Production control method and system for main-grid-free battery piece

Abstract

The invention discloses a main-grid-free battery piece production control method and system, and relates to the technical field of battery piece production control. The production control method for the main-grid-free battery piece comprises the following steps: when each production stage of the main-grid-free battery piece is finished, acquiring corresponding image data and process parameters; based on a pre-trained artificial intelligence recognition model, carrying out recognition analysis on the image data, and generating a production feature set of each production stage; analyzing a production evaluation value of each production stage; comparing and analyzing with a preset evaluation interval of each production stage; if the production evaluation value is in the corresponding production evaluation interval, generating a first control instruction; and if the production evaluation value is outside the corresponding production evaluation interval, a corresponding control instruction is generated according to the corresponding process parameter, the quality monitoring closed-loop control is realized by adopting an image and process dual judgment mechanism, the abnormity identification accuracy and the response efficiency are improved, and the method is suitable for an intelligent precision production line.

Description

Technical Field

[0001] This invention relates to the field of solar cell production control technology, specifically to a method and system for controlling the production of grid-free solar cells. Background Technology

[0002] With the rapid development of the photovoltaic industry, grid-free solar cells have gradually become a key structural form for high-efficiency solar modules due to their higher photoelectric conversion efficiency and better appearance consistency. In grid-free structures, the traditional metal grid is eliminated from the surface of the solar cell, and current collection is achieved through back electrodes or laser transfer processes. This places higher demands on the geometric accuracy, surface uniformity, and welding quality of the front-end manufacturing processes. To ensure the quality of the finished product, it is usually necessary to monitor the condition and assess the quality of the solar cells at each critical process stage (such as laser scribing, laser welding, and cell stacking).

[0003] The limitations of existing technologies include at least the following problems: the current production process of busbarless solar cells lacks a mechanism to combine image detection results with process parameters in real time to form quantifiable production evaluation values. This results in the inability to establish unified and continuous quality judgment standards and cross-stage connection judgments between different production stages. Although image detection exists, it remains at the level of abstract recognition or manual experience judgment. Although process parameters are collected, they are only used for post-event recording. There is no fusion analysis path between the two, making it difficult to form dynamic, objective, and executable quality judgment criteria during the production process. As a result, solar cells with abnormal quality are often mistakenly sent to the next process, increasing rework rate, reducing production line yield, and extending the overall production cycle. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method and system for controlling the production of grid-free solar cells, which solves the problems of disconnect between image recognition results and process parameters, and the lack of unified production evaluation values ​​that lead to non-real-time and inaccurate production decisions in existing technologies.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for controlling the production of grid-free solar cells, comprising the following steps: At the end of each production stage of the gridless solar cell, the corresponding image data and process parameters are acquired; the image data consists of several solar cell pixels, and each solar cell pixel corresponds to a solar cell pixel value. Based on a pre-trained artificial intelligence recognition model, the image data is identified and analyzed to generate a set of production features for each production stage of the gridless solar cell. Based on the production feature set, analyze the production evaluation value of each production stage; The production evaluation value is compared and analyzed with the preset evaluation interval for each production stage; If the production evaluation value is within the corresponding production evaluation interval, a first control command is generated; If the production evaluation value is outside the corresponding production evaluation range, then a corresponding control command is generated based on the corresponding process parameters.

[0006] Furthermore, the artificial intelligence recognition model includes an input layer, a convolutional layer, a pooling layer, a flattening layer, a fully connected layer, and an output layer; the specific steps for generating the production feature set are as follows: In the input layer, the image data is preprocessed to obtain standardized image data; In the convolutional layer, the standardized image data is convolved to generate a convolutional feature map; In the pooling layer, a pooling operation is performed on the convolutional feature map to generate a dimension-reduced pooled feature map; In the flattening layer, the dimensionality-reduced pooling feature map is flattened into one-dimensional data to form a one-dimensional feature vector; In the fully connected layer, the one-dimensional feature vector is weighted to obtain a comprehensive feature. In the output layer, the synthesized features are decoded to generate the production feature set.

[0007] Furthermore, the production feature set includes edge flatness, surface smoothness, dimensional deviation value, surface reflectivity, and surface gloss.

[0008] Furthermore, the specific steps for analyzing the production evaluation values ​​of each production stage are as follows: Read the edge flatness, surface smoothness, and dimensional deviation values ​​for each of the production stages, and perform standardization processing; The standardized edge flatness, surface smoothness, and dimensional deviation values ​​are combined with the surface reflectance and surface gloss of the corresponding production stages for comprehensive analysis to obtain the production evaluation value for each production stage.

[0009] Furthermore, the formula for calculating the production evaluation value is as follows:

[0010] in, , , , , , The following are, in order, the production evaluation values ​​for a certain production stage of a gridless solar cell: edge smoothness, surface smoothness, dimensional deviation, surface reflectivity, and surface gloss. , , , , The coefficients are, in order, the edge smoothness adjustment coefficient, surface smoothness adjustment coefficient, dimensional deviation adjustment coefficient, surface reflection adjustment coefficient, and surface gloss adjustment coefficient for a certain production stage of a gridless solar cell stored in the database.

[0011] Furthermore, the specific steps for generating corresponding control commands based on the corresponding process parameters are as follows: Based on the aforementioned process parameters, analyze the process evaluation values ​​for the corresponding production stage; The process evaluation values ​​are matched and analyzed with the preset process evaluation intervals for the corresponding production stages. If the process evaluation value is within the corresponding process evaluation range, a second control command is generated; If the process evaluation value is outside the corresponding process evaluation range, an alarm command is generated.

[0012] Furthermore, the process parameters include processing temperature, applied pressure, processing speed, path offset, and micro-vibration amplitude values ​​at multiple time points corresponding to the production stage; The specific steps for analyzing the process evaluation values ​​for the corresponding production stage are as follows: Based on the processing temperature value, the applied pressure value, the processing speed value, the path offset value, and the micro-vibration amplitude value, the process feature set corresponding to the production stage is analyzed; Based on the aforementioned process feature set, the process evaluation value corresponding to the production stage is analyzed.

[0013] Furthermore, the process feature set includes processing temperature fluctuation value, application pressure variation range value, processing speed offset value, path offset fluctuation value, and micro-vibration mean square value.

[0014] Furthermore, the formula for calculating the process evaluation value is as follows:

[0015] in, , , , , , The values, in order, represent the process evaluation value, processing temperature fluctuation value, range of applied pressure variation value, processing speed deviation value, path deviation fluctuation value, and mean square value of micro-vibration at a certain production stage of a gridless solar cell. , , , , The following are, in order, the temperature control coefficient, pressure control coefficient, speed control coefficient, path control coefficient, and vibration control coefficient for a certain production stage of gridless solar cells stored in the database.

[0016] A busbarless solar cell production control system includes: The data acquisition unit is used to acquire corresponding image data and process parameters at the end of each production stage of the gridless solar cell. The image feature recognition unit is used to identify and analyze the image data based on a pre-trained artificial intelligence recognition model to generate a production feature set for each production stage of the gridless solar cell. The production evaluation and analysis unit is used to analyze the production evaluation value of each production stage based on the production feature set. The evaluation interval determination unit is used to compare and analyze the production evaluation value with the preset evaluation interval for each production stage. A phased advancement control unit is used to generate a first control command when the production evaluation value is within the corresponding production evaluation interval; The process feedback control unit is used to generate corresponding control commands based on the corresponding process parameters when the production evaluation value is outside the corresponding production evaluation range.

[0017] The present invention has the following beneficial effects: (1) The production control method for gridless solar cells combines the image analysis results of each production stage with the preset feature adjustment coefficients of each stage to form the production evaluation value of each production stage. On this basis, if the evaluation value deviates from the set range, the process evaluation value will be further generated by combining the process parameters of each time sequence, and the control command will be determined accordingly. This dual-layer judgment mechanism enables the system to achieve static quality control through image analysis and capture potential hidden dangers by combining real-time process fluctuations. The logic of control command generation is clear. It can execute equipment shutdown and parameter adjustment, and can also output rework mark, thereby effectively improving the response sensitivity and abnormal identification accuracy in the production process. It is particularly suitable for intelligent production lines that independently track and manage different production links.

[0018] (2) The production control method for grid-free solar cells acquires corresponding image data at the end of each production stage of the grid-free solar cells, and performs multi-level structure analysis through an artificial intelligence recognition model to output the production feature set of each production stage. Since the model has been pre-trained and standardized learning of the image features of each production stage, it can still maintain a high recognition accuracy even in cases of image resolution differences or complex backgrounds. Compared with the traditional method of static quality judgment based solely on edge detection or threshold segmentation, this invention can achieve stable feature extraction at different stages and ensure the traceability of various quality data, enabling the production line system to dynamically respond according to numerical changes, thereby improving the consistency and responsiveness of the overall control cycle.

[0019] (3) The busbarless solar cell production control method extracts the time series fluctuation information of the process parameters in each production stage into a process feature set, including the processing temperature fluctuation value, the range of the applied pressure change, the processing speed offset value, the path offset fluctuation value and the mean square value of micro-vibration, and further generates a process evaluation value based on this, which serves as an important basis for determining whether to enter the next stage or to shut down the equipment. This method makes up for the problem that static image features cannot reflect real-time disturbances. By performing hierarchical fusion judgment on these two types of evaluation values, the production line can not only accurately control the production rhythm based on image quality, but also identify potential drift trends with the help of process dynamic feedback, thereby improving the linkage control capability from image to process.

[0020] (4) This busbarless solar cell production control system, by constructing a complete busbarless solar cell production control system, including data acquisition, image feature recognition, production evaluation analysis, evaluation interval determination, stage advancement control, and process feedback control unit, forms a modular architecture design with a clear structure and well-defined division of labor. The modules use a unified interface protocol to realize information transmission, which has good hardware and software decoupling, making it easy for the system to be flexibly deployed and independently upgraded on different production lines or equipment. Among them, the image feature recognition unit and the evaluation analysis unit rely on the artificial intelligence model running platform and can be directly integrated with edge computing devices, while the control unit can be compatible with mainstream PLC systems or motion control systems for implementation. Through this structured division method, it is convenient to deploy in a distributed manner according to stages and to upgrade and maintain remotely, which significantly improves the engineering feasibility and subsequent maintenance cost control capabilities of the system, and meets the needs of modern production lines for flexible adaptation and rapid integration of intelligent production systems.

[0021] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0022] Figure 1 This is a flowchart of a method for controlling the production of gridless solar cells according to the present invention.

[0023] Figure 2 This is a flowchart illustrating the specific steps involved in analyzing production evaluation values ​​at each production stage in a method for controlling the production of grid-free solar cells according to the present invention.

[0024] Figure 3 This is a block diagram of a gridless solar cell production control system according to the present invention.

[0025] Figure 4 This is a block diagram of an exemplary electronic device according to an embodiment of the present invention. Detailed Implementation

[0026] Please see Figure 1 This invention provides a technical solution: a method for controlling the production of busbarless solar cells, comprising the following steps: at the end of each production stage of the busbarless solar cell, acquiring corresponding image data and process parameters, wherein the generation stage includes a laser scribing stage, a laser welding stage, and a stacking stage (pressing stage); based on a pre-trained artificial intelligence recognition model, recognizing and analyzing the image data to generate a production feature set for each production stage of the busbarless solar cell; analyzing the production evaluation value of each production stage according to the production feature set; comparing the production evaluation value with a preset evaluation interval for each production stage; if the production evaluation value is within the corresponding production evaluation interval, generating a first control command, i.e., driving the busbarless solar cell to enter the next production stage; if the production evaluation value is outside the corresponding production evaluation interval, generating a corresponding control command according to the corresponding process parameters.

[0027] Specifically, the image data consists of several battery cell pixels, and each battery cell pixel corresponds to a battery cell pixel value.

[0028] The image data is acquired through high-resolution optical imaging equipment or X-ray imaging technology.

[0029] The artificial intelligence recognition model includes an input layer, convolutional layer, pooling layer, flattening layer, fully connected layer, and output layer. The generated feature set includes edge flatness, surface smoothness, dimensional deviation value, surface reflectivity, and surface gloss.

[0030] The pre-training steps for an artificial intelligence recognition model are as follows: First, an image training sample library was constructed based on image data from each stage of the production of gridless solar cells. The image data originated from high-resolution visual acquisition equipment on the production line, with an image resolution of no less than 1024×1024 pixels. For each image, combined with manual measurement results, five image feature values ​​were labeled: edge smoothness, surface smoothness, dimensional deviation, surface reflectivity, and surface gloss. The standard deviation of the rate of change of edge curvature is specified for edge smoothness, in millimeters. Surface smoothness is the percentage of high-frequency regions in an image. The size deviation is the difference between the calculated size of the image and the standard size, in millimeters. Surface reflectance is the ratio of the average gray value of an image to the standard gray value; Surface gloss is the ratio of the brightness of the highlight area to the brightness of the entire image.

[0031] Secondly, the image samples undergo uniform preprocessing, including size normalization, grayscale standardization, and image enhancement (such as flipping and adding noise). The processed images serve as the model input, and the image feature values ​​serve as the target output for model training.

[0032] Next, an artificial intelligence recognition model is constructed, whose structure includes an input layer, a convolutional layer, a pooling layer, a flattening layer, a fully connected layer, and an output layer. The model input is the processed image data, and the output is the prediction results of the above five image feature values.

[0033] The output layer employs a regression decoding structure, mapping the comprehensive feature vector generated by the fully connected layer to five specific feature values. Each output neuron corresponds to an image feature, and its value represents the prediction result for that feature.

[0034] During training, manually labeled feature values ​​are used as supervision signals, and minimum mean square error is used as the loss function. Multiple rounds of iterative training are performed until the prediction error converges.

[0035] The specific steps for generating production feature sets for each production stage of busbarless solar cells are as follows: In the input layer of the artificial intelligence recognition model, the image data at the end of each production stage of the busbarless solar cell are preprocessed to obtain standardized image data for each production stage, specifically: Convert the original image to grayscale to remove color interference; Median filtering is applied to grayscale images to eliminate salt-and-pepper noise. Scale and crop the image to a fixed resolution (e.g., 224×224 pixels) to ensure that the input dimensions of subsequent network structures are consistent. In the convolutional layer of the artificial intelligence recognition model, convolutional operations are performed on standardized image data to generate convolutional feature maps for each production stage, specifically as follows: Multiple 3×3 and 5×5 convolutional kernels are applied to extract detailed features (such as texture) and macroscopic structures (such as boundary shapes), respectively. Each convolutional kernel slides across the image to form a feature response map, and the non-linear expressive power is enhanced by the ReLU activation function. In the pooling layer of the artificial intelligence recognition model, pooling operations are performed on the convolutional feature maps to generate dimension-reduced pooled feature maps for each production stage, specifically as follows: Max pooling (e.g., 2×2 window) is applied to extract the most representative edge and texture features from the image; Reduce the dimension of the feature map, reduce computational cost, and retain key response regions; In the flattening layer of the artificial intelligence recognition model, the dimensionality-reduced pooling feature map is flattened into one-dimensional data, forming one-dimensional feature vectors for each production stage, specifically as follows: Each feature map is unfolded into a vector; All feature vectors are concatenated into a complete set of image description vectors for subsequent comprehensive feature learning; In the fully connected layer of the artificial intelligence recognition model, the one-dimensional feature vector is weighted to obtain the comprehensive features of each production stage, specifically: A fully connected network structure is introduced to simulate the global combination relationship between features; Nonlinear enhancement is achieved using the ReLU activation function; Overfitting is prevented and generalization ability is improved by using Dropout layers; In the output layer of the artificial intelligence recognition model, the comprehensive features are decoded to generate production feature sets for each production stage of the busbarless solar cell, specifically as follows: Regarding edge smoothness: The output layer receives the edge structure feature vector generated by the fully connected layer; The curvature statistics in the edge convolution response map are matched and fitted with a preset linear regression function. Input multiple parameters, such as edge offset amplitude and local fluctuation interval length, into a linear combination function; The final decoding yields the edge flatness (unit: mm), which represents the average deviation between the edge fitting line and the true contour.

[0036] Regarding surface smoothness: The output layer receives the high-frequency feature vector of the texture from the fully connected layer; The high-frequency energy values ​​and their proportions are extracted from the feature map using the frequency domain statistics module. Combined with the statistical variance of the texture distribution, it is decoded into surface smoothness (unit: %) through feature weighting. This value represents the proportion of high-frequency textured areas per unit area, reflecting the surface roughness.

[0037] For dimensional deviation values: The output layer utilizes the geometric feature vectors from the fully connected layer; Perform inverse pixel-to-physical mapping decoding on the spacing between edge key points in the image; And based on the calibration coefficients, the actual physical dimensions are restored and compared with the standard dimensions; Output dimensional deviation value (unit: mm), which represents the size of the product size relative to the standard size.

[0038] For surface reflectivity: The output layer extracts the brightness energy center, gray-level mean, and highlight distribution range from the gray-level distribution-related feature vectors; Then, the grayscale response model of the standard reflectance sample is normalized and compared. The degree of reflectivity shift is calculated by fitting a function; Output surface reflectance (unit: gray level or %) represents the deviation of the actual reflectivity from the standard image.

[0039] Regarding surface gloss: The output layer combines the brightness gradient feature vector with the highlight region distribution map; Decoding is performed using a function that compares the brightness of the point with the brightness of the surrounding background. Output surface gloss (unit: ratio, no unit); The higher the value, the brighter the surface; if it is lower than the set value, the surface is darker or contaminated.

[0040] In this implementation scheme, the artificial intelligence model does not output abstract classification labels or confidence scores, but directly generates image feature values, namely edge smoothness, surface smoothness, dimensional deviation, surface reflectivity, and surface gloss. The greatest advantage of this design approach is that the output results can be directly used for production judgment, process correction, and equipment control. Furthermore, each feature value has a clear unit and dimension, which can be used for horizontal comparison, trend analysis, or the establishment of evaluation models, resulting in high data utilization. Compared to traditional models that output binary or multi-class judgments, this feature value output structure significantly improves the feasibility and data value density of the model results, providing a standardized and calculable basis for image-driven quality control, and meeting the automation needs of actual production lines.

[0041] Specifically, such as Figure 2 As shown, the specific steps for analyzing the production evaluation values ​​of each production stage are as follows: Read the edge flatness, surface smoothness, and dimensional deviation values ​​of each production stage of the busbarless solar cell, and perform standardization processing (i.e., unit removal); Combine the standardized edge flatness, surface smoothness, and dimensional deviation values ​​of each production stage of the busbarless solar cell with the surface reflectivity and surface gloss of the corresponding production stage for comprehensive analysis to obtain the production evaluation value of each production stage.

[0042] The specific formula for calculating the production evaluation value of a busbarless solar cell at a certain production stage is as follows: ;in, , , , , , The following are, in order, the production evaluation values ​​for a certain production stage of a gridless solar cell: edge smoothness, surface smoothness, dimensional deviation, surface reflectivity, and surface gloss. , , , , The coefficients are, in order, the edge smoothness adjustment coefficient, surface smoothness adjustment coefficient, dimensional deviation adjustment coefficient, surface reflection adjustment coefficient, and surface gloss adjustment coefficient for a certain production stage of a gridless solar cell stored in the database.

[0043] In the laser scribing stage, since this stage mainly determines the initial outline and cutting boundary of the solar cell, the edge flatness and dimensional deviation values ​​are given a high weight. The edge flatness adjustment coefficient can be set to 0.40, and the dimensional deviation adjustment coefficient can be set to 0.30. The surface smoothness, reflectivity and gloss have a relatively small impact on this stage, and the corresponding coefficients can be set to 0.10, 0.10 and 0.10, respectively.

[0044] During the laser welding stage, the heat treatment of the welding area can easily cause changes in the local surface condition. Therefore, the attention to surface smoothness, reflectivity and gloss increases, and the corresponding adjustment coefficients can be set to 0.30, 0.25 and 0.15 respectively. However, the changes in edge and size are not significant, and the adjustment coefficients can be set to 0.20 and 0.10 respectively.

[0045] In the cell arrangement (pressing) stage, the main focus is on the stability of cell arrangement and pressing, surface texture and overall geometric consistency. Therefore, the surface smoothness adjustment coefficient and the size deviation adjustment coefficient can be set to 0.35, the edge flatness adjustment coefficient can be set to 0.15, and the surface reflection adjustment coefficient and the gloss adjustment coefficient can be set to 0.10 and 0.05, respectively.

[0046] In this implementation plan, different image features are considered at each production stage. By setting specific adjustment coefficients, edge flatness, surface smoothness, dimensional deviation, surface reflectivity, and surface gloss are precisely matched with different process steps, forming an adjustable and controllable evaluation mechanism. For example, in the laser scribing stage, more attention is paid to the shape stability of the cutting boundary, thus increasing the weight of edge flatness. In the welding stage, surface condition changes are the main risk point, so the proportion of smoothness and reflective features is increased. In the assembly stage, dimensional consistency and pressing effect are emphasized, and the proportion of smoothness and dimensional deviation is adjusted accordingly. Thus, the generation of evaluation values ​​is no longer a simple average or fixed formula, but can dynamically adjust the feature contribution according to the actual process focus, which facilitates refined quality control and helps the system quickly locate the source of problems.

[0047] Specifically, the steps for generating corresponding control commands based on the corresponding process parameters are as follows: Based on the process parameters, analyze the process evaluation value of the corresponding production stage and match it with the preset process evaluation range of the corresponding production stage; if the process evaluation value is within the corresponding process evaluation range, generate a second control command to guide the non-busbar solar cell to the rework area; if the process evaluation value is outside the corresponding process evaluation range, generate an alarm command and shut down the relevant equipment.

[0048] In this implementation plan, by matching and analyzing the process evaluation values ​​of each production stage with the corresponding process evaluation ranges, the system can quantitatively judge the state of the solar cells, avoiding subjective human intervention. If the evaluation value is within the range, it indicates that although there is a deviation, it can still be reworked, and the system generates a rework instruction. If it exceeds a reasonable range, an alarm is automatically triggered and the system is shut down to prevent the batch problem from spreading further. This design improves the response efficiency and stability of the production process, which is of practical significance for ensuring product quality and reducing resource waste. It also lays the foundation for subsequent data recording, anomaly tracing, and process optimization, making it suitable for widespread application on production lines with high standardization and automation requirements.

[0049] Specifically, the process parameters include processing temperature, applied pressure, processing speed, path offset, and micro-vibration amplitude values ​​at several times during the production stage.

[0050] The specific steps for analyzing the process evaluation value of the corresponding production stage are as follows: Based on the process parameters, analyze the process feature set of the corresponding production stage; based on the process feature set, analyze the process evaluation value of the corresponding production stage.

[0051] The specific formula for calculating the process evaluation value of a certain production stage is as follows: ;in, , , , , , The values, in order, represent the process evaluation value, processing temperature fluctuation value, range of applied pressure variation value, processing speed deviation value, path deviation fluctuation value, and mean square value of micro-vibration at a certain production stage of a gridless solar cell. , , , , The following are, in order, the temperature control coefficient, pressure control coefficient, speed control coefficient, path control coefficient, and vibration control coefficient for a certain production stage of gridless solar cells stored in the database.

[0052] When calculating the process evaluation value, the processing temperature fluctuation value, the range of the applied pressure change value, the processing speed offset value, the path offset fluctuation value, and the mean square value of micro vibration must first be deunited.

[0053] For the laser scribing stage, which is primarily responsible for cutting the initial boundaries of the solar cells, temperature control and path accuracy are highly dependent on these factors. Since laser power and scanning path directly affect the uniformity of the scribing and the quality of the cut edges, the temperature control adjustment coefficient can be set to 0.35, and the path adjustment coefficient to 0.30. Furthermore, micro-vibration also has a potential impact on cutting accuracy; the vibration adjustment coefficient can be set to 0.20. Pressure and speed are secondary influencing factors in this stage, with corresponding adjustment coefficients set to 0.10 and 0.05, respectively.

[0054] The laser welding stage, which involves laser heat input and weld point formation, is significantly sensitive to temperature stability and pressure. Large temperature fluctuations or improper pressure control during processing can easily lead to weld cracks or incomplete welds; therefore, the temperature control coefficient and pressure control coefficient are set to 0.30 and 0.30, respectively. Since precise welding path positioning is required, the path adjustment coefficient can be set to 0.20; the impact of micro-vibration is secondary, so it is set to 0.10; while speed has a relatively low impact on weld formation, and its corresponding adjustment coefficient can be set to 0.10.

[0055] For the wafer arrangement (pressing) stage, the core issues are the precision of wafer arrangement and the overall uniformity of pressing. Micro-vibration may affect pressing stability, so the vibration adjustment coefficient can be set to 0.30. Path consistency determines the neatness of wafer arrangement, so the path adjustment coefficient is set to 0.25. Pressure control is particularly critical for pressing tightness, so the corresponding adjustment coefficient is set to 0.20. Speed ​​deviation may affect the arrangement sequence, so the adjustment coefficient is set to 0.15. Temperature changes have the least effect at this stage, so the temperature control adjustment coefficient is set to 0.10.

[0056] The process feature set includes processing temperature fluctuation value, application pressure variation range value, processing speed offset value, path offset fluctuation value, and micro-vibration mean square value.

[0057] Specifically, by reading the processing temperature values ​​of the corresponding time series, a temperature change trajectory is constructed for each production stage. After summarizing the processing temperature values ​​at each time point, the standard deviation within the production stage is calculated, thus obtaining the processing temperature fluctuation value for that stage. The larger the fluctuation value, the more unstable the temperature control process is, potentially indicating a risk of lag in temperature control equipment adjustment or uneven thermal coupling. Conversely, a smaller fluctuation value indicates that the temperature control process is stable and reliable during that stage.

[0058] By reading the applied pressure values ​​at multiple time points, the mechanical loading state during the production stage is characterized. The difference between the maximum and minimum values ​​in this sequence is processed to obtain the range of applied pressure changes during the production stage, which is used to measure whether there are sudden changes or fluctuations in pressure output during this stage. When the range is too large, it may indicate that the pressure loader is over-responding or that the workpiece is not contacting evenly, affecting the welding or pressing quality.

[0059] For the processing speed values ​​during the production stage, the actual speed is recorded at each time point, and the system also has the target speed set for that stage as a reference. By calculating the point-by-point difference between the actual speed and the target speed, and taking the absolute value of the average, the processing speed offset value for that stage is obtained. The larger this value, the more systematic the deviation in equipment operating speed or the more lag in control, which may lead to a decrease in local processing quality, requiring adjustment or correction in subsequent processes.

[0060] For path offset values, the system records path offset data at each time point based on the error between the preset standard processing path and the current position of the equipment. The path offset sequence of the entire stage is statistically processed, and its standard deviation is calculated to obtain the path offset fluctuation value of that stage. The larger the value, the worse the path stability, and there is a risk of uncontrollable path offset or abnormal jitter amplitude, which may cause problems such as uneven dicing or welding misalignment.

[0061] To detect the amplitude of micro-vibrations during equipment operation, the system uses a built-in high-frequency accelerometer to collect micro-vibration data in real time. The system squares the amplitude values ​​at each time point, calculates the average, and then takes the square root (RMS value) to obtain the root mean square (RMS) value of the micro-vibration at that stage. This value reflects the dynamic stability of the equipment; a larger RMS value indicates more severe vibration, which may interfere with the stability of the welding points or the uniformity of the pressing process.

[0062] In this implementation plan, key process parameters at different production stages are quantified into a unified set of process characteristics. Adjustment coefficients are then set based on actual process requirements to enable targeted calculation of process evaluation values. Each characteristic value, such as temperature fluctuation, pressure difference, and speed deviation, is constructed based on time-series data, accurately reflecting equipment operating status and process execution stability. Furthermore, by setting adjustment weights for different stages, the evaluation results are more closely aligned with the core control points of each process, avoiding a one-size-fits-all approach. This structure balances universality with stage-specific differences, making it suitable for deployment in automated control systems for real-time monitoring, process early warning, and quality traceability, thereby improving overall process stability and responsiveness.

[0063] Please see Figure 3This invention provides a technical solution: a production control system for busbarless solar cells, comprising: a data acquisition unit, used to acquire corresponding image data and process parameters at the end of each production stage of the busbarless solar cell; an image feature recognition unit, used to identify and analyze the image data based on a pre-trained artificial intelligence recognition model to generate a production feature set for each production stage of the busbarless solar cell; a production evaluation and analysis unit, used to analyze the production evaluation value of each production stage based on the production feature set; an evaluation interval determination unit, used to compare and analyze the production evaluation value with a preset evaluation interval for each production stage; a stage advancement control unit, used to generate a first control command when the production evaluation value is within the corresponding production evaluation interval; and a process feedback control unit, used to generate a corresponding control command based on the corresponding process parameters when the production evaluation value is outside the corresponding production evaluation interval.

[0064] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended foregoing is intended to include both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0065] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from the spirit and scope of the invention. Therefore, if these modifications and variations fall within the scope of the foregoing and equivalent technologies of this invention, this invention also intends to include these modifications and variations. It should be noted that, for the foregoing method embodiments, for the sake of simplicity, they are all described as a series of actions. However, those skilled in the art should understand that this disclosure is not limited to the described order of actions, because according to this disclosure, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to this disclosure.

[0066] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the described module can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0067] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0068] Figure 4A block diagram of an exemplary electronic device capable of implementing embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0069] Device 300 includes a computing unit 301, which can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) 302 or a computer program loaded from storage unit 308 into random access memory (RAM) 303. RAM 303 may also store various programs and data required for the operation of device 300. The computing unit 301, ROM 302, and RAM 303 are interconnected via bus 304. Input / output (I / O) interface 305 is also connected to bus 304.

[0070] Multiple components in device 300 are connected to I / O interface 305, including: input unit 306, such as keyboard, mouse, etc.; output unit 307, such as various types of monitors, speakers, etc.; storage unit 308, such as disk, optical disk, etc.; and communication unit 309, such as network card, modem, wireless transceiver, etc. Communication unit 309 allows device 300 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0071] The computing unit 301 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 301 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 301 performs the various methods and processes described above, such as method 100. For example, in some embodiments, method 100 may be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 308. In some embodiments, part or all of the computer program may be loaded and / or installed on device 300 via ROM 302 and / or communication unit 309. When the computer program is loaded into RAM 303 and executed by the computing unit 301, one or more steps of method 100 described above may be performed.

[0072] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0073] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0074] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0075] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0076] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0077] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0078] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this disclosure can be achieved, and this is not limited herein.

[0079] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for controlling the production of grid-free solar cells, characterized in that, Includes the following steps: At the end of each production stage of the gridless solar cell, the corresponding image data and process parameters are acquired; the image data consists of several solar cell pixels, and each solar cell pixel corresponds to a solar cell pixel value. Based on a pre-trained artificial intelligence recognition model, the image data is identified and analyzed to generate a set of production features for each production stage of the gridless solar cell. Based on the production feature set, analyze the production evaluation value of each production stage; The production evaluation value is compared and analyzed with the preset evaluation interval for each production stage; If the production evaluation value is within the corresponding production evaluation interval, a first control command is generated; If the production evaluation value is outside the corresponding production evaluation range, then a corresponding control command is generated based on the corresponding process parameters.

2. The method according to claim 1, characterized in that, The artificial intelligence recognition model includes an input layer, a convolutional layer, a pooling layer, a flattening layer, a fully connected layer, and an output layer; the specific steps for generating the production feature set are as follows: In the input layer, the image data is preprocessed to obtain standardized image data; In the convolutional layer, the standardized image data is convolved to generate a convolutional feature map; In the pooling layer, a pooling operation is performed on the convolutional feature map to generate a dimension-reduced pooled feature map; In the flattening layer, the dimensionality-reduced pooling feature map is flattened into one-dimensional data to form a one-dimensional feature vector; In the fully connected layer, the one-dimensional feature vector is weighted to obtain a comprehensive feature. In the output layer, the synthesized features are decoded to generate the production feature set.

3. The method according to claim 2, characterized in that, The production feature set includes edge flatness, surface smoothness, dimensional deviation, surface reflectivity, and surface gloss.

4. The method according to claim 3, characterized in that, The specific steps for analyzing the production evaluation values ​​of each production stage are as follows: Read the edge flatness, surface smoothness, and dimensional deviation values ​​for each of the production stages, and perform standardization processing; The standardized edge flatness, surface smoothness, and dimensional deviation values ​​are combined with the surface reflectance and surface gloss of the corresponding production stages for comprehensive analysis to obtain the production evaluation value for each production stage.

5. The method according to claim 4, characterized in that, The formula for calculating the production evaluation value is: in, , , , , , The following are, in order, the production evaluation values ​​for a certain production stage of a gridless solar cell: edge smoothness, surface smoothness, dimensional deviation, surface reflectivity, and surface gloss. , , , , The coefficients are, in order, the edge smoothness adjustment coefficient, surface smoothness adjustment coefficient, dimensional deviation adjustment coefficient, surface reflection adjustment coefficient, and surface gloss adjustment coefficient for a certain production stage of a gridless solar cell stored in the database.

6. The method according to claim 1, characterized in that, The specific steps for generating corresponding control commands based on the aforementioned process parameters are as follows: Based on the aforementioned process parameters, analyze the process evaluation values ​​for the corresponding production stage; The process evaluation values ​​are matched and analyzed with the preset process evaluation intervals for the corresponding production stages. If the process evaluation value is within the corresponding process evaluation range, a second control command is generated; If the process evaluation value is outside the corresponding process evaluation range, an alarm command is generated.

7. The method according to claim 6, characterized in that, The process parameters include processing temperature, applied pressure, processing speed, path offset, and micro-vibration amplitude values ​​at multiple time points in the corresponding production stage. The specific steps for analyzing the process evaluation values ​​for the corresponding production stage are as follows: Based on the processing temperature value, the applied pressure value, the processing speed value, the path offset value, and the micro-vibration amplitude value, the process feature set corresponding to the production stage is analyzed; Based on the aforementioned process feature set, the process evaluation value corresponding to the production stage is analyzed.

8. The method according to claim 7, characterized in that, The process feature set includes processing temperature fluctuation value, application pressure variation range value, processing speed offset value, path offset fluctuation value, and micro-vibration mean square value.

9. The method according to claim 8, characterized in that, The formula for calculating the process evaluation value is: in, , , , , , The values, in order, represent the process evaluation values, processing temperature fluctuation values, range of applied pressure changes, processing speed deviation values, path deviation fluctuation values, and mean square values ​​of micro-vibration at a certain production stage of a gridless solar cell. , , , , The following are, in order, the temperature control coefficient, pressure control coefficient, speed control coefficient, path control coefficient, and vibration control coefficient for a certain production stage of gridless solar cells stored in the database.

10. A production control system for grid-free solar cells, characterized in that, The production control system, applied to the method for controlling the production of grid-less solar cells according to any one of claims 1-9, comprises: The data acquisition unit is used to acquire corresponding image data and process parameters at the end of each production stage of the gridless solar cell. The image feature recognition unit is used to identify and analyze the image data based on a pre-trained artificial intelligence recognition model to generate a production feature set for each production stage of the gridless solar cell. The production evaluation and analysis unit is used to analyze the production evaluation value of each production stage based on the production feature set. The evaluation interval determination unit is used to compare and analyze the production evaluation value with the preset evaluation interval for each production stage. A phased advancement control unit is used to generate a first control command when the production evaluation value is within the corresponding production evaluation interval; The process feedback control unit is used to generate corresponding control commands based on the corresponding process parameters when the production evaluation value is outside the corresponding production evaluation range.

Images