LFP electrode gradient compaction intelligent regulation process method and LFP electrode

Abstract

This application provides a smart control process for gradient compaction of LFP electrodes and an LFP electrode, belonging to the field of LFP electrode preparation technology. It addresses the problems of poor compaction consistency, easy ion channel blockage, high interfacial impedance, and short roller life in related technologies. This process integrates intelligent adaptive control, NiTi / DLC self-healing roller surface, digital twin predictive control, edge computing multi-field coupling control, and interface adaptation optimization, resulting in an LFP electrode with a dense surface layer and a porous inner layer gradient structure, with a surface compaction density of 2.5-2.6 g / cm³. 3 Interfacial impedance ≤ 5Ω, ion diffusion coefficient ≥ 2.5 × 10⁻⁶ ‑10 cm 2 / s, significantly improving performance and reducing costs.

Description

Technical Field

[0001] This application relates to the field of LFP electrode fabrication, and more particularly to an intelligent control process for gradient compaction of LFP electrodes and an LFP electrode. Background Technology

[0002] Lithium-ion batteries, due to their high energy density and long cycle life, have been widely used in new energy vehicles, energy storage, and other fields. Among them, LFP (lithium iron phosphate) electrodes, with their high safety and low cost, have become one of the mainstream electrode types for liquid lithium-ion batteries and emerging solid-state batteries. In solid-state battery systems, the interfacial contact between the electrode and the solid electrolyte, as well as the ion conduction efficiency, are the core bottlenecks that determine battery performance. The electrode compaction process directly affects the electrode density, pore structure, and interfacial compatibility. Therefore, optimizing the electrode compaction process is a key direction for improving the performance of LFP batteries (especially solid-state batteries).

[0003] In existing technologies, LFP electrode gradient compaction processes mostly employ a fixed parameter control mode. This involves preset compaction pressure, roller speed, and other parameters based on experience, using multi-stage roller presses to achieve gradient compaction. Some processes introduce single-dimensional material detection or simple microstructure roller surface design in an attempt to improve compaction, but such improvements are difficult to adapt to the stringent requirements of solid-state batteries for electrode-electrolyte interface contact. Furthermore, existing processes rely heavily on manual sampling for quality control, with an incomplete data traceability system, making precise control throughout the entire process difficult and failing to meet the high process stability requirements of solid-state battery industrialization.

[0004] However, existing technologies have several drawbacks: First, fixed parameter control cannot adapt to differences in particle size and solid content among different batches of LFP materials, easily leading to over-compaction of small-particle materials blocking ion channels and under-compaction of large-particle materials reducing density. This not only affects the consistency of electrode performance but also exacerbates the problem of poor interfacial contact between the electrode and the solid electrolyte in solid-state batteries. Second, there is a lack of efficient parameter optimization mechanisms, resulting in strong control lag and difficulty in coping with process fluctuations caused by changes in multiple environmental fields such as temperature and vibration. Solid-state battery preparation is more sensitive to the process environment, and fluctuations can easily cause interfacial defects. Third, the microstructure roller surface cannot be repaired after wear, resulting in a short service life and increased industrial operation and maintenance costs. Moreover, the worn roller surface will further damage the electrode surface structure, which is not conducive to the interfacial stability of solid-state batteries. Fourth, the interface optimization between the electrode and the electrolyte (including solid electrolyte precursors) is neglected, resulting in high interfacial impedance, which restricts the cycle life of the battery. This problem is particularly prominent in solid-state batteries and has become a key factor limiting the improvement of their cycle performance. Fifth, data traceability is lacking, making it difficult to trace quality problems to their source and lacking a basis for process optimization, which cannot support the quality control requirements of large-scale solid-state battery production. These defects make it difficult for LFP electrodes prepared by existing processes to simultaneously achieve high solid density, high ion conduction efficiency, and long cycle life, and they are even less able to meet the industrialization requirements of high-performance LFP solid-state batteries and high-end liquid lithium-ion batteries. Summary of the Invention

[0005] This application provides an intelligent control process method for gradient compaction of LFP electrodes and an LFP electrode. Through intelligent collaborative control and structural optimization, it can solve the problems of poor adaptability of existing LFP electrode compaction process parameters, control lag, high interface impedance and difficulty in quality traceability, thereby improving the performance consistency and industrial adaptability of LFP electrodes.

[0006] Firstly, this application provides an intelligent control process method for gradient compaction of LFP electrodes. The method includes the following steps: multi-dimensional online detection of LFP materials; inputting the detection data into a digital twin model to obtain predicted process parameters; integrating multi-field data through an edge computing unit to optimize the predicted process parameters in real time; employing a graded self-healing microstructure roller surface for gradient compaction according to preset gradient compaction parameters; subjecting the compacted electrode to electrolyte precursor micro-spraying and infrared curing treatment; and establishing a full-chain data traceability system. The preset gradient compaction parameters include a compaction pressure of 80-195 MPa and a roller speed of 0.5-1.5 m / min, and the response time of the edge computing unit does not exceed 10 ms.

[0007] By adopting the above technical solutions, multi-dimensional material detection provides precise basis for parameter control. Combined with digital twin prediction and edge computing for real-time optimization, dynamic adaptation of process parameters is achieved, effectively avoiding ion channel blockage or insufficient compaction caused by material differences. In particular, it improves the potential for poor interfacial contact between electrodes and solid electrolytes in solid-state batteries. The graded self-healing microstructure roller surface extends service life while ensuring gradient compaction effect and avoids damage to the electrode surface structure by wear rollers, thus helping to improve the interface stability of solid-state batteries. Interface optimization treatment specifically reduces the interfacial impedance between electrodes and electrolytes (including solid electrolyte precursors), breaking through the core bottleneck of high interfacial impedance in solid-state batteries. Full-chain data traceability enables precise quality control, supporting the quality traceability requirements of large-scale solid-state battery production. The synergistic effect of each step effectively solves the problems of poor process parameter adaptability, lagging control, high interfacial impedance, and difficulty in quality traceability in existing processes. It significantly improves the compaction consistency, ion conduction efficiency, and cycle stability of LFP electrodes, reduces industrialization costs, and better meets the stringent requirements of solid-state batteries for electrode preparation processes.

[0008] Furthermore, the multi-dimensional online detection includes simultaneous detection of the particle size D50, solid content, moisture content, and crystal integrity of the LFP material; the particle size D50 is 1.0-2.2 μm, the solid content is 50%-60%, and the moisture content does not exceed 0.5%; the crystal integrity is detected by X-ray diffraction, and the characteristic peak half-width at half-maximum does not exceed 0.2°; the detection data is input into the digital twin model through a standardized interface.

[0009] By adopting the above technical solutions, accurate and synchronous detection of multi-dimensional characteristics of LFP materials can be achieved. Standardized data transmission ensures the compatibility with subsequent control modules, providing comprehensive data support for precise parameter optimization. This further improves the matching degree between process parameters and material characteristics, avoids electrode performance deviations caused by fluctuations in material characteristics, and in particular reduces interface defects caused by material inhomogeneity in solid-state batteries, ensuring electrode performance consistency and laying a material foundation for the high performance of solid-state batteries.

[0010] Furthermore, the multi-field data fused by the edge computing unit includes ambient temperature, roller vibration, and interface voltage; the CNN-LSTM fusion model is used to optimize the predicted process parameters, and the optimized parameter adjustment step size is pressure ±0.5MPa and temperature ±0.5℃; the digital twin model is iterated once every 50 batches, and the edge computing unit feeds back the optimized data to the digital twin model in real time.

[0011] By adopting the above technical solutions, the fusion of multi-field data improves the comprehensiveness of parameter optimization and can accurately address the impact of environmental changes such as temperature and vibration on the process. Solid-state battery fabrication is more sensitive to the process environment, and this optimization mechanism can effectively avoid interface defects caused by environmental fluctuations. The CNN-LSTM model ensures optimization accuracy, and the model iteration mechanism realizes continuous improvement in process control capability, significantly reduces control lag, and further improves process stability and parameter control accuracy, meeting the high requirements of solid-state battery industrialization for process stability.

[0012] Furthermore, the hierarchical self-healing microstructure roller surface includes a NiTi / DLC composite coating, and gradient-distributed primary microbumps and secondary microgrooves; the primary microbumps are distributed in a three-level gradient: the primary roller surface density is 300 bumps / cm². 2 Secondary roller surface density 200-100 particles / cm 2 30 particles / cm³ of surface density on the third-stage roller 2 The main micro-bumps have a diameter of 50-100μm and a height of 15-35μm, while the secondary micro-grooves have a width of 5μm and a depth of 1μm.

[0013] By adopting the above technical solution, the gradient distribution of the main micro-bumps achieves a gradient structure of dense surface and porous inner layer, which ensures high compaction density to improve energy density while retaining inner ion channels to ensure conduction efficiency. At the same time, the optimized surface structure can improve the adhesion to the solid electrolyte. The secondary microgrooves can accurately store electrolyte precursors (including solid electrolyte precursors), providing a structural basis for electrode-electrolyte interface adaptation. The NiTi / DLC composite coating ensures the wear resistance of the roller surface, effectively improves the gradient compaction effect and roller surface durability, avoids roller surface wear from damaging the electrode surface structure, provides structural support for the stable interface required by solid-state batteries, and lays the foundation for high-performance electrodes.

[0014] Furthermore, the self-healing process of the graded self-healing microstructure roller surface is as follows: graded repair is triggered based on the wear grade threshold; when there is slight wear, the roller surface temperature is adjusted to 65°C to perform secondary microgroove self-healing, the slight wear amount is not less than 2μm, and the self-healing time is not more than 5s; when there is severe wear, the roller surface temperature is adjusted to 70°C to perform main micro-bump shape memory repair, the severe wear amount is not less than 5μm, and the repair time is not more than 10s.

[0015] By adopting the above technical solutions, the graded self-repair mechanism enables precise and rapid repair of roller surface wear, ensuring roller surface performance without disassembly and avoiding a decrease in compaction effect caused by roller surface wear. In particular, it can prevent the worn roller surface from damaging the electrode surface structure and affecting the interface contact with the solid electrolyte. At the same time, it further extends the service life of the roller surface, reduces industrial operation and maintenance costs and downtime losses, and ensures the process continuity and stability of large-scale solid-state battery production.

[0016] Furthermore, the coating amount of the electrolyte precursor micro-spraying is dynamically adjusted according to the real-time porosity of the electrode; the coating amount ranges from 0.1 to 0.3 μL / cm. 2 Control accuracy ±0.01μL / cm 2 The uniformity error of the spraying shall not exceed ±5%; the infrared curing parameters are temperature 80℃ and time 5s.

[0017] By adopting the above technical solution, the coating amount is dynamically adapted to the electrode porosity, ensuring that the electrolyte precursor (including solid electrolyte precursor) is uniformly filled into the secondary microgrooves and electrode pores, avoiding excessively high local interfacial impedance caused by uneven electrolyte distribution; low-temperature rapid curing avoids damage to the electrode structure, while promoting the formation of a stable interfacial transition layer, significantly improving the interfacial bonding force between the electrode and the electrolyte (including solid electrolyte), specifically reducing the interfacial impedance problem that is particularly prominent in solid batteries, and greatly improving the cycle stability of the battery.

[0018] Furthermore, the full-chain data traceability system links data through a unique traceability code for each batch of electrodes; the linked data includes material testing data, process control parameters, equipment operation data, and electrode performance testing data; the data storage period is no less than 3 years.

[0019] By adopting the above technical solutions, full-chain data correlation and traceability can be achieved from materials to finished products. Quality problems can be accurately traced to their source, providing data support for process optimization. In particular, it can meet the stringent requirements of solid-state battery industrialization for full life cycle quality control, improve quality control efficiency, and ensure the quality stability of solid-state battery mass production.

[0020] Furthermore, it also includes a wide material adaptation adjustment step: for LFP materials with a particle size D50 of 1.0-1.4μm, a compaction pressure of 80-120MPa and a roller speed of 1.2-1.5m / min are used; for LFP materials with a particle size D50 of 1.4-1.8μm, a compaction pressure of 120-150MPa and a roller speed of 0.8-1.2m / min are used; for LFP materials with a particle size D50 of 1.8-2.2μm, a compaction pressure of 150-195MPa and a roller speed of 0.5-0.8m / min are used.

[0021] By adopting the above technical solutions, precise parameter matching of LFP materials with different particle sizes can be achieved, and material types can be quickly switched without manual adjustment. This avoids electrode performance fluctuations caused by differences in material particle size, and in particular, it can ensure that electrodes prepared from materials with different particle sizes can form good interfacial contact with solid electrolytes. At the same time, it improves the industrialization flexibility and adaptability of the process, and meets the production needs of multiple types of LFP batteries (including solid-state batteries).

[0022] Furthermore, it also includes process stability control steps: real-time monitoring of electrode deformation rate, when the deformation rate is not less than 0.02 mm / s, reducing the roller speed by 0.05 m / min and simultaneously reducing the compaction pressure by 2 MPa; real-time detection of electrode surface flatness, when the surface flatness error is not less than 0.01 mm, triggering parameter backtracking correction.

[0023] By adopting the above technical solutions, rapid response and defect early warning correction in emergency conditions can be achieved, avoiding defects such as electrode cracking caused by process fluctuations. In particular, it can prevent poor interfaces between electrodes with poor surface flatness and solid electrolytes. This further improves the stability and yield of electrode preparation, ensuring the yield and process reliability of solid-state battery industrial production.

[0024] Secondly, this application provides an LFP electrode. It is prepared using the LFP electrode gradient compaction intelligent control process described in any one of the first aspects above; the LFP electrode surface has a secondary microgroove structure, the grooves are filled with an interface transition layer formed by the solidification of an electrolyte precursor, and the electrode cross-section exhibits a gradient structure of dense surface and porous inner layer; the surface compaction density of the LFP electrode is 2.5-2.6 g / cm³. 3 The inner layer porosity is 30%-40%, and the thickness consistency error does not exceed ±0.003mm; the interfacial impedance of the LFP electrode does not exceed 5Ω, and the ion diffusion coefficient is not less than 2.5×10⁻⁶. -10 cm 2 / s.

[0025] By adopting the above technical solution, all the technical advantages of the aforementioned process methods are inherited. The resulting gradient structure not only ensures high compaction density to improve energy density, but also retains the inner porous structure to ensure high ion conduction efficiency. In particular, the design of the surface-level microgrooves and interface transition layer can significantly optimize the interface contact with the solid electrolyte and significantly reduce the interface impedance. The synergistic optimization of various performance parameters enables the LFP electrode to have excellent consistency, cycle stability and energy density. It can not only be directly applied to high-performance liquid lithium-ion batteries, but also meet the core requirements of high-performance LFP solid batteries, significantly improving the overall performance and cycle life of solid batteries.

[0026] In summary, this application has at least the following beneficial effects:

[0027] 1. A smart and collaborative LFP electrode gradient compaction process and corresponding electrode are provided to accurately solve the defects of existing processes. This not only improves the electrode performance and industrial adaptability, but also specifically addresses the core bottlenecks of poor electrode-electrolyte interface contact and high interface impedance in solid-state batteries, thus meeting the industrialization needs of solid-state batteries.

[0028] 2. Achieve precise material-process matching and multi-field coordinated control to ensure consistent electrode performance, reduce interface defects in solid-state batteries caused by material or process fluctuations, and lay the technological foundation for high-performance solid-state batteries.

[0029] 3. By optimizing the self-healing roller surface and interface, the equipment life is extended and the battery cycle stability is significantly improved, especially the interface stability and cycle performance of solid-state batteries, breaking through the key limitations of improving the cycle life of solid-state batteries;

[0030] 4. Full-chain data traceability and wide material compatibility improve quality control efficiency and industrialization flexibility, which can support the quality control and multi-variety production needs of solid-state battery mass production and promote the industrialization process of LFP solid-state batteries.

[0031] It should be understood that the description in the Summary Section is not intended to limit the key or essential features of the embodiments of this application, nor is it intended to restrict the scope of this application. Other features of this application will become readily apparent from the following description. Attached Figure Description

[0032] The above and other features, advantages, and aspects of the embodiments of this application will become more apparent from the accompanying drawings and the following detailed description. In the drawings, the same or similar reference numerals denote the same or similar elements, wherein:

[0033] Figure 1 A flowchart of an intelligent control process for LFP electrode gradient compaction is shown in an embodiment of this application. Detailed Implementation

[0034] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0035] Furthermore, the term "and / or" in this article 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, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0036] This application provides an intelligent control process for gradient compaction of LFP electrodes and an LFP electrode, which can achieve precise material-process matching and multi-field synergistic control, improve electrode performance consistency and cycle stability, extend equipment life, improve quality traceability, and reduce industrialization costs.

[0037] In a first aspect, embodiments of this application disclose a smart control process for gradient compaction of LFP electrodes.

[0038] Figure 1 A flowchart of an intelligent control process for LFP electrode gradient compaction is shown in an embodiment of this application.

[0039] Reference Figure 1 The method specifically includes the following steps:

[0040] S1: Perform multi-dimensional online detection on LFP materials.

[0041] The specific methods in this step include: multi-dimensional online detection, including simultaneous detection of LFP material particle size D50, solid content, moisture content and crystal integrity. The particle size D50 is 1.0-2.2μm, the solid content is 50%-60%, the moisture content is not more than 0.5%, and the crystal integrity is detected by X-ray diffraction, with the characteristic peak half-width not exceeding 0.2°. The detection data is input into the digital twin model through a standardized interface.

[0042] The LFP material processed in this step is a precursor material for the preparation of lithium-ion battery LFP electrodes. It originates from the mixture output of the electrode slurry preparation process. Its core components are a mixture of LFP active material, conductive agent, binder, and solvent. The LFP active material accounts for 60%-80%, the conductive agent 5%-10%, the binder 2%-5%, and the solvent 15%-25%. The material is a fluid slurry with pumpable and online-monitorable physical properties. The core purpose of this step is to provide fundamental data support for intelligent control of subsequent process parameters through comprehensive and accurate material characteristic testing. This ensures precise matching between the subsequent compaction process and the material characteristics, thereby guaranteeing the consistent performance of the final LFP electrode. The conductive agent uses a SuperP:CNT=3:1 composite system (CNTs are single-walled tubes with a diameter of 1-2 nm and a length of 5-10 μm), the binder uses a PVDF-HFP:LA133=2:1 composite system, the solid-liquid ratio of the solvent (NMP) to the solid material is 1:1.2-1:1.3, and the slurry viscosity is controlled at 2000-3000 mPa·s to ensure the structural stability of subsequent coating and compaction.

[0043] The detection process employs an online detection device integrating multiple sensors. This device integrates four main detection modules: a laser particle size analyzer, a near-infrared spectrometer, a Karl Fischer moisture analyzer, and a small online X-ray diffractometer. Each module is connected to a central control unit via a data bus to achieve synchronous acquisition and integration of detection data. Specific operational details are as follows: First, the LFP slurry to be tested is pumped to the detection channel at a stable flow rate (preferably 10-20 mL / min) using a quantitative pumping device. The laser particle size analyzer uses dynamic light scattering technology to detect the particle size D50, covering a range of 0.5-5 μm. Selectable detection parameters include a sampling frequency of 10-20 Hz, preferably 15 Hz, to ensure representativeness of the results. The selectable range for particle size D50 is 1.0-2.2 μm, with 1.4-1.8 μm being preferred. LFP materials with particle sizes within this range can balance compaction density and ion conductivity. The near-infrared spectrometer analyzes the solid content through characteristic spectra, with a detection wavelength range of 1200-2500 nm and a preferred spectral resolution of 8 cm⁻¹. -1 The solid content can be selected in the range of 50%-60%, preferably 55%-58%. This range can avoid the problems of poor compaction caused by excessively thin slurry or uneven coating caused by excessively thick slurry. The Karl Fischer moisture analyzer uses volumetric method to detect the moisture content, with a detection accuracy of 0.01%. The moisture content is strictly controlled to not exceed 0.5%, preferably not exceed 0.3%. Excessive moisture will damage the internal structure of the electrode and affect the electrochemical performance of the battery. A small online X-ray diffractometer is used to detect crystal integrity. It uses Cu target Kα rays, tube voltage 40kV, tube current 30mA, scanning speed preferably 2° / min, scanning range 20°-80°. Crystal integrity is judged by analyzing the full width at half maximum (FWHM) of the LFP characteristic peaks (2θ=18.5°, 35.7°, 40.3°). The FWHM of the characteristic peaks should not exceed 0.2°, preferably not exceed 0.15°. LFP materials with good crystal integrity can improve the cycle stability of the electrode.

[0044] After each module's testing is completed, the test data is transmitted to the digital twin model's database via a standardized Modbus TCP interface, with transmission latency controlled to ≤50ms to ensure data real-time performance. The final output of this step is a standardized test data set containing four core indicators: particle size D50, solid content, moisture content, and crystal integrity. This data set will serve as the core input for subsequent digital twin model prediction of process parameters, achieving a precise correlation between "material characteristics" and "process parameters," and initially solving the problem of poor process adaptability caused by batch-to-batch material variations in traditional processes.

[0045] Between steps S1 and S2, there is a pre-treatment step: the coated LFP wet electrode is subjected to vacuum drying at a temperature of 80℃, a vacuum of -0.09 MPa, and a drying time of 12 h to ensure that the moisture content of the electrode after drying is ≤50ppm. During the drying process, the moisture content data is monitored in real time by an online moisture analyzer, and the detection results are simultaneously incorporated into the input parameter system of the subsequent digital twin model.

[0046] S2: Input the detection data into the digital twin model to obtain the predicted process parameters.

[0047] The data to be processed in this step is the multi-dimensional standardized test data set of LFP material output from step S1, specifically including four core indicators: particle size D50, solid content, moisture content, and crystal integrity. The data format is Modbus TCP standard format with a transmission latency of ≤50ms. This data set is the core input basis for the digital twin model to predict process parameters. The core purpose of this step is to predict the gradient compaction process parameters suitable for the current batch of LFP material in advance based on the virtual mapping relationship constructed from the material characteristic data. This solves the problems of traditional processes relying on empirical parameters and control lag, and provides initial benchmark parameters for the real-time optimization of subsequent edge computing units.

[0048] The digital twin model used in this step is a full-element virtual mapping model based on the LFP electrode gradient compaction process. The model is built on an industrial-grade simulation platform and integrates three core databases: a material properties database, an equipment operation database, and an electrode performance database. The material properties database contains compaction characteristic data for over 1000 batches of LFP materials with different particle sizes and solid contents. The equipment operation database contains correlation data on roller press pressure, roller speed, and roller surface wear. The electrode performance database contains data on the correspondence between compaction parameters and electrode density and porosity. The core algorithm of the model is a multi-factor regression algorithm based on machine learning, with a training sample size of ≥5000 groups, a model prediction accuracy of ≥95%, and an optimal prediction accuracy of ≥98%.

[0049] The specific processing procedure is as follows: First, the standardized detection data transmitted in step S1 is imported into the material characteristic module of the digital twin model through the data interface. The model automatically matches the historical data group in the database that is closest to the current material characteristics, and at the same time calls the current state data of the roller surface in the equipment operation module, including real-time parameters such as roller surface wear and micro-bump height. Then, based on the preset correlation algorithm, the model calculates the initial gradient compaction process parameters suitable for the current material. The parameter types include the compaction pressure, roller speed, and roller surface temperature of the three-stage roller surface. The predicted range of compaction pressure is 80-195MPa, the predicted range of roller speed is 0.5-1.5m / min, and the predicted range of roller surface temperature is 50-80℃. The parameter output step size is pressure ±1MPa, roller speed ±0.1m / min, and temperature ±1℃, with the preferred step size being pressure ±0.5MPa, roller speed ±0.05m / min, and temperature ±0.5℃.

[0050] The digital twin model operates on a collaborative architecture of cloud servers and local edge computing nodes. The cloud server is responsible for iterative training of the model, while the local edge computing nodes are responsible for real-time parameter prediction. The model is updated every 50 batches to ensure that it always keeps pace with the changing trends of the actual production process. The predicted process parameters output by the model are stored in a structured data format, which includes parameter names, numerical ranges, confidence levels, and other information. Parameters with a confidence level ≥90% are directly used as the benchmark for subsequent optimization, while parameters with a confidence level <90% are marked as parameters to be optimized.

[0051] The final output of this step is a set of predicted process parameters for LFP electrode gradient compaction. This parameter set clarifies the basic compaction parameter range for the current batch of material, enabling preliminary matching of material characteristics and avoiding over-compaction or under-compaction of the electrode due to parameter mismatch. Simultaneously, this parameter set provides a precise optimization starting point for the edge computing unit in the subsequent S3 step, significantly shortening the edge computing optimization time, improving the real-time performance of process control, and laying the foundation for intelligent collaborative control throughout the entire process.

[0052] S3: By fusing multi-field data through edge computing units, the predicted process parameters are optimized in real time.

[0053] The specific methods in this step include: the edge computing unit fuses multi-field data including ambient temperature, roller vibration and interface voltage, and uses a CNN-LSTM fusion model to optimize the predicted process parameters. The optimized parameter adjustment step size is pressure ±0.5MPa and temperature ±0.5℃. The digital twin model is iterated once every 50 batches. The edge computing unit feeds back the optimized data to the digital twin model in real time. The response time of the edge computing unit does not exceed 10ms.

[0054] The parameters to be processed in this step are the LFP electrode gradient compaction prediction process parameter set output from step S2. This parameter set includes the initial values ​​of compaction pressure, roller speed, and roller surface temperature for the three-stage roller surfaces. These are theoretical prediction parameters based on material properties and have not yet been adjusted to incorporate dynamic changes in the real-time production environment. The core objective of this step is to leverage the low-latency data processing capabilities of the edge computing unit to integrate multiple real-time data from the production site, thereby instantly correcting the predicted process parameters. This addresses the disconnect between the digital twin model's predicted parameters and actual operating conditions, enabling dynamic and precise control of process parameters and providing optimal execution parameters for subsequent gradient compaction processes.

[0055] The edge computing unit used in this step is an embedded industrial control terminal, equipped with a quad-core industrial-grade processor with a main frequency of ≥2.0GHz, memory of ≥8GB, and multiple data acquisition interfaces. It can simultaneously receive analog and digital signals, and the data acquisition frequency can be selected in the range of 100-1000Hz, with 500Hz being preferred, to ensure the real-time performance and integrity of multiple data. The acquisition of multiple data points relies on various sensors deployed on the production site. Ambient temperature is acquired using a high-precision thermocouple sensor with a measurement range of 20-100℃ and a measurement accuracy of ±0.1℃. The sensor is installed on the frame near the roller surface of the roller press to monitor the ambient temperature changes in the roller pressing area in real time. Rolling vibration is acquired using a triaxial accelerometer with a measurement range of 0-5g and a measurement accuracy of ±0.01g. The sensor is installed on the roller bearing seat of the roller press to monitor the vibration amplitude and frequency during the rolling process in real time. Interface voltage is acquired using a contact voltage sensor with a measurement range of 0-5V and a measurement accuracy of ±0.001V. The sensor is installed on the transmission path of the compaction electrode to monitor the contact voltage changes between the electrode surface and the roller surface in real time, indirectly reflecting the compaction uniformity of the electrode.

[0056] The specific processing procedure is as follows: First, the edge computing unit synchronously collects three types of multi-field data—ambient temperature, roller vibration, and interface voltage—through a standardized data interface. Simultaneously, it receives the predicted process parameter set output from step S2. Data transmission uses the industrial Ethernet protocol, with a transmission delay ≤5ms. Then, the multi-field data and predicted process parameters are input into a pre-trained CNN-LSTM fusion model. The CNN module of this model is responsible for extracting the spatial features of the multi-field data, such as the distribution gradient of ambient temperature and the frequency characteristics of roller vibration. The convolutional kernel size can be selected as 3×3 or 5×5, with 3×3 being preferred. The activation function is the ReLU function, and the pooling method is max pooling. The LSTM module is responsible for learning the sequential patterns of process parameters changing over time. The number of hidden layers can be selected as 2-4 layers, with 3 layers being preferred. The number of neurons in each layer is 128-256, with 256 being preferred, effectively solving the gradient vanishing problem of traditional time series models. Based on the multi-field data features of the input, the model optimizes the predicted process parameters in real time. The optimized parameter adjustment step size is ±0.5MPa for pressure and ±0.5℃ for temperature. The roller speed adjustment step size is set to ±0.05m / min. This adjustment step size can ensure the precision of parameter adjustment and avoid electrode structure damage caused by parameter mutations.

[0057] The response time of the edge computing unit is no more than 10ms, preferably ≤8ms, perfectly matching the cycle time requirements of industrial production lines and ensuring that parameter optimization commands can be sent to the roller press control system in real time. Simultaneously, the edge computing unit aggregates the real-time optimization parameters, multi-field data, and corresponding electrode performance test data for each batch, and feeds this data back to the digital twin model in step S2, with each iteration cycle consisting of 50 batches. This updates the model's training sample library, improves the model's long-term prediction accuracy, and forms a closed-loop control mechanism of "prediction-optimization-feedback-iteration".

[0058] The final output of this step is the optimal set of execution parameters for LFP electrode gradient compaction. This parameter set combines predicted parameters with real-time dynamic correction based on operating conditions, achieving a matching degree of over 99% with actual production needs compared to traditional fixed parameters. This optimal set of execution parameters can directly drive the roller press to perform gradient compaction operations according to the set parameters, effectively avoiding uneven compaction density caused by factors such as ambient temperature fluctuations and roller surface wear. This ensures the precise formation of a dense surface layer and a porous internal structure for the electrode, providing crucial parameter support for improving the energy density and cycle stability of the final LFP electrode.

[0059] S4: Adopt a graded self-healing microstructure roller surface and perform gradient compaction according to preset gradient compaction parameters.

[0060] The specific method of this step includes: a graded self-healing microstructure roller surface including a NiTi / DLC composite coating, and gradient-distributed primary microbumps and secondary microgrooves. The primary microbumps are distributed in a three-level gradient: the primary roller surface density is 300 bumps / cm². 2 Secondary roller surface density 200-100 particles / cm 2 30 particles / cm³ of surface density on the third-stage roller 2 The main micro-bumps have a diameter of 50-100μm and a height of 15-35μm, while the secondary micro-grooves have a width of 5μm and a depth of 1μm. The self-healing process of the graded self-healing microstructure roller surface is based on a wear grading threshold-triggered graded repair. For slight wear, the roller surface temperature is adjusted to 65℃ to initiate self-healing of the secondary micro-grooves. The slight wear is not less than 2μm, and the self-healing time does not exceed 5s. For severe wear, the roller surface temperature is adjusted to 70℃ to initiate shape memory repair of the main micro-bumps. The severe wear is not less than 5μm, and the repair time does not exceed 10s. Preset gradient compaction parameters include a compaction pressure of 8... For LFP materials with a particle size D50 of 1.0-1.4μm, the compaction pressure is 80-120MPa and the roller speed is 1.2-1.5m / min. For LFP materials with a particle size D50 of 1.4-1.8μm, the compaction pressure is 120-150MPa and the roller speed is 0.8-1.2m / min. For LFP materials with a particle size D50 of 1.8-2.2μm, the compaction pressure is 150-195MPa and the roller speed is 0.5-0.8m / min.

[0061] The material to be processed in this step is an LFP electrode sheet after coating and drying processes, derived from the output product of the previous drying process. Its initial state is: thickness 150-200μm, loose and porous surface, inner layer containing an incompletely dense active material particle stacking structure, electrode sheet width adapted to industrial production needs (optional 300-1600mm, preferably 1000mm), possessing the physical characteristics of continuous transport and compaction molding. The core composition of this electrode sheet is consistent with the LFP material detected in step S1, namely a mixed system of LFP active material, conductive agent, and binder, with the moisture content controlled to ≤0.5% after drying. The core purpose of this step is to rely on the special structure of the graded self-healing microstructure roller surface, combined with the optimal gradient compaction parameters output from step S3, to implement precise gradient compaction of the electrode sheet, achieving the ideal structure of "dense surface to improve energy density, porous inner layer to retain ion channels," while ensuring the stability of long-term compaction effect through the roller surface self-healing mechanism, laying the structural foundation for subsequent interface optimization processes. Roll surface wear detection is performed using a laser profilometer, which is activated once every 300m of electrode production. The detection accuracy of the micro-bump height is ±0.1μm. When the wear of the micro-bump height is detected to be ≥2μm (slight wear), the system triggers a polishing warning and automatically adjusts the compaction pressure of the corresponding section by 5-8MPa to compensate for wear loss. The wear detection data is stored in the intelligent system database in real time and is integrated with the roll surface self-repair status data to optimize the pressure compensation coefficient.

[0062] The core equipment used in this step is a three-stage tandem gradient compaction roller press. This equipment integrates three independent roller groups (first-stage, second-stage, and third-stage roller surfaces arranged in series), a servo hydraulic pressurization system, a roller surface temperature control system, a laser wear detection system, and a central control system. Each roller group can independently adjust pressure, roller speed, and temperature parameters. Guide positioning devices are installed at both ends of the roller surface to ensure that the electrode transfer deviation is ≤±0.5mm. Among them, the graded self-healing microstructure roller surface is the core functional component. The roller surface substrate is made of 40Cr alloy material, and the surface is coated with a NiTi / DLC composite coating by magnetron sputtering. The total coating thickness is 50-80μm, of which the NiTi alloy layer is 30-50μm thick to ensure shape memory self-healing performance; the DLC coating is 20-30μm thick to improve surface wear resistance, with a hardness ≥20GPa, preferably ≥25GPa, and a friction coefficient ≤0.15. The main micro-bumps and secondary micro-grooves are formed by laser engraving with a processing accuracy of ±0.1μm. The diameter of the main micro-bumps is preferably 60-80μm and the height is preferably 20-30μm. The width of the secondary micro-grooves is 5μm and the depth is 1μm, which are fixed optimal values. This size can accurately match the particle size of LFP active material, avoiding the situation where the groove is too large, resulting in a loose surface structure, or too small, which would prevent the electrolyte storage function from being realized.

[0063] The specific processing procedure is as follows: First, the electrode sheets to be compacted are smoothly fed into a three-stage series gradient compaction roller press via conveyor rollers. The central control system sends the optimal compaction parameters output by S3 to the servo hydraulic system and drive system of each roller group. For LFP materials of different particle sizes, the parameter adaptation logic is as follows: For small-particle-size materials with a particle size D50 = 1.0-1.4μm, due to the large specific surface area of ​​the particles and their tendency to agglomerate and clog channels, a lower compaction pressure of 80-120MPa (preferably 90-110MPa) and a higher roller speed of 1.2-1.5m / min (preferably 1.3-1.4m / min) are used to avoid over-compaction and damage to the inner pores; for medium-particle-size materials with a particle size D50 = 1.4-1.8μm, a medium compaction pressure of 120MPa is used. -150MPa (preferably 130-140MPa) and medium roller speed 0.8-1.2m / min (preferably 1.0-1.1m / min) are used to balance compaction and porosity; for large-particle materials with a particle size D50 of 1.8-2.2μm, due to the small contact area between particles, a higher compaction pressure of 150-195MPa (preferably 160-180MPa) and a lower roller speed of 0.5-0.8m / min (preferably 0.6-0.7m / min) are used to ensure tight bonding between particles.

[0064] The compaction effect of each roller surface is clearly defined: the density of main micro-bumps on the first-level roller surface is 300 per cm². 2 Under a set pressure, the surface layer of the electrode is initially densified to achieve a surface compaction density of 2.4-2.5 g / cm³. 3 The density of main micro-bumps on the secondary roller surface gradually decreases from 200 to 100 per cm², achieving a pressure transition from the surface to the inner layer and preventing electrode delamination due to sudden pressure changes; the density of main micro-bumps on the tertiary roller surface is 30 per cm². 2 The entire electrode sheet is shaped and compacted to achieve a stable surface compaction density of 2.5-2.6 g / cm³. 3 The inner layer porosity is retained at 30%-40%. Throughout the compaction process, the roller surface temperature control system maintains the roller surface temperature at 50-80℃ (preferably 60-70℃) through the built-in heating unit. This temperature range can improve the plasticity of the electrode material, reduce stress residue during the compaction process, and provide a temperature basis for roller surface self-healing.

[0065] The triggering and execution of the roll surface self-repair mechanism relies on a laser wear detection system and a temperature control system: the laser profilometer scans the roll surface in real time with a sampling frequency of 100Hz and a detection accuracy of ±0.01μm. When a slight wear of ≥2μm is detected, the central control system triggers a slight wear repair command, adjusting the corresponding roll surface temperature to 65℃ (preferably 63-64℃). The austenite transformation characteristics of NiTi alloy are used to achieve self-healing of secondary microgrooves, with a self-healing time of no more than 5s, preferably 3-4s. This process does not require machine shutdown and does not affect the production cycle. When a severe wear of ≥5μm is detected, a severe wear repair command is triggered, adjusting the roll surface temperature to 70℃ (preferably 68-69℃) to achieve shape memory repair of the main micro-bumps. The repair time is no more than 10s, preferably 7-8s. After the repair is completed, the laser profilometer checks again, and after confirming that the repair accuracy meets the standard, the normal compaction parameters are restored.

[0066] The final output of this step is an LFP compacted electrode with a gradient structure, whose core parameters are: surface compaction density 2.5-2.6 g / cm³. 3 The inner layer porosity is 30%-40%, the electrode thickness consistency error is ≤±0.003mm, and the surface layer has a complete microgroove structure with no obvious cracks or delamination defects. The gradient structure of this electrode effectively solves the pain point of traditional compaction processes where "high compaction density and high ion conductivity are difficult to achieve simultaneously." The dense surface structure can improve the battery energy density, while the porous inner layer and secondary microgrooves provide channels and storage space for subsequent electrolyte filling. At the same time, the self-healing mechanism of the roller surface extends the service life of the roller surface from the traditional 1 million m to more than 3.5 million m, significantly reducing the industrial operation and maintenance costs.

[0067] Between S4 and S5 at the headquarters, there is a post-heat preservation process: the compacted electrode is placed in a 60℃ constant temperature chamber for 2 hours to release the internal stress generated during the compaction process. During the heat preservation process, the internal stress data of the electrode is detected in real time by the built-in stress sensor, and the detection data is fed back to the digital twin model for subsequent process parameter iteration and optimization.

[0068] S5: Perform electrolyte precursor micro-spraying and infrared curing treatment on the compacted electrode.

[0069] The specific steps of this method include: dynamically adjusting the spraying amount of the electrolyte precursor micro-spraying according to the real-time porosity of the electrode, with a spraying amount range of 0.1-0.3 μL / cm. 2 Control accuracy ±0.01μL / cm 2 The uniformity error of the spraying should not exceed ±5%, and the infrared curing parameters are 80℃ temperature and 5s time.

[0070] The material to be processed in this step is the LFP gradient compacted electrode sheet output from step S4, which originates from the discharge end of a three-stage series gradient compaction roller press. Its core characteristic is a gradient structure with a "dense surface and porous inner layer," and a surface compaction density of 2.5-2.6 g / cm³. 3 The inner layer porosity is 30%-40%, and the surface layer has a complete microgroove structure. The electrode thickness is reduced to 80-120μm after compaction (adapted to different particle sizes: 100-120μm for small particles, 90-110μm for medium particles, and 80-100μm for large particles). The electrode width is consistent with S4 (300-1600mm, preferably 1000mm). The surface is free of cracks and delamination defects, and it has the physical characteristics of continuous transmission and withstanding micro-spraying and low-temperature curing. The core purpose of this step is to fill the electrode pores and surface layer microgrooves with precise micro-spraying of the electrolyte precursor, and then form a stable interface transition layer through low-temperature infrared curing. This solves the core pain point of "poor electrode-electrolyte interface bonding and high impedance" in semi-solid LFP batteries, while avoiding high-temperature damage to the crystal structure of LFP active materials, thus providing interface protection for the final preparation of high-performance LFP electrodes.

[0071] The core equipment used in this step is an integrated system of "online porosity detection-micro-spraying-infrared curing". This system integrates an online porosity detector, a high-precision micro-spraying unit, an infrared curing unit, and a spraying uniformity detection module sequentially along the electrode transport direction. Each unit is linked through a central control system with a response delay ≤10ms, precisely adapted to the industrial production cycle (15-20m / min). The online porosity detector uses X-ray attenuation method, with a detection range of 10%-60%, a detection accuracy of ±0.5%, and a sampling frequency of 50Hz. It can collect porosity data from different areas of the electrode in real time and transmit it to the micro-spraying unit. The high-precision micro-spraying unit consists of an array of micro-nozzles, a servo metering pump, a precursor storage tank, and a temperature insulation device. The number of nozzles in the micro-nozzle array can be adapted according to the electrode width (30 nozzles for a 300mm wide electrode, 160 nozzles for a 1600mm wide electrode), with nozzle orifice diameters of 50-100μm, preferably 80μm. The distance between the electrodes is adjustable (5-15mm optional, 10mm preferred), and the flow rate of the servo metering pump is adjustable from 0.01-1μL / s with a control accuracy of ±0.001μL / s, ensuring precise control of the coating amount. The infrared curing unit uses a short-wave infrared heater with an adjustable heating power range of 1-5kW. The heating area covers the full width of the electrode and is equipped with a closed-loop temperature control system with a temperature control accuracy of ±0.5℃, which can stably maintain a curing temperature of 80℃. The coating uniformity detection module uses laser-induced fluorescence method with a detection accuracy of ±1%, which can provide real-time feedback on the coating effect and trigger parameter correction.

[0072] The electrolyte precursor used in this step is a special system adapted for semi-solid LFP batteries. Its composition is as follows: 10%-15% lithium salt (such as lithium bis(trifluoromethanesulfonyl)imide, LiTFSI), 60%-70% organic solvent (such as ethylene carbonate and dimethyl carbonate mixed in a volume ratio of 1:1), 5%-10% crosslinking agent (such as polyethylene glycol diacrylate), and 2%-5% additive (such as fluoroethylene carbonate). The precursor has a viscosity range of 5-15 mPa·s (25℃), good fluidity and film-forming properties, and can form a stable ion-conducting interface layer after low-temperature curing.

[0073] The specific processing procedure is as follows: First, the compacted electrode sheet smoothly enters the integrated system via conveyor rollers. The online porosity detector performs real-time porosity detection on the electrode sheet across its entire width and coverage. The detection data is transmitted to the central control system in real time. The system calculates the required coating amount for the current area based on a preset algorithm. When the detected porosity increases by 1%, the control system instructs the servo metering pump to increase the coating amount by 0.02 μL / cm². 2 This ensures that the electrolyte precursor precisely fills the pores without excessive accumulation. Subsequently, the electrode enters the micro-spraying area, where an array of micro-nozzles synchronously sprays the electrolyte precursor along the width of the electrode. The nozzle movement speed is linked to the electrode transport speed (when the electrode transport speed is 15-20 m / min, the nozzle dynamic adjustment frequency is 100 Hz). During spraying, a temperature control device maintains the precursor temperature at 25-30℃ to prevent viscosity changes from affecting spraying accuracy. Ultimately, dynamic control of the spraying amount is achieved within the range of 0.1-0.3 μL / cm², with the preferred spraying amount being 0.15-0.25 μL / cm². 2 This range can balance interfacial bonding force and ion conduction efficiency.

[0074] After coating, the electrode immediately enters the infrared curing unit. A short-wave infrared heater rapidly heats the electrode to 80°C and maintains a stable temperature. The electrode remains at this temperature for 5 seconds (the dwell time is precisely controlled by adjusting the transmission speed; for example, when the electrode transmission speed is 0.083 m / s, the curing area length is set to 0.415 m), achieving rapid cross-linking and curing of the electrolyte precursor. During curing, a closed-loop temperature control system monitors the electrode surface temperature in real time. If fluctuations exceeding ±0.5°C occur, the heater power is immediately adjusted for correction. After curing, a coating uniformity detection module performs a secondary inspection of the electrode. If a uniformity error exceeds ±5%, the system triggers an alarm and fine-tunes the coating parameters to ensure product consistency.

[0075] The final output of this step is a pretreated LFP electrode produced by "gradient compaction + interface curing." Its core parameters are: a uniformly filled, cured interface transition layer within the surface layer's microgrooves and inner pores, with a transition layer thickness of 50-100 nm and a uniformity error ≤ ±5%; an electrode interface impedance ≤ 5 Ω, more than 40% lower than the untreated compacted electrode; and an electrode thickness consistency error maintained at ≤ ± 0.003 mm, with no deformation or adhesion defects. This pretreated electrode exhibits good compatibility with semi-solid electrolytes. Its stable interface transition layer effectively reduces interfacial side reactions during battery cycling, improving cycle stability. Simultaneously, the retained inner porous structure and secondary microgrooves ensure efficient ion transport, laying a core foundation for subsequent battery assembly and performance enhancement. Meanwhile, an interface performance testing step has been added. An electrochemical impedance spectroscopy (detection accuracy ±0.01Ω) is used to detect the electrode-electrolyte interface impedance, and a peel tester (detection accuracy ±0.1N / cm) is used to detect the interface bonding strength, ensuring that the interface bonding strength is ≥5N / cm. The test data is synchronized to the edge computing unit and digital twin model as the basis for interface adaptability optimization. When the interface impedance exceeds 5Ω or the bonding strength is lower than 5N / cm, the system triggers parameter backtracking optimization and adjusts the micro-spraying amount and roller surface microstructure parameters.

[0076] S6: Establish a full-chain data traceability system.

[0077] The specific methods in this step include: a full-chain data traceability system that links data through a unique traceability code for each batch of electrodes. The linked data includes material testing data, process control parameters, equipment operation data, and electrode performance testing data, and the data storage period is no less than 3 years.

[0078] The product to be processed in this step is the LFP pre-treated electrode output from step S5, originating from the discharge end of the integrated system of "online porosity detection-micro-spraying-infrared curing". Its core characteristics are: a gradient structure of "dense surface - porous inner layer", with surface-level micro-grooves and inner-layer pores filled with a cured interface transition layer; interface impedance ≤5Ω; electrode thickness consistency error ≤±0.003mm; electrode width 300-1600mm (preferably 1000mm); and interface optimization processing completed, enabling continuous coding and data association. The core purpose of this step is to assign a unique identifier to each batch of pre-treated electrodes, linking key data throughout the entire process from material input to pre-treatment completion. This addresses the pain points of traditional LFP electrode production, such as "difficulty in tracing quality issues and lack of data support for process optimization," while simultaneously meeting the policy and industry requirements for full lifecycle quality control in the new energy battery industry. This provides a data foundation for subsequent quality traceability and closed-loop process optimization of finished electrodes.

[0079] The core equipment used in this step is a full-chain data traceability system. This system consists of a traceability code generation module, a laser coding unit, a multi-source data acquisition gateway, an industrial distributed database server, and a data management platform. All components achieve real-time data interaction via industrial Ethernet (using the MQTT communication protocol), with a data transmission latency of ≤20ms, adaptable to industrial production cycles. The traceability code generation module uses an encryption algorithm to generate unique batch traceability codes, with the coding rule being "factory code-production line code-production date-batch serial number," and a code length of 18-24 digits. Two coding methods are available: QR code or RFID tag, with QR code being preferred (recognition accuracy ≥99.9%, temperature resistance range -40℃-120℃, suitable for subsequent battery assembly and warehousing environments). The laser coding unit uses a fiber laser coding machine with a coding power of 5-30W (preferably 20W), a coding speed of 100-500 pieces / minute, and coding positions in the inactive areas of the electrode edge (width 5-10mm), with a coding depth of 0.01-0.03mm. This ensures clear traceability codes without damaging the active areas of the electrodes; the multi-source data acquisition gateway has 8-16 data acquisition interfaces, which can simultaneously connect to the testing equipment, control system, and equipment controllers of each step from S1 to S5, and the data acquisition frequency can be selected from 10-100Hz (preferably 50Hz) to ensure data integrity and real-time performance; the industrial distributed database server has a storage capacity of ≥10TB, supports hot backup, and has a data read / write speed of ≥1GB / s to ensure stable storage of massive amounts of production data; the data management platform has functions such as data association, query, statistics, and export, supports multi-dimensional retrieval by traceability code, production date, batch, etc., and has a response time of ≤1s.

[0080] The specific processing procedure is as follows: First, when a single batch of LFP material enters the S1 step for inspection, the traceability code generation module simultaneously generates a unique traceability code for that batch. At the same time, the traceability code is bound to batch information (factory code, production line code, production date, planned output, etc.) and entered into the data management platform. Subsequently, in the S6 step, the pre-processed electrodes are smoothly transported to the laser coding unit via conveyor rollers. The system adjusts the coding rhythm according to the electrode transport speed (maintaining consistency with the previous steps, 15-20 m / min). The laser coding machine accurately prints traceability codes on the inactive areas at the edge of the electrode, printing one traceability code every 1-2 m (preferably 1.5 m), ensuring that each segment of a single roll of electrode can be associated with the corresponding batch through the traceability code.

[0081] After coding is completed, the multi-source data acquisition gateway begins to synchronously collect and integrate the related data from each step: the material detection data includes the LFP material particle size D50, solid content, moisture content, and crystal integrity detection data in step S1; the process control parameters include the digital twin model prediction parameters in step S2, the edge computing optimization parameters in step S3, the three-level gradient compaction parameters (pressure, roller speed, roller surface temperature) and roller surface self-healing records in step S4, and the electrolyte precursor spraying amount, infrared curing parameters, and spraying uniformity detection data in step S5; the equipment operation data covers the operating status parameters (voltage, current, power), fault alarm records, and maintenance records of the core equipment (online detection equipment, digital twin server, edge computing unit, three-level roller press, and micro-spraying-curing integrated system) in each step S1-S5; the electrode performance detection data includes the core performance indicators of LFP electrode surface compaction density, inner layer porosity, interfacial impedance, ion diffusion coefficient, and cycle life in the subsequent finished product testing stage. This part of the data is added to the related data of the corresponding traceability code after the finished product testing is completed.

[0082] All collected data is transmitted to an industrial distributed database server via the MQTT protocol. The system automatically binds the data with the corresponding batch traceability codes, forming a complete data archive of "traceability code - material - process - equipment - performance". Data storage adheres to the requirement of "no less than 3 years", with a preferred storage period of 5 years. Historical data exceeding the storage period can be archived through cold backup to ensure data traceability. Simultaneously, the data management platform updates the associated data for each batch in real time. Production managers can query the full-process data corresponding to a specific traceability code at any time through the platform. When quality issues arise, the traceability code can be used to quickly locate key information such as material characteristics, process parameters, and equipment status of the problematic batch, enabling precise troubleshooting.

[0083] The final output of this step is an LFP pre-processed electrode with a unique traceability code. This electrode retains the core performance parameters of step S5 (interface impedance ≤ 5Ω, thickness consistency error ≤ ±0.003mm, and complete gradient structure) and also possesses full-chain data traceability attributes. Its core benefits are: achieving full lifecycle data traceability from LFP material input to pre-processed electrode output, reducing quality problem investigation time from the traditional 2-4 hours to less than 10 minutes; the associated full-process data can serve as iterative training data for the digital twin model of step S2, providing a basis for continuous process optimization; simultaneously meeting the quality control standards of the new energy battery industry, enhancing the product's market competitiveness and rights protection value, and providing key data support for the subsequent industrial production and quality assurance of finished electrodes. The digital twin model undergoes an algorithm iteration every 50 batches of electrodes, using a gradient boosting tree algorithm to optimize the model's parameter prediction formula. Each iteration improves prediction accuracy by 0.5%-1%; the iteration process is based on 5000+ batches of 'material parameters - process parameters - performance data' in the traceability database, and the iteration process does not affect normal production.

[0084] S7: Process stability control.

[0085] The specific methods in this step include: real-time monitoring of electrode deformation rate; when the deformation rate is not less than 0.02 mm / s, reducing the roller speed by 0.05 m / min and simultaneously reducing the compaction pressure by 2 MPa; real-time monitoring of electrode surface flatness; when the surface flatness error is not less than 0.01 mm, triggering parameter backtracking correction.

[0086] The product processed in this step is the LFP pre-processed electrode with a unique traceability code, output from step S6. Originating from the output end of the full-chain data traceability system, its core characteristics are consistent with the S6 output: a gradient structure of "dense surface - porous inner layer," with a solidified interface transition layer filling the surface-level microgrooves; interface impedance ≤5Ω; electrode thickness consistency error ≤±0.003mm; and electrode width 300-1600mm (preferably 1000mm). It is currently in a continuous industrial production chain, possessing the physical conditions for real-time monitoring and dynamic control. The core purpose of this step is to quickly respond to emergency conditions by real-time monitoring of key mechanical and morphological indicators during electrode transport and forming, avoiding defects such as electrode cracking and deformation caused by process fluctuations. This further ensures the stability of the entire process, guarantees that the final LFP electrode meets high consistency requirements, and improves the yield.

[0087] The core equipment used in this step is the "real-time monitoring-dynamic control" linkage system. This system integrates a deformation rate monitoring unit, a surface flatness monitoring unit, and an emergency control execution module. It is deeply linked with the three-level series gradient compaction roller press control system in the previous S4 step and the full-chain data traceability system in the S6 step. The data transmission adopts a dual-link redundancy design of industrial Ethernet + real-time Ethernet, and the linkage response delay is ≤5ms, ensuring the immediate issuance of control commands in emergency situations. The deformation rate monitoring unit employs a high-precision laser displacement sensor, installed above the transmission path between the discharge end of the three-stage roller press in step S4 and the micro-spraying unit in step S5. The sensor has a measurement range of 0-1 mm / s, a measurement accuracy of ±0.001 mm / s, and a sampling frequency selectable from 100-500 Hz (preferably 300 Hz), enabling real-time capture of deformation rate changes during electrode transmission. The surface flatness monitoring unit uses a line laser profilometer, installed between the infrared curing unit in step S5 and the laser coding unit in step S6. It has a measurement range of 0-5 mm, a measurement accuracy of ±0.001 mm, a scanning width covering the full width of the electrode sheet, and a sampling frequency of 50-200 Hz (preferably 100 Hz), allowing for rapid identification of surface flatness defects such as unevenness and wrinkles. The emergency control execution module is built into the central control system of the three-stage roller press, featuring rapid parameter adjustment and historical parameter retrieval functions. It can directly drive the servo hydraulic system and roller speed drive system to execute control commands.

[0088] The specific processing procedure is as follows: This monitoring and control step runs through the critical transmission stages from S4 to S6, achieving closed-loop control of "real-time monitoring - threshold judgment - emergency response". In the deformation rate monitoring stage, the laser displacement sensor continuously collects displacement change data at a fixed monitoring point on the electrode surface, and the system calculates the deformation rate in real time through a built-in algorithm. When the calculation result shows that the deformation rate is not less than 0.02 mm / s (this threshold is set based on the mechanical properties of the LFP electrode; the elastic modulus of the LFP electrode is ≥15 GPa, and exceeding this deformation rate can easily cause cracking due to stress concentration), the emergency control execution module immediately triggers a first-level emergency response: a speed reduction command is sent to the roller speed drive system of the three-stage roller press to reduce the current roller speed by 0.05 m / min, and a pressure reduction command is sent to the servo hydraulic system to simultaneously reduce the compaction pressure by 2 MPa. After adjustment, the sensor continuously monitors the deformation rate until it drops below 0.02 mm / s, at which point the system automatically returns to the original optimized parameter range. If the deformation rate still fails to meet the standard after adjustment, the system will trigger adjustment again, up to a maximum of 3 consecutive adjustments. If it is still ineffective, an audible and visual alarm will be triggered and production of that batch will be suspended to avoid batch defects.

[0089] In the surface smoothness monitoring stage, the line laser profilometer performs a full-width scan along the electrode width direction, generating real-time three-dimensional profile data of the electrode surface. The system calculates the surface smoothness error by comparing it with a standard profile model. When the detected smoothness error is not less than 0.01mm (this error threshold corresponds to the integrity requirement of the microgroove structure on the electrode surface; exceeding this error will lead to uneven spraying of the electrolyte precursor, affecting the interface bonding effect), the emergency control execution module triggers a level-two emergency response: initiating the parameter retrospective correction function. The system quickly retrieves the historical optimal process parameters from previous production stages of this batch (i.e., the S3 optimization parameters, S4 compaction parameters, and S5 spraying parameters corresponding to the minimum surface smoothness error stage in this batch) using the unique traceability code of step S6. The backtracked parameters are then sent to the corresponding equipment as new execution parameters, while parameter iteration updates are paused until the surface smoothness error drops below 0.01mm, at which point normal parameter optimization logic resumes. During the parameter backtracking process, the system automatically associates the backtracking records, parameters before and after correction, and flatness data with the corresponding traceability codes and incorporates them into the full-chain data archive, providing defect case data support for subsequent process parameter optimization.

[0090] The final output of this step is a highly stable LFP pre-treated electrode, whose core parameters are further optimized compared to previous steps: the electrode cracking rate is reduced from 0.5% in traditional processes to below 0.1%, the surface flatness error is stably controlled within 0.01mm, while retaining the gradient structure of "dense surface - porous inner layer", an interface impedance of ≤5Ω, and a thickness consistency error of ≤±0.003mm. The core benefits of this product are: through rapid response and precise control in emergency situations, it significantly reduces the defect rate caused by process fluctuations, ensuring consistency in mass production; at the same time, the linkage between control data and the traceability system further improves the correlation logic between "process and quality", providing key emergency case data for continuous optimization of the entire process, ultimately ensuring that the produced LFP pre-treated electrode fully meets the high consistency requirements of subsequent finished product processing, laying a stable process foundation for the preparation of high-performance LFP batteries.

[0091] The following describes the verification method for the effectiveness of the intelligent control process of LFP electrode gradient compaction.

[0092] This efficacy verification method aims to systematically verify the effectiveness of the deeply integrated LFP electrode gradient compaction intelligent control process (including the original basic scheme and primary, secondary, and tertiary improvements) by designing a gradient control experiment using the controlled variable method, combined with theoretical derivation and quantitative data. It clarifies the causal relationship between each improvement element of the process method and the performance of the final advanced LFP electrode product, proving that this process method can stably prepare LFP electrodes with a surface-level microgroove structure, gradient compaction morphology, and excellent electrochemical performance. The verification process strictly follows a rigorous procedure of "benchmark setting—variable control—experiment implementation—data comparison—logical derivation—conclusion judgment" to ensure the reliability and persuasiveness of the verification results.

[0093] First, the experimental prerequisites were clearly defined, and all experimental variables were standardized to eliminate interference factors and ensure the effectiveness of comparisons between different schemes. The LFP materials used in the experiments were standardized as follows: LFP active substances were divided into three groups according to particle size (1.0-1.4μm, 1.4-1.8μm, and 1.8-2.2μm); the conductive agent adopted a Super P:CNT=3:1 composite system; the binder adopted a PVDF-HFP:LA133=2:1 composite system; the formula ratio was fixed at LFP:conductive agent:binder=89:3.5:2.5; and the slurry viscosity was controlled at 2000-3000 mPa·s. The basic equipment parameters were standardized as follows: roller diameter (300mm for stage 1, 400mm for stage 2, and 300mm for stage 3); maximum hydraulic system pressure 200MPa; temperature control range 40-100℃; and environmental conditions were standardized as follows: room temperature 25±2℃, humidity 45±5%, and roller unit vibration ≤0.02mm / s. The core testing indicators are divided into two categories: one is the core performance indicators of the product, including surface compaction density, inner layer porosity, thickness consistency error, interfacial impedance, and ion diffusion coefficient; the other is industrialization-related indicators, including electrode cracking rate, roller surface service life, and production cost per Wh. The testing methods for each indicator adopt industry standard methods. Among them, compaction density is measured by the water displacement method, ion diffusion coefficient is measured by electrochemical impedance spectroscopy, thickness consistency is measured by taking the average value of 100 test points with a laser thickness gauge, interfacial impedance is measured by an electrochemical workstation, and cycle life is measured by a battery cycle test system (1C / 1C room temperature).

[0094] Based on the principle of controlled variables, six gradient control experimental schemes were designed to gradually replace the technical means of the baseline scheme with the improved elements of the deeply integrated scheme, in order to verify the independent and synergistic effects of each element. The specific components of each scheme are as follows: The baseline scheme (CK0) adopts conventional fixed parameter gradient compaction + ordinary flat stainless steel roller surface + manual sampling + interface-free optimization process as a blank control baseline; Scheme CK1 replaces the "fixed parameter compaction" of the baseline scheme with an improved intelligent adaptive gradient compaction parameter control system (including online material detection + dynamic parameter matching), and replaces the "flat roller surface" with an improved biomimetic asymmetric microstructure roller surface, verifying the independent effect of the improved elements; Scheme CK2, based on CK1, replaces the "bionic asymmetric microstructure roller surface" with a second-improved NiTi / DLC composite coating roller surface, and adds a second-improved full-process predictive control system based on digital twin, verifying the synergistic effect between the second-improvement elements and the first-improvement elements. Results: The CK3 scheme, based on CK2, replaces the "NiTi / DLC roller surface" with a composite roller surface featuring three improvements: graded micro-bumps and secondary microstructures. It also adds a three-stage improved edge-computing-based real-time control system for thermo-electric-mechanical multi-field coupling, verifying the synergistic effect of the three improved elements (C1+C2) with the preceding improved elements. The CK4 scheme, based on CK3, adds a three-stage improved electrode-electrolyte interface compatibility optimization process (micro-spraying + infrared curing), verifying the independent effect of the C3 element and the preliminary effect of the overall element superposition. The deep fusion scheme (Fusion) integrates all improved elements (A1+A2+B1+B2+C1+C2+C3) and performs synergistic optimization of each element's parameters, verifying the final effect of the deep fusion of all elements. Each scheme has three parallel experiments, and the average data is used to reduce random errors. All testing equipment is pre-calibrated, and outlier data is removed using the Grubbs criterion (significance level α=0.05). Three different batches of materials are selected for repeated experiments to ensure data repeatability.

[0095] After the experiment was conducted, the differences in the effectiveness of each scheme were analyzed by quantitative data comparison. The specific experimental data are shown in the table below: Table 1: Experimental data from 6 groups of gradient control trials

[0096] Based on the above data and theoretical derivation, the logic of how each improvement element of the process method enhances product performance can be clearly defined, thus proving that this process method can produce advanced LFP electrode products. From the perspective of the effects of a single improvement element, the CK1 scheme improves surface compaction density by 8.7%-8.9%, ion diffusion coefficient by 33.3%-46.7%, and electrode cracking rate by 71.4%-85.7% compared to the baseline scheme (CK0). This is because the intelligent adaptive gradient compaction parameter control system, based on closed-loop feedback control theory, dynamically matches compaction parameters by real-time detection of material characteristics such as LFP particle size and solid content, avoiding the problem of "insufficient compaction of large-diameter materials and over-compaction of small-diameter materials" in traditional fixed-parameter compaction. Furthermore, the biomimetic asymmetric microstructure roller surface, based on biomimetic mechanics and contact mechanics theory, achieves point pressure enhancement through gradient-distributed main micro-protrusions. While improving surface densification, it reduces inner layer pressure by 30%-40%, effectively preserving ion channels, thereby achieving a synergistic improvement in compaction density and ion conductivity.

[0097] The combination of secondary improvement elements further optimized product performance. Compared with the CK1 solution, the thickness consistency error of the CK2 solution was reduced by 50%, and the service life of the roller surface was increased by 76.9%-83.3%. The core reason is that the NiTi / DLC composite coating roller surface is based on the shape memory alloy theory. Through temperature control, the height of the micro-bumps can be adaptively adjusted within the range of 15-35μm. After wear, it can be restored by heating to 70℃ and relying on the shape memory effect. The high wear resistance of the DLC coating also further extends the service life of the roller surface. The predictive control system based on digital twins predicts defects such as insufficient porosity and cracking in advance and corrects parameters through real-time mapping between physical entities and virtual models. This solves the lag problem of "detection-feedback-adjustment" in the first improvement, making the defect prediction accuracy rate reach more than 95% and significantly improving product consistency.

[0098] The integration of three improvement elements has achieved a comprehensive leap in product performance. Compared with the CK2 scheme, the thickness consistency error of the CK3 scheme has been further reduced by 20%, and the ion diffusion coefficient has been improved by 4.5%-8.7%. This is due to the composite design of hierarchical micro-bumps and secondary microstructures. The main micro-bumps ensure macroscopic compaction uniformity, while the secondary micro-grooves (5μm wide and 1μm deep) can store electrolyte precursors based on interface chemistry and micro-nano manufacturing theory, increasing the interfacial contact area between the electrode and the electrolyte. The multi-field coupling real-time control system based on edge computing integrates multi-field data such as ambient temperature and roller vibration through local real-time data processing, reducing the parameter control response time from 200ms to less than 50ms. This solves the problem of small delays in the iteration of the digital twin model and controls the compaction uniformity error within ±0.5%. After the addition of an interface compatibility optimization process, the CK4 solution reduces the interface impedance by 33.3%-40% compared to the CK3 solution, and improves the cycle life by 5.6%-11.1%. This is because the micro-spraying process ensures that the electrolyte precursor uniformly covers the electrode surface (spraying amount 0.1-0.3μL / cm², control accuracy ±0.01μL / cm²), and the 80℃ infrared curing, based on polymer chemistry theory, causes the precursor to crosslink and form a stable interface transition layer, effectively improving the solid-liquid interface wetting performance and adhesion strength, and solving the problem of uneven electrolyte wetting in traditional immersion methods.

[0099] The deep fusion approach integrates all improvement elements and achieves synergistic parameter optimization, resulting in LFP electrode products with optimal performance across all aspects, including a surface compaction density of 2.55-2.6 g / cm³. 3 The inner layer porosity is 36-40%, the thickness consistency error is ±0.003mm, the interfacial resistance is 2.5-3.5Ω, and the ion diffusion coefficient is 2.5-2.6×10⁻⁶. -10 cm 2 / s, fully meeting the performance requirements of advanced LFP electrodes. The various improvement elements do not act in isolation, but form a complete closed loop of "material detection - predictive control - gradient compaction - interface optimization - data traceability". With the synergistic support of theories from multiple fields such as cybernetics, materials mechanics, digital twins, and interface science, the core objectives of "surface densification to improve energy density, inner porousness to retain ion channels, and interface optimization to reduce impedance" have been achieved.

[0100] In summary, through quantitative data comparison and rigorous theoretical derivation of gradient control experiments, it can be fully demonstrated that the proposed intelligent control process for gradient compaction of LFP electrodes can stably produce LFP electrode products with preset advanced performance. The effects of each improvement element are traceable and quantifiable. The process scheme with deep integration of all elements has achieved the goal of improving core performance indicators by ≥20% and industrialization indicators by ≥15% compared with the traditional benchmark scheme. The advancement and effectiveness of the process method have been fully verified.

[0101] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, because according to the embodiments of this application, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to this application.

[0102] Secondly, this application discloses an LFP electrode, which is prepared by the LFP electrode gradient compaction intelligent control process described in any of the first aspects of this application. The electrode surface has a secondary microgroove structure, and the grooves are filled with an interface transition layer formed by the solidification of an electrolyte precursor. The electrode cross-section exhibits a significant gradient morphology of dense surface and porous inner layer.

[0103] Its core performance parameters meet the following requirements: surface compaction density is 2.5-2.6 g / cm³. 3 The inner layer porosity is maintained at 30%-40%, and the thickness consistency error does not exceed ±0.003mm; the interfacial impedance does not exceed 5Ω, and the ion diffusion coefficient is not less than 2.5×10⁻⁶. -10 cm 2 / s, combining excellent energy density and ion conduction performance.

[0104] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of disclosure in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the foregoing disclosed concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A method for intelligent control of gradient compaction of LFP electrodes, characterized in that, Includes the following steps: Multi-dimensional online detection of LFP materials; The detection data is input into the digital twin model to obtain the predicted process parameters; By fusing multi-field data through edge computing units, the predicted process parameters can be optimized in real time. A graded self-healing microstructure roller surface is used, and gradient compaction is carried out according to preset gradient compaction parameters; The compacted electrode is subjected to electrolyte precursor micro-spraying and infrared curing treatment; Establish a full-chain data traceability system; The preset gradient compaction parameters include a compaction pressure of 80-195 MPa and a roller speed of 0.5-1.5 m / min, and the response time of the edge computing unit does not exceed 10 ms.

2. The intelligent control process for LFP electrode gradient compaction according to claim 1, characterized in that, The multi-dimensional online detection includes simultaneous detection of LFP material particle size D50, solid content, moisture content and crystal integrity; The particle size D50 is 1.0-2.2 μm, the solid content is 50%-60%, and the moisture content does not exceed 0.5%. The integrity of the crystal was detected by X-ray diffraction, and the full width at half maximum (FWHM) of the characteristic peaks did not exceed 0.2°. The test data is input into the digital twin model through a standardized interface.

3. The intelligent control process for LFP electrode gradient compaction according to claim 1, characterized in that, The edge computing unit integrates multi-field data including ambient temperature, roller vibration, and interface voltage. The predicted process parameters were optimized using a CNN-LSTM fusion model. The optimized parameter control step size was ±0.5MPa for pressure and ±0.5℃ for temperature. The digital twin model iterates every 50 batches, and the edge computing unit feeds back the optimized data to the digital twin model in real time.

4. The intelligent control process for LFP electrode gradient compaction according to claim 1, characterized in that, The graded self-healing microstructure roller surface includes a NiTi / DLC composite coating, as well as gradient-distributed primary microbumps and secondary microgrooves; The main micro-bumps are distributed in a three-level gradient: the first level has a roller surface density of 300 bumps / cm². 2 Secondary roller surface density 200-100 particles / cm 2 30 particles / cm³ of surface density on the third-stage roller 2 ; The main micro-bumps have a diameter of 50-100μm and a height of 15-35μm, while the secondary micro-grooves have a width of 5μm and a depth of 1μm.

5. The intelligent control process for LFP electrode gradient compaction according to claim 4, characterized in that, The self-healing process of the graded self-healing microstructure roller surface is as follows: graded repair is triggered based on the wear graded threshold. When there is slight wear, adjust the roller surface temperature to 65℃ to perform secondary microgroove self-healing. The amount of slight wear is not less than 2μm and the self-healing time is not more than 5s. When severe wear occurs, adjust the roller surface temperature to 70℃ and perform shape memory repair of the main micro-bumps. The amount of severe wear should not be less than 5μm, and the repair time should not exceed 10s.

6. The intelligent control process for LFP electrode gradient compaction according to claim 1, characterized in that, The amount of electrolyte precursor micro-spraying is dynamically adjusted according to the real-time porosity of the electrode. The spraying amount ranges from 0.1 to 0.3 μL / cm. 2 Control accuracy ±0.01μL / cm 2 The uniformity error of the spraying should not exceed ±5%; The infrared curing parameters are: temperature 80℃ and time 5s.

7. The intelligent control process for LFP electrode gradient compaction according to claim 1, characterized in that, The full-chain data traceability system associates data through a unique traceability code for each batch of electrodes; The associated data includes material testing data, process control parameters, equipment operation data, and electrode performance testing data; The data storage period shall not be less than 3 years.

8. The intelligent control process for LFP electrode gradient compaction according to claim 1, characterized in that, It also includes a wide material adaptation adjustment step: For LFP materials with a particle size D50 of 1.0-1.4μm, a compaction pressure of 80-120MPa and a roller speed of 1.2-1.5m / min are adopted. For LFP materials with a particle size D50 of 1.4-1.8μm, a compaction pressure of 120-150MPa and a roller speed of 0.8-1.2m / min are adopted. For LFP materials with a particle size D50 of 1.8-2.2μm, a compaction pressure of 150-195MPa and a roller speed of 0.5-0.8m / min are used.

9. The intelligent control process for LFP electrode gradient compaction according to claim 1, characterized in that, It also includes process stability control steps: The electrode deformation rate is monitored in real time. When the deformation rate is not less than 0.02 mm / s, the roller speed is reduced by 0.05 m / min and the compaction pressure is reduced by 2 MPa simultaneously. The surface flatness of the electrode is monitored in real time. When the surface flatness error is not less than 0.01mm, parameter backtracking correction is triggered.

10. An LFP electrode, characterized in that, The LFP electrode was prepared using the intelligent control process for gradient compaction of the LFP electrode as described in any one of claims 1-9. The LFP electrode surface has a secondary microgroove structure, and the grooves are filled with an interface transition layer formed by the solidification of the electrolyte precursor. The electrode cross-section has a gradient structure of dense surface and porous inner layer. The surface compaction density of the LFP electrode is 2.5-2.6 g / cm³. 3 The inner layer porosity is 30%-40%, and the thickness consistency error does not exceed ±0.003mm; The interfacial impedance of the LFP electrode does not exceed 5Ω, and the ion diffusion coefficient is not less than 2.5×10⁻⁶. -10 cm 2 / s.

Images