Pole piece deviation rectifying system and method

Abstract

The invention relates to the technical field of batteries, and discloses a pole piece deviation rectifying system and method.The system comprises a processor, a deviation rectifying mechanism, a deviation rectifying sensor, an image collector and winding equipment, the deviation correction sensor is used for acquiring the pole piece edge position of the current pole piece material belt and sending the pole piece edge position to the processor; the image collector is used for obtaining the pole piece offset of the historical pole piece material belt in a free end area and sending the pole piece offset to the processor, and the free end area is the area between the deviation correction sensor and the winding equipment; the processor is used for determining a deviation correction target value through a preset deviation correction model based on the pole piece edge position and the pole piece offset, and the preset deviation correction model is a model obtained through training based on historical deviation correction information; and the deviation rectifying mechanism is used for rectifying the deviation of the current pole piece material belt based on the deviation rectifying target value, so that the deviation rectifying accuracy of the pole piece can be improved, and the deviation rectifying effect of the pole piece can be improved.

Description

Technical Field

[0001] This invention relates to the field of battery technology, and more specifically to an electrode correction system and method. Background Technology

[0002] Currently, continuous winding equipment cuts off the cathode electrode during needle switching, resulting in a free end at the head of the cathode electrode of the next cell before it enters the needle. This causes the cathode electrode head to swing out of its direction, so it is necessary to correct the deviation of the cathode electrode head.

[0003] However, the relevant electrode correction methods usually correct the next electrode strip directly based on the offset of the previous electrode strip, without considering the actual situation of the next electrode strip, resulting in low correction accuracy and poor correction effect. Summary of the Invention

[0004] In view of the above problems, the present invention provides an electrode correction system and method, which aims to solve the problem that the relevant electrode correction methods usually correct the next electrode strip directly based on the offset of the previous electrode strip, without considering the actual situation of the next electrode strip, thus resulting in low correction accuracy and poor correction effect.

[0005] In a first aspect, the present invention provides an electrode correction system, the electrode correction system comprising: a processor and a correction mechanism, a correction sensor, an image acquisition device and a winding device arranged sequentially according to the electrode travel direction; The correction sensor is used to obtain the electrode edge position of the current electrode strip and send the electrode edge position to the processor, wherein the electrode strip is the strip corresponding to an electrode of one cell length; The image acquisition device is used to acquire the electrode offset of the historical electrode strip in the free end region and send the electrode offset to the processor, wherein the free end region is the region between the correction sensor and the winding device; The processor is configured to determine a correction target value based on the electrode edge position and the electrode offset using a preset correction model, wherein the preset correction model is a model trained based on historical correction information. The correction mechanism is used to correct the current electrode strip based on the correction target value.

[0006] In the technical solution of this invention embodiment, the electrode edge position of the current electrode strip and the electrode offset of the historical electrode strip in the free end region are used to calculate the correction target value through a pre-trained preset correction model, thereby improving the accuracy of electrode correction and thus improving the electrode correction effect.

[0007] In some embodiments, the historical correction information includes the historical electrode edge position, historical electrode offset, and historical correction target value of the historical electrode material strip, and the historical electrode material strip is the electrode material strip whose correction result is qualified; The correction sensor is also used to obtain the historical electrode edge position of the historical electrode strip and send the historical electrode edge position to the processor; The image acquisition device is also used to acquire the historical electrode offset of the historical electrode strip and send the historical electrode offset to the processor; The processor is further configured to acquire the historical correction target value of the historical electrode strip, and train a correction model based on the historical electrode edge position, the historical electrode offset, and the historical correction target value to obtain a preset correction model.

[0008] In the technical solution of this invention embodiment, a correction model is trained based on the historical electrode edge position, historical electrode offset, and historical correction target value to obtain a preset correction model, thereby ensuring the prediction accuracy of the preset correction model and improving the reliability of the preset correction model.

[0009] In some embodiments, the processor is further configured to perform parameter solving based on the historical electrode edge position, the historical electrode offset, and the historical correction target value to determine the electrode edge position adjustment coefficient, the electrode offset adjustment coefficient, and the model intercept. The processor is also used to construct a preset correction model based on the electrode edge position adjustment coefficient, the electrode offset adjustment coefficient, and the model intercept.

[0010] In the technical solution of this invention embodiment, parameters are solved based on historical electrode edge position, historical electrode offset, and historical correction target value to determine electrode edge position adjustment coefficient, electrode offset adjustment coefficient, and model intercept. A preset correction model is then constructed based on the electrode edge position adjustment coefficient, electrode offset adjustment coefficient, and model intercept, thereby accurately predicting the correspondence between electrode offset and correction target value and improving the accuracy of the preset correction model.

[0011] In some embodiments, the processor is further configured to determine whether the current electrode strip and the previous electrode strip belong to the same correction control group, wherein the correction control group includes a preset number of electrode strips; The processor is further configured to use the correction model of the previous electrode strip as a preset correction model if the current electrode strip and the previous electrode strip belong to the same correction control group.

[0012] In the technical solution of this invention embodiment, when correcting the current electrode strip, it is first determined whether the current electrode strip and the previous electrode strip belong to the same correction control group. If the current electrode strip and the previous electrode strip belong to the same correction control group, there is no need to retrain the correction model. Instead, the correction model of the previous electrode strip is directly used as the preset correction model, thereby reducing the number of modeling times, reducing the processor's computation time, and thus improving the correction efficiency.

[0013] In some embodiments, the processor is further configured to, if the current electrode strip and the previous electrode strip do not belong to the same correction control group, obtain a qualified candidate electrode strip with a correction result, and sort the candidate electrode strips according to the winding order to obtain a sorting result. The processor is further configured to select a preset number of historical electrode strips from the candidate electrode strips according to the sorting result, and train a correction model based on the historical electrode edge position, historical electrode offset and historical correction target value of the historical electrode strips to obtain a preset correction model.

[0014] In the technical solution of this invention embodiment, if the current electrode strip and the previous electrode strip do not belong to the same correction control group, then the electrode strip with qualified correction results is obtained, and the electrode strips to be selected are sorted according to the winding order to obtain a sorting result; based on the sorting result, a preset number of historical electrode strips are selected from the electrode strips to be selected, and a correction model is trained based on the historical electrode edge position, historical electrode offset, and historical correction target value of the historical electrode strips to obtain a preset correction model. This can avoid the deviation of the target value caused by reusing the old model of the previous cell that does not belong to the same correction control group, and ensure the accuracy of the model.

[0015] In some embodiments, the processor is further configured to randomly generate multiple numerical groups and calculate the predicted correction value corresponding to each numerical group, wherein the numerical group includes the number of electrode strips in the correction control group and the number of samples selected from the candidate electrode strips. The processor is further configured to acquire the actual correction value and determine the preset quantity and the preset sample number based on the residual between the predicted correction value and the actual correction value.

[0016] In the technical solution of this invention embodiment, multiple numerical groups are randomly generated, and the predicted correction value corresponding to each numerical group is calculated. The numerical group includes the number of electrode strips in the correction control group and the number of samples selected from the candidate electrode strips. The actual correction value is obtained, and the preset quantity and preset sample number are determined according to the residual between the predicted correction value and the actual correction value, so as to ensure both model accuracy and correction efficiency.

[0017] In some embodiments, the winding device includes a winding needle; The processor is also used to obtain the winding needle corresponding to the current electrode strip and to find the preset correction model corresponding to the winding needle, wherein different winding needles correspond to different preset correction models.

[0018] In the technical solution of this invention embodiment, by obtaining the winding needle corresponding to the current electrode strip and finding the preset correction model corresponding to the winding needle, it is possible to ensure that the correction target value is completely matched with the winding needle, thereby improving the correction adaptability.

[0019] In some embodiments, the processor is further configured to acquire the historical correction value of the previous electrode strip and calculate the difference between the correction target value and the historical correction value. The correction mechanism is further configured to correct the current electrode strip based on the correction target value if the difference is less than a preset threshold.

[0020] In the technical solution of this invention embodiment, the historical correction value of the previous electrode strip is obtained, and the difference between the correction target value and the historical correction value is calculated. Only when the difference is less than a preset threshold is the current electrode strip corrected based on the correction target value. This ensures that the fluctuation of the current correction target value and the historical control parameters are within the range that the equipment and materials can withstand, and maintains the stability of continuous production.

[0021] Secondly, the present invention provides an electrode correction method, comprising: The electrode edge position of the current electrode strip and the electrode offset of the historical electrode strip in the free end region are obtained. The electrode strip is the strip corresponding to an electrode of one cell length, and the free end region is the area between the correction sensor and the winding device. The correction target value is determined by a preset correction model based on the electrode edge position and the electrode offset, wherein the preset correction model is a model trained based on historical correction information. The current electrode strip is corrected based on the correction target value.

[0022] In the technical solution of this invention embodiment, the electrode edge position of the current electrode strip and the electrode offset of the historical electrode strip in the free end region are used to calculate the correction target value through a pre-trained preset correction model, thereby improving the accuracy of electrode correction and thus improving the electrode correction effect.

[0023] In some embodiments, the historical correction information includes the historical electrode edge position, historical electrode offset, and historical correction target value of the historical electrode strip, wherein the historical electrode strip is an electrode strip whose correction result is qualified; before obtaining the current electrode edge position and the historical electrode offset of the electrode strip in the free end region, the method further includes: Obtain the historical electrode edge position, historical electrode offset, and historical correction target value of the historical electrode strip; A correction model is trained based on the historical electrode edge position, the historical electrode offset, and the historical correction target value to obtain a preset correction model.

[0024] In the technical solution of this invention embodiment, a correction model is trained based on the historical electrode edge position, historical electrode offset, and historical correction target value to obtain a preset correction model, thereby ensuring the prediction accuracy of the preset correction model and improving the reliability of the preset correction model.

[0025] In some embodiments, training a correction model based on the historical electrode edge position, the historical electrode offset, and the historical correction target value to obtain a preset correction model includes: Based on the historical electrode edge position, the historical electrode offset, and the historical correction target value, the parameters are solved to determine the electrode edge position adjustment coefficient, the electrode offset adjustment coefficient, and the model intercept. A preset correction model is constructed based on the electrode edge position adjustment coefficient, the electrode offset adjustment coefficient, and the model intercept.

[0026] In the technical solution of this invention embodiment, parameters are solved based on historical electrode edge position, historical electrode offset, and historical correction target value to determine electrode edge position adjustment coefficient, electrode offset adjustment coefficient, and model intercept. A preset correction model is then constructed based on the electrode edge position adjustment coefficient, electrode offset adjustment coefficient, and model intercept, thereby accurately predicting the correspondence between electrode offset and correction target value and improving the accuracy of the preset correction model.

[0027] In some embodiments, before determining the correction target value based on the electrode edge position and the electrode offset using a preset correction model, the method further includes: Determine whether the current electrode strip and the previous electrode strip belong to the same correction control group, wherein the correction control group includes a preset number of electrode strips; If the current electrode strip and the previous electrode strip belong to the same correction control group, then the correction model of the previous electrode strip is used as the preset correction model.

[0028] In the technical solution of this invention embodiment, when correcting the current electrode strip, it is first determined whether the current electrode strip and the previous electrode strip belong to the same correction control group. If the current electrode strip and the previous electrode strip belong to the same correction control group, there is no need to retrain the correction model. Instead, the correction model of the previous electrode strip is directly used as the preset correction model, thereby reducing the number of modeling times, reducing the processor's computation time, and thus improving the correction efficiency.

[0029] In some embodiments, after determining whether the current electrode strip and the previous electrode strip belong to the same correction control group, the method further includes: If the current electrode strip and the previous electrode strip do not belong to the same correction control group, then the correction result is a qualified candidate electrode strip, and the candidate electrode strips are sorted according to the winding order to obtain the sorting result. Based on the sorting results, a preset number of historical electrode strips are selected from the candidate electrode strips. A correction model is trained based on the historical electrode edge position, historical electrode offset, and historical correction target value of the historical electrode strips to obtain the preset correction model.

[0030] In the technical solution of this invention embodiment, if the current electrode strip and the previous electrode strip do not belong to the same correction control group, then the electrode strip with qualified correction results is obtained, and the electrode strips to be selected are sorted according to the winding order to obtain a sorting result; based on the sorting result, a preset number of historical electrode strips are selected from the electrode strips to be selected, and a correction model is trained based on the historical electrode edge position, historical electrode offset, and historical correction target value of the historical electrode strips to obtain a preset correction model. This can avoid the deviation of the target value caused by reusing the old model of the previous cell that does not belong to the same correction control group, and ensure the accuracy of the model.

[0031] In some embodiments, before determining whether the current electrode strip and the previous electrode strip belong to the same correction control group, the method further includes: Multiple numerical groups are randomly generated, and the predicted correction value corresponding to each numerical group is calculated. The numerical group includes the number of electrode strips in the correction control group and the number of samples selected from the candidate electrode strips. Obtain the actual correction value, and determine the preset quantity and the preset sample number based on the residual between the predicted correction value and the actual correction value.

[0032] In the technical solution of this invention embodiment, multiple numerical groups are randomly generated, and the predicted correction value corresponding to each numerical group is calculated. The numerical group includes the number of electrode strips in the correction control group and the number of samples selected from the candidate electrode strips. The actual correction value is obtained, and the preset quantity and preset sample number are determined according to the residual between the predicted correction value and the actual correction value, so as to ensure both model accuracy and correction efficiency.

[0033] In some embodiments, before determining the correction target value based on the electrode edge position and the electrode offset using a preset correction model, the method further includes: Obtain the winding needle corresponding to the current electrode strip and find the preset correction model corresponding to the winding needle. Different winding needles correspond to different preset correction models.

[0034] In the technical solution of this invention embodiment, by obtaining the winding needle corresponding to the current electrode strip and finding the preset correction model corresponding to the winding needle, it is possible to ensure that the correction target value is completely matched with the winding needle, thereby improving the correction adaptability.

[0035] In some embodiments, before correcting the current electrode strip based on the correction target value, the method further includes: Obtain the historical correction value of the previous electrode strip and calculate the difference between the target correction value and the historical correction value; Accordingly, the step of correcting the current electrode strip based on the correction target value includes: If the difference is less than a preset threshold, then the current electrode strip is corrected based on the correction target value.

[0036] In the technical solution of this invention embodiment, the historical correction value of the previous electrode strip is obtained, and the difference between the correction target value and the historical correction value is calculated. Only when the difference is less than a preset threshold is the current electrode strip corrected based on the correction target value. This ensures that the fluctuation of the current correction target value and the historical control parameters are within the range that the equipment and materials can withstand, and maintains the stability of continuous production.

[0037] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0038] Figure 1 This is a top view of an electrode correction system according to some embodiments of the present invention; Figure 2 This is a front view of an electrode correction system according to some embodiments of the present invention; Figure 3 This is a schematic diagram of the correction control group according to some embodiments of the present invention; Figure 4 This is a schematic diagram of the correction control group according to some embodiments of the present invention; Figure 5 This is a schematic diagram of multiple numerical groups for some embodiments of the present invention; Figure 6 This is a flowchart illustrating the operation of an electrode correction system according to some embodiments of the present invention; Figure 7 This is a flowchart of an electrode correction method according to some embodiments of the present invention; Figure 8 This is a flowchart of an electrode correction method according to some embodiments of the present invention; Figure 9 This is a flowchart of an electrode correction method according to some embodiments of the present invention; Figure 10 This is a flowchart of an electrode correction method according to some embodiments of the present invention; Figure 11 This is a flowchart of an electrode correction method according to some embodiments of the present invention.

[0039] The reference numerals in the detailed embodiments are as follows: Processor 100, electrode roll 20, web guiding mechanism 30, web guiding sensor 40, image acquisition device 50, winding device 60, and free end area 70. Detailed Implementation

[0040] The embodiments of the technical solution of the present invention will now be described in detail with reference to the accompanying drawings. These embodiments are merely illustrative of the technical solution of the present invention and are therefore intended to limit the scope of protection of the present invention.

[0041] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the invention; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the invention, are intended to cover non-exclusive inclusion.

[0042] In the description of the embodiments of this invention, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this invention, "multiple" means two or more, unless otherwise explicitly defined.

[0043] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0044] In the description of the embodiments of this invention, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0045] In the description of the embodiments of the present invention, the term "multiple" refers to two or more (including two), similarly, "multiple groups" refers to two or more (including two groups), and "multiple pieces" refers to two or more (including two pieces).

[0046] In the description of the embodiments of the present invention, the technical terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," "counterclockwise," "axial," "radial," and "circumferential" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing the embodiments of the present invention and simplifying the description, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the embodiments of the present invention.

[0047] In the description of the embodiments of this invention, unless otherwise explicitly specified and limited, technical terms such as "installation," "connection," "joining," and "fixing" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in the embodiments of this invention according to the specific circumstances. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0048] Currently, judging from market trends, battery applications are becoming increasingly widespread. Batteries are not only used in energy storage systems such as hydropower, thermal power, wind power, and solar power plants, but also extensively in electric vehicles such as electric bicycles, electric motorcycles, and electric cars, as well as in military equipment and aerospace. With the continuous expansion of battery applications, market demand is also constantly increasing.

[0049] During the battery electrode production process, the continuous winding equipment cuts off the cathode electrode when switching winding needles, resulting in a free end at the head of the cathode electrode of the next cell before entering the winding needle. This causes the cathode electrode head to swing out, so it is necessary to correct the deviation of the cathode electrode head.

[0050] For the first dual-wound cell after replacing the cathode material roll, there is no effective technical solution. The only option is to correct the overhang (OH) of the first dual-wound cell by setting the final overhang target value of the previous cathode material roll. In the winding process, OH refers to the portion of the negative electrode material that extends beyond the coverage area of ​​the positive electrode material during winding. This can lead to uneven local current distribution, causing lithium plating and increasing the risk of lithium dendrite formation, thus affecting battery safety and lifespan.

[0051] However, directly correcting the deviation of the next electrode strip based on the offset of the previous electrode strip does not take into account the actual situation of the next electrode strip, resulting in low correction accuracy and poor correction effect.

[0052] Therefore, in order to solve the above-mentioned technical problems, the target value of correction can be calculated by a pre-trained correction model based on the current electrode edge position of the electrode strip and the electrode offset of the historical electrode strip in the free end region. This can improve the accuracy of electrode correction and thus improve the electrode correction effect.

[0053] In some embodiments, an electrode correction system is proposed, comprising: a processor and a correction mechanism, a correction sensor, an image acquisition device, and a winding device arranged sequentially according to the electrode strip direction; the correction sensor is used to acquire the electrode edge position of the current electrode strip and send the electrode edge position to the processor, wherein the electrode strip is a strip corresponding to an electrode of one cell length; the image acquisition device is used to acquire the electrode offset of a historical electrode strip in the free end region and send the electrode offset to the processor, wherein the free end region is the region between the correction sensor and the winding device; the processor is used to determine a correction target value based on the electrode edge position and the electrode offset using a preset correction model, wherein the preset correction model is a model trained based on historical correction information; the correction mechanism is used to correct the current electrode strip based on the correction target value.

[0054] The electrode alignment system refers to a closed-loop control system used in lithium battery winding production to control the position of the cathode electrode and prevent the electrode head from swinging before entering the winding equipment. Its core is to ensure the consistency of the OH (the portion of the negative electrode extending beyond the positive electrode) at the electrode head through a detection-calculation-adjustment logic, reducing the risk of lithium plating and lithium dendrite formation. The current electrode strip can refer to the first double-wound cell of a new cathode roll after a roll of cathode material has been completely consumed (no remaining electrode sheets to be wound) and replaced with a new roll, after the waste material at the head of the roll has been removed. Alternatively, it can refer to the cell that subsequently enters the winding needle after the previous electrode strip has been cut; this embodiment does not limit this. A double-wound cell can refer to a cell obtained by winding the cathode and anode electrodes together in the winding needle.

[0055] The free end region can refer to the area without a load after the electrode passes through the correction mechanism (such as the correction roller) and the correction sensor and before it reaches the winding equipment (winding needle). The electrode is in a suspended state in this region. If it is not controlled, it is easy to swing its head. This is the core working area of ​​the correction system.

[0056] For ease of understanding, please refer to Figure 1 and Figure 2 This explanation is provided, but does not limit the scope of this application. Figure 1 This is a top view of an electrode correction system according to some embodiments of the present invention. Figure 2 This is a front view of an electrode correction system according to some embodiments of the present invention. As an example, such as... Figure 1 and Figure 2 As shown, the electrode web guiding system includes: a processor 100 and a web guiding mechanism 30, a web guiding sensor 40, an image acquisition device 50, and a winding device 60 arranged sequentially according to the electrode's belt travel direction. The processor 100 can train a preset web guiding model based on historical data, and use real-time electrode position data to deduce the web guiding target value. It can also optimize the target value within the controllable range of the device to avoid damage caused by exceeding the device's capabilities. In a specific implementation, the processor 100 can be a programmable logic controller (PLC). The web guiding mechanism 30 can include core components such as a web guiding roller and a web guiding motor. It can move left and right along the electrode's belt travel direction, adjusting the left and right position of the electrode roll 20 according to the web guiding target value output by the processor 100, keeping the electrode roll 20 at a set reference position. The web guiding sensor 40 can accurately detect the edge position of the current electrode strip 20 on the slitting side, providing real-time position data to the processor for target value calculation. Image acquisition device 50 may refer to a charge-coupled device (CCD) that can detect electrode offset in real time. Winding equipment 60 may include core components such as winding needles, used to wind the calibrated electrode roll 20 into a battery cell. Free end region 70 (e.g. Figure 1The area within the dashed box (in the middle) refers to the unsupported area between the electrode roll 20 and the winding device 60 (winding needle) after passing through the correction mechanism 30 (correction roller) and the correction sensor 40. The electrode roll 20 is suspended in this area; if uncontrolled, it is prone to tipping. This is the core area of ​​the correction system. The electrode edge position refers to the current position of the electrode slitting side, detected in real-time by the correction sensor 40. This is the real-time basis for the processor to determine the degree of electrode offset and calculate the correction target value, directly affecting the accuracy of the target value. The electrode offset can be the deviation between the actual position (e.g., the electrode slitting side) of the historical electrode strip in the free end area, obtained by the image acquisition unit 50 (winding CCD), and the set position (e.g., the baseline). The preset correction model refers to a linear model trained based on historical correction information of cells with acceptable winding CCD results, which can be used to deduce the correction target value of the current cell. The correction target value can refer to the electrode position adjustment value calculated by the processor based on the preset correction model, the real-time electrode edge position, and the historical electrode offset. It is the basis for the correction mechanism to adjust the electrode position. Finally, it can be optimized by the PLC in conjunction with the upper and lower limits of the equipment to ensure that it is within a controllable range.

[0057] This embodiment calculates the correction target value based on the current electrode edge position of the electrode strip and the electrode offset of the historical electrode strip in the free end region through a pre-trained preset correction model, thereby improving the accuracy of electrode correction and thus improving the electrode correction effect.

[0058] In some embodiments, the historical correction information includes the historical electrode edge position, historical electrode offset, and historical correction target value of the historical electrode strip, wherein the historical electrode strip is an electrode strip whose correction result is qualified; the correction sensor is further configured to acquire the historical electrode edge position of the historical electrode strip and send the historical electrode edge position to the processor; the image acquisition device is further configured to acquire the historical electrode offset of the historical electrode strip and send the historical electrode offset to the processor; the processor is further configured to acquire the historical correction target value of the historical electrode strip, and train a correction model based on the historical electrode edge position, the historical electrode offset, and the historical correction target value to obtain a preset correction model.

[0059] To ensure the prediction accuracy and improve the reliability of the preset correction model, in this embodiment, the correction model is trained based on the historical electrode edge position, historical electrode offset, and historical correction target value to obtain the preset correction model.

[0060] The historical electrode edge position can refer to the electrode slitting side position data collected by the correction sensor at the start of the electrode feeding correction, i.e., the parameters in the linear model. It reflects the initial position of the historical electrode strip when it enters the correction stage, and is a key input for linking the initial position and final offset in model training. The historical electrode offset can refer to the deviation between the actual position and the set position of the historical electrode strip in the free end region, as acquired by the image acquisition device (wound CCD), i.e., a parameter in the linear model. It directly reflects the final offset state of the historical electrode strip and is the core basis for judging the correction effect during model training. The historical correction target value can refer to the correction benchmark value set when the historical electrode strip was wound, i.e., the parameter in the linear model. It serves as a target benchmark used to control electrode position during historical production, recording the correspondence between the set target and the actual offset, and is a crucial input parameter for model training. The preset correction model can refer to a model based on historical correction information (…). , , The linear mathematical model or other models obtained through training are not limited in this embodiment.

[0061] In practice, during the continuous winding and consumption of the previous roll of cathode material, each component synchronously collects historical data of qualified electrode sheets according to its division of labor: for each cell that is finally detected as OK by CCD, the correction sensor collects the edge position of the historical electrode sheet strip when its electrode sheet reaches the starting position of the sheet feeding correction. The image acquisition unit (wound CCD) simultaneously acquires the offset of the qualified cell electrode in the free end region (the position of the historical electrode edge) and sends it to the processor; the image acquisition unit (wound CCD) simultaneously acquires the offset of the qualified cell electrode in the free end region (the position of the historical electrode edge) and sends it to the processor; The historical electrode offset is recorded and transmitted to the processor. Simultaneously, the processor actively retrieves the historical correction target value actually executed by the control system during the winding of the qualified battery cell. Complete the single set of historical correction information ( , , The collection of ).

[0062] When the current roll of cathode material is completely used up and a new roll is about to be replaced, the program automatically triggers the algorithm, and the processor retrieves the historical correction information (multiple sets) of all qualified cells from the previous roll. The data (historical correction information) is used as training samples to train the correction model, ultimately obtaining a preset correction model that is supported by historical data and can accurately associate "electrode position - offset - correction target value". During training, the processor automatically filters qualified data that are detected as OK by CCD and removes unqualified data to ensure the validity of the model training data. At the same time, it distinguishes between different winding needles (because each winding needle is different) to train the model separately, further improving the model's specificity.

[0063] This embodiment trains a correction model based on historical electrode edge positions, historical electrode offsets, and historical correction target values ​​to obtain a preset correction model, thereby ensuring the prediction accuracy of the preset correction model and improving its reliability.

[0064] In some embodiments, the processor is further configured to perform parameter solving based on the historical electrode edge position, the historical electrode offset, and the historical correction target value to determine the electrode edge position adjustment coefficient, the electrode offset adjustment coefficient, and the model intercept; the processor is further configured to construct a preset correction model based on the electrode edge position adjustment coefficient, the electrode offset adjustment coefficient, and the model intercept.

[0065] In order to accurately predict the correspondence between electrode offset and correction target value and improve the accuracy of the preset correction model, in this embodiment, parameters are solved based on historical electrode edge position, historical electrode offset and historical correction target value to determine electrode edge position adjustment coefficient, electrode offset adjustment coefficient and model intercept, and a preset correction model is constructed based on electrode edge position adjustment coefficient, electrode offset adjustment coefficient and model intercept.

[0066] If the preset correction model is a linear mathematical model, then the expression is: (in, This represents the historical polarimetric offset. This is the edge of a historical polar sheet. For historical correction target values, This is the electrode edge position adjustment coefficient. This is the electrode offset adjustment coefficient. (This refers to the model intercept), its core function is to infer the target value for the new battery cell's polarization correction by analyzing historical data patterns. The electrode edge position adjustment coefficient can refer to the coefficient in the preset polarization correction model. It is used to measure the position of the edge of historical polar sheets ( ) on historical electrode offset ( The weighting parameters for the degree of influence are calculated by the processor based on historical data, and their values ​​reflect the strength of the correlation between the initial edge position of the electrode and the final correction effect. The electrode offset adjustment coefficient can refer to a coefficient in the preset correction model. It is a measure of historical correction target value ( ) on historical electrode offset ( The weight parameters for the degree of influence are also obtained by the processor through historical data. The model intercept can refer to parameter b in the preset correction model. It is the basic offset compensation value of the linear model, which is obtained by the processor by fitting historical data. It is used to correct the system deviation caused by fixed factors such as inherent equipment errors and slight differences in electrode materials, thereby improving the accuracy of model prediction.

[0067] In the specific implementation, when a roll of cathode material is completely consumed and a new roll is about to be replaced, the processor automatically triggers the parameter solving process. First, it retrieves the historical correction information of all qualified cells in the roll that have been detected as OK by CCD, including the historical electrode edge position of each group of cells. ), historical polarimetric offset ( ) and historical correction target values ​​( The processor automatically removes unqualified data to ensure the validity of the input data. At the same time, the processor distinguishes the winding needles currently used in production (because each winding needle has different structures and precision), and filters the corresponding historical data for different winding needles to avoid parameter deviations caused by differences in winding needles.

[0068] The processor will filter multiple sets of historical data ( , , Input the preset multivariate linear regression algorithm, using a linear model To fit the target, the algorithm minimizes the difference between the predicted value and the actual historical pole offset. The error of the electrode edge position adjustment coefficient is then calculated by determining the error of the electrode edge position adjustment coefficient. ), electrode offset adjustment coefficient ( ) and model intercept ( ).in, The solution reflects right Influence weight, The solution reflects right Influence weight, The solution compensates for the inherent biases of the system, ultimately yielding a set of parameter values ​​that accurately fit the patterns in historical data. Finally, when , , Once the value is determined, the preset correction model is trained and can be used to calculate the correction target value of the current cell electrode.

[0069] When the current cell electrode reaches the starting position of the electrode feeding and correction, the correction sensor collects the current electrode edge position in real time (i.e., the current...). ) and send it to the processor; the processor first obtains the electrode offset of the historical cell electrode ( The mean of ) and then Mean, Current and well-trained , , Substitute into the transformed model The initial correction target value is obtained by reverse calculation. The processor then performs a threshold judgment on the initial target value, excluding values ​​that are outside the controllable range of the device, and obtains the correction target value.

[0070] This embodiment solves for parameters based on historical electrode edge positions, historical electrode offsets, and historical correction target values ​​to determine electrode edge position adjustment coefficients, electrode offset adjustment coefficients, and model intercepts. Based on these parameters, a preset correction model is constructed, which can accurately predict the correspondence between electrode offsets and correction target values, thereby improving the accuracy of the preset correction model.

[0071] In some embodiments, the processor is further configured to determine whether the current electrode strip and the previous electrode strip belong to the same correction control group, wherein the correction control group includes a preset number of electrode strips; the processor is further configured to, if the current electrode strip and the previous electrode strip belong to the same correction control group, use the correction model of the previous electrode strip as a preset correction model.

[0072] To reduce the number of modeling iterations, decrease processor computation time, and improve correction efficiency, in this embodiment, when correcting the current electrode strip, it is first determined whether the current electrode strip and the previous electrode strip belong to the same correction control group. If the current electrode strip and the previous electrode strip belong to the same correction control group, there is no need to retrain the correction model; instead, the correction model of the previous electrode strip is directly used as the preset correction model.

[0073] The preceding electrode strip refers to the immediately preceding cathode electrode strip that has been wound and whose winding CCD inspection result is qualified. It serves as the benchmark for determining the current cell group's affiliation. Its correction model is trained based on historical qualified data from its control group and has been verified to effectively control head offset, providing a basis for model reuse for the current cells in the same group. The correction control group refers to a production unit divided according to a preset number of consecutive qualified CCD inspection electrode strips, based on the stable and predictable nature of continuous qualified cell data in lithium battery winding production. Within the same group, the production environment of the electrode strips (such as winding needle wear, electrode tension, and equipment operating parameters) remains largely unchanged, and the electrode offset (…) ), correction sensor data ( The data patterns are consistent, and the same correction model is used to avoid model failure caused by occasional fluctuations in a single cell. The preset quantity can refer to the number of qualified CCD electrode strips that constitute a correction control group, which is a quantitative standard for dividing control groups.

[0074] For ease of understanding, please refer to Figure 3 This explanation is provided, but does not limit the scope of this application. Figure 3 This is a schematic diagram of the correction control group according to some embodiments of the present invention, as an example. Figure 3 In the winding sequence, the cells are: Cell 1, Cell 2, Cell 3, Cell 4, Cell 5, Cell 6, Cell 7, and Cell 8. Assuming the spin correction control group contains three electrode strips, and the current electrode strip is Cell 8, and Cell 8 belongs to the same spin correction control group as the previous electrode strip (Cell 7), Cell 8 can directly use the spin correction model of Cell 7 as the preset spin correction model to calculate the spin correction target value. The spin correction model of Cell 7 is essentially the spin correction model of the previous electrode strip (Cell 6), meaning that Cells 6, 7, and 8 essentially share the same spin correction model, thus reducing the number of modeling iterations. The spin correction model of Cell 6 is trained based on the historical electrode edge positions, historical electrode offsets, and historical spin correction target values ​​of Cells 1, 2, 3, 4, and 5.

[0075] In this embodiment, when correcting the current electrode strip, it first determines whether the current electrode strip and the previous electrode strip belong to the same correction control group. If the current electrode strip and the previous electrode strip belong to the same correction control group, there is no need to retrain the correction model. Instead, the correction model of the previous electrode strip is directly used as the preset correction model, which can reduce the number of modeling times, reduce the processor's computing time, and thus improve the correction efficiency.

[0076] In some embodiments, the processor is further configured to: if the current electrode strip and the previous electrode strip do not belong to the same correction control group, obtain a candidate electrode strip with a qualified correction result, and sort the candidate electrode strips according to the winding order to obtain a sorting result; the processor is further configured to: select a preset number of historical electrode strips from the candidate electrode strips according to the sorting result, and train a correction model based on the historical electrode edge position, historical electrode offset, and historical correction target value of the historical electrode strips to obtain a preset correction model.

[0077] To avoid target value deviations caused by reusing old models from previous cells that do not belong to the same correction control group, and to ensure model accuracy, in this embodiment, if the current electrode strip and the previous electrode strip do not belong to the same correction control group, the electrode strips with qualified correction results are obtained, and the electrode strips are sorted according to the winding order to obtain the sorting result; based on the sorting result, a preset number of historical electrode strips are selected from the electrode strips, and the correction model is trained based on the historical electrode edge position, historical electrode offset, and historical correction target value of the historical electrode strips to obtain the preset correction model.

[0078] The candidate electrode strips refer to cathode electrodes selected from historical production data that have passed CCD winding inspection. These serve as the foundation for training the new correction model. Selecting only qualified cells avoids interference from anomalous data (such as cells with offsets exceeding ±0.5mm) that could affect model accuracy, ensuring the validity of the model's input data. The winding sequence refers to the actual winding order of the candidate electrode strips during historical production (e.g., production sequence 1, 2, 3…). Sorting in this order ensures that the selected historical electrode strips reflect data variation patterns under continuous production conditions, preventing the model from being unsuitable for the current continuous winding scenario due to random sampling. The sorting result refers to the ordered data list formed by arranging the candidate electrode strips in winding order. This serves as the data pool for subsequently selecting a preset number of historical cells, ensuring the selected samples have temporal continuity and are consistent with the current continuous production scenario for cells. The preset sample size refers to the number of historical electrode strips used to train the new correction model, serving as the standard for the amount of data for model training.

[0079] For ease of understanding, please refer to Figure 4 This explanation is provided, but does not limit the scope of this application. Figure 4 This is a schematic diagram of the correction control group according to some embodiments of the present invention, as an example. Figure 4 In the process, the cells are arranged in the winding sequence as follows: Cell 1, Cell 2, Cell 3, Cell 4, Cell 5, Cell 6, Cell 7, Cell 8, Cell 9, Cell 10, and Cell 11. Assuming the correction control group contains three electrode strips, and the current electrode strip is Cell 9, when the processor determines that the current electrode strip (Cell 9) does not belong to the same correction control group as the previous electrode strip (Cell 8), the processor first selects all cathode electrodes with qualified winding CCD detection results from historical production data as candidate electrode strips. Subsequently, the candidate electrode strips are sorted according to their winding sequence in historical production (e.g., production sequence number from earliest to latest), forming an ordered list of candidate electrode strips (e.g., sorting result: Cell 1 → Cell 2 → Cell 3 → …, corresponding to production sequence number 1 → 2 → 3 → …), ensuring that the subsequently selected samples reflect the data patterns under continuous production conditions. The processor selects samples from the sorted list of candidate electrode strips based on a preset number of samples (e.g., 5), prioritizing samples from the winding sequence. It can select consecutive samples closest to the current cell production sequence (e.g., cells 4, 5, 6, 7, and 8), as the production environment (e.g., winding needle wear, electrode tension) is closest to the current cell's new group environment, maximizing model adaptability. The processor extracts three sets of core data from the selected historical electrode strips, including the historical electrode edge position (i.e.,...). ), historical polarimetric offset (i.e. The historical correction target value (i.e., x2) is used to substitute the average value of each group of data into the linear model formula. The optimal coefficients are calculated using data fitting algorithms (such as the least squares method). , and intercept That is, by adjusting , , The value is chosen to minimize the error between the y-value calculated by the model and the actual historical electrode offset y (e.g., controlling the error within ±0.05mm). When , , Once determined, the new preset correction model is trained and can be used to calculate the correction target value of the current electrode strip.

[0080] In this embodiment, if the current electrode strip and the previous electrode strip do not belong to the same correction control group, the electrode strip with the corrected correction result is selected as a qualified candidate electrode strip. The candidate electrode strips are sorted according to the winding order to obtain a sorting result. Based on the sorting result, a preset number of historical electrode strips are selected from the candidate electrode strips. The correction model is trained based on the historical electrode edge position, historical electrode offset, and historical correction target value of the historical electrode strips to obtain a preset correction model. This can avoid the target value deviation caused by reusing the old model of the previous cell that does not belong to the same correction control group, and ensure the model accuracy.

[0081] In some embodiments, the processor is further configured to randomly generate multiple numerical groups and calculate the predicted correction value corresponding to each numerical group, wherein the numerical group includes the number of electrode strips in the correction control group and the number of samples selected from the candidate electrode strips; the processor is further configured to obtain the actual correction value and determine the preset quantity and the preset number of samples based on the residual between the predicted correction value and the actual correction value.

[0082] To ensure both model accuracy and correction efficiency, this embodiment randomly generates multiple numerical groups and calculates the predicted correction value for each numerical group. The numerical group includes the number of electrode strips in the correction control group and the number of samples selected from the candidate electrode strips. The actual correction value is obtained, and the preset quantity and preset sample number are determined based on the residual between the predicted correction value and the actual correction value.

[0083] In this context, a numerical set refers to a combined unit constructed by the processor to select optimal parameters. It consists of two core parameters: first, the number of electrode strips in the correction control group (i.e., the basic number for dividing the correction control group, also known as the sliding window size); and second, the number of samples selected from the candidate electrode strips (the number of historically qualified cells used to train the correction model). Different combinations of these two parameter values ​​form multiple numerical sets (e.g., "sample count = 5 + sliding window size = 3", "sample count = 6 + sliding window size = 4"), providing a candidate set of parameters for subsequent residual comparison. The predicted correction value can refer to the target correction value of the electrode strip calculated based on the correction model corresponding to a certain numerical set. Specifically, the processor first selects candidate electrode strips (historical cells that passed CCD detection) based on the number of samples in the numerical set, and then uses these samples... (Edge position of historical polar image) (Historical polarity offset) (Historical electrode correction values) Training a linear model Then, the model is applied to the cells corresponding to the number of control cells in the numerical group, and the output is... The predicted correction value reflects the model's control effect on this numerical group. The actual correction value can refer to the target correction value that the candidate electrode strip has actually used in historical production and has been verified to effectively control the electrode head offset. This value is the benchmark for judging the accuracy of the predicted correction value. Since this value has achieved electrode head-free operation and qualified OH in historical production, it can be used as the true value for residual calculation. The residual can refer to the difference between the predicted correction value and the actual correction value of the same electrode strip (which can be the absolute value or the difference of squares). It is the core indicator for measuring the model accuracy corresponding to the numerical group. The smaller the residual, the more the parameter combination of the numerical group (number of cells in the control group, number of samples) is adapted to the production pattern, and the more accurate the model prediction is. The preset quantity can refer to the number of electrode strips in the final determined correction control group, which is a fixed standard for dividing the correction control group. The processor compares the residuals of all numerical groups and selects the number of cells in the control group of the numerical group with the smallest residual as the preset quantity to ensure that the data pattern within the same control group is stable and adapted to the same model. The preset sample size can refer to the final number of samples selected from the candidate electrode strips, which is a fixed sample size for training the preset correction model. Similarly, it is determined by residual optimization, selecting the sample size from the group of values ​​with the smallest residual as the preset sample size, ensuring that the samples for training the model can both cover stable production patterns and avoid computational redundancy caused by too many samples.

[0084] For ease of understanding, please refer to Figure 5 This explanation is provided, but does not limit the scope of this application. Figure 5 This is a schematic diagram of multiple numerical groups for some embodiments of the present invention. Figure 5In the process, the processor first determines the value ranges of two types of parameters based on the production experience of the previous cathode material roll. The first is the candidate value for the number of electrode strips in the correction control group (e.g., 5, 8, 10, balancing data stability and computational efficiency). The second is the candidate value for the number of samples selected from the candidate electrode strips (e.g., 3, 5, 8, ensuring the samples can fit an effective linear model). Then, the candidate values ​​of both types of parameters are fully combined to randomly generate multiple numerical sets, for example: First set of data: number of samples = 5, sliding window size = 3; Second set of data: number of samples = 6, sliding window size = 4; ...... Group n data: number of samples = n, sliding window size = n (to ensure that all possible reasonable combinations are covered and to avoid missing the optimal parameters).

[0085] For each set of values, the processor performs a three-step operation: "sampling-modeling-prediction". 1. Sample selection: From the candidate electrode strips (the previous roll of qualified CCD cells), select consecutive historical electrode strips according to the number of samples (e.g., if the number of samples in the first group of data is 5, then select the 5 qualified cells in the previous roll whose timing is closest to the current time). 2. Modeling: Extracting the selected samples , , Data, substituted into a linear model The specific values ​​for this set of values ​​can be calculated through data fitting (such as least squares method). , , Complete model training; 3. Prediction: Apply this model to the cells corresponding to the number of cells in the control group in the numerical group (e.g., if the sliding window size of the first group of data is 3, then take 3 consecutive cells), and input these cells... and ( Take samples (mean), through the transformation formula Calculate the predicted correction value for each battery cell.

[0086] The processor first retrieves the actual correction value (i.e., historical x 2) of the cells corresponding to the number of cells in the control group from historical data. Then, it calculates the average residual between the predicted correction value and the actual correction value of all cells in each numerical group (e.g., the residuals of the three cells in the first data group are 0.01mm, 0.02mm, ..., with an average of 0.015mm). Finally, it compares the average residuals of all numerical groups and selects the numerical group with the smallest residual. For example, the first data group (sample size = 5, sliding window size = 3) has the smallest average residual of 0.015mm. Therefore, the sliding window size = 3 for this numerical group is determined as the preset number, and the sample size = 5 is determined as the preset sample size.

[0087] In this embodiment, multiple numerical groups are randomly generated, and the predicted correction value corresponding to each numerical group is calculated. The numerical group includes the number of electrode strips in the correction control group and the number of samples selected from the candidate electrode strips. The actual correction value is obtained, and the preset quantity and preset sample number are determined according to the residual between the predicted correction value and the actual correction value, so as to ensure both model accuracy and correction efficiency.

[0088] In some embodiments, the winding device includes a winding needle; the processor is further configured to acquire the winding needle corresponding to the current electrode strip and search for a preset correction model corresponding to the winding needle, wherein different winding needles correspond to different preset correction models.

[0089] To ensure a perfect match between the correction target value and the winding needle and improve the adaptability of the correction, this embodiment obtains the winding needle corresponding to the current electrode strip and searches for the preset correction model corresponding to the winding needle. The winding needle can refer to a cylindrical component in the winding equipment that directly contacts and drives the electrode strip during winding; it is the core load-bearing and driving element for electrode winding.

[0090] In practical implementation, the winding equipment can be configured with multiple winding needles (such as 2 or 3) to achieve continuous production (while one winding needle is winding, another is ready, and there is no production interruption during switching). During the production process of a roll of cathode material, each winding needle will take turns participating in the electrode winding. When a winding needle is activated, it will drive the electrode to rotate and be wound into a battery cell. During switching, the cathode electrode is cut off to prepare for the winding of the next battery cell. Due to the differences in manufacturing, wear, and installation of each winding needle, the same set of correction models cannot be adapted to all winding needles (for example, a severely worn winding needle has a weaker clamping force on the electrode and requires different target values ​​for compensation).

[0091] When the current electrode strip reaches the starting position of the feeding and correction, the winding module will first determine the winding needle that the current electrode strip will use (such as the winding needle enable signal sent by the equipment PLC, or the detection result of the winding needle position sensor), and send the identification information of the winding needle (such as winding needle 1, winding needle 2) to the processor; after receiving it, the processor will complete the identification and acquisition of the winding needle corresponding to the current electrode strip.

[0092] The processor pre-stores preset correction models trained individually for each winding needle during the historical production of electrode strips. During the historical electrode strip production process, the processor collects data according to winding needle classification (e.g., qualified cell data produced by winding needle 1, qualified cell data produced by winding needle 2), and substitutes them into the linear model respectively. The processor trains dedicated models for both the 1st and 2nd curling needles. When the processor obtains the current curling needle identifier, it will directly call the model corresponding to that identifier (e.g., if the current curling needle is 1, then the dedicated model for 1st curling needle will be called), ensuring that the model parameters are completely matched with the characteristics of the current curling needle.

[0093] This embodiment obtains the winding needle corresponding to the current electrode strip and finds the preset correction model corresponding to the winding needle, thereby ensuring that the correction target value is completely matched with the winding needle, and thus improving the correction adaptability.

[0094] In some embodiments, the processor is further configured to acquire the historical correction value of the previous electrode strip and calculate the difference between the correction target value and the historical correction value; the correction mechanism is further configured to perform correction on the current electrode strip based on the correction target value if the difference is less than a preset threshold.

[0095] To ensure that the fluctuations of the current correction target value and historical control parameters are within the acceptable range of equipment and materials, and to maintain the stability of continuous production, this embodiment also obtains the historical correction value of the previous electrode strip and calculates the difference between the correction target value and the historical correction value. Only when the difference is less than a preset threshold is the current electrode strip corrected based on the correction target value.

[0096] The historical correction value refers to the target correction value of the cathode head of the previous electrode strip. This value is a verified and effective correction benchmark from the previous production run, recording the stable parameters of electrode position control in the previous production stage. It serves as a reference benchmark for the current cell correction target value, preventing excessive fluctuations between the current target value and historical control parameters. The preset threshold refers to the upper limit of the difference set by the system based on the physical properties of the winding equipment and the characteristics of the electrode material. It is the standard for judging whether the difference is within the controllable range of the equipment (e.g., set to ±0.1mm). Its value is derived from equipment debugging experience and material tolerance, and must be ensured to be lower than the critical value that may cause equipment damage or electrode head slippage, avoiding irreversible effects.

[0097] In its implementation, after generating the target value for the current electrode strip, the processor automatically retrieves the historical correction value of the previous electrode strip from the historical production data stored in the system. This value has been verified as effective in historical production and belongs to the same continuous production chain as the current cell, making it highly reliable for reference. Subsequently, the processor calculates the difference between the two using the formula: Difference = |Current Target Value - Historical Correction Value|. For example, if the current target value is 0.35mm and the historical correction value is 0.3mm, then the difference is 0.05mm.

[0098] After the processor sends the calculated difference to the correction mechanism, the correction mechanism first retrieves a preset threshold (such as ±0.1mm set by the system) and judges the difference: if the difference is less than the preset threshold (such as 0.05mm < 0.1mm as mentioned above), it means that the current target value fluctuates little with the historical control parameters and is within the range that the equipment can adjust smoothly, which will not cause sudden changes in the electrode position or overload the equipment. At this time, the correction mechanism receives the correction target value and drives the correction roller to move left and right along the electrode belt direction through the correction motor, adjusting the position of the current electrode belt to the baseline that matches the correction target value. After the electrode retainer enters the free end area, there is no head swing. If the difference is greater than or equal to the preset threshold (e.g., the current target value is 0.5mm, the historical correction value is 0.3mm, and the difference is 0.2mm ≥ 0.1mm), it indicates that the current target value and the historical parameters fluctuate too much, which may exceed the equipment's adjustment capability (e.g., the correction roller cannot move too far quickly) or cause the electrode to wrinkle or swing due to sudden position changes. In this case, the correction mechanism will temporarily not execute the target value and will send a feedback signal to the processor, which will then re-optimize the target value (e.g., further adjust it in conjunction with the upper and lower limits of the equipment) until the difference is less than the preset threshold before executing the correction.

[0099] This embodiment obtains the historical correction value of the previous electrode strip and calculates the difference between the correction target value and the historical correction value. Only when the difference is less than a preset threshold is the current electrode strip corrected based on the correction target value. This ensures that the fluctuation of the current correction target value and the historical control parameters are within the acceptable range of the equipment and materials, and maintains the stability of continuous production.

[0100] For ease of understanding, please refer to Figure 6 This explanation is provided, but does not limit the scope of this application. Figure 6 This is a flowchart illustrating the operation of an electrode correction system according to some embodiments of the present invention, as an example. Figure 6 The process includes the following steps: 1. Data collection during winding: During the winding and consumption of a roll of cathode material, collect the CCD result data of the cathode electrode head of each cell with a final CCD result of qualified, the correction sensor data at the beginning of the correction position, and the correction target value data.

[0101] 2. After a roll of cathode material is completely consumed, the program triggers an algorithm to acquire data on the cells produced from the previous roll. The model is trained using this acquired data, and a linear model is output. ,in, This is the CCD result data for the cathode electrode head. This refers to the data from the correction sensor at the start of the wafer feeding correction position. To correct the target value data, , For coefficients / weights, This is the model intercept.

[0102] 3. After the waste single roll at the head of the new cathode material roll is removed, the first double-roll electrode sheet material after the roll change reaches the starting position of the feeding and correction. At this time, the current correction sensor data, combined with the average value of the CCD result data of the qualified cells in the previous roll, is used to deduce a correction target value and output it to the control system. The specific steps are as follows: (1) Triggering algorithm model training when changing volumes: ,enter , , ,get , , ; (2) The transformed algorithm model used when the first double-rolled electrode strip arrives at the position: ,enter , , , , ,get .

[0103] 4. The system performs threshold judgment on the correction target value output by the algorithm model to avoid irreversible impact on equipment and materials caused by exceeding the controllable range of the control system.

[0104] 5. After receiving the predicted target value output by the algorithm, the PLC calculates the data by combining the previous cell's cathode head correction target value and the upper and lower limits of the controllable correction target value, and obtains the final correction target value to perform correction control on the cathode head of the first double-wound cell.

[0105] In some embodiments, such as Figure 7 As shown, an electrode correction method is proposed, including: Step S10: Obtain the current electrode edge position of the electrode strip and the electrode offset of the historical electrode strip in the free end region, wherein the electrode strip is the strip corresponding to an electrode of one cell length.

[0106] Step S20: Determine the correction target value based on the electrode edge position and the electrode offset using a preset correction model, wherein the preset correction model is a model trained based on historical correction information.

[0107] Step S30: Correct the current electrode strip based on the correction target value.

[0108] In this embodiment, the electrode correction method can be applied to an electrode correction system. The electrode correction system can refer to a closed-loop control system used in lithium battery winding production to control the position of the cathode electrode and prevent the electrode head from swinging before entering the winding equipment. Its core is to ensure the consistency of the OH (the portion of the negative electrode extending beyond the positive electrode) at the electrode head through a detection-calculation-adjustment logic, reducing the risk of lithium plating and lithium dendrite formation. The current electrode strip can refer to the first double-wound cell of a new cathode roll after a roll of cathode material has been completely consumed (no remaining electrode strips to be wound) and replaced with a new cathode roll, after the waste material at the head of the roll has been removed. Alternatively, it can refer to the cell that subsequently enters the winding needle after the previous electrode strip has been cut; this embodiment does not limit this. A double-wound cell can refer to a cell obtained by winding the cathode and anode electrodes together in the winding needle.

[0109] The free end region can refer to the area without a load after the electrode passes through the correction mechanism (such as the correction roller) and the correction sensor and before it reaches the winding equipment (winding needle). The electrode is in a suspended state in this region. If it is not controlled, it is easy to swing its head. This is the core working area of ​​the correction system.

[0110] For ease of understanding, please refer to Figure 1 and Figure 2 This explanation is provided, but does not limit the scope of this application. Figure 1 This is a top view of an electrode correction system according to some embodiments of the present invention. Figure 2 This is a front view of an electrode correction system according to some embodiments of the present invention. As an example, such as... Figure 1 and Figure 2As shown, the electrode web guiding system includes: a processor 100 and a web guiding mechanism 30, a web guiding sensor 40, an image acquisition device 50, and a winding device 60 arranged sequentially according to the electrode's belt travel direction. The processor 100 can train a preset web guiding model based on historical data, and use real-time electrode position data to deduce the web guiding target value. It can also optimize the target value within the controllable range of the device to avoid damage caused by exceeding the device's capabilities. In a specific implementation, the processor 100 can be a programmable logic controller (PLC). The web guiding mechanism 30 can include core components such as a web guiding roller and a web guiding motor. It can move left and right along the electrode's belt travel direction, adjusting the left and right position of the electrode roll 20 according to the web guiding target value output by the processor 100, keeping the electrode roll 20 at a set reference position. The web guiding sensor 40 can accurately detect the edge position of the current electrode strip 20 on the slitting side, providing real-time position data to the processor for target value calculation. Image acquisition device 50 may refer to a charge-coupled device (CCD) that can detect electrode offset in real time. Winding equipment 60 may include core components such as winding needles, used to wind the calibrated electrode roll 20 into a battery cell. Free end region 70 (e.g. Figure 1 The area within the dashed box (in the middle) refers to the unsupported area between the electrode roll 20 and the winding device 60 (winding needle) after passing through the correction mechanism 30 (correction roller) and the correction sensor 40. The electrode roll 20 is suspended in this area; if uncontrolled, it is prone to tipping. This is the core area of ​​the correction system. The electrode edge position refers to the current position of the electrode slitting side, detected in real-time by the correction sensor 40. This is the real-time basis for the processor to determine the degree of electrode offset and calculate the correction target value, directly affecting the accuracy of the target value. The electrode offset can be the deviation between the actual position (e.g., the electrode slitting side) of the historical electrode strip in the free end area, obtained by the image acquisition unit 50 (winding CCD), and the set position (e.g., the baseline). The preset correction model refers to a linear model trained based on historical correction information of cells with acceptable winding CCD results, which can be used to deduce the correction target value of the current cell. The correction target value can refer to the electrode position adjustment value calculated by the processor based on the preset correction model, the real-time electrode edge position, and the historical electrode offset. It is the basis for the correction mechanism to adjust the electrode position. Finally, it can be optimized by the PLC in conjunction with the upper and lower limits of the equipment to ensure that it is within a controllable range.

[0111] This embodiment calculates the correction target value based on the current electrode edge position of the electrode strip and the electrode offset of the historical electrode strip in the free end region through a pre-trained preset correction model, thereby improving the accuracy of electrode correction and thus improving the electrode correction effect.

[0112] In some embodiments, such as Figure 8 As shown, the historical correction information includes the historical electrode edge position, historical electrode offset, and historical correction target value of the historical electrode material strip. The historical electrode material strip is the electrode material strip whose correction result is qualified. Before step S10, the following steps are also included: Step S01: Obtain the historical electrode edge position, historical electrode offset, and historical correction target value of the historical electrode strip.

[0113] Step S02: Train a correction model based on the historical electrode edge position, the historical electrode offset, and the historical correction target value to obtain a preset correction model.

[0114] To ensure the prediction accuracy and improve the reliability of the preset correction model, in this embodiment, the correction model is trained based on the historical electrode edge position, historical electrode offset, and historical correction target value to obtain the preset correction model.

[0115] The historical electrode edge position can refer to the electrode slitting side position data collected by the correction sensor at the start of the electrode feeding correction, i.e., the parameters in the linear model. It reflects the initial position of the historical electrode strip when it enters the correction stage, and is a key input for linking the initial position and final offset in model training. The historical electrode offset can refer to the deviation between the actual position and the set position of the historical electrode strip in the free end region, as acquired by the image acquisition device (wound CCD), i.e., a parameter in the linear model. It directly reflects the final offset state of the historical electrode strip and is the core basis for judging the correction effect during model training. The historical correction target value can refer to the correction benchmark value set when the historical electrode strip was wound, i.e., the parameter in the linear model. It serves as a target benchmark used to control electrode position during historical production, recording the correspondence between the set target and the actual offset, and is a crucial input parameter for model training. The preset correction model can refer to a model based on historical correction information (…). , , The linear mathematical model or other models obtained through training are not limited in this embodiment.

[0116] In practice, during the continuous winding and consumption of the previous roll of cathode material, each component synchronously collects historical data of qualified electrode sheets according to its division of labor: for each cell that is finally detected as OK by CCD, the correction sensor collects the edge position of the historical electrode sheet strip when its electrode sheet reaches the starting position of the sheet feeding correction. The image acquisition unit (wound CCD) simultaneously acquires the offset of the qualified cell electrode in the free end region (the position of the historical electrode edge) and sends it to the processor; the image acquisition unit (wound CCD) simultaneously acquires the offset of the qualified cell electrode in the free end region (the position of the historical electrode edge) and sends it to the processor; The historical electrode offset is recorded and transmitted to the processor. Simultaneously, the processor actively retrieves the historical correction target value actually executed by the control system during the winding of the qualified battery cell. Complete the single set of historical correction information ( , , The collection of ).

[0117] When the current roll of cathode material is completely used up and a new roll is about to be replaced, the program automatically triggers the algorithm, and the processor retrieves the historical correction information (multiple sets) of all qualified cells from the previous roll. The data (historical correction information) is used as training samples to train the correction model, ultimately obtaining a preset correction model that is supported by historical data and can accurately associate "electrode position - offset - correction target value". During training, the processor automatically filters qualified data that are detected as OK by CCD and removes unqualified data to ensure the validity of the model training data. At the same time, it distinguishes between different winding needles (because each winding needle is different) to train the model separately, further improving the model's specificity.

[0118] This embodiment trains a correction model based on historical electrode edge positions, historical electrode offsets, and historical correction target values ​​to obtain a preset correction model, thereby ensuring the prediction accuracy of the preset correction model and improving its reliability.

[0119] In some embodiments, training a correction model based on the historical electrode edge position, the historical electrode offset, and the historical correction target value to obtain a preset correction model includes: solving for parameters based on the historical electrode edge position, the historical electrode offset, and the historical correction target value to determine the electrode edge position adjustment coefficient, the electrode offset adjustment coefficient, and the model intercept; and constructing the preset correction model based on the electrode edge position adjustment coefficient, the electrode offset adjustment coefficient, and the model intercept.

[0120] In order to accurately predict the correspondence between electrode offset and correction target value and improve the accuracy of the preset correction model, in this embodiment, parameters are solved based on historical electrode edge position, historical electrode offset and historical correction target value to determine electrode edge position adjustment coefficient, electrode offset adjustment coefficient and model intercept, and a preset correction model is constructed based on electrode edge position adjustment coefficient, electrode offset adjustment coefficient and model intercept.

[0121] If the preset correction model is a linear mathematical model, then the expression is: (in, This represents the historical polarimetric offset. This is the edge of a historical polar sheet. For historical correction target values, This is the electrode edge position adjustment coefficient. This is the electrode offset adjustment coefficient. (This refers to the model intercept), its core function is to infer the target value for the new battery cell's polarization correction by analyzing historical data patterns. The electrode edge position adjustment coefficient can refer to the coefficient in the preset polarization correction model. It is a measure of the position of the edge of historical polar sheets ( ) on historical electrode offset ( The weighting parameters for the degree of influence are calculated by the processor based on historical data, and their values ​​reflect the strength of the correlation between the initial edge position of the electrode and the final correction effect. The electrode offset adjustment coefficient can refer to a coefficient in the preset correction model. It is a measure of historical correction target value ( ) on historical electrode offset ( The weight parameters for the degree of influence are also obtained by the processor through historical data. The model intercept can refer to parameter b in the preset correction model. It is the basic offset compensation value of the linear model, which is obtained by the processor by fitting historical data. It is used to correct the system deviation caused by fixed factors such as inherent equipment errors and slight differences in electrode materials, thereby improving the accuracy of model prediction.

[0122] In the specific implementation, when a roll of cathode material is completely consumed and a new roll is about to be replaced, the processor automatically triggers the parameter solving process. First, it retrieves the historical correction information of all qualified cells in the roll that have been detected as OK by CCD, including the historical electrode edge position of each group of cells. ), historical polarimetric offset ( ) and historical correction target values ​​( The processor automatically removes unqualified data to ensure the validity of the input data. At the same time, the processor distinguishes the winding needles currently used in production (because each winding needle has different structures and precision), and filters the corresponding historical data for different winding needles to avoid parameter deviations caused by differences in winding needles.

[0123] The processor will filter multiple sets of historical data ( , , Input the preset multivariate linear regression algorithm, using a linear model To fit the target, the algorithm minimizes the difference between the predicted value and the actual historical pole offset. The error of the electrode edge position adjustment coefficient is then calculated by determining the error of the electrode edge position adjustment coefficient. ), electrode offset adjustment coefficient ( ) and model intercept ( ).in, The solution reflects right Influence weight, The solution reflects right Influence weight, The solution compensates for the inherent biases of the system, ultimately yielding a set of parameter values ​​that accurately fit the patterns in historical data. Finally, when , , Once the value is determined, the preset correction model is trained and can be used to calculate the correction target value of the current cell electrode.

[0124] When the current cell electrode reaches the starting position of the electrode feeding and correction, the correction sensor collects the current electrode edge position in real time (i.e., the current electrode edge position). ) and send it to the processor; the processor first obtains the historical electrode offset of the battery cell ( The mean of ) and then Mean, Current and well-trained , , Substitute into the transformed model The initial correction target value is obtained by reverse calculation. The processor then performs a threshold judgment on the initial target value, excluding values ​​that are outside the controllable range of the device, and obtains the correction target value.

[0125] This embodiment solves for parameters based on historical electrode edge positions, historical electrode offsets, and historical correction target values ​​to determine electrode edge position adjustment coefficients, electrode offset adjustment coefficients, and model intercepts. Based on these parameters, a preset correction model is constructed, which can accurately predict the correspondence between electrode offsets and correction target values, thereby improving the accuracy of the preset correction model.

[0126] In some embodiments, such as Figure 9 As shown, before step S20, the procedure further includes: Step S11: Determine whether the current electrode strip and the previous electrode strip belong to the same correction control group, wherein the correction control group includes a preset number of electrode strips.

[0127] Step S12: If the current electrode strip and the previous electrode strip belong to the same correction control group, then the correction model of the previous electrode strip is used as the preset correction model.

[0128] To reduce the number of modeling iterations, decrease processor computation time, and improve correction efficiency, in this embodiment, when correcting the current electrode strip, it is first determined whether the current electrode strip and the previous electrode strip belong to the same correction control group. If the current electrode strip and the previous electrode strip belong to the same correction control group, there is no need to retrain the correction model; instead, the correction model of the previous electrode strip is directly used as the preset correction model.

[0129] The preceding electrode strip refers to the immediately preceding cathode electrode strip that has been wound and whose winding CCD inspection result is qualified. It serves as the benchmark for determining the current cell group's affiliation. Its correction model is trained based on historical qualified data from its control group and has been verified to effectively control head offset, providing a basis for model reuse for the current cells in the same group. The correction control group refers to a production unit divided according to a preset number of consecutive qualified CCD inspection electrode strips, based on the stable and predictable nature of continuous qualified cell data in lithium battery winding production. Within the same group, the production environment of the electrode strips (such as winding needle wear, electrode tension, and equipment operating parameters) remains largely unchanged, and the electrode offset (…) ), correction sensor data ( The data patterns are consistent, and the same correction model is used to avoid model failure caused by occasional fluctuations in a single cell. The preset quantity can refer to the number of qualified CCD electrode strips that constitute a correction control group, which is a quantitative standard for dividing control groups.

[0130] For ease of understanding, please refer to Figure 3 This explanation is provided, but does not limit the scope of this application. Figure 3 This is a schematic diagram of the correction control group according to some embodiments of the present invention, as an example. Figure 3 In the winding sequence, the cells are: Cell 1, Cell 2, Cell 3, Cell 4, Cell 5, Cell 6, Cell 7, and Cell 8. Assuming the spin correction control group contains three electrode strips, and the current electrode strip is Cell 8, and Cell 8 belongs to the same spin correction control group as the previous electrode strip (Cell 7), Cell 8 can directly use the spin correction model of Cell 7 as the preset spin correction model to calculate the spin correction target value. The spin correction model of Cell 7 is essentially the spin correction model of the previous electrode strip (Cell 6), meaning that Cells 6, 7, and 8 essentially share the same spin correction model, thus reducing the number of modeling iterations. The spin correction model of Cell 6 is trained based on the historical electrode edge positions, historical electrode offsets, and historical spin correction target values ​​of Cells 1, 2, 3, 4, and 5.

[0131] In this embodiment, when correcting the current electrode strip, it first determines whether the current electrode strip and the previous electrode strip belong to the same correction control group. If the current electrode strip and the previous electrode strip belong to the same correction control group, there is no need to retrain the correction model. Instead, the correction model of the previous electrode strip is directly used as the preset correction model, which can reduce the number of modeling times, reduce the processor's computing time, and thus improve the correction efficiency.

[0132] In some embodiments, such as Figure 10 As shown, after step S11, the method further includes: Step S12': If the current electrode strip and the previous electrode strip do not belong to the same correction control group, then obtain the qualified candidate electrode strips with correction results, and sort the candidate electrode strips according to the winding order to obtain the sorting result.

[0133] Step S13': Select a preset number of historical electrode strips from the candidate electrode strips according to the sorting results, and train a correction model based on the historical electrode edge position, historical electrode offset and historical correction target value of the historical electrode strips to obtain a preset correction model.

[0134] To avoid target value deviations caused by reusing old models from previous cells that do not belong to the same correction control group, and to ensure model accuracy, in this embodiment, if the current electrode strip and the previous electrode strip do not belong to the same correction control group, the electrode strips with qualified correction results are obtained, and the electrode strips are sorted according to the winding order to obtain the sorting results. Based on the sorting results, a preset number of historical electrode strips are selected from the electrode strips, and the correction model is trained based on the historical electrode edge position, historical electrode offset, and historical correction target value of the historical electrode strips to obtain the preset correction model.

[0135] The candidate electrode strips refer to cathode electrodes selected from historical production data that have passed CCD winding inspection. These serve as the foundation for training the new correction model. Selecting only qualified cells avoids interference from anomalous data (such as cells with offsets exceeding ±0.5mm) that could affect model accuracy, ensuring the validity of the model's input data. The winding sequence refers to the actual winding order of the candidate electrode strips during historical production (e.g., production sequence 1, 2, 3…). Sorting in this order ensures that the selected historical electrode strips reflect data variation patterns under continuous production conditions, preventing the model from being unsuitable for the current continuous winding scenario due to random sampling. The sorting result refers to the ordered data list formed by arranging the candidate electrode strips in winding order. This serves as the data pool for subsequently selecting a preset number of historical cells, ensuring the selected samples have temporal continuity and are consistent with the current continuous production scenario for cells. The preset sample size refers to the number of historical electrode strips used to train the new correction model, serving as the standard for the amount of data for model training.

[0136] For ease of understanding, please refer to Figure 4 This explanation is provided, but does not limit the scope of this application. Figure 4 This is a schematic diagram of the correction control group according to some embodiments of the present invention, as an example. Figure 4 In the process, the cells are arranged in the winding sequence as follows: Cell 1, Cell 2, Cell 3, Cell 4, Cell 5, Cell 6, Cell 7, Cell 8, Cell 9, Cell 10, and Cell 11. Assuming the correction control group contains three electrode strips, and the current electrode strip is Cell 9, when the processor determines that the current electrode strip (Cell 9) does not belong to the same correction control group as the previous electrode strip (Cell 8), the processor first selects all cathode electrodes with qualified winding CCD detection results from historical production data as candidate electrode strips. Subsequently, the candidate electrode strips are sorted according to their winding sequence in historical production (e.g., production sequence number from earliest to latest), forming an ordered list of candidate electrode strips (e.g., sorting result: Cell 1 → Cell 2 → Cell 3 → …, corresponding to production sequence number 1 → 2 → 3 → …), ensuring that the subsequently selected samples reflect the data patterns under continuous production conditions. The processor selects samples from the sorted list of candidate electrode strips based on a preset number of samples (e.g., 5), prioritizing samples from the winding sequence. It can select consecutive samples closest to the current cell production sequence (e.g., cells 4, 5, 6, 7, and 8), as the production environment (e.g., winding needle wear, electrode tension) is closest to the current cell's new group environment, maximizing model adaptability. The processor extracts three sets of core data from the selected historical electrode strips, including the historical electrode edge position (i.e.,...). ), historical polarimetric offset (i.e. The historical correction target value (i.e., x2) is used to substitute the average value of each group of data into the linear model formula. The optimal coefficients are calculated using data fitting algorithms (such as the least squares method). , and intercept That is, by adjusting , , The value is chosen to minimize the error between the y-value calculated by the model and the actual historical electrode offset y (e.g., controlling the error within ±0.05mm). When , , Once determined, the new preset correction model is trained and can be used to calculate the correction target value of the current electrode strip.

[0137] In this embodiment, if the current electrode strip and the previous electrode strip do not belong to the same correction control group, the electrode strip with the corrected correction result is selected as a qualified candidate electrode strip. The candidate electrode strips are sorted according to the winding order to obtain a sorting result. Based on the sorting result, a preset number of historical electrode strips are selected from the candidate electrode strips. The correction model is trained based on the historical electrode edge position, historical electrode offset, and historical correction target value of the historical electrode strips to obtain a preset correction model. This can avoid the target value deviation caused by reusing the old model of the previous cell that does not belong to the same correction control group, and ensure the model accuracy.

[0138] In some embodiments, before step S11, the method further includes: randomly generating multiple numerical groups and calculating the predicted correction value corresponding to each numerical group, wherein the numerical group includes the number of electrode strips in the correction control group and the number of samples selected from the candidate electrode strips; obtaining the actual correction value, and determining the preset quantity and the preset number of samples based on the residual between the predicted correction value and the actual correction value.

[0139] To ensure both model accuracy and correction efficiency, this embodiment randomly generates multiple numerical groups and calculates the predicted correction value for each numerical group. The numerical group includes the number of electrode strips in the correction control group and the number of samples selected from the candidate electrode strips. The actual correction value is obtained, and the preset quantity and preset sample number are determined based on the residual between the predicted correction value and the actual correction value.

[0140] In this context, a numerical set refers to a combined unit constructed by the processor to select optimal parameters. It consists of two core parameters: first, the number of electrode strips in the correction control group (i.e., the basic number for dividing the correction control group, also known as the sliding window size); and second, the number of samples selected from the candidate electrode strips (the number of historically qualified cells used to train the correction model). Different combinations of these two parameter values ​​form multiple numerical sets (e.g., "sample count = 5 + sliding window size = 3", "sample count = 6 + sliding window size = 4"), providing a candidate set of parameters for subsequent residual comparison. The predicted correction value can refer to the target correction value of the electrode strip calculated based on the correction model corresponding to a certain numerical set. Specifically, the processor first selects candidate electrode strips (historical cells that passed CCD detection) based on the number of samples in the numerical set, and then uses these samples... (Edge position of historical polar image) (Historical polarity offset) (Historical electrode correction values) Training a linear model Then, the model is applied to the cells corresponding to the number of control cells in the numerical group, and the output is... The predicted correction value reflects the model's control effect on this numerical group. The actual correction value can refer to the target correction value that the candidate electrode strip has actually used in historical production and has been verified to effectively control the electrode head offset. This value is the benchmark for judging the accuracy of the predicted correction value. Since this value has achieved electrode head-free operation and qualified OH in historical production, it can be used as the true value for residual calculation. The residual can refer to the difference between the predicted correction value and the actual correction value of the same electrode strip (which can be the absolute value or the difference of squares). It is the core indicator for measuring the model accuracy corresponding to the numerical group. The smaller the residual, the more the parameter combination of the numerical group (number of cells in the control group, number of samples) is adapted to the production pattern, and the more accurate the model prediction is. The preset quantity can refer to the number of electrode strips in the final determined correction control group, which is a fixed standard for dividing the correction control group. The processor compares the residuals of all numerical groups and selects the number of cells in the control group of the numerical group with the smallest residual as the preset quantity to ensure that the data pattern within the same control group is stable and adapted to the same model. The preset sample size can refer to the final number of samples selected from the candidate electrode strips, which is a fixed sample size for training the preset correction model. Similarly, it is determined by residual optimization, selecting the sample size from the group of values ​​with the smallest residual as the preset sample size, ensuring that the samples for training the model can both cover stable production patterns and avoid computational redundancy caused by too many samples.

[0141] For ease of understanding, please refer to Figure 5 This explanation is provided, but does not limit the scope of this application. Figure 5 This is a schematic diagram of multiple numerical groups for some embodiments of the present invention. Figure 5In the process, the processor first determines the value ranges of two types of parameters based on the production experience of the previous cathode material roll. The first is the candidate value for the number of electrode strips in the correction control group (e.g., 5, 8, 10, balancing data stability and computational efficiency). The second is the candidate value for the number of samples selected from the candidate electrode strips (e.g., 3, 5, 8, ensuring the samples can fit an effective linear model). Then, the candidate values ​​of both types of parameters are fully combined to randomly generate multiple numerical sets, for example: First set of data: number of samples = 5, sliding window size = 3; Second set of data: number of samples = 6, sliding window size = 4; ...... Group n data: number of samples = n, sliding window size = n (to ensure that all possible reasonable combinations are covered and to avoid missing the optimal parameters).

[0142] For each set of values, the processor performs a three-step operation: "sampling-modeling-prediction". 1. Sample selection: From the candidate electrode strips (the previous roll of qualified CCD cells), select consecutive historical electrode strips according to the number of samples (e.g., if the number of samples in the first group of data is 5, then select the 5 qualified cells in the previous roll whose timing is closest to the current time). 2. Modeling: Extracting the selected samples , , Data, substituted into a linear model The specific values ​​for this set of values ​​can be calculated through data fitting (such as least squares method). , , Complete model training; 3. Prediction: Apply this model to the cells corresponding to the number of cells in the control group in the numerical group (e.g., if the sliding window size of the first group of data is 3, then take 3 consecutive cells), and input these cells... and ( Take samples (mean), through the transformation formula Calculate the predicted correction value for each battery cell.

[0143] The processor first retrieves the actual correction value (i.e., historical x 2) of the cells corresponding to the number of cells in the control group from historical data. Then, it calculates the average residual between the predicted correction value and the actual correction value of all cells in each numerical group (e.g., the residuals of the three cells in the first data group are 0.01mm, 0.02mm, ..., with an average of 0.015mm). Finally, it compares the average residuals of all numerical groups and selects the numerical group with the smallest residual. For example, the first data group (sample size = 5, sliding window size = 3) has the smallest average residual of 0.015mm. Therefore, the sliding window size = 3 for this numerical group is determined as the preset number, and the sample size = 5 is determined as the preset sample size.

[0144] In this embodiment, multiple numerical groups are randomly generated, and the predicted correction value corresponding to each numerical group is calculated. The numerical group includes the number of electrode strips in the correction control group and the number of samples selected from the candidate electrode strips. The actual correction value is obtained, and the preset quantity and preset sample number are determined according to the residual between the predicted correction value and the actual correction value, so as to ensure both model accuracy and correction efficiency.

[0145] In some embodiments, before step S20, the method further includes: obtaining the winding needle corresponding to the current electrode strip and finding the preset correction model corresponding to the winding needle, wherein different winding needles correspond to different preset correction models.

[0146] To ensure a perfect match between the correction target value and the winding needle and improve the adaptability of the correction, this embodiment obtains the winding needle corresponding to the current electrode strip and searches for the preset correction model corresponding to the winding needle. The winding needle can refer to a cylindrical component in the winding equipment that directly contacts and drives the electrode strip during winding; it is the core load-bearing and driving element for electrode winding.

[0147] In practical implementation, the winding equipment can be configured with multiple winding needles (such as 2 or 3) to achieve continuous production (while one winding needle is winding, another is ready, and there is no production interruption during switching). During the production process of a roll of cathode material, each winding needle will take turns participating in the electrode winding. When a winding needle is activated, it will drive the electrode to rotate and be wound into a battery cell. During switching, the cathode electrode is cut off to prepare for the winding of the next battery cell. Due to the differences in manufacturing, wear, and installation of each winding needle, the same set of correction models cannot be adapted to all winding needles (for example, a severely worn winding needle has a weaker clamping force on the electrode and requires different target values ​​for compensation).

[0148] When the current electrode strip reaches the starting position of the feeding and correction, the winding module will first determine the winding needle that the current electrode strip will use (such as the winding needle enable signal sent by the equipment PLC, or the detection result of the winding needle position sensor), and send the identification information of the winding needle (such as winding needle 1, winding needle 2) to the processor; after receiving it, the processor will complete the identification and acquisition of the winding needle corresponding to the current electrode strip.

[0149] The processor pre-stores preset correction models trained individually for each winding needle during the historical production of electrode strips. During the historical electrode strip production process, the processor collects data according to winding needle classification (e.g., qualified cell data produced by winding needle 1, qualified cell data produced by winding needle 2), and substitutes them into the linear model respectively. The processor trains dedicated models for both the 1st and 2nd curling needles. When the processor obtains the current curling needle identifier, it will directly call the model corresponding to that identifier (e.g., if the current curling needle is 1, then the dedicated model for 1st curling needle will be called), ensuring that the model parameters are completely matched with the characteristics of the current curling needle.

[0150] This embodiment obtains the winding needle corresponding to the current electrode strip and finds the preset correction model corresponding to the winding needle, thereby ensuring that the correction target value is completely matched with the winding needle, and thus improving the correction adaptability.

[0151] In some embodiments, such as Figure 11 As shown, before step S30, the procedure further includes: Step S21: Obtain the historical correction value of the previous electrode strip and calculate the difference between the correction target value and the historical correction value; Accordingly, step S30 includes: Step S30': If the difference is less than a preset threshold, then the current electrode strip is corrected based on the correction target value.

[0152] To ensure that the fluctuations of the current correction target value and historical control parameters are within the acceptable range of equipment and materials, and to maintain the stability of continuous production, this embodiment also obtains the historical correction value of the previous electrode strip and calculates the difference between the correction target value and the historical correction value. Only when the difference is less than a preset threshold is the current electrode strip corrected based on the correction target value.

[0153] The historical correction value refers to the target correction value of the cathode head of the previous electrode strip. This value is a verified and effective correction benchmark from the previous production run, recording the stable parameters of electrode position control in the previous production stage. It serves as a reference benchmark for the current cell correction target value, preventing excessive fluctuations between the current target value and historical control parameters. The preset threshold refers to the upper limit of the difference set by the system based on the physical properties of the winding equipment and the characteristics of the electrode material. It is the standard for judging whether the difference is within the controllable range of the equipment (e.g., set to ±0.1mm). Its value is derived from equipment debugging experience and material tolerance, and must be ensured to be lower than the critical value that may cause equipment damage or electrode head slippage, avoiding irreversible effects.

[0154] In its implementation, after generating the target value for the current electrode strip, the processor automatically retrieves the historical correction value of the previous electrode strip from the historical production data stored in the system. This value has been verified as effective in historical production and belongs to the same continuous production chain as the current cell, making it highly reliable for reference. Subsequently, the processor calculates the difference between the two using the formula: Difference = |Current Target Value - Historical Correction Value|. For example, if the current target value is 0.35mm and the historical correction value is 0.3mm, then the difference is 0.05mm.

[0155] After the processor sends the calculated difference to the correction mechanism, the correction mechanism first retrieves a preset threshold (such as ±0.1mm set by the system) and judges the difference: if the difference is less than the preset threshold (such as 0.05mm < 0.1mm as mentioned above), it means that the current target value fluctuates little with the historical control parameters and is within the range that the equipment can adjust smoothly, which will not cause sudden changes in the electrode position or overload the equipment. At this time, the correction mechanism receives the correction target value and drives the correction roller to move left and right along the electrode belt direction through the correction motor, adjusting the position of the current electrode belt to the baseline that matches the correction target value. After the electrode retainer enters the free end area, there is no head swing. If the difference is greater than or equal to the preset threshold (e.g., the current target value is 0.5mm, the historical correction value is 0.3mm, and the difference is 0.2mm ≥ 0.1mm), it indicates that the current target value and the historical parameters fluctuate too much, which may exceed the equipment's adjustment capability (e.g., the correction roller cannot move too far quickly) or cause the electrode to wrinkle or swing due to sudden position changes. In this case, the correction mechanism will temporarily not execute the target value and will send a feedback signal to the processor, which will then re-optimize the target value (e.g., further adjust it in conjunction with the upper and lower limits of the equipment) until the difference is less than the preset threshold before executing the correction.

[0156] This embodiment obtains the historical correction value of the previous electrode strip and calculates the difference between the correction target value and the historical correction value. Only when the difference is less than a preset threshold is the current electrode strip corrected based on the correction target value. This ensures that the fluctuation of the current correction target value and the historical control parameters are within the acceptable range of the equipment and materials, and maintains the stability of continuous production.

[0157] For ease of understanding, please refer to Figure 6 This explanation is provided, but does not limit the scope of this application. Figure 6 This is a flowchart illustrating the operation of an electrode correction system according to some embodiments of the present invention, as an example. Figure 6 The process includes the following steps: 1. Data collection during winding: During the winding and consumption of a roll of cathode material, collect the CCD result data of the cathode electrode head of each cell with a final CCD result of qualified, the correction sensor data at the beginning of the correction position, and the correction target value data.

[0158] 2. After a roll of cathode material is completely consumed, the program triggers an algorithm to acquire data on the cells produced from the previous roll. The model is trained using this acquired data, and a linear model is output. ,in, This is the CCD result data for the cathode electrode head. This refers to the data from the correction sensor at the start of the wafer feeding correction position. To correct the target value data, , For coefficients / weights, This is the model intercept.

[0159] 3. After the waste single roll at the head of the new cathode material roll is removed, the first double-roll electrode sheet material after the roll change reaches the starting position of the feeding and correction. At this time, the current correction sensor data, combined with the average value of the CCD result data of the qualified cells in the previous roll, is used to deduce a correction target value and output it to the control system. The specific steps are as follows: (1) Triggering algorithm model training when changing volumes: ,enter , , ,get , , ; (2) The transformed algorithm model used when the first double-rolled electrode strip arrives at the position: ,enter , , , , ,get .

[0160] 4. The system performs threshold judgment on the correction target value output by the algorithm model to avoid irreversible impact on equipment and materials caused by exceeding the controllable range of the control system.

[0161] 5. After receiving the predicted target value output by the algorithm, the PLC calculates the data by combining the previous cell's cathode head correction target value and the upper and lower limits of the controllable correction target value, and obtains the final correction target value to perform correction control on the cathode head of the first double-wound cell.

[0162] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. These modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention. In particular, as long as there is no structural conflict, the various technical features mentioned in the embodiments can be combined in any way. The present invention is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. An electrode correction system, characterized in that, The electrode correction system includes: a processor and a correction mechanism, a correction sensor, an image acquisition device, and a winding device arranged sequentially according to the electrode travel direction; The correction sensor is used to obtain the electrode edge position of the current electrode strip and send the electrode edge position to the processor, wherein the electrode strip is the strip corresponding to an electrode of one cell length; The image acquisition device is used to acquire the electrode offset of the historical electrode strip in the free end region and send the electrode offset to the processor, wherein the free end region is the region between the correction sensor and the winding device; The processor is configured to determine a correction target value based on the electrode edge position and the electrode offset using a preset correction model, wherein the preset correction model is a model trained based on historical correction information. The correction mechanism is used to correct the current electrode strip based on the correction target value.

2. The electrode correction system as described in claim 1, characterized in that, The historical correction information includes the historical electrode edge position, historical electrode offset, and historical correction target value of the historical electrode material strip. The historical electrode material strip is the electrode material strip whose correction result is qualified. The correction sensor is also used to obtain the historical electrode edge position of the historical electrode strip and send the historical electrode edge position to the processor; The image acquisition device is also used to acquire the historical electrode offset of the historical electrode strip and send the historical electrode offset to the processor; The processor is further configured to acquire the historical correction target value of the historical electrode strip, and train a correction model based on the historical electrode edge position, the historical electrode offset, and the historical correction target value to obtain a preset correction model.

3. The electrode correction system as described in claim 2, characterized in that, The processor is also used to solve for parameters based on the historical electrode edge position, the historical electrode offset, and the historical correction target value, and to determine the electrode edge position adjustment coefficient, the electrode offset adjustment coefficient, and the model intercept. The processor is also used to construct a preset correction model based on the electrode edge position adjustment coefficient, the electrode offset adjustment coefficient, and the model intercept.

4. The electrode correction system as described in claim 1, characterized in that, The processor is further configured to determine whether the current electrode strip and the previous electrode strip belong to the same correction control group, wherein the correction control group includes a preset number of electrode strips; The processor is further configured to use the correction model of the previous electrode strip as a preset correction model if the current electrode strip and the previous electrode strip belong to the same correction control group.

5. The electrode correction system as described in claim 4, characterized in that, The processor is further configured to, if the current electrode strip and the previous electrode strip do not belong to the same correction control group, obtain a qualified candidate electrode strip with a correction result, and sort the candidate electrode strips according to the winding order to obtain a sorting result. The processor is further configured to select a preset number of historical electrode strips from the candidate electrode strips according to the sorting result, and train a correction model based on the historical electrode edge position, historical electrode offset and historical correction target value of the historical electrode strips to obtain a preset correction model.

6. The electrode correction system as described in claim 5, characterized in that, The processor is further configured to randomly generate multiple numerical groups and calculate the predicted correction value corresponding to each numerical group, wherein the numerical group includes the number of electrode strips in the correction control group and the number of samples selected from the candidate electrode strips. The processor is further configured to acquire the actual correction value and determine the preset quantity and the preset sample number based on the residual between the predicted correction value and the actual correction value.

7. The electrode correction system as described in any one of claims 1 to 6, characterized in that, The winding equipment includes a winding needle; The processor is also used to obtain the winding needle corresponding to the current electrode strip and to find the preset correction model corresponding to the winding needle, wherein different winding needles correspond to different preset correction models.

8. The electrode correction system as described in any one of claims 1 to 6, characterized in that, The processor is also configured to acquire the historical correction value of the previous electrode strip and calculate the difference between the correction target value and the historical correction value. The correction mechanism is further configured to correct the current electrode strip based on the correction target value if the difference is less than a preset threshold.

9. A method for correcting electrode polarization, characterized in that, include: The electrode edge position of the current electrode strip and the electrode offset of the historical electrode strip in the free end region are obtained. The electrode strip is the strip corresponding to an electrode of one cell length, and the free end region is the area between the correction sensor and the winding device. The correction target value is determined by a preset correction model based on the electrode edge position and the electrode offset, wherein the preset correction model is a model trained based on historical correction information. The current electrode strip is corrected based on the correction target value.

10. The electrode correction method as described in claim 9, characterized in that, The historical correction information includes the historical electrode edge position, historical electrode offset, and historical correction target value of the historical electrode material strip. The historical electrode material strip is the electrode material strip whose correction result is qualified. Before obtaining the current electrode edge position and the historical electrode offset in the free end region, the method further includes: Obtain the historical electrode edge position, historical electrode offset, and historical correction target value of the historical electrode strip; A correction model is trained based on the historical electrode edge position, the historical electrode offset, and the historical correction target value to obtain a preset correction model.

11. The electrode correction method as described in claim 10, characterized in that, The step of training a correction model based on the historical electrode edge position, the historical electrode offset, and the historical correction target value to obtain a preset correction model includes: Based on the historical electrode edge position, the historical electrode offset, and the historical correction target value, the parameters are solved to determine the electrode edge position adjustment coefficient, the electrode offset adjustment coefficient, and the model intercept. A preset correction model is constructed based on the electrode edge position adjustment coefficient, the electrode offset adjustment coefficient, and the model intercept.

12. The electrode correction method as described in claim 9, characterized in that, Before determining the correction target value based on the electrode edge position and the electrode offset using a preset correction model, the method further includes: Determine whether the current electrode strip and the previous electrode strip belong to the same correction control group, wherein the correction control group includes a preset number of electrode strips; If the current electrode strip and the previous electrode strip belong to the same correction control group, then the correction model of the previous electrode strip is used as the preset correction model.

13. The electrode correction method as described in claim 12, characterized in that, After determining whether the current electrode strip and the previous electrode strip belong to the same correction control group, the method further includes: If the current electrode strip and the previous electrode strip do not belong to the same correction control group, then the correction result is a qualified candidate electrode strip, and the candidate electrode strips are sorted according to the winding order to obtain the sorting result. Based on the sorting results, a preset number of historical electrode strips are selected from the candidate electrode strips. A correction model is trained based on the historical electrode edge position, historical electrode offset, and historical correction target value of the historical electrode strips to obtain the preset correction model.

14. The electrode correction method as described in claim 13, characterized in that, Before determining whether the current electrode strip and the previous electrode strip belong to the same correction control group, the method further includes: Multiple numerical groups are randomly generated, and the predicted correction value corresponding to each numerical group is calculated. The numerical group includes the number of electrode strips in the correction control group and the number of samples selected from the candidate electrode strips. Obtain the actual correction value, and determine the preset quantity and the preset sample number based on the residual between the predicted correction value and the actual correction value.

15. The electrode correction method according to any one of claims 9 to 14, characterized in that, Before determining the correction target value based on the electrode edge position and the electrode offset using a preset correction model, the method further includes: Obtain the winding needle corresponding to the current electrode strip and find the preset correction model corresponding to the winding needle. Different winding needles correspond to different preset correction models.

16. The electrode correction method according to any one of claims 9 to 14, characterized in that, Before correcting the current electrode strip based on the correction target value, the method further includes: Obtain the historical correction value of the previous electrode strip and calculate the difference between the target correction value and the historical correction value; Accordingly, the step of correcting the current electrode strip based on the correction target value includes: If the difference is less than a preset threshold, then the current electrode strip is corrected based on the correction target value.

Images