A lithium battery current collector intelligent manufacturing method and system based on ultra-thin aluminum foil defects

Abstract

The application belongs to the technical field of tension adjustable control device, and discloses a lithium battery current collector intelligent manufacturing method and system based on ultra-thin aluminum foil defects, wherein the ultra-thin aluminum foil is a metal aluminum foil with a hole dense area and a thickness uneven defect; in the manufacturing process, the holes are first filled with hot-pressed conductive resin powder and fixed by a base film to avoid stress concentration in the hole dense area; then the coating gap and pressure of the coating and detection head are controlled to compensate for the thickness change of the ultra-thin aluminum foil; finally, the tension parameters in the processing section are jointly adjusted by the multiple floating rollers and the coating and detection head, so that the finally prepared carbon-coated positive electrode current collector foil has uniform thickness, and conductive adhesive coating layers are formed on the upper and lower surfaces of the ultra-thin aluminum foil. The application fully utilizes the inherent defects of the ultra-thin aluminum foil, combines defect detection with tension control, realizes real-time feedback and dynamic adjustment, ensures efficient operation of the production line, and reduces the rate of defective products and production cost.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of tension-adjustable control devices, in particular to a lithium battery current collector intelligent manufacturing method and system based on ultra-thin aluminum foil defects. BACKGROUND

[0002] Ultra-thin aluminum foil generally refers to aluminum foil with a thickness of 1-8 μm (partially ≤4 μm for high-end applications), which is an upgraded product of traditional aluminum foil (≥20 μm). When it is applied to the manufacture of lithium battery positive current collectors, it has the following main inherent defects: 1. Microscopic hole defects, which are manifested as: 1-8 μm diameter through holes (pinholes) that cause electrolyte penetration and battery short circuit; the main causes are: impurities (such as aluminum oxide particles) embedded in the base material during rolling; local stress concentration caused by micro-cracks on the surface of the roller; uneven annealing process (porosity increases significantly when temperature fluctuates >5℃); 2. Surface wrinkles and uneven thickness, which are manifested as: thickness fluctuation >±0.5 μm (±0.2 μm for high-end requirements), and wrinkle height >2 μm; the main causes are: inaccurate rolling tension control; roller thermal deformation; interlayer misplacement during aluminum foil winding; 3. Deterioration of mechanical properties, which are manifested as: tensile strength <100 MPa (ideal value 120-150 MPa), and elongation <2%; the main causes are: uneven grain size; cold rolling work hardening not fully eliminated; excessive content of impurity elements.

[0003] When ultra-thin aluminum foil is applied to the manufacture of lithium battery current collectors, it is generally recognized in the art that the smaller the pinhole density, the better, which usually includes the following reasons: 1. Avoiding internal short circuit of the battery; 2. Ensuring the mechanical properties of the current collector; 3. Improving the electrochemical performance of the battery. If there are too many pinholes, the active material will not adhere firmly, and it will easily fall off during charging and discharging, reducing the number of active materials participating in the reaction and the specific capacity and energy density of the battery. Numerous existing documents can prove this point, for example, CN116093440A discloses a lithium battery aluminum foil pinhole control method, and CN119101831A discloses a high-toughness, low-pinhole-rate, heat-resistant power battery aluminum foil produced by a short process and a preparation method thereof, which records that the aluminum foil current collector, as an important component of the power battery, plays a role in loading electrode active material, converging current and leading out, and high electrode compaction density is beneficial to improving the energy density of the power battery, which requires the aluminum foil current collector to have high strength, toughness, low pinhole rate and other characteristics to ensure that there are no problems such as broken strips and defects during the battery production coating and rolling process.

[0004] The existing power battery aluminum foil mainly uses 1060 alloy as raw material, and the comprehensive mechanical property indexes such as tensile strength and elongation of the ultra-thin aluminum foil prepared are low, and the pinhole rate is high (there are many pinhole dense areas), which greatly affects the yield of high-performance power battery products and increases the production cost. Therefore, in the example 1 provided by the prior art, the pinhole rate of the aluminum foil with a thickness of 12 μm is reduced to 10 / m 2 The pinhole rate of the comparative example is 30 / m 2 and above. However, due to the ductility of the material itself, the complexity of the processing technology and the influence of environmental factors, it is actually difficult to accurately control the pinhole rate of the ultra-thin aluminum foil in the production process, resulting in a large number of pinhole defects. On the one hand, it leads to local damage, tearing and belt breaking caused by uneven tension during the preparation of the current collector, which seriously affects the quality, performance and yield of the current collector product. Therefore, there is a technical contradiction between the decrease of the material thickness and the increase of the pinhole density and the decrease of the mechanical properties of the material in the ultra-thin aluminum foil, and it is difficult to completely solve it in the production process of the ultra-thin aluminum foil, which belongs to the inherent defect.

[0005] In view of these problems, the conventional lithium battery current collector preparation technology is also difficult to solve. The carbon-coated aluminum foil (also known as carbonized aluminum foil) has several steps in its manufacturing process: surface pretreatment, coating formula preparation, film coating, drying, and post-processing. For example, CN117059819A discloses a preparation method of a high-rate carbon-coated aluminum foil current collector for lithium batteries. However, it still faces several technical difficulties or contradictions: 1. Insufficient defect detection capability. The existing technology lacks detection precision and speed, and is insufficient in detecting the full-area pinhole density, distribution position, and thickness variation of ultra-thin aluminum foil. It is difficult to achieve real-time detection and targeted processing of ultra-thin aluminum foil in the processing process (cannot adjust the subsequent preparation parameters in advance to address defects). 2. Insufficient tension control capability. Traditional tension control mainly relies on static parameter setting and simple feedback adjustment, which cannot adapt to dynamic changes in the production process and material properties. It cannot capture and adapt to dynamic changes in tension distribution in real time, especially the thickness of ultra-thin aluminum foil, the position and density of pinholes, the ductility of the material, the coating speed, and temperature changes, which will affect the tension distribution. Traditional tension control methods cannot capture these dynamic changes in real time, and lack the ability to control the tension of ultra-thin aluminum foil throughout its length. It cannot make fine adjustments to sections with high pinhole density and thin thickness. If the mechanical properties of the worst section are to be maintained without breaking, a smaller tension is required, which will significantly reduce production efficiency. If a larger tension is used, it will easily cause the worst section to break, affecting the yield and production efficiency. 3. Insufficient closed-loop optimization control capability. It cannot achieve coordination between defect detection and tension control, coating thickness compensation, and other functions, making it difficult to fundamentally solve the inherent processing defect problem of ultra-thin aluminum foil. 4. Cannot automatically overcome sensor abnormalities. Single-type or single-path sensors are usually used, especially tension sensors. If the sensor malfunctions, it cannot be detected in time until a belt breakage or other failure occurs, resulting in a large amount of scrap foil and other losses.

[0006] Existing research shows that carbon-coated aluminum foil has been widely used as a positive current collector for lithium batteries. Compared to bare foil materials, carbon-coated aluminum foil can significantly improve the electrochemical performance and interface stability of the battery, thereby enhancing the overall performance and reliability of the lithium battery. However, due to the presence of various technical defects or contradictions in ultra-thin aluminum foil, conventional techniques and production equipment and processes cannot be directly applied to the process of preparing carbon-coated aluminum foil from ultra-thin aluminum foil. They also cannot achieve the technical effect of simultaneously improving production efficiency and product quality, and cannot meet the needs of the lithium battery manufacturing industry. SUMMARY

[0007] In view of the above technical problems, the present application provides a kind of lithium battery current collector intelligent manufacturing method and system based on ultra-thin aluminum foil defect, for the inherent defects such as pinhole dense area of ultra-thin aluminum foil itself, synchronous improvement preparation process and equipment, focus on improving detection ability, tension control ability, and using deep learning technology, the defects are reasonably utilized and controlled, solve the technical difficulties existing in the process of preparing defect ultra-thin aluminum foil into carbon-coated aluminum foil, can realize the problems of improving production efficiency and product quality at the same time, to meet the needs of lithium battery manufacturing industry.

[0008] The present application provides the following technical solutions:

[0009] A kind of lithium battery current collector intelligent manufacturing method based on ultra-thin aluminum foil defect, characterized in that, the ultra-thin aluminum foil is less than 8 μm, with hole dense area (overall pinhole rate is greater than 5%) and thickness uneven defect metal aluminum foil;In the manufacturing process, use hot-pressing conductive resin powder to fill the hole and be fixed by substrate film, avoid stress concentration in hole dense area;The thickness change of the ultra-thin aluminum foil is controlled by the coating gap and pressure compensation of coating and detection head, and the tension parameters in each processing section are adjusted by multiple floating rollers cooperating with coating and detection head, so that the thickness of the finally prepared carbon-coated positive current collector foil is uniform, and conductive adhesive coating is formed on the upper and lower surfaces of the ultra-thin aluminum foil, comprising the following steps:

[0010] S1, setting intelligent manufacturing system

[0011] According to the front and rear order, set unwinding machine A, coating machine B, winding machine C, and control unit for controlling the coordinated operation of the unwinding machine A, coating machine B and winding machine C;

[0012] Base material film unwinding roller, aluminum foil unwinding roller and hot-pressing roller, powder spraying head and front vision detector are arranged in the unwinding machine A;

[0013] Front floating roller, rear vision detector, coating and detection head, middle floating roller and rear floating roller are arranged in the coating machine B;

[0014] S2, first section manufacturing and control

[0015] After feeding to winding roller and fixing joint, make base material film unwinding roller and aluminum foil unwinding roller uniform speed, synchronous unwinding, winding roller uniform speed winding, base material film is below, aluminum foil is above, at the same time, in the gap between base material film and aluminum foil, powder spraying head is sprayed to the bonding surface between base material film and aluminum foil;

[0016] The base film, the sprayed conductive resin powder and the aluminum foil form a sandwich structure. When the aluminum foil passes through the contact surface between the aluminum foil unwinding roller and the hot pressing roller, the resin in the conductive resin powder melts, so that the conductive resin powder and the base film and the aluminum foil form an integrated structure. The conductive resin powder fills the holes of the ultra-thin aluminum foil and makes the thickness uniform, so that the bottom composite aluminum foil is obtained.

[0017] The front vision detection camera and the front vision detection light source of the front vision detector detect the hole dense area and the thickness defect of the bottom composite aluminum foil leaving the first section, obtain defect detection image data, and transmit the defect detection image data to the control unit in real time.

[0018] S3, second section manufacturing and control

[0019] The bottom composite aluminum foil passes through the front floating roller, the rear vision detector, the coating and detection head, the middle floating roller and the rear floating roller in sequence.

[0020] The coating and detection head non-uniformly coats the conductive paste on the upper surface of the bottom composite aluminum foil, and detects the gap between the coating and detection head and the bottom composite aluminum foil and the conductive paste pressure of the coating surface.

[0021] The control unit continuously locates and tracks the hole dense area and the thickness defect of the bottom composite aluminum foil in the first section, and when the defect concentrated area section passes through the second section, the tension of the bottom composite aluminum foil passing through is controlled by adjusting the relative height of the front floating roller, the middle floating roller and the rear floating roller, and the gap between the coating and detection head and the bottom composite aluminum foil.

[0022] After the bottom composite aluminum foil passes through the coating and detection head and is coated with conductive paste, it is dried, and the carbon coating layer and the bottom composite aluminum foil are compacted when passing through the rear floating roller, forming a carbon-coated positive electrode current collector foil.

[0023] S4, third section manufacturing and control

[0024] The carbon-coated positive electrode current collector foil is uniformly wound by the winding roller of the winding machine, and the manufacturing of the current section foil is completed, that is, the manufacturing of the carbon-coated positive electrode current collector foil in the single control section length between the unwinding roller and the winding roller.

[0025] S5, continuous manufacturing and control of multiple control sections

[0026] Repeat steps S2-S4 to continuously manufacture and control the carbon positive electrode current collector foil in each control section length in sequence, and uniformly wind by the winding roller for subsequent processing.

[0027] A lithium battery current collector intelligent manufacturing system based on ultra-thin aluminum foil defects is used to implement the lithium battery current collector intelligent manufacturing method based on ultra-thin aluminum foil defects.

[0028] Compared with the prior art, the manufacturing method and system provided by the present application have at least the following remarkable beneficial effects:

[0029] 1、The present application makes full use of the large number of hole (pinhole) dense areas, thickness unevenness and other defects inherent in ultra-thin aluminum foil, and simultaneously improves the intelligent manufacturing method and system for lithium battery current collectors in view of the various inherent defects possessed by ultra-thin aluminum foil, constructs a double-sided conductive coating, and uses deep learning and other technologies to focus on improving the detection capability of aluminum foil defects, tension control capability and the like, realizes reasonable utilization and control of defects, can solve various technical difficulties existing in the process of preparing the defect ultra-thin aluminum foil into carbon-coated aluminum foil, can maintain a high processing speed, can simultaneously realize the problems of improving production efficiency and product quality, and can meet the needs of the lithium battery manufacturing industry.

[0030] 2、The present application introduces a base film, conductive resin powder spraying and hot pressing composite mechanism and process into the unwinding machine, hot-presses the three into one, strengthens the strength of the ultra-thin aluminum foil when it actually leaves the ultra-thin aluminum foil material roll, makes up for the process defects of low strength caused by a large number of pinholes, is beneficial to subsequent process control, and can maintain a high processing speed.

[0031] 3、The present application significantly improves the batch, rapid detection precision and speed of a large number of micro-holes (pinholes) defects, through the paired visual detection camera and visual detection light source, the image signals in a larger area are collected in sections, usually the width of the ultra-thin aluminum foil collected each time is 0.5-1m, the light emitted by the high-brightness visual detection light source passes through the ultra-thin aluminum foil and the base film, at the same time, a large number of micro-holes (hole dense areas) and thickness data of the ultra-thin aluminum foil in the section are collected at one time (the holes with relatively high brightness are pinhole defects, and the areas with relatively high brightness are areas with relatively thin thickness); the high-definition camera (high-resolution industrial camera) and the strip light source cooperate with each other, significantly improve the image quality, and ensure that the tiny defects can also be clearly captured.

[0032] 4、The present application significantly improves the adaptability and efficiency of defect detection through deep machine learning, multi-scale feature extraction and dynamic transfer learning technologies. A multi-scale feature pyramid network (FPN) and a deep learning model are used to extract global semantic information and local detail information, which can accurately identify various types of micro-hole dense areas, thickness unevenness and other defects.

[0033] 5. The control unit of the present application realizes the tracking of defect location changes and the dynamic adjustment and optimization of tension within each control section length of the carbon positive electrode current collector foil, fully utilizes the in-transit time and distance of the ultra-thin aluminum foil of each section, and predicts the relative height and action start-stop time of the front floating roller, the middle floating roller and the rear floating roller, as well as the gap and pressure between the coating and detection head and the bottom layer composite aluminum foil, to control the tension parameters of the aluminum foil passing through each place, keep the unwinding and winding speed unchanged, and perform closed-loop intelligent control, so as to improve the product quality of the obtained carbon-coated positive electrode current collector and the consistency of the final lithium battery.

[0034] 6. The present application realizes closed-loop cooperative optimization of defect detection and tension control, realizes real-time feedback and dynamic adjustment of aluminum foil tension at different positions within the section through the combination of closed-loop control module, defect detection and tracking and tension control; after detecting defects, the system can complete the adjustment of process parameters in advance without the need to reduce the speed, which can avoid the expansion of defects or the generation of defective products; through multiple iterations and closed-loop optimization, the control unit can continuously improve the tension control parameters, reduce the defective rate and improve the production efficiency.

[0035] 7. The present application also improves the environmental adaptability and robustness of the system, corrects the influence of temperature drift and roller wear on the elastic modulus through the physical constraint compensation module, ensures the precision and stability of tension control, and makes the system maintain high-precision detection and tension control in a dynamically changing environment, which has strong adaptability and significantly better robustness than the prior art.

[0036] 8. The present application realizes dynamic transfer learning, can quickly adapt to new types of defect detection tasks, and can be applied to different batches and defect-specific ultra-thin aluminum foil processing. Through the transfer learning technology, the system can quickly identify new types of defects without retraining the model, which significantly reduces the detection cost and time.

[0037] 9. The present application significantly reduces the defective rate and production cost. Through the combination of high-precision defect detection, tracking and dynamic tension regulation, the system can significantly reduce the defective rate, and the real-time feedback and closed-loop optimization mechanism further improves the production efficiency and reduces the production cost.

[0038] 10. The present application provides a new intelligent and automated solution, realizes the full-process intelligentization and automation of defect detection, tracking and tension control, reduces the dependence on manual experience, and through the integration of high-precision optical imaging, deep learning defect recognition and adaptive tension control, the system can realize efficient and accurate production management.

[0039] 11. The present application can realize accurate tension control operation, through setting multiple pressure sensors and sliding resistance scales, comparing and analyzing the data collected at each place, discovering data drift, error or abnormality in time, avoiding distortion and accumulation easily occurred in single data collection approach, and ensuring long time safe and accurate operation of the system.

[0040] 12. The defect area real-time detection and feedback control mechanism provided by the present application can timely detect, locate and fully utilize the inherent defects of the ultra-thin aluminum foil, convert the defects into elements for improving the performance of the carbon-coated positive electrode current collector foil, greatly reduce the difficulty of manufacturing the ultra-thin aluminum foil, improve the production efficiency, and also improve the economic benefits of enterprises.

[0041] 13. The present application is easy to integrate into various lithium battery production complete production lines, can adapt to the production needs of different types and specifications of lithium batteries, can better meet the needs of modern production for high efficiency, high quality and automation, and has a wider application prospect. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 It is a component structure schematic diagram of the lithium battery current collector intelligent manufacturing system based on the ultra-thin aluminum foil defects of the embodiment of the present application.

[0043] Figure 2 It is a component structure schematic diagram of the control program module of the control unit in the lithium battery current collector intelligent manufacturing system based on the ultra-thin aluminum foil defects of the embodiment of the present application.

[0044] Figure 3 It is a partial control flow schematic diagram of the control unit in the lithium battery current collector intelligent manufacturing method based on the ultra-thin aluminum foil defects of the embodiment of the present application.

[0045] Figure 4 It is a partial control principle schematic diagram of the control unit in the lithium battery current collector intelligent manufacturing method based on the ultra-thin aluminum foil defects of the embodiment of the present application.

[0046] In the figure:

[0047] A, unwinding machine; B, coating and drying machine; C, winding machine

[0048] 1, base material film unwinding roller; 2, base material film;

[0049] 3, powder spraying head; 4, conductive resin powder; 5, aluminum foil unwinding roller; 6, hot pressing roller;

[0050] 71, front vision detection camera; 72, front vision detection light source;

[0051] 8, ultra-thin aluminum foil;

[0052] 9. Front floating roller; 10. Front floating roller height adjustment cylinder;

[0053] 101. Rear vision inspection camera; 102. Rear vision inspection light source;

[0054] 11. Coating and detection head;

[0055] 12. Middle floating roller; 13. Middle floating roller height adjustment cylinder;

[0056] 14. Rear floating roller; 15. Rear floating roller height adjustment cylinder;

[0057] 16. Take-up roller. Detailed Implementation

[0058] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0059] Example 1

[0060] See appendix Figures 1-4 The intelligent manufacturing method for lithium battery current collectors based on defects in ultra-thin aluminum foil provided in this embodiment uses ultra-thin aluminum foil with a thickness of less than 8μm (4μm in this embodiment), containing a large number of densely packed pores (micropores) and uneven thickness defects (the area of ​​the densely packed pores accounts for more than or equal to 5%). During the manufacturing process, the pores are first filled with hot-pressed conductive resin powder and fixed by a substrate film to avoid stress concentration in the densely packed pore areas. Then, the coating gap and pressure of the coating and detection head are controlled to compensate for the thickness variation of the ultra-thin aluminum foil. Finally, multiple floating rollers, in conjunction with the coating and detection head, jointly adjust the tension parameters within the processing section to ensure that the final carbon-coated positive electrode current collector foil has a uniform thickness and forms a conductive bonding coating on both the upper and lower surfaces of the ultra-thin aluminum foil, resulting in a high-performance carbon-coated positive electrode current collector. The specific steps include the following:

[0061] S1. Set up an intelligent manufacturing system

[0062] In the order of their arrangement, unwinding machine A, coating machine B, and winding machine C are set up, along with a control unit (PLC controller + host computer) that coordinates the operation of unwinding machine A, coating machine B, and winding machine C.

[0063] The unwinding machine A is equipped with a substrate film unwinding roller 1, an aluminum foil unwinding roller 5, a hot press roller 6, a powder spraying head 3, and a front vision detector (including a front vision detection camera 71 and a front vision detection light source 72).

[0064] The coating machine B is equipped with a front floating roller 9, a rear vision detector (including a rear vision detection camera 101 and a rear vision detection light source 102), a coating and detection head 11, a middle floating roller 12 and a rear floating roller 14.

[0065] The base film unwinding roller 1, the aluminum foil unwinding roller 5, the hot pressing roller 6 and the winding roller 16 are all provided with pressure sensors (i.e. the tension sensors shown in the figure) for collecting the aluminum foil tension change data in real time. Figure 4 The surface temperature of the hot pressing roller 6 is 150-180 °C, and the pressure is 1-3 MPa, so as to ensure that the conductive resin powder 4 passing through the contact surface is melted and forms an integrated structure with the ultra-thin aluminum foil 8 and the base film 2.

[0066] The base film in the embodiment is a PP film.

[0067] The conductive resin powder 4 is a mixed dry powder particle prepared by carbon-based materials and resin material particles. The carbon-based materials include conductive graphite, conductive carbon black, graphene or carbon nanotubes, etc. The resin includes epoxy resin, polyester resin, etc. The ratio is that the carbon-based materials are 1-3 parts, the resin materials are 7-9 parts, and other additives such as plasticizers and curing agents are 1 part.

[0068] Specific scheme one of the conductive resin powder 4: conductive carbon black-epoxy resin system

[0069] Components and mass fraction ratio: conductive carbon black 2 parts, epoxy resin 8 parts, and other additives (plasticizer and curing agent, filler) 1 part. Under the ratio, the better conductive performance can be ensured, the reliable bonding effect can be realized, and the comprehensive cost is relatively controllable.

[0070] Specific scheme two of the conductive resin powder 4: graphene-polyester resin system

[0071] Components and mass fraction ratio: graphene 1.5 parts, polyester resin 8.5 parts, and other additives (plasticizer and curing agent, filler) 1 part. The ratio can exert the high conductive advantage of graphene, and at the same time, utilize the stability characteristics of polyester resin to improve the comprehensive performance of the product.

[0072] Specific scheme three of the conductive resin powder 4: carbon nanotube-epoxy resin system

[0073] Components and mass fraction ratio: carbon nanotube 2.5 parts, epoxy resin 7.5 parts, and other additives (plasticizer and curing agent, filler) 1 part. In the scheme, the proportion of carbon nanotube is appropriately increased, which can further enhance the conductive performance, and is suitable for the lithium battery current collector scene with higher conductive requirement.

[0074] The plasticizer and the curing agent, and the filler can adopt conventional components and proportions (such as a mass ratio of 0.1:0.1:0.8). The plasticizer includes phthalate plasticizers such as di(2-ethylhexyl) phthalate (DEHP), dibutyl phthalate (DBP), and diethyl phthalate (DEP); and also includes aliphatic dibasic acid ester plasticizers or epoxy plasticizers.

[0075] The curing agent is an amine curing agent, including aliphatic amines such as ethylenediamine and diethylenetriamine, or aromatic amines such as m-phenylenediamine and diaminodiphenylmethane.

[0076] The filler is an inorganic filler, including calcium carbonate, talc, or silicon dioxide.

[0077] S2, first section manufacturing and control

[0078] The first section is mainly a processing section within the range of the unwinding machine A;

[0079] After feeding the ultra-thin aluminum foil and the PP film to the winding roller 16 and fixing the joints, the base film unwinding roller 1 and the aluminum foil unwinding roller 5 are uniformly and synchronously unwound, and the winding roller 16 is uniformly wound. The base film 2 is below, the aluminum foil 8 is above, and the conductive resin powder 4 is sprayed from the powder spraying head 3 to the bonding surface between the base film 2 and the aluminum foil 8 in the gap between the base film 2 and the aluminum foil 8.

[0080] The base film 2, the sprayed conductive resin powder 4, and the aluminum foil 8 form a sandwich structure. When passing through the contact surface of the aluminum foil unwinding roller 5 and the hot press roller 6, the resin in the conductive resin powder 4 melts, so that the conductive resin powder 4 and the base film 2 and the aluminum foil 8 form an integrated structure. The conductive resin powder 4 fills the holes in the ultra-thin aluminum foil and has a uniform thickness, obtaining a bottom layer composite aluminum foil, so that the inherent defect structure of the ultra-thin aluminum foil 8 is targetedly strengthened.

[0081] The front vision detection camera (industrial camera) 71 and the front vision detection light source (high-brightness strip-shaped industrial light source) of the front vision detector detect the hole and thickness defects of the bottom layer composite aluminum foil leaving the first section, obtain defect detection image data, and transmit the defect detection image data to the control unit in real time.

[0082] S3, second section manufacturing and control

[0083] The second section is mainly a processing section within the range of the coating and drying machine B;

[0084] The bottom layer composite aluminum foil passes through the front floating roller 9, the rear vision detector (rear vision detection camera 101 and rear vision detection light source 102), the coating and detection head 11, the middle floating roller 12, and the rear floating roller 13 in sequence.

[0085] The conductive paste is non-uniformly coated on the upper surface of the bottom layer composite aluminum foil by the coating and detection head 11, and the gap between the coating and detection head 11 and the bottom layer composite aluminum foil and the conductive paste pressure of the coating surface are detected; for the area with thinner thickness, the gap between the detection head 11 and the bottom layer composite aluminum foil is increased to increase the coating thickness and reduce the tension of the ultra-thin aluminum foil at this position;

[0086] According to the detection data of the front vision detector, the control unit continuously locates and tracks the hole dense area and thickness defect of the first section of the bottom layer composite aluminum foil, and when the defect concentrated section passes through the second section, the secondary detection and confirmation are performed through the front vision detector, the control parameter adjustment instruction is triggered, and the control unit adjusts the relative height of the front floating roller 9, the middle floating roller 12 and the rear floating roller 14, and the gap between the coating and detection head 11 and the bottom layer composite aluminum foil, to control the tension and coating thickness of the bottom layer composite aluminum foil passing through the coating and detection head 11;

[0087] The control unit specifically adjusts the height of the corresponding floating roller by controlling the front floating roller height adjusting cylinder 10, the middle floating roller height adjusting cylinder 13 and the rear floating roller height adjusting cylinder 15; increasing the height of the front and middle floating rollers will increase the tension of the bottom layer composite aluminum foil passing through the floating roller, and decreasing the height of the front and middle floating rollers will decrease the tension of the bottom layer composite aluminum foil passing through the floating roller; the adjustment effect of the rear floating roller is opposite;

[0088] After the bottom layer composite aluminum foil is coated with conductive paste by the coating and detection head 11, it is dried, and the carbon coating layer and the bottom layer composite aluminum foil are compacted when passing through the rear floating roller 15 to form the carbon-coated positive electrode current collector foil;

[0089] S4, third section manufacturing and control

[0090] The third section is mainly a processing section within the range of the winding machine C;

[0091] The carbon-coated positive electrode current collector foil is uniformly wound by the winding roller of the winding machine, and the manufacturing of the current section foil is completed, that is, the manufacturing of the carbon-coated positive electrode current collector foil within the length of a single control section between the unwinding roller and the winding roller;

[0092] S5, continuous manufacturing and control of multiple control sections

[0093] Repeat steps S2-S4 to continuously manufacture and control the carbon-coated positive electrode current collector foil within the length of the first to third control sections, and sequentially uniformly wind by the winding roller for subsequent slitting and other processing.

[0094] The control unit includes a PLC controller and an upper computer, the PLC controller is used to collect detection data in each section, and real-time transmission to the upper computer, and according to the control signal returned by the upper computer, the manufacturing parameters in each section are actually controlled;

[0095] The host computer is internally provided with an intelligent control program, which includes a data receiving module, a defect identification module, a defect classification module, a defect positioning module, a defect feature analysis module, a tension control module, a model predictive control module, a PID control module, a physical constraint compensation module and a closed-loop control module.

[0096] The PLC controller and the host computer generate intelligent control signals according to detection data processing, feature extraction and defect judgment, and send them to the PLC controller for execution, and then collect feedback data, and again perform detection data processing, feature extraction, defect judgment and defect tracking data to generate intelligent control signals, to realize closed-loop intelligent control.

[0097] In the intelligent control program, the data receiving module is used to receive various real-time detection data and control feedback data.

[0098] The defect identification module includes multiple deep learning algorithm models, which are used to analyze and predict the detection data, especially the defect detection image data, and identify the type, position and size of the micro-hole defect feature data.

[0099] The defect tracking analysis module tracks and analyzes the defect area according to the sensing data of the front and rear vision detectors and the winding speed, and determines the time point of reaching the coating and detection head.

[0100] The predictive control module determines the tension adjustment strategy and adjustment parameters according to the defect feature data and tracking analysis data.

[0101] The tension overall control module is used to control the balance of foil speed, tension and defects in the four sections to avoid aluminum foil cracking or belt breakage.

[0102] The PID control module is used to perform real-time adjustment of the tension adjustment parameters.

[0103] The physical constraint compensation module is used to correct the influence of temperature drift and roller wear on tension control.

[0104] The closed-loop control unit optimizes the manufacturing control data in the next control section length according to the manufacturing detection, control data and feedback results in the previous control section length, and iterates continuously to improve manufacturing precision and efficiency.

[0105] The control unit generates intelligent control signals according to detection data processing, feature extraction and defect judgment, and sends them to the PLC controller for execution, and then collects feedback data, and again performs detection data processing, feature extraction, defect judgment and defect tracking data to generate intelligent control signals, to realize closed-loop intelligent control, specifically as follows:

[0106] The PLC controller and the upper computer first process the received data to extract defect feature data, which is handed over to a defect identification module and a defect classification module to identify and classify the aluminum foil defects, a defect positioning module is used to determine the coordinate position of the defects on the aluminum foil, the results are sent to a defect tracking analysis module for tracking processing, and a predictive control module is used to output control signals, and the relative height and action start and stop time of the front floating roller, the middle floating roller and the rear floating roller, and the gap and pressure between the coating and detection head and the bottom layer composite aluminum foil are jointly adjusted by the tension overall control module, the PID control module and the physical constraint compensation module, so that the tension parameters of the aluminum foil passing through each place are controlled, the unwinding and winding speeds are kept unchanged, closed-loop intelligent control is performed, the aluminum foil processing in multiple control section lengths is repeated, and the product quality of the obtained carbon-coated positive electrode current collector and the consistency of the final lithium battery are improved.

[0107] The image data is analyzed by a defect identification module based on a deep machine learning model to identify the type, position and size of the micro-hole dense area defects. The module can accurately identify multiple types of defects and output detailed information of the defects through a deep learning model and a multi-scale feature extraction technology. The deep learning defect identification module includes a defect classification model that classifies defects such as holes, scratches and stains using a deep learning model convolutional neural network (CNN) for classification, supports transfer learning and can quickly adapt to new types of defect detection tasks. In the defect identification module, the function of the aluminum foil defect classification model is as follows:

[0108] Formula 1

[0109] In the formula:

[0110] F: feature map;

[0111] Wc and bc : weights and biases of the classification network;

[0112] P class : probability distribution of defect categories;

[0113] The defect positioning module is used to determine the position of the defects, specifically through a target detection algorithm (such as Faster R-CNN, YOLO) to locate the specific position of the defects and calibrate the coordinate information. The defect positioning and size analysis unit analyzes the size (such as diameter, area) of the defects, extracts the boundary box and size information of the defects through image segmentation technology (such as U-Net). A multi-scale feature pyramid network (FPN) algorithm is used to extract global semantic information and local detail information to ensure detection accuracy. Multi-stage feature maps are extracted through a bottom-up path, high-level feature maps and low-level feature maps are fused through a top-down path, and a multi-scale feature pyramid is generated.

[0114] In the prediction control module, the control function is as follows:

[0115] Formula 2

[0116] In the formula, T is a predicted tension value, T is a current tension value, T is a target tension value, and T is an actual tension value.

[0117] T future : predicted tension value;

[0118] T curren t： current tension value;

[0119] U： control input;

[0120] A,B : system matrix;

[0121] E : external disturbance;

[0122] In the PID control module, the control function is as follows:

[0123] Formula 3

[0124] In the formula, T is a predicted tension value, T is a current tension value, T is a target tension value, and T is an actual tension value.

[0125] U(t) : output of controller;

[0126] e (t) : tension error, difference between target tension and actual tension;

[0127] K p , K i , K d : proportional coefficient, integral coefficient, and differential coefficient, respectively;

[0128] : error e(t) integral with respect to time.

[0129] The defect identification module uses multiple deep learning models to classify and identify defects in images, including: using a multi-scale feature pyramid network (FPN) to extract multi-scale features of defects, using a convolutional neural network (CNN) for image analysis, and using federated learning for sharing defect data among multiple processing sections.

[0130] ​​The tension overall control module can also realize tension adjustment in combination with the following mode: a prediction control module (MPC) generates a tension adjustment strategy by tracking and predicting the dynamic influence of defects on the tension of the underlying composite aluminum foil at the coating and detection head 11; the PID control unit executes real-time adjustment of the tension parameters according to the strategy generated by the MPC; and the physical constraint compensation module corrects the elastic modulus by the following formula:

[0131] ε t =ε 0 ×[1+α(T−T 0 )+β⋅τ]t

[0132] wherein, ε t is the corrected elastic modulus, ε 0 is the initial elastic modulus, α is a temperature drift coefficient, T is the current temperature, T 0 is a standard temperature, β is a wear coefficient, τ is the cumulative wear amount, t is a time factor.

[0133] The front visual detector and the rear visual detector each include a high-definition digital camera and a strip-shaped light source; the high-definition digital camera is arranged above the aluminum foil processing route; the strip-shaped light source is arranged below the aluminum foil processing route, and the light emitted thereby covers the entire width of the aluminum foil processing route and is within the effective data acquisition range of the high-definition digital camera, can provide high-brightness uniform illumination, can make the light penetrate the foil and enhance the contrast of the defect image, and can detect even smaller micropores and thickness differences; the high-definition digital camera acquires image data of the aluminum foil passing through the effective data acquisition range thereof (generally 0.5-1 m width of the aluminum foil is acquired at a time), and sends the image data to the control unit to determine the defects existing in the aluminum foil.

[0134] The front floating roller, the middle floating roller and the rear floating roller each include a floating roller and a height adjustment cylinder, and a linkage sliding resistance gauge is arranged on the height adjustment cylinder; when the height adjustment cylinder drives the floating roller to move up and down, the sliding resistance gauge synchronously feeds back actual data of control execution to the control unit, so as to compare and verify the data of other sensors.

[0135] As Figure 1As shown, the lithium battery current collector intelligent manufacturing system based on ultra-thin aluminum foil defects provided by the application is used to implement the lithium battery current collector intelligent manufacturing method based on ultra-thin aluminum foil defects; in sequence from front to back, a unwinding machine A, a coating machine B, a winding machine C are arranged in sequence, and a control unit for coordinating operation of the unwinding machine A, the coating machine B and the winding machine C is arranged;

[0136] A base material film unwinding roller, an aluminum foil unwinding roller and a hot pressing roller, a powder spraying head and a front visual detector are arranged in the unwinding machine A.

[0137] A front floating roller, a rear visual detector, a coating and detection head, a middle floating roller and a rear floating roller are arranged in the coating machine B.

[0138] The winding machine C is provided with a winding roller 16.

[0139] The control unit of the embodiment of the application comprises a PLC controller and an upper computer, and the upper computer is provided with a control program, and the running steps of the control program itself are as follows:

[0140] A1: data preprocessing: denoising, enhancing and edge detection are performed on the visual image of the ultra-thin aluminum foil to generate a high-quality defect detection input image;

[0141] A2: feature extraction: multi-scale defect features are extracted through a three-level feature pyramid network (FPN) to capture global semantic information and local detail information;

[0142] Multi-stage feature maps can be extracted from the backbone network through a bottom-up path; through a top-down path and a horizontal connection, high-level feature maps and low-level feature maps are fused to generate a multi-scale feature pyramid;

[0143] A3: classification and determination: a dynamic transfer learning model is used to classify and grade the defects, mark coordinates, and then output defect types, positions and size parameters;

[0144] A4: data feedback: the defect information is introduced into a tension overall control module, cooperates with a PID control module, a physical constraint compensation module and the like, control instructions are issued in advance, and the tension parameter of the foil passing through the coating and detection head 11 is adjusted in real time to avoid defects from being enlarged or defective products from being generated when passing through the coating and detection head 11. The tension borne by the foil passing through the coating and detection head 11 is the largest in the three sections, and is also the position most prone to tearing and belt breakage.

[0145] The high-precision optical imaging module is a basic module of the whole system, is used to capture images on the surface of the ultra-thin aluminum foil in real time, and provides high-quality input data for subsequent defect identification. Through the combination of a high-resolution industrial camera and a strip light source, the module can capture images with a resolution of 12K or more, so that dense areas formed by small defects such as holes with a diameter of less than 0.1 mm can also be clearly captured.

[0146] The embodiment of the present application makes multi-aspect and synergistic improvements on tension unevenness and micro-hole defect dense area in the production of ultra-thin aluminum foil, can capture and track defect features and positions in real time, can identify tiny defects, and improves detection accuracy and speed; through deep learning and closed-loop control, etc., dynamic tension segmentation adjustment is performed, and micro-hole dense area and other defects are reasonably applied, so that the prepared carbon-coated positive electrode current collector foil has uniform thickness, and conductive adhesive coating is formed on the upper and lower surfaces of the ultra-thin aluminum foil, greatly improving the conductivity, safety performance of the positive electrode current collector, and the volume energy density and consistency of the lithium battery.

[0147] The high-definition camera in the embodiment specifically adopts a high-resolution industrial camera for obtaining a high-precision optical image penetrating the thickness of the ultra-thin aluminum foil, and the resolution reaches 12K or more, which is much higher than the resolution (usually 2K-8K) of a traditional industrial camera, and can capture more subtle defects, for example, a FLIR Blackfly S BFS-U3-51S5C industrial camera, according to the width of the aluminum foil, one or more cameras can be installed for full-coverage online detection; the resolution reaches 12K (i.e., the horizontal pixel number is about 12,000 pixels), which can provide extremely high image details, compared with a traditional industrial camera (2K-8K resolution), a 12K camera can capture a micro-hole dense area formed by smaller defects (such as holes with a diameter less than 0.1 mm); high-speed shooting: supports high-speed shooting, adapts to the needs of high-speed production lines, and ensures real-time performance.

[0148] The convolutional neural network (CNN) works in cooperation with the federated learning unit. The role of the convolutional neural network (CNN) is as follows: in the identification of ultra-thin aluminum foil defects, CNN is used as a core image analysis tool, which automatically extracts local features in the image through the convolution kernel in the convolution layer sliding on the image. For example, in the identification of micro-hole dense area defects, the convolution kernel can capture texture, shape and other feature information related to the defects in the image. With the deepening of the network level, the features are down-sampled through the pooling layer, which reduces the data volume while retaining the key features, and then the features are mapped to different categories through the fully connected layer, so as to realize the classification of defects.

[0149] The feature extraction principle and advantages of the multi-scale feature pyramid network (FPN), the multi-scale feature extraction process: FPN extracts multi-stage feature maps from the backbone network in a bottom-up manner. In this process, with the deepening of the network level, the size of the feature map gradually decreases, but the semantic information gradually increases. For example, in the early network layers, local detail information of the image can be captured, such as the edges and textures of tiny defects; in deeper network layers, more global semantic information can be obtained, such as the position and distribution of defects in the entire aluminum foil image.

[0150] Multi-scale feature fusion mechanism: Through the top-down path and lateral connection, FPN fuses high-level feature maps with low-level feature maps to generate a multi-scale feature pyramid. Specifically, after up-sampling operation, high-level feature maps are added element by element or spliced with low-level feature maps of the corresponding layer, so as to combine global semantic information with local detail information. This fusion mechanism makes the feature pyramid contain rich feature information at different scales, which can better adapt to defects of different sizes and shapes. For example, when identifying micro-hole defects in ultra-thin aluminum foil, for the hole dense area formed by the micro-hole with a diameter less than 0.1 mm, the low-level feature map can provide clearer edge information, while the high-level feature map can determine its position in the whole image, and the combination of the two can improve the accuracy of defect recognition.

[0151] Defect recognition specific process, classification process based on deep learning model: After processing by CNN and FPN, feature maps F containing rich feature information are obtained. These feature maps are input into the classification network, and the classification network performs weighted summation and nonlinear transformation operations on the features through the learned weights Wc and biases bc to calculate the probability distribution Pclass of the defect class. For example, in a classification task with n defect classes, Pclass is a vector of length n, and each element represents the probability of the corresponding defect class. By sorting Pclass, the class with the highest probability is selected as the classification result of the defect.

[0152] Positioning method and principle: The defect positioning model uses the feature maps output by the deep learning model to determine the specific position of the defect in the ultra-thin aluminum foil image. A common method is to find the area where the feature value is significantly higher than the background area by threshold segmentation or region growing operations on the feature map. These areas are the areas where the defects are located. Then, by calculating the bounding box or contour information of these areas, the position coordinates of the defects are determined.

[0153] Defect size analysis, size analysis method and significance: The defect size analysis unit determines the size of the defect by analyzing the area size occupied by the defect in the feature map or the original image. For example, the actual size of the defect can be obtained by calculating the number of pixels in the defect area and combining the resolution information of the image. This is of great significance for evaluating the impact of defects on the quality of ultra-thin aluminum foil products. Different sizes of defects may have different degrees of impact on the performance of the product.

[0154] Prediction principle based on defect characteristics: The Model Predictive Control Unit (MPC) analyzes the characteristic information of defects in the production process of ultra-thin aluminum foil, such as the type, location, size, and distribution of defects, to predict the dynamic impact of these defects on tension. For example, when a specific type of micro-hole dense area defect is detected in a certain area of the ultra-thin aluminum foil, the MPC will predict the tension trend in that area based on the characteristics of the defect and the process parameters of the aluminum foil production (such as winding speed, aluminum foil thickness, etc.). This is based on a deep understanding of the physical laws in the production process of ultra-thin aluminum foil and the establishment of mathematical models, through the learning and analysis of a large amount of historical data, to determine the relationship between defect characteristics and tension changes.

[0155] Generation process of tension adjustment strategy: After obtaining the prediction results of the dynamic impact of defects on tension, the MPC generates the corresponding tension adjustment strategy according to the preset objective function and control constraints. The objective function usually considers multiple factors, such as minimizing tension fluctuations, ensuring stable aluminum foil quality, and improving production efficiency. Control constraints include physical limitations of the system, such as the output power range of the motor and the measurement range of the tension sensor. For example, if it is predicted that the tension in a certain area will decrease, which may cause the aluminum foil to relax and produce new defects, the MPC will calculate the required tension value according to the objective function and constraints, and generate the corresponding control instructions to adjust the tension control parameters.

[0156] PID control principle: The PID control unit adjusts the tension parameters in real time according to the tension adjustment strategy generated by the MPC. PID control is a classic feedback control algorithm that calculates the control input U(t) by performing proportional, integral, and differential operations on the system error e(t), where e(t) represents the tension error, i.e., the difference between the target tension and the actual tension; Kp is the proportional coefficient, which amplifies the influence of the current tension error on the control input, enabling the system to quickly respond to changes in error; Ki is the integral coefficient, which accumulates past tension errors to eliminate steady-state errors in the system; and Kd is the differential coefficient, which predicts the trend of tension error changes and adjusts the control input in advance to improve the dynamic performance of the system.

[0157] Real-time adjustment process: In actual operation, the PID control unit continuously obtains the current tension error e(t) and calculates the control input U(t) according to the set Kp, Ki, and Kd coefficients. For example, when a deviation between the actual tension and the target tension is detected, the PID control unit will amplify the current deviation according to the proportional coefficient Kp to produce an immediate control effect; at the same time, the integral coefficient Ki will accumulate the past deviation, so that the control input can gradually eliminate the steady-state error; the derivative coefficient Kd will adjust the control input in advance according to the rate of change of the deviation to avoid overshoot or oscillation of the system. Through this real-time adjustment process, the PID control unit can accurately adjust the tension parameter to the target value, thereby ensuring the stability and consistency of the tension in the production process of ultra-thin aluminum foil.

[0158] Elasticity modulus correction formula analysis: The physical constraint compensation module corrects the elasticity modulus by the formula εt=ε0×[1+α(T−T0)+β⋅τ]t. Wherein, εt is the corrected elasticity modulus, ε0 is the initial elasticity modulus, α is the temperature drift coefficient, reflecting the influence degree of temperature change on the elasticity modulus; T is the current temperature, T0 is the standard temperature; β is the wear coefficient, used to describe the influence of roller wear on the elasticity modulus; τ is the cumulative wear amount, t is the time factor. This formula considers the influence of temperature drift and roller wear on the elasticity modulus.

[0159] Correction process and significance: In the production process of ultra-thin aluminum foil, changes in temperature and wear of the roller will cause changes in the elasticity modulus, affecting the accuracy of tension control. The physical constraint compensation module monitors the current temperature T and cumulative wear τ in real time, and calculates the corrected elasticity modulus εt according to the pre-determined temperature drift coefficient α and wear coefficient β using the above formula. Then the corrected elasticity modulus is applied to the tension control model, so that the tension control parameters can be adjusted according to the actual material properties, thereby improving the accuracy and stability of tension control. For example, when the temperature rises, the elasticity modulus of the material may decrease, and through the correction of the compensation unit, the tension control parameters can be appropriately increased to maintain the stability of the aluminum foil tension.

[0160] Model prediction and PID control synergy: The combination of MPC and PID control fully utilizes the advantages of both. MPC can make forward-looking predictions and plans for tension changes caused by defects, allowing the system to adjust tension parameters in advance to avoid defect expansion or defective products; PID control can track and correct tension errors in real time to ensure the accuracy and rapid response of tension control. This synergy allows the tension control unit to maintain high stability and accuracy in complex and variable production environments.

[0161] Physical constraint compensation ensures the effectiveness of the system: The physical constraint compensation module can correct the temperature drift and roller wear and other factors, eliminating the influence of these external factors on tension control, further improving the accuracy and stability of tension control. For example, during long-term production, the wear of the roller will cause uneven tension distribution. Through real-time correction by the physical constraint compensation module, the tension control unit can always maintain optimal working conditions, ensuring the consistency of product quality.

[0162] Ability to respond to dynamic changes: In actual production, the production environment and working conditions of ultra-thin aluminum foil may change, such as temperature fluctuations, equipment aging, etc. The adaptive tension control module involved in this claim can sense these changes in real time and automatically adjust the tension parameters through MPC prediction, real-time adjustment of PID control, and correction by the physical constraint compensation module, enabling the system to adapt to dynamic changes in the environment and maintain stable production performance.

[0163] Improved system robustness: Since the system can automatically compensate for various external disturbances and internal changes affecting tension control, it has stronger robustness. Even when faced with some unpredictable factors, such as minor differences in raw materials, sudden equipment failures, etc., the system can minimize the impact on tension control through its own adjustment mechanism, ensuring continuous production and stable product quality.

[0164] Real-time feedback unit: The main function of the real-time feedback unit is to complete process adjustment immediately after detecting defects to ensure timely optimization of tension during ultra-thin aluminum foil production, avoiding further expansion of defects or production of defective products.

[0165] Real-time processing algorithm: After the tension control module receives defect information, a real-time processing algorithm is used to analyze and process the data. This algorithm can quickly analyze defect information, determine the type, location, and size of defects, and calculate the tension parameters that need to be adjusted according to the pre-set control strategy. For example, for different types of defects, the algorithm will determine whether to increase or decrease tension and the specific amplitude of adjustment based on the degree and direction of its impact on tension.

[0166] Fast execution mechanism: The tension control module adjusts tension based on the calculated tension parameters through a fast execution mechanism. These execution mechanisms usually include motors, pneumatic devices, hydraulic devices, etc., which can respond to control instructions in a short time to achieve accurate adjustment of tension parameters. For example, a motor can quickly change its output power according to control input U(t) to adjust the tension and running speed during ultra-thin aluminum foil production.

[0167] The process optimization unit is used to optimize the tension control parameters according to historical defect data to reduce the scrap rate and improve the production quality of ultra-thin aluminum foil. Data collection and storage: the system records various process parameters and defect data in real time during production, including tension parameters, rolling speed, aluminum foil thickness, defect type, location and size, etc. These data are stored in the database to form a large historical defect data set. For example, during long-term production, each detected defect information and corresponding tension control parameters are recorded in detail to form a rich data resource.

[0168] Data analysis and mining: the process optimization unit uses data analysis techniques to deeply mine and analyze the historical defect data set, and finds the relationship and rules between defect compensation and tension control parameters through statistical analysis, machine learning, etc. For example, by analyzing a large amount of data, it is found that when the tension is controlled within a certain range under the conditions of certain hot pressing speed, hot pressing temperature and pressure and corresponding aluminum foil thickness, the compensation effect on defects such as hole dense area is better.

[0169] Optimization model establishment and solution: based on the results of data analysis, the process optimization unit establishes an optimization model of tension control parameters. The model takes reducing the scrap rate as the objective function and various process parameters and constraint conditions as constraints, and solves the optimal tension control parameter combination through mathematical optimization algorithm. For example, intelligent optimization algorithms such as genetic algorithm and particle swarm algorithm are used to find the tension control parameter settings that minimize the scrap rate under the premise of meeting production requirements and equipment limitations.

[0170] Parameter update and application: once the optimized tension control parameters are obtained, the process optimization unit updates these parameters to the tension control module and applies them in actual production. At the same time, the system continues to monitor the defect situation and tension control effect in the production process, and continuously adjusts and optimizes the tension control parameters to adapt to changes in the production environment and improve production efficiency.

[0171] The main working principle and process are as follows.

[0172] 1. During operation, the data of each tension sensor is fed back to the PLC controller, and the PLC controller calibrates the data of the two sensors;

[0173] 2. When the tension sensor senses during operation, the position of the corresponding floating roller changes, driving the position of the sliding resistance ruler to change, and the driven roller in each floating roller (the driven roller) also feeds back the corresponding data to the PLC controller; Figure 4

[0174] ​3. The PLC controller comprehensively judges whether the real-time feedback tension of the two sensors is accurate (whether it exceeds the allowed error range); before starting, the staff can set the error range of the previously verified tension data in the PLC controller;

[0175] 4. When the two sensor data are compared and exceed the allowed error range, the tension data abnormality control process is automatically triggered, the PLC controller automatically switches the PID closed-loop control algorithm to the closed-loop control based on the smaller data to avoid tearing or wrinkling or belt breakage, and stops after completing the processing of the current roll;

[0176] 7. After stopping, detect each tension sensor, recalibrate the tension and set the tension data error range, and then enter the processing of the next roll.

[0177] A kind of super-thin aluminum foil intelligent coating manufacturing method based on multi-modal perception, comprising the following steps:

[0178] S1. Data preprocessing

[0179] Data preprocessing: the visual image collected in the production process of super-thin aluminum foil is preliminarily processed, and high-quality image data is generated, which provides basic data for subsequent defect detection analysis and control.

[0180] Specific steps:

[0181] 1. Denoising: use image denoising algorithm (such as Gaussian filter, non-local mean denoising or wavelet transform) to remove random noise in the image, and the denoising algorithm can preserve image details while reducing noise interference on defect detection.

[0182] 2. Image enhancement: use histogram equalization or gamma correction method to enhance the contrast of the image, so that the defect area (hole dense area, thickness super-thin area) is more obvious, for uneven illumination, adaptive histogram equalization (CLAHE) can be used to enhance local contrast.

[0183] 3. Edge detection: use edge detection algorithm (such as Canny edge detection or Sobel operator) to extract edge information in the image, edge detection can help identify the outline of the defect, and provide support for subsequent feature extraction and classification.

[0184] Optimization for image characteristics of super-thin aluminum foil: denoising and enhancement algorithms are optimized for the high reflectivity and small defect characteristics of super-thin aluminum foil to ensure image quality.

[0185] Edge detection and defect detection: edge detection extracts the outline of the defect, which provides support for subsequent multi-scale feature extraction.

[0186] S2. Feature extraction

[0187] Feature extraction is the core step of defect detection, through a three-level feature pyramid network (FPN) to extract multi-scale defect features, capture global semantic information and local detail information.

[0188] Specific steps:

[0189] 1. Bottom-up path: the backbone network (such as ResNet50) extracts feature maps C2, C3, C4, C5 layer by layer, the resolution of the feature map gradually decreases, and the semantic information gradually increases.

[0190] 2. Top-down path: starting from the high-level feature map C5, the resolution is enlarged by upsampling: Fup(x,y)=Upsample(C5), the upsampling feature map is fused with the low-level feature map C4: P4=Fusion(Fup(x,y),C4), the fused feature map is further extracted by convolution operation.

[0191] 3. Final feature pyramid: output multi-scale feature maps P2, P3, P4, P5, each feature map corresponds to different resolution and semantic information.

[0192] S3. Classification and determination

[0193] Classification and determination is based on the extracted features to classify defects and output the type, location and size parameters of the defects.

[0194] Specific steps:

[0195] 1. Defect classification: use deep learning model (such as convolutional neural network, CNN) to classify defects, classification results include hole dense area, ultra-thin area, scratch, stain and other defect types.

[0196] 2. Defect positioning: use target detection algorithm (such as Faster R-CNN, YOLO) to locate the specific position of the defect, output the boundary box coordinates of the defect area.

[0197] 3. Defect size analysis: use image segmentation technology (such as U-Net) to extract the boundary box and size information of the defect area, output the diameter, area and other size parameters of the defect.

[0198] S4. Data feedback

[0199] Data feedback is to input the defect detection results into the control unit to adjust the tension parameters in real time, to efficiently compensate for defects and avoid defects from expanding or producing defective products, the specific steps are:

[0200] 1. Defect information transmission: Real-time transmission of defect type, area location and size parameters to the tension overall control module, achieving cooperative optimization of defect detection and tension control through closed-loop control.

[0201] 2. Tension parameter adjustment: Model predictive control unit (MPC) predicts tension adjustment strategy based on defect characteristics, and PID control unit performs real-time adjustment of tension parameters with an accuracy of 0.001N.

[0202] 3. Physical constraint compensation: Corrects the impact of temperature drift and roller wear on tension control, ensuring the accuracy and stability of tension adjustment.

[0203] Dynamic adjustment of tension precision: The combination of model predictive control and PID control can adjust tension parameters in real time according to defect characteristics, with an accuracy of 0.001N. This high-precision tension adjustment can effectively avoid the expansion of micro-hole defects or the generation of new defective products caused by uneven tension, improving product quality.

[0204] 4. Enhance the adaptive ability and robustness of the system to improve the ability to respond to different batches of ultra-thin aluminum foil materials and equipment changes: The physical constraint compensation module can correct the impact of temperature drift and roller wear on tension control, so that the system can maintain stable tension control performance under different environmental conditions, material batches and equipment states. For example, in the case of different batches of foil materials, large temperature changes or severe roller wear, the system can still accurately adjust tension parameters to ensure the continuity of production and the consistency of product quality.

[0205] Ability to quickly adapt to new defect types: Dynamic transfer learning model enables the system to quickly learn and adapt to new batches of foil defects without retraining the entire model. This greatly improves the flexibility and adaptability of the system, reducing maintenance costs and time.

[0206] The advantages of real-time feedback and closed-loop optimization: By feeding back the defect detection results to the tension overall control module in real time and dynamically adjusting, the system realizes closed-loop cooperative optimization of defect detection and tension control.

[0207] The above-mentioned embodiments of the present application have been measured on an internal test production line, and the actual test shows that the maximum diameter of the winding and unwinding is 1500mm, the maximum bearing capacity is 2000KG, and the running speed can reach 250m / min. The real-time detection and feedback mechanism provided by the present application can timely detect, locate and fully utilize the inherent defects of the ultra-thin aluminum foil, and convert the defects into elements for improving the performance of the carbon-coated positive electrode current collector foil, greatly reducing the difficulty of manufacturing the ultra-thin aluminum foil, improving the production efficiency, and improving the economic benefit of the enterprise. The present application is easy to integrate into various lithium battery production complete production lines, can adapt to the production needs of different types and specifications of lithium batteries, can better meet the needs of modern production for high efficiency, high quality and automation, and has a wider application prospect.

[0208] Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor belong to the scope of protection of the present application.

Claims

1. A method for intelligent manufacturing of lithium battery current collector based on ultra-thin aluminum foil defects, characterized by, The super-thin aluminum foil is a metal aluminum foil with a thickness less than 8 microns, a hole dense area and a thickness uneven defect; in the manufacturing process, the holes are first filled with a hot-pressed conductive resin powder and fixed by a base film to avoid stress concentration in the hole dense area; then the coating gap and pressure of the coating and detection head are controlled to compensate for the thickness change of the super-thin aluminum foil; finally, a plurality of floating rollers cooperate with the coating and detection head to jointly adjust the tension parameters in the processing section, so that the finally prepared carbon-coated positive electrode current collector foil has a uniform thickness and forms a conductive adhesive coating on the upper and lower surfaces of the super-thin aluminum foil, comprising the following steps: S1, setting an intelligent manufacturing system The unwinding machine A, the coating machine B and the winding machine C are arranged in sequence, and a control unit is arranged to control the coordinated operation of the unwinding machine A, the coating machine B and the winding machine C; A base film unwinding roller, an aluminum foil unwinding roller and a hot-pressing roller, a powder spraying head and a front vision detector are arranged in the unwinding machine A; A front floating roller, a rear vision detector, a coating and detection head, a middle floating roller and a rear floating roller are arranged in the coating machine B; S2, first section manufacturing and control After being fed to the winding roller and fixed, the base film unwinding roller and the aluminum foil unwinding roller are unwound at a constant speed, the winding roller is wound at a constant speed, the base film is below and the aluminum foil is above, and at the same time, the conductive resin powder is sprayed from the powder spraying head to the joint surface between the base film and the aluminum foil in the gap between the base film and the aluminum foil; When the base film, the sprayed conductive resin powder and the aluminum foil form a sandwich structure passing through the contact surface of the aluminum foil unwinding roller and the hot-pressing roller, the resin in the conductive resin powder melts, so that the conductive resin powder, the base film and the aluminum foil form an integrated structure, the conductive resin powder fills the holes of the super-thin aluminum foil and makes the thickness uniform, and a bottom layer composite aluminum foil is obtained; The front vision detection camera and the front vision detection light source of the front vision detector detect the hole dense area and the thickness defect of the bottom layer composite aluminum foil leaving the first section, obtain defect detection image data, and transmit the defect detection image data to the control unit in real time; S3, second section manufacturing and control The bottom layer composite aluminum foil passes through the front floating roller, the rear vision detector, the coating and detection head, the middle floating roller and the rear floating roller in sequence; The bottom layer composite aluminum foil is non-uniformly coated with conductive slurry by the coating and detection head, and the gap between the coating and detection head and the bottom layer composite aluminum foil and the conductive slurry pressure of the coating surface are detected; The control unit continuously locates and tracks the hole dense area and the thickness defect of the bottom layer composite aluminum foil in the first section, and when the defect concentrated section passes through the second section, the relative height of the front floating roller, the middle floating roller and the rear floating roller and the gap between the coating and detection head and the bottom layer composite aluminum foil are adjusted to control the tension when the bottom layer composite aluminum foil passes through; After the bottom layer composite aluminum foil is coated with conductive slurry by the coating and detection head, it is dried, the carbon coating and the bottom layer composite aluminum foil are compacted by the rear floating roller, and a carbon-coated positive electrode current collector foil is formed; S4, third section manufacturing and control The carbon-coated positive electrode current collector foil is wound at a constant speed by the winding roller of the winding machine, and the manufacturing of the current section foil is completed, that is, the manufacturing of the carbon-coated positive electrode current collector foil in the single control section between the unwinding roller and the winding roller. S5, multiple control sections are continuously manufactured and controlled Steps S2-S4 are repeated to continuously manufacture and control the carbon positive electrode current collector foil in each control section length, and the winding roller is uniformly wound for subsequent processing.

2. The method for intelligent manufacturing of lithium battery current collector based on defects of ultra-thin aluminum foil according to claim 1, characterized in that, The control unit includes a PLC controller and an upper computer, the PLC controller is used for collecting detection data in each section, and real-time transmission is given to the upper computer, and the manufacturing parameters in each section are actually controlled according to the control signal returned by the upper computer; The upper computer is internally provided with an intelligent control program, which includes: a data receiving module, a defect identification module, a defect classification module, a defect positioning module, a defect feature analysis module, a tension control module, a model predictive control module, a PID control module, a physical constraint compensation module and a closed-loop control module; The PLC controller and the upper computer generate intelligent control signals according to detection data processing, feature extraction and defect judgment, and send them to the PLC controller for execution, and then collect feedback data, and again perform detection data processing, feature extraction, defect judgment and defect tracking data generation Intelligent control signal, realize closed-loop intelligent control. 3.The method of claim 2, wherein the method further comprises, In the intelligent control program, The data receiving module is used for receiving various real-time detection data and control feedback data; The defect identification module includes a plurality of deep learning algorithm models, which are used for analyzing and predicting defect detection image data, identifying the type, position and size of the feature data of the micro-hole dense area defect; The defect tracking analysis module tracks and analyzes the defect area according to the winding speed and the sensing data of the front and rear vision detectors, and determines the time point of reaching the coating and detection head; The predictive control module determines the tension adjustment strategy and adjustment parameters according to the defect feature data and tracking analysis data; The tension overall control module is used to control the balance of foil speed, tension and defects in the four sections, so as to avoid aluminum foil cracking or belt breaking; The PID control module is used for real-time adjustment of the tension adjustment parameters; The physical constraint compensation module is used to correct the influence of temperature drift and roller wear on tension control; The closed-loop control unit optimizes the manufacturing control data in the next control section length according to the manufacturing detection, control data and feedback results in the previous control section length, and iterates continuously to improve manufacturing precision and efficiency. 4.The method of claim 3, wherein the method further comprises, after the step of forming the aluminum foil into the aluminum sheet, the step of: The control unit generates intelligent control signals according to detection data processing, feature extraction and defect judgment, and sends them to the PLC controller for execution, and then collects feedback data, and again performs detection data processing, feature extraction, defect judgment and defect tracking data generation Intelligent control signal, realize closed-loop intelligent control, specifically: ​ The PLC controller is connected with the upper computer, and first, the received data is processed to extract defect feature data, which is sent to a defect identification module and a defect classification module to identify and classify the aluminum foil defects, a defect positioning module is used to determine the coordinate position of the defects on the aluminum foil, and the result is sent to a defect tracking analysis module for tracking processing, and then a predictive control module is used to output a control signal, and the tension overall control module, the PID control module and the physical constraint compensation module are used to jointly adjust the relative height and action start-stop time of the front floating roller, the middle floating roller and the rear floating roller, and the gap and pressure between the coating and detection head and the bottom layer composite aluminum foil, so that the tension parameters of the aluminum foil passing through each place are controlled, the unwinding and winding speeds are kept unchanged, closed-loop intelligent control is performed, the aluminum foil in multiple control section lengths is repeatedly processed, and the product quality of the obtained carbon-coated positive electrode current collector and the consistency of the final lithium battery are improved.

5. The method for intelligent manufacturing of lithium battery current collector based on defects of ultra-thin aluminum foil according to claim 3, characterized in that, In the defect identification module, the aluminum foil defect classification model function is formula 1 as follows: Formula 1 In the formula: F: feature map; Wc and bc : weights and biases of the classification network; P class : probability distribution of defect categories; In the predictive control module, the predictive control function is formula 2 as follows: Formula 2 In the formula: T future : predicted tension value; T curren t： Current tension value; U： control input; A, B : system matrix; E : external disturbance; In the PID control module, the control function is formula 3 as follows: Formula 3 In the formula: U(t) : output of the controller; e (t) : tension error, difference between target tension and actual tension; K p , K i , K d : proportional, integral, derivative coefficients, respectively : error e(t) Integral over time. 6.The method of claim 3, wherein the method further comprises, after the step of forming the first and second aluminum layers, the step of: The defect identification module uses multiple deep learning models to classify and identify defects, including: using a multi-scale feature pyramid network (FPN) to extract multi-scale features of defects, using a convolutional neural network (CNN) for image analysis, and using federated learning for sharing defect data in multiple processing sections. ​ 7. The method for intelligent manufacturing of lithium battery current collector based on ultra-thin aluminum foil defects according to any one of claims 1 to 6, characterized in that, The front visual detector and the rear visual detector both include a high-definition digital camera and a bar light source. The high-definition digital camera is arranged above the aluminum foil processing route, and the bar light source is arranged below the aluminum foil processing route, and the light emitted by the bar light source covers the entire width of the aluminum foil processing route and is within the effective data acquisition range of the high-definition digital camera; the high-definition digital camera acquires image data of the aluminum foil passing through the effective data acquisition range thereof and sends the image data to the control unit to determine the defects existing in the aluminum foil.

8. The method for intelligent manufacturing of lithium battery current collector based on ultra-thin aluminum foil defects according to any one of claims 1 to 6, characterized in that, The front floating roller, the middle floating roller and the rear floating roller each include a floating roller and a height adjusting cylinder, and a linked sliding resistance ruler is arranged on the height adjusting cylinder; when the height adjusting cylinder drives the floating roller to move up and down, the sliding resistance ruler feeds back control execution data to the control unit.

9. The method for intelligent manufacturing of lithium battery current collector based on defects of ultra-thin aluminum foil according to any one of claims 1 to 6, characterized in that, Pressure sensors are arranged on the base film unwinding roller, the aluminum foil unwinding roller, the hot press roller and the winding roller to acquire aluminum foil tension change data in real time; the surface temperature of the hot press roller is 150-180 DEG C, and the pressure is 1-3 MPa.

10. A lithium battery current collector intelligent manufacturing system based on ultra-thin aluminum foil defects, characterized by, The application is used to implement the intelligent manufacturing method of lithium battery current collector based on ultra-thin aluminum foil defects according to any one of claims 1-9.

Images