Solar cell, welding system and welding method

Abstract

The invention provides a solar cell, a welding system and a welding method of the solar cell. The welding method comprises the following steps that firstly, a plasma cleaning method and a chemical etching method are combined to conduct pretreatment on an electrode area of the cell; 2, accurately aligning the welding strip to a preset position of an electrode area, and adjusting the position of the welding strip in advance through a dynamic supplement mechanism; thirdly, metallurgical bonding of the welding strip and the electrode area of the battery piece is completed within the set temperature range through the pulse current welding technology; fourthly, the welding state is fed back in real time through multiple sensors, and welding parameters are dynamically adjusted based on the welding state; and 5, after the battery piece is welded, inputting multi-modal data of the battery piece by utilizing a pre-trained defect classification model, and outputting a defect detection result. According to the method, the offset position of the battery piece is pre-judged in advance through a dynamic supplement mechanism, dynamic compensation is carried out on the welding strip, the positioning precision is improved, and the offset is reduced.

Description

Technical Field

[0001] The present invention relates to the field of solar cell welding, and in particular to a solar cell, a welding system and a welding method. Background Art

[0002] In the traditional crystalline silicon photovoltaic module manufacturing process, the first step is to weld ribbons to the cells and connect them in series to form a cell string. Specifically, the method of forming a cell string includes: alternating the ribbons and cells in sequence, with the length of the ribbon approximately twice the width of the cell, so that half of the ribbon is on the front of one cell and the other half is on the back of the adjacent cell. In this way, a cell string containing multiple cells with the upper and lower surfaces (i.e., positive and negative electrodes) connected in series is formed. The cell string is then transported to the welding station via an adsorption track. Generally, multiple pressing wires or pressing blocks are used to compact the upper and lower ribbons to the cell. The ribbons are then welded to the surface electrodes of the cell through infrared heating.

[0003] In the field of photovoltaic cell manufacturing, the welding process is a core link that affects the efficiency and reliability of the components. Traditional processes have significant bottlenecks in electrode pretreatment, welding ribbon alignment, welding thermal management, dynamic parameter control, and defect detection. For example, in the existing technology, surface contaminants are not thoroughly removed during the pretreatment of the cell, resulting in low welding strength. When aligning the welding ribbon with the cell, the problem of deviation caused by the thermal expansion of the cell is ignored, resulting in offset welding position. In addition, the welding parameters are solidified during the welding process, and the quality cannot be guaranteed in the event of an emergency. Summary of the Invention

[0004] In response to the technical deficiencies in the background art, the present invention proposes a solar cell, a welding system, and a welding method, which solve the above technical problems and meet practical needs. The specific technical solutions are as follows:

[0005] A solar cell welding method comprises the following steps:

[0006] Step 1: Pre-treating the electrode area of the cell by combining plasma cleaning and chemical etching;

[0007] Step 2: Accurately align the soldering ribbon to the pre-set position of the electrode area, and adjust the position of the soldering ribbon in advance through the dynamic replenishment mechanism;

[0008] Step 3: Complete the metallurgical bonding of the welding ribbon and the battery cell electrode area within a set temperature range through a pulse current welding process;

[0009] Step 4: Real-time feedback of welding status is obtained through multiple sensors, and welding parameters are dynamically adjusted based on the welding status;

[0010] Step 5: After the battery cell is welded, the pre-trained defect classification model is used to input the multimodal data of the battery cell and output the defect detection results.

[0011] Furthermore, step one also includes:

[0012] The electrode surface is bombarded by high-energy plasma, and the high-energy ions physically strip off the organic matter on the electrode surface, and the hydrogen in the high-energy plasma reduces some of the metal oxides on the electrode surface;

[0013] Prepare an etching solution, which is prepared by adding 0.5%-3% dilute hydrochloric acid and 0.01% benzotriazole. Immerse the electrode surface after plasma cleaning in the etching solution for a period of time, and use ultrasound to assist the immersion process;

[0014] After impregnation, the electrode surface was rinsed with deionized water and then dried with nitrogen.

[0015] Furthermore, step 2 also includes:

[0016] Obtaining feature data of the geometric center lines of the welding strip grid lines and the electrode main grid, and using an edge detection algorithm to make the welding strip grid lines coincide with the center lines of the electrode main grid;

[0017] An offset prediction model is established based on the historical data of thermal expansion of the cell. The data of the welding ribbon and the electrode busbar are input to obtain compensation data. Based on the compensation data, the robot arm is controlled to deflect at the pre-calculated position.

[0018] When acquiring the characteristic data of the geometric center line of the welding strip and the electrode main grid, edge features are extracted by a multi-light source fusion imaging method, and the multi-light source fusion imaging method includes:

[0019] 0° orthophoto obtains clear main grid and blurred secondary grid images;

[0020] Oblique illumination at 30° allows for clear secondary grid and low-contrast main grid images;

[0021] Oblique illumination at 60° obtains an image with bright edges of the welding strip and dark background;

[0022] The three-angle images are synthesized into HDR images to extract edge features.

[0023] Furthermore, the offset prediction model is:

[0024]

[0025] Among them, T(x,y) is the coordinate position after deformation, which means the new position of the original point (x,y) after deformation; a0,a x ,a y is the coefficient of the linear part, describing the rigid deformation such as overall translation, rotation and scaling; ωi is the weight coefficient, which indicates the contribution of each control point to the local deformation; U(r) represents the radial basis function; ||(x,y)-(x i ,y i )|| is the Euclidean distance, which means the distance between the current point (x, y) and the control point (x i ,y i ) distance; n is the number of control points, usually the intersection points of the secondary grid with obvious features or artificial marking points are selected.

[0026] Furthermore, step three also includes:

[0027] A high current pulse is input to heat the pre-treated electrode surface. This is the breakdown stage, and the resistance change of the battery cell is monitored in real time through several resistance probes set around the battery cell.

[0028] When the resistance probe detects that the resistance fluctuation of the battery cell reaches the preset range, the high current pulse input is stopped and the medium current continuous pulse input is started to heat the electrode surface. This is the metallurgical stage.

[0029] After the preset conditions are met in the metallurgical stage, the tempering stage begins. At this time, low current pulses are used to heat the electrode surface to maintain the electrode surface temperature. After heating for a preset time, the heating ends and the cooling stage begins. The preset conditions include temperature, heating time, energy accumulation, and the morphology of the weld spot on the electrode surface.

[0030] During the metallurgical stage, when the resistance probe detects that the resistance fluctuation is greater than the preset range, the pulse current is increased and the heating time is extended.

[0031] Furthermore, the step 4 includes:

[0032] The temperature, solder joint morphology and local resistivity of the solder joints on the electrode surface are collected in real time using an infrared thermal imager, a high-speed camera and a resistance probe. The timestamps of the temperature, solder joint morphology and resistivity are synchronized to achieve data synchronization among multiple sensors.

[0033] The defect recognition model is trained based on historical data of solder joint morphology, and the data of solder joint morphology is imported into the defect recognition model to obtain the defect type and defect probability;

[0034] A dual-threshold trigger mechanism is established. When the resistivity change rate reaches a preset value or the defect probability of the solder joint morphology is greater than a preset value, the trigger parameters are adjusted, and the priority of the solder joint morphology is higher than the priority of the resistivity and temperature.

[0035] Furthermore, the step 4 further includes:

[0036] The temperature of the electrode surface is monitored by an infrared thermal imager. During the metallurgical stage, when the temperature exceeds a first preset temperature, the metallurgical stage is terminated and the tempering stage is entered; when the temperature is lower than a second preset temperature, the pulse current is increased and the pulse interval is shortened;

[0037] The weld morphology on the electrode surface is monitored in real time by a high-speed camera. During the metallurgical stage, when the weld morphology captured by the high-speed camera passes the morphology detection, the tempering stage is started; during the tempering stage, when the surface smoothness of the weld morphology captured by the high-speed camera reaches the preset value and there are no visible cracks or collapses, the tempering stage is ended and the cooling stage begins.

[0038] Furthermore, the step five includes:

[0039] The weld spot morphology of the electrode surface after welding is completed is obtained, and the weld spot morphology data is imported into the defect recognition model. When an unidentified defect is detected, the image is automatically captured and a training sample is generated to update the defect recognition model. When a defect is identified, a detection report containing the defect type and defect probability is generated.

[0040] A solar cell welding system is used to implement the above-mentioned solar cell welding method.

[0041] A solar cell is manufactured using the above-mentioned solar cell welding method.

[0042] Compared with the prior art, the solar cell, welding system and welding method provided by the present invention have the following beneficial effects:

[0043] The present invention realizes an ultra-clean interface through the synergistic effect of plasma cleaning and chemical etching, thoroughly removes pollutants on the electrode surface, improves welding bonding strength, predicts the position of battery cell offset in advance through a dynamic replenishment mechanism, dynamically compensates for the welding strip, improves positioning accuracy, and reduces offset. By real-time detection of the resistivity, morphological characteristics and temperature changes of the welding points during the welding process, multi-parameter fusion is used to realize intelligent control of welding parameters, thereby optimizing energy consumption while improving welding quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 The figure is a flow chart of a solar cell welding method in the present invention. DETAILED DESCRIPTION

[0045] In the description of the present invention, it should be understood that the terms "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "center," and "inner" are used to indicate positions or locations based on those shown in the accompanying drawings. These terms are intended solely to facilitate the description of the present invention and simplify the description. They are not intended to indicate or imply that the devices or components referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limiting the present invention. Furthermore, the terms "first," "second," and the like are used for descriptive purposes only and should not be construed to indicate or imply relative importance or implicitly specify the number of the technical features referred to. Thus, features defined as "first," "second," and the like may explicitly or implicitly include one or more of such features. In the description of the present invention, it should be noted that, unless otherwise expressly specified or limited, the terms "mounted," "connected," and "connected" should be understood broadly. For example, they can refer to fixed connections, removable connections, or integral connections; they can refer to direct connections, indirect connections through an intermediary, or internal connections between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0046] The following describes the implementation of the present invention in conjunction with the accompanying drawings and relevant embodiments. The implementation of the present invention is not limited to the following embodiments, and the present invention relates to the relevant necessary components in this technical field, which should be regarded as the common knowledge in this technical field and can be known and mastered by technical personnel in this technical field.

[0047] See Figure 1 , a solar cell welding method, comprising the following steps:

[0048] Step 1: Pre-treat the electrode area of the cell using a combination of plasma cleaning and chemical etching. High-energy ion bombardment removes surface organic contaminants, followed by hydrogen reduction of metal oxides. Then, an etchant dissolves the remaining oxides, plasma cleaning exposes the oxide layer, and chemical etching precisely removes the oxides. The total process takes less than 90 seconds. First, plasma cleaning is used to treat the electrode area. The cell is placed in a vacuum chamber of a plasma cleaning machine. A mixture of argon and hydrogen (4:1 by volume) is introduced. A high-frequency electric field ionizes the gases to form a plasma. This process takes between 5 and 30 seconds. The high-energy particles in the plasma physically sputter and chemically react with contaminants on the electrode surface, stripping them from the surface and expelling them from the chamber, thus cleaning the electrode surface. This process not only removes surface impurities but also creates a microscopic roughness on the electrode surface, increasing the surface area and improving wettability for subsequent soldering. Chemical etching is then performed. Based on the characteristics of the battery cell electrode material, a specific etching solution is prepared and the battery cell is immersed in the etching solution. The etching solution chemically reacts with the oxide layer on the electrode surface and part of the base material, dissolving the surface oxide layer and irregular protrusions, making the electrode surface smoother and more uniform. At the same time, it further activates the active sites on the electrode surface, creating favorable conditions for the metallurgical bonding of the solder and electrode during the subsequent welding process. The combination of plasma cleaning and chemical etching can significantly improve the cleanliness and activity of the electrode surface, reduce the contact resistance of the welding interface, and improve welding quality.

[0049] Step 2: Precisely align the solder ribbon to the pre-set position of the electrode area, adjusting its position in advance using a dynamic replenishment mechanism. A visual positioning system captures real-time images of the cell electrode area and solder ribbon, extracting the edge features of the ribbon and electrode. The system is equipped with multiple high-resolution CCD cameras, capturing clear images of the cell and solder ribbon from different angles. Image processing algorithms analyze the captured images to identify the position, shape, and size of the cell electrode area, as well as the position and posture of the solder ribbon. An offset prediction model predicts position offset within the next 50 ms, and a six-axis robotic arm performs feedforward compensation to reduce this offset.

[0050] Step 3: Metallurgically bond the solder ribbon to the cell electrode area using a pulsed current welding process within a set temperature range. This process achieves a metallurgical bond between the ribbon and the cell electrode area by applying a periodically varying pulsed current to the welding area within a set temperature range. The welding equipment utilizes a dedicated pulsed welding power supply capable of precisely controlling parameters such as the peak current, frequency, and pulse width. The pulsed current welding process heats the electrode area using three-stage pulse control. During welding, a closed-loop temperature control device monitors the weld zone temperature in real time and adjusts the pulse frequency accordingly to maintain a temperature between 180°C and 220°C. During welding, as the pulsed current passes through the contact area between the ribbon and the electrode, the resistive heat generated by the current rapidly heats and melts the solder, simultaneously causing diffusion and metallurgical reactions to occur on the electrode surface. Due to the cyclical variation of the pulsed current, significant heat is generated during the peak current phase, causing the solder to melt rapidly and form a good wetting and bonding with the electrode. During the valley current phase, the weld zone cools rapidly, promoting metallurgical reactions and forming a stable intermetallic compound layer. By precisely controlling the parameters of the pulse current and the welding temperature, the welding process is carried out within the set temperature range, which can not only ensure that the solder is fully melted and the metallurgical reaction is completed, but also avoid damage to the battery cell or degradation of the electrode material performance due to excessive temperature, thereby achieving high-quality welding connections and improving the electrical performance and mechanical strength of the battery cell.

[0051] Step 4: Real-time feedback of the welding status is obtained through multiple sensors, and the welding parameters are dynamically adjusted based on the welding status; the resistivity of the electrode area is measured in real time through the installed four-probe array, the installed high-speed camera is combined with the defect recognition model to monitor the morphology of the welding area of the electrode area, and the installed infrared thermal imager is used to detect the temperature gradient in real time. The welding parameters are adjusted in real time based on the resistivity, morphology monitoring data and temperature gradient, thereby adjusting the current size and frequency of the pulse current and the pressure applied to the welding area by the pressure control device.

[0052] Step 5. After the cell is welded, the pre-trained defect classification model is used to input the multimodal data of the cell and output the defect detection results. The defect classification model is trained by collecting 5,000 high-fidelity defect images and can output 12 types of defects and three levels of defect grades. The multimodal data includes the solder joint morphology, thermal images, and resistance distribution maps captured by the camera. The solder joint morphology includes the welding shape of the solder ribbon, the appearance of the solder joint, and the presence of defects such as cold solder joints, leaking solder joints, and solder accumulation. The defect classification model can automatically extract effective features from the multimodal data, perform comprehensive analysis and judgment, and output the defect detection results of the cell, accurately identifying various types of welding defects and their locations, providing a basis for subsequent quality control and product screening, and improving the production yield and product quality of solar cells.

[0053] Where necessary, step one also includes:

[0054] The electrode surface is bombarded by high-energy plasma, and the high-energy ions physically strip away the organic matter on the electrode surface. The hydrogen in the high-energy plasma reduces some of the metal oxides on the electrode surface. The plasma is composed of a mixture of argon and hydrogen. Argon acts as an inert carrier gas to provide high-energy ion bombardment, and hydrogen acts as a reducing gas to participate in the chemical reaction. The RF power frequency is 13.56 MHz, and the power density is 0.8-1.2 W / cm 2 , the vacuum degree of the cavity is maintained at 10 -2 pa, ensuring uniform and stable plasma. High-energy argon ions bombard the electrode surface with a kinetic energy of 500-1000eV, physically sputtering away organic matter and breaking CC bonds. The active hydrogen atoms in the hydrogen plasma react with metal oxides to generate metal elements and water, reducing the oxygen content on the electrode surface.

[0055] An etching solution is prepared by adding 0.5%-3% dilute hydrochloric acid and 0.01% benzotriazole. The electrode surface after plasma cleaning is immersed in the etching solution for a period of time, and ultrasound is used to assist the immersion process. When the concentration of dilute hydrochloric acid is less than 0.5%, it cannot effectively remove the oxide layer. When the concentration of dilute hydrochloric acid is greater than 3%, it will cause excessive loss of electrode thickness. Benzotriazole can form a monomolecular protective film on the silver surface, thereby inhibiting the corrosion of pure silver by hydrochloric acid.

[0056] After the impregnation is completed, the electrode surface is rinsed with deionized water and then blown dry with nitrogen. Ultrasonic assisted rinsing can be used during rinsing to dissolve the AgCI precipitate and avoid residual particles that cause cold solder joints. After the impregnation is completed, the electrode is immediately immersed in deionized water for a three-stage countercurrent rinse. The rinsing time for each stage is 1-3 minutes, and the water flow rate is controlled at 0.5-1m / s. The residual etching solution and dissolved products are completely removed by continuous replacement. After rinsing is completed, high-pressure nitrogen is used for drying. The nitrogen pressure is set to 0.3-0.6MPa, the air flow temperature is controlled at 40-60℃, and the electrode surface is blown in a laminar flow. This drying process can not only quickly evaporate water, but also use the inert environment of nitrogen to prevent the electrode surface from oxidizing again, ultimately obtaining a clean, dry and highly active welding surface, laying a solid foundation for subsequent welding processes.

[0057] It should be noted that step 2 also includes:

[0058] Obtaining feature data of the geometric center lines of the welding strip grid lines and the electrode main grid, and using an edge detection algorithm to make the welding strip grid lines coincide with the center lines of the electrode main grid;

[0059] An offset prediction model is established based on the historical data of thermal expansion of the cell. The data of the welding ribbon and the electrode busbar are input to obtain compensation data. Based on the compensation data, the robot arm is controlled to deflect at the pre-calculated position.

[0060] When acquiring the characteristic data of the geometric center line of the welding strip and the electrode main grid, edge features are extracted by a multi-light source fusion imaging method, and the multi-light source fusion imaging method includes:

[0061] 0° orthogonal illumination captures a clear busbar image and a blurred secondary grid image. Illuminating the cell surface perpendicularly, the light evenly covers the electrode area, resulting in a clear busbar outline in the image. However, the secondary grid, due to direct light, creates overlapping shadows and a blurred image. This setting highlights the busbar geometry and facilitates subsequent analysis of the busbar centerline.

[0062] 30° oblique illumination captures clear images of the secondary grille and low-contrast main grille. Projecting light at a 30° angle creates a distinct shadow contrast on the secondary grille, clearly revealing its structural details. However, the main grille's surface reflects unevenly due to the angle of the light, reducing image contrast. Images at this angle focus on capturing the edge features of the secondary grille, complementing the orthophoto image.

[0063] An image with a 60° oblique angle captures a brightly lit edge against a dark background. Increasing the incident angle to 60° reveals a brighter area at the edge of the ribbon due to reflected light, while the background appears dark due to the lack of direct light. This strong contrast between light and dark makes the outline of the ribbon stand out in the image, facilitating accurate identification of the ribbon edge.

[0064] The three-angle images are synthesized into an HDR image to extract edge features. After completing the image acquisition from the three angles, they are combined into a high dynamic range image. During the HDR synthesis process, an image fusion algorithm is used to integrate the advantageous features of the images from each angle, eliminating the problem of information loss in a single image. Subsequently, the HDR image is processed using the Canny edge detection algorithm to extract the clear edges of the solder strip grid and the electrode busbar. The geometric centerline of the two is then calculated using the least squares method, providing key data support for subsequent precise alignment.

[0065] It should be noted that the offset prediction model is:

[0066]

[0067] Among them, T(x,y) is the coordinate position after deformation, which means the new position of the original point (x,y) after deformation; a0,a x ,a y is the coefficient of the linear part, describing the rigid deformation such as overall translation, rotation and scaling; ω iis the weight coefficient, which indicates the contribution of each control point to the local deformation; U(r) represents the radial basis function; ||(x,y)-(x i ,y i )|| is the Euclidean distance, which means the distance between the current point (x, y) and the control point (x i ,y i ) distance; n is the number of control points, usually the sub-grid intersection points with obvious features or artificial marking points are selected to ensure that the model can accurately reflect the local deformation of the battery cell.

[0068] In one embodiment of the present invention, step three further includes:

[0069] A high current pulse is input to heat the pre-treated electrode surface. This is the breakdown stage. The resistance change of the cell is monitored in real time by several resistance probes set around the cell. The current value of the high current pulse is 150A, the pulse width is 0.5ms, and the probe configuration is a four-probe array distributed at the four corners of the cell. The resistance change is detected in real time. When the resistance fluctuation range is ≥0.3mΩ / cm 2 , indicating that the surface barrier layer has broken down, marking the end of the breakdown phase. A high current pulse is introduced through the contact point between the welding ribbon and the electrode, instantly generating a large amount of Joule heat, rapidly raising the electrode surface temperature to 300-500°C. Under this high temperature, the tiny oxide layer and contaminants remaining on the electrode surface are rapidly decomposed or vaporized. Simultaneously, the metal atoms on the electrode surface gain sufficient energy to become active, forming a highly active surface state, creating conditions for subsequent metallurgical reactions.

[0070] When the resistance probe detects that the cell's resistance fluctuations reach a preset range, the high-current pulse input ceases and a medium-current continuous pulse begins to heat the electrode surface. This enters the metallurgical stage. The medium-current pulses are 100A, 3ms wide, 200Hz frequent, and have a 60% duty cycle. The metallurgical stage is controlled by a combination of multiple parameters. A temperature sensor monitors the temperature changes in the weld area in real time to ensure that the temperature remains within a reasonable range. Energy accumulation is calculated by integrating the input current and time, reflecting the total energy input during the welding process. Simultaneously, a high-speed camera captures the solder joint morphology on the electrode surface in real time to observe solder wetting, spreading, and solder joint formation. When the temperature reaches the preset metallurgical temperature, the heating time meets the set value, the energy accumulation reaches the target value, and the solder joint morphology meets quality standards, the preset conditions are considered met and the system enters the tempering stage.

[0071] After the preset conditions are met during the metallurgical stage, the tempering stage begins. Low-current pulses are used to heat the electrode surface, maintaining the surface temperature. Heating is terminated after a preset heating time, and the cooling stage begins. Preset conditions include temperature, heating time, energy accumulation, and the weld morphology on the electrode surface. The tempering stage uses low-current pulses to heat the electrode surface. The current is set at 20A, the pulse frequency is reduced to 500Hz, and the pulse width is maintained at 1ms. The primary purpose of this stage is to eliminate internal stresses generated during welding through gentle heating, stabilize the crystal structure of the intermetallic compound, and further enhance the mechanical strength and electrical performance of the weld. The electrode surface temperature is maintained at 180±3°C. The heating time is set based on the size of the weld and the material properties, typically 300-500ms. After the preset heating time, the system stops heating and the weld area enters a natural cooling stage. Cooling is accelerated by forced air or water cooling to rapidly solidify the weld and complete the metallurgical bond between the ribbon and the cell electrode area.

[0072] During the metallurgical stage, when the resistance probe detects that the resistance fluctuation is greater than the preset range, the pulse current is increased and the heating time is extended.

[0073] It should be noted that the step 4 includes:

[0074] Using infrared thermal imagers, high-speed cameras, and resistance probes, the system collects real-time temperature, solder joint topography, and local resistivity data from the solder joints on the electrode surfaces. The timestamps for these temperatures, solder joint topography, and resistivity are synchronized to achieve multi-sensor data synchronization. A three-dimensional monitoring system is constructed using these cameras. The infrared thermal imager, an uncooled device with a wavelength range of 3-5μm, can monitor the electrode surface with a temperature resolution of 0.1°C, capturing 30-60 thermal images per second to capture real-time changes in the solder joint temperature field. The high-speed camera, with a frame rate of 5000fps and a microscope lens, clearly records the microscopic topography evolution of the solder joint during formation. The resistance probe, with a sampling frequency of 10-50kHz, acquires real-time local resistivity data from the electrodes. To achieve precise synchronization of heterogeneous data from multiple sources, the system utilizes a timestamp-based synchronization algorithm. At the data acquisition end, each sensor is equipped with a high-precision clock module to ensure that the timestamp error generated by each sensor's data is within ±0.2ms. After the collected data is transmitted to the central processing unit, a timestamp alignment algorithm chronologically matches the temperature, solder joint topography, and resistivity data to construct a welding status dataset containing spatiotemporal information. This synchronization mechanism enables the system to analyze the welding process from a multi-dimensional perspective. For example, by combining the distribution of high-temperature areas at a specific moment with changes in solder joint topography, it can determine whether local overheating is causing solder spatter.

[0075] A defect recognition model was trained based on historical data of solder joint topography. This data was imported into the defect recognition model to determine defect types and probabilities. The historical data consisted of 100,000 solder joint images with 12 defect labels, 5% of which were small sample defects. StyleGAN2 was then used to generate 5,000 defect images, covering variations in lighting, angle, and occlusion. During the model training phase, the raw image data was first preprocessed, including normalization, contrast enhancement, and noise removal, to improve data quality. The processed data was then divided into training, validation, and test sets. During training, model performance was optimized by adjusting hyperparameters such as the number of network layers, convolution kernel size, and pooling layer parameters. A cross-entropy loss function was used as the training objective, and stochastic gradient descent was used for parameter updates. After multiple rounds of iterative training, the model achieved a defect recognition accuracy of over 95% on the test set. During the actual welding process, the real-time collected solder joint morphology images are input into the trained model, and the model outputs the probability values of various defects. For example, the probability of judging that the current solder joint has a cold solder defect is 80%, providing a quantitative basis for subsequent decision-making.

[0076] A dual threshold trigger mechanism is established. When the resistivity change rate reaches a preset value or the defect probability of the solder joint morphology is greater than a preset value, the parameter adjustment is triggered. The solder joint morphology takes precedence over the resistivity and temperature. When the morphology defect probability is greater than 95%, the parameter adjustment is triggered immediately. If only the resistivity exceeds the limit, a delay of 20ms is made to confirm that there is no abnormality in the morphology before adjustment. When the resistivity change rate is ≥0.5mΩ / cm 2 When the probability of morphological defects is greater than 95%, parameter adjustment is triggered, which indicates that the electrode surface is oxidized or has a cold solder joint.

[0077] It should be noted that the step 4 also includes:

[0078] The electrode surface temperature is monitored using an infrared thermal imager. During the metallurgical stage, if the temperature exceeds the first preset temperature, the metallurgical stage is terminated and the tempering stage begins. If the temperature falls below the second preset temperature, the pulse current is increased and the pulse interval is shortened. The first preset temperature is 220°C. When the temperature exceeds this threshold, it indicates a risk of overheating during the welding process, which may cause damage to the cell substrate or deterioration of the intermetallic compound performance. The system immediately terminates the metallurgical stage and enters the tempering stage. The second preset temperature is 180°C. If the temperature falls below this value, it indicates insufficient heating. The system automatically increases the pulse current by 15A and shortens the pulse interval from 3ms to 2ms to quickly raise the welding temperature.

[0079] High-speed cameras monitor the weld joint morphology on the electrode surface in real time. During the metallurgical stage, the tempering stage begins when the weld joint morphology captured by the high-speed camera passes the morphology test. During the tempering stage, the tempering stage ends and the cooling stage begins when the surface smoothness of the weld joint captured by the high-speed camera reaches a preset value and no visible cracks or collapses are observed. The high-speed camera continuously monitors the weld joint morphology during the metallurgical stage. The system incorporates a built-in weld joint morphology detection algorithm based on morphology and deep learning, evaluating the weld joint's area, shape, and edge roughness in multiple dimensions. When the weld joint morphology meets preset standards, the tempering stage is triggered. During the tempering stage, the weld joint morphology continues to be monitored. When the surface smoothness reaches a preset value of Ra 0.8-1.6μm and no visible cracks, collapses, or other defects are observed, the system determines that the weld quality meets the standard, ends the tempering stage, and enters the cooling stage. During the cooling process, the wind speed of the air cooling device and the flow rate of the water cooling system are controlled in a closed loop to control the cooling rate at 5-15℃ / s, ensuring the stability of the solder joint crystallization process and avoiding internal stress caused by excessive cooling.

[0080] It should be noted that the step five includes:

[0081] The weld topography of the electrode surface after welding is captured and fed into the defect recognition model. If an unidentified defect is detected, an image is automatically captured and used to generate a training sample to update the defect recognition model. If a defect is identified, a test report is generated that includes the defect type and probability. After welding, the weld topography of the electrode surface is captured using a high-resolution industrial camera with an optical zoom lens. The camera's resolution should be at least 5 megapixels, and it is equipped with a circular shadowless light source and a coaxial light source to ensure uniform illumination and shadow-free lighting on the weld surface. During capture, an automated robotic arm controls the camera's position, capturing multi-angle images of each cell solder joint to obtain three-dimensional structural information. The captured image data undergoes lossless compression and is transmitted to the central processing unit, providing clear and complete raw data for subsequent defect recognition. If the model detects an unidentified defect type during solder joint image analysis, the system automatically captures the image and marks it as an "unknown defect" sample. This triggers a manual review process, where professional technicians conduct a detailed image analysis to determine the defect type and feature description. Once the image is labeled, the sample is included in the training dataset.

[0082] In one embodiment of the present invention, the present application further provides a solar cell welding system for use in the above-mentioned solar cell welding method.

[0083] In one embodiment of the present invention, the present application further provides a solar cell, which is manufactured using the above-mentioned solar cell welding method.

[0084] The present invention realizes an ultra-clean interface through the synergistic effect of plasma cleaning and chemical etching, thoroughly removes pollutants on the electrode surface, improves welding bonding strength, predicts the position of battery cell offset in advance through a dynamic replenishment mechanism, dynamically compensates for the welding strip, improves positioning accuracy, and reduces offset. By real-time detection of the resistivity, morphological characteristics and temperature changes of the welding points during the welding process, multi-parameter fusion is used to realize intelligent control of welding parameters, thereby optimizing energy consumption while improving welding quality.

[0085] The above description is only a preferred embodiment of the present invention. It should be pointed out that ordinary technicians in this technical field can make several improvements and modifications without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A solar cell welding method, characterized in that: The following steps are involved: Step 1: Pre-treating the electrode area of the cell by combining plasma cleaning and chemical etching; Step 2: Accurately align the soldering ribbon to the pre-set position of the electrode area, and adjust the position of the soldering ribbon in advance through the dynamic replenishment mechanism; Step 3: Complete the metallurgical bonding of the welding ribbon and the battery cell electrode area within a set temperature range through a pulse current welding process; Step 4: Real-time feedback of welding status is obtained through multiple sensors, and welding parameters are dynamically adjusted based on the welding status; Step 5: After the battery cell is welded, the pre-trained defect classification model is used to input the multimodal data of the battery cell and output the defect detection results.

2. A solar cell welding method according to claim 1, characterized in that: Step 1 also includes: The electrode surface is bombarded by high-energy plasma, and the high-energy ions physically strip off the organic matter on the electrode surface, and the hydrogen in the high-energy plasma reduces some of the metal oxides on the electrode surface; Prepare an etching solution, which is prepared by adding 0.5%-3% dilute hydrochloric acid and 0.01% benzotriazole. Immerse the electrode surface after plasma cleaning in the etching solution for a period of time, and use ultrasound to assist the immersion process; After impregnation, the electrode surface was rinsed with deionized water and then dried with nitrogen.

3. A solar cell welding method according to claim 1, characterized in that: Step 2 also includes: Obtaining feature data of the geometric center lines of the welding strip grid lines and the electrode main grid, and using an edge detection algorithm to make the welding strip grid lines coincide with the center lines of the electrode main grid; An offset prediction model is established based on the historical data of thermal expansion of the cell. The data of the welding ribbon and the electrode busbar are input to obtain compensation data. Based on the compensation data, the robot arm is controlled to deflect at the pre-calculated position. When acquiring the characteristic data of the geometric center line of the welding strip and the electrode main grid, edge features are extracted by a multi-light source fusion imaging method, and the multi-light source fusion imaging method includes: 0° orthophoto obtains clear main grid and blurred secondary grid images; Oblique illumination at 30° allows for clear secondary grid and low-contrast main grid images; Oblique illumination at 60° obtains an image with bright edges of the welding strip and dark background; The three-angle images are synthesized into HDR images to extract edge features.

4. A solar cell welding method according to claim 3, characterized in that: The offset prediction model is: Among them, T(x,y) is the coordinate position after deformation, which means the new position of the original point (x,y) after deformation; a0,a x ,a y is the coefficient of the linear part, describing the rigid deformation such as overall translation, rotation and scaling; ω i is the weight coefficient, which indicates the contribution of each control point to the local deformation; U(r) represents the radial basis function; ||(x,y)-(x i ,y i )|| is the Euclidean distance, which means the distance between the current point (x, y) and the control point (x i ,y i ) distance; n is the number of control points, usually the intersection points of the secondary grid with obvious features or artificial marking points are selected.

5. A solar cell welding method according to claim 1, characterized in that: Step three also includes: A high current pulse is input to heat the pre-treated electrode surface. This is the breakdown stage, and the resistance change of the battery cell is monitored in real time through several resistance probes set around the battery cell. When the resistance probe detects that the resistance fluctuation of the battery cell reaches the preset range, the high current pulse input is stopped and the medium current continuous pulse input is started to heat the electrode surface. This is the metallurgical stage. After the preset conditions are met in the metallurgical stage, the tempering stage begins. At this time, low current pulses are used to heat the electrode surface to maintain the electrode surface temperature. After heating for a preset time, the heating ends and the cooling stage begins. The preset conditions include temperature, heating time, energy accumulation, and the morphology of the weld spot on the electrode surface. During the metallurgical stage, when the resistance probe detects that the resistance fluctuation is greater than the preset range, the pulse current is increased and the heating time is extended.

6. A solar cell welding method according to claim 5, characterized in that: The fourth step includes: The temperature, solder joint morphology and local resistivity of the solder joints on the electrode surface are collected in real time using an infrared thermal imager, a high-speed camera and a resistance probe. The timestamps of the temperature, solder joint morphology and resistivity are synchronized to achieve data synchronization among multiple sensors. The defect recognition model is trained based on historical data of solder joint morphology, and the data of solder joint morphology is imported into the defect recognition model to obtain the defect type and defect probability; A dual-threshold trigger mechanism is established. When the resistivity change rate reaches a preset value or the defect probability of the solder joint morphology is greater than a preset value, the trigger parameters are adjusted, and the priority of the solder joint morphology is higher than the priority of the resistivity and temperature.

7. A solar cell welding method according to claim 6, characterized in that: The step 4 further includes: The temperature of the electrode surface is monitored by an infrared thermal imager. During the metallurgical stage, when the temperature exceeds a first preset temperature, the metallurgical stage is terminated and the tempering stage is entered; when the temperature is lower than a second preset temperature, the pulse current is increased and the pulse interval is shortened; The weld morphology on the electrode surface is monitored in real time by a high-speed camera. During the metallurgical stage, when the weld morphology captured by the high-speed camera passes the morphology detection, the tempering stage is started; during the tempering stage, when the surface smoothness of the weld morphology captured by the high-speed camera reaches the preset value and there are no visible cracks or collapses, the tempering stage is ended and the cooling stage begins.

8. A solar cell welding method according to claim 5, characterized in that: The step five includes: The weld spot morphology of the electrode surface after welding is completed is obtained, and the weld spot morphology data is imported into the defect recognition model. When an unidentified defect is detected, the image is automatically captured and a training sample is generated to update the defect recognition model. When a defect is identified, a detection report containing the defect type and defect probability is generated.

9. A solar cell welding system, characterized in that: Used to implement a solar cell welding method as described in any one of claims 1-8.

10. A solar cell, characterized in that: The solar cell is manufactured using the solar cell welding method according to any one of claims 1 to 8.

Images