A method and system for battery pack based single cell inspection

Abstract

The application discloses a single battery inspection method and system based on a battery pack, relates to the technical field of battery inspection, and solves the technical problem that the prior art does not consider the differentiated characteristics of batteries in different operating states such as charging, discharging and standby, that the whole battery pack is monitored and analyzed, and that the abnormality of a single battery cannot be identified in time, and that the service life of the battery pack is reduced; real-time data of the battery pack and a plurality of single batteries are acquired; the operating state of the battery is analyzed based on the real-time data of the battery pack; an evaluation coefficient of the single battery is analyzed according to the operating state and the real-time data of the plurality of single batteries; the inspection cycle of the plurality of single batteries is analyzed based on the evaluation coefficient; and the behavior difference of the battery under different working conditions is fully considered, so that the evaluation result is more scientific and accurate.

Description

Technical Field

[0001] This application belongs to the field of battery inspection and relates to single-cell battery inspection technology for storage battery packs, specifically a method and system for single-cell battery inspection based on storage battery packs. Background Technology

[0002] A battery bank is a DC power system composed of two or more individual batteries connected together in a specific way. The core value of conducting individual cell inspections of a battery bank lies in eliminating potential problems at an early stage. Due to differences in manufacturing, usage environment, and aging, the performance of individual cells within a battery bank gradually diverges, resulting in lagging cells. If only overall voltage and current are monitored, some cells with high internal resistance and insufficient capacity are often masked until the overall capacity drops sharply or even a sudden failure occurs. Individual cell inspections can accurately pinpoint the voltage, internal resistance, and temperature status of each cell, promptly identifying lagging cells and preventing them from dragging down the performance of the entire bank. Simultaneously, by continuously tracking data changes, battery degradation trends can be predicted, enabling a shift from post-failure repair to preventative maintenance, effectively extending the lifespan of the entire bank, reducing the risk of unplanned downtime, and ensuring the reliable operation of the power supply system.

[0003] In existing technologies, individual cell inspections of battery packs are typically based on simple threshold alarms. Each cell is inspected only when the overall battery voltage exceeds a fixed range, thus identifying any abnormal or faulty cells. However, existing technologies do not consider the differentiated characteristics of batteries during charging, discharging, and standby. Monitoring and analyzing the entire battery pack fails to promptly identify anomalies in individual cells, reducing the battery's lifespan.

[0004] This application provides a method and system for inspecting a single battery cell based on a battery pack, in order to solve the above-mentioned technical problems. Summary of the Invention

[0005] This application aims to solve at least one of the technical problems existing in the prior art; to this end, this application proposes a single-cell battery inspection method and system based on a battery pack, to solve the technical problem that the prior art does not consider the differentiated characteristics of batteries in different operating states such as charging, discharging, and standby; and cannot timely identify abnormalities in individual cells when monitoring and analyzing the entire battery pack, thus reducing the service life of the batteries.

[0006] To achieve the above objectives, the first aspect of this application provides a method for inspecting a single cell of a battery pack, comprising: Acquire real-time data of the battery pack and several individual cells; The battery's operating status is analyzed based on real-time data from the battery pack. The evaluation coefficient of a single battery is analyzed based on its operating status and real-time data from several individual batteries. The inspection cycle of several individual batteries was analyzed based on the evaluation coefficient.

[0007] Preferably, the real-time data analysis of the battery's operating status based on the battery pack includes: Retrieve real-time data of the battery pack; the real-time data of the battery pack includes: real-time total current and real-time total voltage; A first-order low-pass filter is used to smooth the real-time data of the battery pack; the average current and voltage change rate are calculated within a set time window based on the processed real-time data of the battery pack. The filtering formula is as follows: ;in, This represents the real-time data of the battery pack sampled for the kth time. This is the value after the k-th filtering. This is the value after the (k-1)th filtering. These are the filter coefficients; The formula for calculating the average current is: The formula for calculating the voltage change rate is: ;in, The length of the window; Let be the real-time total current value at time j; The real-time total voltage value at time k; This represents the real-time total voltage value at kN. The sampling period; The average current and voltage change rate within a set time period are integrated into a state analysis sequence in chronological order; the state analysis model is called, and the state analysis sequence is input into the state analysis function to obtain the operating state of the battery pack; the operating state includes: charging, discharging, and standby; the state analysis model is built based on an artificial intelligence model.

[0008] Preferably, the state analysis model is constructed based on an artificial intelligence model, including: Select a model framework and deep learning algorithm from the artificial intelligence library; construct the model based on the model framework and deep learning algorithm to obtain the initial model; Obtain the standard dataset; the standard dataset includes standard input data consistent with the content attributes of the state analysis sequence, and standard output data consistent with the content attributes of the running state; The standard dataset is divided into training, validation, and test sets according to a set ratio; the initial model is trained using the training set; the internal parameters of the initial model are adjusted using the validation set; and the initial model is tested using the test set to obtain test metrics. Obtain the standard range of indicators; compare the test indicators with the standard range of indicators; if all test indicators are within the standard range of indicators, mark the initial model as the state analysis model; otherwise, rebuild and retrain the initial model.

[0009] Preferably, the step of analyzing the evaluation coefficient of a single battery cell based on its operating status and real-time data from several individual cells includes: Retrieve real-time data of several individual cells and the operating status of the battery pack; the real-time data of individual cells includes: real-time voltage, real-time current, and usage time; When the battery pack is charging, the charging characteristic indicators of individual cells are analyzed, and the charging evaluation coefficient of the battery is calculated based on the charging characteristic indicators. When the battery pack is in a discharging state, the discharge characteristic indicators of individual cells are analyzed, and the discharge evaluation coefficient of the battery is calculated based on the discharge characteristic indicators. When the battery pack is in standby mode, the standby characteristic indicators of individual cells are analyzed, and the standby evaluation coefficient of the battery is calculated based on the standby characteristic indicators. The evaluation coefficients for each operating state are adjusted based on usage time.

[0010] Preferably, the analysis of the charging characteristic indicators of a single battery cell and the calculation of the battery's charging evaluation coefficient based on the charging characteristic indicators include: Retrieve real-time data from several individual cells; calculate the average real-time voltage and standard deviation of several individual cells; calculate the voltage balance deviation of each individual cell based on the voltage deviation analysis function; The expression for the voltage deviation analysis function is: ;in, This represents the real-time voltage of the i-th battery cell; This represents the average value of the real-time voltage. The standard deviation of voltage; It is a very small constant; Based on formula Calculate the voltage change rate deviation for each battery cell; where, Let be the rate of change of voltage in the i-th cell; This represents the average rate of change of voltage within the battery pack. The standard deviation of the rate of change of voltage; Calculate the rate of change of real-time current. When the rate of change of real-time current is less than a set rate of change threshold, then based on the formula... Calculate the relative deviation of the charging internal resistance of each battery cell; where, Let be the internal resistance of the i-th battery during charging. ; This represents the average value of the charging internal resistance; This represents the standard deviation of the charging internal resistance within the battery pack. Based on formula Calculate the charging evaluation factor for each battery cell; where, , and The weighting coefficient is greater than 0.

[0011] Preferably, the analysis of the discharge characteristic indicators of a single battery cell and the calculation of the battery's discharge evaluation coefficient based on the discharge characteristic indicators include: Retrieve real-time data from several individual cells; recalculate the voltage balance deviation of each individual cell based on the voltage deviation analysis function; based on the formula... Recalculate the voltage change rate deviation for each battery cell; Based on formula Recalculate the relative deviation of the charging internal resistance of each battery cell; based on the formula Calculate the discharge evaluation coefficient for each battery cell; where, .

[0012] Preferably, the step of analyzing the standby characteristic indicators of a single battery cell and calculating the battery's standby evaluation coefficient based on the standby characteristic indicators includes: Retrieve real-time data from several individual battery cells; calculate the static voltage deviation of each individual battery cell based on the voltage deviation analysis function; The coefficient of variation of the voltage sampling sequence during the standby period is calculated based on the voltage fluctuation analysis function; the expression of the voltage fluctuation analysis function is: ;in ; Let be the real-time voltage of the i-th battery at time k; This represents the total number of sampling times. Based on formula The coefficient of variation for each battery cell was standardized; among them, This represents the average value of the coefficient of variation within the battery pack. This represents the standard deviation of the coefficient of variation within the battery pack. Based on formula Calculate the standby evaluation coefficient for each battery cell; where, and The weighting coefficient is greater than 0.

[0013] Preferably, the correction of the evaluation coefficient for each operating state based on usage time includes: Retrieve the usage time from the real-time data of each battery; calculate the time tolerance factor based on the tolerance factor analysis function; the expression for the tolerance analysis factor is: ;in, This is the allowable deviation for the new battery; This is the aging rate coefficient; For usage time; It is a logarithmic function with base e; Based on formula The voltage balance deviation in charging and discharging states, as well as the voltage settling deviation in standby state, are corrected; the evaluation coefficient for each operating state is recalculated based on the corrected indicators.

[0014] Preferably, the step of analyzing the inspection cycle of several individual batteries based on evaluation coefficients includes: Retrieve the corrected evaluation coefficients; obtain the inspection cycle library; where the evaluation coefficients include: charging evaluation coefficients, discharging evaluation coefficients, and standby evaluation coefficients; The corrected evaluation coefficients are matched with the inspection cycle library to obtain the inspection cycle corresponding to each operating state. The shortest inspection cycle is selected as the inspection cycle of the corresponding single battery cell.

[0015] The second aspect of this application provides a single-cell battery inspection system based on a battery pack, including: a data acquisition module, a status assessment module, and a cycle determination module; The data acquisition module is used to acquire real-time data of the battery pack and several individual batteries. The status assessment module is used to analyze the operating status of the battery based on real-time data of the battery pack; and to analyze the evaluation coefficient of each individual battery based on the operating status and real-time data of several individual batteries. The cycle determination module is used to analyze the inspection cycle of several individual batteries based on the evaluation coefficient.

[0016] Compared with the prior art, the beneficial effects of this application are: 1. This application smooths the real-time total current and total voltage of the battery pack using a first-order low-pass filter, calculates the average current and voltage change rate within a set time window, constructs a state analysis sequence, and then inputs it into a state analysis model built based on an artificial intelligence model to accurately identify whether the battery pack is currently in a charging, discharging, or standby state. Based on this, corresponding characteristic indicators are extracted for different operating states: in the charging state, voltage equalization deviation, voltage change rate deviation, and relative deviation of charging internal resistance are calculated; in the discharging state, the corresponding deviation indicators are recalculated; and in the standby state, voltage static deviation and voltage fluctuation variation coefficient are analyzed. This state-based evaluation method fully considers the behavioral differences of the battery under different operating conditions, making the evaluation results more scientific and accurate. Simultaneously, the method also introduces a tolerance factor based on usage time to correct the evaluation coefficients, further considering the cumulative effect of battery aging over time, and avoiding evaluation bias caused by different battery usage years.

[0017] 2. This application matches the modified charging evaluation coefficient, discharging evaluation coefficient, and standby evaluation coefficient with an inspection cycle library to determine the corresponding inspection cycle for each battery under each operating state, and selects the shortest cycle as the final inspection cycle for that battery. This means that batteries in poor condition and with higher risk will receive more frequent inspections, while batteries in good condition can have their inspection intervals appropriately extended, thereby significantly reducing the manpower and equipment costs of inspections while ensuring safe battery operation. Furthermore, the state analysis model is trained, validated, and tested based on a standard dataset, ensuring the model's generalization ability and reliability. The introduction of a minimal constant in the voltage deviation analysis function avoids division-to-zero anomalies, and the standardization of the coefficient of variation eliminates dimensional differences. These detailed designs enhance the robustness and engineering practicality of the method. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram illustrating the overall steps of the method described in this application; Figure 2 This is a schematic diagram showing the connection of the system structure modules in this application. Detailed Implementation

[0020] The technical solutions of this application will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0021] Example 1: Please see Figure 1 The first aspect of this application provides a method for inspecting a single cell battery based on a battery pack, including: S100: Acquire real-time data of the battery pack and several individual cells.

[0022] In this embodiment, real-time data acquisition is fundamental to inspection. Data sources can be direct reading from the Battery Management System (BMS) or acquisition via independently deployed voltage, current, and temperature sensors. Battery pack data typically includes macroscopic parameters such as total voltage, total current, and ambient temperature, while individual cell data includes at least the terminal voltage of each cell, and may also include microscopic parameters such as cell temperature or internal resistance, depending on specific needs. It should be understood that the method described in this invention is applicable to various battery packs composed of multiple cells connected in series, including but not limited to lead-acid batteries, lithium-ion batteries, and nickel-cadmium batteries. As long as they are connected in series and require individual cell consistency management, the technical solution of this embodiment can be applied. By simultaneously acquiring data at both the overall pack and individual cell levels, complete information input is provided for subsequent status identification and differentiated assessment.

[0023] S200 analyzes the battery's operating status based on real-time data from the battery pack.

[0024] The core of this step lies in utilizing the electrical characteristics of the entire battery pack to determine the current operating environment. Because the internal electrochemical reaction mechanisms of batteries differ significantly under different states such as charging, discharging, and standby, the abnormal behavior of individual cells also varies drastically. For example, at the end of charging, the voltage may rise too quickly, while during discharging, it may exhibit abnormal voltage drops. Therefore, before evaluating individual cells, the operating state of the entire battery pack must be clearly defined, serving as the trigger condition for selecting the evaluation strategy. This macroscopic-to-microscopic analytical logic effectively avoids the misjudgment problems caused by ignoring the operating condition context in traditional methods.

[0025] S300: Analyzes the evaluation coefficient of a single battery cell based on its operating status and real-time data from several individual cells.

[0026] After determining the operating state, this step uses the state information as a routing factor and combines it with real-time data from individual cells to calculate an evaluation coefficient characterizing the health of each cell. This evaluation coefficient is not a simple threshold comparison of a single physical quantity, but a quantitative indicator that integrates multi-dimensional features. Specifically, the physical quantities involved in the calculation and their weighting relationships are adjusted accordingly when the operating state differs, ensuring that the evaluation results accurately reflect the battery performance under the current operating conditions. For example, in the charging state, more attention may be paid to voltage balance and polarization effects, while in the standby state, the focus may be on self-discharge rate or static voltage stability. In this way, the evaluation model is dynamically adapted to actual electrochemical behavior.

[0027] S400, based on evaluation coefficient analysis, determines the inspection cycle of several individual battery cells.

[0028] This is the decision output stage of this method. The system maps the evaluation coefficient obtained in step S300 to specific inspection time intervals, thus completing the closed loop from "state awareness" to "operation and maintenance execution". The higher the evaluation coefficient, the greater the deviation of the individual battery from its normal state or the higher the risk of degradation, and the shorter the corresponding inspection cycle should be to track its changing trend at a higher frequency; conversely, for individual batteries in good condition, the inspection interval can be appropriately extended to save system resources. This dynamic periodic adjustment mechanism based on risk assessment can significantly improve operation and maintenance efficiency while ensuring safety, compared with traditional fixed-cycle inspections.

[0029] In one possible implementation, the inspection cycle of several individual battery cells is analyzed based on evaluation coefficients, including: Retrieve the corrected evaluation coefficients; obtain the inspection cycle library; where the evaluation coefficients include: charging evaluation coefficients, discharging evaluation coefficients, and standby evaluation coefficients; The corrected evaluation coefficients are matched with the inspection cycle library to obtain the inspection cycle corresponding to each operating state. The shortest inspection cycle is selected as the inspection cycle of the corresponding single battery cell.

[0030] In summary, this embodiment establishes a data-driven adaptive inspection framework through the orderly execution of the four steps described above. This framework not only clarifies the data flow logic between each step—namely, real-time data-driven state analysis, state-triggered differentiated evaluation, and evaluation results determining the inspection strategy—but also reserves ample room for expansion in the implementation of subsequent specific algorithms through the definition of higher-level concepts.

[0031] Example 2: In this embodiment, the process of analyzing the battery's operating status based on real-time data from the battery pack in step S200 of Embodiment 1 is specifically defined. Analyzing the battery's operating status based on real-time data from the battery pack includes: retrieving real-time data from the battery pack; wherein, the real-time data from the battery pack includes: real-time total current and real-time total voltage. Specifically, the real-time total current and real-time total voltage are the most fundamental physical quantities characterizing the overall energy flow direction and intensity of the battery pack. Their sampling frequency should satisfy the Nyquist sampling theorem, and is typically set between 1Hz and 10Hz to balance data volume and dynamic capture capability.

[0032] A first-order low-pass filter is used to smooth the real-time data of the battery pack. Based on the processed real-time data, the average current and voltage change rate are calculated within a set time window. Because battery packs are often affected by electromagnetic interference, contact resistance fluctuations, or sudden load changes in actual operating environments, the original acquired signals are often superimposed with high-frequency noise. Directly using this for status judgment can easily lead to false triggering. This embodiment uses a first-order low-pass filter because it effectively suppresses high-frequency noise while preserving the DC component and low-frequency trend of the signal, and its low computational complexity makes it suitable for real-time computation in embedded systems. Compared to higher-order filters, the first-order low-pass filter does not suffer from waveform distortion caused by excessive phase lag, achieving an optimal balance between response speed and noise suppression.

[0033] The filtering formula is: ;in, This represents the real-time data of the battery pack sampled for the kth time. This is the value after the k-th filtering. This is the value after the (k-1)th filtering. These are the filter coefficients.

[0034] In practical implementation, the filter coefficients The optimal value range is 0.05 to 0.2. If the value is too small, the filtering effect is good, but the system response is sluggish, potentially missing rapidly switching conditions; if the value is too large, the noise suppression capability decreases. For example, in a data center UPS scenario, the value can be... Setting it to 0.1 can filter out millisecond-level spike interference and track the real changes in load current within seconds.

[0035] The formula for calculating the average current is: ; The formula for calculating the rate of change of voltage is: ; in, The length of the window; Let be the real-time total current value at time j; The real-time total voltage value at time k; This represents the real-time total voltage value at kN. The sampling period.

[0036] A sliding window is introduced to calculate the average current and voltage change rate, aiming to further eliminate the randomness of single-point data. The window length should be matched with the battery's time constant; for example, for lead-acid batteries, 30 to 60 sampling points (corresponding to a time span of 30 seconds to 1 minute) can be used to cover the main process of battery polarization establishment. The voltage change rate is calculated using a differential approximation method, which has better noise immunity than direct differentiation and can accurately reflect the evolution trend of battery terminal voltage over time. This is a key criterion for distinguishing between charging (voltage rise), discharging (voltage fall), and standby (voltage stability).

[0037] The average current and voltage change rate over a set time period are integrated into a state analysis sequence in chronological order. The state analysis model is then invoked, and the state analysis sequence is input into the state analysis function to obtain the operating state of the battery pack. The operating states include charging, discharging, and standby. The state analysis model is built based on an artificial intelligence model. It is important to emphasize that the "state analysis sequence" is not an abstract data set, but rather a time-series feature vector carrying specific physical meaning. This sequence transforms discrete electrical measurements into spatiotemporally correlated pattern expressions, enabling the artificial intelligence model to learn the evolution of current and voltage combinations under different operating conditions, rather than simply memorizing static thresholds. This design ensures that state recognition is based on an objective mapping of electrochemical behavior, rather than subjectively set rules, thus meeting the requirement of fully disclosed technical solutions.

[0038] Furthermore, the state analysis model is built upon an artificial intelligence model, including: selecting a model framework and deep learning algorithm from an AI library; and constructing the model based on the model framework and deep learning algorithm to obtain an initial model. For example, LSTM (Long Short-Term Memory) or GRU (Gated Recurrent Unit) can be chosen as the basic framework, as these types of recurrent neural networks are naturally suitable for handling the aforementioned time-dependent state analysis sequences. Of course, in other implementations, a Transformer architecture or a one-dimensional convolutional neural network (1D-CNN) can also be used, as long as it has the ability to extract temporal features.

[0039] Obtain a standard dataset; this dataset includes standard input data consistent with the content attributes of the state analysis sequence, and standard output data consistent with the content attributes of the running state. The standard dataset is the cornerstone of model training and can originate from laboratory calibration tests, historical operation and maintenance records, or simulation-generated data. Crucially, the standard input data must be a feature sequence after the aforementioned filtering and windowing calculations, not the raw voltage and current values, to ensure consistency between the training and inference environments; the standard output data consists of the corresponding real-world state labels (e.g., 0 for standby, 1 for charging, and 2 for discharging).

[0040] The standard dataset is divided into training, validation, and test sets according to a predetermined ratio. The initial model is trained using the training set; the internal parameters of the initial model are adjusted using the validation set; and the initial model is tested using the test set to obtain test metrics. Typical ratios are 7:2:1 or 8:1:1. The training set is used to fit the mapping relationship between features and states; the validation set is used to monitor overfitting and adjust hyperparameters (such as learning rate and number of hidden layer nodes) during training; the test set is only used in the final evaluation to objectively reflect the model's generalization ability. Test metrics include, but are not limited to, accuracy, precision, recall, and F1 score, which together constitute a quantitative scale for measuring model performance.

[0041] The process involves: obtaining the standard range of performance indicators; comparing the test indicators with the standard range; marking the initial model as a state analysis model if all test indicators fall within the standard range; otherwise, retraining the initial model. The standard range of performance indicators is a pre-set acceptance threshold based on the security requirements of the actual application scenario, such as requiring an accuracy of at least 98% and a recall of at least 95%. Only when the model meets all key performance indicators is it allowed to be deployed to the inspection system. This rigorous verification mechanism ensures that the online state analysis model has reliable engineering applicability and avoids subsequent evaluation errors due to model defects. Through the above complete construction and verification process, this invention deeply integrates artificial intelligence technology with battery physical characteristics, forming an interpretable, verifiable, and highly accurate operating state identification scheme.

[0042] Example 3: In this embodiment, the process of analyzing the evaluation coefficient of a single battery cell based on the operating status and real-time data of several individual batteries in step S300 of embodiment 1 is specifically defined. Analyzing the evaluation coefficient of a single battery cell based on the operating status and real-time data of several individual batteries includes: retrieving real-time data of several individual batteries and the operating status of the battery pack; wherein, the real-time data of a single battery cell includes: real-time voltage, real-time current, and usage time. Specifically, the data retrieval action here should be strictly synchronized with the overall data acquisition in embodiment 2 in terms of timing to ensure that the operating condition background on which the individual cell evaluation is based is accurate. In addition to conventional electrical quantities, this embodiment specifically introduces the dimension of "usage time". Usage time refers to the cumulative effective working time of a single battery cell since it has been put into operation or last reset. It is not only a timestamp but also a fundamental physical quantity characterizing the degree of battery aging. In engineering implementation, this parameter can be directly provided by the internal timer of the battery management system or calculated from the historical charge-discharge cycle count. The purpose of introducing usage time is to provide a benchmark for subsequent evaluation coefficient correction, because the performance tolerance of batteries will naturally increase over time, and ignoring this factor will lead to a significant increase in the misjudgment rate of old batteries.

[0043] When the battery pack is charging, the charging characteristic indicators of individual cells are analyzed, and the charging evaluation coefficient of the battery is calculated based on these indicators. When the battery pack is discharging, the discharging characteristic indicators of individual cells are analyzed, and the discharging evaluation coefficient of the battery is calculated based on these indicators. When the battery pack is in standby mode, the standby characteristic indicators of individual cells are analyzed, and the standby evaluation coefficient of the battery is calculated based on these indicators. These three state-specific processing mechanisms constitute the core of the differentiated evaluation system of this invention. State-specific evaluation is necessary because the electrochemical behavior of batteries differs fundamentally under different operating conditions, and a single evaluation model cannot cover all anomalies in all scenarios. For example, during charging, the battery primarily undergoes redox reactions and ion insertion processes. At this stage, voltage differences between cells primarily reflect capacity inconsistencies or the strength of polarization effects. Therefore, charging characteristic indicators focus on voltage balance and the relative deviation of charging internal resistance. During discharging, the battery outputs energy, and lagging cells often exhibit excessively rapid voltage drops or abnormally high internal resistance. If the charging state evaluation criteria are applied in this case, it's easy to miss the problem of reduced load-carrying capacity due to increased internal resistance. As for standby, although the external current is zero, internal micro-short circuits or accelerated self-discharge may still exist. These anomalies are masked by the load current under dynamic operating conditions and can only be effectively detected under static conditions through voltage fluctuation rate or coefficient of variation. Through this state-driven routing mechanism, the system can automatically switch to the evaluation algorithm best suited to the current physical process, thereby reducing the risk of false alarms while maintaining detection sensitivity. It should be understood that although this embodiment lists three typical states: charging, discharging, and standby, in other embodiments, if the application scenario includes a special operation and maintenance mode (such as active balancing maintenance or pulse testing), the corresponding evaluation branch can also be extended on this basis. As long as it follows the logical framework of "state recognition - feature matching - coefficient calculation", it falls within the protection scope of this invention.

[0044] The evaluation coefficients for each operating state are corrected based on usage time. This step serves as a bridge between the static evaluation model and dynamic lifecycle management. Regardless of the battery's operating state, its original evaluation coefficients reflect the deviation from the ideal state at the current moment. However, for a battery that has been in service for many years, a certain degree of performance degradation is a normal result of physical aging, not an inevitable symptom of failure. If the evaluation coefficients are not corrected for the time dimension, the system will generate more and more false alarms as the battery's service life increases, leading maintenance personnel to lose trust in the system and even disable the alarm function. Therefore, this embodiment uses usage time as a global correction factor applied to the evaluation coefficients for each state. This correction is not a simple linear compensation, but rather simulates the nonlinear law of the tolerance boundary gradually widening during battery aging, allowing the evaluation criteria to adaptively adjust as the battery's lifecycle evolves. In this way, it retains a keen awareness of sudden and accelerated degradation while avoiding excessive intervention in the normal aging process, thus achieving a balance between the scientific nature and economic efficiency of the inspection strategy throughout the entire lifecycle. Subsequent embodiments will further detail the specific characteristic index calculation formulas for each state and the correction function expression based on usage time.

[0045] Example 4: In this embodiment, the process of analyzing the charging characteristic indicators of individual cells and calculating the charging evaluation coefficient in Example 3 when the battery pack is in a charging state is specifically defined. Analyzing the charging characteristic indicators of individual cells and calculating the charging evaluation coefficient based on these indicators includes: retrieving real-time data from several individual cells; calculating the average real-time voltage and voltage standard deviation of several individual cells; and calculating the voltage balance deviation of each individual cell based on a voltage deviation analysis function. Specifically, under charging conditions, the dispersion of individual cell voltage is the most intuitive indicator reflecting the consistency of the battery pack. However, directly using the voltage difference is easily affected by the voltage fluctuations of the entire pack. Therefore, this embodiment adopts a statistical standardization method, dividing the absolute deviation of the individual cell voltage from the mean by the voltage standard deviation to obtain a dimensionless relative deviation. This processing method can effectively eliminate common voltage drift caused by overall changes in charging current, allowing the evaluation results to focus only on the differences between individual cells.

[0046] The expression for the voltage deviation analysis function is: ;in, This represents the real-time voltage of the i-th battery cell; This represents the average value of the real-time voltage. The standard deviation of voltage; It is a minimal constant In this formula, a minimum constant is introduced into the denominator. This has significant engineering defense implications. In practical calculations, when the battery pack is in a highly uniform ideal state or when sampling accuracy is limited, the voltage standard deviation... It may approach zero or even become zero due to numerical truncation. If no addition... Division operations will result in division by zero exceptions or numerical overflow, which will cause the inspection program to crash. The value is usually set to 10. -6 Up to 10 -9 The magnitude ensures mathematical continuity without substantially affecting the calculation results within the normal range, thus guaranteeing the robustness of the algorithm in embedded devices.

[0047] Based on formula Calculate the voltage change rate deviation for each battery cell; where, Let be the rate of change of voltage in the i-th cell; This represents the average rate of change of voltage within the battery pack. The standard deviation of the voltage change rate is denoted as . The voltage change rate deviation reflects the difference in polarization establishment rate of individual cells during charging. Towards the end of charging, underperforming cells often exhibit abnormal voltage rise rates (too fast or too slow) due to difficulties in active material conversion or increased internal resistance. By capturing this dynamic response characteristic, this embodiment can identify potentially degraded cells before significant differentiation in static voltage, extending the assessment from "static result evaluation" to "dynamic process evaluation."

[0048] Calculate the rate of change of real-time current. When the rate of change of real-time current is less than a set rate of change threshold, then based on the formula... Calculate the relative deviation of the charging internal resistance of each battery cell; where, Let be the internal resistance of the i-th battery during charging. ; This represents the average value of the charging internal resistance; This represents the standard deviation of the internal resistance during charging within the battery pack. It's important to note here that the formula... It is not the absolute DC internal resistance in the traditional sense, but rather a "differential internal resistance" or "relative internal resistance" characterizing the deviation of a single cell from the group average level. Its molecule... This approach eliminates the influence of common-mode factors such as ambient temperature and SOC reference on the terminal voltage, ensuring that the calculation results purely reflect the impedance characteristics of the individual cell relative to the group. Furthermore, this embodiment sets a current change rate threshold as a calculation criterion. This is a prerequisite because internal resistance estimation relies on the quasi-steady-state assumption of Ohm's law. During transient processes with drastic fluctuations in charging current, the electrochemical polarization and concentration polarization inside the battery have not yet reached equilibrium, and the line inductance will generate an induced electromotive force. At this time, the "internal resistance" obtained by directly dividing the terminal voltage difference by the current contains a large number of spurious components and cannot truly characterize the battery's health state. Only when the rate of change of current is below a threshold (e.g., 0.05 A / s) does the system determine that it has entered the quasi-steady-state region and initiate internal resistance deviation calculation, thereby avoiding misjudgments caused by dynamic disturbances.

[0049] Based on formula Calculate the charging evaluation factor for each battery cell; where, , and Weighting coefficients are greater than 0. This formula uses the Euclidean norm to integrate the deviation indicators of the three dimensions into a comprehensive score. Compared to a simple weighted summation, the square root method is more sensitive to extreme outliers; that is, if any dimension shows a significant deviation, the final evaluation coefficient will increase significantly. This aligns with the risk management logic of the "weakest link effect" in battery safety monitoring. Weighting coefficients , and It can be flexibly configured according to the specific battery chemistry system and application scenario. For example, for valve-regulated lead-acid batteries, since the voltage difference at the end of charging is most sensitive to changes in internal resistance, it can be... Set it to a larger value (e.g., 0.5), and , Each is set to 0.25; however, for lithium-ion batteries, if more attention is paid to the voltage plateau differences caused by capacity consistency, the value can be appropriately increased. The proportion of [something]. This parameterized design allows the same algorithm framework to be adapted to different types of battery packs, enhancing the versatility and practicality of the technical solution.

[0050] Example 5: In this embodiment, the process of analyzing the discharge characteristic indicators of individual cells and calculating the discharge evaluation coefficient when the battery pack is in a discharging state, as described in Example 3, is specifically defined. Analyzing the discharge characteristic indicators of individual cells and calculating the discharge evaluation coefficient based on these indicators includes: retrieving real-time data from several individual cells; recalculating the voltage equalization deviation of each individual cell based on a voltage deviation analysis function; and applying the formula... Recalculate the voltage change rate deviation for each battery cell. Specifically, although the mathematical expressions of the two formulas above are completely consistent with the charging state calculation formula in Example 4, the necessity of "recalculation" must be emphasized in this example. This is because after the battery switches from the charging state to the discharging state, the direction of its internal electrochemical reaction reverses, and the dynamic response characteristics of the terminal voltage also change accordingly. For example, a cell that performs normally at the end of charging may experience a voltage drop first at the beginning of discharging due to its high internal resistance. Therefore, the system cannot directly reuse the charging evaluation data from the previous moment, but must recalculate based on the real-time sampled value at the current discharging moment, substituting it into the formula to ensure that the evaluation result can truly reflect the consistency level of the cells under the discharging condition. In addition, although the formula structure is the same, its physical meaning has changed: the voltage equalization deviation at this time This characterizes the voltage drop dispersion during the discharge process, while the voltage change rate deviation... This reflects the difference in dynamic response speed of each cell under load impact, and these two together constitute the basic dimensions for evaluating the load-carrying capacity stability of the battery pack.

[0051] Based on formula Recalculate the relative deviation of the charging internal resistance of each battery cell; based on the formula Calculate the discharge evaluation coefficient for each battery cell; where, It is important to note here that although this step logically follows the calculation framework for the relative deviation of internal resistance, the core variables are adjusted specifically for the discharge conditions. The definition has been critically adjusted, which is the most fundamental technical difference between this embodiment and Embodiment 4. Firstly, regarding the polarity reversal of the molecule term, i.e., the polarity reversal during charging... When it becomes a discharge This is determined by the electrochemical discharge characteristics of the battery. During discharge, the terminal voltage of a single cell with poor performance or increased internal resistance will decrease. It will be lower than the average voltage of the entire group. ,lead to It is a positive value. If the charging formula is incorrectly applied... The calculation result will be negative, which not only violates the non-negativity property of resistance as a physical scalar, but may also mask abnormal directional information in subsequent absolute or square calculations, leading to evaluation distortion. By reversing polarity, it is ensured that in discharge scenarios, the more backward the cell, the higher its calculated differential internal resistance. The larger the deviation, the better it aligns with the assessment logic that "the greater the deviation, the worse the health."

[0052] Secondly, regarding the use of the absolute value of current in the denominator term. The design primarily aims to adapt to the different definitions of current direction used by various battery management systems (BMSs). In engineering practice, some BMSs define the discharge current as a negative value to distinguish the energy flow direction, while others uniformly define it as a positive value. Regardless of whether the original real-time current I is positive or negative, the internal resistance calculation must be based on the magnitude of the current, not its direction. Introducing absolute value operations eliminates the interference of the current sign on the internal resistance estimation results, ensuring the algorithm's universality and robustness across different hardware platforms. Finally, the discharge evaluation coefficient... Although the calculation structure is similar to the charging evaluation coefficient To maintain consistency, both methods employ weighted Euclidean norm fusion of multidimensional biases, but their input parameters have been entirely replaced with feature values ​​adapted to the discharge scenario. This means... Instead of being a metric for measuring battery charging acceptance, it has been transformed into a comprehensive score characterizing the battery's ability to maintain voltage stability and consistent energy output under load conditions. This "isomorphic but heterogeneous parameter" design strategy reduces the complexity of the system algorithm implementation and achieves accurate coverage of risks across all operating conditions by reconstructing the physical meaning of the parameters, effectively preventing missed detections or false alarms caused by applying a single model across different operating conditions.

[0053] Example 6: In this embodiment, the process of analyzing the standby characteristic indicators of individual cells and calculating the standby evaluation coefficient when the battery pack is in standby mode, as described in Embodiment 3, is specifically defined. Analyzing the standby characteristic indicators of individual cells and calculating the standby evaluation coefficient based on these indicators includes: retrieving real-time data from several individual cells; and calculating the voltage settling deviation of each individual cell based on the voltage deviation analysis function. Specifically, although the voltage deviation analysis function SV_i from Embodiments 4 and 5 is reused here, the physical meaning of this indicator undergoes a fundamental change under standby conditions. During the dynamic charging and discharging process, SV_i mainly reflects the polarization voltage difference or capacity inconsistency caused by current flow; however, in the standby state, the external current is zero, and the ohmic voltage drop disappears. At this time, the difference in cell voltage mainly stems from the difference in self-discharge rate or the presence of internal micro-short circuits. Therefore, this embodiment redefines SV_i as "voltage settling deviation" to characterize the deviation of the battery from its thermodynamic equilibrium state under no-load conditions. This strategy of reusing the same mathematical tools but giving them different physical interpretations simplifies the complexity of algorithm implementation and ensures that the most representative discrete voltage features can be extracted under different operating conditions.

[0054] The coefficient of variation of the voltage sampling sequence during the standby period is calculated based on the voltage fluctuation analysis function; the expression of the voltage fluctuation analysis function is: ; in ; Let be the real-time voltage of the i-th battery at time k; The total number of sampling times; coefficient of variation This embodiment introduces a key indicator to address the unique blind spot in static degradation detection during standby. Because the absolute voltage change during standby is extremely small (potentially only in the millivolt or even microvolt range), directly using the standard deviation is easily affected by the voltage reference level, making it difficult to compare the stability of individual cells under different SOCs. The coefficient of variation (CV) constructs a dimensionless measure of relative fluctuation by normalizing the standard deviation to the percentage of the mean. This allows the system to sensitively detect abnormal fluctuations that, while having small absolute amplitudes, are significant relative to their own reference, such as intermittent micro-short circuits caused by micropore damage in the separator, or localized instability caused by electrolyte drying. Compared to simple voltage threshold judgment, the CV can reveal the evolution trend of hidden faults inside the battery from a time-series statistical perspective, effectively filling the static monitoring gap that cannot be covered by dynamic operating conditions.

[0055] Based on formula The coefficient of variation for each battery cell was standardized; among them, This represents the average value of the coefficient of variation within the battery pack. This represents the standard deviation of the coefficient of variation within the battery pack. The defensive consideration behind this step is to eliminate environmental common-mode interference. In actual computer rooms or energy storage power station environments, periodic fluctuations in ambient temperature or temperature drift in the BMS acquisition circuitry can cause the voltage fluctuation rate of the entire battery pack to rise or fall synchronously. If only the original coefficient of variation is considered... During assessment, it is highly susceptible to misinterpreting normal fluctuations across the entire group as individual abnormalities when environmental temperatures change drastically. This can be addressed by subtracting the group mean. Divide by the within-group standard deviation Standardized indicators The common drift components of the group are removed, and only the specific fluctuation information of the individual cell relative to the average level of the group is retained. Only when the fluctuation of a certain cell deviates significantly from the statistical law of the group will its standardized value show a high amplitude, thereby greatly reducing the false alarm rate caused by external environmental factors and improving the robustness of the inspection system in complex field environments.

[0056] Based on formula Calculate the standby evaluation coefficient for each battery cell; where, and The weighting coefficients are greater than 0. Unlike the evaluation models in Examples 4 and 5, which include three dimensions—voltage, rate of change, and internal resistance—the standby evaluation coefficients in this example are different. Only voltage settling deviation was incorporated. and standardized coefficient of variation Two metrics. This structural simplification is not a functional deficiency, but a necessary choice based on the standby electrochemical characteristics. Since the current is constant at zero in standby mode, the internal resistance parameter loses its physical definition and cannot be calculated; forcibly introducing it would only introduce noise. Simultaneously, the voltage change rate also approaches zero after a long period of rest, losing its discriminative power. Therefore, the evaluation focus shifts entirely to the absolute position of the static voltage. With time stability Above. Regarding weight configuration, and The settings should reflect an emphasis on static stability. For example, for lead-acid battery packs used for long-term float charging standby, the risk of micro-short circuits often appears before a significant voltage drop. In this case, the settings can be adjusted accordingly. Set a higher value (e.g., 0.6 to 0.7) to prioritize responses to abnormal fluctuations; however, for lithium battery packs that are frequently used and where the consistency of remaining capacity is of greater concern, the value can be appropriately increased. The proportion of [something]. Through this two-dimensional evaluation architecture adapted to static operating conditions, the present invention achieves seamless monitoring coverage of the entire life cycle and all operating states of the battery.

[0057] Example 7: In this embodiment, the process of correcting the evaluation coefficients for each operating state based on usage time in Embodiment 3 is specifically defined. Correcting the evaluation coefficients for each operating state based on usage time includes: retrieving the usage time from the real-time data of each battery cell; and calculating the time tolerance factor based on the tolerance factor analysis function. Specifically, usage time... This is a key dimension characterizing the cumulative effect of battery aging, and its data source should be consistent with the aforementioned embodiments. Tolerance factor It is not a simple scaling factor, but a dynamic threshold regulator that simulates the evolution of battery performance tolerance throughout its entire life cycle. In engineering implementation, this factor is calculated independently of the real-time electrical sampling period and is typically updated on an hourly or daily basis to reduce the computational load on the processor.

[0058] The expression for the tolerance analysis factor is: ;in, This is the allowable deviation for the new battery; This is the aging rate coefficient; For usage time; It is a logarithmic function with base e. The reason why this embodiment uses a logarithmic function form instead of a nonlinear function is based on the electrochemical mechanism of battery aging. In the initial stage of battery operation ( Initial inconsistencies (relatively small) due to differences in manufacturing processes quickly become apparent, and early aging processes such as SEI film formation are rapid. Therefore, the allowable deviation range should increase rapidly over time. However, as the battery enters its stable service life and later degradation phase, the performance degradation rate slows down, and the expansion of the tolerance boundary should also decelerate. The "steep at first, then gradual" characteristic of the logarithmic function perfectly matches this physical law. In contrast, if linear correction is used, either insufficient correction in the initial stage leads to frequent false alarms, or excessive correction in the later stage increases the risk of missed detections. Parameters This represents the inherent dispersion benchmark of a new battery at the time of manufacture, usually obtained from the packing tolerances provided by the manufacturer or statistical analysis of measured data; for example, 0.005 can be used for a 2V lead-acid battery. (Parameter) This reflects the degree of accelerated aging under specific application scenarios, and can be calibrated by fitting historical operation and maintenance data, such as in high-temperature or high-rate charging and discharging scenarios. The value should be increased accordingly to accommodate the need for faster tolerance expansion.

[0059] Based on formula The voltage equalization deviation during charging and discharging, as well as the voltage settling deviation during standby, are corrected. Based on the corrected indicators, the evaluation coefficients for each operating state are recalculated. It is important to note here that the corrected deviation terms... The denominator in the original calculation formulas of Examples 4 to 6 was directly replaced (the original denominator was the statistical standard deviation). The physical significance of this replacement lies in the fact that, as batteries age, the benchmark for evaluating whether a single cell is "abnormal" is no longer the current statistical dispersion of the group, but rather a "tolerance band" that incorporates the time dimension. When the voltage deviation of a certain cell is less than the tolerance factor at the current moment... When the deviation is less than 1, it will be considered a normal fluctuation in the subsequent weighted fusion. Conversely, only when the deviation significantly exceeds the reasonable boundary after considering aging factors will a deviation contribution greater than 1 be generated, thereby triggering a high-risk warning. This mechanism effectively solves the technical problem that traditional fixed thresholds or pure statistical thresholds lose their discriminative power in the later stages of battery life due to the increase in overall dispersion, or frequently cause false alarms due to the lack of consideration for aging.

[0060] To more intuitively verify the rationality of this correction mechanism, a set of specific numerical calculation examples are provided below. Assume that the allowable deviation for a new battery of a certain type is... Aging rate coefficient When the battery is in brand new condition ( )hour, At this point, the tolerance is extremely narrow, and any tiny voltage deviation will be sensitively detected, meeting the maintenance requirements for maintaining high consistency in new batteries; when the battery reaches the mid-term of service (e.g. (hours) The tolerance has increased by approximately 100 times, meaning the system automatically adapts to the performance differentiation that inevitably occurs after two years of battery pack use, avoiding misjudging normal aging variations as faults; when the battery enters its final stage (e.g. (hours) Although the usage time increased fourfold, the tolerance only increased by about twofold, demonstrating the convergence control of the logarithmic function over later-stage risks and preventing excessive tolerance from masking the true accelerated degradation. It should be understood that the above values ​​are merely illustrative examples to explain the principles of this invention; in practical applications, and The specific values ​​should be calibrated according to the battery type, application scenario, and operation and maintenance strategy. As long as the core concept of logarithmic tolerance increasing non-linearly over time is followed, it falls within the protection scope of this invention. Through the correction mechanism of this embodiment, the inspection system can maintain the dynamic adaptability of the evaluation standard throughout its entire life cycle, ensuring both refined management in the new battery stage and economical operation and maintenance in the old battery stage, significantly enhancing the robustness of the technical solution in different service stages.

[0061] Example 8: Please see Figure 2 Another embodiment of this application provides a single-cell battery inspection system based on a battery pack, including: a data acquisition module, a status assessment module, and a cycle determination module; The data acquisition module is used to acquire real-time data of the battery pack and several individual batteries. The status assessment module is used to analyze the operating status of the battery based on real-time data of the battery pack; and to analyze the evaluation coefficient of each individual battery based on the operating status and real-time data of several individual batteries. The cycle determination module is used to analyze the inspection cycle of several individual batteries based on the evaluation coefficient.

[0062] Some of the data in the above formula are calculated by removing dimensions and taking their numerical values. The formula is the closest to the real situation obtained by software simulation of a large amount of collected data. The preset parameters and preset thresholds in the formula are set by those skilled in the art according to the actual situation or obtained through simulation of a large amount of data.

[0063] The working principle of this application is as follows: This application acquires real-time data of the battery pack and several individual cells; analyzes the operating status of the battery based on the real-time data of the battery pack; analyzes the evaluation coefficient of the individual cells based on the operating status and the real-time data of several individual cells; and analyzes the inspection cycle of several individual cells based on the evaluation coefficient.

[0064] The above embodiments are only used to illustrate the technical methods of this application and are not intended to limit it. Although this application has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of this application without departing from the spirit and scope of the technical methods of this application.

Claims

1. A method for inspecting a single battery cell in a battery pack, characterized in that, include: Acquire real-time data of the battery pack and several individual cells; The battery's operating status is analyzed based on real-time data from the battery pack. The evaluation coefficient of a single battery is analyzed based on its operating status and real-time data from several individual batteries. The inspection cycle of several individual batteries was analyzed based on the evaluation coefficient.

2. The method for inspecting a single battery cell based on a battery pack according to claim 1, characterized in that, The real-time data analysis of the battery pack's operating status includes: Retrieve real-time data of the battery pack; the real-time data of the battery pack includes: real-time total current and real-time total voltage; A first-order low-pass filter is used to smooth the real-time data of the battery pack; the average current and voltage change rate are calculated within a set time window based on the processed real-time data of the battery pack. The filtering formula is as follows: ;in, This represents the real-time data of the battery pack sampled for the kth time. This is the value after the k-th filtering. This is the value after the (k-1)th filtering. These are the filter coefficients; The formula for calculating the average current is: The formula for calculating the voltage change rate is: ;in, The length of the window; Let be the real-time total current value at time j; The real-time total voltage value at time k; This represents the real-time total voltage value at kN. The sampling period; The average current and voltage change rate within a set time period are integrated into a state analysis sequence in chronological order; the state analysis model is called, and the state analysis sequence is input into the state analysis function to obtain the operating state of the battery pack; the operating state includes: charging, discharging, and standby; the state analysis model is built based on an artificial intelligence model.

3. The method for inspecting a single battery cell based on a battery pack according to claim 2, characterized in that, The state analysis model is built based on an artificial intelligence model and includes: Select a model framework and deep learning algorithm from the artificial intelligence library; construct the model based on the model framework and deep learning algorithm to obtain the initial model; Obtain the standard dataset; the standard dataset includes standard input data consistent with the content attributes of the state analysis sequence, and standard output data consistent with the content attributes of the running state; The standard dataset is divided into training, validation, and test sets according to a set ratio; the initial model is trained using the training set; the internal parameters of the initial model are adjusted using the validation set; and the initial model is tested using the test set to obtain test metrics. Obtain the standard range of indicators; compare the test indicators with the standard range of indicators; if all test indicators are within the standard range of indicators, mark the initial model as the state analysis model; otherwise, rebuild and retrain the initial model.

4. The method for inspecting a single battery cell based on a battery pack according to claim 1, characterized in that, The method of analyzing the evaluation coefficient of a single battery cell based on its operating status and real-time data from several individual cells includes: Retrieve real-time data of several individual cells and the operating status of the battery pack; the real-time data of individual cells includes: real-time voltage, real-time current, and usage time; When the battery pack is charging, the charging characteristic indicators of individual cells are analyzed, and the charging evaluation coefficient of the battery is calculated based on the charging characteristic indicators. When the battery pack is in a discharging state, the discharge characteristic indicators of individual cells are analyzed, and the discharge evaluation coefficient of the battery is calculated based on the discharge characteristic indicators. When the battery pack is in standby mode, the standby characteristic indicators of individual cells are analyzed, and the standby evaluation coefficient of the battery is calculated based on the standby characteristic indicators. The evaluation coefficients for each operating state are adjusted based on usage time.

5. A method for inspecting a single battery cell based on a battery pack according to claim 4, characterized in that, The analysis of the charging characteristic indicators of a single battery cell, and the calculation of the battery's charging evaluation coefficient based on these indicators, includes: Retrieve real-time data from several individual cells; calculate the average real-time voltage and standard deviation of several individual cells; calculate the voltage balance deviation of each individual cell based on the voltage deviation analysis function; The expression for the voltage deviation analysis function is: ;in, This represents the real-time voltage of the i-th battery cell; This represents the average value of the real-time voltage. The standard deviation of voltage; It is a very small constant; Based on formula Calculate the voltage change rate deviation for each battery cell; where, Let be the rate of change of voltage in the i-th cell; This represents the average rate of change of voltage within the battery pack. The standard deviation of the rate of change of voltage; Calculate the rate of change of real-time current. When the rate of change of real-time current is less than a set rate of change threshold, then based on the formula... Calculate the relative deviation of the charging internal resistance of each battery cell; where, Let be the internal resistance of the i-th battery during charging. ; This represents the average value of the charging internal resistance; This represents the standard deviation of the charging internal resistance within the battery pack. Based on formula Calculate the charging evaluation factor for each battery cell; where, , and The weighting coefficient is greater than 0.

6. The method for inspecting a single battery cell based on a battery pack according to claim 4, characterized in that, The analysis of the discharge characteristic indicators of a single battery cell, and the calculation of the battery's discharge evaluation coefficient based on these indicators, includes: Retrieve real-time data from several individual cells; recalculate the voltage balance deviation of each individual cell based on the voltage deviation analysis function; based on the formula... Recalculate the voltage change rate deviation for each battery cell; Based on formula Recalculate the relative deviation of the charging internal resistance of each battery cell; based on the formula Calculate the discharge evaluation coefficient for each battery cell; where, .

7. A method for inspecting a single battery cell based on a battery pack according to claim 4, characterized in that, The analysis of standby characteristic indicators of a single battery cell, and the calculation of the battery's standby evaluation coefficient based on these indicators, includes: Retrieve real-time data from several individual battery cells; calculate the static voltage deviation of each individual battery cell based on the voltage deviation analysis function; The coefficient of variation of the voltage sampling sequence during the standby period is calculated based on the voltage fluctuation analysis function; the expression of the voltage fluctuation analysis function is: ;in ; Let be the real-time voltage of the i-th battery at time k; This represents the total number of sampling times. Based on formula The coefficient of variation for each battery cell was standardized; among them, This represents the average value of the coefficient of variation within the battery pack. This represents the standard deviation of the coefficient of variation within the battery pack. Based on formula Calculate the standby evaluation coefficient for each battery cell; where, and The weighting coefficient is greater than 0.

8. A method for inspecting a single battery cell based on a battery pack according to claim 4, characterized in that, The correction of the evaluation coefficients for each operating state based on usage time includes: Retrieve the usage time from the real-time data of each battery; calculate the time tolerance factor based on the tolerance factor analysis function; the expression for the tolerance analysis factor is: ;in, This is the allowable deviation for the new battery; This is the aging rate coefficient; For usage time; It is a logarithmic function with base e; Based on formula The voltage balance deviation in charging and discharging states, as well as the voltage settling deviation in standby state, are corrected; the evaluation coefficient for each operating state is recalculated based on the corrected indicators.

9. A method for inspecting a single battery cell based on a battery pack according to claim 1, characterized in that, The method for analyzing the inspection cycle of several individual battery cells based on evaluation coefficients includes: Retrieve the corrected evaluation coefficients; obtain the inspection cycle library; where the evaluation coefficients include: charging evaluation coefficients, discharging evaluation coefficients, and standby evaluation coefficients; The corrected evaluation coefficients are matched with the inspection cycle library to obtain the inspection cycle corresponding to each operating state. The shortest inspection cycle is selected as the inspection cycle of the corresponding single battery cell.

10. A single-cell battery inspection system based on a battery pack, applied to the single-cell battery inspection method based on a battery pack as described in any one of claims 1-9, characterized in that, include: Data acquisition module, status assessment module, and cycle determination module; The data acquisition module is used to acquire real-time data of the battery pack and several individual batteries. The status assessment module is used to analyze the operating status of the battery based on real-time data from the battery pack. The evaluation coefficient of a single battery is analyzed based on its operating status and real-time data from several individual batteries. The cycle determination module is used to analyze the inspection cycle of several individual batteries based on the evaluation coefficient.

Images