Electronic device for detecting battery abnormalities and its operating method

Abstract

An electronic device according to an embodiment disclosed herein can acquire input data through a sensing circuit, acquire standardized data for the input data by standardizing each column vector of the input data, acquire judgment criteria data based on the standardized data, and determine a first abnormal state and / or a second abnormal state of each of M battery cells based on values ​​indicated by entries of each row vector of the judgment criteria data. The first abnormal state may be determined based on a learning-based model, and the second abnormal state may be determined based on a method other than the learning-based model.

Description

[Technical Field]

[0001] [Cross-reference of related applications] The embodiments disclosed herein claim priority under Korean Patent Application No. 10-2022-0113110, filed on September 6, 2022, and include all the contents disclosed in the said Korean Patent Application as part of this Specification.

[0002] The embodiments disclosed herein relate to an electronic device for detecting battery abnormalities and a method for operating the same. [Background technology]

[0003] Due to their high applicability across product lines and electrical characteristics such as high energy density, secondary batteries are widely used not only in portable devices but also in electric vehicles (EVs) and hybrid electric vehicles (HEVs) powered by electric drive sources.

[0004] Currently, widely used types of rechargeable batteries include lithium-ion batteries, lithium polymer batteries, nickel-cadmium batteries, nickel-metal hydride batteries, and nickel-zinc batteries. The operating voltage of a single rechargeable battery cell is approximately 2.5 to 4.5 volts (V). Therefore, when a higher output voltage is required, multiple battery cells are connected in series to form a battery pack. Alternatively, depending on the required charge and discharge capacity of the battery pack, many battery cells may be connected in parallel to form a battery pack. Consequently, the number of battery cells included in a battery pack can be set in various ways depending on the required output voltage or charge and discharge capacity. [Overview of the Initiative] [Problems that the invention aims to solve]

[0005] A battery pack may contain multiple battery cells connected in series and / or parallel. Therefore, even if there is an abnormal battery cell among the multiple battery cells, it may not be easy to detect the abnormality of that battery cell.

[0006] This may necessitate a method for detecting abnormal battery cells among multiple battery cells.

[0007] The technical problems of the embodiments disclosed herein are not limited to those mentioned above, and any other technical problems not mentioned above should be clearly understood by those skilled in the art from the following description. [Means for solving the problem]

[0008] An electronic device according to one embodiment disclosed herein includes a battery module containing M battery cells, a sensing circuit in which M is an integer of 2 or more that acquires a state value relating to the state of each of the M battery cells, and a processor, wherein the electronic device acquires input data through the sensing circuit, the input data is representable as an M × N matrix, the M × N entries in the input data represent the state value of each of the M battery cells, N represents the number of time periods for acquiring the state value of each of the M battery cells, the processor acquires standardized data for the input data by standardizing each column vector of the input data, each column vector consists of entries acquired in the same unit time, the processor acquires judgment criterion data based on the standardized data, and determines a first state abnormality and / or a second state abnormality of each of the M battery cells based on the values ​​indicated by the entries in each row vector of the judgment criterion data, the first state abnormality is determined based on a learning platform model, and the second state abnormality may be determined based on a method other than the learning platform model.

[0009] In one embodiment, the processor may identify a row vector in the row vector of the judgment criterion data that has at least one value less than or equal to a reference critical standardized score, obtain an output vector for the identified row vector based on the learning platform model, and determine the first state abnormality of the battery cell corresponding to the identified row vector based on the output vector.

[0010] In one embodiment, the processor obtains the standardized data by removing the offset from each entry of the row vector of the input data and standardizing each of the column vectors of the input data from which the offset has been removed, wherein the offset is set for each of the row vectors of the input data, and the offset may be set to the value of the first entry of each entry of the row vector of the input data.

[0011] In one embodiment, the processor may obtain the standardized data by flattening each entry of the row vector of the input data and standardizing each of the column vectors of the flattened input data.

[0012] In one embodiment, the processor obtains the interval average value of each row vector of the standardized data as the criterion data, and determines the first and / or second state abnormalities of each of the M battery cells based on the interval average value of each row vector of the criterion data, wherein the interval average value is the average value of the entries of each row vector of the standardized data that are included in different time intervals, and the time intervals do not overlap with each other.

[0013] In one embodiment, the processor may acquire the change amount of each entry in the row vector of the standardized data as the judgment criterion data, and determine the first state abnormality and / or second state abnormality for each of the M battery cells based on the change amount of each entry in the row vector of the judgment criterion data.

[0014] In one embodiment, the processor may identify the sum of the entries in each of the row vectors of the judgment criterion data, identify row vectors in the judgment criterion data that have a sum less than or equal to a critical sum, and determine that the battery cell corresponding to the identified row vector is in the second abnormal state.

[0015] In one embodiment, the learning platform model is an autoencoder, and the autoencoder can be learned based on the state values ​​of a normal battery cell.

[0016] The operation method of an electronic device according to one embodiment disclosed herein includes the operation of acquiring input data through a sensing circuit of the electronic device, the input data being representable as an M×N matrix, where the M×N entries of the input data represent the state value of each of the M battery cells of the electronic device, where N represents the number of time periods for acquiring the state value of each of the M battery cells, the operation of acquiring standardized data for the input data by standardizing each of the column vectors of the input data, each of the column vectors consisting of entries acquired in the same unit time, the operation of acquiring judgment criterion data based on the standardized data, and the operation of determining a first state abnormality and / or a second state abnormality of each of the M battery cells based on the values ​​indicated by the entries of each row vector of the judgment criterion data, wherein the first state abnormality is determined based on a learning platform model, and the second state abnormality may be determined based on a method other than the learning platform model.

[0017] In one embodiment, the operation of determining the first state abnormality and / or the second state abnormality includes an operation of identifying a row vector having at least one value less than or equal to a reference critical normalization score among the row vectors of the determination reference data, an operation of obtaining an output vector for the identified row vector based on the learning base model, and an operation of determining the first state abnormality of the battery cell corresponding to the identified row vector based on the output vector.

[0018] In one embodiment, the operation of obtaining the normalization data includes an operation of removing an offset of each entry of the row vector of the input data, and an operation of obtaining the normalization data by normalizing each column vector of the input data from which the offset has been removed. The offset is set for each row vector of the input data, and the offset may be set to the value of the first entry among the entries of each row vector of the input data.

[0019] In one embodiment, the operation of obtaining the normalization data may include an operation of flattening each entry of the row vector of the input data, and an operation of obtaining the normalization data by normalizing each column vector of the flattened input data.

[0020] In one embodiment, the operation of obtaining the determination reference data includes an operation of obtaining an interval average value of each row vector of the normalization data as the determination reference data. The operation of determining the first state abnormality and / or the second state abnormality includes an operation of determining the first state abnormality and / or the second state abnormality of each of the M battery cells based on the interval average value of each row vector of the determination reference data. The interval average value is an average value of the entries of each row vector of the normalization data included in different time intervals, and the time intervals may not overlap with each other.

[0021] In one embodiment, the operation to acquire the judgment criterion data includes the operation to acquire the change amount of each entry in the row vector of the standardized data as the judgment criterion data, and the operation to determine the first state abnormality and / or the second state abnormality may include the operation to determine the first state abnormality and / or the second state abnormality for each of the M battery cells based on the change amount of each entry in the row vector of the judgment criterion data.

[0022] In one embodiment, the operation for determining the first and / or second state abnormality may include the operation of identifying the sum of the entries in each of the row vectors of the judgment criterion data, the operation of identifying a row vector in the judgment criterion data that has a sum less than or equal to a critical sum, and the operation of determining that the battery cell corresponding to the identified row vector is in the second state abnormality. [Effects of the Invention]

[0023] The various embodiments of the electronic devices and their operating methods disclosed herein can detect an abnormal battery cell from among multiple battery cells.

[0024] The effects of the electronic device for detecting battery abnormalities and its operating method disclosed herein are not limited to those mentioned above, and any other effects not mentioned herein will be clearly understood by those skilled in the art through the disclosure herein. [Brief explanation of the drawing]

[0025] [Figure 1] This is a block diagram of an electronic device in a network environment according to various embodiments of the present disclosure. [Figure 2] This is a flowchart showing the operation of an electronic device according to one embodiment of the present disclosure. [Figure 3] This is a flowchart showing the operation of an electronic device according to one embodiment of the present disclosure. [Figure 4a] This diagram illustrates the input data. [Figure 4b] This diagram illustrates smoothed input data. [Figure 4c] This diagram illustrates input data with the offset removed. [Figure 4d] This diagram illustrates standardized data of input data with the offset removed. [Figure 4e] This diagram illustrates standardized data with the offset removed. [Figure 4f] This figure illustrates interval-averaged data for standardized data. [Figure 4g] This figure illustrates the change in interval average data. [Figure 5] This is a flowchart showing the operation of an electronic device according to one embodiment of the present disclosure. [Figure 6] This is a flowchart showing the operation of an electronic device according to one embodiment of the present disclosure.

[0026] In relation to the description of the figure, the same or similar reference numerals may be used for the same or similar components. [Modes for carrying out the invention]

[0027] Embodiments of the present invention are described below with reference to the drawings. However, this should not be understood as limiting the present invention to any particular embodiment, but rather as including various modifications, equivalents, and / or alternatives to embodiments of the present invention.

[0028] The embodiments and terminology used herein should be understood not as limiting the technical features described herein to any particular embodiment, but as including a variety of modifications, equivalents, or substitutions of such embodiments. In relation to the descriptions of the figures, similar or related reference numerals may be used for similar or related components. A singular noun corresponding to an item may include one or more of such items unless otherwise clearly indicated in the context.

[0029] In this document, each of the phrases such as "A or B," "at least one of A and B," "at least one of A or B," "A, B or C," "at least one of A, B and C," and "at least one of A, B or C" may include any one of the items listed together with the phrase, or any possible combination thereof. Terms such as "first," "second," "first," "second," "A," "B," "(a)," or "(b)" may be used merely to distinguish one component from other components and, unless otherwise stated, do not limit the component in any other way (e.g., importance or order).

[0030] Wherever a component (e.g., Component 1) is referred to as being "functionally" or "communically" connected to another component (e.g., Component 2), or even without such terms, as being "coupled," "connected," or "coupled," it means that the component can be connected to the other component directly (e.g., by wire or wireless) or indirectly (e.g., via Component 3).

[0031] The methods described herein in various embodiments may be provided as part of a computer program product. A computer program product may be traded as a commodity between a seller and a buyer. A computer program product may be distributed in the form of a device-readable recording medium (e.g., compact disc read-only memory (CD-ROM)) or online (e.g., download or upload) through an application store or directly between two user devices. In the case of online distribution, at least a portion of the computer program product may be temporarily recorded or temporarily generated on a device-readable recording medium such as the memory of a manufacturer's server, an application store server, or an intermediary server.

[0032] According to the embodiments disclosed herein, each of the aforementioned components (e.g., modules or programs) may include one or more individuals, some of which may be separated and arranged in other components. According to the embodiments disclosed herein, one or more components or operations in the aforementioned components may be omitted, or one or more other components or operations may be added. Generally or additionally, multiple components (e.g., modules or programs) may be integrated into a single component. In such a case, the integrated component may perform one or more functions of each of the multiple components in the same or similar manner as they were performed by the components in the multiple components prior to the integration. According to the embodiments disclosed herein, operations performed by modules, programs or other components may be performed sequentially, in parallel, iteratively, or heuristically, or one or more of the operations may be performed in a different order, omitted, or one or more other operations may be added.

[0033] Figure 1 is a block diagram of an electronic device 101 according to one embodiment of the present disclosure.

[0034] As shown in Figure 1, the electronic device 101 may include a battery module 110, a sensor circuit 120, a communication circuit 130, a memory 140, and a processor 150. Depending on the embodiment, at least one component of the electronic device 101 shown in Figure 1 (e.g., the battery module 110) may be replaced by another component (e.g., a battery cell containing multiple battery modules 110). Depending on the embodiment, at least one component of the electronic device 101 shown in Figure 1 (e.g., the battery module 110) may consist of multiple units. Depending on the embodiment, at least one component of the electronic device 101 shown in Figure 1 (e.g., the sensor circuit 120 or the communication circuit 130) may be excluded from the electronic device 101. Depending on the embodiment, the electronic device 101 may further include at least one component other than the components shown in Figure 1 (e.g., a power unit (e.g., a motor), a display, an input device, or an output device).

[0035] In one embodiment, the electronic device 101 may be a battery management system (BMS). When the electronic device 101 is embodied as a battery management system, it can supply power from the battery module 110 to external components (e.g., a motor).

[0036] In one embodiment, the electronic device 101 may be a battery swapping station (BSS). When the electronic device 101 is embodied as a battery swapping station, it may include multiple slots for storing and / or charging multiple battery modules 110.

[0037] In one embodiment, the battery module 110 can supply power to one or more components of the electronic device 101. In one embodiment, the battery module 110 may be detachable from the electronic device 101.

[0038] In one embodiment, the battery module 110 may include one or more battery cells 111, 113, or 115. The one or more battery cells 111, 113, or 115 may be included in the battery module 110 in an electrically connected manner. For example, the one or more battery cells 111, 113, or 115 may be connected in series and / or in parallel. Depending on the embodiment, the one or more battery cells 111, 113, or 115 may be included in the battery module 110 in an electrically isolated manner.

[0039] In one embodiment, the sensor circuit 120 can acquire information related to the battery module 110. In one embodiment, the sensor circuit 120 can acquire values ​​(or information) related to the state of one or more battery cells 111, 113, or 115. In one embodiment, the state-related values ​​may represent one or more values ​​for the voltage, current, resistance, state of charge (SOC), state of health (SOH), or temperature of the battery cell, or a combination thereof. Hereinafter, the state-related values ​​may be referred to as "state values".

[0040] In one embodiment, the sensor circuit 120 can provide information (e.g., status values) for each of the one or more battery cells 111, 113, or 115 to the processor 150.

[0041] In one embodiment, the communication circuit 130 can establish a wired communication channel and / or a wireless communication channel between the electronic device 101 and the external electronic device 102, and send / receive data to and from the external electronic device 102 through the established communication channel.

[0042] In one embodiment, the communication circuit 130 can perform communication based on at least one radio access technology (RAT). In one embodiment, the communication circuit 130 can send / receive data with an external electronic device 102 using at least one RAT.

[0043] In one embodiment, the memory 140 may include volatile memory and / or non-volatile memory.

[0044] In one embodiment, the memory 140 can store data used by at least one component of the electronic device 101 (e.g., the processor 150 or the sensor circuit 120). For example, the data may include software (or associated instructions), input data, or output data. In one embodiment, the instructions may cause the electronic device 101 to perform an operation defined by the instructions when executed by the processor 150.

[0045] In one embodiment, the processor 150 can execute software to control at least one other component (e.g., hardware or software component) of the electronic device 101 connected to the processor 150, and can perform a variety of data processing or calculations.

[0046] The following describes how an electronic device 101 according to one embodiment of the present disclosure determines an abnormality in one or more battery cells 111, 113, or 115 included in a battery module 110.

[0047] In one embodiment, the electronic device 101 can acquire the state values ​​(e.g., one or more values ​​related to voltage, current, resistance, charge status, health status, or temperature) of one or more battery cells 111, 113, or 115. In one embodiment, the electronic device 101 can acquire the state values ​​of one or more battery cells 111, 113, or 115 during a specified time interval (e.g., 180 seconds (sec)). In one embodiment, the electronic device 101 can acquire the state values ​​of one or more battery cells 111, 113, or 115 at a specified time period (e.g., 0.1 seconds) during a specified time interval. For example, the electronic device 101 can acquire the voltage values ​​of one or more battery cells 111, 113, or 115. Another example is that the electronic device 101 can acquire the current values ​​of one or more battery cells 111, 113, or 115.

[0048] In one embodiment, the state values ​​of one or more battery cells 111, 113, or 115 acquired at a specified time interval during a specified time interval may be referred to as "input data". In one embodiment, the input data may be represented by a matrix (e.g., an M × N matrix) consisting of N values ​​for each of the M battery cells 111, 113, or 115. For example, the input data may consist of N voltage values ​​for each of the M battery cells 111, 113, or 115. Another example is that the input data may consist of N values ​​for current, resistance, charge state, health state, or temperature for each of the M battery cells 111, 113, or 115. For example, an entry in the input data may represent the voltage value for each of the battery cells 111, 113, or 115. Another example is that an entry in the input data may represent a value for one of the resistance, charge state, health state, or temperature for each of the battery cells 111, 113, or 115. Here, M may be the number of battery cells 111, 113, or 115, and N may be the number of time periods required to acquire the state values ​​of battery cells 111, 113, or 115. For example, if the number of battery cells 111, 113, or 115 is 100, the specified time interval is 180 seconds, and the specified time period is 0.1 seconds, then M may be 100 and N may be 1800.

[0049] In one embodiment, input data (D input ) can be expressed as shown in Formula 1 below. [Formula 1]

number

[0050] In one embodiment, the electronic device 101 can acquire normalized data based on the input data. In one embodiment, the normalized data for arbitrary matrix data may be acquired based on the normalization between the entries of each column vector of the arbitrary matrix data. In one embodiment, the normalization may be to convert the value into a normalized score (or Z - score). For example, the normalization can be based on Equation 2 below. [Equation 2] [Number] In Equation 2, z i,j may be the normalized score of the j - th entry of the i - th row vector of the arbitrary matrix data. vi,j m may be the value of the j-th entry of the i-th row vector of an arbitrary matrix data. j σ may be the mean of the entries in the j-th column vector of an arbitrary matrix data. j `i` may be the standard deviation of the entry in the j-th column vector of the arbitrary matrix data. Here, `i` may be an integer less than or equal to the number of rows in the arbitrary matrix data (1 or more), and `j` may be an integer less than or equal to the number of columns in the arbitrary matrix data (1 or more).

[0051] In one embodiment, the electronic device 101 can determine a first and / or second abnormal state of battery cells 111, 113, or 115 based on standardized data. In one embodiment, the electronic device 101 can determine a first and / or second abnormal state of battery cells 111, 113, or 115 using criterion data based on standardized data. In one embodiment, the criterion data may be the standardized data itself. In one embodiment, the criterion data may be data from which data processing (e.g., offset removal, interval averaging, change amount acquisition, and / or concatenation) has been performed on the standardized data. In one embodiment, a first abnormal state may mean an abnormal state in which the instantaneous voltage of the battery cell drops when the battery cell is being charged. In one embodiment, a second abnormal state may mean an abnormal state in which the slope of the voltage of the battery cell decreases (or is low) when the battery cell is being charged. In one embodiment, a first abnormal state may be determined based on a learning-based model, and a second abnormal state may be determined based on a method other than a learning-based model (e.g., a rule-based model).

[0052] The following describes how the electronic device 101 determines the first abnormal state of the battery cell 111, 113, or 115.

[0053] In one embodiment, the electronic device 101 can identify row vectors containing entries in each row vector of the judgment criterion data that have a value less than or equal to the reference critical standardized score. In another embodiment, the electronic device 101 can identify row vectors containing entries in each row vector of the judgment criterion data that have a value greater than or equal to the reference critical standardized score. In yet another embodiment, the electronic device 101 can identify row vectors containing entries in each row vector of the judgment criterion data that have a value outside the range of the reference critical standardized score. In one embodiment, the reference critical standardized score may be a score experimentally determined to extract candidates that may be first state abnormalities. Hereinafter, an example is given in which the electronic device 101 identifies row vectors containing entries in each row vector of the judgment criterion data that have a value less than or equal to the reference critical standardized score.

[0054] In one embodiment, the electronic device 101 can obtain an output vector of a learning base model based on row vectors of criterion data. In one embodiment, the electronic device 101 can obtain an output vector for row vectors containing entries having values ​​less than or equal to a criterion critical standardized score based on the learning base model. In one embodiment, the learning base model may be an autoencoder. In one embodiment, the autoencoder may be an artificial neural network trained to reconstruct criterion data by normal behavior. In one embodiment, the autoencoder may be an artificial neural network trained in an unsupervised manner. In one embodiment, the autoencoder may include an encoder and a decoder. In one embodiment, the encoder of the autoencoder can compress the input to a lower dimension, and the decoder of the autoencoder can reconstruct the compressed output of the autoencoder to its original dimension. This allows the input and output of the autoencoder to have the same dimension, and the output of the autoencoder may have a form that includes the reconstruction error (or reconstruction loss) in the input of the autoencoder.

[0055] In one embodiment, the electronic device 101 can determine a first state abnormality of a battery cell based on the output vector of a learning base model. In one embodiment, the electronic device 101 can determine a first state abnormality of a battery cell based on the restoration error of an autoencoder. In one embodiment, the restoration error for an arbitrary row vector may be the RMS (root mean square) value of the vector obtained by subtracting the output vector for the arbitrary row vector from the arbitrary row vector.

[0056] In one embodiment, the electronic device 101 can identify row vectors in the judgment criterion data in which the restoration error exceeds a specified reference restoration error. In one embodiment, the electronic device 101 can determine that the battery cell associated with the row vector exceeding the reference restoration error is in a first state abnormality. In one embodiment, the reference restoration error may be an error experimentally determined for determining the first state abnormality.

[0057] The following describes how the electronic device 101 determines the second abnormal state of the battery cell 111, 113, or 115.

[0058] In one embodiment, the electronic device 101 can determine a second state abnormality of battery cells 111, 113, or 115 based on the sum of the entries in each row vector of the judgment criterion data. In one embodiment, the electronic device 101 can identify row vectors in the row vector whose sum is less than or equal to a specified criterion sum. In another embodiment, the electronic device 101 can identify row vectors in the row vector whose sum is greater than or equal to a specified criterion sum. In yet another embodiment, the electronic device 101 can identify row vectors in the row vector whose sum is outside the range of the specified criterion sum. Hereinafter, we will exemplify an example in which the electronic device 101 identifies row vectors in the row vector whose sum is greater than or equal to a specified criterion sum. In one embodiment, the criterion sum may be a value experimentally determined for determining the second state abnormality.

[0059] In one embodiment, the electronic device 101 can determine that a battery cell associated with a row vector whose sum is less than or equal to a specified reference sum is in a second abnormal state.

[0060] In one embodiment, if the electronic device 101 detects a battery cell exhibiting an abnormal condition, it may provide a notification to the user via an output device (or display, communication circuit 130). In one embodiment, the notification provided to the user may include a notification prompting the user to check for an abnormality in the battery cell, a notification prompting the user to replace the battery cell, or a combination thereof.

[0061] Figure 1 illustrates that the electronic device 101 acquires standardized data from the input data and determines a first state abnormality and / or a second state abnormality based on the standardized data's judgment criteria; however, this is merely an illustration. Depending on the embodiment, the electronic device 101 can perform processing to modify the input data and determine a first state abnormality and / or a second state abnormality based on the processed standardized data from the input data. Furthermore, the electronic device 101 can perform processing to modify the standardized data from the input data and determine a first state abnormality and / or a second state abnormality based on the processed standardized data (or judgment criteria data). Depending on the embodiment, the processing of the input data and / or the processing of the standardized data may be performed selectively.

[0062] The following describes an example in which an electronic device 101 according to one embodiment of this disclosure processes input data.

[0063] In one embodiment, the electronic device 101 can obtain smoothed data (or moving average data) of the input data. In one embodiment, the smoothing of arbitrary matrix data may be performed between each entry of the row vector of the arbitrary matrix data. For example, the smoothing of arbitrary matrix data may be performed based on the following equation 3. [Formula 3]

number

[0064] In one embodiment, the electronic device 101 can acquire smoothed data (or moving average data) of the input data, acquire standardized data of the acquired smoothed data (or moving average data), and determine a first abnormal state and / or a second abnormal state based on the standardized data.

[0065] In one embodiment, the electronic device 101 can obtain data from which the offset has been removed from the input data. In one embodiment, the offset of the arbitrary matrix data may be set for each row vector of the arbitrary matrix data. In one embodiment, the offset for the arbitrary matrix data may be the value of the first entry among the entries of each row vector of the arbitrary matrix data. In one embodiment, the removal of the offset for the arbitrary matrix data may be performed by subtracting the offset of each row vector from the entry of each row vector of the arbitrary matrix data.

[0066] In one embodiment, the electronic device 101 can acquire data from which the offset of the input data has been removed, acquire standardized data from the acquired data from which the offset has been removed, and determine a first abnormal state and / or a second abnormal state based on the standardized data.

[0067] In one embodiment, the electronic device 101 can sequentially perform offset removal and smoothing on the input data. In one embodiment, the electronic device 101 can obtain standardized data from the data obtained by sequentially performing offset removal and smoothing, and can determine a first abnormal state and / or a second abnormal state based on the standardized data. Depending on the embodiment, offset removal may be performed on the input data first, followed by smoothing, or smoothing may be performed on the input data first, followed by offset removal.

[0068] The following describes an example in which an electronic device 101 according to one embodiment of the present disclosure processes standardized data to obtain judgment criterion data. The standardized data described below may be standardized data of input data, standardized data of smoothed input data, standardized data of input data with the offset removed, or standardized data of input data that has been smoothed and the offset removed.

[0069] In one embodiment, the electronic device 101 can obtain judgment criterion data by removing the offset from the standardized data. In one embodiment, the electronic device 101 can determine a first abnormal state and / or a second abnormal state based on the standardized data from which the offset has been removed (i.e., judgment criterion data).

[0070] In one embodiment, the electronic device 101 can acquire interval average data of standardized data as criterion data. In one embodiment, the interval average may be the average of each entry in the row vector of arbitrary matrix data at a specified time interval (e.g., 18 seconds). In one embodiment, the time intervals do not need to overlap. For example, if the specified time interval is 18 seconds, the specified time interval is 180 seconds, and the specified time period is 0.1 seconds, the number of columns in the interval average data is 10, and the columns of interval average data can represent the average of the 180 columns of the standardized data.

[0071] In one embodiment, the electronic device 101 can determine a first abnormal state and / or a second abnormal state based on interval average data (i.e., judgment criterion data).

[0072] In one embodiment, the electronic device 101 can acquire change amount data of standardized data as criterion data. In one embodiment, the change amount may be the difference between adjacent entries in each row vector of arbitrary matrix data. For example, the change amount of an arbitrary entry may be the value obtained by subtracting the value of the entry prior to the arbitrary entry from the value of the arbitrary entry.

[0073] In one embodiment, the electronic device 101 can determine a first abnormal state and / or a second abnormal state based on change amount data (i.e., judgment criterion data).

[0074] In one embodiment, the electronic device 101 can acquire concatenated standardized data as judgment criterion data. In one embodiment, the concatenation of arbitrary matrix data may be by concatenating arbitrary matrix data with other matrix data in the column direction. In one embodiment, the concatenation of arbitrary matrix data may be by sequentially concatenating arbitrary matrix data excluding the first column, the first column of the arbitrary matrix data, and the last column of the arbitrary matrix data. In one embodiment, the electronic device 101 can determine a first state abnormality and / or a second state abnormality based on the concatenated data (i.e., judgment criterion data).

[0075] In one embodiment, the electronic device 101 can acquire judgment criterion data by sequentially performing offset removal, interval averaging, change amount acquisition, and concatenation on standardized data. In one embodiment, the electronic device 101 can determine a first abnormal state and / or a second abnormal state based on the judgment criterion data acquired by sequentially performing offset removal, interval averaging, change amount acquisition, and concatenation on standardized data. The order of processing between offset removal, interval averaging, change amount acquisition, and concatenation on standardized data may be set differently for each embodiment.

[0076] Figure 2 is a flowchart showing the operation of an electronic device according to one embodiment of the present disclosure. The operation shown in Figure 2 may be performed by the electronic device 101 shown in Figure 1.

[0077] As shown in Figure 2, in operation 210, the electronic device 101 can acquire input data. In one embodiment, the input data may include the state values ​​of one or more battery cells 111, 113, or 115 acquired at specified time periods during a specified time interval. In one embodiment, the input data may be represented by a matrix (e.g., an M × N matrix) consisting of N values ​​for each of the M battery cells 111, 113, or 115. For example, an entry in the input data may represent the voltage value of each battery cell 111, 113, or 115. Another example is that an entry in the input data may represent a value for one of the following: resistance, charge state, health state, or temperature of each battery cell 111, 113, or 115, where M is the number of battery cells 111, 113, or 115, and N may be the number of time periods for acquiring the state values ​​of the battery cells 111, 113, or 115.

[0078] In operation 220, the electronic device 101 can acquire standardized data. In one embodiment, the standardized data may be data standardized between each entry of the column vector of input data.

[0079] In operation 225, the electronic device 101 can acquire criterion data based on standardized data. In one embodiment, the criterion data may be the standardized data itself. In one embodiment, the criterion data may be data from which data processing (e.g., offset removal, interval averaging, change amount acquisition and / or concatenation) has been performed on the standardized data.

[0080] In operation 230, the electronic device 101 can determine a first state abnormality. In one embodiment, the electronic device 101 can obtain the output vector of the learning base model based on the row vector of the judgment criterion data. In one embodiment, the electronic device 101 can determine a first state abnormality of the battery cell based on the output vector of the learning base model.

[0081] In one embodiment, the electronic device 101 can determine a first abnormal state of the battery cell based on the restoration error of the autoencoder. In one embodiment, the restoration error for an arbitrary row vector may be the RMS value of the vector obtained by subtracting the output vector for the arbitrary row vector from the arbitrary row vector.

[0082] In one embodiment, the electronic device 101 can identify row vectors in the judgment criterion data in which the restoration error exceeds a specified reference restoration error. In one embodiment, the electronic device 101 can determine that the battery cell associated with the row vector exceeding the reference restoration error is in a first abnormal state.

[0083] In operation 240, the electronic device 101 can determine a second state abnormality. In one embodiment, the electronic device 101 can determine a second state abnormality of battery cells 111, 113, or 115 based on the sum of the entries in each row vector of the judgment criterion data. In one embodiment, the electronic device 101 can identify row vectors in the row vector whose sum is less than or equal to a specified criterion sum. In one embodiment, the electronic device 101 can determine that a battery cell associated with a row vector whose sum is less than or equal to a specified criterion sum is in a second state abnormality.

[0084] Figure 2 illustrates an example in which the electronic device 101 acquires standardized data from the input data and determines a first state abnormality and / or a second state abnormality based on the standardized data's judgment criteria; however, this is merely an example. Depending on the embodiment, the electronic device 101 can perform processing to modify the input data and determine a first state abnormality and / or a second state abnormality based on the processed standardized data from the input data. Furthermore, the electronic device 101 can determine a first state abnormality and / or a second state abnormality based on judgment criteria data acquired by performing processing to modify the standardized data from the input data. Depending on the embodiment, the processing of the input data and / or the processing of the standardized data may be performed selectively.

[0085] In one embodiment, the method by which the electronic device 101 processes input data may include offset removal from the input data and smoothing of the input data. In one embodiment, the method by which the electronic device 101 processes standardized data to obtain judgment criterion data may include offset removal from the standardized data, interval averaging of the standardized data, and acquisition of change amount from the standardized data.

[0086] Figure 2 shows that operations 230 and 240 are both performed, but this is merely an example. Depending on the embodiment, operations 230 and 240 may be performed selectively. In some embodiments, either operation 230 or operation 240 may not be performed. For example, the electronic device 101 may perform operation 230 and not operation 240 in order to determine only the first abnormal state. Another example is the electronic device 101, which may perform operation 240 and not operation 230 in order to determine only the second abnormal state.

[0087] Figure 3 is a flowchart illustrating the operation of an electronic device according to one embodiment of the present disclosure. Figure 4a is a diagram illustrating input data. Figure 4b is a diagram illustrating smoothed input data. Figure 4c is a diagram illustrating offset-removed input data. Figure 4d is a diagram illustrating standardized data of offset-removed input data. Figure 4e is a diagram illustrating standardized data of offset-removed data. Figure 4f is a diagram illustrating interval-averaged data of standardized data. Figure 4g is a diagram illustrating change amount data of interval-averaged data. Operations 310, 320, and 330 in Figure 3 may be included in operation 220 in Figure 2. Operations 340, 350, and 360 in Figure 3 may be included in operation 225 in Figure 2. The operations in Figure 3 may be performed by the electronic device 101 shown in Figure 1. In one embodiment, the change amount data in Figure 4g may be understood as criterion data obtained by the operations in Figure 3.

[0088] As shown in Figure 3, in operation 310, the electronic device 101 can perform smoothing on the input data.

[0089] In one embodiment, the input data may include state values ​​for one or more battery cells 111, 113, or 115 acquired at specified time intervals within a specified time interval. In one embodiment, the input data may be represented by a matrix (e.g., an M × N matrix) consisting of N values ​​for each of the M battery cells 111, 113, or 115. In one embodiment, the state values ​​may be one or more values ​​relating to voltage, current, resistance, charge state, health state, or temperature.

[0090] As shown in Figure 4a, graph 401 can represent the input data 410. The input data 410 may be the voltage values ​​of M battery cells 111, 113, or 115 acquired during a time interval of 180 seconds.

[0091] In one embodiment, the smoothing of the input data 410 may be performed between each entry of the row vector of the input data. For example, the smoothing of the input data 410 may be performed based on the formula 3.

[0092] As shown in Figure 4b, Graph 402 can be seen to represent the smoothed input data 420. Compared to Graph 401, Graph 402 shows that the noise in the input data 420 has been reduced by smoothing.

[0093] In operation 320, the electronic device 101 can remove the offset from the smoothed input data 420. In one embodiment, the offset from the smoothed input data 420 may be set for each row vector of the smoothed input data 420. In one embodiment, the offset for the smoothed input data 420 may be the value of the first entry among the entries of each row vector of the smoothed input data 420. In one embodiment, the removal of the offset from the smoothed input data 420 may be done by subtracting the offset of each row vector from each entry of the row vector of the smoothed input data 420.

[0094] As shown in Figure 4c, Graph 403 can show the input data 430 with the offset removed. Compared to Graph 402, Graph 403 shows that the initial value of the input data 430 starts from 0.

[0095] In operation 330, the electronic device 101 can perform standardization on the offset-removed input data 430. In one embodiment, the standardized data for the input data 430 can be obtained based on standardization between each entry in the column vector of the input data 430. In one embodiment, standardization may be the conversion of values ​​into standardized scores (or Z scores). For example, standardization may be based on formula 2.

[0096] As shown in Figure 4d, Graph 404 can show the standardized data 440 of the input data 430.

[0097] In operation 340, the electronic device 101 can remove the offset from the standardized data 440.

[0098] In one embodiment, the offset of the standardized data 440 may be set for each row vector of the standardized data 440. In one embodiment, the offset for the standardized data 440 may be the value of the first entry among the entries of each row vector of the standardized data 440. In one embodiment, offset removal for the standardized data 440 may be performed by subtracting the offset of each row vector from each entry of the row vector of the standardized data 440.

[0099] As shown in Figure 4e, Graph 405 can show the standardized data 450 with the offset removed.

[0100] In operation 350, the electronic device 101 can perform interval averaging of the offset-removed standardized data 450. In one embodiment, the interval average may be the average of each entry in the row vector of the standardized data 450 over a specified time interval (e.g., 18 seconds). For example, if the specified time interval is 18 seconds, the specified time interval is 180 seconds, and the specified time period is 0.1 seconds, then there are 10 columns of interval-averaged data, and the columns of interval-averaged data may represent the average of the 180 columns of the standardized data.

[0101] As shown in Figure 4f, Graph 406 can show the interval-averaged data 460. Compared to Graph 405 in Figure 4e, it can be seen that Graph 406 has 10 intervals.

[0102] In operation 360, the electronic device 101 can obtain the change in the interval average data 460. In one embodiment, the change may be the difference between adjacent entries in each row vector of the interval average data 460. For example, the change in an arbitrary entry may be the value obtained by subtracting the values ​​of entries prior to the arbitrary entry from the value of the arbitrary entry.

[0103] As shown in Figure 4g, Graph 407 can represent the change data 470. Compared to Graph 407 in Figure 4g, Graph 407 has 9 intervals. This may be because there is one entry in Graph 407 for which no change data is obtained (for example, the last entry).

[0104] Subsequently, the electronic device 101 can determine a first and / or second state abnormality of the battery cell 111, 113, or 115 based on the change amount data 470.

[0105] Figure 5 is a flowchart illustrating the operation of an electronic device according to one embodiment of the present disclosure. The operation in Figure 5 may be included in operation 230 of Figure 2. The operation in Figure 5 may be performed by the electronic device 101 shown in Figure 1.

[0106] The operation shown in Figure 5 may be performed for each row vector.

[0107] As shown in Figure 5, in operation 510, the electronic device 101 can determine whether a row vector contains entries less than or equal to the reference critical standardized score. In one embodiment, the electronic device 101 can identify row vectors containing entries among the entries of each row vector of the judgment criterion data that have values ​​less than or equal to the reference critical standardized score. Depending on the embodiment, the row vector may be a row vector of change amount data 470 shown in Figure 4f. Depending on the embodiment, the row vector may be a row vector of data obtained by concatenating the change amount data 470. In one embodiment, the row vector may be a row vector of data obtained by sequentially concatenating the change amount data 470 excluding the first column, the first column of the change amount data 470, and the last column of the change amount data 470.

[0108] In operation 520, the electronic device 101 can obtain an output vector of row vectors based on a learning base model. In one embodiment, the learning base model may be an autoencoder. In one embodiment, the autoencoder may be an artificial neural network trained to reconstruct judgment criterion data based on normal behavior. In one embodiment, the output vector may have the same dimensions as the row vectors and may include reconstruction error (or reconstruction loss).

[0109] In operation 530, the electronic device 101 can determine whether the restoration error between the row vector and the output vector exceeds the reference restoration error. In one embodiment, the restoration error between the row vector and the output vector may be the RMS value of the vector obtained by subtracting the row vector from the output vector.

[0110] In one embodiment, if the restoration error exceeds the reference restoration error, the electronic device 101 can perform operation 540. In one embodiment, if the restoration error does not exceed the reference restoration error, the electronic device 101 can perform operation 550.

[0111] In operation 540, the electronic device 101 can determine that the first state abnormality is present. In one embodiment, the electronic device 101 can determine that the battery cell corresponding to the row vector is in the first state abnormality.

[0112] In operation 550, the electronic device 101 can determine that the first state abnormality is not present. In one embodiment, the electronic device 101 can determine that the battery cell corresponding to the row vector is not in the first state abnormality.

[0113] Figure 6 is a flowchart illustrating the operation of an electronic device according to one embodiment of the present disclosure. The operation in Figure 6 may be included in operation 240 of Figure 2. The operation in Figure 6 may be performed by the electronic device 101 shown in Figure 1.

[0114] The operation shown in Figure 6 may be performed for each row vector.

[0115] As shown in Figure 6, in operation 610, the electronic device 101 can identify the sum of the entries in the row vector. Depending on the embodiment, the row vector may be the row vector of the change amount data 470 shown in Figure 4f.

[0116] In operation 620, the electronic device 101 can determine whether the sum is less than or equal to the standard sum.

[0117] In one embodiment, if the sum is less than or equal to the reference sum, the electronic device 101 can perform operation 630. In one embodiment, if the sum is not less than or equal to the reference sum, the electronic device 101 can perform operation 640.

[0118] In operation 630, the electronic device 101 can determine that the second state abnormality is present. In one embodiment, the electronic device 101 can determine that the battery cell corresponding to the row vector is in the second state abnormality.

[0119] In operation 640, the electronic device 101 can determine that the second state abnormality is not present. In one embodiment, the electronic device 101 can determine that the battery cell corresponding to the row vector is not in a second state abnormality.

Claims

1. An electronic device, A battery module containing M battery cells, where M is an integer of 2 or more, and A sensing circuit that acquires a state value related to the state of each of the M battery cells, A processor is included, and the processor is Input data is acquired through the sensing circuit, the input data can be represented as an M x N matrix, the M x N entries in the input data represent the state value of each of the M battery cells, and N represents the number of time periods required to acquire the state value of each of the M battery cells. By standardizing each of the column vectors of the input data, standardized data for the input data is obtained, and each of the column vectors consists of entries obtained in the same unit of time. Based on the standardized data, judgment criterion data is obtained, Based on the values ​​indicated by each entry in the row vector of the aforementioned judgment criteria data, the first and / or second state abnormalities of each of the M battery cells are determined. The aforementioned first state anomaly is determined based on the learning-based model, The second abnormal state is determined based on a method other than the learning platform model. electronic equipment.

2. The aforementioned processor, Among the row vectors of the aforementioned judgment criteria data, identify a row vector that has at least one value less than or equal to the standard critical standardized score, Based on the aforementioned learning infrastructure model, an output vector is obtained for the identified row vector. Based on the output vector, the first abnormal state of the battery cell corresponding to the identified row vector is determined. The electronic device according to claim 1.

3. The aforementioned processor, Remove the offset from each entry of the row vector of the input data, The standardized data is obtained by standardizing each of the column vectors of the input data from which the offset has been removed. The offset is set for each of the row vectors of the input data, The offset is set to the value of the first entry among the entries in each row vector of the input data. The electronic device according to claim 1.

4. The aforementioned processor, The entries of each row vector of the input data are flattened, The standardized data is obtained by standardizing each of the column vectors of the flattened input data. The electronic device according to claim 1.

5. The aforementioned processor, The interval mean values ​​of each row vector of the standardized data are obtained as the judgment criterion data. Based on the interval average value of each row vector of the judgment criterion data, the first abnormal state and / or second abnormal state of each of the M battery cells are determined. The interval average is the average value of each entry in the row vector of the standardized data that is included in different time intervals. The aforementioned time intervals do not overlap with each other. The electronic device according to claim 1.

6. The aforementioned processor, The change in the entry amount for each row vector of the standardized data is obtained as the judgment criterion data. Based on the amount of change in each entry of the row vector of the judgment criterion data, the first state abnormality and / or second state abnormality of each of the M battery cells are determined. The electronic device according to claim 1.

7. The aforementioned processor, Identify the sum of the entries in each row vector of the judgment criterion data, Among the row vectors of the judgment criteria data, row vectors having a sum value less than or equal to the critical sum value are identified. The battery cell corresponding to the identified row vector is determined to be in the second abnormal state. The electronic device according to claim 1.

8. The aforementioned learning platform model is an autoencoder, The autoencoder is learned based on the state values ​​of a normal battery cell. The electronic device according to claim 1.

9. A method for operating an electronic device, The operation involves acquiring input data through the sensing circuit of the electronic device, wherein the input data can be represented by an M × N matrix, and the M × N entries in the input data represent the state value of each of the M battery cells of the electronic device, and N represents the number of time periods required to acquire the state value of each of the M battery cells. The operation involves obtaining standardized data for the input data by standardizing each of the column vectors of the input data, where each of the column vectors consists of entries acquired in the same unit of time. The operation of obtaining judgment criterion data based on the standardized data, and The operation includes determining a first state abnormality and / or a second state abnormality for each of the M battery cells based on the value indicated by each entry in the row vector of the judgment criterion data, The aforementioned first state anomaly is determined based on the learning-based model, The second abnormal state is determined based on a method other than the learning platform model. method.

10. The operation for determining the first and / or second status abnormality is: An operation to identify row vectors in the aforementioned judgment criterion data that have at least one value less than or equal to the criterion critical standardized score. Based on the aforementioned learning infrastructure model, the operation of obtaining an output vector for the identified row vector, and The operation includes determining the first abnormal state of the battery cell corresponding to the identified row vector based on the output vector, The method according to claim 9.

11. The operation to obtain the aforementioned standardized data is: The operation of removing the offset from each entry of the row vector of the input data, and This includes an operation to obtain the standardized data by standardizing each of the column vectors of the input data from which the offset has been removed, The offset is set for each of the row vectors of the input data, The offset is set to the value of the first entry among the entries in each row vector of the input data. The method according to claim 9.

12. The operation to obtain the aforementioned standardized data is: The operation of flattening each entry in the row vector of the input data, and The operation includes obtaining the standardized data by standardizing each of the column vectors of the flattened input data, The method according to claim 9.

13. The operation to acquire the aforementioned judgment criterion data is as follows: This includes the operation of obtaining the interval average value of each row vector of the standardized data as the judgment criterion data, The operation for determining the first and / or second status abnormality is: The operation includes determining a first state abnormality and / or a second state abnormality for each of the M battery cells based on the interval average value of each of the row vectors of the judgment criterion data, The interval average is the average value of each entry in the row vector of the standardized data that is included in different time intervals. The aforementioned time intervals do not overlap with each other. The method according to claim 9.

14. The operation to acquire the aforementioned judgment criterion data is as follows: This includes the operation of obtaining the change amount of each entry in the row vector of the standardized data as the judgment criterion data, The operation for determining the first and / or second status abnormality is: The operation includes determining a first state abnormality and / or a second state abnormality for each of the M battery cells based on the amount of change of the entry for each of the row vectors of the judgment criterion data, The method according to claim 9.

15. The operation for determining the first and / or second status abnormality is: An operation to identify the sum of the entries in each row vector of the judgment criterion data, An operation to identify row vectors in the row vectors of the judgment criteria data that have a sum value less than or equal to a critical sum value, and The operation includes determining that the battery cell corresponding to the identified row vector is in the second abnormal state, The method according to claim 9.

Images