Battery management system, battery pack, energy storage system and battery management method

Abstract

A battery management system according to the present disclosure includes a voltage measurement circuit and a control unit. The voltage measurement circuit is configured to generate voltage signals indicating the cell voltage of each of a plurality of serially connected battery cells. The control unit determines an observation matrix including a plurality of observation voltage vectors, the plurality of observation voltage vectors indicating the voltage history of each of the plurality of battery cells in a moving window having a predetermined size. The control unit determines a recovery matrix including a plurality of recovery voltage vectors corresponding to the plurality of observation voltage vectors in a one-to-one relationship. The control unit detects an abnormality of each of the plurality of battery cells based on a plurality of absolute error vectors indicating the difference between the plurality of observation voltage vectors and the plurality of recovery voltage vectors.

Description

Technical Field

[0001] This application claims the benefit of Korean Patent Application No. 10-2020-0087749, filed with the Korean Intellectual Property Office on Jul. 15, 2020, the disclosure of which is incorporated herein by reference in its entirety.

[0002] This disclosure relates to the abnormal detection of battery cells. Background Art

[0003] Recently, the demand for portable electronic products such as laptop computers, video cameras, and mobile phones has been rapidly increasing, and with the widespread development of electric vehicles, energy storage accumulators for energy storage, robots, and satellites, many studies are being conducted on high-performance batteries that can be repeatedly charged and discharged.

[0004] Currently, commercially available batteries include nickel-cadmium batteries, nickel-metal hydride batteries, nickel-zinc batteries, lithium batteries, etc., and among them, lithium batteries have little or no memory effect, so they are attracting more attention than nickel-based batteries due to their advantages of being able to be recharged conveniently, having a very low self-discharge rate, and having a high energy density.

[0005] Recently, with the widespread popularity of applications that require high voltage, battery packs including a plurality of battery cells connected in series have been widely used. As the number of battery cells included in the battery pack increases, the possibility of an abnormality occurring in the battery cell increases. Therefore, the need for a diagnostic technique for accurately detecting an abnormality in a battery cell has grown.

[0006] The prior art monitors cell information (e.g., voltage, current, temperature) including a plurality of parameters associated with the state of a battery cell, and detects an abnormality in the battery cell based on the operating state of the battery cell (e.g., charging, discharging, resting) and the monitored cell information.

[0007] However, this abnormal detection method requires a battery management system (BMS) to monitor the cell information of the battery cell using a plurality of sensors, so abnormal detection requires a large amount of calculation and a long time. In particular, in a structure in which power is supplied from the battery cell to the BMS, electrical energy of the battery cell may always be consumed during the operation of the BMS for abnormal detection.

[0008] In addition, the prior art detects an abnormality in a battery cell based on a rapid change in the cell information of the battery cell in a short time. However, the cell information of a defective battery cell does not always change rapidly in a short time, and may tend to change slowly over a long period of time, so an abnormality in the battery cell may not be detected at an appropriate time. Summary of the Invention

[0009] Technical problem

[0010] The present disclosure is designed to solve the above problems, and thus the present disclosure aims to provide a battery management system, a battery management method, a battery pack, and an energy storage system that use the cell voltage of each of a plurality of battery cells included in the battery pack as a single parameter for anomaly detection.

[0011] The present disclosure also aims to provide a battery management system, a battery management method, a battery pack, and an energy storage system that generate an observation matrix as a data set including a plurality of observation voltage vectors indicating changes in the cell voltage of each of a plurality of battery cells, use at least one of a plurality of principal components of the observation matrix to recover the observation matrix, and accurately detect anomalies in each battery cell based on the difference between the data set before recovery and the data set after recovery.

[0012] These and other objects and advantages of the present disclosure can be understood from the following description and will be apparent from the embodiments of the present disclosure. Additionally, it will be readily understood that the objects and advantages of the present disclosure can be achieved by the means and combinations thereof set forth in the appended claims.

[0013] Technical solution

[0014] A battery management system according to one aspect of the present disclosure is configured to detect anomalies in each of a plurality of battery cells connected in series. The battery management system includes a voltage measurement circuit and a control unit. The voltage measurement circuit is configured to generate a voltage signal indicating the cell voltage of each battery cell. The control unit is configured to determine an observation matrix including a plurality of observation voltage vectors, the plurality of observation voltage vectors indicating the voltage history of each of a plurality of battery cells measured multiple times in time series within a moving window having a predetermined size. The control unit is configured to determine a recovery matrix including a plurality of recovery voltage vectors corresponding to the plurality of observation voltage vectors in a one-to-one relationship. The control unit is configured to determine a plurality of absolute error vectors indicating the difference between the plurality of observation voltage vectors and the plurality of recovery voltage vectors. The control unit is configured to detect anomalies in each of the plurality of battery cells based on the plurality of absolute error vectors.

[0015] The control unit can be configured to extract a first sub-matrix, a second sub-matrix, and a third sub-matrix from the observation matrix using a matrix decomposition algorithm. The first sub-matrix is an orthogonal matrix including a plurality of principal component vectors indicating variance information of the observation matrix. The second sub-matrix is a diagonal matrix including a plurality of singular values indicating descriptive factor information for the plurality of principal component vectors of the variance information. The third sub-matrix is an orthogonal matrix including a plurality of coefficient vectors indicating dependence information of a plurality of observed voltage vectors on the plurality of principal component vectors. The control unit can be configured to select at least one of the plurality of singular values using a predetermined recovery condition. The control unit can be configured to transform the first sub-matrix, the second sub-matrix, and the third sub-matrix into a first approximation matrix, a second approximation matrix, and a third approximation matrix based on at least one selected singular value. The control unit can be configured to determine a plurality of recovered voltage vectors by multiplying the first approximation matrix, the second approximation matrix, and the third approximation matrix.

[0016] The control unit can be configured to select each of a threshold number of singular values from the plurality of singular values in descending order.

[0017] The control unit can be configured to: select the first to k-th singular values from the plurality of singular values when the ratio of the k-th singular value to the sum of the plurality of singular values is equal to or greater than a first threshold ratio and the ratio of the u-th singular value to the sum of the plurality of singular values is less than the first threshold ratio. k is a natural number less than n, u is k + 1, n is the total number of battery cells, and the k-th singular value is the k-th largest singular value among the plurality of singular values.

[0018] The control unit can be configured to: select the first to q-th singular values from the plurality of singular values when the ratio of the sum of the first to q-th singular values to the sum of the plurality of singular values is equal to or greater than a second threshold ratio and the ratio of the sum of the first to p-th singular values to the sum of the plurality of singular values is less than the second threshold ratio. q is a natural number greater than or equal to 2 and less than n, p is q - 1, n is the total number of battery cells, and the q-th singular value is the q-th largest singular value among the plurality of singular values.

[0019] The control unit can be configured to: detect each battery cell corresponding to each absolute error vector among the plurality of absolute error vectors that includes at least one component outside a predetermined absolute error range as defective.

[0020] The control unit can be configured to: determine a relative error range based on the plurality of absolute error vectors. The control unit can be configured to: detect each battery cell corresponding to each absolute error vector among the plurality of absolute error vectors that includes at least one component outside the relative error range as defective.

[0021] The control unit may be configured to output an error message indicating an impossible recovery situation when the ratio of the maximum value to the minimum value among a plurality of singular values is less than a preset ratio.

[0022] A battery pack according to another aspect of the present disclosure includes a battery management system.

[0023] An energy storage system according to still another aspect of the present disclosure includes a battery pack.

[0024] A battery management method according to yet another aspect of the present disclosure is used to detect an abnormality in each of a plurality of battery cells connected in series. The battery management method includes: determining an observation matrix including a plurality of observation voltage vectors, the plurality of observation voltage vectors indicating the voltage history of each of the plurality of battery cells measured a plurality of times in time series in a moving window having a predetermined size; determining a recovery matrix including a plurality of recovery voltage vectors corresponding to the plurality of observation voltage vectors in a one-to-one relationship; determining a plurality of absolute error vectors indicating the difference between the plurality of observation voltage vectors and the plurality of recovery voltage vectors; and detecting an abnormality in each of the plurality of battery cells based on the plurality of absolute error vectors.

[0025] Beneficial effect

[0026] According to at least one of the embodiments of the present disclosure, only the cell voltage other than current or temperature is used to detect an abnormality in each of the plurality of battery cells included in a battery pack, thereby reducing the computational amount, time, and power required for abnormality detection.

[0027] According to at least one of the embodiments of the present disclosure, when detecting an abnormality in each battery cell, after generating an observation matrix as a data set including a plurality of observation voltage vectors indicating the time-related change of the cell voltage of each of the plurality of battery cells, and using at least one of the plurality of principal components of the observation matrix to recover the observation matrix, the difference between the data set before recovery and the data set after recovery can be used to accurately detect an abnormality in each of the plurality of battery cells.

[0028] The effects of the present disclosure are not limited to the effects mentioned above, and those skilled in the art will clearly understand these and other effects from the appended claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The drawings illustrate preferred embodiments of the present disclosure and, together with the following detailed description of the present disclosure, are used to provide a further understanding of the technical aspects of the present disclosure. Therefore, the present disclosure should not be construed as being limited to the drawings.

[0030] Figure 1 is a diagram exemplarily showing the configuration of an energy storage system according to the present disclosure.

[0031] Figure 2 It is a graph exemplarily showing the change in the single - cell voltage of a battery cell over time.

[0032] Figure 3 It is in the description as an indication Figure 2 The figure referred to when describing an exemplary observation matrix of a data set indicating the voltage history of the battery cell shown.

[0033] Figure 4 It is in the description as a recovery Figure 3 The figure referred to when describing an exemplary recovery matrix which is the result of the observation matrix.

[0034] Figure 5 The figure referred to when describing the voltage change indicated by the observation matrix and the voltage change indicated by the recovery matrix.

[0035] Figure 6 It is in the description Figure 5 The figure referred to when describing the error between the observation matrix and the recovery matrix shown.

[0036] Figure 7 It is a flowchart exemplarily showing a battery management method according to a first embodiment of the present disclosure.

[0037] Figure 8 It is a flowchart exemplarily showing a battery management method according to a second embodiment of the present disclosure. Detailed Description of the Invention

[0038] Hereinafter, preferred embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. Before the description, it should be understood that the terms or words used in this specification and the appended claims should not be construed as limited to the general and dictionary meanings, but should be interpreted based on the meanings and concepts corresponding to the technical aspects of the present disclosure on the basis of the principle that the inventor is allowed to appropriately define the terms to obtain the best illustration.

[0039] Therefore, the embodiments described herein and the illustrations shown in the drawings are only the most preferred embodiments of the present disclosure, and are not intended to fully describe the technical aspects of the present disclosure. Thus, it should be understood that various other equivalents and modifications may have been made thereto at the time of the filing of this application.

[0040] Terms including ordinal numbers such as "first", "second", etc. are used to distinguish one element from another among various elements, and are not intended to limit the elements by these terms.

[0041] Unless the context clearly indicates otherwise, it should be understood that the term "comprising", as used in this specification, specifies the presence of the stated elements, but does not preclude the presence or addition of one or more other elements. Additionally, as used herein, the term "control unit" refers to a processing unit for at least one function or operation, and this can be implemented by hardware and software alone or in combination.

[0042] Further, throughout the specification, it should be further understood that when an element is referred to as being "connected to" another element, it can be directly connected to the other element or there can be intermediate elements.

[0043] Figure 1 is a diagram exemplarily showing the configuration of the energy storage system 1 according to the present disclosure.

[0044] Reference Figure 1 , the energy storage system 1 includes a battery pack 10, a switch 20, and a power conversion system 30.

[0045] The battery pack 10 includes a positive terminal P+, a negative terminal P−, a cell group 11, and a battery management system 100. The cell group 11 includes a plurality of battery cells BC 1 ~BC n . The reference numeral 'n' is a natural number of 2 or more indicating the total number of battery cells. Hereinafter, when providing a description common to the plurality of battery cells BC 1 ~BC n , the reference numeral "BC" is used to indicate the battery cell.

[0046] The positive and negative terminals of each battery cell BC are electrically coupled to another battery cell BC through a conductor such as a bus bar. The battery cell BC can be a lithium-ion battery cell. The battery cell BC is not limited to a specific type and includes any type of battery cell that can be repeatedly charged.

[0047] The switch 20 is installed on the power line PL of the battery pack 10. When the switch 20 is turned on, power transfer is possible from any one of the battery pack 10 and the power conversion system 30 to the other. The switch 20 can be implemented as at least one of known switching devices such as a relay and a field effect transistor (FET). The control unit 140 can control the on / off of the switch 20 according to the condition of the cell group 11.

[0048] The power conversion system 30 is operatively coupled to the battery management system 100 via the superior controller 2. Operatively coupled means directly / indirectly connected to send and receive signals in one or both directions. The power conversion system 30 can generate DC power for charging the cell group 11 from the AC power supplied by the power grid 40. The power conversion system 30 can generate AC power from the DC power from the battery pack 10.

[0049] The battery management system 100 includes a voltage measurement circuit 110 and a control unit 140. The battery management system 100 may further include at least one of a current sensor 120, a temperature sensor 130, or an interface unit 150.

[0050] The voltage measurement circuit 110 is arranged to be electrically connectable to the positive and negative terminals of each battery cell BC. The voltage measurement circuit 110 is configured to measure the voltage across each battery cell BC and output a signal indicating the measured voltage to the control unit 140.

[0051] The current sensor 120 is serially electrically connected to the cell group 11 via the power line PL. For example, a shunt resistor or a Hall effect device can be used as the current sensor 120. The current sensor 120 is configured to measure the current flowing through the cell group 11 and output a signal indicating the measured current to the control unit 140.

[0052] The temperature sensor 130 is arranged within a predetermined distance range from the cell group 11. For example, a thermocouple can be used as the temperature sensor 130. The temperature sensor 130 is configured to measure the temperature of the cell group 11 and output a signal indicating the measured temperature to the control unit 140.

[0053] The control unit 140 is operatively coupled to the switch 20, the voltage measurement circuit 110, the current sensor 120, the temperature sensor 130, and / or the interface unit 150.

[0054] The control unit 140 can be implemented in hardware using at least one of an application specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), a field programmable gate array (FPGA), a microprocessor, or an electrical unit for performing other functions.

[0055] The control unit 140 may have a memory embedded therein. The memory may pre-store various programs and data necessary for executing the battery management method according to an embodiment as described below. The memory may include at least one type of storage medium such as, for example, flash memory type, hard disk type, solid state drive (SSD) type, silicon disk drive (SDD) type, multimedia card micro type, random access memory (RAM), static random access memory (SRAM), read only memory (ROM), electrically erasable programmable read only memory (EEPROM), or programmable read only memory (PROM).

[0056] The interface unit 150 may be coupled to the upper controller 2 of the energy storage system 1 to enable communication. The interface unit 150 may send messages from the upper controller 2 to the control unit 140 and messages from the control unit 140 to the upper controller 2. Messages from the control unit 140 may include information for notifying an abnormality of each battery cell BC. Communication between the interface unit 150 and the upper controller 2 may use, for example, a wired network such as a local area network (LAN), a controller area network (CAN), and a daisy chain, and / or a short-range wireless network such as Bluetooth, Zigbee, and Wi-Fi. The interface unit 150 may include output devices (e.g., a display, a speaker) to provide the information received from the control unit 140 and / or the upper controller 2 in a recognizable format. The upper controller 2 may control the power conversion system 30 based on the cell information (e.g., the cell voltage, current, temperature, SOC, abnormality of each battery cell) collected by communicating with the battery management system 100.

[0057] The control unit 140 may execute a diagnostic mode for detecting an abnormality of a plurality of battery cells BC 1 ~BC n The period during which the control unit 140 executes the diagnostic mode may be referred to as a "diagnostic period". The control unit 140 may execute the diagnostic mode while the cell group 11 is maintained in a predetermined diagnosable state (e.g., a rest state, a constant current charging state, a constant voltage charging state).

[0058] Figure 2 is a graph exemplarily showing the change in the cell voltage of a battery cell over time, and Figure 3 is a diagram referred to when describing an exemplary observation matrix as a data set indicating the voltage history of the battery cell shown in Figure 2 The control unit 140 determines a plurality of battery cells BC at a preset time interval based on the voltage signal from the voltage measurement circuit 110

[0059] ~BC 1 ~BC nthe voltage value of the individual voltage of each of them, and record the determined voltage value in the memory. As described below, the preset time interval may be equal to the time length of the abnormality detection period (timing).

[0060] The control unit 140 determines an observation matrix X using a moving window 200 of a predetermined size, and the observation matrix X includes a plurality of observation voltage vectors X 1 ~X n , and the plurality of observation voltage vectors X 1 ~X n indicate the changes in the individual voltages of a plurality of battery cells BC 1 ~BC n measured at a preset time interval in the moving window 200. The size of the moving window 200 may be preset or adjustable by the control unit 140.

[0061] In the moving window 200, the individual voltage of the battery cell BC may be measured m times (m is a natural number of 2 or more) in time series by the voltage measurement circuit 110, and the measured individual voltage may be recorded in the memory by the control unit 140. For example, when the size of the moving window 200 = 200 seconds and the time interval = 1 second, m = 200, so the individual voltage of each battery cell BC is measured 200 times in the moving window 200.

[0062] Reference Figure 2 , curve 210 exemplarily shows a plurality of battery cells BC 1 ~BC n among which the individual voltage of the j-th battery cell BC j changes over time. In Figure 2 , t 1 and t m are the start time point and the end time point of the moving window 200 respectively. Curve 210 may be based on the individual voltage measured when the j-th battery cell BC j is in a defective state. The defective state may be, for example, a state that triggers an abnormal behavior of the individual voltage such as an internal short circuit.

[0063] Hereinafter, the abnormality detection operation according to the present disclosure will be described based on the j-th battery cell BC j . The description of the j-th battery cell BC j can be commonly applied to the remaining battery cells BC 1 ~BC n .

[0064] Reference Figure 3, the observation matrix X is an m×n matrix including m rows and n columns. Hereinafter, for convenience of description, it is assumed that m is greater than n, i is a natural number of 1 or more and m or less, and j is a natural number of 1 or more and n or less.

[0065] The n column vectors of the observation matrix X can be in one-to-one correspondence with a plurality of observation voltage vectors X 1 ~X n . That is to say, each of the plurality of observation voltage vectors X 1 ~X n is a column vector of the observation matrix X having m elements (measured cell voltages). The j-th observation voltage vector X j is a time series array of the cell voltages of the j-th battery cell BC j measured m times in the moving window 200, that is, the time series of the measured cell voltages of the j-th battery cell BC j . The j-th observation voltage vector X j can be the j-th column vector of the observation matrix X. Refer to Figure 2 , in the observation matrix X, "x ij " is an element (which can be called "data" or "component") indicating the cell voltage of the j-th battery cell BC j measured at the i-th time in the moving window 200. That is to say, x ij can be the measured cell voltage of the j-th battery cell BC j indexed to the i-th measurement time of the moving window 200.

[0066] The control unit 140 can extract a first sub-matrix A, a second sub-matrix B, and a third sub-matrix C from the observation matrix X using matrix factorization T . The matrix factorization algorithm can include, for example, singular value decomposition (SVD) and principal component analysis (PCA). In this specification, the superscript 'T' on the right side of a matrix indicates a transposed matrix. As shown, the product of multiplying the first sub-matrix A, the second sub-matrix B, and the third sub-matrix C T is equal to the observation matrix X.

[0067] The first sub-matrix A is an m×m matrix. The second sub-matrix B is an m×n matrix. The third sub-matrix C T is an n×n matrix.

[0068] The first sub-matrix A is an orthogonal matrix and includes a plurality of principal component vectors A 1 ~A m . The plurality of principal component vectors A 1 ~A mEach principal component vector in can be called a "left singular vector" and can be a column vector of the first sub-matrix A with m elements. That is, the first sub-matrix A can be expressed as follows.

[0069] A = [A 1 A 2 ...A m , A i = [a 1i a 2i ...a mi T

[0070] Among the multiple principal component vectors A 1 ~A m The principal component vectors A 1 ~A n indicate the variance information of the observation matrix X. The remaining principal component vectors A 1 ~A m may be redundant when describing the variance information of the observation vector X. The j-th principal component vector A n+1 ~A m corresponds to the axial direction with the j-th largest variance of the elements of the observation matrix X. That is, when the elements of the observation matrix X are mapped onto the axes of each of the multiple principal component vectors A j ~A 1 ~A m once, the variance of the elements of the observation matrix X along the axis of the j-th principal component vector A j can be the j-th largest variance.

[0071] As the magnitude of the variance of the j-th principal component vector A j increases, it indicates that the j-th principal component vector A j has a greater descriptive factor for the distribution of the elements of the observation matrix X. As the descriptive factor of the j-th principal component vector A j increases, the j-th principal component vector A j contains a larger amount of information related to the common voltage behavior characteristics (e.g., normal voltage behavior characteristics) of multiple battery cells BC 1 ~BC n in the moving window 200. On the contrary, as the magnitude of the variance of the j-th principal component vector A j decreases, the descriptive factor is lower. That is, the j-th principal component vector A j contains a larger amount of information related to the noise characteristics (e.g., defective state).

[0072] The second sub-matrix B is a diagonal matrix and includes multiple singular values b 11～ b nn ​Elements of the main diagonal. That is, the second sub-matrix B can be expressed as follows.

[0073] B = [B 1 B 2 ...B n , B j = [b 1j b 2j ...b mj T

[0074] where i ≠ j and b ij is 0. b jj is the j-th singular value.

[0075] That is, among the total m×n elements of the second sub-matrix B, the values of the elements except for the n elements on the main diagonal are 0. The singular value b jj indicates the description factor of the j-th principal component vector A j . Multiple singular values b 11 ~b nn can satisfy the following relationship: b 11 ≥b 22 ≥...≥b nn ≥0. That is, multiple singular values b 11 ~b nn can be called the first singular value to the n-th singular value in descending order, and b jj can be the j-th largest singular value among multiple singular values b 11 ~b nn .

[0076] Multiple singular values b 11 ~b nn indicate the description factor information of multiple principal component vectors A 1 ~A n .

[0077] The third sub-matrix C T is an orthogonal matrix and includes multiple coefficient vectors C 1 T ~C n T . Each of the multiple coefficient vectors C 1 T ~C n T can be called a "right singular vector" and can be a row vector of the third sub-matrix C T with n components. The third sub-matrix C T can be expressed as follows.

[0078] C T = [C​1 C 2 ...C n T =[C 1 T ;C 2 T ;...;C n T

[0079] Multiple coefficient vectors C 1 T ~C n T indicate multiple observed voltage vectors X 1 ~X n for multiple principal component vectors A 1 ~A n dependency information. That is, the cell voltage of the j-th battery cell BC j is affected by the j-th principal component vector A 1～ A n among the multiple principal component vectors A j to the extent set by the j-th coefficient vector C j T Set.

[0080] The observation matrix X can be equal to the product of multiplying the first sub-matrix A, the second sub-matrix B, and the third sub-matrix C T and satisfies the relationship of Equation 1 below.

[0081] <Equation 1>

[0082]

[0083] In Equation 1, A j is regarded as an (m×1) matrix, and C j T is regarded as a (1×n) matrix.

[0084] The control unit 140 determines multiple restored voltage vectors X 1 ~X n corresponding to the multiple observed voltage vectors X 1 ~X n in a one-to-one relationship 1 '~X n '.

[0085] Specifically, the control unit 140 selects, from multiple singular values b 11 ~b nn the number of singular values that are 1 or more and less than n and satisfy one or more predetermined restoration conditions. The restoration conditions are used to obtain n singular values b 11 ​​~b nn At least one singular value among them is to be used for approximating (restoring) the observation matrix X. The restoration conditions can be as follows.

[0086] (1) Multiple singular values b in descending order 11 ~b nn The threshold number of singular values among them

[0087] For example, when the threshold number is r, a natural number less than n, the singular values b 11 ~b nn can be selected from the multiple singular values b 11 ~b rr .

[0088] (2) Singular values whose ratio (description factor) to the sum of the multiple singular values b 11 ~b nn is equal to or greater than the first threshold ratio (for example, 0.04)

[0089] For example, k is a natural number less than n, where u = k + 1, b kk / (b 11 +b 22 +...+b nn ) ≥ the first threshold ratio > b uu / (b 11 +b 22 +...+b nn ), the singular values b 11 ~b kk .

[0090] (3) The partial sum of the multiple singular values b in descending order 11 ~b nn and the ratio of the sum of the multiple singular values b 11 ~b nn is equal to or greater than the second threshold ratio (for example, 0.997)

[0091] For example, q is a natural number less than n and 2 or more, where p = q - 1, (b 11 +b 22 +...+b pp +b qq ) / (b 11 +b 22 +...+b nn ) ≥ the second threshold ratio > (b 11 +b 22 +...+b pp ) / (b 11 +b 22 +...+b nn ), the singular values b 11~b qq 。

[0092] When at least one singular value that satisfies at least one of the above restoration conditions (1), (2), and (3) is selected, the control unit 140 can approximate a plurality of observed voltage vectors X 1 ~X n , to determine a plurality of restored voltage vectors X 1 '~X n '。

[0093] Before selecting the singular values that satisfy the (one or more) restoration conditions, the control unit 140 can calculate the maximum value b 11 ~b nn of a plurality of singular values b 11 and the ratio of the minimum value b nn . When the ratio of the maximum value b11 to the minimum value b nn is less than a predetermined ratio (e.g., 200%), the control unit 140 can output an error message indicating an impossible restoration situation. The impossible restoration situation is a situation where there is no explicit difference in the description factors among a plurality of principal component vectors A 1 ~A n . That is, in the impossible restoration situation, none of the plurality of principal component vectors A 1 ~A n sufficiently includes information related to the common voltage behavior characteristics of a plurality of battery cells BC 1 ~BC n . The reason for the impossible restoration situation can be, for example, a failure of the voltage measurement circuit and an abnormality of a number of battery cells BC 1 ~BC n exceeding a predetermined ratio among the plurality of battery cells BC.

[0094] When the ratio of the maximum value b 11 to the minimum value b nn is less than a predetermined ratio, the control unit 140 can increase the size of the moving window 200 by a predetermined time in the next cycle. The reason for increasing the size of the moving window 200 is to sufficiently reflect the common voltage behavior characteristics of a plurality of battery cells BC 1 ~BC n in the observation vector X.

[0095] Hereinafter, when w is a natural number less than n, it is assumed that singular values b 11 ~b nn are selected from a plurality of singular values b 11 ~b ww by the (one or more) restoration conditions. From a plurality of singular values b 11 ~bnn The singular value b selected from 11 ~b ww Each of them can be called a singular value of interest. Multiple singular values b 11 ~b nn Among them, the remaining singular values other than the singular values of interest are singular values that are not very relevant to the common voltage behavior characteristics of multiple battery cells BC 1 ~BC n .

[0096] Figure 4 is a figure referred to when describing an exemplary recovery matrix that is the result of the observation matrix for recovery Figure 3 , Figure 5 is a figure referred to when describing the voltage change indicated by the observation matrix and the voltage change indicated by the recovery matrix, and Figure 6 is a figure referred to when describing Figure 5 the error between the observation matrix and the recovery matrix shown

[0097] Referring to Figure 4 , the control unit 140 can, based on the singular values of interest b 11 ~b ww transform the first sub-matrix A, the second sub-matrix B, and the third sub-matrix C T into the first approximation matrix A', the second approximation matrix B', and the third approximation matrix C T '.

[0098] The first approximation matrix A' is an m×w matrix including each of the first to w principal component vectors A 1 ~A w of the first sub-matrix A as column vectors. That is, the first approximation matrix A' is the result of removing the principal component vectors A w+1 ~A n from the first sub-matrix A

[0099] The second approximation matrix B' is a w×w matrix including the singular values of interest b 11 ~b ww as the main diagonal components. That is, the second approximation matrix B' is the result of changing the remaining main diagonal components of the second sub-matrix B other than the singular values of interest b 11 ~b ww to 0

[0100] The third approximation matrix C T ' is a matrix including the first to w coefficient vectors C T of the third sub-matrix C 1 T ~C w TEach of which is a w×n matrix as a row vector. That is, the third approximation matrix C T ' is obtained by removing the row vectors C T from the third sub-matrix C w+1 T ~C n T .

[0101] The control unit 140 can determine a plurality of restored voltage vectors X T ' based on the first approximation matrix A', the second approximation matrix B', and the third approximation matrix C 1 '~X n '. Specifically, the control unit 140 can determine the restoration matrix X' by multiplying the first approximation matrix A', the second approximation matrix B', and the third approximation matrix C T '. The restoration matrix X' is the result of restoring the observation matrix X using only the first to m principal component vectors A 1 ~A m among which the first to w principal component vectors A 11 ~A ww corresponding to the singular values of interest b 1 ~b w .

[0102] The restoration matrix X' is an m×n matrix and includes a plurality of restored voltage vectors X 1 '~X n ' as column vectors. The restoration matrix X' can satisfy the relationship of Equation 2 below.

[0103] [Equation 2]

[0104]

[0105] In Equation 2, A j is regarded as an (m×1) matrix, and C j T is regarded as a (1×n) matrix.

[0106] The control unit 140 detects an abnormality of the j-th battery cell BC j based on the corresponding observed voltage vector X j and the restored voltage vector X j .

[0107] Reference Figure 5 , curve 510 indicates the change of the cell voltage of the j-th battery cell BC j over time and is equal to Figure 2 curve 210. Curve 520 indicates the j-th battery cell BC indicated by the restored voltage vector X j ' jThe change of the monomer voltage over time.

[0108] The control unit 140 can determine the first to nth absolute error vectors corresponding to the first to n battery monomers BC in a one-to-one relationship. 1 ~BC n The jth absolute error vector corresponds to the difference between the jth observed voltage vector X and the jth restored voltage vector X'. j and the jth restored voltage vector X j '. Figure 6 The curve 610 indicates the jth absolute error vector. That is, the curve 610 indicates the change over time of the difference between the curve 510 and the curve 520.

[0109] The jth absolute error vector 610 includes each of the first to m voltage differences ΔV~ΔV as components. The ith voltage difference ΔV is x - x', that is, the difference between the ith element x of the observed voltage vector X and the ith element x' of the restored voltage vector X'. The ith element x indicates the monomer voltage of the jth battery monomer BC measured at the ith time in the moving window 200. The ith element x' indicates the restoration result of the ith element x. 1j ~ΔV mj in each. The ith voltage difference ΔV ij is x ij -x ij ', that is, the difference between the ith element x of the observed voltage vector X j and the ith element x ij of the restored voltage vector X j '. The ith element x ij indicates the monomer voltage of the jth battery monomer BC measured at the ith time in the moving window 200. The ith element x ij indicates the measurement of the jth battery monomer BC at the ith time in the moving window 200. The ith element x j ' indicates the restoration result of the ith element x ij . ij

[0110] The control unit 140 can determine the relative error range R at each measurement timing of the monomer voltage in the moving window 200 based on the first to nth absolute error vectors. 2 .

[0111] At each measurement timing (e.g., time point t i ), the median of the relative error range R 2 can be the average of the elements of the first to nth absolute error vectors indexed to the corresponding measurement timing. The upper limit of the relative error range R 2 can be the value obtained by adding the product of the standard deviation of the elements of the first to nth absolute error vectors multiplied by a predetermined weight (e.g., 3) to the median. The lower limit of the relative error range R 2 can be the value obtained by subtracting the product of the standard deviation of the components of the first to nth absolute error vectors multiplied by a predetermined weight (e.g., 3) from the median. For example, assuming the weight = 3 and σ i is indexed to the time point t iStandard deviations of components of the first to nth absolute error vectors. At time point t i , relative error range R 2 Median of = (ΔV i1 +ΔV i2 +...+ΔV in ) / n = ΔV i , relative error range R 2 Upper limit of = ΔV i +3σ i , and lower limit of relative error range R 2 = ΔV i -3σ i .

[0112] The control unit 140 can determine whether the jth battery cell BC 1 is defective by comparing each element of the jth absolute error vector with at least one of a predetermined absolute error range R 2 or a relative error range R j .

[0113] In an example, when at least one of the first to m elements ΔV 1j to ΔV mj of the jth absolute error vector is outside the absolute error range R l , the control unit 140 can detect that the jth battery cell BC j is defective. The absolute error range R 1 can be preset to -1.0 to 1.0 mV considering the voltage resolution of the voltage measurement circuit 110.

[0114] In another example, when at least one of the first to m elements ΔV 1j to ΔV mj of the jth absolute error vector is outside the relative error range R 2 , the control unit 140 detects that the jth battery cell BC j is defective.

[0115] In yet another example, when at least one of the first to m elements ΔV 1j to ΔV mj of the jth absolute error vector is simultaneously outside both the absolute error range R 1 and the relative error range R 2 , the control unit 140 can detect that the jth battery cell BC j is defective.

[0116] When an abnormality of at least one battery cell BC is detected, the control unit 140 can perform a predetermined protection operation. For example, when it is determined that the jth battery cell BC jWhen there is a defect, the control unit 140 can turn off the switch 20.

[0117] Figure 7 is a flowchart exemplarily showing a battery management method according to a first embodiment of the present disclosure. The method can be repeated at a predetermined time interval Figure 7 of the method.

[0118] Reference Figures 1 to 7 , in step S710, the control unit 140 determines an observation matrix X including a plurality of observation voltage vectors X 1 ~X n . The plurality of observation voltage vectors X 1 ~X n indicate the voltage history of each of a plurality of battery cells BC 1 ~BC n in a moving window 200 having a predetermined size.

[0119] In step S720, the control unit 140 uses a matrix decomposition algorithm to extract from the observation matrix X a first sub-matrix A including a plurality of principal component vectors A l ~A m , a second sub-matrix B including a plurality of singular values b l ~A m indicating description factors of each of the plurality of principal component vectors A ll ~b nn , and a third sub-matrix C including a plurality of coefficient vectors C 1 T ~C n T . That is, the control unit 140 decomposes the observation matrix X in the form of a multiplication of the first sub-matrix A, the second sub-matrix B, and the third sub-matrix C T by applying a matrix decomposition algorithm to the observation matrix X. T of the observation matrix X.

[0120] In step S730, the control unit 140 selects at least one of the plurality of singular values b 11 ~b nn of the second sub-matrix B. The selection of the singular values can use at least one of the above recovery conditions.

[0121] In step S740, the control unit 140 transforms the first sub-matrix A, the second sub-matrix B, and the third sub-matrix C 11 ~b ww into a first approximation matrix A', a second approximation matrix B', and a third approximation matrix C T ' respectively based on at least one selected singular value b T '.

[0122] In step S750, the control unit 140 determines a recovery matrix X' including a plurality of recovery voltage vectors X' to X' that correspond to a plurality of observation voltage vectors X to X in a one-to-one relationship by multiplying a first approximation matrix A', a second approximation matrix B', and a third approximation matrix C'. T by multiplying a first approximation matrix A', a second approximation matrix B', and a third approximation matrix C' l ~X n corresponding to a plurality of observation voltage vectors X l '~X n 's recovery matrix X'.

[0123] In step S760, the control unit 140 determines a plurality of absolute error vectors indicating the differences between the plurality of observation voltage vectors X to X and the plurality of recovery voltage vectors X' to X'. The plurality of absolute error vectors correspond to the plurality of observation voltage vectors X to X in a one-to-one relationship. 1 ~X n and the plurality of recovery voltage vectors X 1 '~X n '. The plurality of absolute error vectors correspond to the plurality of observation voltage vectors X to X in a one-to-one relationship 1 ~X n .

[0124] In step S762, the control unit 140 determines a relative error range R based on the plurality of absolute error vectors. Since each absolute error vector has m elements, the control unit 140 can determine m relative error ranges R corresponding to the m elements of each absolute error vector, that is, a vector of the relative error range R. Step S762 can be selectively omitted from the method. 2 . Since each absolute error vector has m elements, the control unit 140 can determine m relative error ranges R corresponding to the m elements of each absolute error vector 2 , that is, the relative error range R 2 's vector. Can be from Figure 7 The method of can be selectively omitted step S762.

[0125] In step S770, the control unit 140 determines whether at least one element of each absolute error vector is outside a predetermined absolute error range R and / or a relative error range R. A value of "yes" in step S770 indicates that at least one battery cell BC is detected as defective. For example, as shown in 1 and / or relative error range R 2 outside. A value of "yes" in step S770 indicates that at least one battery cell BC is detected as defective. For example, since at time point t in the moving window 200 as shown in Figure 6 the element ΔV of the j-th absolute error vector 610 i is outside both the absolute error range R ij and the relative error range R 1 and relative error range R 2 both outside, so the control unit 140 can determine that the j-th battery cell BC j is defective.

[0126] In step S780, the control unit 140 activates a predetermined protection operation. In the example, the control unit 140 disconnects the switch 20. In another example, the control unit 140 outputs a diagnostic message indicating information (e.g., identification number) of each battery cell BC detected as defective. The interface unit 150 may send the diagnostic message to the upper controller 2 or output visual information and / or auditory information corresponding to the diagnostic message.

[0127] Figure 8 is a flowchart exemplarily showing a battery management method according to a second embodiment of the present disclosure.

[0128] In Figure 8 the method, steps S810 to S880 are the same as steps S710 to S780 of Figure 7 and redundant descriptions are omitted.

[0129] Figure 8 The method of Figure 7 is different from the method of Figure 8 because the method of

[0130] also includes steps S822 and S824. ll ~b nn Whether the maximum ratio of 11 ~b nn is equal to or greater than a preset ratio is determined by the control unit 140 in step S822. The maximum ratio is the ratio of the maximum value b 11 among the plurality of singular values b nn ~b 1 ~A n to the minimum value b

[0131] When the value of step S822 is "No", it indicates that there is no principal component vector among the plurality of principal component vectors A

[0132] having a description factor large enough to be used in voltage recovery from the observation matrix X to the recovery matrix X'. When the value of step S822 is "No", the method moves to step S824. When the value of step S822 is "Yes", the method moves to step S830.

[0133] Although the present disclosure has been described above with respect to a limited number of embodiments and figures, the present disclosure is not limited thereto, and it will be apparent to those skilled in the art that various modifications and changes can be made thereto within the technical scope of the present disclosure and the equivalent scope of the appended claims.

[0134] Additionally, since those skilled in the art can make many substitutions, modifications, and changes to the present disclosure described above without departing from the technical aspects of the present disclosure, the present disclosure is not limited by the above-described embodiments and figures, and some or all of the embodiments can be selectively combined to allow various modifications.

Claims

1. A battery management system for detecting an abnormality in each of a plurality of serially-connected battery cells, comprising: a voltage measurement circuit configured to generate a voltage signal indicative of a cell voltage of each battery cell; and a control unit configured to determine an observation matrix including a plurality of observation voltage vectors based on the voltage signal, the plurality of observation voltage vectors indicating a voltage history of each of the plurality of battery cells measured a plurality of times in time series in a moving window having a predetermined size, wherein the control unit is configured to: determine a recovery matrix including a plurality of recovery voltage vectors corresponding to the plurality of observation voltage vectors in a one-to-one relationship, determine a plurality of absolute error vectors indicative of a difference between the plurality of observation voltage vectors and the plurality of recovery voltage vectors, and detect an abnormality in each of the plurality of battery cells based on the plurality of absolute error vectors.

2. The battery management system according to claim 1, wherein the control unit is configured to: extract a first sub-matrix, a second sub-matrix, and a third sub-matrix from the observation matrix using a matrix factorization algorithm, the first sub-matrix being an orthogonal matrix including a plurality of principal component vectors indicative of variance information of the observation matrix, the second sub-matrix being a diagonal matrix including a plurality of singular values indicative of description factor information of the plurality of principal component vectors for the variance information, the third sub-matrix being an orthogonal matrix including a plurality of coefficient vectors indicative of dependence information of the plurality of observation voltage vectors on the plurality of principal component vectors, and the control unit is configured to: select at least one of the plurality of singular values using a predetermined recovery condition, transform the first sub-matrix, the second sub-matrix, and the third sub-matrix into a first approximation matrix, a second approximation matrix, and a third approximation matrix based on at least one selected singular value, and determine the plurality of recovery voltage vectors by multiplying the first approximation matrix, the second approximation matrix, and the third approximation matrix.

3. The battery management system according to claim 2, wherein the control unit is configured to select a threshold number of singular values from the plurality of singular values in descending order.

4. The battery management system according to claim 2, wherein the control unit is configured to: select the first to k-th singular values from the plurality of singular values when a ratio of the k-th singular value to a sum of the plurality of singular values is equal to or greater than a first threshold ratio and a ratio of the u-th singular value to the sum of the plurality of singular values is less than the first threshold ratio, and k is a natural number less than n, u is k + 1, n is a total number of the plurality of battery cells, and the k-th singular value is the k-th largest singular value among the plurality of singular values.

5. The battery management system according to claim 2, wherein The control unit is configured to: select the first to q-th singular values from the plurality of singular values when the ratio of the sum of the first to q-th singular values to the sum of the plurality of singular values is equal to or greater than a second threshold ratio and the ratio of the sum of the first to p-th singular values to the sum of the plurality of singular values is less than the second threshold ratio, and q is a natural number greater than or equal to 2 and less than n, p is q - 1, n is the total number of the plurality of battery cells, and the q-th singular value is the q-th largest singular value among the plurality of singular values.

6. The battery management system according to claim 1, wherein, the control unit is configured to: detect as defective each battery cell corresponding to each absolute error vector among the plurality of absolute error vectors that includes at least one component outside a predetermined absolute error range.

7. The battery management system according to claim 6, wherein, the control unit is configured to: determine a relative error range based on the plurality of absolute error vectors, and detect as defective each battery cell corresponding to each absolute error vector among the plurality of absolute error vectors that includes at least one component outside the relative error range.

8. The battery management system according to claim 2, wherein, the control unit is configured to output an error message indicating an irrecoverable situation when the ratio of the maximum value to the minimum value among the plurality of singular values is less than a preset ratio.

9. A battery pack including the battery management system according to any one of claims 1 to 8.

10. An energy storage system including the battery pack according to claim 9.

11. A battery management method for detecting an abnormality in each of a plurality of serially connected battery cells, comprising: determining an observation matrix including a plurality of observation voltage vectors indicating the voltage history of each of the plurality of battery cells measured multiple times in time series in a moving window having a predetermined size; determining a recovery matrix including a plurality of recovery voltage vectors corresponding to the plurality of observation voltage vectors in a one-to-one relationship; determining a plurality of absolute error vectors indicating the difference between the plurality of observation voltage vectors and the plurality of recovery voltage vectors; and and detecting an abnormality in each of the plurality of battery cells based on the plurality of absolute error vectors.

Images