Apparatus and methods for diagnosing battery status

Abstract

The apparatus for diagnosing battery state according to embodiments of the present disclosure is an apparatus capable of tracking and diagnosing battery state based on various characteristic values ​​of the battery. According to aspects of the present disclosure, battery state can be diagnosed periodically to prevent accidents caused by battery defects in advance. Furthermore, according to one aspect of the present disclosure, battery state can be tracked and diagnosed based on a curve of the battery obtained in each predetermined cycle. That is, since the time point at which the battery state is diagnosed as abnormal and the battery state at that time can be specified and stored, specific diagnostic information about the battery state can be provided.

Description

Technical Field

[0001] This application claims priority to Korean Patent Application No. 10-2020-0128194, filed in Korea on October 5, 2020, the disclosure of which is incorporated herein by reference.

[0002] This disclosure relates to apparatus and methods for diagnosing battery status, and more specifically, to apparatus and methods for diagnosing battery status. Background Technology

[0003] Recently, demand for portable electronic products such as laptops, cameras, and mobile phones has increased dramatically, and electric vehicles, energy storage batteries, robots, and satellites have seen significant development. Therefore, research is actively underway on high-performance batteries that allow for repeated charging and discharging.

[0004] Currently available batteries include nickel-cadmium batteries, nickel-metal hydride batteries, nickel-zinc batteries, and lithium batteries. Among them, lithium batteries have attracted much attention due to their almost non-existent memory effect compared to nickel-based batteries, as well as their very low self-discharge rate and high energy density.

[0005] Traditionally, battery information such as voltage, current, and / or temperature is obtained, and the battery's condition is diagnosed based on a simple comparison of this information with a standard voltage or standard voltage range. For example, the battery's condition is diagnosed as normal, overvoltage, or undervoltage depending on whether its voltage falls within the standard voltage range.

[0006] However, in reality, since fires can still occur even when the battery's voltage, current, and / or temperature fall within standard ranges, a technique is needed that accurately diagnoses battery condition from a statistical perspective by tracking changes in the battery curve, rather than by simply comparing characteristic values ​​(e.g., voltage values) with reference values ​​(e.g., standard voltage or standard voltage range). Summary of the Invention

[0007] Technical issues

[0008] This disclosure aims to address the problems in the related technologies, and therefore aims to provide apparatus and methods for diagnosing battery status, which can track and diagnose battery status based on various characteristic values ​​of the battery.

[0009] These and other objects and advantages of this disclosure may be understood from the following detailed description and will become more apparent from the exemplary embodiments of this disclosure. Furthermore, it will be readily understood that the objects and advantages of this disclosure may be achieved by the means shown in the appended claims and combinations thereof.

[0010] Technical solution

[0011] An apparatus for diagnosing battery state according to aspects of this disclosure may include: a curve generation unit configured to obtain battery information including at least one of voltage, capacity, internal resistance, SOC, and SOH for each of a plurality of batteries, and to generate a plurality of battery curves for each of the plurality of batteries in each cycle based on the obtained battery information; a feature extraction unit configured to extract a plurality of feature values ​​for each of the plurality of batteries using the battery information and the plurality of battery curves generated by the curve generation unit; a principal component calculation unit configured to calculate a plurality of principal components for the plurality of feature values ​​extracted by the feature extraction unit; and a battery classification unit configured to... The system is configured to calculate the distribution degree of each principal component representing the distribution of the multiple batteries based on multiple feature values ​​extracted for each of the multiple batteries, select at least one of the multiple principal components as a target component based on the result of comparing the calculated distribution degree of each principal component with a preset standard value, and classify each of the multiple batteries into any one of the multiple groups based on the selected at least one target component; and a battery state diagnosis unit is configured to update the classification history of each group to which each of the multiple batteries is classified by the battery classification unit in each cycle, and diagnose the state of each of the multiple batteries based on the updated classification history.

[0012] The battery classification unit can be configured to classify a selected target component into any one of a plurality of groups based on the reference cell corresponding to the anomalous cell being set as the reference cell.

[0013] The battery status diagnostic unit can be configured to set the group to which the reference cell belongs among multiple groups as the abnormal group and the remaining groups as the normal group.

[0014] The classification history can be provided to each of the multiple batteries and can be configured to include the groups classified in the previous cycle, the groups classified in the current cycle, the frequency of changes to the abnormal groups, and the frequency of changes to the normal groups.

[0015] The battery state diagnostic unit can be configured to identify a target battery among a plurality of batteries whose group in the previous cycle is different from the group in the current cycle, and diagnose the state of the target battery as normal or abnormal based on the frequency of change to an abnormal group or the frequency of change to a normal group for the identified target battery.

[0016] The battery status diagnostic unit can be configured to diagnose the status of the target battery as abnormal when the frequency of changes to the abnormal group is equal to or greater than a predetermined frequency, provided that the target battery is classified into an abnormal group in the current cycle.

[0017] The battery status diagnostic unit can be configured to diagnose the target battery as being in a normal state when the frequency of change to the normal group is equal to or greater than a predetermined frequency, provided that the target battery is classified into the normal group in the current cycle.

[0018] The battery classification unit can be configured to select at least one of a plurality of principal components, and select at least one principal component as the target component when the sum of the distribution degrees of each principal component corresponding to the selected at least one principal component is equal to or greater than a preset standard value.

[0019] The battery classification unit can be configured to calculate a sum by sequentially adding the distribution degrees of each principal component corresponding to each of the multiple principal components from the largest one until the sum becomes equal to or greater than a preset standard value.

[0020] The battery classification unit can be configured to generate a representative model in each loop that classifies multiple batteries and a reference cell into multiple groups.

[0021] The battery classification unit can be configured to generate a target set including at least one selected target component, generate multiple target subsets including at least one target component from the generated target set, generate at least one classification model for classifying multiple batteries and a reference cell for each of the multiple target subsets, and set any one of the multiple classification models generated for the multiple target subsets as the representative model of the corresponding loop.

[0022] The battery classification unit can be configured to calculate the classification degree of multiple batteries and reference cells for each of the multiple generated classification models, and set the classification model with the smallest calculated classification degree among the multiple generated classification models as the representative model of the corresponding loop.

[0023] The battery classification unit can be configured to calculate the inner product of multiple batteries and reference cells for at least one target component included in each of multiple target subsets, and generate multiple classification models for classifying the multiple batteries and reference cells based on the size of the multiple calculated inner products.

[0024] The battery classification unit can be configured to control each of multiple classification models such that the total number of groups classified is equal to or less than a preset standard number.

[0025] According to another aspect of this disclosure, a battery pack may include means for diagnosing battery status according to another aspect of this disclosure.

[0026] An energy storage system according to another aspect of this disclosure may include means for diagnosing battery status according to another aspect of this disclosure.

[0027] A method for diagnosing battery state according to another aspect of this disclosure may include: a curve generation step, generating a battery curve for each of a plurality of batteries representing the correspondence between voltage and capacity of each of the plurality of batteries in each cycle; a feature extraction step, extracting a plurality of feature values ​​from each of the plurality of battery curves generated in the curve generation step; a principal component calculation step, calculating a plurality of principal components for the plurality of feature values ​​extracted in the feature extraction step; a principal component distribution degree calculation step, calculating the distribution degree of each principal component representing the distribution of the plurality of batteries based on the plurality of feature values ​​extracted for each of the plurality of batteries; a target component selection step, selecting at least one of the plurality of principal components as a target component based on the result of comparing the distribution degree of each principal component calculated in the principal component distribution degree calculation step with a preset standard value; a battery classification step, classifying each of the plurality of batteries into any one of a plurality of groups based on the at least one target component selected in the target component selection step; and a battery state diagnosis step, updating the classification history of the groups to which each of the plurality of batteries was classified in the battery classification step in each cycle, and diagnosing the state of each of the plurality of batteries based on the updated classification history.

[0028] Beneficial effects

[0029] According to this disclosure, battery status can be periodically diagnosed to prevent accidents caused by battery defects in advance.

[0030] Furthermore, according to one aspect of this disclosure, battery state can be tracked and diagnosed based on the battery curves obtained in each predetermined cycle. That is, since the time point at which the battery state is diagnosed as abnormal and the battery state at that time can be specified and stored, specific diagnostic information about the battery state can be provided.

[0031] The effects of this disclosure are not limited to those described above, and those skilled in the art will clearly understand from the appended claims other effects not mentioned herein. Attached Figure Description

[0032] The accompanying drawings illustrate preferred embodiments of the present disclosure and, together with the foregoing disclosure, serve to provide a further understanding of the technical features of the present disclosure; therefore, the present disclosure should not be construed as limited to the drawings.

[0033] Figure 1 This is a schematic diagram of an apparatus for diagnosing battery status according to an embodiment of the present disclosure.

[0034] Figure 2 This is a schematic diagram of a voltage-differential capacity curve in a battery curve generated by an apparatus for diagnosing battery state according to an embodiment of the present disclosure.

[0035] Figure 3 This is a diagram schematically illustrating an example of the distribution degree of each principal component calculated by an apparatus for diagnosing battery state according to an embodiment of the present disclosure.

[0036] Figure 4 and Figure 5 This is a schematic diagram illustrating a classification model generated in the current cycle by a device for diagnosing battery state according to an embodiment of the present disclosure.

[0037] Figure 6 This diagram schematically illustrates an exemplary configuration of a battery pack including a device for diagnosing battery status according to embodiments of the present disclosure.

[0038] Figure 7 This is a schematic diagram illustrating a method for diagnosing battery status according to another embodiment of the present disclosure. Detailed Implementation

[0039] It should be understood that the terms used in the specification and appended claims should not be construed as limited to their general and dictionary meanings, but rather as being interpreted based on their meanings and concepts corresponding to the technical aspects of this disclosure, on the basis that the inventors are permitted to properly define the terms for the best interpretation.

[0040] Therefore, the description presented herein is merely a preferred example for illustrative purposes only and is not intended to limit the scope of this disclosure. It should be understood that other equivalents and modifications may be made thereto without departing from the scope of this disclosure.

[0041] Furthermore, in describing this disclosure, detailed descriptions are omitted here when it is believed that detailed descriptions of relevant known elements or functions would obscure the key subject matter of this disclosure.

[0042] Ordinal terms such as “first” and “second” can be used to distinguish elements among various elements, but are not intended to restrict elements by terminology.

[0043] Throughout this specification, when a part is referred to as “comprising” or “including” any element, unless otherwise expressly stated, it means that the part may further include other elements without excluding them.

[0044] Furthermore, throughout this specification, when a part is referred to as "connected" to another part, it is not limited to the case where they are "directly connected," but also includes the case where they are "indirectly connected" by means of another element inserted between them.

[0045] Preferred embodiments of the present disclosure will be described in detail below with reference to the accompanying drawings.

[0046] Figure 1 This is a schematic diagram of an apparatus 100 for diagnosing battery status according to an embodiment of the present disclosure.

[0047] refer to Figure 1 According to embodiments of the present disclosure, the apparatus 100 for diagnosing battery status may include a curve generation unit 110, a feature value extraction unit 120, a principal component calculation unit 130, a battery classification unit 140, and a battery status diagnosis unit 150.

[0048] Here, a battery refers to a physically separable, independent unit with a negative terminal and a positive terminal. For example, a pouch-shaped lithium polymer unit can be considered a battery.

[0049] The curve generation unit 110 can be configured to obtain battery information including at least one of the following: voltage, capacity, internal resistance, SOC (state of charge), and SOH (state of health) of each of a plurality of batteries.

[0050] Specifically, the battery information obtained by the curve generation unit 110 can be information obtained during the charging process of the battery.

[0051] For example, the battery voltage and capacity can be measured while the battery is being charged. Furthermore, the battery's state of charge (SOC) can be estimated based on the battery voltage and / or charging current. Additionally, the battery's internal resistance can be estimated based on the battery voltage and SOC. For example, the battery's internal resistance can be estimated based on the EKF (Extended Kalman Filter) and ECM (Equivalent Circuit Model) used for the battery, according to the battery voltage and SOC. Furthermore, the battery's state of equilibrium (SOH) can be estimated based on the estimated SOC.

[0052] Additionally, the curve generation unit 110 can be configured to generate multiple battery curves for each of the multiple batteries for each cycle based on the obtained battery information.

[0053] Here, the loop can be set to a predetermined time interval. The loop setting can be changed. The loop can also be temporarily changed. For example, the loop can be set by default to have a first time interval. If a specific event occurs before the arrival of the (n+1)th loop according to the first time interval since the nth loop, the curve generation unit 110 can obtain battery information and multiple battery curves. Furthermore, if the nth loop arrives, the curve generation unit 110 can obtain battery information and multiple battery curves. Here, the specific event can be the case where a battery state diagnosis request is input from the outside to the device 100 for diagnosing battery state.

[0054] The curve generation unit 110 can generate a voltage-SOC curve that represents the relationship between the battery voltage (V) and the battery SOC, and a voltage-capacity curve that indicates the relationship between the battery voltage (V) and capacity (Q).

[0055] Additionally, the curve generation unit 110 can generate a differential voltage curve representing the relationship between the battery's charging time (t) and differential voltage (dV / dt). Here, the differential voltage can be a value obtained by differentiating the battery's voltage value [V] with respect to the charging time value [sec].

[0056] Furthermore, the curve generation unit 110 can generate a voltage-differential capacity curve representing the relationship between battery voltage (V) and differential capacity (dQ / dV). Here, differential capacity can be a value obtained by differentiating the battery capacity value [mAh] with respect to the battery voltage value [V].

[0057] The curve generation unit 110 can generate at least one of the voltage-SOC curve, differential voltage curve, voltage-capacity curve, and voltage-differential capacity curve as a battery curve for each of the multiple batteries.

[0058] Figure 2 This is a schematic diagram of a voltage-differential capacity curve in a battery curve generated by a device 100 for diagnosing battery status according to an embodiment of the present disclosure.

[0059] refer to Figure 2 The curve generation unit 110 can generate a voltage-differential capacity curve of the battery's voltage and differential capacity based on the voltage and capacity values ​​obtained while the battery is being charged.

[0060] The feature extraction unit 120 can be configured to extract multiple feature values ​​for each of the multiple batteries by using battery information and multiple battery curves generated by the curve generation unit 110.

[0061] For example, in Figure 2 In one embodiment, the feature extraction unit 120 can determine the charging start point I, the first feature f1, the second feature f2, the third feature f3, the fourth feature f4 and the charging end point F from the voltage-differential capacity curve generated by the curve generation unit 110.

[0062] Here, the first feature f1, the second feature f2, and the fourth feature f4 can be points where the instantaneous rate of change of the differential capacitance with respect to voltage is 0. Preferably, the instantaneous rate of change of the differential capacitance with respect to voltage based on each of the first feature f1, the second feature f2, and the fourth feature f4 can change from a positive value to a negative value. For example, in Figure 2In this embodiment, the voltage value of the first feature f1 selected by the feature extraction unit 120 can be V1, and the differential capacity value can be dQ1. Additionally, the voltage value of the second feature f2 can be V2, and the differential capacity value can be dQ2. Furthermore, the voltage value of the fourth feature f4 can be V4, and the differential capacity value can be dQ4.

[0063] Furthermore, the third feature f3 can be a point where the voltage value (V3) is located between the voltage value (V2) of the second feature f2 and the voltage value (V4) of the fourth feature f4, and where the differential capacity of the third feature f3 changes by a predetermined amount relative to the instantaneous rate of change of the voltage. In particular, the third feature f3 is related to the degradation of the battery positive electrode and can appear in the voltage region between the second feature f2 and the fourth feature f4.

[0064] The feature extraction unit 120 can extract various factors that can be obtained from the voltage-differential capacity curve, such as the voltage value at the charging start point I, the voltage values ​​and differential capacity values ​​of the first to fourth features f1, f2, f3, and f4, the voltage value at the charging end point F, the average change in the differential capacity of the voltage between the charging start point I and each feature f1, f2, f3, and f4, and the area of ​​the differential capacity of the voltage between the charging start point I and each feature f1, f2, f3, and f4, as the feature values ​​of the battery in the voltage-differential capacity curve.

[0065] For example, in Figure 2 In this embodiment, the voltage value at the charging start point I can be VI, and the voltage value at the charging end point F can be VF. Additionally, the differential capacity value at the charging start point I can be dQI, and the differential capacity value at the charging end point F can be dQF.

[0066] For example, the average change in the differential capacity of the voltage between the charging start point I and the first characteristic f1 can be calculated using the formula "(dQ1-dQI)÷(V1-VI)". Additionally, the area of ​​the differential capacitance between the charging start point I and the first characteristic f1 relative to the voltage can be calculated based on the area from VI to V1 based on dQI. In a similar manner, various characteristic values ​​of the battery can be extracted from the voltage-differential capacity curve.

[0067] Furthermore, the feature extraction unit 120 can extract the SOC at the charging start point I and the SOC at the charging end point F from the voltage-SOC curve generated by the curve generation unit 110. That is, the feature extraction unit 120 can further extract the lowest SOC (SOC at the charging start point I) and the highest SOC (SOC at the charging end point F) of the battery as feature values.

[0068] Furthermore, the feature extraction unit 120 can further extract the minimum differential voltage value, maximum differential voltage value, and average differential voltage value from the differential voltage curve generated by the curve generation unit 110 as feature values.

[0069] Furthermore, the feature extraction unit 120 can further extract the battery's SOH and internal resistance [Ω] from the battery information as feature values.

[0070] The principal component calculation unit 130 can be configured to calculate multiple principal components for multiple eigenvalues ​​extracted by the eigenvalue extraction unit 120.

[0071] Specifically, the principal component calculation unit 130 can calculate multiple principal components based on the dispersion of multiple feature values ​​extracted for each of the multiple cells using a principal component analysis method. That is, here, a principal component can refer to a unique vector of multiple extracted feature values.

[0072] For example, suppose the feature extraction unit 120 extracts N feature values ​​for each of the multiple batteries. The principal component calculation unit 130 can calculate N principal components from the N feature values ​​using principal component analysis. That is, the principal component calculation unit 130 can calculate the first principal component PC1 to the nth principal component (PCN).

[0073] As a more specific example, suppose the eigenvalues ​​of the first battery are a1, b1, and c1, the eigenvalues ​​of the second battery are a2, b2, and c2, and the eigenvalues ​​of the third battery are a3, b3, and c3. Referring to the Euclidean coordinate system, the eigenvalues ​​of the first battery can be expressed as (a1, b1, c1), the eigenvalues ​​of the second battery can be expressed as (a2, b2, c2), and the eigenvalues ​​of the third battery can be expressed as (a3, b3, c3). The principal component calculation unit 130 can calculate the first principal component PC1, the second principal component PC2, and the third principal component PC3 for (a1, b1, c1), (a2, b2, c2), and (a3, b3, c3).

[0074] The battery classification unit 140 can be configured to calculate the distribution degree of each principal component representing the distribution of the multiple batteries based on multiple feature values ​​extracted for each of the multiple batteries.

[0075] First, the battery classification unit 140 can calculate the distribution value for each of the multiple principal components by projecting multiple feature values ​​of each of the multiple batteries onto each of the multiple principal components.

[0076] For example, as in the previous embodiments, assume that the feature extraction unit 120 extracts three feature values ​​for the first battery, the second battery, and the third battery, and that the first principal component PC1, the second principal component PC2, and the third principal component PC3 are calculated by the principal component calculation unit 130. The battery classification unit 140 can calculate the distribution values ​​of the first to third batteries for the first principal component PC1 by projecting the feature values ​​of the first to third batteries onto the first principal component PC1. That is, the battery classification unit 140 can project (a1, b1, c1), (a2, b2, c2), and (a3, b3, c3) onto the first principal component PC1. For example, assume that the feature value 1 of the first battery projected onto the first principal component PC is (a1', b1', c1'), the feature value of the second battery projected onto it is (a2', b2', c2'), and the feature value of the third battery projected onto it is (a3', b3', c3'). The battery classification unit 140 can calculate the maximum distance among the distances between (a1', b1', c1') and (a2', b2', c2'), the distances between (a1', b1', c1') and (a3', b3', c3'), and the distances between (a2', b2', c2') and (a3', b3', c3'), as the distribution values ​​for the first to third batteries of the first principal component PC1. In this way, the battery classification unit 140 can calculate the distribution values ​​for the first to third batteries of the second principal component PC2 and the distribution values ​​for the first to third batteries of the third principal component PC3.

[0077] Furthermore, the battery classification unit 140 can calculate the distribution degree of each principal component based on the distribution value calculated for each of the multiple principal components. Specifically, the battery classification unit 140 can calculate the distribution degree of each principal component using Equation 1 below.

[0078] [Equation 1]

[0079]

[0080] In Equation 1, n is the number of principal components calculated by the principal component calculation unit 130, and is a natural number. PCi refers to the i-th principal component, and PCj refers to the j-th principal component. For example, PC1 refers to the first principal component PC1, PC2 refers to the second principal component PC2, and PCn refers to the n-th principal component PCn. That is, i and j are temporary variables used to specify the corresponding principal components, and j is a natural number between 1 and n.

[0081] Additionally, in Equation 1, PCjv is the distribution degree of each principal component. For example, PC1v is the distribution degree of the first principal component PC1, PC2v is the distribution degree of the second principal component PC2, and PCnv is the distribution degree of the nth principal component PCn.

[0082] Furthermore, in Equation 1, Var(PCi) refers to the set of distances between the eigenvalues ​​of the multiple cells projected onto the i-th principal component. Additionally, max{Var(PCi)} is the distribution value of the i-th principal component. That is, max{Var(PCi)} represents the maximum distance among the distances between the eigenvalues ​​of the multiple cells projected onto the i-th principal component.

[0083] Figure 3 This is a diagram schematically illustrating an example of the distribution degree of each principal component calculated by an apparatus 100 for diagnosing battery state according to an embodiment of the present disclosure.

[0084] refer to Figure 3 The first principal component PC1, the second principal component PC2, the third principal component PC3, the fourth principal component PC4, and the fifth principal component PC5 can be calculated by the principal component calculation unit 130. Here, the distribution degree of the first principal component PC1 is 55%, the distribution degree of the second principal component PC2 is 25%, the distribution degree of the third principal component PC3 is 11%, the distribution degree of the fourth principal component PC4 is 6%, and the distribution degree of the fifth principal component PC5 is 3%.

[0085] The battery classification unit 140 can be configured to select at least one of a plurality of principal components as a target component based on the result of comparing the calculated distribution degree for each principal component with a preset standard value.

[0086] Preferably, the battery classification unit 140 can be configured to select at least one of a plurality of principal components, and select the at least one principal component as the target component when the sum of the distribution degrees of each principal component corresponding to the at least one selected principal component is equal to or greater than a preset standard value.

[0087] More specifically, the battery classification unit 140 can be configured to calculate a sum by sequentially adding the distribution degrees of each principal component corresponding to each of the plurality of principal components from the largest one until the sum becomes equal to or greater than a preset standard value.

[0088] For example, in Figure 3In this embodiment, it is assumed that the standard value is preset to 90%. The battery classification unit 140 may first select a first principal component PC1 and compare the distribution degree (55%) of the first principal component PC1 with the standard value (90%). Since the distribution degree (55%) of the first principal component PC1 is less than the standard value (90%), the battery classification unit 140 may select a second principal component PC2 and calculate the sum of the distribution degrees of the first and second principal components PC1 and PC2 as 80%. Since the sum of the distribution degrees (80%) of the first and second principal components PC1 and PC2 is also less than the standard value (90%), the battery classification unit 140 may select a third principal component PC3 and calculate the sum of the distribution degrees of the first to third principal components PC1 to PC3 as 91%. Since the sum of the distribution degrees (91%) of the first to third principal components PC1 to PC3 is greater than the standard value (90%), the battery classification unit 140 may select the first principal component PC1, the second principal component PC2, and the third principal component PC3 as target components.

[0089] Additionally, the battery classification unit 140 can be configured to classify each of the multiple batteries into one of multiple groups based on at least one selected target component. That is, the target component is selected to classify the multiple batteries and can be some or all of multiple principal components.

[0090] Specifically, the battery classification unit 140 can be configured to classify multiple batteries into one of multiple groups based on at least one selected target component being set as a reference cell corresponding to the abnormal cell.

[0091] For example, the reference cell is a battery of the same type as the multiple batteries and may be a degraded cell in an EOL (end-of-life) state. The battery classification unit 140 can classify the multiple batteries and the reference cell into one of multiple groups based on at least one selected target component.

[0092] The battery status diagnostic unit 150 can be configured to update the classification history of each group to which a plurality of batteries are classified by the battery classification unit 140 in each cycle.

[0093] Preferably, the battery classification unit 140 can be configured to classify a reference cell corresponding to an abnormal cell into one of a plurality of groups based on at least one selected target component. Furthermore, the battery state diagnosis unit 150 can be configured to designate the group to which the reference cell belongs as an abnormal group and the remaining groups as normal groups.

[0094] That is, the group to which the reference individual is classified among multiple groups can be set as the abnormal group, and the remaining groups among multiple groups other than the abnormal group can be set as the normal group.

[0095] Additionally, a classification history can be provided for each of the multiple batteries. That is, a classification history can be provided for each of the multiple batteries. Furthermore, the classification history can be configured to include the groups classified in the previous loop, the groups classified in the current loop, the frequency of changes to abnormal groups, and the frequency of changes to normal groups.

[0096] For example, the battery state diagnostic unit 150 can update the classification history of each of multiple batteries in each cycle. The updated classification history can include the groups classified in the previous cycle and the groups classified in the current cycle. Additionally, the frequency of changes to abnormal groups or normal groups can be updated based on the group classification result of the batteries in the current cycle.

[0097] In addition, the battery status diagnostic unit 150 can be configured to diagnose the status of each of multiple batteries based on an updated classification history.

[0098] Specifically, the battery state diagnostic unit 150 can be configured to identify a target battery among a plurality of batteries that is different from the group classified in the previous cycle and the group classified in the current cycle.

[0099] Here, a target battery refers to a battery whose group in the previous loop differs from the group in the current loop. Depending on the classification results, a target battery identified in the current loop may not exist, or at least it may exist. For example, a battery classified as an anomalous group in the previous loop but as a normal group in the current loop can be identified as a target battery. Conversely, a battery classified as a normal group in the previous loop but as an anomalous group in the current loop can be identified as a target battery.

[0100] Additionally, the battery status diagnostic unit 150 can be configured to diagnose the status of a target battery as normal or abnormal based on the frequency of change to the abnormal group or the frequency of change to the normal group for a given target battery.

[0101] Here, the frequency of changing to the anomalous group is the number of times a cell was classified as a normal group in the previous loop but is classified as an anomalous group in the current loop. Conversely, the frequency of changing to the normal group is the number of times a cell was classified as an anomalous group in the previous loop but is classified as a normal group in the current loop. That is, in the case of the target battery, the frequency of changing to the anomalous group or the frequency of changing to the normal group can be updated based on the group classification result. For example, the frequency of changing can be increased by one.

[0102] Preferably, when the target battery is classified as an abnormal group in the current cycle, the battery state diagnosis unit 150 can be configured to diagnose the state of the target battery as abnormal if the frequency of change to the abnormal group is equal to or greater than a predetermined frequency.

[0103] Conversely, when the target battery is classified as a normal group in the current cycle, the battery state diagnostic unit 150 can be configured to diagnose the target battery as a normal state if the frequency of change to the normal group is equal to or greater than a predetermined frequency.

[0104] For example, the predetermined frequency can be preset to two or more. When the battery is identified as the target battery multiple times, the battery status diagnosis unit 150 can diagnose the battery status as abnormal or normal.

[0105] That is, the battery status diagnostic unit 150 does not determine the battery status based on the fact that the battery classification result changes from abnormal to normal or from normal to abnormal only once, but can determine the battery status when the battery classification result changes multiple times. Therefore, the accuracy of the battery status determination by the battery status diagnostic unit 150 can be very high.

[0106] Since the apparatus 100 for diagnosing battery state according to embodiments of the present disclosure diagnoses the state of each of a plurality of batteries based on principal component analysis and battery classification, it has the advantage of being able to consider various aspects when diagnosing battery state. Furthermore, compared to prior art methods that diagnose battery state by simply comparing battery characteristic values ​​with predetermined reference values, the apparatus 100 for diagnosing battery state can improve the accuracy of battery state diagnosis.

[0107] At the same time, multiple reference units can be provided. In this case, each of the multiple reference units can be configured to be degraded for different reasons.

[0108] For example, the first reference monomer can be configured to degrade due to the loss of positive electrode reaction area, and the second reference monomer can be configured to degrade due to the loss of available lithium. Furthermore, the third reference monomer can be configured to degrade due to the loss of negative electrode reaction area.

[0109] In this case, the state of the battery classified into the same group as each reference cell can be diagnosed in more detail.

[0110] For example, referring to the preceding embodiments, the state of a battery classified into the same group as the first reference cell can be diagnosed as a state of positive electrode reaction area loss. Furthermore, the state of a battery classified into the same group as the second reference cell can be diagnosed as a state of usable lithium loss. Additionally, the state of a battery classified into the same group as the third reference cell can be diagnosed as a state of negative electrode reaction area loss.

[0111] Since the apparatus 100 for diagnosing battery status according to embodiments of the present disclosure includes a plurality of reference cells configured to be degraded for various reasons, there is an advantage in diagnosing the status of a plurality of batteries more specifically.

[0112] For example, each of the curve generation unit 110, feature extraction unit 120, principal component calculation unit 130, battery classification unit 140, and battery state diagnosis unit 150 included in the apparatus 100 for diagnosing battery state according to embodiments of the present disclosure may be configured as a processor known in the art.

[0113] As another example, the device 100 for diagnosing battery status, including a curve generation unit 110, a feature extraction unit 120, a principal component calculation unit 130, a battery classification unit 140, and a battery status diagnosis unit 150, can be housed in a single processor. That is, the curve generation unit 110, feature extraction unit 120, principal component calculation unit 130, battery classification unit 140, and battery status diagnosis unit 150 can be divided into functional components of a single processor. Alternatively, each of the curve generation unit 110, feature extraction unit 120, principal component calculation unit 130, battery classification unit 140, and battery status diagnosis unit 150 can be implemented as a unit core housed in a single processor.

[0114] Furthermore, the processor may optionally include application-specific integrated circuits (ASICs), another chipset, logic circuits, registers, communication modems, and data processing devices, as known in the art, to execute the various control logics disclosed below. Additionally, when the control logic is implemented in software, the processor can be implemented as a collection of program modules. In this case, the program modules can be stored in memory and executed by the processor. The memory can be located inside or outside the processor and can be connected to the processor by various well-known means.

[0115] refer to Figure 1 The apparatus 100 for diagnosing battery status according to embodiments of the present disclosure may further include a storage unit 160.

[0116] Here, storage unit 160 can store programs, data, etc., required for diagnosing battery status according to this disclosure. That is, storage unit 160 can store data necessary for the operation and function of each component of the device 100 for diagnosing battery status, data generated during the execution of operations or functions, etc. There are no particular limitations on the type of storage unit 160, as long as it is a known information storage device capable of recording, erasing, updating, and retrieving data. As examples, information storage devices may include RAM, flash memory, ROM, EEPROM, registers, etc. Furthermore, storage unit 160 can store program code that defines the procedures that can be executed by the device 100 for diagnosing battery status.

[0117] For example, storage unit 160 can store the classification history of multiple batteries. Furthermore, storage unit 160 can store multiple battery curves and multiple feature values, thereby storing various information about multiple batteries.

[0118] That is, according to the device 100 for diagnosing battery status, changes in the status of each of multiple batteries can be tracked and diagnosed. Therefore, when a corresponding battery experiences an accident, the cause of the accident can be analyzed from various aspects based on the various information about the corresponding battery stored in the storage unit 160.

[0119] The following will describe in detail how the battery classification unit 140 classifies multiple batteries and reference cells into any of the multiple groups based on the target composition.

[0120] The battery classification unit 140 can be configured to generate a representative model in each cycle for classifying multiple batteries and a reference cell into any of the multiple groups.

[0121] Specifically, the battery classification unit 140 can be configured to generate a target set that includes at least one selected target component.

[0122] For example, suppose in Figure 3 In this embodiment, the standard value is set to 90%. Since the sum of the distribution degrees of the first principal component PC1 (55%), the second principal component PC2 (25%), and the third principal component PC3 (11%) is 91%, the first principal component PC1, the second principal component PC2, and the third principal component PC3 can be selected as target components. Furthermore, the first principal component PC1, the second principal component PC2, and the third principal component PC3 selected above can be configured to form a target set.

[0123] Furthermore, the battery classification unit 140 can be configured to generate multiple subsets of targets that include at least one target component from the generated target set.

[0124] In the above embodiment, the battery classification unit 140 can generate multiple target subsets by selecting at least one of the first principal component PC1, the second principal component PC2, and the third principal component PC3 included in the target set.

[0125] For example, if described using tabular notation, the target set is {first principal component PC1, second principal component PC2, third principal component PC3}. Multiple target subsets can be {first principal component PC1}, {second principal component PC2}, {third principal component PC3}, {first principal component PC1, second principal component PC2}, {first principal component PC1, third principal component PC3}, {second principal component PC2, third principal component PC3}, and {first principal component PC1, second principal component PC2, third principal component PC3}. Here, since multiple batteries are classified by target components, it should be noted that empty sets are excluded from the target subsets. Therefore, the number of target subsets generated by the battery classification unit 140 is 2. n -1, where n is the total number of target components. That is, in the above embodiment, since there are 3 target components, therefore according to "2 3 The formula "-1" indicates that the number of target subsets is 7.

[0126] The battery classification unit 140 can be configured to generate at least one classification model for each of a plurality of target subsets for classifying a plurality of batteries and a reference cell.

[0127] Specifically, the battery classification unit 140 can generate at least one classification model for a target subset. Furthermore, the number of classification models generated for each target subset can be independent for each subset. That is, one classification model can be generated for one target subset, while 10 classification models can be generated for another target subset.

[0128] Figure 4 and Figure 5 This is a schematic diagram illustrating a classification model generated in the current cycle by a device 100 for diagnosing battery status according to an embodiment of the present invention.

[0129] Specifically, Figure 4 and 5 The classification model shown is generated in the current loop and can be a different classification model randomly generated for a subset of targets.

[0130] For example, refer to Figure 4 and Figure 5 The classification model generated by the battery classification unit 140 can be a decision tree. That is, the classification model includes a root node RN, internal nodes IN, and terminal nodes TN. The root node RN and internal nodes IN can include classification conditions for classifying multiple batteries, and the terminal nodes can include information about the classified batteries. Each terminal node can correspond to a group to which a battery can be classified. However, some of the multiple terminal nodes can be blank nodes where the battery has not been classified.

[0131] For example, in Figure 4 In this embodiment, it is assumed that the reference cell is classified as the first terminal node TN1. The battery state diagnosis unit 150 can diagnose the state of the battery among the plurality of batteries that is classified as the first terminal node TN1 as an abnormal state.

[0132] As another example, in Figure 4 In this embodiment, it is assumed that a first reference cell with a loss of positive electrode reaction area is classified as a first terminal node TN1, and a second reference cell with a loss of available lithium is classified as a fourth terminal node TN4. The battery state diagnosis unit 150 can diagnose the state of the battery classified as a first terminal node TN1 as an abnormal state due to the loss of positive electrode reaction area, and diagnose the state of the battery classified as a fourth terminal node TN4 as an abnormal state due to the loss of available lithium.

[0133] In addition, Figure 4 and 5 In this embodiment, the classification conditions included in the root node RN and internal nodes IN can be arbitrarily generated by the battery classification unit 140. For example, Figure 4 The root node RN of the classification model can include classification conditions associated with the first principal component PC1, and Figure 5 The root node RN of the classification model can include classification conditions associated with the third principal component PC3. Therefore, classification models generated for the same target subset may differ. Specific details regarding the classification conditions included in the root node RN and internal nodes IN will be described later.

[0134] The battery classification unit 140 can be configured to set any one of the multiple classification models generated for multiple target subsets as the representative model of the corresponding loop.

[0135] Specifically, the battery classification unit 140 can be configured to calculate the classification degree of multiple batteries and reference cells for each of the multiple generated classification models.

[0136] Battery classification unit 140 can calculate the classification degree of each classification model using Equation 2 below. Here, the classification degree of a classification model is an indicator of the classification performance of the classification model. That is, the classification degree of a classification model can be an indicator of the extent to which the classification model classifies multiple batteries and a reference cell. For example, the classification degree of a classification model can be Gini impurity.

[0137] [Equation 2]

[0138]

[0139] Here, GI(n) is the classification degree of the classification model, and n is a temporary variable used to indicate each of the multiple classification models generated for a target subset. Additionally, i is a temporary variable indicating the terminal node, and m is the number of terminal nodes included in the classification model. Furthermore, Pi is the ratio of the number of batteries and / or reference cells classified to the i-th terminal node to the total number of batteries and reference cells.

[0140] For example, Figure 4 The classification degree of the classification model can be determined by the formula "1-(P1)". 2 +P2 2 +P3 2 +P4 2 +P5 2 +P6 2 To calculate, and Figure 5 The classification degree of the classification model can be determined by the formula "1-(P1)". 2 +P2 2 +P3 2 +P4 2 +P5 2 +P6 2 +P7 2 +P8 2 )" to calculate.

[0141] As a specific example, assume there are five cells in total and one reference cell. That is, m in Equation 2 can be 6. Furthermore, assume the reference cell is... Figure 4 In the classification model, a battery is classified into the first terminal node TN1, and a battery is classified into each of the second to sixth terminal nodes TN2, TN3, TN4, TN5, and TN6. The total number of batteries and reference cells is 6, and the number of batteries and / or reference cells classified into each of the first to sixth terminal nodes TN1, TN2, TN3, TN4, TN5, and TN6 is 1. Figure 4 In this embodiment, P1 to P6 can all be "1 ÷ 6". Therefore, Figure 4 The classification degree of a classification model can be determined according to the equation "1-{(1÷6)}". 2 The calculation of "5÷6" is performed using the formula "×6".

[0142] As another example, suppose the reference monomer is in Figure 5 In the classification model, a battery is classified into the first terminal node TN1, and a battery is classified into each of the third to seventh terminal nodes TN3, TN4, TN5, TN6, and TN7. Figure 5 In this embodiment, P1 and P3 through P7 can all be "1 ÷ 6". Furthermore, P2 and P8 can be "0" because there is no classified battery and / or reference cell. Therefore, Figure 5The classification degree of a classification model can be determined according to the equation "1-{(1÷6)}". 2 The calculation of "5÷6" is performed using the formula "×6".

[0143] Furthermore, the battery classification unit 140 can be configured to set the classification model with the lowest calculated classification degree among multiple generated classification models as the representative model of the corresponding cycle.

[0144] According to Equation 2, since the calculated classification degree is low, it can be determined that the classification model uniformly classifies multiple batteries and reference cells. Therefore, the battery classification unit 140 can set the classification model with the lowest calculated classification degree among the multiple classification models generated for multiple target subsets as the representative model of the corresponding loop.

[0145] For example, in the nth loop, suppose p target subsets are generated, and q classification models are generated for each target subset. That is, in the nth loop, a total of p×q classification models can be generated. The battery classification unit 140 can calculate the classification degree for each of the p×q classification models, and set the classification model with the lowest calculated classification degree as the representative model for the nth loop.

[0146] If among the p×q classification models, there exist multiple classification models with the lowest calculated classification degree, then the battery classification unit 140 can set the classification model with the largest number of non-blank terminal nodes among the multiple classification models with the lowest calculated classification degree as the representative model. In other words, the battery classification unit 140 can set the classification model with the largest number of terminal nodes to which the battery and / or reference cell are classified as the representative model among the multiple classification models with the lowest calculated classification degree.

[0147] That is, although the classification degree calculated based on Equation 2 is the same, it can be determined that due to the reduction in the number of non-blank terminal nodes, multiple batteries and reference cells are uniformly classified. Therefore, the battery classification unit 140 can set a representative model by considering the number of non-blank terminal nodes and the classification degree of multiple classification models.

[0148] If the calculated classification degree is the lowest, and there are multiple classification models with the same number of non-blank terminals, the battery classification unit 140 can arbitrarily select any one of the multiple classification models and set it as the representative model.

[0149] For example, referring to the preceding embodiments, Figure 4 and 5 The classification model has a classification degree of "5 ÷ ​​6". Additionally, Figure 4 The classification model contains 6 non-blank terminal nodes TN1, TN2, TN3, TN4, TN5, and TN6. Additionally, Figure 5 The number of non-blank terminal nodes TN1, TN3, TN4, TN5, TN6, and TN7 in the classification model is 6. Therefore, the battery classification unit 140 can classify the non-blank terminal nodes TN1, TN3, TN4, TN5, TN6, and TN7. Figure 4 Classification model or Figure 5 The classification model is set as the representative model for the current loop.

[0150] In addition, the battery status diagnosis unit 150 can diagnose the status of batteries that are classified into the same group as the reference cell as abnormal based on the representative model set by the battery classification unit 140.

[0151] That is, the apparatus 100 for diagnosing battery status according to embodiments of the present disclosure can generate multiple classification models in one cycle and set a representative model by comparing the classification degrees among the multiple generated classification models. Furthermore, since battery status can be diagnosed based on the set representative model, the apparatus 100 for diagnosing battery status can more accurately and statistically diagnose battery status compared to prior art methods that diagnose battery status by simply comparing battery information.

[0152] The following text will describe in detail the classification criteria included in the root node RN and the internal nodes IN.

[0153] The battery classification unit 140 can be configured to calculate the inner product of multiple batteries and a reference cell for at least one target component included in each of multiple target subsets.

[0154] Here, the target component can be a unique vector of multiple feature values ​​extracted from multiple batteries. For each of these target components, the inner product of the batteries for the target component can be obtained by calculating the inner product of the vectors of the battery's feature values.

[0155] For example, suppose the target subset is {first principal component PC1, second principal component PC2}. Also, suppose a first cell, a second cell, and a reference cell are defined. For ease of description, the starting point of the first principal component PC1 is referred to as point A, and the starting point of the second principal component PC2 is referred to as point B.

[0156] First, the battery classification unit 140 can calculate the inner product of the first battery, the inner product of the second battery, and the inner product of the reference cell for the first principal component PC1.

[0157] Specifically, the battery classification unit 140 calculates a first temporary vector, where the starting point is point A and the ending point is the point where the feature value of the first battery is located. Furthermore, the battery classification unit 140 can calculate the inner product between the first principal component PC1 and the first temporary vector to calculate the inner product of the first battery with respect to the first principal component PC1.

[0158] Similarly, battery classification unit 140 calculates a second temporary vector, where the starting point is point A and the ending point is the point where the feature value of the second battery is located. Furthermore, battery classification unit 140 can calculate the inner product between the first principal component PC1 and the second temporary vector to calculate the inner product of the second battery with respect to the first principal component PC1.

[0159] In the manner described above, the battery classification unit 140 can calculate the inner product of a reference cell for the first main component PC1, the inner product of a first cell for the second main component PC2, the inner product of a second cell for the second main component PC2, and the inner product of a reference cell for the second main component PC2.

[0160] Furthermore, the battery classification unit 140 can be configured to generate multiple classification models for classifying multiple batteries and a reference cell based on the size of multiple calculated inner products.

[0161] Specifically, the battery classification unit 140 can select one of multiple target subsets to generate a classification model. Furthermore, the battery classification unit 140 can arbitrarily select one of the target components belonging to the selected target subset.

[0162] Furthermore, the battery classification unit 140 can classify the multiple batteries and the reference cell based on the result of comparing the inner product of the multiple batteries for the selected target component and the inner product of the reference cell with an arbitrarily set comparison value. Here, the arbitrarily set comparison value can be a value set by the battery classification unit 140.

[0163] That is, the classification criteria used to compare the size of the inner product of multiple cells and reference cells for the selected target component with the size of an arbitrarily set comparison value can be included in the root node RN or the inner node IN of the classification model, and multiple cells and reference cells can be classified according to the classification results.

[0164] Subsequently, the battery classification unit 140 can arbitrarily select one of the target components belonging to the selected target subset and reclassify multiple batteries and reference cells. That is, the target components selected by the battery classification unit 140 can overlap. In addition, since it is an arbitrarily selected value, the arbitrarily set comparison value can be changed.

[0165] For example, in Figure 5 In one embodiment, when the target subset is {first principal component PC1, second principal component PC2, third principal component PC3}, the battery classification unit 140 can arbitrarily select the first principal component PC1 among the first principal component PC1, the second principal component PC2, and the third principal component PC3 for the root node RN.

[0166] The battery classification unit 140 can calculate the inner product of multiple batteries and the inner product of a reference cell for the first principal component PC1. Furthermore, the battery classification unit 140 can arbitrarily generate and set a first comparison value for the first principal component PC1. The battery classification unit 140 can classify the multiple batteries and reference cells in the root node RN based on the result of comparing the size of the inner product of the multiple batteries and the reference cell for the first principal component PC1 with the size of the arbitrarily set comparison value.

[0167] Subsequently, the battery classification unit 140 can again arbitrarily select the first principal component PC1 from the target subset for the first internal node IN1. Furthermore, the battery classification unit 140 can arbitrarily generate and set a second comparison value for the first principal component PC1. Here, the first comparison value and the second comparison value are randomly generated values ​​and can be independent of each other. Preferably, the first comparison value and the second comparison value can be different from each other.

[0168] The battery classification unit 140 can classify some batteries in the first internal node IN1 based on the result of comparing the size of the inner product of some batteries (batteries and / or reference cells classified into the first internal node IN1 according to the classification result in the root node RN) with the size of the second comparison value.

[0169] In this way, Figure 5 In one embodiment, the battery classification unit 140 can generate a classification model including a root node RN, six internal nodes IN1, IN2, IN3, IN4, IN5, IN6 and eight terminal nodes TN1, TN2, TN3, TN4, TN5, TN6, TN7, TN8.

[0170] Preferably, the battery classification unit 140 can be configured to control each of the multiple classification models such that the total number of groups to be classified is less than or equal to a preset standard number.

[0171] That is, the battery classification unit 140 can control the number of terminal nodes included in the classification model to be less than or equal to the standard number.

[0172] For example, if there is no limit to the number of terminal nodes, the classification model can be extended until multiple batteries and reference cells are classified into a single group. In this case, there is a problem of excessive waste of system resources required to generate a classification model.

[0173] Furthermore, referring to Equation 2, if multiple batteries and reference cells are classified into separate groups, the classification degrees of multiple classification models can all be the same. For example, in multiple classification models, if at most one battery or reference cell is included in each of the multiple groups, then since Pi in Equation 2 is 0 or "1 ÷ (number of multiple batteries + number of reference cells)", the classification degrees of multiple classification models can all be the same.

[0174] In this case, since there are no batteries classified into the same group as the reference cell, it is possible to conclude that the battery's condition cannot be diagnosed as abnormal.

[0175] Therefore, since the device 100 for diagnosing battery status according to the embodiments of this disclosure controls the total number of terminal nodes included in the classification model—that is, the total number of groups to be classified—to be no less than or equal to a preset standard number, the system resources required to generate the classification model can be saved and the battery status can be accurately diagnosed.

[0176] The device 100 for diagnosing battery state according to this disclosure can be applied to a BMS (Battery Management System). That is, a BMS according to this disclosure may include the aforementioned device 100 for diagnosing battery state. In this configuration, at least some of the components of the device 100 for diagnosing battery state can be implemented by supplementing or adding functionality included in a conventional BMS. For example, the curve generation unit 110, feature extraction unit 120, principal component calculation unit 130, battery classification unit 140, and battery state diagnosis unit 150 of the device 100 for diagnosing battery state can be implemented as components of a BMS.

[0177] Furthermore, the device 100 for diagnosing battery status according to this disclosure can be provided to a battery pack. For example, a battery pack according to this disclosure may include the device 100 for diagnosing battery status as described above, and at least one battery cell. Additionally, the battery pack may further include electrical devices (relays, fuses, etc.), a casing, etc.

[0178] Figure 6 This is a diagram schematically illustrating an exemplary configuration of a battery pack 1 including a device 100 for diagnosing battery status according to an embodiment of the present disclosure.

[0179] For example, refer to Figure 6The battery pack 1 may include multiple batteries B1, B2, B3, B4, a measurement unit 10, a charging and discharging unit 20, and a device 100 for diagnosing battery status. The charging and discharging unit 20 can charge the multiple batteries. Furthermore, the measurement unit 10 can measure the voltage and current of each of the multiple batteries while they are being charged. Additionally, the measurement unit 10 can measure the capacity of each of the multiple batteries. Furthermore, the measurement unit 10 can estimate the internal resistance, state of charge (SOC), and state of equilibrium (SOH) of each of the multiple batteries. The curve generation unit 110 can receive battery information from the measurement unit 10 and generate voltage-SOC curves, voltage-capacity curves, differential voltage curves, and voltage-differential capacity curves based on the received battery information.

[0180] Furthermore, the device 100 for diagnosing battery status according to embodiments of the present disclosure can be provided in an energy storage system. That is, the energy storage system according to the present disclosure may include the device 100 for diagnosing battery status.

[0181] For example, the device 100 for diagnosing battery status according to embodiments of the present disclosure can be provided in an energy storage system to periodically diagnose the status of each battery included in the energy storage system. Therefore, even if a fire occurs in the energy storage system, the timing and cause of the fire can be easily identified using the information stored in the device 100 for diagnosing battery status.

[0182] Figure 7 This diagram schematically illustrates a method for diagnosing battery status according to another embodiment of the present disclosure. Here, each step of the method for diagnosing battery status can be performed by a device 100 for diagnosing battery status.

[0183] The following text will briefly describe the content that overlaps with the previously described content.

[0184] refer to Figure 7 The method for diagnosing battery status may include a curve generation step (S100), a feature extraction step (S200), a principal component calculation step (S300), a principal component distribution degree calculation step (S400), a target component selection step (S500), a battery classification step (S600), and a battery status diagnosis step (S700).

[0185] The curve generation step (S100) is a step of generating a battery curve for each of the multiple batteries, representing the correspondence between the voltage and capacity of each of the multiple batteries in each cycle, and can be performed by the curve generation unit 110.

[0186] For example, in the curve generation step (S100), a voltage-SOC curve, a voltage-capacity curve, a differential voltage curve, and a voltage-differential capacity curve can be generated for each of the multiple batteries.

[0187] The feature value extraction step (S200) is a step of extracting multiple feature values ​​from each of the multiple battery curves generated in the curve generation step (S100), and can be performed by the feature value extraction unit 120.

[0188] The principal component calculation step (S300) is a step of calculating multiple principal components for multiple eigenvalues ​​extracted in the eigenvalue extraction step (S200), and can be executed by the principal component calculation unit 130.

[0189] For example, the principal component calculation unit 130 can calculate multiple principal components by using a principal component analysis method based on the distribution of multiple feature values ​​extracted for each of the multiple cells.

[0190] The principal component distribution degree calculation step (S400) is a step of calculating the distribution degree of each principal component representing the distribution of the multiple batteries based on the multiple feature values ​​extracted for each of the multiple batteries, and can be performed by the battery classification unit 140.

[0191] The target component selection step (S500) is a step of selecting at least one of a plurality of principal components as a target component based on the result of comparing the distribution degree of each principal component calculated in the principal component distribution degree calculation step (S400) with a preset standard value, and can be executed by the battery classification unit 140.

[0192] For example, in Figure 3 In this embodiment, the standard value can be preset to 90%. Furthermore, since the sum of the distribution degree of the first principal component PC1 (55%), the distribution degree of the second principal component PC2 (25%), and the distribution degree of the third principal component PC3 (11%) is equal to or greater than the standard value (90%), the battery classification unit 140 can select the first principal component PC1, the second principal component PC2, and the third principal component PC3 as target components.

[0193] The battery sorting step (S600) is a step of sorting each of the plurality of batteries into any one of the plurality of groups based on at least one target component selected in the target component selection step (S500), and can be performed by the battery sorting unit 140.

[0194] The battery state diagnosis step (S700) is a step that updates the classification history of each of the multiple batteries to which it was classified in the battery classification step (S600) in each cycle, and diagnoses the state of each of the multiple batteries based on the updated classification history, and can be executed by the battery state diagnosis unit 150.

[0195] For example, the battery state diagnosis unit 150 can identify a target battery as one that is in a different group than the one classified in the previous cycle and is classified in the current cycle. Subsequently, when the target battery is classified into an abnormal group in the current cycle, the battery state diagnosis unit 150 can be configured to diagnose the target battery as having an abnormal state if the frequency of changes to the abnormal group is equal to or greater than a predetermined frequency. Conversely, when the target battery is classified into a normal group in the current cycle, the battery state diagnosis unit 150 can be configured to diagnose the target battery as having a normal state if the frequency of changes to the normal group is equal to or greater than a predetermined frequency.

[0196] The embodiments of this disclosure described above are not necessarily implemented by apparatus and methods, but may also be implemented by a program for implementing functions corresponding to the configuration of this disclosure, or by a recording medium on which the program is recorded. Such implementations can be readily implemented by those skilled in the art based on the description of the above embodiments.

[0197] This disclosure has been described in detail. However, it should be understood that while the detailed description and specific examples indicate preferred embodiments of this disclosure, they are given by way of illustration only, as various changes and modifications within the scope of this disclosure will become apparent to those skilled in the art from this detailed description.

[0198] Additionally, 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, and the present disclosure is not limited to the above embodiments and drawings, and each embodiment can be selectively combined in part or in whole to allow for various modifications.

[0199] (Reference symbols)

[0200] 1: Battery pack

[0201] 10: Measurement Unit

[0202] 20: Charging and discharging unit

[0203] 100: Device for diagnosing battery status

[0204] 110: Curve generation unit

[0205] 120: Feature Value Extraction Unit

[0206] 130: Principal Component Calculation Unit

[0207] 140: Battery Classification Unit

[0208] 150: Battery Status Diagnostic Unit

[0209] 160: Storage unit

Claims

1. A device for diagnosing battery status, comprising: A curve generation unit is configured to obtain battery information including at least one of voltage, capacity, internal resistance, SOC and SOH for each of a plurality of batteries, and to generate a plurality of battery curves for each of the plurality of batteries based on the obtained battery information in each cycle; A feature extraction unit is configured to extract multiple feature values ​​for each of the plurality of batteries by using the battery information and the plurality of battery curves generated by the curve generation unit. A principal component calculation unit is configured to calculate a plurality of principal components for the plurality of feature values ​​extracted by the feature extraction unit; A battery classification unit is configured to calculate the distribution degree of each principal component representing the distribution of the plurality of batteries based on the plurality of feature values ​​extracted for each of the plurality of batteries, select at least one of the plurality of principal components as a target component based on the result of comparing the calculated distribution degree of each principal component with a preset standard value, and classify each of the plurality of batteries into any one of the plurality of groups based on the selected at least one target component. as well as A battery state diagnostic unit is configured to update the classification history of each group to which the plurality of batteries are classified by the battery classification unit in each cycle, and to diagnose the state of each of the plurality of batteries based on the updated classification history.

2. The apparatus for diagnosing battery status according to claim 1, in, The battery classification unit is configured to classify a selected target component into any of the plurality of groups based on the fact that at least one target component will be set as a reference cell corresponding to the anomalous cell. The battery status diagnostic unit is configured to set the group to which the reference cell belongs among the plurality of groups as an abnormal group and set the remaining groups as normal groups.

3. The apparatus for diagnosing battery status according to claim 2, in, The classification history is provided to each of the plurality of batteries and is configured to include groups classified in the previous cycle, groups classified in the current cycle, the frequency of change to the abnormal group, and the frequency of change to the normal group.

4. The apparatus for diagnosing battery status according to claim 3, in, The battery state diagnostic unit is configured to identify a target battery among the plurality of batteries whose group in the previous cycle is different from the group in the current cycle, and to diagnose the state of the target battery as normal or abnormal based on the frequency of change to the abnormal group or the frequency of change to the normal group for the identified target battery.

5. The apparatus for diagnosing battery status according to claim 4, in, The battery state diagnostic unit is configured to: when the target battery is classified into the abnormal group in the current cycle, and when the frequency of changes to the abnormal group is equal to or greater than a predetermined frequency, diagnose the state of the target battery as abnormal, and The battery state diagnostic unit is configured to diagnose the state of the target battery as normal when the frequency of change to the normal group is equal to or greater than the predetermined frequency, provided that the target battery is classified into the normal group in the current cycle.

6. The apparatus for diagnosing battery status according to claim 2, in, The battery classification unit is configured to select at least one of the plurality of principal components, and when the sum of the distribution degrees of each principal component corresponding to the selected at least one principal component is equal to or greater than the preset standard value, the selected at least one principal component is selected as the target component.

7. The apparatus for diagnosing battery status according to claim 6, in, The battery classification unit is configured to calculate a sum by sequentially adding the distribution degrees of each principal component corresponding to each of the plurality of principal components from the largest one, until the sum becomes equal to or greater than the preset standard value.

8. The apparatus for diagnosing battery status according to claim 6, in, The battery classification unit is configured to generate, in each cycle, a representative model that classifies the plurality of batteries and the reference cell into the plurality of groups.

9. The apparatus for diagnosing battery status according to claim 8, in, The battery classification unit is configured to generate a target set including at least one selected target component, generate multiple target subsets including at least one target component from the generated target set, generate at least one classification model for classifying the multiple batteries and the reference cell for each of the multiple target subsets, and set any one of the multiple classification models generated for the multiple target subsets as a representative model for the corresponding cycle.

10. The apparatus for diagnosing battery status according to claim 9, in, The battery classification unit is configured to calculate the classification degree of the plurality of batteries and the reference cell for each of the plurality of generated classification models, and to set the classification model with the smallest calculated classification degree among the plurality of generated classification models as the representative model of the corresponding cycle.

11. The apparatus for diagnosing battery status according to claim 9, in, The battery classification unit is configured to calculate the inner product of the plurality of batteries and the reference cell for at least one target component included in each of the plurality of target subsets, and to generate the plurality of classification models for classifying the plurality of batteries and the reference cell based on the size of the plurality of calculated inner products.

12. The apparatus for diagnosing battery status according to claim 9, in, The battery classification unit is configured to control each of the plurality of classification models such that the total number of classified groups is equal to or less than a preset standard number.

13. A battery pack comprising the means for diagnosing battery status according to any one of claims 1 to 12.

14. An energy storage system comprising a means for diagnosing battery status according to any one of claims 1 to 12.

15. A method for diagnosing battery state, comprising: The curve generation step generates a battery curve for each of the plurality of batteries, representing the correspondence between the voltage and capacity of each of the plurality of batteries in each cycle; The feature extraction step extracts multiple feature values ​​from each of the plurality of battery curves generated in the curve generation step. The principal component calculation step calculates multiple principal components for the multiple eigenvalues ​​extracted in the eigenvalue extraction step; The principal component distribution degree calculation step involves calculating the distribution degree of each principal component representing the distribution of the multiple batteries based on the multiple feature values ​​extracted for each of the multiple batteries. The target component selection step involves selecting at least one of the plurality of principal components as the target component based on the result of comparing the distribution degree of each principal component calculated in the principal component distribution degree calculation step with a preset standard value. The battery classification step, based on the at least one target component selected in the target component selection step, classifies each of the plurality of batteries into any one of a plurality of groups; as well as The battery status diagnosis step updates the classification history of each group to which each of the plurality of batteries was classified in the battery classification step in each cycle, and diagnoses the status of each of the plurality of batteries based on the updated classification history.

Images