model, it can not only monitor the degree of fit of the model during the training process, but also help evaluate the model's predictive ability for the data, thereby further optimizing the model and ensuring it has higher prediction accuracy and stability in practical applications.

[0054]

[0055] Among them, SS res Represents the residual sum of squares, which is the sum of squares of the differences between the model predictions and the actual observed values. tot represents the total sum of squares, which is the sum of the squares of the differences between the actual observations and the mean of the actual observations.

[0056] This optional embodiment optimizes welding parameter prediction by training a neural network model, which not only improves the prediction accuracy and stability of the model, but also significantly reduces production costs and improves production efficiency, providing strong technical support for intelligent and efficient production in the battery manufacturing industry.

[0057] As an optional embodiment, the training sample is input into the original neural network model, and the original neural network model adjusts the parameters of the original neural network model according to a predetermined loss function, including: inputting the sample material data into the original neural network model, and the original neural network model determines the predicted welding parameters; the original neural network model determines the value of the loss function based on the predicted welding parameters and the sample welding parameters; and adopts a back propagation algorithm to adjust the parameters of the original neural network model according to the value of the loss function.

[0058] Optionally, in this optional embodiment, the specific steps of the process of training the original neural network model to predict welding parameters are as follows: First, the training sample material data containing information such as the number of tab layers, foil thickness, and weld mark depth is input into the original neural network model. These data are the basis for model training and are used to learn the mapping relationship between different material properties and welding parameters. Then, the original neural network model performs calculations based on the input material data through its internal parameters and outputs predicted welding parameters, such as energy, pressure, and amplitude. At this time, the model attempts to predict the optimal welding parameters for given material properties based on its current knowledge. The predicted welding parameters output by the model can then be compared with the actual welding parameters in the training sample, and the gap between the predicted value and the true value can be quantified by a loss function (such as mean square error MSE). The smaller the value of the loss function, the closer the model's prediction is to the actual value, and the better the performance of the model.

[0059] Next, the back propagation algorithm (BP) is used to adjust the parameters of the model. The BP algorithm calculates the gradient of the loss function relative to the model parameters, propagates the error back from the output layer to the input layer, and updates the weights and biases of the model to minimize the loss function. This process is iterative. Each time a training sample passes through the model, the parameters are adjusted according to the calculated gradient to gradually optimize the model performance.

[0060] The above steps constitute a training loop, and the model will repeat these steps until the loss function reaches the preset convergence criterion or the number of training iterations is completed. This ensures that the model can continue to learn from the data and continuously improve the accuracy of predicting welding parameters.

[0061] This optional embodiment inputs the training samples into the original neural network model and adjusts the model parameters through the loss function and the back propagation algorithm, which can significantly improve the accuracy and reliability of welding parameter prediction, reduce production costs, and improve the controllability and consistency of the welding process, thereby promoting the intelligent transformation and efficiency improvement of the battery manufacturing industry.

[0062] As an optional embodiment, a back propagation algorithm is used to adjust the parameters of the original neural network model according to the value of the loss function, including: determining the gradient of the value of the loss function; and using an adaptive moment estimation algorithm to adjust the parameters of the original neural network model according to the gradient of the value of the loss function.

[0063] Optionally, in order to deal with problems such as exponential growth of parameters, overfitting, and slow training speed that may occur during model training, the following training methods can be used to prevent the above problems from occurring while improving the performance and training efficiency of the model. For example, the adaptive moment estimation algorithm (Adam) can be used as an optimizer during the training process of the backpropagation algorithm. Adam combines the advantages of momentum optimization and adaptive learning rate to effectively update network parameters and accelerate the convergence of the model. The core principle of Adam is to consider the first-order moment (mean) estimation and the second-order moment (variance) estimation of the gradient at the same time each time the parameters are updated. Specifically, Adam uses exponentially weighted moving average to estimate the first-order moment and second-order moment of the gradient, and performs bias correction on these estimates.

[0064] In addition, BN batch normalization can be used to process the features so that their mean is 0 and their variance is 1. The purpose is to speed up the training of the neural network and enhance the robustness and generalization ability of the model. The main purpose of BN is to deal with the problem of internal covariate shift in deep neural network training. This shift refers to the change in the distribution of input data in each layer, making network learning more difficult. BN improves the stability of the network by normalizing the features of each batch of data and eliminating the change in the distribution of input data.

[0065] As a specific embodiment, the present invention proposes a CNN-LSTM model, which is applied to the prediction of ultrasonic welding parameters of battery cell tabs, which significantly improves the welding quality and production efficiency. Through real-time data processing and predictive optimization, the automation and precise control of the welding process are realized, which has brought significant technological progress and economic benefits to battery manufacturing and related industries.

[0066] Among them, the CNN-LSTM model can realize parameter prediction in various scenarios through training with different data sets. The model combines the feature extraction capability of convolutional neural network (CNN) and the time series analysis capability of long short-term memory network (LSTM), and can effectively process complex data in the battery cell manufacturing process. By training data sets in different scenarios, the model can learn the relationship between various manufacturing parameters, so as to predict the best parameter combination, optimize the manufacturing process, and improve the quality and performance of the battery cell. Whether it is welding, coating, winding or drying, the CNN-LSTM model can predict the best parameters by analyzing the input data, ensure the stability and consistency of the manufacturing process, and ultimately improve the overall performance of the battery cell.

[0067] Figure 3 is a schematic diagram of the structure of a neural network model provided according to an optional embodiment of the present invention, such as Figure 3 As shown in the figure, the CNN-LSTM model combines a convolutional neural network (CNN) and a long short-term memory network (LSTM) to obtain prediction data. First, the data is preprocessed and normalized to improve the calculation speed and efficiency of the CNN network. The preprocessed data is fed into the CNN network for feature extraction and analysis. Among them, maxpool1D is a one-dimensional maximum pooling operation used to reduce the size of the input data and extract the most significant features. It implements the pooling operation by taking the maximum value on each sliding window of the input data. Flatten refers to the operation of converting a multidimensional array into a one-dimensional array, which is usually used to flatten the output of a convolutional layer or a pooling layer as the input of a fully connected layer. The flatten operation will retain the order of the input data, but will rearrange it into a one-dimensional array. The flatten operation is usually performed between the convolutional layer and the fully connected layer of the neural network. The extracted features are then input into the LSTM network, which outputs the prediction results.

[0068] Figure 4 is a schematic diagram of a welding parameter prediction process according to an optional embodiment of the present invention, such as Figure 4 As shown, when predicting welding parameters, the input data is first obtained, and then input into the welding parameter prediction model, and the welding parameter prediction model outputs the prediction result. Among them, the welding parameter prediction model is obtained by training the CNN-LSTM model with sample data that has been preprocessed.

[0069] It should be noted that, for the above-mentioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should know that the present invention is not limited by the described action sequence, because according to the present invention, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present invention.

[0070] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method for determining welding parameters according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for enabling a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in each embodiment of the present invention.

[0071] According to an embodiment of the present invention, a device for determining welding parameters for implementing the above-mentioned method for determining welding parameters is also provided. Figure 5 is a structural block diagram of a device for determining welding parameters according to an embodiment of the present invention, such as Figure 5 As shown, the device for determining welding parameters includes: an acquisition module 52 and a determination module 54. The device for determining welding parameters is described below.

[0072] The acquisition module 52 is used to acquire material data of the battery cell to be welded, wherein the material data represents the structure and performance of the battery cell to be welded.

[0073] The determination module 54 is connected to the acquisition module 52, and is used to input the material data into a pre-trained neural network model, and the neural network model outputs the welding parameters of the battery cells to be welded, wherein the neural network model is obtained by training the original neural network model through training samples, and the training samples include sample material data and sample welding parameters of the sample battery cells.

[0074] It should be noted that the acquisition module 52 and the determination module 54 correspond to steps S202 to S204 in the embodiment, and the examples and application scenarios implemented by the multiple modules and the corresponding steps are the same, but are not limited to the contents disclosed in the above embodiment. It should be noted that the above modules, as part of the device, can be run in the computer terminal 10 provided in the embodiment.

[0075] An embodiment of the present invention may provide a computer device. Optionally, in this embodiment, the computer device may be located in at least one network device among multiple network devices of a computer network. The computer device includes a memory and a processor.

[0076] Among them, the memory can be used to store software programs and modules, such as program instructions / modules corresponding to the method and device for determining welding parameters in the embodiment of the present invention. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, the above-mentioned method for determining welding parameters is realized. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory may further include a memory remotely arranged relative to the processor, and these remote memories can be connected to the computer terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0077] The processor can call the information and application programs stored in the memory through the transmission device to perform the following steps: obtaining material data of the battery cells to be welded, wherein the material data characterizes the structure and performance of the battery cells to be welded; inputting the material data into a pre-trained neural network model, and the neural network model outputs the welding parameters of the battery cells to be welded, wherein the neural network model is obtained by training the original neural network model through training samples, and the training samples include sample material data and sample welding parameters of the sample battery cells.

[0078] Optionally, the processor may also execute the program code of the following steps: inputting material data into a pre-trained neural network model, and having the neural network model output welding parameters of the battery cells to be welded, including: inputting material data into a convolutional network unit included in the neural network model, and having the convolutional network unit extract material features from the material data, wherein the material features characterize the material properties of the battery cells to be welded; inputting material features into a long short-term memory network unit included in the neural network model, and having the long short-term memory network unit output welding parameters.

[0079] Optionally, the processor may also execute program code of the following steps: the neural network model is trained in the following manner: the training sample is input into the original neural network model, and the parameters of the original neural network model are adjusted according to a predetermined loss function; the neural network model is determined according to the parameters that make the loss function meet predetermined conditions.

[0080] Optionally, the processor may also execute the program code of the following steps: inputting the training sample into the original neural network model, and the original neural network model adjusting the parameters of the original neural network model according to a predetermined loss function, including: inputting the sample material data into the original neural network model, and the original neural network model determines the predicted welding parameters; the original neural network model determines the value of the loss function based on the predicted welding parameters and the sample welding parameters; and using the back propagation algorithm to adjust the parameters of the original neural network model according to the value of the loss function.

[0081] Optionally, the processor may also execute the program code of the following steps: using a back-propagation algorithm to adjust the parameters of the original neural network model according to the value of the loss function, including: determining the gradient of the value of the loss function; using an adaptive moment estimation algorithm to adjust the parameters of the original neural network model according to the gradient of the value of the loss function.

[0082] Optionally, the processor may also execute the program code of the following steps: the material data includes at least one of the following: the number of tab layers, the foil thickness and the weld mark depth, wherein the number of tab layers is the number of layers of metal foil included in the tab of the battery cell to be welded, the foil thickness is the thickness of the metal foil constituting the tab of the battery cell to be welded, and the weld mark depth is the depth of the weld mark that needs to be formed on the tab of the battery cell to be welded.

[0083] By adopting the embodiment of the present invention, a solution for determining welding parameters is provided. By obtaining the material data of the battery cell to be welded, wherein the material data characterizes the structure and performance of the battery cell to be welded; the material data is input into a pre-trained neural network model, and the neural network model outputs the welding parameters of the battery cell to be welded, wherein the neural network model is obtained by training the original neural network model through training samples, and the training samples include sample material data and sample welding parameters of the sample battery cell, so as to achieve the purpose of accurately predicting the optimal welding parameters according to the material data of the battery cell, greatly improve the efficiency and welding quality of the welding process, and further solve the technical problem of low efficiency of the method for determining the welding parameters used in welding the tabs of the battery cell.

[0084] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the hardware related to the terminal device through a program, and the program can be stored in a non-volatile storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0085] The embodiment of the present invention further provides a non-volatile storage medium. Optionally, in this embodiment, the non-volatile storage medium can be used to store the program code executed by the method for determining welding parameters provided in the above embodiment.

[0086] Optionally, in this embodiment, the non-volatile storage medium may be located in any computer terminal in a computer terminal group in a computer network, or in any mobile terminal in a mobile terminal group.

[0087] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for executing the following steps: obtaining material data of the battery cells to be welded, wherein the material data characterizes the structure and performance of the battery cells to be welded; inputting the material data into a pre-trained neural network model, and the neural network model outputs welding parameters of the battery cells to be welded, wherein the neural network model is obtained by training an original neural network model through training samples, and the training samples include sample material data and sample welding parameters of the sample battery cells.

[0088] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for executing the following steps: inputting material data into a pre-trained neural network model, and having the neural network model output welding parameters of the battery cells to be welded, including: inputting material data into a convolutional network unit included in the neural network model, and having the convolutional network unit extract material features from the material data, wherein the material features characterize the material properties of the battery cells to be welded; inputting material features into a long short-term memory network unit included in the neural network model, and having the long short-term memory network unit output welding parameters.

[0089] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for executing the following steps: the neural network model is trained in the following manner: the training sample is input into the original neural network model, and the parameters of the original neural network model are adjusted according to a predetermined loss function; the neural network model is determined according to the parameters that make the loss function meet predetermined conditions.

[0090] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for executing the following steps: inputting the training sample into the original neural network model, and the original neural network model adjusts the parameters of the original neural network model according to a predetermined loss function, including: inputting the sample material data into the original neural network model, and the original neural network model determines the predicted welding parameters; the original neural network model determines the value of the loss function based on the predicted welding parameters and the sample welding parameters; and using the back propagation algorithm to adjust the parameters of the original neural network model according to the value of the loss function.

[0091] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: using a back-propagation algorithm to adjust the parameters of the original neural network model according to the value of the loss function, including: determining the gradient of the value of the loss function; using an adaptive moment estimation algorithm to adjust the parameters of the original neural network model according to the gradient of the value of the loss function.

[0092] Optionally, in this embodiment, the non-volatile storage medium is configured to store program codes for performing the following steps: the material data includes at least one of the following: the number of tab layers, the foil thickness and the weld mark depth, wherein the number of tab layers is the number of layers of metal foil included in the tab of the battery cell to be welded, the foil thickness is the thickness of the metal foil constituting the tab of the battery cell to be welded, and the weld mark depth is the depth of the weld mark that needs to be formed on the tab of the battery cell to be welded.

[0093] An embodiment of the present invention further provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the method for determining welding parameters in each embodiment of the present application.

[0094] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0095] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0096] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0097] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0098] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0099] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a non-volatile storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0100] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for determining welding parameters, characterized in that: include: Acquire material data of the battery core to be welded, wherein the material data characterizes the structure and performance of the battery core to be welded; The material data is input into a pre-trained neural network model, and the neural network model outputs the welding parameters of the battery cell to be welded, wherein the neural network model is obtained by training an original neural network model through training samples, and the training samples include sample material data and sample welding parameters of the sample battery cell.

2. The method according to claim 1, characterized in that The step of inputting the material data into a pre-trained neural network model, and outputting welding parameters of the battery core to be welded by the neural network model, comprises: Inputting the material data into a convolutional network unit included in the neural network model, and extracting material features from the material data by the convolutional network unit, wherein the material features characterize the material properties of the battery core to be welded; The material characteristics are input into the long short-term memory network unit included in the neural network model, and the welding parameters are output by the long short-term memory network unit.

3. The method according to claim 1, characterized in that The neural network model is trained in the following way: Inputting the training sample into the original neural network model, and adjusting the parameters of the original neural network model according to a predetermined loss function; The neural network model is determined according to the parameters that make the loss function meet the predetermined conditions.

4. The method according to claim 3, characterized in that The step of inputting the training sample into the original neural network model, wherein the original neural network model adjusts the parameters of the original neural network model according to a predetermined loss function, comprises: Inputting the sample material data into the original neural network model, and determining the predicted welding parameters by the original neural network model; The original neural network model determines the value of the loss function according to the predicted welding parameters and the sample welding parameters; A back propagation algorithm is used to adjust the parameters of the original neural network model according to the value of the loss function.

5. The method according to claim 4, characterized in that The back propagation algorithm is used to adjust the parameters of the original neural network model according to the value of the loss function, including: determining a gradient of a value of the loss function; An adaptive moment estimation algorithm is used to adjust the parameters of the original neural network model according to the gradient of the value of the loss function.

6. The method according to any one of claims 1 to 5, characterized in that The material data includes at least one of the following: the number of tab layers, the foil thickness and the weld mark depth, wherein the number of tab layers is the number of layers of metal foil included in the tab of the battery cell to be welded, the foil thickness is the thickness of the metal foil constituting the tab of the battery cell to be welded, and the weld mark depth is the depth of the weld mark that needs to be formed on the tab of the battery cell to be welded.

7. A device for determining welding parameters, characterized in that: include: An acquisition module, used to acquire material data of the battery core to be welded, wherein the material data represents the structure and performance of the battery core to be welded; A determination module is used to input the material data into a pre-trained neural network model, and the neural network model outputs the welding parameters of the battery cell to be welded, wherein the neural network model is obtained by training an original neural network model through training samples, and the training samples include sample material data and sample welding parameters of the sample battery cell.

8. A non-volatile storage medium, characterized in that: The non-volatile storage medium includes a stored program, wherein when the program is executed, the device where the non-volatile storage medium is located is controlled to execute the method for determining welding parameters according to any one of claims 1 to 6.

9. A computer device, characterized in that: include: Memory and processor, The memory stores a computer program; The processor is used to execute the computer program stored in the memory, and when the computer program is run, the processor executes the method for determining welding parameters according to any one of claims 1 to 6.

10. A computer program product comprising computer instructions, characterized in that: The computer instructions are executed by a processor to implement the method for determining welding parameters as described in any one of claims 1 to 6.

Images