Battery cell fault diagnosis method in battery cluster, cloud server and storage medium

Abstract

The invention provides a fault diagnosis method for battery cells in a battery cluster, a cloud server and a storage medium. The fault diagnosis method for the battery cells in the battery cluster comprises the following steps: further calculating a temperature fluctuation parameter and a voltage fluctuation parameter by performing parameter acquisition on the whole battery cluster and each single battery cell; then obtaining corresponding correlation coefficients based on the temperature fluctuation parameter, the voltage fluctuation parameter and the extracted target characteristic value; and obtaining a battery cell diagnosis result based on the normalized correlation coefficient. The fault diagnosis method does not depend on complex models and high computing power, and hardware and labor cost is reduced; the fault diagnosis process is fast and efficient, and maintenance personnel can take measures in time; the method is easy to implement in an existing battery management system, large-scale transformation is not needed, and application and popularization are facilitated.

Description

Technical Field

[0001] The present application relates to the technical field of battery cell fault diagnosis, and in particular to a battery cell fault diagnosis method within a battery cluster, a cloud server, and a storage medium. Background Art

[0002] Energy storage batteries are mainly used to store and release electrical energy in order to provide electricity when needed; they are widely used in many fields, such as renewable energy storage, electric vehicles, emergency power supplies, grid regulation, etc.

[0003] A battery cluster is usually composed of multiple cells. Connecting multiple cells in series or in parallel can increase the voltage and capacity of the battery cluster to meet the needs of different equipment. However, a cell failure will affect the performance of the entire battery cluster. Timely diagnosis of cell failure is beneficial to maintaining the efficiency and life of the battery cluster, and can help maintenance personnel take timely measures to avoid greater losses.

[0004] However, existing battery cell fault diagnosis methods usually require the use of more complex model algorithms to label the raw data, and then rely on deep learning, artificial intelligence algorithms, etc. This method requires high computing power, large capital investment, and high labor costs. Summary of the invention

[0005] In order to solve the above technical problems, the present application provides a method for diagnosing cell faults in a battery cluster, a cloud server and a storage medium.

[0006] Specifically, the present application provides a method for diagnosing cell faults in a battery cluster, comprising the following steps:

[0007] S100: Collecting parameters of the battery cluster as a whole and each single cell based on the BMS to obtain diagnostic parameters; wherein the diagnostic parameters at least include single cell temperature, single cell voltage and initial characteristic parameters.

[0008] S200: Acquire a temperature fluctuation parameter and a voltage fluctuation parameter based on the single cell temperature and the single cell voltage, respectively.

[0009] S300: extracting a plurality of target characteristic values ​​from the initial characteristic parameter based on the window time, and calculating a correlation coefficient between each of the target characteristic values ​​and the voltage fluctuation parameter and the temperature fluctuation parameter.

[0010] S400: normalizing the same type of correlation coefficients, determining the type of battery cell fault to which the normalized correlation coefficients belong, and obtaining a battery cell diagnosis result based on the determination result.

[0011] In the above technical scheme, compared with the existing methods, the fault diagnosis method of the present application does not rely on complex models and high computing power, and can be implemented in a relatively simple environment; since it reduces the dependence on deep learning and artificial intelligence algorithms, it reduces hardware and labor costs, making fault diagnosis more economical; real-time parameter collection and simple fluctuation analysis make the fault diagnosis process fast and efficient, making it convenient for maintenance personnel to take timely measures; the method is easy to implement in existing battery management systems, without the need for large-scale modifications, and is easy to promote and apply.

[0012] Furthermore, after S100, the method further includes: performing data cleaning on the diagnostic parameters.

[0013] In the above technical solution, data cleaning can remove noise and erroneous data, improve the accuracy of parameters, and make subsequent analysis more reliable.

[0014] Furthermore, after completing data cleaning, it also includes:

[0015] It is determined whether the single cell temperature and the single cell voltage correspond to each other. If so, the process proceeds to S200 ; otherwise, the process proceeds to S200 after performing temperature interpolation calculation on the single cell temperature.

[0016] In the above technical solution, by judging the correspondence between temperature and voltage, the consistency between data is ensured and the possibility of misdiagnosis caused by mismatched data is reduced; if mismatch is found, the temperature data is corrected by temperature interpolation calculation, which can more accurately reflect the actual working state of the battery cell, thereby improving the accuracy of fault diagnosis.

[0017] Further, the target characteristic value includes at least a single cell characteristic value and a battery cluster characteristic value; the step of extracting multiple target characteristic values ​​from the initial characteristic parameter based on the window time in S300 includes:

[0018] Based on the window time, a first feature extraction and a second feature extraction are performed on the single battery cell and the battery cluster respectively to obtain a single battery cell feature value and a battery cluster feature value.

[0019] In the above technical solution, multi-level feature extraction enables a more in-depth analysis of the status of the battery cluster as a whole and each single cell, which helps to identify potential risks and failure modes.

[0020] Furthermore, the first feature extraction specifically includes:

[0021] A first voltage value having a cycle number not greater than a preset cycle number is obtained based on the window time, and a second voltage value is obtained based on the window time; wherein the first voltage value includes at least a maximum voltage value, a minimum voltage value, a voltage value before charging, and a voltage value after charging; and the second voltage value includes at least a voltage mean, a voltage mode, and a voltage median.

[0022] In the above technical solution, the first voltage value covers multiple key parameters such as the maximum, minimum, and voltages before and after charging, which helps to fully reflect the working status and charging and discharging process of the battery cell; the obtained second voltage value (such as mean, mode and median) provides a multi-dimensional perspective on voltage distribution, making the analysis more detailed and accurate; the limitation on the number of cycles allows the analysis to focus on a specific working cycle, which helps to discover the trend of battery performance changes over time.

[0023] Furthermore, the second feature extraction specifically includes:

[0024] The target current value, target SOC and battery capacity are obtained based on the window time; wherein the target current value includes at least the average non-zero current, the average charging current and the average discharging current; the target SOC includes at least the SOC before charging, the SOC after charging, the SOC after discharging, the charging SOC change, the discharging SOC change, the window cycle number and the cumulative cycle number; the battery capacity includes at least the charging capacity and the discharging capacity.

[0025] In the above technical scheme, by extracting the average current value (including non-zero current, charging current, and discharging current), the influence of the current behavior of the battery cluster under different operating conditions on the single battery can be fully understood, and the battery usage status can be comprehensively evaluated; including the SOC before and after charging, the SOC after discharging, etc., it can accurately track the changes in the state of charge of the battery cluster and its charging and discharging process, which is helpful to monitor the health status and efficiency of the battery; by obtaining the charging capacity and discharging capacity, the actual capacity and remaining capacity of the battery can be accurately calculated, providing a reliable basis for the evaluation of battery performance.

[0026] Furthermore, the acquisition of the window cycle number and the cumulative cycle number specifically includes:

[0027] A data sample is selected based on the SOC before charging, the SOC after charging or the SOC after discharging, and the number of valid frames of the data sample and the timing number of the current window time are obtained; the number of window cycles is obtained based on the data sample, the number of valid frames and the timing number; and the number of window cycles is accumulated and summed based on multiple timing numbers to obtain the cumulative number of cycles.

[0028] In the above technical solution, by screening valid data samples, the accuracy of the cycle count can be ensured to avoid erroneous counting caused by invalid data; by accumulating and summing the cumulative number of cycles of multiple time series numbers, the overall usage status of the battery can be evaluated more scientifically, supporting the formulation of reasonable maintenance strategies.

[0029] Further, the cell fault type includes at least a first fault type and a second fault type; the cell diagnosis result includes at least a first fault diagnosis result and a second fault diagnosis result. The cell fault type to which the normalized correlation coefficient is determined in S400, so as to obtain the cell diagnosis result based on the determination result, specifically includes:

[0030] The normalized correlation coefficient is used as the diagnostic coefficient; if the diagnostic coefficient is in the first interval, the single cell corresponding to the diagnostic coefficient is biasedly marked; if the diagnostic coefficient is in the second interval, the single cell corresponding to the diagnostic coefficient is determined to belong to the first fault type, and a first fault diagnosis result corresponding to the first fault type is obtained; if the diagnostic coefficient is in the third interval, the single cell corresponding to the diagnostic coefficient is determined to belong to the second fault type, and a second fault diagnosis result corresponding to the first fault type is obtained.

[0031] In the above technical solution, by defining the first fault type and the second fault type, the classification of battery cell faults is made more systematic, which is convenient for targeted diagnosis and processing; the normalized correlation coefficient is used as a diagnostic coefficient, which can effectively reduce the impact caused by different data scales, thereby improving the accuracy of diagnosis; this classification and judgment method can be applied to battery systems of different types and sizes, enhancing the flexibility and adaptability of the system.

[0032] Furthermore, it should be noted that the diagnostic parameters are not limited to the single cell temperature, single cell voltage and initial characteristic parameters. Those skilled in the art can also use corresponding sensors to measure, for example, the internal stress and internal pressure of the single cell, so as to derive the maximum pressure value, the minimum pressure value, the pressure before or after charging or discharging, etc. Through multi-dimensional data collection, the types of diagnostic results can be divided more finely in the future, which is conducive to more comprehensive clustering and maintenance.

[0033] More preferably, when the number of battery clusters is large, clustering methods such as hierarchical clustering and K-means+ can be used to cluster the diagnostic coefficients of abnormal batteries so as to give diagnostic results in batches.

[0034] Based on the same concept, the present application also provides a cloud server, which uses the battery cell fault diagnosis method in a battery cluster to perform fault diagnosis on each single battery cell in the battery cluster to obtain corresponding battery cell diagnosis results.

[0035] In the above technical solution, the cloud server can centrally manage the battery cells in multiple battery clusters, facilitate unified monitoring and fault diagnosis, and improve management efficiency; the cloud server has good scalability and can increase computing resources as needed to adapt to the diagnostic needs of more battery clusters; the cloud server can also provide a friendly user interface, making it easier for maintenance personnel to access and understand the diagnostic results of the battery cells, thereby improving the operating experience.

[0036] Based on the same concept, the present application also provides a computer-readable storage medium storing a computer program, which implements the method for diagnosing faults of battery cells in a battery cluster when executed by the cloud server.

[0037] Compared with the prior art, the beneficial effects of this application are:

[0038] The present application further calculates the temperature fluctuation parameters and the voltage fluctuation parameters by collecting parameters of the battery cluster as a whole and each single cell; then obtains the corresponding correlation coefficient based on the temperature fluctuation parameters and the voltage fluctuation parameters and the extracted target characteristic values; and then obtains the cell diagnosis result based on the normalized correlation coefficient.

[0039] The fault diagnosis method of the present application does not rely on complex models and high computing power, which reduces hardware and labor costs; and the fault diagnosis process is fast and efficient, which facilitates maintenance personnel to take timely measures; the method is easy to implement in existing battery management systems, does not require large-scale modifications, and is easy to promote and apply. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 This is a flow chart of the method for diagnosing faults of cells in a battery cluster described in this application.

[0041] Figure 2 It is a schematic diagram of the normalized correlation coefficients of the voltage fluctuation parameter and the temperature fluctuation parameter described in the embodiment.

[0042] Figure 3 It is a schematic diagram of the normalized correlation coefficient between the voltage fluctuation parameter and the maximum voltage value described in the embodiment.

[0043] Figure 4 It is a schematic diagram of the normalized correlation coefficient between the voltage fluctuation parameter and the minimum voltage value described in the embodiment.

[0044] Figure 5 It is a schematic diagram of the normalized correlation coefficient between the voltage fluctuation parameter and the daily average discharge current described in the embodiment.

[0045] Figure 6 It is a schematic diagram of the normalized correlation coefficient between the voltage fluctuation parameter and the number of window cycles described in the embodiment.

[0046] Figure 7 It is a schematic diagram of the normalized correlation coefficient between the voltage fluctuation parameter and the cumulative number of cycles described in the embodiment. DETAILED DESCRIPTION

[0047] The following is a further detailed description of a method for diagnosing faults of cells in a battery cluster, a cloud server and a storage medium of the present application in conjunction with specific embodiments and drawings.

[0048] See also Figure 1 The present application provides a method for diagnosing cell faults in a battery cluster, comprising the following steps:

[0049] S100: Collecting parameters of the battery cluster as a whole and each single cell based on the BMS to obtain diagnostic parameters; wherein the diagnostic parameters at least include single cell temperature, single cell voltage and initial characteristic parameters.

[0050] In some embodiments, the number of single cells in the battery cluster is 240 (there are 5 PACKs in the battery cluster, and each PACK has 48 single cells), and the window time is set to 24 hours.

[0051] Based on BMS, the timing voltage, current, temperature, etc. in the battery cluster are collected with high precision, and simple real-time calculations are performed on the BMS side or the SOC, cumulative discharge power and cumulative charge power are collected and uploaded to the cloud server for data calculation.

[0052] In addition, it should be noted that the present application is not limited to 240 single cells, and those skilled in the art can apply the technical solution of the present application to any battery cluster.

[0053] It should be noted that the diagnostic parameters are not limited to the single cell temperature, single cell voltage and initial characteristic parameters. Those skilled in the art can also use corresponding sensors to measure, for example, the internal stress and internal pressure of the single cell to derive the maximum pressure value, the minimum pressure value, the pressure before or after charging or discharging, etc. Through multi-dimensional data collection, the types of diagnostic results can be divided more finely in the future, which is conducive to more comprehensive clustering and maintenance.

[0054] Furthermore, after S100, the method further includes: performing data cleaning on the diagnostic parameters.

[0055] In some embodiments, data cleaning is performed based on the diagnostic parameters collected with high precision or the diagnostic parameters obtained after preliminary calculation by the BMS to remove incomplete data frames and obtain complete and valid data frames.

[0056] Furthermore, after completing data cleaning, it also includes:

[0057] It is determined whether the single cell temperature and the single cell voltage correspond to each other. If so, the process proceeds to S200 ; otherwise, the process proceeds to S200 after performing temperature interpolation calculation on the single cell temperature.

[0058] In some embodiments, performing temperature interpolation calculation on the temperature of the single cell includes:

[0059] The arrangement order of the single cells in the battery cluster is determined according to the actual flow direction of the current in the battery cluster where the single cells are located; then the temperature measurement points are marked as known temperature measurement points and unknown temperature measurement points respectively, the actual flow path distance between any known temperature measurement point and the unknown temperature measurement point is obtained, and the weight coefficient is calculated according to the distance; when it is determined that the distance between the known temperature measurement point and the unknown temperature measurement point meets the preset interpolation distance threshold, a preset rule interpolation operation is performed according to the temperature of the known temperature measurement point and the weight coefficient, and the temperature value of the unknown temperature measurement point is calculated, and the temperature value is the temperature of the single cell after the temperature interpolation calculation.

[0060] It should be noted that those skilled in the art may also use other temperature interpolation methods and are not limited to the above one.

[0061] S200: Acquire a temperature fluctuation parameter and a voltage fluctuation parameter based on the single cell temperature and the single cell voltage, respectively.

[0062] In some embodiments, multi-frame voltage parameters and temperature parameters of all single cells are collected based on the window time, and then the fluctuation parameters are obtained based on the range or standard deviation corresponding to the multi-frame voltage parameters and temperature parameters of the multiple single cells within the window time; each fluctuation parameter is further normalized by normal distribution to obtain multiple standard fluctuation parameters; finally, the obtained standard fluctuation parameters are checked for quality, so that when the test passes, the final temperature fluctuation parameters and voltage fluctuation parameters can be obtained based on each standard fluctuation parameter.

[0063] It should be noted that those skilled in the art may also use other fluctuation parameter calculation methods and are not limited to the above one.

[0064] S300: extracting a plurality of target characteristic values ​​from the initial characteristic parameter based on the window time, and calculating a correlation coefficient between each of the target characteristic values ​​and the voltage fluctuation parameter and the temperature fluctuation parameter.

[0065] Further, the target characteristic value includes at least a single cell characteristic value and a battery cluster characteristic value; the step of extracting multiple target characteristic values ​​from the initial characteristic parameter based on the window time in S300 includes:

[0066] Based on the window time, a first feature extraction and a second feature extraction are performed on the single battery cell and the battery cluster respectively to obtain a single battery cell feature value and a battery cluster feature value.

[0067] Furthermore, the first feature extraction specifically includes:

[0068] A first voltage value having a cycle number not greater than a preset cycle number is obtained based on the window time, and a second voltage value is obtained based on the window time; wherein the first voltage value includes at least a maximum voltage value, a minimum voltage value, a voltage value before charging, and a voltage value after charging; and the second voltage value includes at least a voltage mean, a voltage mode, and a voltage median.

[0069] In some embodiments, the number of cycles is preferably 1, and those skilled in the art may also select a suitable number of cycles according to actual application requirements without being limited thereto; wherein, the number of cycles 1 includes the number of times the charging and discharging processes cannot be greater than 2 at the same time.

[0070] The above first voltage value and the second voltage value can be obtained based on the timing voltage initially collected by the BMS, and the above voltage values ​​all correspond to single cells.

[0071] Furthermore, the second feature extraction specifically includes:

[0072] The target current value, target SOC and battery capacity are obtained based on the window time; wherein the target current value includes at least the average non-zero current, the average charging current and the average discharging current; the target SOC includes at least the SOC before charging, the SOC after charging, the SOC after discharging, the charging SOC change, the discharging SOC change, the window cycle number and the cumulative cycle number; the battery capacity includes at least the charging capacity and the discharging capacity.

[0073] In some embodiments, the above target current value can be obtained by the current collected by the BMS, and the above partial target SOC and battery capacity can be obtained based on the calculated or collected SOC, cumulative discharge current and cumulative charge capacity; and the above target current value, target SOC and battery capacity all correspond to the battery cluster as a whole.

[0074] Furthermore, the acquisition of the window cycle number and the cumulative cycle number specifically includes:

[0075] A data sample is selected based on the SOC before charging, the SOC after charging or the SOC after discharging, and the number of valid frames of the data sample and the timing number of the current window time are obtained; the number of window cycles is obtained based on the data sample, the number of valid frames and the timing number; and the number of window cycles is accumulated and summed based on multiple timing numbers to obtain the cumulative number of cycles.

[0076] In some embodiments, the “window cycle number” can be obtained by using the change value of SOC within the selected window time. When SOC is expressed in percentage, the window cycle number of the currently selected data sample or data sample group can be calculated by the following formula:

[0077]

[0078] Wherein, N represents the number of valid frames in the selected window time, and m is the timing number of the current window.

[0079] The "cumulative number of cycles" is the cumulative sum of all "window cycle numbers" before and after that time node.

[0080]

[0081] Furthermore, the correlation coefficient between each of the target characteristic values ​​and the voltage fluctuation parameter and the temperature fluctuation parameter is calculated.

[0082]

[0083] Among them, X and Y represent two parameters, X i Represents the samples in parameter X, Y i represents the sample in parameter Y, ρ X,Y represents the correlation coefficient between X and Y. represents the sample mean of parameter X, Represents the sample mean of parameter Y.

[0084] For example, the correlation coefficient between the voltage fluctuation parameter and the temperature fluctuation parameter of a single cell is calculated: Vf- Tf i,i∈[1,240].

[0085] The correlation coefficient between the voltage fluctuation parameter of a single cell and the maximum voltage value of the single cell: ρ Vf-Vmax i,i∈[1,240].

[0086] The correlation coefficient between the voltage fluctuation parameter of a single cell and the minimum voltage value of a single cell: ρ Vf-Vmin i,i∈[1,240].

[0087] The correlation coefficient between the voltage fluctuation parameter of a single cell and the average discharge current of the battery cluster as a whole:

[0088] The correlation coefficient between the voltage fluctuation parameter of a single cell and the window cycle number of the battery cluster as a whole: ρ Vf-window-Cycle i,i∈[1,240].

[0089] Correlation coefficient between the voltage fluctuation parameter of a single cell and the cumulative number of cycles of the battery cluster as a whole: ρ Vf-cumulative-Cycle i,i∈[1,240].

[0090] Each single cell obtained 6 correlation coefficients, and a total of 1440 correlation coefficients were obtained for 240 cells.

[0091] S400: normalizing the same type of correlation coefficients, determining the type of battery cell fault to which the normalized correlation coefficients belong, and obtaining a battery cell diagnosis result based on the determination result.

[0092] In some embodiments, the same correlation coefficient of each single cell is normalized:

[0093]

[0094] For example, the battery cluster has 240 (i.e., Q in the formula) single cells. The fluctuation parameters and minimum values ​​of the 240 cells are calculated by the methods of steps S200 and S300, and are recorded as V, V Tf i,i∈[1,240], then we get 240 sets of data. For each set of data, we use the above correlation coefficient formula to calculate the corresponding correlation coefficient and get ρ Vf-Tf i,i∈[1,240], for ρ Vf- Tf i is normalized according to the following formula:

[0095]

[0096] At this time, the correlation coefficient between the normalized voltage fluctuation parameter and the temperature fluctuation parameter of each single cell is obtained: ρ Vf-Tf Zi,i∈[1,240], similarly we get: Vf-Vmax Zi,i∈[1,240],ρ Vf-Vmin Zi,i∈[1,240], ρ Vf-window-Cycle Zi,i∈[1,240],ρ Vf-cumulative-Cycle Zi,i∈[1,240].

[0097] Further, the cell fault type includes at least a first fault type (i.e., type D) and a second fault type (i.e., type U); the cell diagnosis result includes at least a first fault diagnosis result (corresponding to the diagnosis result of type D) and a second fault diagnosis result (corresponding to the diagnosis result of type U). The cell fault type to which the normalized correlation coefficient belongs is determined in S400, so as to obtain the cell diagnosis result based on the determination result, specifically including:

[0098] The normalized correlation coefficient is used as the diagnostic coefficient; if the diagnostic coefficient is in the first interval, the single cell corresponding to the diagnostic coefficient is biasedly marked; if the diagnostic coefficient is in the second interval, the single cell corresponding to the diagnostic coefficient is determined to belong to the first fault type, and a first fault diagnosis result corresponding to the first fault type is obtained; if the diagnostic coefficient is in the third interval, the single cell corresponding to the diagnostic coefficient is determined to belong to the second fault type, and a second fault diagnosis result corresponding to the first fault type is obtained.

[0099] In some embodiments, the normalized correlation coefficient of each single cell is obtained, and the single cells whose coefficient is outside the interval (-R, R) are marked as biased, where the single cells in the interval (-∞, -R] are marked as ρ Vf-Vmin D type, the interval [R, +∞) is denoted as ρ Vf-Vmin U class, where ρ Vf-Vmin represents the correlation coefficient, D means it is significantly lower than the average level, and U means it is significantly higher than the average level.

[0100] Although according to the definition of the correlation coefficient, when the absolute value of the correlation coefficient is less than 0.2 or 0.1, there is no obvious correlation between the two, or a very weak correlation is shown, but due to the increase in the amount of data, it tends to be stable. For example, the maximum voltage often appears at the end of charging, and it often shows transient characteristics and disappears quickly, but the voltage fluctuation is also positively correlated with it. Therefore, after the above normalization, a cluster sample with an obvious correlation bias can still be obtained in a smaller correlation coefficient.

[0101] At this point, based on the definition of correlation, we can analyze the reasons why some cells fluctuate too much, for example:

[0102] When the correlation coefficient between the normalized voltage fluctuation and the average discharge current of a single cell belongs to category U, the correlation between the fluctuation and the average discharge current of the cell is at a relatively high level within the cluster. In other words, the voltage fluctuation of the cell is closely related to the average discharge current, and the size of the discharge current will obviously affect its fluctuation characteristics.

[0103] When the correlation coefficient between the normalized voltage fluctuation and the average discharge current of a single cell belongs to Class D, the correlation between the fluctuation and the average discharge current of the cell is at a low level within the cluster. In other words, the voltage fluctuation of the cell has a small relationship with the average discharge current, and the size of the average discharge current will not significantly affect its fluctuation characteristics. When the U-type cells in the cluster exceed a certain proportion (for example, 3%), the cluster discharge current should be reduced.

[0104] Similarly, when the correlation coefficient between the normalized voltage fluctuation and the cumulative number of cycles of a single cell belongs to category U, the aging level of the cell is obviously at a higher level within the cluster, otherwise it is at a lower level.

[0105] When the correlation coefficient between the normalized temperature fluctuation and the cumulative number of cycles of a single cell belongs to the U category, the temperature difference level of the cell's location changes over time and is obviously at a higher level within the cluster, otherwise it is at a lower level.

[0106] When the correlation coefficient between the normalized voltage fluctuation and temperature fluctuation of a single cell belongs to category U, the temperature control and heat dissipation of the cell are at a poor level within the cluster, otherwise they are at a good level.

[0107] I won’t go into details about this and more.

[0108] Among them, R is preferably 3, and those skilled in the art can also select other values ​​according to actual application requirements and are not limited to this.

[0109] When diagnosing a single cell, a problematic cell often shows abnormalities in multiple parameters. The diagnosis result is made based on the data calculated based on the "raw data after cleaning the example" provided by the data support and the normalized correlation coefficient calculated based on it.

[0110] like Figure 2 As shown, the results show that the voltage fluctuation parameters and temperature fluctuation parameters in the battery cluster show an obvious regular distribution. Except for the poor temperature control at the location of PACK1, the cells in other PACKs are regularly distributed. This result obviously cannot give diagnostic advice for the cells, but it is of great significance for the optimization of the heat dissipation system and the optimization of the thermal management of PACK1.

[0111] like Figure 3 As shown, considering that the voltage fluctuation parameters and the maximum voltage have little correlation, it is necessary to refer to other coefficients for comprehensive consideration. However, judging from this coefficient alone, the voltage fluctuation of the cell numbered 204 has a smaller correlation with the maximum voltage, and the maximum transient characteristic cluster of cells numbered 204, 171, and 174 is more obvious.

[0112] like Figure 4 As shown, the coefficient of cells 1, 79, and 204 is obviously larger within the cluster. The coefficient is biased and has no PACK regularity, reflecting a discrete characteristic with a proportion of more than 4%. The battery in this cluster needs to adjust the lower voltage threshold to extend its service life.

[0113] like Figure 5 As shown, this coefficient reflects that the impedance of the three battery cells No. 171, No. 174, and No. 204 is relatively small.

[0114] like Figure 6 As shown, it is reflected that the voltage fluctuation of cells 4, 69, and 204 has a high correlation with the number of cycles within the window time. By comparing the correlation between the voltage fluctuation and the maximum voltage value, the minimum voltage value, the discharge current, and the number of cycles within the window time, it can be determined that the capacity of cell 204 should be greater than the cluster average level, and its actual SOC is lower than the cluster average level.

[0115] like Figure 7 As shown, it can be seen that monomer No. 76 and monomer No. 204 have obvious outlier characteristics, among which the aging level of monomer No. 76 is slightly higher than the average level in the cluster, but does not exceed the threshold of R=3. The aging level of monomer No. 204 is significantly lower than the average level in the cluster. Considering the large capacity and low SOC level of monomer No. 204, the aging of monomer No. 76 is significant, and the lower discharge threshold should be temporarily increased.

[0116] Preferably, when there are a large number of battery clusters, clustering methods such as hierarchical clustering and K-means+ can be used to cluster the normalized correlation coefficients of abnormal batteries so as to provide diagnostic results in batches.

[0117] Furthermore, the present application also provides a cloud server, which uses the battery cell fault diagnosis method in a battery cluster to perform fault diagnosis on each single battery cell in the battery cluster to obtain corresponding battery cell diagnosis results.

[0118] Furthermore, the present application also provides a computer-readable storage medium storing a computer program, which implements the method for diagnosing faults of battery cells in a battery cluster when executed by the cloud server.

[0119] In some embodiments, the storage medium stores several computer programs for enabling the cloud server to execute all or part of the steps of the methods described in various embodiments of the present application.

[0120] The storage medium may include a U disk, a mobile hard disk, a read-only memory, a random access memory, a magnetic disk or an optical disk, and other media that can store program codes.

[0121] In summary, the present application provides a method for diagnosing faults of battery cells in a battery cluster, a cloud server, and a storage medium; by collecting parameters of the battery cluster as a whole and each single battery cell, the temperature fluctuation parameters and voltage fluctuation parameters are further calculated; then the corresponding correlation coefficients are obtained based on the temperature fluctuation parameters and voltage fluctuation parameters and the extracted target characteristic values; and the battery cell diagnosis results are obtained based on the normalized correlation coefficients. The fault diagnosis method of the present application does not rely on complex models and high computing power, which reduces hardware and labor costs; and the fault diagnosis process is fast and efficient, which facilitates maintenance personnel to take timely measures; the method is easy to implement in existing battery management systems, does not require large-scale transformation, and is easy to promote and apply.

[0122] Although example embodiments have been described herein with reference to the accompanying drawings, it should be understood that the above example embodiments are merely exemplary and are not intended to limit the scope of the present application to this. Those of ordinary skill in the art may make various changes and modifications therein without departing from the scope and spirit of the present application. All these changes and modifications are intended to be included within the scope of the present application as required by the appended claims.

[0123] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0124] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed.

[0125] The various component embodiments of the present application can be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. It should be understood by those skilled in the art that a microprocessor or digital signal processor (DSP) can be used in practice to implement some or all functions of some modules according to the embodiments of the present application. The application can also be implemented as a device program (e.g., computer program and computer program product) for executing a part or all of the methods described herein. Such a program implementing the present application can be stored on a computer-readable medium, or can have the form of one or more signals. Such a signal can be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

[0126] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0127] Although the description of the present application is carried out in conjunction with the above specific embodiments, it is obvious that those skilled in the art can make many substitutions, modifications and changes based on the above content. Therefore, all such substitutions, improvements and changes are included in the spirit and scope of the appended claims.

Claims

1. A method for diagnosing cell faults in a battery cluster, characterized in that: The following steps are involved: S100: collecting parameters of the battery cluster as a whole and each single cell based on the BMS to obtain diagnostic parameters; wherein the diagnostic parameters at least include single cell temperature, single cell voltage and initial characteristic parameters; S200: Acquire a temperature fluctuation parameter and a voltage fluctuation parameter based on the single cell temperature and the single cell voltage respectively; S300: extracting a plurality of target characteristic values ​​from the initial characteristic parameter based on the window time, and calculating the correlation coefficient between each of the target characteristic values ​​and the voltage fluctuation parameter and the temperature fluctuation parameter; S400: normalizing the same type of correlation coefficients, determining the type of battery cell fault to which the normalized correlation coefficients belong, and obtaining a battery cell diagnosis result based on the determination result.

2. The method for diagnosing cell faults in a battery cluster according to claim 1, characterized in that: After S100, the method further includes: Data cleaning is performed on the diagnostic parameters.

3. The method for diagnosing cell faults in a battery cluster according to claim 2, characterized in that: After completing data cleaning, it also includes: It is determined whether the single cell temperature and the single cell voltage correspond to each other. If so, the process proceeds to S200 ; otherwise, the process proceeds to S200 after performing temperature interpolation calculation on the single cell temperature.

4. The method for diagnosing cell faults in a battery cluster according to claim 1, characterized in that: The target characteristic value includes at least a single cell characteristic value and a battery cluster characteristic value; the step of extracting multiple target characteristic values ​​from the initial characteristic parameters based on the window time in S300 includes: Based on the window time, a first feature extraction and a second feature extraction are performed on the single battery cell and the battery cluster respectively to obtain a single battery cell feature value and a battery cluster feature value.

5. The method for diagnosing cell faults in a battery cluster according to claim 4, characterized in that: The first feature extraction specifically includes: Acquire a first voltage value whose cycle number is not greater than a preset cycle number based on the window time, and acquire a second voltage value based on the window time; Among them, the first voltage value at least includes a maximum voltage value, a minimum voltage value, a voltage value before charging and a voltage value after charging; the second voltage value at least includes a voltage mean, a voltage mode and a voltage median.

6. The method for diagnosing cell faults in a battery cluster according to claim 4, characterized in that: The second feature extraction specifically includes: Acquire a target current value, a target SOC and a battery capacity based on the window time; Among them, the target current value includes at least the average non-zero current, the average charging current and the average discharging current; the target SOC includes at least the SOC before charging, the SOC after charging, the SOC after discharging, the charging SOC change, the discharging SOC change, the window cycle number and the cumulative cycle number; the battery capacity includes at least the charging capacity and the discharging capacity.

7. The method for diagnosing cell faults in a battery cluster according to claim 6, characterized in that: The acquisition of the window cycle number and the cumulative cycle number specifically includes: Selecting a data sample based on the SOC before charging, the SOC after charging, or the SOC after discharging, and obtaining the number of valid frames of the data sample and the timing number of the current window time; Obtaining the number of window cycles based on the data samples, the number of valid frames and the timing number; The number of cycles of each window is accumulated and summed based on multiple time sequence numbers to obtain a cumulative number of cycles.

8. The method for diagnosing cell faults in a battery cluster according to claim 1, characterized in that: The cell fault type includes at least a first fault type and a second fault type; the cell diagnosis result includes at least a first fault diagnosis result and a second fault diagnosis result. The cell fault type to which the normalized correlation coefficient belongs in S400 is determined, and the cell diagnosis result is obtained based on the determination result, specifically including: The normalized correlation coefficient was used as the diagnostic coefficient; If the diagnostic coefficient is in the first interval, a single cell corresponding to the diagnostic coefficient is subjected to a paranoid marking; If the diagnosis coefficient is in the second interval, determining that the single cell corresponding to the diagnosis coefficient belongs to a first fault type, and obtaining a first fault diagnosis result corresponding to the first fault type; If the diagnosis coefficient is in the third interval, it is determined that the single battery cell corresponding to the diagnosis coefficient belongs to the second fault type, and a second fault diagnosis result corresponding to the first fault type is obtained.

9. A cloud server, characterized in that: The cloud server uses the method for diagnosing faults of cells in a battery cluster as described in any one of claims 1 to 8 to perform fault diagnosis on each single cell in the battery cluster to obtain corresponding cell diagnosis results.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by the cloud server, the method for diagnosing faults of cells in a battery cluster as described in any one of claims 1 to 8 is implemented.

Images