Storage battery grouping online capacity checking method and system fusing health degree evolution model

Abstract

The invention discloses a storage battery grouping online capacity checking method and system fusing a health degree evolution model, and relates to the technical field of storage battery management, and the method comprises the following steps: S1, building a sensitive baseline, injecting a controlled short pulse, and measuring a voltage hysteresis area and a platform repetition rate for a tiny circulation working condition at the end of each charging and discharging period; generating a memory early warning index; s2, after the memory early warning index is obtained, an induction test is executed, and a grading threshold value is formed by comparing the capacity hysteresis residual error, the internal resistance transition rate and the open-circuit voltage deviation in the short-time charging and discharging process. According to the method, high-precision online health accounting and grouping differentiation operation limitation are realized through multi-parameter dynamic detection and grading judgment, and synchronous iteration of a management system and working conditions is guaranteed by matching operation result real-time write-back and baseline self-adaptive updating. The system can identify high-risk monomers in advance and dynamically group and reconstruct the high-risk monomers, thereby preventing failure and improving the safety and energy efficiency utilization rate of the battery pack.

Description

Technical Field

[0001] This invention relates to the field of battery management technology, specifically to a method and system for online capacity assessment of battery groups that integrates a health evolution model. Background Technology

[0002] "Online Capacity Assessment for Battery Groups Based on Integrated Health Evolution Model" refers to the process of battery management where a mathematical model reflecting the dynamic changes in battery health (i.e., health evolution) is constructed based on continuous observation of the long-term operating status of each individual battery or battery pack. This model is then combined with online monitoring data under actual operating conditions to achieve real-time calculation of remaining battery capacity (i.e., online capacity assessment). Based on this, batteries with similar performance or matching health status are automatically grouped into the same working group according to their current health and capacity status, optimizing the overall energy utilization and operational consistency of the battery pack. This method not only dynamically reflects the process of battery degradation and capacity loss, improving the scientific nature of group management, but also effectively reduces energy efficiency losses and operational risks caused by uneven battery performance, achieving efficient and intelligent management throughout the entire battery lifecycle.

[0003] Existing technologies have the following shortcomings: Current technologies for battery capacity calculation and health status assessment typically rely on dynamic monitoring of historical charge / discharge data and current operating status of individual battery cells. However, existing technologies struggle to sensitively identify the memory effect generated by battery cells subjected to prolonged low-rate charge / discharge cycles. Some battery cells, under frequent shallow charge / discharge or low-rate cycling conditions, gradually accumulate a memory effect, causing a gradual decrease in their actual usable capacity. This process is subtle and difficult for conventional health evolution models to detect in a timely manner. Because existing models fail to effectively identify this hidden degradation during dynamic tracking, online capacity calculations consistently overestimate the actual remaining energy level of such batteries. When the system faces sudden high-rate discharge demands, the actual release energy of the overestimated battery group is far lower than the model prediction. In extreme cases, this can lead to the synchronous failure of the entire battery group within a short period, causing a momentary power outage for critical loads. In severe cases, it can lead to core equipment downtime and system paralysis, posing significant safety hazards and operational risks.

[0004] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for online capacity assessment of battery groups based on a fusion health evolution model, in order to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an online capacity verification method for battery grouping based on a health evolution model, comprising the following steps: S1, at the end of each charge-discharge cycle, establishes a sensitive baseline for micro-cycle conditions, injects controlled short pulses, measures the voltage hysteresis area and plateau repetition rate, and generates a memory warning index; S2, after obtaining the memory warning index, performs an induced test, and forms a graded threshold by comparing the capacity hysteresis residual, internal resistance transition rate and open circuit voltage deviation during the short-term charge and discharge process; S3 embeds the grading threshold into the online capacity calculation process, sets a dynamic threshold based on the prediction residual, identifies high-risk battery cells with abnormal performance in real time, and outputs dynamic grouping instructions. S4, upon receiving the dynamic grouping instruction, implements operational limiting measures, including load transfer, rate limit setting, charge / discharge depth control, and local temperature management, to suppress memory effect accumulation and maintain stable grouped battery capacity; S5, after completing the operation limits, write the operation results back to the sensitive baseline, update the classification threshold, sampling period and pulse amplitude according to the residual change trend, and generate risk stratification data corresponding to different load levels; S6 automatically initiates a pre-inspection process after generating risk stratification data and before high-rate discharge is triggered. It calls the risk stratification data and simulates the same level of operating conditions to check the stability of the groups. When signs of capacity collapse are found, it immediately performs group reconstruction, forming a closed-loop process from baseline construction, degradation identification, threshold embedding, limitation control, threshold update to pre-inspection verification.

[0007] Preferably, step S1 includes: After completing one charge-discharge cycle, a high-precision data acquisition device is used to record the charging current, discharging current, charging voltage, discharging voltage, ambient temperature, battery casing temperature, estimated state of charge, and energy input and output of the target battery cell. The sampling period is no less than five times per second, and all raw data are stored in real time. Using the obtained energy input and output data, the charging capacity and discharging capacity of this cycle are calculated respectively, and compared with the previous cycle and historical data item by item to form sensitive baselines for maximum charging capacity, maximum discharging capacity, charging and discharging current variation range, temperature variation range, state of charge variation trend and terminal voltage variation range. After the sensitive baseline is established, a current pulse with an amplitude of 5% to 10% of the rated capacity is applied to the target battery cell. The pulse lasts for ten seconds. During the injection, the changes in battery terminal voltage and current are collected. The sampling frequency is no less than one hundred times per second to ensure the integrity of the voltage dynamic response data. Using high-resolution terminal voltage data obtained from short-pulse excitation, hysteresis curves of current input / output and voltage change are plotted, the hysteresis area is calculated, and the number of voltage plateau segments is counted. Combined with the parameters of the current cycle and historical sensitive baselines, a memory warning index is generated, and this memory warning index is used for the health grading of subsequent processes.

[0008] Preferably, step S2 includes: When the memory warning index is obtained and it is initially judged that the target battery cell has a memory effect accumulation trend, a short-term excitation charge-discharge experiment is performed on the battery using a precision adjustable current source. After the battery temperature reaches environmental equilibrium, it is first charged with a constant current of 20% of the nominal capacity for five minutes, and then discharged with a constant current of the same amount for five minutes. The charging and discharging current, voltage and temperature are collected in real time, and the sampling frequency is not less than one hundred times per second. After the experiment, the actual charging capacity and the actual discharging capacity were accumulated respectively. The capacity hysteresis residual was obtained by direct comparison, and its variation range and trend in the continuous period were analyzed. At the instants when charging switches to discharging and discharging switches to resting, the terminal voltage change data are collected respectively, the maximum jump value is counted and the equivalent internal resistance is calculated accordingly, and the internal resistance transition rate is further calculated. At the same time, the open circuit voltage is measured 30 minutes after the end of discharging to obtain the open circuit voltage deviation of this cycle. The capacity hysteresis residual, internal resistance transition rate, and open circuit voltage deviation are compared with sensitive baselines and historical data item by item. Based on the deviation magnitude, the data are divided into normal range, warning range, and danger range to form a classification threshold. Finally, the health level and risk classification of the battery cell are determined.

[0009] Preferably, step S3 includes: During each actual charge and discharge cycle, a high-precision acquisition device is used to collect the charging capacity, discharging capacity, terminal voltage, terminal current and shell temperature of each battery cell. The sampling frequency is maintained at more than 100 times per second, and the data is compared with sensitive baselines and historical data one by one. Based on the grading thresholds formed in the previous cycle, the capacity hysteresis residual, internal resistance transition rate, and open circuit voltage deviation are compared with the thresholds in detail to determine whether a single cell has crossed the normal, safe, warning, or dangerous range, and high-risk single cells are identified in real time based on the predicted residual dynamic threshold. High-risk cells are removed from the original group and a separate high-risk group is formed. The maximum discharge current is set to 60% of the nominal capacity, the depth of charge and discharge is limited to 30% to 80% of the rated capacity, and strict operating restrictions are implemented with the temperature fluctuation range not exceeding three degrees Celsius. The capacity calculation results of all new groups are verified, and the output energy and consistency changes are analyzed. The grouping results are used as the reference baseline for the next cycle. If abnormal cells are found, they are fed back to the subsequent dynamic threshold setting stage to achieve a continuous closed loop in the grouping adjustment and calculation process.

[0010] Preferably, step S4 includes: Upon receiving a dynamic instruction from the high-risk group, the output current of the high-risk group is gradually reduced, while the output current of the healthy group is simultaneously increased until the load of the high-risk group drops to 30%. The healthy group then takes over the remaining load. The entire process is completed within sixty seconds, ensuring that the voltage difference does not exceed 0.2 volts and that the load switching is smooth without any drops. For high-risk groups, a maximum rate limit is set. The maximum discharge current of a single cell shall not exceed 0.5 times the nominal capacity, and the charging current shall not exceed 0.3 times. Checks are conducted every ten minutes, and if any limit is exceeded, the rate is reduced to 0.4 times. Strictly control the depth of charge and discharge for high-risk groups. Set the discharge cutoff voltage to 3.1 volts and the maximum charging cutoff voltage to 4.0 volts. The depth of discharge shall not exceed 60% of the rated capacity and the charging capacity shall not exceed 85%. The sampling interval shall not exceed 30 seconds. The corresponding charging and discharging operation shall be terminated when the remaining capacity reaches the threshold. A thermal sensor is attached to the surface of each high-risk unit's casing, and temperature data is uploaded every ten seconds. When the unit temperature exceeds the ambient temperature by eight degrees Celsius, the fan is activated. If the temperature does not recover, the cooling plate is activated simultaneously until the temperature recovers. If the temperature is abnormal, load reduction, shortening of charge and discharge cycles, and extension of cooling time are implemented to ensure the stability of the high-risk group capacity.

[0011] Preferably, step S5 includes: After the high-risk group operation restriction measures are implemented, the actual total charging capacity, total discharging capacity, terminal voltage, discharging current, charging current, surface temperature and dynamic residual changes of each battery cell are summarized, and all raw data are numbered and archived, corresponding one-to-one with the data of the previous cycle and the sensitive baseline. The difference between the collected capacity retention, capacity hysteresis residual, internal resistance transition rate and open circuit voltage deviation and the sensitive baseline is calculated. If the parameter decreases, the new data point is written to the sensitive baseline. If the parameter fluctuates or deteriorates, a high-risk mark is made on the sensitive baseline, and the write-back is performed immediately after the data is synchronized. Based on the residual change trend of this period, the classification threshold, sampling period and excitation pulse amplitude are dynamically updated. When the residual decreases, the threshold and sampling period are relaxed, and when the residual increases, the threshold and period are tightened and the pulse excitation amplitude is adjusted. All parameters are updated and then incorporated into the health assessment process of the next period. By combining the new sensitive baseline and classification threshold, all battery cells and groups are risk-stratified and divided into three categories: high risk, medium risk and low risk. The stratification results are associated with all collected data and uploaded to the battery management center for use in grouping strategies and load allocation, ensuring a closed loop of monitoring and management.

[0012] Preferably, step S6 includes: Before planning high-rate discharge, based on the risk stratification results of the previous cycle, the capacity retention, capacity hysteresis residual, internal resistance transition rate, open circuit voltage deviation, temperature rise amplitude and grouping status of each battery cell are called up one by one to establish a pre-inspection data table and organize it into groups. Based on the pre-detection data, the operating conditions of each group were simulated under high-rate discharge conditions. The discharge current of the group was gradually increased, and the terminal voltage drop rate, discharge capacity, capacity hysteresis residual and surface temperature rise rate were recorded. The remaining capacity, internal resistance transition rate, temperature rise and open-circuit voltage recovery of all cells were monitored. Based on the simulation data, risk checks are performed on each group. If a single cell’s capacity decreases by more than 1% every five minutes, its internal resistance jump rate increases by more than 4%, its terminal voltage is lower than three volts, or its temperature rises by more than forty-five degrees Celsius, then the high-risk cell is removed from the main load group and regrouped according to its health level. The new grouping, individual unit simulation parameters, remaining capacity, temperature, internal resistance changes, and all key indicators are incorporated into the new cycle baseline database. All grouping optimization, capacity calculation, risk assessment, and pre-inspection processes are based on this database to achieve optimal grouping configuration and closed-loop management throughout the entire process.

[0013] The battery grouping online capacity assessment system, which integrates a health evolution model, includes a micro-cycle sensitive baseline and memory warning generation module, a multi-parameter hierarchical threshold identification module, an online capacity calculation and dynamic grouping module, a refined operation limitation and grouping management module, an operation result write-back and adaptive update module, and a high-rate pre-detection and grouping reconstruction module. The micro-cycle sensitive baseline and memory warning generation module establishes a sensitive baseline for micro-cycle conditions at the end of each charge-discharge cycle, injects controlled short pulses, measures the voltage hysteresis area and plateau repetition rate, and generates a memory warning index. The multi-parameter graded threshold identification module, after acquiring the memory warning index, performs an induced test and forms a graded threshold by comparing the capacity hysteresis residual, internal resistance transition rate and open circuit voltage deviation during the short-term charge and discharge process. The online capacity calculation and dynamic grouping module embeds the grading threshold into the online capacity calculation process, sets a dynamic threshold based on the prediction residual, identifies high-risk battery cells with abnormal performance in real time, and outputs dynamic grouping instructions. The refined operation restriction and group management module, upon receiving a dynamic grouping instruction, implements operation restriction measures, including load transfer, rate limit setting, charge / discharge depth control, and local temperature management, to suppress the accumulation of memory effect and maintain the stability of the grouped battery capacity. The result write-back and adaptive update module, after completing the operation constraints, writes the operation results back to the sensitive baseline, updates the classification threshold, sampling period and pulse amplitude according to the residual change trend, and generates risk stratification data corresponding to different load levels; The high-rate pre-inspection and group reconstruction module automatically starts the pre-inspection process before high-rate discharge is triggered after generating risk stratification data. It calls the risk stratification data and simulates the same level of operating conditions to check the group stability. When signs of capacity collapse are found, group reconstruction is immediately performed, forming a closed-loop process from baseline construction, degradation identification, threshold embedding, limit control, threshold update to pre-inspection verification.

[0014] The technical effects and advantages provided by the present invention in the above technical solution are as follows: This invention achieves high-precision online calculation of the true health status of batteries through dynamic detection and hierarchical discrimination of multi-dimensional physical quantities. It can also automatically implement differentiated operating restrictions for battery groups with different risk levels, effectively suppressing the further spread of adverse degradation. Real-time write-back of operating results and adaptive updates of sensitive baselines ensure that the entire health assessment and grouping management system is always iterated in sync with actual operating conditions. Ultimately, under extreme load scenarios such as high-rate discharge, the system can proactively identify and isolate high-risk cells, dynamically reconstruct groups, and avoid sudden failures and major safety hazards caused by capacity overestimation or group imbalance. This achieves closed-loop management throughout the entire process, from health data collection, risk identification, group optimization to intelligent control, significantly improving the safety, operational consistency, and energy efficiency of the battery pack throughout its entire lifecycle. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0016] Figure 1 This is a flowchart of the online capacity assessment method for battery grouping based on the present invention, which integrates a health evolution model.

[0017] Figure 2 This is a schematic diagram of the module of the battery grouping online capacity system that integrates the health evolution model of the present invention. Detailed Implementation

[0018] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0019] This invention provides, for example Figure 1 The online capacity allocation method for battery groups based on the fusion health evolution model, as shown, includes the following steps: S1, at the end of each charge-discharge cycle, establishes a sensitive baseline for micro-cycle conditions, injects controlled short pulses, measures the voltage hysteresis area and plateau repetition rate, and generates a memory warning index; To address the challenge of sensitively identifying the memory effect generated in individual battery cells under minute charge-discharge cycles, a specific method is proposed to provide early warning of the memory effect using active excitation and multidimensional response measurement, with each charge-discharge cycle as a unit. This method enables the identification of early signs of degradation within the battery through clear and operable steps, providing fundamental data support for subsequent battery capacity assessment and group management. The specific steps are as follows: After each complete charge-discharge cycle, detailed operational data is collected from the target battery cell. Specifically, a high-precision data acquisition device is used to record the battery's charging current, discharging current, charging voltage, discharging voltage, ambient temperature, battery casing temperature, estimated SOC (State of Charge), and actual energy input and output during the cycle. To ensure data accuracy, the acquisition cycle should be no less than five times per second, and all raw data must be stored in real time. After acquisition, the obtained energy input and output data for the cycle are used to calculate the current charging and discharging capacity, which are then compared and analyzed item by item with the previous cycle and historical operating data to form a sensitive baseline for the current cycle's characteristics. This sensitive baseline specifically includes: the maximum charging capacity, maximum discharging capacity, charging and discharging current variation range, temperature variation range, SOC trend, and battery terminal voltage variation range for the current cycle. By normalizing the above baseline parameters with historical data sequences, the specific operating state experienced by the target battery cell during the current cycle can be accurately characterized, providing detailed background information for subsequent physical excitation and response measurement.

[0020] After establishing the sensitive baseline, immediately apply a short pulse signal with preset parameters to the target battery cell. Specifically, using an external current source or pulse generator, inject a single current pulse with an amplitude of 5%–10% of the rated capacity after the battery is in a quiescent state and its voltage and current are stable. The pulse duration is set to 10 seconds. The pulse type can be a constant current pulse, and the specific value should be adjusted according to the actual battery type and safety operating standards. During the pulse injection, high-frequency sampling of voltage changes across the battery terminals is performed using a synchronous measuring device. A sampling frequency of at least 100 times per second is recommended to ensure the complete voltage dynamic response process is recorded. Before and after pulse application, multiple data points should be recorded, including battery terminal voltage, terminal current, surface temperature, and ambient temperature. All data should be uploaded in real time and backed up locally. This operation actively stimulates the internal reaction mechanism of the battery while ensuring safe operation, obtaining crucial data for subsequent memory effect assessment. Unlike existing technologies that rely solely on passive acquisition, this method actively introduces external stimuli to make the micro-states such as changes in battery internal resistance, electrode reactivity, and polarization processes externally measurable, providing a reliable physical signal source for identifying the accumulation of early memory effects.

[0021] Based on the high-resolution terminal voltage response data of the battery obtained under the aforementioned short-pulse excitation, quantitative physical quantity extraction and analysis were carried out. The specific steps were as follows: First, the hysteresis curves of current input / output and voltage change were plotted using the raw acquired data, and the voltage hysteresis loop area formed during the pulse excitation process was calculated. The voltage hysteresis area was determined by accurately calculating the closed region enclosed by the current-voltage curve in the forward and reverse phases using numerical integration, obtaining the actual area data in volt-amperes. Subsequently, the number of terminal voltage plateau segments occurring within the current cycle was counted, and the start and end current and voltage ranges of the plateau segments were analyzed to obtain the voltage plateau repetition rate, which is equal to the ratio of the number of plateau segments occurring within the current cycle to the number of historical typical plateau segments. To improve measurement accuracy, the hysteresis area and plateau repetition rate measured in the current cycle were compared with sensitive baseline parameters one by one, and their variation amplitude and rate of change were calculated. Through the above data processing, the structural changes inside the battery caused by shallow charge-discharge cycles, such as increased polarization, reduced reaction area of ​​electrode active materials, and increased charge transfer barriers, can be effectively reflected. This process does not rely on traditional single capacity estimation, and can achieve quantitative analysis of microscopic physical quantities, thereby improving the accuracy of identifying early memory effects.

[0022] Based on the obtained voltage hysteresis area and plateau repetition rate, a memory warning index is generated by combining sensitive baseline parameters from the current cycle and historical cycles. The generation process involves weighted comparison of the current cycle's voltage hysteresis area, plateau repetition rate, charge / discharge capacity, and SOC change trends with historical baseline data. Through multi-parameter synthesis, a numerical indicator reflecting the potential accumulation level of memory effect in a single battery cell is obtained. If the memory warning index for the current cycle shows a significant increase compared to the previous cycle and deviates from the historical average baseline, the battery is considered to have an initial risk of memory effect. At this point, the memory warning index serves as the core criterion for triggering induced testing and further health grading in subsequent steps, and is transmitted in real-time to the next process stage.

[0023] The purpose of this step is to achieve early, highly sensitive identification of the memory effect under micro-cycle conditions by actively introducing controlled short-pulse signals into individual battery cells at the end of each charge-discharge cycle, combined with the establishment of a sensitive baseline and high-precision measurement of multiple physical quantities. Unlike traditional methods that rely solely on historical charge-discharge data or conventional monitoring parameters, this step can capture potential degradation characteristics accumulated due to shallow charge-discharge and low-rate cycling before battery performance shows significant decline. This is achieved through refined collection of parameters such as charge-discharge capacity, terminal voltage changes, voltage hysteresis area, and plateau repetition rate. Quantitative analysis of voltage hysteresis area and plateau repetition rate not only reveals microscopic changes such as increased internal polarization and reduced electrode reactivity but also dynamically reflects structural anomalies in the battery. Ultimately, the generation of the memory warning index provides a scientific and quantifiable input basis for subsequent targeted induced testing, health grading identification, and group management, fundamentally improving the scientific rigor and real-time nature of battery health status management and effectively avoiding the risks of capacity overestimation and sudden failure under traditional methods.

[0024] S2, after obtaining the memory warning index, performs an induced test, and forms a graded threshold by comparing the capacity hysteresis residual, internal resistance transition rate and open circuit voltage deviation during the short-term charge and discharge process; After obtaining the memory warning index of a single battery cell, a multi-dimensional quantitative analysis of the battery's capacity hysteresis residual, internal resistance transition rate, and open-circuit voltage deviation under micro-cycle conditions is further conducted through a precise and specific induced testing process. This analysis leads to the formation of operable and repeatable graded thresholds. The specific steps are as follows: Once the memory warning index has been obtained and it is preliminarily determined that the target battery cell exhibits a tendency for memory effect accumulation, a short-term excitation charge-discharge experiment is performed on the battery according to preset experimental parameters. Specifically, after the battery temperature reaches environmental equilibrium and is left to rest for 15 minutes, a precision adjustable current source is used to continuously charge the battery at a current 0.2 times its nominal capacity for 5 minutes using a constant current. Immediately afterwards, the current is switched to a discharge current of 0.2 times the nominal capacity for 5 minutes of constant current discharge. During this period, charging current, discharging current, terminal voltage, casing temperature, and ambient temperature are collected in real time at a sampling frequency of no less than 100 times per second to ensure that all dynamic processes are completely recorded. Through the above charge-discharge operation, local response changes occur in the electrodes, separator, and electrolyte inside the battery, stimulating physical characteristics that are not easily detected by conventional monitoring methods under long-term shallow charge-discharge cycles.

[0025] After the short-term charge-discharge experiment, the obtained raw data were calculated in detail. First, the actual charging capacity and actual discharging capacity were accumulated separately, accurate to 0.001 ampere-hours. Then, using a point-to-point analysis method, the total capacity released at the end of the discharge was directly compared with the input capacity of the charge just completed, yielding the capacity hysteresis residual. The capacity hysteresis residual is the difference between the charged energy and the released energy, measured in ampere-hours. This value reflects the energy loss and polarization reaction that occurs in the battery under short-term dynamic excitation. By analyzing the variation amplitude and trend of the capacity hysteresis residual over multiple consecutive cycles, latent signs of enhanced polarization and reduced reversible reaction within the battery can be revealed.

[0026] Throughout the aforementioned short-duration charge and discharge process, especially at the instant the current switches from charging to discharging and at the end of discharging before switching to rest, the sudden change amplitude of the battery terminal voltage is recorded. Specifically, 5 seconds of data are collected before and after each current switch, the maximum voltage jump value at the instant of current change is statistically analyzed, and the equivalent internal resistance at each moment is calculated based on Ohm's law. The equivalent internal resistances of the two switches from charging to discharging and from discharging to rest are compared to calculate the internal resistance transition rate of this cycle. This value reflects the charge transfer efficiency, polarization degree, and electrode material state changes of the battery under different operating conditions. During this process, it is also necessary to compare the open-circuit voltages at both ends of the battery before and after the short-duration excitation, especially measuring the difference between the open-circuit voltage before excitation and the open-circuit voltage 30 minutes after the end of discharging, as the open-circuit voltage deviation for this cycle. The open-circuit voltage deviation reflects the rate of equilibrium recovery of the microscopic reaction within the battery and the stability of the electrode surface state.

[0027] The measured results of three specific physical quantities—capacity hysteresis residual, internal resistance transition rate, and open-circuit voltage deviation—are compared item by item with previously established sensitive baselines and historical data. Based on the deviation between the actual measured value and the normal range of the baseline, each parameter is divided into normal, warning, and danger ranges. For example, if the capacity hysteresis residual is greater than 0.005 AH for three consecutive measurements, the internal resistance transition rate is greater than 10%, and the open-circuit voltage deviation is greater than 10 mV, it is considered a high-risk state; if only one indicator exceeds the limit, it is considered a moderate-risk state; and if all parameters are within the normal range, it is considered a safe state. The above-mentioned classification threshold results will be comprehensively analyzed with the memory warning index obtained in the previous cycle to ultimately determine the current health level and risk classification of the battery cell, providing a clear stratification basis for subsequent online capacity calculation and grouping strategy adjustments.

[0028] The purpose of this step is to systematically and comprehensively quantify and evaluate the battery's true capacity hysteresis residual, internal resistance transition rate, and open-circuit voltage deviation under dynamic operating conditions by performing short-term charge-discharge excitation experiments with clearly defined parameters on battery cells that have already exhibited memory warnings. Unlike traditional health assessments that rely solely on static voltage, current data, or single charge-discharge capacity, this step can specifically stimulate and expose latent changes such as internal polarization, reduced reactivity, and structural degradation within the battery in a controlled environment. Through multidimensional and continuous measurement of key physical quantities, the previously difficult-to-capture memory effect accumulation state is transformed into specific, directly comparable, and categorizable values. This not only improves the sensitivity of identifying early abnormal states of the battery but also provides a scientific quantitative basis for subsequent health grading, risk threshold setting, and group management, effectively avoiding systemic risks caused by capacity overestimation and abnormal cell grouping, thereby enhancing the safety and operational consistency of the entire battery pack.

[0029] S3 embeds the grading threshold into the online capacity calculation process, sets a dynamic threshold based on the prediction residual, identifies high-risk battery cells with abnormal performance in real time, and outputs dynamic grouping instructions. After determining the grading thresholds for capacity hysteresis residual, internal resistance transition rate, and open-circuit voltage deviation of individual battery cells, these grading thresholds are further integrated into the online capacity calculation process. Through dynamic threshold settings and real-time multi-parameter discrimination, accurate identification and grouping adjustment of high-risk battery cells are achieved, ensuring the stability and safety of the overall battery pack performance. The specific steps are as follows: After obtaining the aforementioned individual cell classification thresholds, a capacity calculation is performed on each individual battery cell within each actual charge / discharge cycle. Specifically, using high-precision current and voltage acquisition devices, the actual charging capacity, actual discharging capacity, battery terminal voltage, terminal current, and casing surface temperature of each battery cell are collected in real time during the current cycle, with a sampling frequency maintained at over 100 times per second. All the collected data must be stored synchronously and compared one by one with existing sensitive baselines and historical data sequences. During capacity calculation, the input power during the charging phase and the output power during the discharging phase of each battery cell are strictly checked to ensure that all data remain accurate and consistent within the same physical unit. At the same time, based on the classification thresholds formed in the previous cycle, the capacity hysteresis residual, internal resistance transition rate, and open-circuit voltage deviation collected in the current cycle are compared in detail with their classification thresholds to determine whether the current operating status of the battery cell has crossed the normal, safe, warning, or dangerous range.

[0030] Based on the data comparison results above, a dynamic threshold for the predicted residual is set in real time for the current cycle. Specifically, this involves first calculating the difference between the actual capacity and theoretical capacity of each battery cell throughout the entire charging and discharging process, and then comparing this difference with the historical average and the classification threshold. For example, if the capacity hysteresis residual of a cell exceeds 0.005 amps for three consecutive cycles, the internal resistance transition rate is greater than 10% of the previous cycle, and the open-circuit voltage deviation is greater than 10 millivolts, then the cell is marked as high-risk in the current cycle. This judgment standard does not rely on subjective settings but is closely based on the aforementioned classification threshold, ensuring the objectivity and repeatability of each judgment. Cells judged to be in a high-risk state are promptly marked in the database and compared horizontally with the calculation results of all other cells to ensure that all high-risk cells are accurately identified in the current cycle.

[0031] After identifying high-risk cells, grouping adjustments are implemented based on health level and operating status. Specifically, all cells marked as high-risk in this cycle are removed from existing groups and placed in a separate high-risk group. Strict restrictions are placed on the subsequent operating strategy of this group, such as setting its maximum allowable discharge current to 60% of its nominal capacity, strictly controlling the depth of charge / discharge between 30% and 80% of its rated capacity, and strengthening real-time monitoring of the internal battery casing temperature to ensure temperature fluctuations do not exceed 3 degrees Celsius. For the remaining cells not identified as high-risk, they are divided into several new groups with consistent health levels based on their current capacity consistency, internal resistance, and voltage fluctuation range. The load allocation and operating current of the new groups are determined by the average health level within the group; high-health groups can be assigned higher loads, while low-health groups adopt more conservative operating strategies. The formation of each group is determined by both real-time data from this cycle and the aforementioned grading thresholds, ensuring the rationality and consistency of the grouping.

[0032] After completing all grouping adjustments for this cycle, the online capacity calculation results for all groups and individual cells are summarized and verified. Specifically, this includes checking the consistency of actual charging capacity, discharging capacity, terminal voltage changes, and temperature data for all cells within the new group, and performing a full comparison with the previous cycle's grouping results, calculation residuals, and health classifications. The analysis focuses on the output energy, consistency changes, and temperature equilibrium levels of each group after this cycle's adjustments. If the abnormal residuals of high-risk cells are effectively suppressed after this cycle's grouping adjustments, and the parameters within each group tend to be consistent, the current grouping results are used as the reference baseline for the next cycle, and the sensitive baseline data is updated. If some cells are found to continue exhibiting large capacity residuals, large internal resistance fluctuations, or abnormal temperatures within the new group, these cells are promptly re-included in the high-risk monitoring queue, and their detailed data is fed back to the subsequent dynamic threshold setting stage.

[0033] The purpose of this step is to directly integrate multi-dimensional classification thresholds such as capacity hysteresis residual, internal resistance transition rate, and open-circuit voltage deviation into the online capacity calculation process of each cycle. Combined with real-time collected physical quantities such as charge / discharge capacity, voltage, and temperature, targeted prediction residual thresholds are dynamically set to identify high-risk battery cells exhibiting abnormal performance in real time. Through high-frequency data comparison and classification judgment, it can not only respond quickly when early signs of battery performance degradation appear, but also accurately distinguish cells in different health states and output group adjustment instructions based on the identification results. This effectively avoids the misallocation of high-risk cells due to inaccurate capacity estimation, and prevents inconsistent cell health states from leading to overall battery pack performance degradation and system safety hazards. Simultaneously, through real-time group adjustment, targeted management and load allocation optimization of high-risk cells are achieved, improving the overall energy efficiency and operational consistency of the battery pack, ensuring that the entire battery system continues to operate in a highly efficient, safe, and controllable state throughout its entire lifecycle.

[0034] S4, upon receiving the dynamic grouping instruction, implements operational limiting measures, including load transfer, rate limit setting, charge / discharge depth control, and local temperature management, to suppress memory effect accumulation and maintain stable grouped battery capacity; Upon receiving the dynamic grouping instruction, a series of detailed and specific operational restriction measures are implemented for battery cells identified as high-risk groups. These measures work together across multiple dimensions, including load allocation, rate setting, charge / discharge depth control, and temperature management, to effectively suppress the continuous accumulation of memory effects and ensure stable and safe grouped capacity. The implementation process of each step is described in detail below: Upon receiving dynamic instructions from the high-risk group, load transfer measures are immediately implemented. The specific method involves smoothly transferring part or all of the electrical load currently handled by the high-risk group to a healthier group, following a pre-set switching sequence. During operation, the output current of the high-risk group is gradually reduced in 1A increments, while the output current of the healthy group is simultaneously increased. Each adjustment lasts 10 seconds, until the load on the high-risk group drops to 30% of its original level, at which point the healthy group takes over the remaining load. Throughout this process, the voltage changes between the high-risk and healthy groups are continuously monitored to ensure the voltage difference does not exceed 0.2V, and the entire load transfer process is completed within 60 seconds, guaranteeing uninterrupted power supply without any momentary drops. After the load transfer is complete, the healthy group carries the main load, while the high-risk group retains only a small amount of auxiliary or backup load, thereby reducing its discharge pressure and effectively preventing sudden disconnections caused by inflated capacity or severe polarization of the high-risk battery under high load.

[0035] After load transfer, a rate cap is set for high-risk groups. Specifically, the maximum allowable discharge current of a single cell in a high-risk group is set to no more than 0.5C of its nominal capacity, and the maximum allowable charging current is set to no more than 0.3C of its nominal capacity. For example, for a battery with a rated capacity of 100Ah, its discharge current must not exceed 50A, and its charging current must not exceed 30A. Within this range, all cells in each high-risk group must be monitored by an independent current shunt. If the current of a single cell exceeds the set value, its current load is immediately reduced via an electronic switch or bypass circuit to ensure that its discharge and charging rates remain within a safe range. Simultaneously, the execution of the rate setting is checked every 10 minutes. If a cell is found to exceed the set rate twice consecutively, it is automatically reduced to 0.4C until compliance is achieved, preventing new polarization exacerbation and memory effects induced by high-rate operation.

[0036] Based on the set upper limit of the charging rate, strict control of charge and discharge depth is implemented for high-risk groups. Specifically, the allowable discharge cut-off voltage for each battery in the high-risk group is set to 3.1V (taking a typical lithium battery as an example), the maximum charging cut-off voltage is limited to 4.0V, the discharge depth does not exceed 60% of the rated capacity, and the charging termination capacity does not exceed 85% of the rated capacity. All charging and discharging processes use fixed-point sampling, with each sampling interval not exceeding 30 seconds. After each charging and discharging process, the remaining capacity is immediately checked. If any single cell has reached the threshold, the corresponding charging and discharging operation is immediately terminated to prevent over-discharge or over-charge. After each charge and discharge cycle, the change in remaining capacity is recorded. After accumulating 3 cycles, if the depth deviates from the set threshold, the charge and discharge termination point is lowered accordingly to ensure that the battery is always in a shallow cycle state, effectively delaying capacity loss and structural degradation.

[0037] To further ensure the safe operation of high-risk battery groups, localized temperature management measures are implemented for all high-risk battery groups. Specifically, a thermistor is attached to the surface of each high-risk cell's casing to collect temperature data in real time, automatically uploading it to the central data acquisition terminal every 10 seconds. When the cell temperature exceeds the ambient temperature by 8 degrees Celsius, the air-cooled fan at that cell's location is immediately activated, with the fan speed linearly adjusted according to the temperature rise. If the temperature does not return to within 6 degrees Celsius of the ambient temperature within 30 seconds, the cooling plate is simultaneously activated, increasing the circulating coolant flow to its maximum and continuously cooling until the cell temperature returns to normal. For cells exhibiting persistently high temperatures, a 15-minute cooling period is scheduled before the next charge / discharge cycle to prevent further heat accumulation and deterioration. All temperature data is analyzed in conjunction with charge / discharge data. When abnormal temperatures occur simultaneously with abnormal capacity or rate, priority is given to measures such as reducing load, shortening charge / discharge cycles, and extending cooling time to ensure that temperature control and electronic control work together to effectively break the chain of memory effect accumulation.

[0038] The purpose of this step is to immediately implement multiple specific operational restrictions on high-risk battery groups after identification, including load transfer, rate cap setting, strict control of charge / discharge depth, and dynamic local temperature management. This reduces the workload and environmental stress of high-risk battery cells at the source, minimizing the continuous accumulation and spread of memory effects. By gradually and smoothly distributing the main load to healthier groups, high-risk groups only bear limited, low-intensity backup loads, and their charge / discharge current and capacity range are strictly limited, effectively slowing down internal polarization and the degradation rate of active materials. Simultaneously, combined with high-frequency temperature monitoring and active heat dissipation or cooling regulation, the risk of thermal runaway can be mitigated in a timely manner, preventing further performance degradation due to rising temperatures. These measures work together to ensure the stability of capacity retention in high-risk groups and the overall safety of the battery pack during operation, effectively improving energy utilization efficiency and providing a solid foundation for subsequent dynamic optimization of battery groups and full life-cycle health management.

[0039] S5, after completing the operation limits, write the operation results back to the sensitive baseline, update the classification threshold, sampling period and pulse amplitude according to the residual change trend, and generate risk stratification data corresponding to different load levels; After implementing operational restrictions for high-risk groups, the actual operational results of this cycle are systematically written back and dynamically optimized adaptively. Through step-by-step, specific technical operations, real-time updates of sensitive baselines and adaptive adjustments to tiered thresholds and monitoring parameters are achieved. Based on this, risk stratification data for different load levels is generated, providing a highly reliable foundation for subsequent grouping and scheduling. The technical implementation process of each step is described in detail below: After all high-risk group operational restrictions were implemented, detailed operational data for each group and individual cell within the current cycle were comprehensively collected and organized. During operation, the actual total charging capacity, total discharging capacity, entire process of terminal voltage changes, high-frequency collected discharge current, charging current, surface temperature, and dynamic residual changes for each battery cell were first summarized. Particular attention was paid to the specific changes in capacity retention, capacity hysteresis residual, internal resistance transition rate, and open-circuit voltage deviation after load transfer, rate limits, depth of charge / discharge, and temperature management for high-risk groups. All raw data were numbered and archived chronologically, and matched item by item with data from the previous cycle and the original sensitive baseline to establish a detailed dataset for the current cycle, ensuring data completeness and providing a true and detailed basis for subsequent baseline correction.

[0040] All collected key operating parameters are compared one by one with the previously established sensitive baselines, and the data is written back in a timely manner based on the changes. In practice, the capacity retention, capacity hysteresis residual, internal resistance transition rate, and open-circuit voltage deviation for the current cycle are first categorized and organized by group and individual cell, and then the difference is directly calculated with the corresponding historical parameters on the sensitive baseline. If it is found that the capacity hysteresis residual in the high-risk group of the current cycle has decreased by more than 0.002 amp-hours compared to the previous cycle, or the internal resistance transition rate has decreased from 12% to 9%, the new data point is directly written to the corresponding position on the sensitive baseline as the latest evaluation benchmark. If it is found that a parameter fluctuates drastically or continues to deteriorate, such as the capacity hysteresis residual increasing instead of decreasing, a high-risk mark is made on the sensitive baseline as a warning for subsequent grouping and tightening of control thresholds. All write-back operations should be performed immediately after data synchronization is completed to ensure that the sensitive baseline always reflects the most accurate current operating health status of the battery.

[0041] Based on the trend of residual changes in the current cycle, all classification thresholds, monitoring sampling periods, and excitation pulse amplitudes related to health assessment are dynamically updated. In implementation, the rate of change of each group and individual unit in key parameters between the current cycle and the previous baseline is first calculated. When the rate of decrease in capacity hysteresis residual, internal resistance transition rate, or open-circuit voltage deviation exceeds a predetermined threshold (e.g., the decrease exceeds 10% for two consecutive cycles), the capacity hysteresis residual classification threshold is increased from 0.005 AH to 0.006 AH, and the sampling period is correspondingly widened from once every 30 seconds to once every 45 seconds to reduce system load. When the residual increase rate accelerates or the internal resistance transition rate change exceeds historical extremes, the classification threshold and risk judgment range are tightened, the sampling period is shortened to 15 seconds, and the pulse excitation amplitude is increased by 10% to improve the sensitivity to abnormal trends. After all parameters are updated, they are immediately incorporated into the online monitoring and health assessment process for the next cycle, achieving an adaptive closed loop of monitoring and feedback.

[0042] Based on the updated sensitive baseline, grading thresholds, and monitoring parameters, all battery cells and groups in this cycle are risk-stratified. Specifically, all cells are mapped one-to-one with the new baseline and thresholds based on key indicators such as capacity retention, capacity hysteresis residual, internal resistance transition rate, and open-circuit voltage deviation in this cycle, and are classified into high-risk, medium-risk, and low-risk categories according to the range of values. For cells classified as high-risk, it is explicitly required that they can only be assigned to auxiliary loads or backup circuits in the next load scheduling cycle; medium-risk cells can participate in regular loads; and low-risk cells are prioritized for allocation to main loads and high-rate tasks. The above stratification results are all correlated with the collected charge / discharge data, temperature data, rate settings, charge / discharge depth, etc., and uploaded to the battery management center in real time as an important input basis for grouping strategies and future scenario predictions. All risk stratification data are automatically updated with each cycle to ensure that subsequent health management measures are highly consistent with actual battery operation.

[0043] The purpose of this step is to comprehensively and promptly write back the actual operating results of each cycle after the end of operational constraints to the original sensitive baseline. Combined with the changing trends of key parameters such as residuals, dynamic adjustments are made to various grading thresholds, sampling periods, and pulse excitation amplitudes to achieve adaptive optimization of the battery health management strategy. By meticulously archiving, comparing, and correcting physical indicators such as capacity hysteresis residuals, internal resistance transition rates, and open-circuit voltage deviations, not only is it ensured that the sensitive baseline reflects the latest operating status, but it can also automatically tighten or loosen grading thresholds and adjust sampling frequency and pulse intensity in response to changes in risk enhancement or health recovery, thereby improving monitoring sensitivity or reducing system load. Finally, combining the updated baseline and thresholds, all individual cells and groups are stratified into multi-dimensional indicators, forming risk stratification data for different load levels. This provides a scientific basis for group optimization and high-risk early warning in the next cycle, constructing a closed-loop health management system throughout the entire lifecycle. This process ensures the real-time and targeted nature of battery system health judgments, effectively improving operational safety and maintenance efficiency.

[0044] S6, after generating risk stratification data, automatically starts the pre-inspection process before high-rate discharge is triggered, calls the risk stratification data and simulates the same level of operating conditions, checks the group stability, and immediately performs group reconstruction when signs of capacity collapse are found, forming a closed-loop process from baseline construction, degradation identification, threshold embedding, limit control, threshold update to pre-inspection verification. After generating risk stratification data, for high-rate discharge scenarios, a management closed loop of self-inspection and dynamic optimization is formed through detailed group pre-inspection, operating condition simulation, group risk verification and reconstruction, ensuring that each battery group still has sufficient stability and safety redundancy under high stress loads. The specific implementation steps are as follows: Before planning high-rate discharge operations, all individual battery cells and groups are pre-checked according to the risk stratification results established in the previous cycle. During operation, the latest capacity retention, capacity hysteresis residual, internal resistance transition rate, open-circuit voltage deviation, historical temperature rise, and group status from the previous cycle are retrieved for each battery cell. All cells are grouped into high-risk, medium-risk, and low-risk groups, and detailed operating data for all cells within each group are compiled. For high-risk groups, the capacity retention decrease, capacity hysteresis residual change, maximum internal resistance transition rate, and peak temperature are summarized for the most recent three cycles. A dedicated pre-check data table is created for all data, organized by group, cell, and time series, ensuring that every subsequent operating condition simulation and risk verification is based on complete physical quantities, leaving no information blind spots.

[0045] Based on the above data table, a full-process simulation of the actual operating conditions was conducted for each group under high-rate discharge conditions. During actual operation, the total load was proportionally allocated to each group according to the upcoming high-rate discharge current setting, and the discharge current of each group was gradually increased. The rate of voltage drop, actual discharge capacity, capacity hysteresis residual change, and surface temperature rise rate of each cell under the simulated load were recorded. Throughout the simulation, not only was the overall output capability of the group examined, but the actual remaining capacity, extreme values ​​of internal resistance transition rate, duration of temperature rise, and open-circuit voltage recovery curve of each individual battery were also monitored. If, during the simulation, a cell was found to have a capacity drop rate greater than 1% every 5 minutes, an internal resistance transition rate increase exceeding 4%, a terminal voltage drop below 3.0V, or a surface temperature exceeding 45°C, that cell was immediately classified as a potential collapse risk. The entire simulation process was required to cover all load levels, progressing continuously from low to high rates, to fully verify the actual response of the groups under various extreme load conditions.

[0046] Based on the simulation data, a detailed risk check and reconstruction are performed on the groupings. Specifically, the total capacity, minimum terminal voltage, average temperature rise, and individual cell internal resistance fluctuation range of each group under simulation are rigorously compared. If two or more indicators of a single cell in a group exceed the safety threshold, all high-risk cells in that group are immediately removed from the main load group and reclassified into a standby group, or a controlled operation scheme is set separately according to parameters such as capacity, internal resistance, and temperature rise. Groups that do not exhibit sudden capacity drops, abnormal internal resistance, or excessively rapid temperature rise are retained as main load groups. After each group reconstruction, the new composition of each group, the simulated limit parameters of each cell, the remaining capacity, and the historical changes in temperature and internal resistance are all incorporated into the baseline database for the new cycle for subsequent dynamic calculations and group optimization. In this way, each grouping is based on the most realistic and rigorous high-rate operating condition verification data, ensuring that all battery cells are assigned to a group and load level that matches their health level.

[0047] The system automatically summarizes all details of sensitive baselines, grading thresholds, simulated operating conditions for each group, capacity residuals, and risk reconstruction within the current cycle. It stores all data, including the latest grouping results, various physical indicators under simulated operating conditions, risk stratification levels, collapse risk cell identifiers, adjusted grouping allocation schemes, and dynamic thresholds and sampling parameters for the next cycle. All subsequent grouping optimizations, online capacity calculations, risk grading determinations, operational restrictions, and pre-inspection processes are based solely on the latest archived data, ensuring that each group can automatically complete closed-loop self-checks and optimal group configuration before each high-rate discharge or other extreme operating condition. In this way, the entire process is technically detailed, data traceability is clear, and key criteria are fully repeatable, systematically eliminating risks such as capacity overestimation, group imbalance, and sudden failures, providing a scientific guarantee for high safety and high reliability management of the battery pack throughout its entire lifecycle.

[0048] The purpose of this step is to automatically conduct targeted pre-inspection of battery groups and high-intensity operating condition simulations using the latest generated risk stratification data before high-rate discharge is triggered. This involves examining key indicators such as capacity retention, internal resistance changes, terminal voltage, and temperature rise of each group and individual cell under high stress loads, identifying potential risks of capacity collapse or group consistency imbalance in advance. If any group exhibits signs of instability in the simulation, such as a sudden drop in capacity, a jump in internal resistance, or abnormal temperature rise, the system immediately performs group reconstruction, promptly isolating high-risk cells and reallocating loads and operating strategies. This completely avoids sudden disconnections, output interruptions, or even systemic failures caused by capacity overestimation or group imbalance. This step, through a closed-loop implementation of dynamic pre-inspection, simulation verification, and group reconstruction, shifts group management from static prediction to dynamic testing and real-time response, effectively improving the stability and safety redundancy of the entire battery pack under extreme conditions, providing a solid guarantee for intelligent and safe operation throughout its entire lifecycle.

[0049] The online capacity assessment method for battery groups, which integrates the aforementioned health evolution model, significantly improves the sensitivity and timeliness of identifying memory effects and latent capacity degradation in individual battery cells under long-term micro-cycle conditions. This method not only achieves high-precision online calculation of the battery's true health status through dynamic detection and hierarchical discrimination of multi-dimensional physical quantities, but also automatically implements differentiated operational restrictions for battery groups with different risk levels, effectively suppressing the further spread of adverse degradation. Real-time write-back of operational results and adaptive updates of sensitive baselines ensure that the entire health assessment and group management system iterates synchronously with actual operating conditions. Ultimately, under extreme load scenarios such as high-rate discharge, the system can proactively identify and isolate high-risk cells, dynamically reconstruct groups, and avoid sudden failures and major safety hazards caused by capacity overestimation or group imbalance. This achieves closed-loop management throughout the entire process, from health data collection, risk identification, group optimization to intelligent control, significantly improving the safety, operational consistency, and energy efficiency of the battery pack throughout its entire lifecycle.

[0050] This invention provides, for example Figure 2 The battery grouping online capacity assessment system shown in the diagram, which integrates a health evolution model, includes a micro-cycle sensitive baseline and memory warning generation module, a multi-parameter hierarchical threshold identification module, an online capacity calculation and dynamic grouping module, a refined operation limitation and grouping management module, an operation result write-back and adaptive update module, and a high-rate pre-detection and grouping reconstruction module. The micro-cycle sensitive baseline and memory warning generation module establishes a sensitive baseline for micro-cycle conditions at the end of each charge-discharge cycle, injects controlled short pulses, measures the voltage hysteresis area and plateau repetition rate, and generates a memory warning index. The multi-parameter graded threshold identification module, after acquiring the memory warning index, performs an induced test and forms a graded threshold by comparing the capacity hysteresis residual, internal resistance transition rate and open circuit voltage deviation during the short-term charge and discharge process. The online capacity calculation and dynamic grouping module embeds the grading threshold into the online capacity calculation process, sets a dynamic threshold based on the prediction residual, identifies high-risk battery cells with abnormal performance in real time, and outputs dynamic grouping instructions. The refined operation restriction and group management module, upon receiving a dynamic grouping instruction, implements operation restriction measures, including load transfer, rate limit setting, charge / discharge depth control, and local temperature management, to suppress the accumulation of memory effect and maintain the stability of the grouped battery capacity. The result write-back and adaptive update module, after completing the operation constraints, writes the operation results back to the sensitive baseline, updates the classification threshold, sampling period and pulse amplitude according to the residual change trend, and generates risk stratification data corresponding to different load levels; The high-rate pre-inspection and group reconstruction module automatically starts the pre-inspection process before high-rate discharge is triggered after generating risk stratification data. It calls the risk stratification data and simulates the same level of operating conditions to check the group stability. When signs of capacity collapse are found, group reconstruction is immediately performed, forming a closed-loop process from baseline construction, degradation identification, threshold embedding, limit control, threshold update to pre-inspection verification.

[0051] The battery grouping online capacity verification method based on the integrated health evolution model provided in this embodiment of the invention is implemented through the aforementioned battery grouping online capacity verification system based on the integrated health evolution model. For details of the specific methods and processes of the battery grouping online capacity verification system based on the integrated health evolution model, please refer to the embodiments of the battery grouping online capacity verification method based on the integrated health evolution model, which will not be repeated here.

[0052] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A battery grouping online capacity verification method integrating a health evolution model, characterized in that, Includes the following steps: S1, at the end of each charge-discharge cycle, establishes a sensitive baseline for micro-cycle conditions, injects controlled short pulses, measures the voltage hysteresis area and plateau repetition rate, and generates a memory warning index; S2, after obtaining the memory warning index, performs an induced test, and forms a graded threshold by comparing the capacity hysteresis residual, internal resistance transition rate and open circuit voltage deviation during the short-term charge and discharge process; S3 embeds the grading threshold into the online capacity calculation process, sets a dynamic threshold based on the prediction residual, identifies high-risk battery cells with abnormal performance in real time, and outputs dynamic grouping instructions. S4, upon receiving the dynamic grouping instruction, implements operational limiting measures, including load transfer, rate limit setting, charge / discharge depth control, and local temperature management, to suppress memory effect accumulation and maintain stable grouped battery capacity; S5, after completing the operation limits, write the operation results back to the sensitive baseline, update the classification threshold, sampling period and pulse amplitude according to the residual change trend, and generate risk stratification data corresponding to different load levels; S6 automatically initiates a pre-inspection process after generating risk stratification data and before high-rate discharge is triggered. It calls the risk stratification data and simulates the same level of operating conditions to check the stability of the groups. When signs of capacity collapse are found, it immediately performs group reconstruction, forming a closed-loop process from baseline construction, degradation identification, threshold embedding, limitation control, threshold update to pre-inspection verification.

2. The online capacity verification method for battery grouping based on the integrated health evolution model according to claim 1, characterized in that, Step S1 includes: After completing one charge-discharge cycle, a high-precision data acquisition device is used to record the charging current, discharging current, charging voltage, discharging voltage, ambient temperature, battery casing temperature, estimated state of charge, and energy input and output of the target battery cell. The sampling period is no less than five times per second, and all raw data are stored in real time. Using the obtained energy input and output data, the charging capacity and discharging capacity of this cycle are calculated respectively, and compared with the previous cycle and historical data item by item to form sensitive baselines for maximum charging capacity, maximum discharging capacity, charging and discharging current variation range, temperature variation range, state of charge variation trend and terminal voltage variation range. After the sensitive baseline is established, a current pulse with an amplitude of 5% to 10% of the rated capacity is applied to the target battery cell. The pulse lasts for ten seconds. During the injection, the changes in battery terminal voltage and current are collected. The sampling frequency is no less than one hundred times per second to ensure the integrity of the voltage dynamic response data. Using high-resolution terminal voltage data obtained from short-pulse excitation, hysteresis curves of current input / output and voltage change are plotted, the hysteresis area is calculated, and the number of voltage plateau segments is counted. Combined with the parameters of the current cycle and historical sensitive baselines, a memory warning index is generated, and this memory warning index is used for the health grading of subsequent processes.

3. The online capacity verification method for battery grouping based on the integrated health evolution model according to claim 1, characterized in that, Step S2 includes: When the memory warning index is obtained and it is initially judged that the target battery cell has a memory effect accumulation trend, a short-term excitation charge-discharge experiment is performed on the battery using a precision adjustable current source. After the battery temperature reaches environmental equilibrium, it is first charged with a constant current of 20% of the nominal capacity for five minutes, and then discharged with a constant current of the same amount for five minutes. The charging and discharging current, voltage and temperature are collected in real time, and the sampling frequency is not less than one hundred times per second. After the experiment, the actual charging capacity and the actual discharging capacity were accumulated respectively. The capacity hysteresis residual was obtained by direct comparison, and its variation range and trend in the continuous period were analyzed. At the instants when charging switches to discharging and discharging switches to resting, the terminal voltage change data are collected respectively, the maximum jump value is counted and the equivalent internal resistance is calculated accordingly, and the internal resistance transition rate is further calculated. At the same time, the open circuit voltage is measured 30 minutes after the end of discharging to obtain the open circuit voltage deviation of this cycle. The capacity hysteresis residual, internal resistance transition rate, and open circuit voltage deviation are compared with sensitive baselines and historical data item by item. Based on the deviation magnitude, the data are divided into normal range, warning range, and danger range to form a classification threshold. Finally, the health level and risk classification of the battery cell are determined.

4. The online capacity verification method for battery grouping based on the integrated health evolution model according to claim 1, characterized in that, Step S3 includes: During each actual charge and discharge cycle, a high-precision acquisition device is used to collect the charging capacity, discharging capacity, terminal voltage, terminal current and shell temperature of each battery cell. The sampling frequency is maintained at more than 100 times per second, and the data is compared with sensitive baselines and historical data one by one. Based on the grading thresholds formed in the previous cycle, the capacity hysteresis residual, internal resistance transition rate, and open circuit voltage deviation are compared with the thresholds in detail to determine whether a single cell has crossed the normal, safe, warning, or dangerous range, and high-risk single cells are identified in real time based on the predicted residual dynamic threshold. High-risk cells are removed from the original group and a separate high-risk group is formed. The maximum discharge current is set to 60% of the nominal capacity, the depth of charge and discharge is limited to 30% to 80% of the rated capacity, and strict operating restrictions are implemented with the temperature fluctuation range not exceeding three degrees Celsius. The capacity calculation results of all new groups are verified, and the output energy and consistency changes are analyzed. The grouping results are used as the reference baseline for the next cycle. If abnormal cells are found, they are fed back to the subsequent dynamic threshold setting stage to achieve a continuous closed loop in the grouping adjustment and calculation process.

5. The online capacity verification method for battery grouping based on the integrated health evolution model according to claim 1, characterized in that, Step S4 includes: Upon receiving a dynamic instruction from the high-risk group, the output current of the high-risk group is gradually reduced, while the output current of the healthy group is simultaneously increased until the load of the high-risk group drops to 30%. The healthy group then takes over the remaining load. The entire process is completed within sixty seconds, ensuring that the voltage difference does not exceed 0.2 volts and that the load switching is smooth without any drops. For high-risk groups, a maximum rate limit is set. The maximum discharge current of a single cell shall not exceed 0.5 times the nominal capacity, and the charging current shall not exceed 0.3 times. Checks are conducted every ten minutes, and if any limit is exceeded, the rate is reduced to 0.4 times. Strictly control the depth of charge and discharge for high-risk groups. Set the discharge cutoff voltage to 3.1 volts and the maximum charging cutoff voltage to 4.0 volts. The depth of discharge shall not exceed 60% of the rated capacity and the charging capacity shall not exceed 85%. The sampling interval shall not exceed 30 seconds. The corresponding charging and discharging operation shall be terminated when the remaining capacity reaches the threshold. A thermal sensor is attached to the surface of each high-risk unit's casing, and temperature data is uploaded every ten seconds. When the unit temperature exceeds the ambient temperature by eight degrees Celsius, the fan is activated. If the temperature does not recover, the cooling plate is activated simultaneously until the temperature recovers. If the temperature is abnormal, load reduction, shortening of charge and discharge cycles, and extension of cooling time are implemented to ensure the stability of the high-risk group capacity.

6. The online capacity verification method for battery grouping based on the integrated health evolution model according to claim 1, characterized in that, Step S5 includes: After the high-risk group operation restriction measures are implemented, the actual total charging capacity, total discharging capacity, terminal voltage, discharging current, charging current, surface temperature and dynamic residual changes of each battery cell are summarized, and all raw data are numbered and archived, corresponding one-to-one with the data of the previous cycle and the sensitive baseline. The difference between the collected capacity retention, capacity hysteresis residual, internal resistance transition rate and open circuit voltage deviation and the sensitive baseline is calculated. If the parameter decreases, the new data point is written to the sensitive baseline. If the parameter fluctuates or deteriorates, a high-risk mark is made on the sensitive baseline, and the write-back is performed immediately after the data is synchronized. Based on the residual change trend of this period, the classification threshold, sampling period and excitation pulse amplitude are dynamically updated. When the residual decreases, the threshold and sampling period are relaxed, and when the residual increases, the threshold and period are tightened and the pulse excitation amplitude is adjusted. All parameters are updated and then incorporated into the health assessment process of the next period. By combining the new sensitive baseline and classification threshold, all battery cells and groups are risk-stratified and divided into three categories: high risk, medium risk and low risk. The stratification results are associated with all collected data and uploaded to the battery management center for use in grouping strategies and load allocation, ensuring a closed loop of monitoring and management.

7. The online capacity verification method for battery grouping based on the integrated health evolution model according to claim 1, characterized in that, Step S6 includes: Before planning high-rate discharge, based on the risk stratification results of the previous cycle, the capacity retention, capacity hysteresis residual, internal resistance transition rate, open circuit voltage deviation, temperature rise amplitude and grouping status of each battery cell are called up one by one to establish a pre-inspection data table and organize it into groups. Based on the pre-detection data, the operating conditions of each group were simulated under high-rate discharge conditions. The discharge current of the group was gradually increased, and the terminal voltage drop rate, discharge capacity, capacity hysteresis residual and surface temperature rise rate were recorded. The remaining capacity, internal resistance transition rate, temperature rise and open-circuit voltage recovery of all cells were monitored. Based on the simulation data, risk checks are performed on each group. If a single cell’s capacity decreases by more than 1% every five minutes, its internal resistance jump rate increases by more than 4%, its terminal voltage is lower than three volts, or its temperature rises by more than forty-five degrees Celsius, then the high-risk cell is removed from the main load group and regrouped according to its health level. The new grouping, individual unit simulation parameters, remaining capacity, temperature, internal resistance changes, and all key indicators are incorporated into the new cycle baseline database. All grouping optimization, capacity calculation, risk assessment, and pre-inspection processes are based on this database to achieve optimal grouping configuration and closed-loop management throughout the entire process.

8. A battery grouping online capacity assessment system integrating a health evolution model, used to implement the battery grouping online capacity assessment method integrating a health evolution model as described in any one of claims 1-7, characterized in that, This includes modules for generating sensitive baselines and memory warnings for micro-cycles, identifying multi-parameter hierarchical thresholds, calculating online capacity and dynamically grouping data, managing refined operational limits and grouping data, writing back operational results and adaptive updating data, and performing high-rate pre-detection and grouping reconstruction. The micro-cycle sensitive baseline and memory warning generation module establishes a sensitive baseline for micro-cycle conditions at the end of each charge-discharge cycle, injects controlled short pulses, measures the voltage hysteresis area and plateau repetition rate, and generates a memory warning index. The multi-parameter graded threshold identification module, after acquiring the memory warning index, performs an induced test and forms a graded threshold by comparing the capacity hysteresis residual, internal resistance transition rate and open circuit voltage deviation during the short-term charge and discharge process. The online capacity calculation and dynamic grouping module embeds the grading threshold into the online capacity calculation process, sets a dynamic threshold based on the prediction residual, identifies high-risk battery cells with abnormal performance in real time, and outputs dynamic grouping instructions. The refined operation restriction and group management module, upon receiving a dynamic grouping instruction, implements operation restriction measures, including load transfer, rate limit setting, charge / discharge depth control, and local temperature management, to suppress the accumulation of memory effect and maintain the stability of the grouped battery capacity. The result write-back and adaptive update module, after completing the operation constraints, writes the operation results back to the sensitive baseline, updates the classification threshold, sampling period and pulse amplitude according to the residual change trend, and generates risk stratification data corresponding to different load levels; The high-rate pre-inspection and group reconstruction module automatically starts the pre-inspection process before high-rate discharge is triggered after generating risk stratification data. It calls the risk stratification data and simulates the same level of operating conditions to check the group stability. When signs of capacity collapse are found, group reconstruction is immediately performed, forming a closed-loop process from baseline construction, degradation identification, threshold embedding, limit control, threshold update to pre-inspection verification.

Images