Internal resistance-capacity-self-discharge three-dimensional consistency grading method and system for aviation cylindrical quasi-solid-state battery cell and matching rule

Abstract

The invention discloses an internal resistance-capacity-self-discharge three-dimensional consistency grading method and system for an aviation cylindrical quasi-solid-state lithium cell and a matching rule. The invention aims to solve the technical problems that an aviation battery pack only needs thousands of to thousands of battery cells in a single time, the candidate range is far smaller than a million level of vehicle regulations, and extremely high consistency still needs to be obtained under the same batch production condition to meet intrinsic safety and airworthiness verification repeatability. The method comprises the following steps: pre-grouping an empty cell mass reference narrow window, performing multi-stage liquid injection and liquid supplement closed loop by taking a mass drop point after liquid injection as a control target, performing vacuum degassing infiltration, performing in-situ curing on an A / B bi-component flame-retardant quasi-solid electrolyte, and performing consistency control on a blended LLZO ceramic electrolyte. And forming a three-dimensional grading result of an internal resistance label, a capacity label and a self-discharge label through three-stage gating grading of steady-state open-circuit voltage and internal resistance gating-high-temperature standing self-discharge gating-fine capacity grading gating after formation, and generating a grouping scheme according to a rule that the internal resistance is preferentially matched in parallel, and the capacity is preferentially matched in series and the self-discharge is preferentially matched in series. The system provides a threshold window, a closed-loop frequency upper limit, a judgment rule, a secondary or segmented secondary fitting model form and a whole process tracing field, and computability, verifiability and auditing of consistent manufacturing and matching are achieved. By adopting the method, the parallel circulation and hot spot risk can be obviously reduced, the series wooden barrel effect is inhibited, and the consistency and safety margin of the high-voltage large-series-parallel aviation battery pack are improved.

Description

[0001] [Technical Field] This invention relates to a method for sorting, grading, and matching cylindrical quasi-solid-state lithium-ion battery cells (e.g., 18650, 21700) before assembly into packs for aviation applications. Specifically, it belongs to the technical field of manufacturing and quality control, encompassing "cell consistency assessment—grading—matching rules—pack verification." Aviation battery packs and automotive power batteries differ fundamentally in their usage boundaries: aviation scenarios are more sensitive to unit energy mass, unit volume power, wide-temperature rate performance, and failure consequences, while also imposing more stringent requirements on "intrinsic safety" and "provable consistency." Consistency is not merely about a single indicator like capacity or voltage; rather, it requires that within the same SOC window, key variables such as the cell's equivalent internal resistance, polarization behavior, usable capacity, static voltage drift, self-discharge rate, thermal response, and aging rate be statistically controllable and traceable. If the consistency of the cells is insufficient, some cells in the same series-parallel structure may reach the voltage / temperature threshold prematurely, triggering protection strategies, resulting in unusable capacity and accelerated lifespan differentiation. In extreme cases, it may cause individual cell thermal runaway and increase the risk of thermal propagation. In aviation applications, the tolerance for such risks is much lower than that for ground transportation vehicles.

[0002] [Background Technology] A key prerequisite for achieving high consistency in the current automotive power battery industry is "uniformity in ultra-large-scale, long-cycle, and strongly coupled manufacturing." Taking a typical high-consistency cylindrical battery cell as an example, its consistency is based first on the uniformity of the materials: the particle size and specific surface area of ​​the positive and negative electrode main materials, the dispersion state of the binder / conductive agent, the solvent system, the additive system, the pore size of the separator and the uniformity of the coating, and the thickness and surface condition of the copper / aluminum foil all need to be stable; secondly, it comes from the uniformity of the process: large-volume continuous homogenization (e.g., thousand-liter stirring and circulating degassing), continuous coating and online thickness closed-loop control, consistent rolling density and springback, consistent slitting burrs and tab size, consistent winding tension and alignment, consistent baking moisture content and drying curve, consistent liquid injection volume and wetting path, consistent standing wetting time, consistent formation / capacity testing current and temperature, and consistent sorting criteria. More importantly, automotive battery cell production lines typically operate on a statistical scale of "millions of cells in the same batch," allowing for multi-level sorting within a huge sample pool to eliminate or redistribute inferior cells, thereby compressing differences in capacity, internal resistance, and voltage to a smaller range. This "ultra-large sample pool + multi-level sorting" model is inherently limited in aviation battery packs: aviation battery packs often use cells in the thousands (e.g., 1800 cells), and due to model iterations, trial production batches, and airworthiness certification schedules, it is difficult to maintain the continuous, homogeneous production of millions of cells like in the automotive industry in the long term. As a result, even if the goal is to match the consistency indicators of automotive battery cells, aviation battery cells often face the challenges of "insufficient sample pool leading to insufficient selection range, inadequate elimination of inferior cells, and insufficient sorting dimensions leading to the introduction of implicit differences into the pack."

[0003] In existing technologies, cell sorting and grouping typically relies on several explicit indicators: open-circuit voltage, DC internal resistance, AC impedance, capacity (single or multiple capacity grading), self-discharge (static voltage drop), and physical dimensions. While numerous domestic and international patents cover these approaches, they generally lean towards the context of large-scale manufacturing for electric vehicles / energy storage, emphasizing "rapid grading using a few measurable indicators within a given liquid injection and formation window." For the "stronger three-dimensional consistency (internal resistance-capacity-self-discharge coupling consistency)" and "approaching large sample pool consistency from small sample pools" required for aerospace quasi-solid-state cells, there is still a lack of systematic solutions that can be directly implemented, examined, and reproduced.

[0004] For example, Chinese patent CN108899596A, "Method for Sorting and Grouping Consistent Power Lithium Battery Cells," proposes to screen and group cells based on several electrical performance parameters. However, its shortcomings are as follows: Firstly, it emphasizes the static rule of "grouping cells according to thresholds and then combining them," which makes it difficult to characterize the "polarization differences and static drift differences" that are more prominent in aviation cells under high-rate, wide-temperature, and quasi-solid-state electrolyte systems. Secondly, it often treats "self-discharge" as a short-term static or single-time criterion, which makes it difficult to solve the capacity and voltage drift differentiation problem of aviation cells under a longer static window (such as a month of high-temperature static storage), and it does not treat the difference in liquid injection volume / wetting as a core controllable variable for consistency.

[0005] For example, Chinese patent CN102760914A, "A Method for Sorting and Grouping Lithium-ion Power Battery Cells", sorts and groups cells based on parameters such as voltage and internal resistance. However, its shortcomings are: it mainly addresses the conventional manufacturing fluctuations of liquid power batteries, has too few sorting dimensions, and lacks an operational path for "how to compensate for insufficient material and process uniformity through subsequent processes". Especially in the case of small sample cells for aviation battery cells, it is difficult to compress the capacity difference to a narrower range by simply relying on voltage / internal resistance thresholds, and it is even more difficult to compress the self-discharge dispersion to a provable level.

[0006] For example, Chinese patent CN112397788A, "Method, Apparatus and Storage Medium for Sorting and Grouping Lithium-ion Batteries," proposes a process-oriented implementation of sorting and grouping. Its shortcomings are: it focuses more on process or system implementation rather than taking the coupling relationship of "internal resistance-capacity-self-discharge" as the core modeling object. For aerospace quasi-solid-state cells, "liquid injection / immersion-in-situ polymerization-formation curve" has a stronger coupling effect on the final consistency. If the sorting logic of conventional liquid cells is still used, it is easy to see the phenomenon of "initial screening indicators are qualified, but long-term static self-discharge differentiation and later grouping differentiation".

[0007] Regarding self-discharge consistency, the recent Chinese patent CN119959810A, "A Screening Method for Self-Discharge Consistency of Lithium-ion Batteries," has included self-discharge as a screening object, which is a beneficial progress. However, its shortcomings are that self-discharge screening is often treated as "end-stage elimination" rather than "three-dimensional grading linked to capacity and internal resistance." If the slight difference in liquid injection volume, wettability, electrode liquid distribution, and the difference in SEI / CEI formation after formation are not incorporated into the same set of controllable processes and criteria, the dispersion of self-discharge will still be amplified in the long-term static and high-rate cycling of aerospace applications.

[0008] In addition, patents such as CN103506328A and CN104218267A, which focus on battery sorting and grouping or consistency control, propose different parameter combinations and process designs. Their common shortcomings are: most of them are based on "single measurement + rule grading", and lack a unified three-dimensional consistency framework for the more critical "multi-stage, multi-timescale retest closed loop" of aerospace quasi-solid-state cells (such as initial OCV / IR grading, retest after formation, retest after replenishment, retest after high temperature standing, and final capacity retest). They also do not provide a patentable path to approximate the consistency of a large sample pool with a small sample pool through "manufacturing side precision improvement + back-end retest closed loop".

[0009] In the field of primary or special batteries, CN112958487B "A screening and grouping method for primary lithium primary batteries" provides a screening and grouping approach, but its target system, failure mechanism and process window are significantly different from those of aerospace quasi-solid-state secondary cells, and it still does not incorporate "the nonlinear coupling between capacity and self-discharge caused by slight differences in liquid injection volume" into the core control.

[0010] Internationally, numerous patents have also been filed related to aerospace and high-reliability battery systems. US Patent 10958083B2, "Battery pack with reduced voltage variance," focuses on reducing voltage variance through battery pack-level strategies. However, its shortcoming lies in its emphasis on "system-level variance mitigation" rather than "minimizing variance during cell manufacturing and sorting." If aerospace battery packs primarily rely on system-level strategies (balancing, current limiting, and protection window contraction) to address consistency issues, it often comes at the cost of sacrificing usable capacity / power and increasing system complexity, failing to fundamentally resolve the thermal and lifespan variations caused by minute differences in cell performance.

[0011] US Patent US20230020002A1, "Cell matching across multiple characteristics," proposes a multi-characteristic matching approach. However, its shortcomings lie in the fact that the matching characteristics are more reflected in the "combination of data dimensions," which may not necessarily fall into the complete link of "process precision improvement + mass conservation weighing closed loop + multi-stage retesting" that can be implemented in aerospace quasi-solid-state cells. If no replicable implementation path is given for key process variables such as liquid injection / replenishment accuracy, wetting and in-situ polymerization uniformity, and formation curve consistency, algorithm matching alone may still be limited by data separability and noise in a small sample pool.

[0012] At the level of cell characterization, US patent US8965721B2, "DC impedance measurement for a cell", emphasizes impedance measurement and characterization. Its shortcomings are: impedance measurement can reflect partial polarization and interface state, but it cannot directly replace the control of variables with longer time scales and stronger coupling with liquid distribution, such as "capacity-self-discharge-fluid filling adequacy". If aerospace quasi-solid-state cells rely solely on impedance or internal resistance grading, self-discharge dispersion and capacity drift may still occur in long-term static and high-temperature environments.

[0013] Related to electrolyte filling, US patent US8568911B2, "Liquid electrolyte storage battery and method for filling an electrolyte into a case," involves electrolyte filling and filling methods. However, its shortcomings are as follows: the electrolyte filling process for automotive power batteries often prioritizes production capacity and yield, and the accuracy of electrolyte filling is often aimed at meeting the needs of large-scale manufacturing. For aerospace quasi-solid-state cells, to approach the consistency of millions of samples in a small sample pool, slight differences in electrolyte volume and droplet loss will be amplified into capacity differences and self-discharge differences. However, most existing electrolyte filling patents do not take the approach of "defining the target of electrolyte volume-capacity deviation with a high-precision weighing closed loop" as the core.

[0014] In summary, while existing technologies cover multiple aspects such as cell sorting and grouping, impedance measurement, system balancing, and electrolyte injection processes, several key gaps remain in addressing the mass production consistency issues of aerospace cylindrical quasi-solid-state cells: First, there is a lack of a grading framework with "internal resistance-capacity-self-discharge three-dimensional coupling consistency" as the core objective; existing technologies primarily rely on two-dimensional or single-dimensional threshold grading. Second, there is a lack of an engineering approach to "approaching the consistency of large sample pools (millions of samples) under small sample pool (thousands of samples) conditions," especially a lack of a process and criteria that integrates "weighing accuracy improvement, closed-loop electrolyte injection / replenishment, consistency of the three-step formation curve, and long-term high-temperature static retesting" into a reproducible, examineable, and patentable process. Third, there is a lack of technical solutions that incorporate the in-situ polymerization uniformity, gas generation suppression, and self-discharge control of aerospace quasi-solid-state flame-retardant electrolytes into the same consistency grading system. Against this backdrop, a "three-dimensional consistency grading method, system, and selection rules for internal resistance-capacity-self-discharge" for aerospace cylindrical quasi-solid-state batteries is proposed. This transforms the "necessary selection of a small number of batteries from a large sample pool" into a "patentable selection algorithm + process + criterion." Supported by feasible high-precision weighing and liquid replenishment closed-loop, staged OCV / IR / capacity retesting, and long-term high-temperature static retesting, it achieves provable consistency close to or even better than that of automotive batteries under limited sample pool conditions. This is a technical problem that urgently needs to be solved in this field and has significant engineering and industrial value. Summary of the Invention: The technical problem this invention aims to solve is: under the constraints of "small quantity per pack, limited batch selection range, and extremely high airworthiness safety threshold" of cylindrical quasi-solid-state batteries for aviation, how can we still achieve an engineerably reproducible three-dimensional consistency classification and selection of "internal resistance-capacity-self-discharge" without the "million-level batch consistency foundation" of automotive power batteries? This would enable high-voltage aviation battery packs composed of 1800 or 2700 cylindrical cells to maintain intrinsic safety under conditions such as high rate, wide temperature range, frequent low-altitude takeoffs and landings, and vibration and shock. Furthermore, this invention elevates the empirical practice of "selecting 2700 cells from 500,000" into a patentable algorithm, process, criterion, and system.

[0015] Existing technologies generally use "voltage, capacity, and internal resistance" as the main parameters for battery pack assembly, but their fundamental starting point is mostly geared towards the large-scale production of automobiles / energy storage, assuming that cells in the same batch come from a sufficiently large base of consistent coating, rolling, slitting, winding, and formation processes. Relevant authorized patents typically employ a sorting combination of "static parameters + dynamic curve characteristics" (e.g., thresholding based solely on static voltage, AC internal resistance, capacity rating, and discharge plateau voltage, or grouping dynamic curves after clustering). However, these methods generally suffer from three shortcomings: First, they fail to establish "effective quality consistency of active materials in the cells" as a quantifiable and controllable prior condition, resulting in subsequent finer sorting still being dominated by the "weakest link" effect of "active material deviation." Second, "self-discharge" is often judged based on coarse criteria such as short-term static voltage decay or single high-temperature static storage, making it difficult to preemptively eliminate the most dangerous "micro-short circuit / latent self-discharge abnormal cells" in aviation scenarios. Third, they fail to treat "liquid / quasi-solid electrolyte introduction amount" as a key controllable variable for capacity consistency, and they have not established quantitative mapping and compensation rules among the four factors of "liquid injection amount—effective capacity—internal resistance—self-discharge," thus failing to reduce capacity differences to 1 / 5 or even more stringent than those in automotive power battery sorting in small-sample aviation cells.

[0016] To address the aforementioned pain points, this invention proposes a three-dimensional consistency grading method, system, and selection rules. Its core is not simply "measuring more parameters," but rather decomposing the formation mechanism of aviation battery cell consistency into three engineerable controllable layers: The first layer is the prior constraint of "consistency between material and electrode effective load," which, through high-precision weighing and areal density closed-loop control of electrode / semi-finished battery cells, compresses the root cause of capacity differences at the source; the second layer is the "consistency of quasi-solid-state flame-retardant electrolyte introduction and in-situ solidification," which, through metering and compensation with an accuracy one order of magnitude higher than that used in automotive electrolyte injection, stabilizes the nonlinear sensitive segment near full capacity; the third layer is a multi-stage verification process of "formation—resting—high-temperature aging—re-resting—re-capacity assessment," which solidifies the aviation-grade safety red line of self-discharge into executable rejection rules through two OCV (open-circuit voltage) gates, one capacity gate, and one internal resistance gate. Ultimately, the battery pack is assembled in series and parallel using a three-dimensional hierarchical "individual fingerprint" to ensure that the entire battery pack maintains temperature consistency, voltage consistency, and aging consistency under maximum current, maximum thermal stress, and maximum vibration stress conditions.

[0017] To achieve the above objectives, the present invention adopts the following technical solution.

[0018] The present invention establishes a three-dimensional consistency grading index system and grading space. It defines the consistency of a single battery cell suitable for aerospace assembly as a point in a three-dimensional coordinate system: the X-axis represents internal resistance consistency, the Y-axis represents capacity consistency, and the Z-axis represents self-discharge consistency. The three-dimensional coordinate system does not use only a single measurement value, but rather stable values ​​after "specified SOC, specified temperature, and specified dormancy time window," and includes a short-pulse dynamic characteristic verification.

[0019] Internal resistance consistency comprises two parts: AC internal resistance and DC equivalent internal resistance. AC internal resistance is used to quickly screen for welding contact defects and electrode structure discrepancies; DC internal resistance reflects the difference in I²R heating under high-rate operating conditions. Capacity consistency is judged by the average and variance of multiple capacity tests to avoid single-test errors. Self-discharge consistency is based on the voltage drift rate during the "fully charged – voltage drop back to specified – high-temperature rest – secondary OCV measurement" process, along with the necessary "equivalent self-discharge current" criterion. This focuses on capturing abnormal self-discharge cells induced by the most critical factors in aviation, such as hidden micro-short circuits, electrode burrs, metallic foreign objects, and abnormal micropores in the separator.

[0020] In terms of engineering specifications, for aerospace 18650 / 21700 cylindrical quasi-solid-state cells, this invention provides directly executable example thresholds (not limited to these): single-cell capacity grading steps can be achieved in the 5–20 mAh range; single-cell DC internal resistance grading steps can be achieved in the 0.05–0.20 mΩ range; single-cell OCV grading steps can be achieved in the 0.1–0.5 mV range; the self-discharge drift rate threshold can be set to ensure that the OCV drift after 30 days of high-temperature static storage does not exceed a certain millivolt threshold, or the equivalent self-discharge current does not exceed the microamp–ten microamp range. These thresholds can be parameterized according to the cell system, capacity level, operating temperature range, and aircraft mission spectrum.

[0021] The process and key innovations of the three-dimensional consistency grading method. The method of this invention includes at least the following steps (which can be tailored according to production line capacity, but key gates should not be deleted).

[0022] Step 1: "Source Selection" and Semi-finished Product Status Definition for Preparing Aerospace-Grade Battery Cells. On automotive power battery cell mass production lines, each batch contains cells that meet automotive standards but still fall short of aerospace standards. To achieve aerospace-grade consistency without altering the automotive production line's cycle time, this invention proposes the concept of "preparing aerospace-grade battery cells": Before liquid injection and sealing, or at a traceable semi-finished product stage before or after sealing, cells within the most stable window of materials, electrodes, and winding conditions from the same batch are selected and marked as preparing aerospace-grade battery cells. Key process parameters such as electrode roll number, coating batch, roll gap, slitting tool lifespan, winding tension window, and drying dew point window are linked to these parameters using a traceable code. The purpose of this approach is to establish "aerospace consistency" based on the "mechanism of forming millions of automotive-grade consistency," rather than performing post-hoc statistical remediation on a small sample.

[0023] Step 2: High-precision weighing gate control for consistent cell effective load (first gate).

[0024] Traditional battery pack assembly methods typically screen for voltage, internal resistance, and capacity at the finished cell level, neglecting to consider the "effective mass of active material" as a strong prior constraint. This invention introduces high-precision weighing before electrolyte filling to measure the mass of semi-finished cells or key components (such as cores and electrode assemblies). The weighing resolution is at least an order of magnitude higher than conventional automotive practices. By dividing the cells into several "mass differential bands," subsequent electrochemical sorting is performed within these bands, suppressing the upper limit of capacity dispersion at its source. The theoretical basis lies in the fact that, under the same chemical system and the same electrode areal density control, the consistency of the effective loading of active material is the primary factor for capacity consistency, and its influence is often stronger than subsequent OCV / internal resistance screening.

[0025] Step 3: High-precision injection, replenishment, and closed-loop compensation of aerospace quasi-solid-state flame-retardant electrolyte (the second gate). This invention further discovers that in the sensitive region near full capacity, the amount of electrolyte introduced has a significant nonlinear impact on achievable capacity, polarization resistance, and self-discharge risk, and this relationship can be fitted in engineering using a quadratic function or a piecewise quadratic function. Taking 18650 as an example, when the injection volume deviation is small, the capacity deviation is not significant; however, when the injection volume approaches the threshold for complete pore wetting, even a small injection deviation can lead to amplified deviations in capacity and internal resistance. The injection precision commonly found in automotive cells (e.g., on the order of 0.1 g) is insufficient for aerospace quasi-solid-state electrolytes. This invention employs high-precision metering and a closed-loop strategy of "weighing—injection—re-weighing—replenishment—re-weighing" to achieve post-injection quality consistency comparable to or even exceeding the "million-level batch consistency" level; and it inversely maps the "target capacity deviation" to the "replenishment quality fine-tuning amount," allowing capacity consistency to approach or even surpass the automotive cell pairing threshold within a small sample pool.

[0026] Step 4: Consistency Control of Quasi-Solid-State In-Situ Curing (The Third Gate). Aerospace flame-retardant quasi-solid-state systems often involve in-situ polymerization / gelation curing processes at around 70°C. This invention emphasizes that the consistency of the curing process is the second major factor contributing to the consistency of self-discharge and internal resistance: insufficient curing leads to ion channel instability and localized side reactions; excessive or uneven curing causes pore blockage and increased polarization. Therefore, this invention sets quantitative windows for the curing temperature field, time window, clamping method, furnace atmosphere, dew point, and cleanliness, and binds these to the cell code, so that any subsequent anomalies can be traced back to differences in the curing curve.

[0027] Step 5: Three-step formation and two OCV gating + one high-temperature self-discharge gating (the fourth gate). This invention employs an improved three-step formation process (e.g., low-rate pre-activation—medium-rate film formation—constant voltage setting, with rate and time configured according to the system) to reduce formation differences; after formation, the first OCV steady-state classification is performed to eliminate abnormal points; then, the cell is fully charged and dropped back to a specified voltage (e.g., a storage voltage of around 3.6V), left to stand for one month under specified high-temperature conditions, and the second OCV is measured, combined with necessary equivalent self-discharge current criteria to eliminate abnormal self-discharge cells. This step directly addresses aviation safety: even if the capacity and internal resistance of a self-discharge abnormal cell are qualified, it may still induce a chain risk of local overcharge / over-discharge and thermal runaway in high-voltage series connection.

[0028] Step 6: Third Capacity Matching and 3D Fingerprint Database Establishment (Fifth Gate). Finally, the cells that have passed the aforementioned gate control undergo a third precise capacity matching, establishing the cell's 3D fingerprint: capacity level, internal resistance level, and self-discharge level. This fingerprint is then written into the database to generate matching rules that can be executed by the system: first, clustering by parallel groups in 3D space to ensure the minimum 3D distance within the same parallel group; then, matching by OCV and internal resistance slope in the series dimension to ensure that the voltage drift and thermal drift of the series link are consistent. This rule transforms "grouping" from an empirical task into a repeatable algorithmic execution.

[0029] III. System Implementation. For engineering implementation, this invention also provides a system corresponding to the above method, including: a high-precision metering unit (weighing and calibration), a micro-liquid injection and replenishment unit, an in-situ curing control unit, a formation and capacity testing unit, an internal resistance and impedance testing unit, a high-temperature static placement and self-discharge determination unit, a data acquisition and three-dimensional grading algorithm unit, and an execution unit that outputs a series-parallel assembly list according to selection rules. The system can be deployed on the aerospace battery cell process island next to the automotive battery cell mass production line to achieve consistent manufacturing of aerospace battery cells through "cutting—precise control—precise sorting—traceability".

[0030] IV. Beneficial Effects and Quantitative Comparison with Existing Authorized Patents. Compared with existing authorized patents, this invention has at least the following beneficial effects.

[0031] First, this invention prioritizes "consistency of effective loading of active material" as the primary gate control, coupling it with subsequent closed-loop compensation during liquid injection. This allows for the compression of capacity consistency into a smaller subset of the stringent thresholds for automotive power batteries, even in aviation battery cells with sample sizes far smaller than one million. While many existing grouping patents mention combining static and dynamic parameters, they typically do not treat "liquid injection quality" as a controllable variable for capacity consistency, nor do they form a closed-loop compensation mechanism. Consequently, capacity differences are primarily eliminated post-processing, making it difficult to achieve both yield and consistency. Taking a typical static + dynamic grouping scheme as an example, its core remains "test-screen-group," while this invention employs "control-compensation-verification-group": first, control and compensation are implemented, then verification is performed using a multi-stage evidence chain, and finally, assembly is performed.

[0032] Secondly, this invention elevates self-discharge from an "optional" to a "mandatory gating item for aviation," and establishes an executable rejection rule through "one month of high-temperature static storage + secondary OCV gating + necessary equivalent self-discharge current criteria." Existing patents contain self-discharge detection and estimation methods, but these often serve battery health assessment or quality monitoring, rather than forming a three-dimensional classification along with capacity and internal resistance to directly drive the output of the battery pack pairing list. This invention couples self-discharge with battery pack pairing, directly reducing the risk of thermal runaway in series high-voltage aviation battery packs under long-term storage and extreme operating conditions.

[0033] Third, this invention is particularly well-suited to the industrial reality of aviation, which requires "small batch sizes and extremely high safety standards": by cutting off the prepared aviation battery cells next to the automotive production line with a capacity of millions, and implementing high-precision weighing, liquid injection, curing, and three-dimensional grading in the aviation process island, even for a set of aviation battery packs with only 1,800 or 2,700 cells, the consistency level can be close to or even exceed that of automotive battery packs, significantly reducing the resource waste and delivery uncertainty caused by "having to select 2,700 cells out of 500,000".

[0034] V. Implementation method (including manpower, machinery, materials, methods, and environment) using a Ka-52 electrified 135 series 20 parallel (maximum 567V), 80Ah, approximately 40kWh aviation battery pack as an example. The following embodiments are for illustrative purposes only and not for limiting the invention. The objective of this embodiment is to form a cell consistency cell that meets the requirements of high-voltage aviation applications and to assemble a 135S20P battery pack accordingly. The highest voltage of this battery pack is approximately 4.2V × 135 = 567V, the nominal voltage is approximately 3.7V × 135 ≈ 500V, and the nominal energy is approximately 500V × 80Ah = 40kWh. The parallel number 20 indicates that each series stage consists of 20 cells connected in parallel, and the number of series stages is 135, so the total number of cells required is 135 × 20 = 2700. If the battery pack is divided into 3 modules, with 900 cells in each module, then each module can be designed as (135S(20P) / 3) or segmented according to structure and connected in series by high voltage busbars. In short, the consistency of the cells must be prioritized to ensure the three-dimensional consistency of each parallel group; otherwise, the differences in I²R heating and self-discharge will be amplified under aviation conditions.

[0035] (I) Personnel Requirements. The recommended personnel configuration for implementing this invention should include at least: 1-2 process engineers with experience in lithium-ion cell manufacturing (coating, rolling, slitting, winding, assembly, drying, electrolyte injection, formation), and an understanding of the coupling relationship between areal density, porosity, compaction density, and capacity / internal resistance; 1 metrology and quality engineer familiar with mass metrology, balance calibration, traceability, uncertainty assessment, and statistical process control; 1-2 electrochemical testing engineers familiar with formation and capacity testing equipment, EIS / pulse internal resistance testing, OCV curves, and self-discharge criteria; 1 data and algorithm engineer capable of outputting three-dimensional hierarchical, clustering, and selection rules, and solidifying these rules into a traceable assembly list; and 1 safety and environmental engineer responsible for fire and explosion prevention, solvent / electrolyte management, cleanliness and dew point control, electrostatic protection, and emergency response.

[0036] (II) Equipment and Instrument Requirements. High-precision weighing and calibration: It is recommended to use an analytical balance with a reading of 0.01 mg or higher for critical gate weighing, along with matching weights and a constant temperature and humidity weighing chamber; for batch turnover, a 0.1 mg-level balance can be provided separately.

[0037] Micro-volume injection and replenishment: Employ a step-controlled micro-metering pump or a time-pressure precision dispensing / injection platform with closed-loop weighing compensation capability; the injection needle and flow channel must be compatible with the viscosity range of flame-retardant quasi-solid electrolyte.

[0038] Vacuum drying and dew point control: vacuum oven or vacuum drying room, dew point monitoring and cleanliness control; used for degassing and wetting control before and after liquid injection.

[0039] In-situ curing and temperature control: Programmable temperature-controlled oven or circulating air oven with temperature uniformity verification capability; used for in-situ curing at around 70℃ and subsequent static setting.

[0040] Formation and capacity separation equipment: multi-channel formation cabinet and capacity separation cabinet, with high-precision metering capability for low current; can record characteristics such as constant current and constant voltage inflection point, plateau capacity, and coulombic efficiency.

[0041] Internal resistance / impedance testing: Equipment that supports AC internal resistance (e.g., 1kHz) and DC pulse internal resistance testing, and has impedance spectrum testing capability when necessary.

[0042] High-temperature static setting and self-discharge determination: constant temperature chamber, temperature stability meets long-term static setting requirements; equipped with OCV acquisition and equivalent self-discharge current testing device (optional).

[0043] Data collection and traceability: Barcode / RFID and database systems ensure that each battery cell is traceable throughout the entire process from cutting to assembly.

[0044] (III) Material System and Key Parameters. This embodiment uses a high-nickel ternary cathode and silicon-based anode system as an example. It should be emphasized that the specific material formulation can be implemented according to the applicant's existing system. The focus of protection of this invention is on the consistency grading method, system and selection rules, rather than being limited to a single chemical formulation.

[0045] Cathode: 9-series high-nickel layered oxide system (can be NCM or NCA path), requiring controlled batch D50, specific surface area and moisture content; the bonding system and conductive system are designed and configured for high rate and safety.

[0046] Anode: Graphite-silicon composite anode system, with silicon content and pre-lithiation / film formation strategy configured according to cycle life and first-efficiency requirements; key control is given to the uniformity of electrode surface density and the compaction density window.

[0047] Diaphragm: High heat-resistant microporous diaphragm or coated diaphragm system, with a focus on controlling porosity consistency and thermal shrinkage index.

[0048] Current collectors: The thickness and surface roughness of aluminum foil and copper foil are controlled; for high-magnification aerospace applications, it is recommended to strictly control burrs and cleanliness.

[0049] Electrolyte: Flame-retardant quasi-solid electrolyte system (can be gel / in-situ polymerization type), the key lies in the accuracy of liquid injection metering, the consistency of degassing and wetting, the consistency of curing temperature field and the suppression of side reaction gas generation.

[0050] Structural components and packaging: The cylindrical steel shell and cap system are used to withstand low pressure differentials and external pressure changes; the present invention ensures the basis for subsequent welding and thermal consistency of the battery pack based on the cell consistency method.

[0051] (iv) Process steps and key points of quantitative control (method). Selecting preliminary aviation battery cells: Select battery cells with stable process windows next to the automotive mass production line, requiring the same coating roll number range, the same roll gap window, and the same winding tension window.

[0052] Pre-filling weighing gate: Weigh the semi-finished cells and divide them into quality sub-classes. For example, divide the quality distribution into several narrow intervals, control the quality difference in each interval to the milligram level or even more strictly, and remove abnormal points at the end.

[0053] High-precision liquid injection and replenishment closed-loop: First, inject the baseline volume, then weigh it again, and replenish the liquid according to the difference between the "target mass and the measured mass"; after replenishment, weigh it again until it falls within the threshold. This process simultaneously records the total injected mass and uses the injected mass as the input variable for the subsequent capacity deviation compensation model.

[0054] Degassing and in-situ curing: After liquid injection, degassing and in-situ curing are carried out under vacuum conditions, followed by in-situ curing according to the curing curve. The curing temperature, time and atmosphere must fall within the window; the curing curve is bound to the cell code.

[0055] Three-step formation: low-rate pre-activation, medium-rate film formation, and constant voltage tuning are adopted in three stages; the voltage, current, and time window of each stage are set according to the system, and the inflection point and coulombic efficiency are recorded.

[0056] First OCV gating and internal resistance gating: After formation, allow the system to stand until the OCV reaches steady state, measure the OCV and AC / DC internal resistance, and eliminate abnormal points.

[0057] Storage voltage drop and high-temperature resting: Drop the battery cell from full charge to storage voltage (3.6V in example), place it in a constant temperature chamber and rest at high temperature for 30 days; take it out and measure the OCV for the second time, calculate the drift rate or equivalent self-discharge index, and discard the battery cells with abnormal self-discharge.

[0058] The third stage of precise capacity classification and three-dimensional grading and warehousing: precise capacity classification of gated cells is performed to form three-dimensional grades; a three-dimensional label is output for each cell.

[0059] Selective configuration rule output: Prioritize parallel groups, select the 20 units with the smallest distance in three-dimensional space to form a parallel group, forming 135 parallel groups; then match the OCV and internal resistance slope according to the series dimension to generate the assembly list of 135S20P.

[0060] Post-assembly verification: After assembly, low-rate equalization charge and discharge and one high-rate pulse verification are performed to confirm that the differential pressure, parallel circulating current and temperature rise of each string are consistent and fall within the aviation threshold.

[0061] (V) Environmental and Safety Controls (Environment). Dew Point and Cleanliness: The dew point and particle cleanliness of the filling and packaging areas must be controlled to reduce the risk of self-discharge caused by metal foreign objects and moisture; Static Electricity and Fire Prevention: Static electricity protection, ventilation, and fire protection configurations in areas where solvents and electrolytes are used shall be implemented in accordance with the specifications; Metrological Environment: The weighing area shall be kept at constant temperature and humidity to avoid errors caused by airflow and vibration in milligram-level weighing; Data Traceability: Each cell shall be traceable at least to the electrode roll number, filling quality, curing curve, formation curve, two OCV measurements, and three capacity records.

[0062] (VI) Implementation Results. Through the above process, under the realistic aviation conditions where "the number of selectable cells is far less than one million", this invention transforms consistency from "relying on mass rejection after the fact" to "relying on quality gating + liquid injection closed-loop compensation + curing consistency + self-discharge gating before the fact" to ensure consistency. This enables the 2,700 cells ultimately used in the 135S20P high-voltage aviation battery pack to converge simultaneously on four main lines: capacity difference, internal resistance difference, OCV difference, and self-discharge difference. This significantly reduces the risks of parallel circulating current, series drift, hot spot concentration, and abnormal self-discharge chain amplification, providing a foundation for subsequent liquid-free cooling and heat management and high-reliability welding. [Attached Image Description] Figure 1 This is a schematic diagram of the consistency control system structure for the manufacturing and selection of aerospace cylindrical quasi-solid-state battery cells in an embodiment of the present invention. Brief explanation of the reference numerals in the attached drawings: 1-Instruction manual drawing (assembly); 2-Pressure gauge; 3-Precision injection pump / needle; 4-Fixed crossbeam; 5-Clamping assembly; 6-Test cell / liquid tank; 7-Support frame; 8-Pipeline / sensor connection point; 9-Workbench; 10-Vacuum / protective cover; 13-High-precision electronic balance / weighing unit; 14-Vacuum pump assembly; 15-Conveying track; 17-Support feet / adjustment mechanism; 18-Control host / drive unit. This device includes a frame support structure, a sealed injection chamber, an injection needle and its positioning and sealing assembly, a vacuum and pressurization gas circuit system, a valve group and control module, and a human-machine interface and data recording unit. It is used to complete pre-vacuuming, segmented injection, final precision replenishment, pressure stabilization and wetting, and safety pressure relief processes within the same fixture, and provides traceable process data support for subsequent in-situ curing and consistent grouping. Further explanation of the reference numerals in the attached diagram (Figure:): 2—Pressure gauge or pressure sensor assembly, used to monitor the pressure of the injection chamber or pressure stabilizing chamber in real time, serving as a criterion for segmented pressure control and safety interlocking. 3—Multi-station injection head or injection needle assembly, used to seal and connect with the injection ports of each cell to achieve parallel injection, supporting staged operations of "rapid filling + final precision replenishment". 4—Injection needle positioning plate or distribution plate, used to limit the relative position, coaxiality, and insertion depth consistency of each injection needle, reducing sealing failure and droplet splashing caused by eccentricity. 5—Sealing and clamping assembly or needle seat seal, used to maintain a reliable seal between the injection needle and the cell injection port under vacuum and pressurization conditions, preventing leakage, air intake, and backflow. 6—Cell clamp and sealed injection chamber (transparent cavity), used to accommodate and position the vertically placed cylindrical cell array, forming a sealed space that can be vacuumed and pressurized, and facilitating process observation. 7—Gas or liquid connection lines, used to connect the vacuum source, pressure source, valve group, and injection tank, enabling switching between vacuuming, segmented pressurization, pressure stabilization, and pressure relief circuits. 8—Frame support plate and column frame, providing structural support and assembly reference for the device, ensuring overall rigidity and repeatability of the multi-station injection process. 9—Injection tank base or cell tray assembly, providing bottom positioning and support for the cell array and forming a sealed assembly reference with the injection tank. 10—Cylindrical cells (aircraft-grade cells to be injected), used for segmented injection, replenishment weighing, and pressure stabilization wetting in this device. 13—Instrument mounting platform or intermediate support frame, used to mount process control, display, recording, and weighing / traceability-related components, and enhance overall support stability. 14—Vacuum pump or gas pump (vacuum and pressure source), providing power for pre-vacuuming, differential pressure-driven injection, and pressure stabilization wetting. 15—Control and display terminal (human-machine interface), used to set stage parameters, display pressure / status, and record process data such as batch, workstation, and stage switching. 17—Leveling and vibration-damping feet or grounding feet, used for leveling and vibration isolation of the device, reducing the impact of external vibrations on sealing stability and the consistency of final precision liquid replenishment. 18—Gas circuit valve group and electrical control module (valve island / control box assembly), used for valve control and interlocking of actions such as vacuuming, pressurizing, stabilizing, depressurizing, and switching, and responsible for signal acquisition and safety protection.

[0063] This invention relates to the manufacturing and selection consistency control of cylindrical quasi-solid-state lithium battery cells for aviation, and particularly to a grading method, system and selection rules based on internal resistance, capacity and self-discharge as three-dimensional core indicators and a quality measurement closed loop as the main line. This enables aviation battery packs to achieve matching consistency close to or even better than that of large-scale automotive-grade battery cells under the same batch production conditions of millions of cells, even in engineering scenarios where only thousands of cells are needed per batch. This significantly reduces the risks of circulating current, hot spots and discrete drift after being connected in series and parallel, and improves the intrinsic safety and repeatable airworthiness verification capability of aviation battery packs. The technical concept of this invention is based on the following verifiable physical and electrochemical facts: For high-energy-density cylindrical cells with a high-nickel cathode and silicon-based anode system, the final achievable capacity is not only affected by the quality and pore structure of the electrode active material, but also by the effective wetting of the electrolyte, the formation of ion migration channels, and the byproducts of interfacial reactions. Among these factors, the "effective amount of electrolyte added and its spatial distribution" exhibits a significant nonlinear relationship with capacity release and internal resistance formation. Especially in the threshold range close to the point where the pores are fully wetted and the interfacial reaction tends to saturate, small deviations in the electrolyte are amplified into discrete differences in capacity and self-discharge. In the low addition range deviating from the threshold, the capacity is less sensitive to electrolyte deviations, exhibiting piecewise nonlinear characteristics. Based on this, this invention transforms the "liquid injection-liquid replenishment-curing" process from traditional volume or time control to a closed-loop control centered on mass measurement. It then uses a quadratic function or a piecewise quadratic function to fit the mapping relationship between "mass deviation and capacity deviation," thereby forming traceable replenishment criteria and selection rules.

[0064] Specifically, this invention proposes a consistent manufacturing and selection process for aerospace cylindrical quasi-solid-state battery cells, comprising six stages: pre-screening, precision electrolyte injection and closed-loop replenishment, in-situ curing, three-stage gating and grading, and final selection rules. In the pre-screening stage, a quality benchmark is established for empty battery cells from the same batch (characterized by being formed under the same raw materials, coating, and rolling conditions as those used in automotive-grade mass production lines, but with aerospace-specific process windows reserved in the sealing and electrolyte systems). The resolution and uncertainty of the weighing system used should be less than one-tenth of the target consistency budget. The weighing environment must meet the conditions of constant temperature and humidity, low airflow disturbance, vibration resistance, and electrostatic suppression. Weighing data is bound and archived with the unique identifier of the battery cell. By narrow-window grading of the empty battery cell quality distribution, cells with similar effective active material loadings are grouped into the same candidate group, reducing capacity dispersion caused by slight differences in electrode surface density and winding tension from the source, providing convergent initial conditions for subsequent electrolyte injection consistency.

[0065] In the precision injection and closed-loop replenishment stages, this invention does not use "a fixed volume injected per syringe" as the control target, but rather "the mass landing point after injection," and employs a multi-stage dosing strategy from coarse to fine to balance efficiency and accuracy. The first stage is a rapid dosing stage, where the dosing step size is set according to the proportion of the target injection mass, preferably a percentage of the target injection mass, to quickly approach the target range. The second stage is a fine dosing stage, where the dosing step size converges to a thousandth-level or smaller proportion of the target injection mass, to compress the mass deviation within the first threshold window. The third stage is a micro-compensation stage, where the dosing step size further converges to a ten-thousandth-level proportion of the target injection mass, and a fixed waiting time is executed after each micro-dosing to eliminate transient errors caused by droplet fall and needle residue, ultimately ensuring that the mass deviation after injection meets the second threshold window. The aforementioned "threshold window" is not arbitrarily set, but is derived from the consistency of target capacity, the consistency of target internal resistance, and the consistency of target self-discharge. The second threshold window should be smaller than the first threshold window, and the combined error of "mass measurement uncertainty + needle residual deviation + evaporation / splash loss fluctuation" should still be less than one-third of the threshold window, thus ensuring stable closed-loop convergence. To achieve closed-loop operation, this invention performs a re-weighing after each addition, calculates the mass deviation Δm, and maps Δm to the expected capacity deviation ΔQ_est and the expected self-discharge deviation ΔSD_est through a fitting model. The fitting model is preferably a quadratic function ΔQ_est = a·(Δm)^2 + b·Δm + c, or a piecewise quadratic function in different addition intervals, where the segment boundaries correspond to the "sensitive interval near the effective wetting threshold of the electrolyte." When Δm falls within the first threshold but not the second threshold, the system automatically enters the micro-compensation stage; when Δm exceeds the first threshold, the system determines it to be an abnormal addition or weighing, triggers anomaly interception, and records the cause code; when Δm falls within the second threshold, the system determines the injection quality is qualified and enters the degassing and wetting step. Through the above-mentioned "quality closed loop + nonlinear mapping + threshold criterion", this invention transforms the capacity consistency that originally relied on screening millions of batches into a controllable consistency convergence process that can be achieved within a smaller candidate set.

[0066] In the degassing and in-situ curing stages, this invention proposes a two-component in-situ gelation / curing consistency control strategy for quasi-solid systems. The two-component system comprises a first component A and a second component B, where A is a monomer or oligomer system capable of in-situ polymerization, and B is a crosslinking agent, initiator, or synergistic curing component. After mixing, the two components undergo in-situ polymerization / crosslinking within a controlled temperature window to form a gel or quasi-solid network. To improve pressure resistance and suppress dendrite risk, this invention incorporates solid ceramic electrolyte particles into the electrolyte system, preferably garnet-type lithium lanthanum zirconium oxide (LLZO) or its doped variants, and can further employ surface modification to improve its dispersion and interfacial compatibility within the polymer network. The preferred particle size distribution of the LLZO particles is submicron to micrometer, with a D50 ranging from hundreds of nanometers to several hundred nanometers. The amount added aims to enhance ion channels and improve withstand voltage, and is within a certain weight percentage range based on the mass fraction of the quasi-solid phase after curing. To avoid particle agglomeration causing local impedance dispersion, this invention requires the mixing process to meet specified shear power density and mixing time windows, and uses the rheological properties and particle dispersion uniformity of the mixed sample as release criteria. The in-situ curing temperature and time are controlled using a window-based method, preferably completing curing in the medium temperature range to ensure sufficient electrolyte network formation and controlled side reactions and gas generation. The curing process uses multi-point temperature monitoring and records the temperature integral at the location of each cell, thereby compressing the "curing degree dispersion" within the target window. To ensure consistent wetting and degassing before curing, this invention employs a programmable vacuum-backpressure cycle for degassing and wetting. An upper limit is set for the number of cycles to avoid over-processing and side reactions, while a lower limit is set to ensure sufficient removal of microbubbles. After degassing and wetting, the cells are placed in a consistent settling window before curing to allow the liquid phase to redistribute and reach a quasi-steady state, thereby reducing interfacial impedance dispersion after curing. Each of the above steps is linked to a unique cell identifier to form traceability fields, including at least: empty cell quality baseline, target liquid injection quality, Δm sequence for each addition, fitted model version number, threshold window version number, degassing and wetting program number, curing temperature integral, initial OCV after curing, internal resistance after curing, and subsequent gating results, forming an auditable chain of evidence for consistency throughout the entire process.

[0067] In the three-stage gating and selection rule stage, this invention upgrades the consistency screening of aviation battery cells from a single indicator to a three-dimensional consistency grading of "internal resistance-capacity-self-discharge". The first stage gating is post-formation steady-state open-circuit voltage and internal resistance gating. After the battery cell is placed in a specified resting window, the OCV is collected and the AC internal resistance or DC pulse equivalent internal resistance is tested to form the first round of grading. The second stage gating is high-temperature resting self-discharge gating. After the battery cell is charged to a specified SOC and then falls back to a specified voltage platform, it is placed in a high-temperature resting window for a specified duration. The OCV drift and self-discharge equivalent indicators are measured and the second round of grading is carried out to eliminate potential micro-short circuit and high-side reaction cells. The third stage gating is precise capacity grading gating. In a high-precision capacity grading system, a high-resolution capacity value and coulombic efficiency are obtained with a consistent charge and discharge regime to form the final grading label. The core of the three-dimensional classification is to compress the allowable discreteness of each indicator to a window that is more stringent than automotive-grade, and to provide "parallel priority matching rules" and "series drift suppression rules": in the parallel dimension, priority is given to matching internal resistance to ensure uniform current distribution in parallel branches to suppress I²R hot spots; in the series dimension, priority is given to matching capacity and self-discharge to ensure consistent cascade voltage platforms and avoid the barrel effect from prematurely triggering undervoltage or overvoltage boundaries; for cell packs intended for high-rate operation, additional tags for rate internal resistance slope or pulse temperature rise are added to match transient high-power conditions such as takeoff and landing and go-around. Through the above three-dimensional classification and selection rules, this invention solidifies the engineering experience of "selecting a small number of cells for grouping from a very large number of candidate cells" into a patentable algorithmic process, criterion system, and traceability field system, transforming the consistency of aviation battery packs from a "probabilistic event" into a "calculable, verifiable, and traceable engineering result".

[0068] Taking the aviation battery pack used in the Ka-52 two-seat electric helicopter as an example, this battery pack adopts a 135 series 20 parallel structure, with a nominal voltage of approximately 500 volts, a capacity of several ampere-hours, and an energy level of tens of kilowatt-hours. A single pack requires thousands of cells. This type of high-voltage, large-parallel battery pack is extremely sensitive to consistency. If the parallel internal resistance is too large, it will form obvious parallel circulating currents and hot spots under high-current conditions such as takeoff, landing, and go-around, which will lead to local temperature rise, accelerated aging, and induce abnormal self-discharge. If the series capacity and self-discharge are too large, some series will reach the undervoltage or overvoltage boundary prematurely under multi-cycle conditions, weakening the safety margin of the entire pack. This invention achieves high consistency in battery pack grouping even with a significantly smaller candidate range than the million-unit batch size for automotive applications. This is achieved through pre-grouping empty cells within a narrow quality window, closed-loop liquid injection and replenishment, in-situ curing consistency control of two components with LLZO blending, and a three-stage three-dimensional gating system encompassing formation, high-temperature settling, and precise capacity assessment. This results in more uniform current distribution in the parallel branches of the final grouped cells, more consistent cascade drift, and more controllable overall thermal management. Furthermore, it provides a foundation for electrochemical consistency in thermal uniformity under subsequent liquid-cooling-free structures. This invention emphasizes that the above consistency does not rely on a single absolute liquid injection value or a single curing temperature point, but rather on a systematic approach of "closed-loop quality measurement + nonlinear mapping fitting + threshold window + gating system + traceable evidence chain." Therefore, even with different cell specifications, different energy level battery packs, or different aerospace operating conditions, the same consistency target can be achieved by updating the threshold window and fitting model parameters, demonstrating significant engineering applicability and scalability.

[0069] This invention comprises three parts at the system level: a hardware system, a software system, and methodological rules for the consistent manufacturing and selection of aerospace cylindrical quasi-solid-state battery cells. The hardware system includes at least one high-precision weighing unit, at least one micro-volume injection and replenishment closed-loop unit, at least one vacuum degassing and impregnation unit, at least one in-situ curing temperature control unit, at least one electrical performance acquisition and grading unit, and a traceability database interconnected with the data from the aforementioned units. The software system includes a control module for quality closed-loop replenishment, a fitting module for nonlinear mapping of quality deviation and capacity deviation, a version control module for threshold window management, an algorithm module for generating three-dimensional grading and selection rules, and a recording module for solidifying traceability fields and auditing outputs throughout the entire process. The methodological rules include pre-screening rules, injection and replenishment closed-loop rules, in-situ curing consistency rules, three-stage gating grading rules, and selection rules for series-parallel grouping. Through the synergy of the aforementioned systems and rules, this invention can still obtain repeatable, verifiable, and traceable high-consistency aviation battery cell groups, even when the demand for aviation battery pack cells is far less than that of automotive-grade million-level batch production. This provides a foundation for the thermal consistency, circulating current suppression, and airworthiness verification of high-voltage, large-series-parallel battery packs.

[0070] One of the key innovations of this invention is the transformation of liquid injection consistency control from traditional volumetric metering control to closed-loop control centered on mass metering, and the introduction of a nonlinear mapping fitting model and threshold window criterion. The threshold window includes at least two levels: a first threshold determines whether to enter the fine compensation process, and a second threshold determines whether the aerospace-grade consistency target has been achieved, allowing entry into the degassing, wetting, and solidification processes. The size of the threshold window is jointly determined by the target capacity consistency target, the target internal resistance consistency target, the self-discharge consistency target, and the combined uncertainty budget of the weighing and liquid injection systems. The second threshold is strictly smaller than the first threshold, and the second threshold should ensure that the combined error of the weighing system uncertainty, the residual error of the liquid injection system, and the mass fluctuation caused by environmental evaporation and micro-splashing are still significantly smaller than the second threshold, thus guaranteeing the reliability and auditability of closed-loop convergence. This invention further specifies the upper limit of the number of closed-loop cycles and the anomaly interception mechanism. When a battery cell reaches the upper limit of the number of cycles during the closed-loop liquid replenishment process but still cannot converge to the second threshold, or when phenomena such as poor repeatability, drift anomaly, or electrostatic interference anomaly occur, the system automatically determines that the battery cell enters the isolation queue and generates an anomaly cause code. At the same time, the status of the same batch of equipment and the environmental status are locked and audited to prevent the anomaly from spreading.

[0071] The second key innovation of this invention lies in proposing a nonlinear mapping fitting model between electrolyte injection quality deviation and capacity deviation, and its application. For high-energy-density cylindrical cells with high-nickel cathodes and silicon-based anodes, when the effective electrolyte dosage is in the low dosage range, the capacity is less sensitive to dosage deviation. However, when the effective electrolyte dosage approaches the threshold range where the pores are sufficiently wetted and the interface reaction tends to saturate, the capacity's sensitivity to dosage deviation significantly increases and exhibits nonlinear amplification characteristics. Based on this, this invention preferably uses a quadratic function or a piecewise quadratic function as the fitting model, with the quality deviation Δm as the independent variable and the capacity deviation estimate ΔQ_est as the dependent variable, to establish a fitting relationship and use it for electrolyte replenishment decisions. The piecewise boundaries are obtained through experimental calibration and are related to the specific cell system, electrode pore structure, and degree of solidification network formation. This invention can also introduce a self-discharge sensitivity term into the fitting model, simultaneously mapping Δm to the self-discharge deviation estimate ΔSD_est, for use as a priori label for subsequent three-dimensional grading. The fitted model has version number and applicable scope fields in the system, and is bound to the raw material batch, process window version, and equipment status version, thereby enabling the model to be transferred and auditable between different batches.

[0072] The third key innovation of this invention lies in proposing an in-situ curing consistency control strategy for an A / B two-component flame-retardant quasi-solid electrolyte, and introducing an LLZO-type solid ceramic electrolyte to enhance pressure resistance and suppress dendrite formation. Simultaneously, the "mixing-dispersion-curing temperature integral" is incorporated into a traceability field. Component A contains monomers or oligomers capable of in-situ polymerization, along with necessary solvents and flame-retardant synergistic components. Component B contains a crosslinking agent, an initiation system, or a synergistic curing component. After mixing A and B within a controlled ratio range, a specified operable time window is achieved, forming a continuous gel or quasi-solid network within the intermediate-temperature curing window. To ensure aerospace consistency, this invention specifies the rheological index window and reactivity window after A / B mixing, and uses the intra-batch dispersion of rheological index and gelation time as one of the release criteria. The LLZO-type ceramic electrolyte is preferably garnet-type lithium lanthanum zirconium oxide or its doped variants, with a preferred particle size distribution in the submicron to micron range. The amount added is within a certain weight percentage range, aiming to improve withstand voltage and controllable interfacial impedance. To avoid particle agglomeration leading to increased local impedance and abnormal self-discharge, this invention requires a dispersion process that meets a specified shear power density and specifies an upper limit for particle agglomeration after dispersion. Surface modification is introduced when necessary to improve compatibility within the polymer network. The curing process employs temperature and time window control, using the curing temperature integral as the core traceability field to ensure consistent curing degree for each cell, thereby reducing rate performance and gas generation dispersion caused by insufficient or excessive curing.

[0073] The fourth key innovation of this invention lies in proposing a three-stage gating system: "post-formation steady-state OCV and internal resistance gating—high-temperature static self-discharge gating—precise capacity grading gating," and applying the grading results to the selection rules for high-voltage, large-series-parallel aviation battery packs. The first stage of gating involves collecting open-circuit voltage and testing equivalent internal resistance after a specified static window, eliminating cells with abnormal internal resistance or OCV and forming the first round of grading. The second stage of gating places the cells in a specified high-temperature window for a specified duration, measuring OCV drift, self-discharge equivalent indicators, and necessary impedance changes, eliminating cells with potential micro-short circuits or excessive side reactions and forming the second round of grading. The third stage of gating involves charging and discharging under a consistent regime in a high-resolution precise capacity grading system to obtain high-resolution capacity values ​​and coulombic efficiency, forming the third round of grading. All three stages of gating require recording the test equipment status, calibration status, and environmental window as traceability fields, ensuring the three-dimensional grading is auditable and reproducible. The final selection rules are based on the principle of prioritizing internal resistance in parallel connections and prioritizing capacity and self-discharge in series connections. In addition, rate tags can be added according to the characteristics of aviation power conditions. For example, pulse internal resistance slope or pulse temperature rise tags can be introduced for transient high-power conditions such as take-off, landing and go-around, so that the heat generation of the battery pack under high current is more consistent, thereby reducing the probability of parallel circulating current and local hot spots.

[0074] Regarding the specific implementation methods,Taking an aviation battery pack used in a two-seat electric helicopter as an example, this battery pack adopts a high-voltage series-parallel structure, with over a hundred series connections and over ten parallel connections, a total energy level of tens of kilowatt-hours, and thousands of cells. This type of battery pack is significantly more sensitive to consistency than automotive-grade battery packs. This is because aviation operating conditions include high-rate transients, repeated takeoff and landing cycles, and go-arounds, and require higher safety redundancy. If the cell capacity is too variable, the cascade bottleneck effect will prematurely trigger undervoltage or overvoltage boundaries, weakening the overall safety margin of the pack. If the cell internal resistance is too variable, uneven current distribution in parallel branches will create I²R heating differences, which will be amplified into hot spots, further accelerating aging and self-discharge anomalies. Therefore, this embodiment first reserves aviation cells under the same batch of materials, coating, and rolling conditions as automotive-grade mass production lines. These aviation cells have a process window for performing aviation-specific flame-retardant quasi-solid liquid injection and in-situ curing before sealing. Subsequently, a baseline distribution of empty battery cell quality was established for the reserved aviation battery cells. Cells with similar effective active material loading were grouped into the same candidate group using a narrow window grading system. The narrow window width was derived from the target capacity consistency and should be significantly smaller than the quality grading width used in automotive-grade screening to ensure convergence of subsequent electrolyte injection consistency. Next, a multi-stage strategy of coarse addition, fine addition, and micro-compensation was implemented in the precision electrolyte injection and replenishment closed-loop system. In the coarse addition stage, a large step size was used to rapidly approach the target quality range. In the fine addition stage, the quality deviation was compressed to within the first threshold. In the micro-compensation stage, multiple small-step compensations and re-weighings were performed to bring the quality deviation to within the second threshold. Each addition, each re-weighing, and the quality deviation Δm sequence were bound to a unique cell identifier and written into the database. The replenishment decision is not based on fixed empirical values, but rather on a fitted model that calculates ΔQ_est based on Δm and provides replenishment suggestions, making the replenishment process calculable and auditable. The fitted model uses a quadratic function or piecewise quadratic function, and the model parameters and piecewise boundaries are determined using calibration data from the same system. The model version number and applicable scope are bound to the material and process window of this batch. After completing the liquid injection closed loop, the cell enters the vacuum degassing and impregnation program. The program number, number of vacuum cycles, vacuum window, and back pressure window are written as traceability fields. After degassing and impregnation, in-situ curing of the A / B two-component flame-retardant quasi-solid electrolyte is performed. A and B are mixed in a specified ratio range and injected and distributed within a specified rheological window. The curing temperature window and time window are controlled, and the curing temperature integral and multi-point temperature uniformity verification results are written as traceability fields. LLZO-type ceramic electrolyte particles are mixed into the electrolyte system to improve pressure resistance and suppress dendrites. The particle size distribution and addition amount are within a specified range, and a uniform distribution is formed under specified dispersion conditions. The shear conditions, dispersion time window, and agglomeration index of the dispersion process are written as traceability fields.After curing, the cells enter the first stage of gating, where steady-state OCV and internal resistance are collected to form the first round of grading. Subsequently, they enter high-temperature static gating, where they are statically placed within a specified window for a specified duration, and OCV drift and self-discharge equivalence indicators are measured to form the second round of grading. Finally, they enter precise capacity gating to obtain high-resolution capacity, forming the third round of grading. The equipment calibration status, environmental window, and test system version number for all three stages of gating are recorded in the traceability field. Ultimately, three-dimensional grading labels are used for grouping: parallel groups prioritize internal resistance matching, and series groups prioritize capacity and self-discharge matching, combined with rate tags for optimized matching under aviation transient conditions. Through the above implementation method, the requirement to "select a small number of cells for grouping from a very large number of candidate cells" can be transformed into a patentable algorithmic process, criterion system, and traceable evidence chain system. This allows for aviation-grade consistent grouping even within a candidate set of thousands to tens of thousands, significantly improving the safety margin and repeatability of airworthiness verification for high-voltage, large-series-parallel battery packs.

[0075] The beneficial effects of this invention are reflected in the following aspects: First, this invention improves the consistency of liquid injection from volume measurement to mass measurement closed-loop convergence through closed-loop liquid injection and replenishment. Furthermore, it maps mass deviation to capacity and self-discharge deviation estimates using a quadratic or piecewise quadratic fitting model, enabling calculable and auditable replenishment decisions. This allows the cell capacity consistency to meet aerospace requirements within a relatively small candidate set. Second, this invention introduces A / B dual-component in-situ curing and LLZO blending into the consistency control framework, solidifying the mixing rheological window, dispersion and agglomeration index, and curing temperature integral as traceability fields, significantly reducing the common problems in quasi-solid-state systems. The invention addresses several key challenges: first, it eliminates the risks of solidification dispersion, interface impedance dispersion, and gas generation dispersion; second, it proposes a three-dimensional classification system with three-stage gating, which simultaneously controls internal resistance, capacity, and self-discharge, avoiding the "surface consistency but deep inconsistency" caused by screening based on a single indicator; and third, it directly transforms the classification results into the safety benefits of series and parallel connection through selection rules, reducing the risks of parallel circulating current and hot spots and suppressing the series bottleneck effect; fourth, it solidifies the above process into a traceable field system, providing a complete evidence chain for the consistency of aviation battery cells, facilitating airworthiness verification, quality auditing, failure tracing, and continuous improvement, and has significant engineering promotion value.

[0076] The core problem this project aims to solve is not simply to "press the electrolyte higher" or "press it for longer," but rather to upgrade electrolyte injection from traditional "single-parameter control" to "step-by-step quality target control" under the small-batch conditions of aerospace cylindrical quasi-solid-state cells. In other words, pressure is only one of the execution methods. The real control objective of this project is to ensure that each cell reaches the preset electrolyte injection quality threshold at every stage, and to make the consistency of this threshold auditable, traceable, and closed-loop across the entire batch. This will compress the random fluctuations in subsequent formation, settling, self-discharge screening, and capacity grading into the allowable window for aerospace battery packs.

[0077] Therefore, this project divides the entire injection process into at least three consecutive stages. Each stage is defined by a "stage target injection volume," a "stage allowable error window," a "stage duration window," and a "stage transition criterion." The stage target injection volume is the primary control object, the stage duration window is the secondary control object, and the pressure difference is merely one of the execution parameters to achieve the target injection volume within a given time window. In other words, this project no longer considers "setting the pressure to a fixed value and injecting until it stops flowing" as acceptable. Instead, it considers "within a specified time window, according to a specified segmented target injection volume curve, advancing the electrolyte mass increment to the specified endpoint, and converging the error to the final window through a weighing-based replenishment closed loop" as acceptable. This clearly expresses in the project documentation that we are pursuing consistent and segmented controllable injection quality, rather than a specific pressure figure.

[0078] In the first stage of electrolyte injection, the goal of this project is to rapidly establish the main permeation pathway within the cell, allowing the electrolyte to enter the main volume fraction region of the core's pore network within a short time. The focus of this stage is not on achieving final precision, but rather on achieving a rapid and non-uniform electrolyte injection efficiency. Therefore, the control logic for the first stage is primarily based on the target injection volume, constrained by the stage duration, while allowing for relatively higher driving pressure differentials or higher injection flow rates to meet the time window requirements. As long as the mass increment is confirmed to reach the target window at the end of the first stage, and no abnormal abrupt changes occur in the pressure and flow rate curves during the injection process (e.g., pulsations caused by suspected air leakage, splashing, or needle valve blockage), the first stage is considered successful, and the process proceeds to the second stage. The direct significance of this approach is to concentrate and compress the main time losses in traditional isobaric electrolyte injection, characterized by "slow inlet establishment and prolonged initial stages," while preventing fluctuations from the initial stages from carrying over to the later stages. This is because a mass landing point confirmation is necessary after the first stage, and any execution differences from the initial stages are truncated at the stage boundary.

[0079] In the second stage of electrolyte injection, this project aims to promote uniform wetting and pore filling within the core, shifting the electrolyte distribution from "rapid entry" to "uniform diffusion and elimination of preferential channels." The control logic in this stage still prioritizes the target injection volume, but places greater emphasis on the stage duration window. This is because wetting inherently requires a certain residence time for capillary action, diffusion, and pore redistribution to take effect. Traditional methods often use a fixed pressure to prolong the entire process, resulting in neither speed nor uniformity. This project sets the second stage in a "liquid injection advance—short residence—re-advance" rhythm, ensuring that the electrolyte, after reaching the target mass increment, must undergo a specified wetting residence window before proceeding to the next step. Pressure in this stage is no longer maximized but serves as an auxiliary measure to maintain stable penetration and avoid forming new preferential channels. The key is to confirm, through stage weighing and flow decay criteria, that the cell's absorption of the electrolyte enters a predictable decay range and the mass increment stably falls within the second stage's target window before proceeding to the third stage. Through this combination of "segmented target quantity + time window + absorption behavior criterion", the second stage achieves engineered control of uniformity, rather than relying on luck or experience.

[0080] In the third stage of electrolyte injection, this project defines the goal as precise and consistent electrolyte injection and error convergence near the full capacity range. This is because, towards the end of the injection phase, small deviations in electrolyte mass can be amplified into capacity deviations, self-discharge differences, and internal resistance differences, thereby compromising the consistency of the aviation battery pack. The control strategy in this stage no longer relies on a fixed time or pressure, but instead uses a "mass approximation window" as the transition condition: when the weighing result after the second stage enters the approximation window to the final target electrolyte injection mass increment, the system automatically switches to the final precision control mode. In precision control mode, the electrolyte injection is performed in finer steps, and each step must be verified by weighing. If necessary, a replenishment loop is initiated until the error falls within the final mass window. The "segmented control" is most directly reflected here: the allowable single-push volume in the third stage is much smaller than that in the first two stages, and the duration window of the third stage has changed from "as fast as possible" to "as stable as possible within the allowable cycle time." An upper limit is set on the number of closed-loop cycles; exceeding this limit results in the cell being judged as abnormal and discarded, ensuring that batch statistical consistency is not dragged down by a few abnormal samples. Through the quality landing point closed-loop in the final stage, this project explicitly incorporates the quality errors caused by unavoidable splashing, needle residue, and valve cavity retention in traditional liquid injection into the control object, so that final consistency no longer depends on worker feel or accidental equipment conditions, but on closed-loop convergence.

[0081] The above three-stage process is guided by a key principle: the qualification of any stage is not solely determined by "reaching the set pressure" or "reaching the set time," but rather by "achieving the stage target injection volume and falling within the window," with "the stage duration falling within the allowable window" and "no abnormalities in the injection absorption behavior curve" as constraints. In this way, pressure becomes an execution variable serving the quality objective and time window, rather than the sole fulcrum of the process. Ultimately, this project formed a segmented control path for injection quality, progressing from coarse to fine, from fast to stable, and from advancement to convergence. The initial stage uses a large drive and a large injection volume to ensure efficiency; the middle stage uses alternating advancement and dwell to ensure uniformity; and the final stage uses fine steps and a closed-loop weighing replenishment system to ensure consistency.

[0082] In terms of project results, this project can be attributed to avoid falling into the narrow description of "only talking about pressure": By breaking down the liquid injection process into three stages with clearly defined target quantities, and by reducing ineffective waiting in the first stage, reducing hidden rework time caused by uneven wetting in the middle stage, and achieving convergence in the final stage through weighing closed loop instead of multiple trial and error, the overall liquid injection cycle time can be significantly shortened under the premise of safety and controllability. Tests show that the total liquid injection time can be shortened by about 20%. At the same time, since each stage has clear quality landing point confirmation and traceability records, the error convergence in the final stage compresses the liquid injection differences between cells into a smaller window, thereby providing more stable prior conditions for subsequent OCV screening, static self-discharge screening, and capacity-based precise grouping. This allows small-batch aviation cells to approach or even surpass the "one million cells in the same batch consistency" effect relied upon by large-scale automotive cell systems.

Claims

1. First claim: A three-dimensional consistency grading method for internal resistance-capacity-self-discharge of aerospace cylindrical quasi-solid-state lithium battery cells, characterized in that, It includes pre-screening, quality closed-loop injection and liquid replenishment, degassing and infiltration, in-situ curing of A / B two-component, three-stage gated grading, and generating steps of matching rules for grouping. Among them, the quality closed-loop injection and liquid replenishment takes the mass landing point after injection as the control target, forms a closed loop through multi-stage dosing and reweighing, sets at least two-level threshold windows and sets the upper limit of the closed-loop times. When the mass deviation Δm converges within the second threshold window, it is determined that the injection quality is qualified, otherwise, abnormal interception is triggered; The three-stage gated grading includes steady-state open-circuit voltage and internal resistance gating after formation, high-temperature standing self-discharge gating, and fine grading capacity gating in sequence, and obtains a three-dimensional grading result of internal resistance label, capacity label and self-discharge label; The matching rules generate a grouping scheme based on the basic principles of preferentially matching internal resistance in parallel, preferentially matching capacity and self-discharge in series. Second claim: The method according to the first claim, characterized in that, The pre-screening includes establishing a quality benchmark distribution of empty battery cells and grading them according to a narrow window, and classifying battery cells with similar effective loadings of active substances into the same candidate group. The width of the narrow window is determined by back-inference of the target capacity consistency, and is used to reduce the initial discreteness of the subsequent injection and curing consistency convergence. Third claim: The method according to the first claim, characterized in that, The quality closed-loop injection and liquid replenishment includes three stages: rough dosing, fine dosing and micro-compensation. The dosing step in the rough dosing stage is set according to the proportion of the target injection quality to quickly approach the target range. The dosing step in the fine dosing stage converges to a smaller proportion to make the mass deviation fall into the first threshold window. The micro-compensation stage uses multiple small-step dosing and fixed waiting for reweighing to make the mass deviation fall into the second threshold window, and the second threshold window is smaller than the first threshold window. Fourth claim: The method according to the first claim, characterized in that, A non-linear mapping fitting model for liquid replenishment decision-making is set, which maps the mass deviation Δm to the capacity deviation estimate ΔQ_est and / or the self-discharge deviation estimate ΔSD_est. The fitting model is in the form of a quadratic function or a piecewise quadratic function, and has model version number and applicable range fields, which are used for auditing and updating under different batches and different process windows. Fifth claim: The method according to the first claim, characterized in that, In the in-situ curing step of A / B two-component, component A contains monomers or oligomers that can be in-situ polymerized and flame retardant synergistic components, component B contains cross-linking agents and / or initiating systems and / or synergistic curing components. After A and B are mixed within the specified ratio range, a quasi-solid network is formed within a controlled temperature window, and the rheological index window, gelation time window and curing temperature integral after mixing are used as consistency traceability fields. Sixth claim: The method according to claim 5, characterized in that, Solid ceramic electrolyte particles are blended in the quasi-solid system to improve the voltage resistance and inhibit dendrites. The solid ceramic electrolyte particles are garnet-type lithium lanthanum zirconium oxide LLZO or its doped variants. The particle size distribution and the addition amount are within the specified range, and shear conditions and upper limit of agglomeration index are set for the dispersion process. The dispersion conditions and agglomeration index are used as consistency traceability fields. Seventh claim: The method according to the first claim, characterized in that, The high-temperature standing self-discharge gating in the three-stage gated grading includes placing the battery cells in a specified high-temperature window and standing for a specified time window, measuring the open-circuit voltage drift and self-discharge equivalent index and combining them with the internal resistance change to eliminate potential micro-short circuit or side reaction abnormal battery cells. Eighth claim: The method according to the first claim, characterized in that, A full-process traceability field system is established, wherein the traceability fields include at least the empty cell quality benchmark, liquid injection target quality, quality deviation Δm sequence, threshold window version number, fitting model version number, degassing and wetting procedure number, curing temperature integral, steady-state open-circuit voltage after formation, internal resistance test results, open-circuit voltage drift and self-discharge equivalent index after high-temperature static storage, precise capacity value and coulombic efficiency, and group selection rule output results, which are used for airworthiness verification, quality audit and failure traceability. Ninth claim: A three-dimensional consistency hierarchical system for implementing the method described in any one of claims 1 to 8, characterized in that, It includes a high-precision weighing unit, a micro-liquid injection and replenishment closed-loop unit, a degassing and wetting unit, an in-situ curing temperature control unit, an electrical performance acquisition and grading unit, and a traceability database and control software system. The control software system includes a threshold window management module, a closed-loop replenishment control module, a nonlinear mapping fitting module, a three-stage gating and grading module, a selection rule generation module, and a traceability audit module.

Images