Unit cell health degree identification method applied to series-parallel battery pack

Abstract

The application discloses a unit cell health degree identification method applied to a hybrid battery pack, relates to the technical field of battery health state monitoring, and generates a standard operation data set by collecting operation data of each unit cell in the hybrid battery pack and pre-processing the operation data; extracts a voltage-state of charge curve of each unit cell from the standard operation data set, generates a corresponding health degree characteristic sequence; obtains estimated current according to a current distribution inversion model, and generates corresponding current distribution characteristic parameters; generates corresponding unit cell health degrees according to the health degree characteristic sequence and the current distribution characteristic parameters, and then identifies abnormal unit cells; outputs the unit cell health degrees and generates alarm information for the abnormal unit cells; and the application realizes quantitative evaluation of the health states of each unit cell in the hybrid battery pack and automatic identification of abnormal unit cells, and improves the intelligent level of battery management.

Description

Technical Field

[0001] This invention relates to the field of battery health status monitoring technology, specifically a method for identifying the health status of unit cells in a hybrid battery pack. Background Technology

[0002] Hybrid battery packs are composed of multiple cell units connected in series and in parallel. They are widely used in energy storage power stations, electric vehicles and backup power supplies. Due to differences in manufacturing processes, uneven operating temperature distribution and different degrees of cycle aging, the health status of each cell unit in a hybrid battery pack varies significantly. The performance degradation of individual cell units can drag down the usable capacity of the entire battery pack and bring safety hazards.

[0003] Existing health assessment methods are mostly geared towards the overall battery pack or battery module level, making it difficult to pinpoint the individual cell level. Furthermore, the current of individual cells within parallel branches cannot be directly measured, leading to significant errors in health assessment based on traditional single-cell models. In addition, existing methods lack the ability to pre-diagnose trends in individual cell health and provide support for health management decisions, making it impossible to identify degradation risks in advance and develop targeted maintenance strategies.

[0004] How to accurately identify the health status of each cell in a hybrid battery pack, and how to invert the cell status based on measurable data to achieve pre-diagnosis and health management are problems we need to solve. To this end, we now provide a method for identifying the health status of cells in hybrid battery packs. Summary of the Invention

[0005] The purpose of this invention is to provide a method for identifying the health status of unit cells in a hybrid battery pack.

[0006] The objective of this invention can be achieved through the following technical solution: a method for identifying the health status of individual cells in a hybrid battery pack, comprising the following steps:

[0007] Step S1: Collect the operating data of each cell in the hybrid battery pack, preprocess the operating data, and generate a standard operating dataset;

[0008] Step S2: Extract the voltage-state-of-charge curve of each cell from the standard operating dataset and then generate the corresponding health feature sequence;

[0009] Step S3: Obtain the estimated current based on the pre-built current distribution inversion model, and then generate the corresponding current distribution characteristic parameters;

[0010] Step S4: Generate the health status of each cell based on the health status feature sequence and current distribution feature parameters, and identify abnormal cells based on the health status of each cell.

[0011] Step S5: Output the health status of the cell and generate alarm information for abnormal cells.

[0012] Furthermore, the process of collecting operational data from each cell within the hybrid battery pack includes:

[0013] Each battery cell is connected to the sampling interface of the battery management system via an independent voltage sampling line and temperature sensor. The main circuit position of each parallel branch is connected to the sampling interface of the battery management system via a current sensor. The battery management system synchronously collects the terminal voltage, battery temperature, and branch current of each parallel branch of each battery cell according to a preset sampling period through the above sampling interface, and estimates the state of charge of each battery cell in real time. The battery management system also stores the manufacturer's technical parameters of each battery cell in advance, including the rated capacity.

[0014] The corresponding operating data is obtained by integrating the terminal voltage of each cell, the branch current of each parallel branch, the cell temperature of each cell, and the state of charge.

[0015] Furthermore, the process of preprocessing the runtime data to generate a standard runtime dataset includes:

[0016] The collected operational data is processed sequentially with outlier removal, filtering and smoothing, and timing alignment. For the terminal voltage, branch current, and battery temperature of each cell, it is determined whether they exceed their respective preset reasonable value ranges. Operational data that exceeds the corresponding reasonable value ranges are identified as outliers and removed.

[0017] After outlier removal, the operational data after outlier removal are filtered and smoothed. The operational data of each cell after outlier removal and filtering and smoothing are interpolated and aligned according to a unified time base so that each operational data has a consistent sampling time on the time axis. The operational data after time alignment are integrated into a standard operational dataset.

[0018] Furthermore, the process of extracting the voltage-state-of-charge curves of each cell from the standard operating dataset includes:

[0019] The terminal voltage and state of charge of each cell in a complete charge-discharge cycle are obtained from the standard operating dataset. Taking a complete charge-discharge cycle as the extraction unit, starting from the start time of a complete charge-discharge cycle, the corresponding values ​​of the state of charge and terminal voltage at each sampling time in the complete charge-discharge cycle are read sequentially to form a set of discrete data points with the state of charge as the independent variable and the terminal voltage as the dependent variable.

[0020] To eliminate the influence of measurement noise on the curve shape, the discrete data point sequence is filtered and smoothed to obtain a smoothed discrete data point sequence. The data points in the smoothed discrete data point sequence are sorted in ascending order of their state of charge values. For data points with the same or similar state of charge values, the average value of their terminal voltage is taken as the terminal voltage corresponding to that state of charge data point, thereby obtaining the voltage-state of charge curve corresponding to the complete charge-discharge cycle.

[0021] If a single battery cell has multiple complete charge-discharge cycles, the corresponding voltage-state-of-charge curve is extracted for each complete charge-discharge cycle. The extracted voltage-state-of-charge curves are then averaged to generate the voltage-state-of-charge curve for the corresponding battery cell.

[0022] Furthermore, the process of generating the corresponding health feature sequence includes:

[0023] Health characteristic parameters are extracted from the voltage-charge state curve. These health characteristic parameters include average open-circuit voltage, voltage plateau interval length, polarization voltage, capacity decay, and temperature rise characteristic parameters. The average open-circuit voltage is obtained based on the average value of the terminal voltage of the unit cell within the range of state of charge variation. The voltage change rate curve is obtained by differentiating the voltage-charge state curve. Continuous state of charge intervals where the absolute value of the voltage change rate is less than a preset change rate threshold are marked as voltage plateau intervals. The state of charge lengths of all voltage plateau intervals are summed to obtain the voltage plateau interval length.

[0024] The polarization voltage is obtained based on the difference between the terminal voltage and the open circuit voltage of the cell. The current actual usable capacity of the cell is obtained based on the branch current. The capacity decay is obtained based on the difference between the current actual usable capacity and the rated capacity of the cell. The temperature rise characteristic parameters of each cell are obtained based on the battery temperature in the standard operating dataset.

[0025] The health characteristic parameters of each cell are standardized, and the health characteristic sequence of each cell is generated after the standardization process.

[0026] Furthermore, the process of constructing the current distribution inversion model includes:

[0027] Select unit cells with the same model and specifications as the same type of unit cells, and conduct internal resistance characteristic tests under different temperatures, different states of charge and different current conditions. The internal resistance characteristic test covers the battery temperature range and state of charge range that the unit cells may encounter in actual operation. Record the terminal voltage, battery temperature, state of charge and actual current under each test condition to form a test dataset.

[0028] Based on the experimental dataset, with terminal voltage, battery temperature, and state of charge as input variables and actual current as output variable, a multinomial regression method is used to fit the mapping relationship between input and output variables to obtain a current allocation inversion model. The terminal voltage, battery temperature, and state of charge of the cell are then input into the current allocation inversion model to obtain the corresponding estimated current.

[0029] Furthermore, the process of obtaining the estimated current based on the pre-built current distribution inversion model and then generating the corresponding current distribution characteristic parameters includes:

[0030] For each cell in the parallel branch, the terminal voltage, cell temperature and state of charge of the cell are input into the current distribution inversion model. The current distribution inversion model outputs the estimated current of the cell. The sum of the estimated currents of all cells in the parallel branch is equal to the branch current of the corresponding parallel branch. Then, current distribution characteristic parameters are generated based on the estimated current.

[0031] Furthermore, the process of generating the cell health status based on the health characteristic sequence and current distribution characteristic parameters includes:

[0032] The capacity retention rate is obtained based on the capacity decay in the health feature sequence. The relative internal resistance level of each cell is obtained based on the current distribution feature parameters of the cell. The current distribution health contribution is generated based on the relative internal resistance level.

[0033] The voltage-state-of-charge curve of each cell in the current charge-discharge cycle is dynamically time-normalized and compared with the corresponding reference voltage-state-of-charge curve to obtain the curve shape difference. Based on the curve shape difference, the curve shape health contribution is generated. The cell health is obtained by comprehensively considering the capacity retention rate, current distribution health contribution, and curve shape health contribution.

[0034] Furthermore, the process of identifying abnormal battery cells based on the health status of each cell includes:

[0035] Obtain the health status of all cells in the hybrid battery pack, generate the average and standard deviation of the health status of all cells, set a health status deviation threshold based on the average and standard deviation of the health status of all cells, compare the health status of each cell with the health status deviation threshold, and mark the corresponding cell as a suspected abnormal cell when the health status of a cell is less than the health status deviation threshold.

[0036] A second confirmation is performed on suspected abnormal cell batteries. The terminal voltage and battery temperature of the suspected abnormal cell battery within the most recent preset time period are retrieved, as well as the terminal voltage and battery temperature of other cells batteries in the same parallel branch within the same time period, to form the corresponding terminal voltage curve and battery temperature curve respectively.

[0037] The terminal voltage curve and battery temperature curve of the suspected abnormal cell are compared with the terminal voltage curve and battery temperature curve of other cells in the same parallel branch.

[0038] When the deviation between the terminal voltage curve of a suspected abnormal cell and the terminal voltage curve of other cells in the parallel branch is continuously greater than a preset voltage deviation threshold, and the deviation between the cell temperature curve and the cell temperature curve of other cells in the parallel branch is continuously greater than a preset temperature deviation threshold, the suspected abnormal cell is marked as an abnormal cell.

[0039] Furthermore, the process of outputting the health status of the battery cells and generating alarm information for abnormal battery cells includes:

[0040] The health status and sorting results of each battery cell are output to the display terminal in list form. The health status is classified into corresponding levels according to the health status of the battery cells. Alarm information is generated for the battery cells marked as abnormal and pushed to the battery management system. The alarm information includes the health status value, health level, abnormality type and handling suggestions of the abnormal battery cells.

[0041] The abnormality types include capacity decay abnormality, current distribution abnormality, curve shape deviation abnormality, and temperature abnormality. Capacity decay abnormality is triggered when the capacity retention rate is lower than a preset capacity threshold. Current distribution abnormality is triggered when the current distribution health contribution is lower than a preset current distribution threshold. Curve shape deviation abnormality is triggered when the curve shape health contribution is lower than a preset shape threshold. Temperature abnormality is triggered when the temperature rise characteristic parameter exceeds a preset temperature rise threshold.

[0042] The proposed solutions include replacement recommendations, equalization maintenance recommendations, and reduced power operation recommendations. Replacement recommendations are given for abnormal cell units exhibiting abnormal capacity decay or abnormal curve deviation. Equalization maintenance recommendations are given for abnormal cell units with abnormal current distribution. Reduced power operation recommendations are given for abnormal cell units with abnormal temperature. When multiple abnormal types exist for the same abnormal cell unit, the final solution is determined by prioritizing replacement recommendations over equalization maintenance recommendations, and equalization maintenance recommendations over reduced power operation recommendations.

[0043] Compared with the prior art, the beneficial effects of the present invention are as follows: Based on the physical constraint that the terminal voltage of each cell in the parallel branch of the hybrid battery pack is the same, the estimated current of each cell is obtained through the current distribution inversion model, and then the current distribution characteristic parameters are calculated and the relative internal resistance level of the cell is obtained by inversion. No additional internal resistance measurement equipment is required, which reduces the system hardware cost. At the same time, the morphological comparison of the voltage-state of charge curve is introduced, which overcomes the influence of the state of charge estimation deviation on the point-to-point direct comparison, realizes the comprehensive quantitative evaluation of capacity decay, internal resistance growth and polarization increase, and significantly improves the accuracy of health assessment.

[0044] The cell health score is a weighted fusion of three dimensions: capacity retention rate, current distribution health contribution, and curve shape health contribution. It comprehensively reflects the battery health status from three perspectives: capacity decay, internal resistance increase, and polarization increase. The anomaly detection threshold is adaptively determined by the statistical relationship between the mean and standard deviation of the health score, avoiding the problem of insufficient adaptability of fixed thresholds at different aging stages. Combined with the secondary confirmation mechanism of terminal voltage curve and battery temperature curve, the false alarm rate is effectively reduced and the reliability of anomaly identification is improved.

[0045] This invention forms a fully automated closed-loop process that can be directly deployed in existing battery management systems to achieve online assessment and anomaly diagnosis of the health status of hybrid battery packs without manual intervention or additional hardware modifications, and has good engineering promotion value. Attached Figure Description

[0046] 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.

[0047] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0048] like Figure 1 As shown, the method for identifying the health status of unit cells in a hybrid battery pack includes the following steps;

[0049] Step S1: Collect the operating data of each cell in the hybrid battery pack, preprocess the operating data, and generate a standard operating dataset;

[0050] Step S2: Extract the voltage-state-of-charge curve of each cell from the standard operating dataset and then generate the corresponding health feature sequence;

[0051] Step S3: Obtain the estimated current based on the pre-built current distribution inversion model, and then generate the corresponding current distribution characteristic parameters;

[0052] Step S4: Generate the health status of each cell based on the health status feature sequence and current distribution feature parameters, and identify abnormal cells based on the health status of each cell.

[0053] Step S5: Output the health status of the cell and generate alarm information for abnormal cells.

[0054] It should be further explained that, in the specific implementation process, the process of collecting operational data from each cell within the hybrid battery pack includes:

[0055] The hybrid battery pack is composed of multiple unit cells connected in series and in parallel.

[0056] Each cell is connected to the sampling interface of the battery management system through an independent voltage sampling line and temperature sensor. The main circuit position of each parallel branch is connected to the sampling interface of the battery management system through a current sensor.

[0057] The battery management system synchronously collects the terminal voltage, battery temperature and branch current of each parallel branch of each battery cell through the above sampling interface according to a preset sampling period, and estimates the state of charge of each battery cell in real time. The battery management system refers to an electronic control unit that monitors and manages the hybrid battery pack.

[0058] The battery management system also pre-stores the manufacturer's technical parameters for each battery cell, including the rated capacity, which are obtained from the technical manual provided by the battery manufacturer.

[0059] The corresponding operating data is obtained by integrating the terminal voltage of each cell, the branch current of each parallel branch, the cell temperature of each cell, and the state of charge.

[0060] It should be further explained that, in the specific implementation process, the process of preprocessing the runtime data to generate a standard runtime dataset includes:

[0061] The collected runtime data is processed sequentially, including outlier removal, filtering and smoothing, and timing alignment.

[0062] For the terminal voltage, branch current and battery temperature of each cell, determine whether they exceed their respective preset reasonable value range, and identify the operating data that exceeds the corresponding reasonable value range as abnormal values ​​and remove them.

[0063] After outlier removal, each piece of operational data after outlier removal is filtered and smoothed using a moving average filtering method.

[0064] Specifically:

[0065] ;

[0066] in, To preset the sliding window length, This is the index of the current sampling time. For the first The running data after filtering and smoothing at each sampling time;

[0067] The operating data of each cell are calculated point by point according to the above formula to complete the filtering and smoothing process.

[0068] The operational data of each cell after outlier removal and filtering and smoothing are interpolated and aligned according to a unified time base to ensure that each operational data has a consistent sampling time on the time axis. The operational data after time alignment are then integrated into a standard operational dataset.

[0069] It should be further explained that, in the specific implementation process, the process of extracting the voltage-state-of-charge curves of each cell from the standard operating dataset includes:

[0070] The terminal voltage and state of charge of each cell are obtained from the standard operating dataset during a complete charge-discharge cycle, and the terminal voltage and state of charge correspond one-to-one on the time axis.

[0071] Using a complete charge-discharge cycle as the extraction unit, starting from the beginning of a complete charge-discharge cycle, the corresponding values ​​of the state of charge and terminal voltage at each sampling moment within the complete charge-discharge cycle are read sequentially to form a set of discrete data points with the state of charge as the independent variable and the terminal voltage as the dependent variable.

[0072] The complete charge-discharge cycle refers to the discharge process in which the state of charge decreases from near 100% to near 0%, or the charging process in which the state of charge increases from near 0% to near 100%.

[0073] To eliminate the influence of measurement noise on the curve shape, the discrete data point sequence is filtered and smoothed to obtain a smoothed discrete data point sequence.

[0074] The data points in the smoothed discrete data point sequence are sorted in ascending order of their state of charge (SOC) values. For data points with the same or similar SOC values, the average value of their terminal voltages is taken as the terminal voltage corresponding to that SOC data point, thereby obtaining the voltage-SOC curve corresponding to the complete charge-discharge cycle. The voltage-SOC curve is a mapping relationship between the terminal voltage and the continuous change of the SOC. Each data point contains the corresponding SOC and terminal voltage.

[0075] If a single battery cell has multiple complete charge-discharge cycles, the corresponding voltage-state-of-charge curve is extracted for each complete charge-discharge cycle. The extracted voltage-state-of-charge curves are then averaged to generate the voltage-state-of-charge curve for the corresponding battery cell.

[0076] The averaging process involves sampling at equal intervals within the 0% to 100% state of charge range, with a preset state of charge step size. For each sampling point, the arithmetic mean of the terminal voltages of multiple curves at that sampling point is taken as the representative value of the corresponding sampling point, and these values ​​are sequentially connected to generate the voltage-state of charge curve of the corresponding cell.

[0077] It should be further explained that, in the specific implementation process, the process of generating the corresponding health feature sequence includes:

[0078] Health characteristic parameters are extracted from the voltage-charge state curve. These health characteristic parameters include average open-circuit voltage, voltage plateau interval length, polarization voltage, capacity decay, and temperature rise characteristic parameters.

[0079] The average open-circuit voltage is obtained by averaging the terminal voltage of the cell over the range of its state of charge variation, and is denoted as . ,Right now:

[0080] ;

[0081] in, This represents the upper limit of the range of changes in the state of charge. This represents the lower limit of the range of changes in the state of charge. The state of charge is The corresponding open-circuit voltage at that time;

[0082] The open-circuit voltage refers to the terminal voltage measured when the cell is in a static state, i.e., when no current flows through it. The state of charge variation range refers to the range of the state of charge of the cell from the initial value to the final value during a single charge and discharge cycle.

[0083] Differentiating the voltage-charge state curve yields the voltage rate of change curve. Continuous charge state intervals where the absolute value of the voltage rate of change is less than a preset threshold are marked as voltage plateau intervals. The charge state lengths of all voltage plateau intervals are summed to obtain the length of the voltage plateau interval, denoted as . ,Right now:

[0084] ;

[0085] in, This represents the total number of voltage plateau intervals. For the first The final state of charge of each voltage plateau interval For the first The initial state of charge of each voltage plateau range;

[0086] The polarization voltage is obtained by the difference between the terminal voltage and the open-circuit voltage of the cell, denoted as . ,Right now:

[0087] ;

[0088] in, The total number of sampling points. For the first The state of charge corresponding to each sampling point The state of charge is Terminal voltage at time The state of charge is Open-circuit voltage at that time;

[0089] The current actual usable capacity of the cell is obtained based on the branch current, denoted as . ,Right now:

[0090] ;

[0091] in, Indicates the branch current;

[0092] The capacity decay is obtained by measuring the difference between the current usable capacity and the rated capacity of the cell, and is denoted as: ,Right now:

[0093] ;

[0094] in, This refers to the rated capacity of the cell unit.

[0095] Simultaneously, based on the battery temperatures in the standard operating dataset, the temperature rise characteristic parameters of each cell are calculated. These temperature rise characteristic parameters refer to the temperature rise of the cell during the charging and discharging process, denoted as... ,Right now:

[0096] ;

[0097] in, This represents the highest temperature value of a single cell during the charging and discharging process. This represents the temperature of the cell at the start of charging and discharging.

[0098] The health characteristic parameters of each cell are standardized. The standardization process refers to mapping each health characteristic parameter to a unified value range according to a preset benchmark value and a proportional coefficient. The benchmark value of each health characteristic parameter is determined based on the standard value of the same type of cell under healthy conditions, and the proportional coefficient is preset according to the physical meaning of each health characteristic parameter. After the standardization process is completed, the health characteristic sequence of each cell is generated.

[0099] It should be further explained that, in the specific implementation process, the process of constructing the current distribution inversion model includes:

[0100] Select unit cells with the same model and specifications as the same type of unit cells, and conduct internal resistance characteristic tests under different temperatures, different states of charge and different current conditions. The internal resistance characteristic test is used to establish the correspondence between the terminal voltage, battery temperature, state of charge and actual current of the unit cell.

[0101] The actual current refers to the current value flowing through the cell in the internal resistance characteristic test, which is obtained by direct measurement and is used to characterize the real current load state of the cell under the corresponding test conditions.

[0102] During the internal resistance characteristic test, the battery temperature range and state of charge range that the cell may encounter in actual operation are covered. The terminal voltage, battery temperature, state of charge and actual current under each test condition are recorded to form a test dataset.

[0103] Based on the experimental dataset, the terminal voltage, battery temperature and state of charge are used as input variables, and the actual current is used as the output variable. The multinomial regression method is used to fit the mapping relationship between the input variables and the output variables to obtain the current distribution inversion model.

[0104] The terminal voltage, battery temperature, and state of charge input current of the cell are assigned to the inversion model to obtain the corresponding estimated current.

[0105] It should be further explained that, in the specific implementation process, the process of obtaining the estimated current based on the pre-built current distribution inversion model and then generating the corresponding current distribution characteristic parameters includes:

[0106] For each cell in the parallel branch, the terminal voltage, cell temperature and state of charge of the cell are input into the current distribution inversion model. The current distribution inversion model outputs the estimated current of the cell. The sum of the estimated currents of all cells in the parallel branch is equal to the branch current of the corresponding parallel branch.

[0107] The current distribution characteristic parameters are calculated based on the estimated current, and denoted as follows: ,Right now:

[0108] ;

[0109] in, The total number of cell units in the parallel branch. Let i be the estimated current of the i-th cell. This refers to the branch current of the parallel branch.

[0110] It should be further explained that, in the specific implementation process, the process of generating the cell health status based on the health status feature sequence and current distribution feature parameters includes:

[0111] The capacity retention rate is calculated based on the capacity decay in the health feature sequence, denoted as . ,Right now:

[0112] ;

[0113] Based on the current distribution characteristic parameters of the individual cells, the relative internal resistance level of each individual cell is obtained, denoted as . ,Right now:

[0114] ;

[0115] The health contribution of current distribution is calculated based on the relative internal resistance level and denoted as follows: ,Right now:

[0116] ;

[0117] in, The preset current distribution sensitivity coefficient;

[0118] The voltage-state-of-charge curve of each cell in the current charge-discharge cycle is dynamically time-corrected and compared with the corresponding reference voltage-state-of-charge curve. The reference voltage-state-of-charge curve is the voltage-state-of-charge curve of the cell in a brand new state, which is pre-stored in the battery management system.

[0119] Specifically:

[0120] The voltage-state curve of the current charge-discharge cycle is discretized into a first voltage sequence, and the reference voltage-state curve is discretized into a second voltage sequence. The Euclidean distance between corresponding points in the first and second voltage sequences is calculated as a distance metric between points. A dynamic programming method is used to search for the optimal regularized path between the first and second voltage sequences, minimizing the cumulative distance on the regularized path. The minimum cumulative distance is taken as the curve shape difference degree, denoted as . ;

[0121] The health contribution of curve morphology is calculated based on the degree of difference in curve morphology, and denoted as . ,Right now:

[0122] ;

[0123] in, This is the preset morphology sensitivity coefficient;

[0124] The health of a single cell is determined by comprehensively considering its contribution to capacity retention, current distribution health, and curve morphology health, and is denoted as... ,Right now:

[0125] ;

[0126] in, , , These are preset weighting coefficients corresponding to capacity retention rate, current distribution health contribution, and curve morphology health contribution, respectively, satisfying... The relationship.

[0127] It should be further explained that, in the specific implementation process, the process of identifying abnormal cell batteries based on the health status of each cell battery includes:

[0128] Obtain the cell health status of all cells within the hybrid battery pack, denoted as . ,in, Q, This represents the total number of individual cells within the hybrid battery pack.

[0129] Generate the average and standard deviation of the health status of all cell units, denoted as . and ,Right now:

[0130] ;

[0131] ;

[0132] A health deviation threshold is set based on the average and standard deviation of the health status of all cell units, denoted as . ,Right now:

[0133] ;

[0134] The health status of each cell is compared with the health status deviation threshold. When the health status of a cell is less than the health status deviation threshold, the cell is marked as a suspected abnormal cell.

[0135] A second confirmation was conducted on suspected abnormal battery cells;

[0136] Specifically:

[0137] The terminal voltage and battery temperature of the suspected abnormal cell within the most recent preset time period are retrieved. At the same time, the terminal voltage and battery temperature of other cells in the parallel branch are retrieved within the same time period, and the terminal voltage curve and battery temperature curve of the corresponding cell are constructed respectively.

[0138] The terminal voltage curve and battery temperature curve of the suspected abnormal cell are compared with the terminal voltage curve and battery temperature curve of other cells in the same parallel branch.

[0139] The preset duration is set in advance according to the battery type and operating conditions;

[0140] When the deviation between the terminal voltage curve of a suspected abnormal cell and the terminal voltage curve of other cells in the parallel branch is continuously greater than a preset voltage deviation threshold, and the deviation between the battery temperature curve and the battery temperature curve of other cells in the parallel branch is continuously greater than a preset temperature deviation threshold, the suspected abnormal cell is marked as an abnormal cell.

[0141] The term "continuously greater than" means that the deviation is continuously true within a preset continuous judgment period.

[0142] It should be further explained that, in the specific implementation process, the process of outputting the health status of the unit battery and generating alarm information for abnormal unit batteries includes:

[0143] The health status and sorting results of each battery cell are output to the display terminal in list form. The display terminal refers to the display device connected to the battery management system, which displays the health status value and health status level of each battery cell from high to low.

[0144] When the health of a cell is greater than or equal to the preset good health threshold, it is marked as good level; when the health of a cell is between the preset good health threshold and the preset deterioration health threshold, it is marked as normal level; when the health of a cell is less than or equal to the preset deterioration health threshold, it is marked as deterioration level.

[0145] An alarm message is generated for the cells marked as abnormal cells. The alarm message includes the health value, health level, abnormality type and handling suggestions of the abnormal cells.

[0146] The abnormality types include capacity decay abnormality, current distribution abnormality, curve shape deviation abnormality, and temperature abnormality.

[0147] When the capacity retention rate is lower than the preset capacity threshold, a capacity decay anomaly is triggered. When the current distribution health contribution is lower than the preset current distribution threshold, a current distribution anomaly is triggered. When the curve shape health contribution is lower than the preset shape threshold, a curve shape deviation anomaly is triggered. When the temperature rise characteristic parameter exceeds the preset temperature rise threshold, a temperature anomaly is triggered.

[0148] The proposed solutions include replacement recommendations, balanced maintenance recommendations, and reduced power operation recommendations.

[0149] Replacement recommendations are given for abnormal cell units exhibiting abnormal capacity decay and abnormal curve shape deviation; equalization maintenance recommendations are given for abnormal cell units with abnormal current distribution; and reduced power operation recommendations are given for abnormal cell units with abnormal temperature.

[0150] When multiple abnormal types exist in the same abnormal cell, the final handling recommendation is determined in the following order: replacement recommendation takes precedence over equalization maintenance recommendation, and equalization maintenance recommendation takes precedence over power reduction operation recommendation.

[0151] Alarm information is pushed to the battery management system, and at the same time, audible and visual alerts are provided through the display terminal.

[0152] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications or equivalent substitutions made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for identifying the health status of unit cells in a hybrid battery pack, characterized in that, Includes the following steps: Step S1: Collect the operating data of each cell in the hybrid battery pack, preprocess the operating data, and generate a standard operating dataset; Step S2: Extract the voltage-state-of-charge curve of each cell from the standard operating dataset and then generate the corresponding health feature sequence; Step S3: Obtain the estimated current based on the pre-built current distribution inversion model, and then generate the corresponding current distribution characteristic parameters; Step S4: Generate the health status of each cell based on the health status feature sequence and current distribution feature parameters, and identify abnormal cells based on the health status of each cell. Step S5: Output the health status of the cell and generate alarm information for abnormal cells.

2. The method for identifying the health of unit cells in a hybrid battery pack according to claim 1, characterized in that, The process of collecting operational data from each cell within a hybrid battery pack includes: Each battery cell is connected to the sampling interface of the battery management system via an independent voltage sampling line and temperature sensor. The main circuit position of each parallel branch is connected to the sampling interface of the battery management system via a current sensor. The battery management system synchronously collects the terminal voltage, battery temperature, and branch current of each parallel branch of each battery cell according to a preset sampling period through the above sampling interface, and estimates the state of charge of each battery cell in real time. The battery management system also stores the manufacturer's technical parameters of each battery cell in advance, including the rated capacity. The corresponding operating data is obtained by integrating the terminal voltage of each cell, the branch current of each parallel branch, the cell temperature of each cell, and the state of charge.

3. The method for identifying the health of unit cells in a hybrid battery pack according to claim 2, characterized in that, The process of preprocessing runtime data to generate a standard runtime dataset includes: The collected operational data is processed sequentially with outlier removal, filtering and smoothing, and timing alignment. For the terminal voltage, branch current, and battery temperature of each cell, it is determined whether they exceed their respective preset reasonable value ranges. Operational data that exceeds the corresponding reasonable value ranges are identified as outliers and removed. After outlier removal, the operational data after outlier removal are filtered and smoothed. The operational data of each cell after outlier removal and filtering and smoothing are interpolated and aligned according to a unified time base so that each operational data has a consistent sampling time on the time axis. The operational data after time alignment are integrated into a standard operational dataset.

4. The method for identifying the health of unit cells in a hybrid battery pack according to claim 3, characterized in that, The process of extracting the voltage-state-of-charge curves of each cell from the standard operating dataset includes: The terminal voltage and state of charge of each cell in a complete charge-discharge cycle are obtained from the standard operating dataset. Taking a complete charge-discharge cycle as the extraction unit, starting from the start time of a complete charge-discharge cycle, the corresponding values ​​of the state of charge and terminal voltage at each sampling time in the complete charge-discharge cycle are read sequentially to form a set of discrete data points with the state of charge as the independent variable and the terminal voltage as the dependent variable. To eliminate the influence of measurement noise on the curve shape, the discrete data point sequence is filtered and smoothed to obtain a smoothed discrete data point sequence. The data points in the smoothed discrete data point sequence are sorted in ascending order of their state of charge values. For data points with the same or similar state of charge values, the average value of their terminal voltage is taken as the terminal voltage corresponding to that state of charge data point, thereby obtaining the voltage-state of charge curve corresponding to the complete charge-discharge cycle. If a single battery cell has multiple complete charge-discharge cycles, the corresponding voltage-state-of-charge curve is extracted for each complete charge-discharge cycle. The extracted voltage-state-of-charge curves are then averaged to generate the voltage-state-of-charge curve for the corresponding battery cell.

5. The method for identifying the health of unit cells in a hybrid battery pack according to claim 4, characterized in that, The process of generating the corresponding health feature sequence includes: Health characteristic parameters are extracted from the voltage-charge state curve. These health characteristic parameters include average open-circuit voltage, voltage plateau interval length, polarization voltage, capacity decay, and temperature rise characteristic parameters. The average open-circuit voltage is obtained based on the average value of the terminal voltage of the unit cell within the range of state of charge variation. The voltage change rate curve is obtained by differentiating the voltage-charge state curve. Continuous state of charge intervals where the absolute value of the voltage change rate is less than a preset change rate threshold are marked as voltage plateau intervals. The state of charge lengths of all voltage plateau intervals are summed to obtain the voltage plateau interval length. The polarization voltage is obtained based on the difference between the terminal voltage and the open circuit voltage of the cell. The current actual usable capacity of the cell is obtained based on the branch current. The capacity decay is obtained based on the difference between the current actual usable capacity and the rated capacity of the cell. The temperature rise characteristic parameters of each cell are obtained based on the battery temperature in the standard operating dataset. The health characteristic parameters of each cell are standardized, and the health characteristic sequence of each cell is generated after the standardization process.

6. The method for identifying the health of unit cells in a hybrid battery pack according to claim 5, characterized in that, The process of constructing the current distribution inversion model includes: Select unit cells with the same model and specifications as the same type of unit cells, and conduct internal resistance characteristic tests under different temperatures, different states of charge and different current conditions. The internal resistance characteristic test covers the battery temperature range and state of charge range that the unit cells may encounter in actual operation. Record the terminal voltage, battery temperature, state of charge and actual current under each test condition to form a test dataset. Based on the experimental dataset, with terminal voltage, battery temperature, and state of charge as input variables and actual current as output variable, a multinomial regression method is used to fit the mapping relationship between input and output variables to obtain a current allocation inversion model. The terminal voltage, battery temperature, and state of charge of the cell are then input into the current allocation inversion model to obtain the corresponding estimated current.

7. The method for identifying the health of unit cells in a hybrid battery pack according to claim 6, characterized in that, The process of obtaining the estimated current and generating the corresponding current distribution characteristic parameters based on the pre-built current distribution inversion model includes: For each cell in the parallel branch, the terminal voltage, cell temperature and state of charge of the cell are input into the current distribution inversion model. The current distribution inversion model outputs the estimated current of the cell. The sum of the estimated currents of all cells in the parallel branch is equal to the branch current of the corresponding parallel branch. Then, current distribution characteristic parameters are generated based on the estimated current.

8. The method for identifying the health of unit cells in a hybrid battery pack according to claim 7, characterized in that, The process of generating the health status of a unit cell based on its health characteristic sequence and current distribution characteristic parameters includes: The capacity retention rate is obtained based on the capacity decay in the health feature sequence. The relative internal resistance level of each cell is obtained based on the current distribution feature parameters of the cell. The current distribution health contribution is generated based on the relative internal resistance level. The voltage-state-of-charge curve of each cell in the current charge-discharge cycle is dynamically time-normalized and compared with the corresponding reference voltage-state-of-charge curve to obtain the curve shape difference. Based on the curve shape difference, the curve shape health contribution is generated. The cell health is obtained by comprehensively considering the capacity retention rate, current distribution health contribution, and curve shape health contribution.

9. The method for identifying the health of unit cells in a hybrid battery pack according to claim 8, characterized in that, The process of identifying abnormal battery cells based on the health status of each cell includes: Obtain the health status of all cells in the hybrid battery pack, generate the average and standard deviation of the health status of all cells, set a health status deviation threshold based on the average and standard deviation of the health status of all cells, compare the health status of each cell with the health status deviation threshold, and mark the corresponding cell as a suspected abnormal cell when the health status of a cell is less than the health status deviation threshold. A second confirmation is performed on suspected abnormal cell batteries. The terminal voltage and battery temperature of the suspected abnormal cell battery within the most recent preset time period are retrieved, as well as the terminal voltage and battery temperature of other cells in the same parallel branch within the same time period, to form the corresponding terminal voltage curve and battery temperature curve respectively. The terminal voltage curve and battery temperature curve of the suspected abnormal cell are compared with the terminal voltage curve and battery temperature curve of other cells in the same parallel branch. When the deviation between the terminal voltage curve of a suspected abnormal cell and the terminal voltage curve of other cells in the parallel branch is continuously greater than a preset voltage deviation threshold, and the deviation between the cell temperature curve and the cell temperature curve of other cells in the parallel branch is continuously greater than a preset temperature deviation threshold, the suspected abnormal cell is marked as an abnormal cell.

10. The method for identifying the health of unit cells in a hybrid battery pack according to claim 9, characterized in that, The process of outputting the health status of the unit battery and generating alarm information for abnormal unit batteries includes: The health status and sorting results of each battery cell are output to the display terminal in list form. The health status is classified into corresponding levels according to the health status of the battery cells. Alarm information is generated for the battery cells marked as abnormal and pushed to the battery management system. The alarm information includes the health status value, health level, abnormality type and handling suggestions of the abnormal battery cells. The abnormality types include capacity decay abnormality, current distribution abnormality, curve shape deviation abnormality, and temperature abnormality. Capacity decay abnormality is triggered when the capacity retention rate is lower than a preset capacity threshold. Current distribution abnormality is triggered when the current distribution health contribution is lower than a preset current distribution threshold. Curve shape deviation abnormality is triggered when the curve shape health contribution is lower than a preset shape threshold. Temperature abnormality is triggered when the temperature rise characteristic parameter exceeds a preset temperature rise threshold. The proposed solutions include replacement recommendations, equalization maintenance recommendations, and reduced power operation recommendations. Replacement recommendations are given for abnormal cell units exhibiting abnormal capacity decay or abnormal curve deviation. Equalization maintenance recommendations are given for abnormal cell units with abnormal current distribution. Reduced power operation recommendations are given for abnormal cell units with abnormal temperature. When multiple abnormal types exist for the same abnormal cell unit, the final solution is determined by prioritizing replacement recommendations over equalization maintenance recommendations, and equalization maintenance recommendations over reduced power operation recommendations.

Images