Data analysis device and operating method thereof

Abstract

According to an embodiment disclosed in the present document, a data analysis device may comprise: an information acquisition unit for acquiring a plurality of data sets, each of which includes time series data for a plurality of parameters; and a controller for extracting a feature of each of the plurality of data sets on the basis of a correlation between the plurality of parameters, extracting a pattern of each of the plurality of data sets on the basis of similarity between the extracted features, and classifying the plurality of data sets into a plurality of clusters on the basis of the pattern.

Description

Data analysis device and its operating method

[0001] Cross-citation with related applications

[0002] This invention claims the benefit of priority from Korean Patent Application No. 10-2023-0106325, filed on August 14, 2023, and all contents of the document in that Korean Patent Application are incorporated herein by reference.

[0003] Technology field

[0004] The embodiments disclosed in this document relate to a data analysis device and a method of operating the same.

[0005] Recently, research and development on secondary batteries has been actively conducted. Here, secondary batteries are rechargeable and include both conventional Ni / Cd and Ni / MH batteries, as well as recent lithium-ion batteries. Among secondary batteries, lithium-ion batteries have the advantage of having a much higher energy density than conventional Ni / Cd and Ni / MH batteries. Furthermore, lithium-ion batteries can be manufactured in a compact and lightweight form, making them suitable for use as power sources for mobile devices. Recently, their use has expanded to include power sources for electric vehicles, attracting attention as a next-generation energy storage medium.

[0006] Typically, batteries are monitored and their operation controlled by a battery management system (BMS). The data collected from these BMSs encompasses a wide variety of factors and is so vast that even when a battery malfunction occurs, it can be difficult to analyze the cause of the problem.

[0007] One purpose of the embodiments disclosed in this document is to provide a data analysis device and an operating method thereof that can effectively identify the cause of a battery abnormality by classifying battery data.

[0008] The technical problems of the embodiments disclosed in this document are not limited to the technical problems mentioned above, and other technical problems not mentioned will be clearly understood by those skilled in the art from the descriptions below.

[0009] According to an embodiment disclosed in the present document, a data analysis device may include an information acquisition unit that acquires a plurality of data sets, each of which includes time series data for a plurality of parameters, and a controller that extracts features of each of the plurality of data sets based on correlations between the plurality of parameters, extracts patterns of each of the plurality of data sets based on similarities between the extracted features, and classifies the plurality of data sets into a plurality of clusters based on the patterns.

[0010] According to an embodiment, the controller can extract the feature as a matrix representing a correlation between the plurality of parameters.

[0011] According to an embodiment, the correlation between the plurality of parameters can be expressed as the Pearson correlation coefficient of each pair of parameters.

[0012] According to an embodiment, the controller can calculate correlation coefficients between features of the plurality of data sets and generate a correlation coefficient matrix composed of the correlation coefficients.

[0013] According to an embodiment, the controller can extract the pattern by applying a classification algorithm to the correlation coefficient matrix.

[0014] According to an embodiment, the classification algorithm may be a K-means algorithm.

[0015] According to an embodiment, the controller can extract the pattern based on the result of applying the classification algorithm while changing the K value in the correlation coefficient matrix.

[0016] According to an embodiment, the controller can analyze the characteristics of each of the plurality of clusters.

[0017] According to an embodiment, the controller can analyze anomaly occurrence history related to a data set included in each of the plurality of clusters and estimate a parameter related to the anomaly occurrence history.

[0018] According to an embodiment, the controller may extract the features by setting some of the time series data as a window section for each of the plurality of data sets.

[0019] According to an embodiment disclosed in the present document, a method of operating a data analysis device may include the steps of obtaining a plurality of data sets, each of which includes time series data for a plurality of parameters, extracting features of each of the plurality of data sets based on correlations between the plurality of parameters, extracting patterns of each of the plurality of data sets based on similarities between the extracted features, and classifying the plurality of data sets into a plurality of clusters based on the patterns.

[0020] According to an embodiment, the step of extracting features of each of the plurality of data sets may be characterized by extracting the features as a matrix representing correlations between the plurality of parameters.

[0021] According to an embodiment, the step of extracting a pattern of each of the plurality of data sets may include the step of calculating correlation coefficients between features of the plurality of data sets to generate a correlation coefficient matrix composed of the correlation coefficients, and the step of applying a classification algorithm to the correlation coefficient matrix to extract the pattern.

[0022] According to an embodiment, the classification algorithm is a K-means algorithm, and the step of extracting the pattern by applying the classification algorithm may be characterized by extracting the pattern according to a result of applying the classification algorithm while changing the K value in the correlation coefficient matrix.

[0023] According to an embodiment, the operating method of the data analysis device may further include a step of analyzing an abnormality occurrence history related to a data set included in each of the plurality of clusters and estimating a parameter related to the abnormality occurrence history.

[0024] The data analysis device and its operation method according to the embodiments disclosed in this document can facilitate the analysis of massive and unstructured data by extracting and classifying patterns of a data set.

[0025] In addition, the data analysis device and its operation method according to the embodiments disclosed in this document can provide a starting point for identifying the cause of anomaly occurrence by classifying similar data sets.

[0026] In addition, various effects may be provided, either directly or indirectly, through this document.

[0027] FIG. 1 is a diagram showing the configuration of a data analysis system according to one embodiment disclosed in this document.

[0028] FIG. 2 is a block diagram showing the configuration of a data analysis device according to one embodiment disclosed in this document.

[0029] FIG. 3 is a block diagram showing the configuration of a controller according to one embodiment disclosed in this document.

[0030] FIG. 4 is a diagram showing an example of extracted features according to one embodiment disclosed in this document.

[0031] FIG. 5 is a diagram showing an example of a correlation coefficient matrix according to one embodiment disclosed in this document.

[0032] FIG. 6 is a drawing showing an example of an extracted pattern according to one embodiment disclosed in this document.

[0033] FIG. 7 is a flowchart for explaining an operation method of a data analysis device according to one embodiment disclosed in this document.

[0034] Figure 8 is a flowchart illustrating a process for extracting a pattern according to one embodiment disclosed in this document.

[0035] FIG. 9 is a block diagram showing the hardware configuration of a computing system for performing an operation method of a data analysis device according to one embodiment disclosed in this document.

[0036] Hereinafter, various embodiments of the present invention will be described with reference to the attached drawings. However, this is not intended to limit the present invention to specific embodiments, and it should be understood that the present invention encompasses various modifications, equivalents, and / or alternatives of the embodiments.

[0037] In this document, the singular form of a noun corresponding to an item may include one or more of said items, unless the context clearly indicates otherwise. In this document, 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 each include any one of the items listed together in that phrase, or all possible combinations thereof. Terms such as "first," "second," or "first" or "second" may be used merely to distinguish the corresponding element from other corresponding elements, and do not limit the corresponding elements in any other respect (e.g., importance or order). When a component (e.g., a first component) is referred to as being “coupled” or “connected” to another component (e.g., a second component), with or without the terms “functionally” or “communicatively,” it means that the component can be connected to the other component directly (e.g., wired), wirelessly, or through a third component.

[0038] Each component (e.g., a module or a program) described in this document may include one or more entities. According to various embodiments, one or more components or operations of the components may be omitted, or one or more other components or operations may be added. Alternatively or additionally, a plurality of components (e.g., a module or a program) may be integrated into a single component. In such a case, the integrated component may perform one or more functions of each of the plurality of components identically or similarly to those performed by the corresponding component among the plurality of components prior to integration. According to various embodiments, the operations performed by a module, program, or other component may be executed sequentially, in parallel, iteratively, or heuristically, or one or more of the operations may be executed in a different order, omitted, or one or more other operations may be added.

[0039] The term "module" or "part" used in this document may include a unit implemented in hardware, software, or firmware, and may be used interchangeably with terms such as logic, logic block, component, or circuit. A module may be an integral component, or a minimum unit or part of such a component that performs one or more functions. For example, according to one embodiment, a module may be implemented in the form of an application-specific integrated circuit (ASIC).

[0040] Various embodiments of the present document may be implemented as software (e.g., a program or an application) including one or more instructions stored in a machine-readable storage medium (e.g., memory). For example, a processor of the device may call at least one instruction among the one or more instructions stored from the storage medium and execute it. This enables the device to operate to perform at least one function according to the at least one instruction called. The one or more instructions may include code generated by a compiler or code executable by an interpreter. The machine-readable storage medium may be provided in the form of a non-transitory storage medium. Here, 'non-transitory' only means that the storage medium is a tangible device and does not contain a signal (e.g., electromagnetic waves), and this term does not distinguish between cases where data is stored semi-permanently and cases where it is stored temporarily in the storage medium.

[0041]

[0042] FIG. 1 is a diagram showing the configuration of a data analysis system according to one embodiment disclosed in this document.

[0043] Referring to FIG. 1, the data analysis system may include a battery management system (BMS, 10, 20, 30) and a data analysis device (100). Although FIG. 3 illustrates three BMSs, this is not limited thereto, and the BMSs (10, 20, 30) may be configured to include n (n is a natural number). Each BMS may be located inside the battery pack, or may be located outside the battery pack in relation to a configuration in which the battery pack is mounted or connected, such as a vehicle or a charger.

[0044] The data analysis device (100) can acquire a data set of time-series data from each battery management system (10, 20, 30). The data analysis device (100) can classify the acquired multiple data sets into multiple clusters. More specifically, the data analysis device (100) can classify the multiple clusters by extracting features of each of the multiple data sets and extracting patterns of each data set from the extracted features.

[0045] The data analysis device (100) can extract patterns and classify data sets acquired in real time from each battery management device. In an embodiment, the data analysis device (100) can learn the process of extracting features from data sets, extracting patterns from features, and classifying data sets from patterns. When a new data set is acquired, the data analysis device (100) can classify it using the learned model.

[0046] The data analysis device (100) can learn the relationship between battery data and battery abnormalities using the classification results of the data set. For example, the data analysis device (100) can classify the data set and analyze the occurrence history of abnormalities in the classified clusters to identify the cause of the abnormality.

[0047] Through this, the data analysis device (100) can facilitate the analysis of massive, unstructured data by extracting and classifying patterns from the data set. Furthermore, the data analysis device (100) can classify similar data sets to provide a starting point for identifying the cause of anomaly occurrences.

[0048] FIG. 2 is a block diagram showing the configuration of a data analysis device according to one embodiment disclosed in this document.

[0049] Referring to FIG. 2, the data analysis device (100) may include an information acquisition unit (110) and a controller (120).

[0050] The information acquisition unit (110) may include a communication circuit that can establish a wired communication channel and / or a wireless communication channel between components inside and outside the data analysis device (100), such as a server, a vehicle, a battery management system, etc., and transmit and receive various types of data through the established communication channel.

[0051] The information acquisition unit (110) can acquire multiple data sets. Each of the multiple data sets may include time-series data for multiple parameters. In an embodiment, the information acquisition unit (110) can acquire multiple data sets from multiple battery management systems. For example, the information acquisition unit (110) can acquire each data set from different battery management systems.

[0052] The plurality of parameters may include various parameters related to the state of the battery. For example, the plurality of parameters may comprehensively include parameters related to the state of the battery (e.g., current, voltage, temperature, etc.), parameters derived from the state of the battery (e.g., SOC, SOH, DOD, voltage deviation, etc.), parameters related to the vehicle in which the battery is installed when the battery is installed in a vehicle (e.g., driving time, driving distance, speed, acceleration, etc.), parameters related to external factors that may affect the state of the battery (e.g., outside temperature, etc.), etc., but are not limited to the examples described above.

[0053] The controller (120) can extract features of each of the plurality of data sets based on correlations between multiple parameters. For example, the controller (120) can extract features of each of the plurality of data sets based on various operations (e.g., covariance, joint probability, etc.) that can calculate correlations between multiple parameters from time-series data included in each data set.

[0054] According to an embodiment, the controller (120) can extract features as a matrix representing correlations between multiple parameters. For example, the features of each of the multiple data sets can be expressed as a matrix representing correlations. For example, if the data set acquired by the information acquisition unit (110) includes time series data for h (where h is a natural number) parameters, the features extracted by the controller (120) can be expressed as an h*h matrix. The correlations between the multiple parameters can be calculated based on various operations (e.g., covariance, joint probability, etc.) from data for any two parameters.

[0055] According to an embodiment, the correlation between multiple parameters can be expressed as the Pearson correlation coefficient of each parameter pair. For example, if the data set acquired by the information acquisition unit (110) includes time series data for h parameters (h is a natural number), the number of each parameter pair can be h^2. In this case, the features of each data set can be expressed as an h*h matrix, and the value of each component can be the Pearson correlation coefficient of the parameter pair.

[0056] In an embodiment, the Pearson correlation coefficient can be defined as in [Mathematical Formula 1] below.

[0057]

[0058] Here, r xy is the Pearson correlation coefficient for the x parameter and the y parameter, n is the number of data included in the time series data of each parameter, and x i is the ith data in the time series data of the x parameter, y i is the ith data in the time series data of the y parameter, is the mean of the x parameter data, means the average of the y parameter data.

[0059] That is, the Pearson correlation coefficient can be calculated from the covariance and standard deviation of two parameters. The Pearson correlation coefficient has a normalized value and can have a value between -1 and 1, for example. For example, the value of the Pearson correlation coefficient between identical parameters is 1, and accordingly, the values ​​of the elements located on the diagonal in the h*h matrix can be 1.

[0060] According to an embodiment, the controller (120) may extract patterns from each of the plurality of data sets based on the similarity between the extracted features. For example, the controller (120) may cluster data sets having similar features. The controller (120) may extract patterns from each of the data sets based on the clustering results.

[0061] In another example, the controller (120) can analyze the characteristics of each data set and convert the characteristics of each data set into a pattern. For example, the controller (120) can define multiple patterns and determine the pattern of each data set based on the characteristics of each data set. In this case, the controller (120) can learn the process of determining the pattern from the characteristics of the data set.

[0062] According to an embodiment, the controller (120) can calculate a correlation coefficient between features of multiple data sets. The controller (120) can calculate the correlation coefficient based on various operations (e.g., covariance, joint probability, etc.). For example, the controller (120) can calculate the correlation coefficient between features based on [Mathematical Equation 2] below.

[0063]

[0064] Here, A and B each represent feature matrices, r is the correlation coefficient between features of A and B, and A nm is the matrix value of (n,m) elements in feature matrix A, B nmis the matrix value of (n,m) elements in the B feature matrix, is the mean of the feature matrix A, means the average of the B feature matrix.

[0065] According to an embodiment, the controller (120) may generate a correlation coefficient matrix composed of correlation coefficients. For example, if the data set is i (i is a natural number), the correlation coefficient matrix may be an i*i matrix.

[0066] According to an embodiment, the controller (120) can extract patterns of each of the plurality of data sets by applying a classification algorithm to the correlation coefficient matrix. The classification algorithm can include various algorithms, such as a decision tree-based algorithm such as the XGBoost algorithm, a K-Nearest Neighbor classification (K-NN) algorithm such as the K-means algorithm, and a Support Vector Machine (SVM) algorithm.

[0067] In one embodiment, the classification algorithm may be the K-means algorithm. The K-means algorithm may be an algorithm that classifies multiple data sets into K clusters. For example, the K-means algorithm may be an algorithm that classifies data sets based on the average distance between the center of each cluster and the data set.

[0068] According to an embodiment, the controller (120) can change the K value in the correlation coefficient matrix and extract a pattern based on the result of applying the classification algorithm. In the K-means algorithm, the K value can mean the number of classification clusters. The controller (120) can express the result classified according to the K value as a number. For example, when K = 2, each data set can belong to either the first cluster or the second cluster. In this case, the controller (120) can process the classification result values ​​of the data sets belonging to the first cluster as 0 and the classification result values ​​of the data sets belonging to the second cluster as 1.

[0069] In this way, the controller (120) can accumulate classification results by changing the K value. For example, depending on the K value, the controller (120) can determine the classification result value of each data set as any one value from 0 to K-1. The controller (120) can accumulate the classification result values ​​and express the pattern of each data set in the form of a list of classification result values. For example, the controller (120) can express the pattern by listing the classification result values ​​in ascending order according to the K value.

[0070] In an embodiment, the controller (120) can classify multiple data sets into multiple clusters based on patterns. For example, the controller (120) can classify multiple data sets into multiple clusters based on pattern similarity. For example, the controller (120) can classify data sets with identical patterns into the same cluster.

[0071] In an embodiment, the controller (120) can analyze the characteristics of each of the plurality of clusters. For example, the controller (120) can analyze the common characteristics of the data sets belonging to each of the plurality of clusters.

[0072] According to an embodiment, the controller (120) can analyze abnormality occurrence histories related to data sets included in each of a plurality of clusters. The abnormality occurrence histories may include various abnormality histories related to batteries, such as fire occurrence histories, voltage abnormalities, and lithium deposition histories.

[0073] The controller (120) can classify data sets into multiple clusters and analyze abnormality occurrence histories associated with the data sets included in each cluster to estimate the cause of the abnormality. For example, if identical abnormality occurrence histories exist within the same cluster, the controller (120) can estimate that the cause of the corresponding abnormality occurrence histories is similar or identical. The abnormality occurrence histories associated with the data sets may include, for example, abnormality histories that occurred in batteries managed by the battery management device from which the corresponding data set was acquired.

[0074] According to an embodiment, the controller (120) can estimate parameters related to the abnormal occurrence history. The controller (120) can estimate parameters among a plurality of parameters that are determined to be related to the abnormal occurrence history. The controller (120) can estimate the parameters by learning the relationship between patterns in the data set and the abnormal occurrence history.

[0075] According to an embodiment, the controller (120) may extract features by setting a window section for each of a plurality of data sets. Here, each of the plurality of data sets may include time-series data, and the time-series data may include a plurality of time sections. The controller (120) may target data corresponding to a certain time period among the time-series data as a target for feature extraction. For example, the controller (120) may set a section of the time-series data in each data set with high data reliability as a window section, excluding a section with inaccurate data (e.g., an initial section of driving, etc.).

[0076] FIG. 3 is a block diagram showing the configuration of a controller according to one embodiment disclosed in this document.

[0077] Referring to FIG. 3, the controller (120) may include a feature extraction module (121) and a classification module (123). The controller (120) may classify each data set into multiple clusters using the feature extraction module (121) and the classification module (123). The controller (120) may further include a cause identification module (125).

[0078] The feature extraction module (121) can extract features from each of a plurality of data sets. In addition, the feature extraction module (121) can generate a correlation coefficient matrix from the extracted features.

[0079] The classification module (123) can extract patterns for each of the multiple data sets from the extracted features. The classification module (123) can extract patterns by applying a classification algorithm to the features of each data set. For example, the classification module (123) can include a program that performs a K-means algorithm, and can extract patterns for each of the multiple data sets by changing the K value and performing clustering. The classification module (123) can classify the multiple data sets based on the patterns.

[0080] The cause identification module (125) can analyze the classification results of the data set and the abnormality occurrence history in each cluster to determine the cause of the abnormality based on the abnormality occurrence history. For example, the cause identification module (125) can estimate a parameter among multiple parameters that is determined to be related to the abnormality occurrence history.

[0081] According to an embodiment, each of the feature extraction module (121), the classification module (123), and the cause identification module (125) may include a neural network model. The feature extraction module (121) may learn a process of extracting features of each data set. The classification module (123) may learn a process of extracting patterns from features and a process of classifying data sets according to the patterns. The cause identification module (125) may learn a process of estimating the cause of anomaly occurrence based on the anomaly occurrence history related to the classification results of the data set and the anomaly occurrence history related to the data sets belonging to each cluster.

[0082] The controller (120) can classify a new data set even when a new data set is acquired using the learned feature extraction module (121), classification module (123), and cause identification module (125), and can estimate the cause of the abnormality when an abnormality occurs in a battery related to the data set.

[0083] FIG. 4 is a diagram showing an example of extracted features according to one embodiment disclosed in this document.

[0084] Referring to FIG. 4, the controller (120) can extract features from each of multiple data sets. Each feature can be represented as a matrix representing correlations between multiple parameters. For example, the value of each element of the feature matrix can be the Pearson correlation coefficient for each pair of parameters.

[0085] For example, in FIG. 4, each row and each column represent parameters (f1 to fn), and each element of the matrix can be expressed as a Pearson correlation coefficient between the parameters corresponding to the row and column in which the element is located. In an embodiment, the elements in the feature matrix can be expressed as numbers, or can be distinguished by color as in FIG. 4.

[0086] FIG. 5 is a diagram showing an example of a correlation coefficient matrix according to one embodiment disclosed in this document.

[0087] Referring to FIG. 5, the controller (120) can generate a correlation coefficient matrix from the features of each data set. In FIG. 5, M1 to M4 represent the feature matrices of each data set, and each element represents a correlation coefficient between the features. The numbers and values ​​shown in FIG. 5 are merely examples and are not limited thereto. As shown in FIG. 5, the correlation coefficient of each data set can be a normalized value between -1 and 1.

[0088] FIG. 6 is a drawing showing an example of an extracted pattern according to one embodiment disclosed in this document.

[0089] Referring to FIG. 6, the controller (120) can extract a pattern of each data set. A to N represent data sets, and the pattern of each data set is indicated in the 6th column of the matrix illustrated in FIG. 6.

[0090] In an embodiment, the controller (120) can classify data sets by changing the K value and accumulate the classification results. For example, the controller (120) can determine the classification result as any one of 0 to K-1 depending on the K value. In FIG. 6, the first to fifth columns represent classification results obtained by setting the K value to 2 to 6, respectively. The controller (120) can generate a pattern by listing the classification result values ​​for each data set. For example, the pattern of the A data set can be defined as 00044, and the pattern of the H data set can be defined as 10112.

[0091] The controller (120) can classify multiple data sets based on patterns. For example, the controller (120) can classify data sets F through J as belonging to the same cluster because they have the same pattern of 10112.

[0092] The controller (120) can estimate parameters related to the abnormality occurrence history by analyzing the abnormality occurrence history related to the data sets included in each of the plurality of clusters. The 8th column in Fig. 6 indicates whether an abnormality has occurred in each data set. The controller (120) can estimate parameters related to the abnormality occurrence history by comprehensively analyzing the abnormality occurrence history in the data sets and the classifications to which the data sets belong. For example, the controller (120) can estimate that the abnormality occurrence in the F, H, and J data sets in which the same abnormality has occurred among the F to J data sets having the same pattern as 10112 has the same or similar cause.

[0093] In Fig. 6, the abnormality occurrence history is simply expressed as whether or not an abnormality occurred, but the data management device (100) may include various information related to the abnormality occurrence, such as the number of times an abnormality occurred, the type of abnormality occurrence history, and the time of the abnormality occurrence, in the abnormality occurrence history.

[0094] Additionally, the controller (120) can manage and store the location of the BMS or the location where each data set was acquired, as shown in the 7th column of FIG. 6.

[0095] FIG. 7 is a flowchart illustrating the operation method of a data analysis device according to one embodiment disclosed in this document. The embodiment illustrated in FIG. 7 is merely one embodiment, and the order of steps according to various embodiments of the present invention may differ from that illustrated in FIG. 7, and some of the steps illustrated in FIG. 7 may be omitted, the order of steps may be changed, or steps may be merged.

[0096] Referring to FIG. 7, the operating method of the data analysis device may include a step of obtaining a plurality of data sets, each of which includes time series data for a plurality of parameters (S100), a step of extracting features of each of the plurality of data sets based on correlations between the plurality of parameters (S200), a step of extracting patterns of each of the plurality of data sets based on similarities between the extracted features (S300), a step of classifying the plurality of data sets into a plurality of clusters based on the patterns (S400), and a step of analyzing abnormality occurrence history related to data sets included in each of the plurality of clusters and estimating parameters related to the abnormality occurrence history (S500).

[0097] In step S100, the information acquisition unit (110) can acquire multiple data sets. The information acquisition unit (110) can acquire time series data acquired from each of the multiple battery management devices as a data set.

[0098] At step S200, the controller (120) can extract features of each of the plurality of data sets based on correlations between the plurality of parameters. For example, the controller (120) can express the features as a matrix representing Pearson correlation coefficients for the plurality of parameters.

[0099] At step S300, the controller (120) can extract patterns from each of the plurality of data sets based on the similarity between the extracted features. For example, the controller (120) can generate a correlation coefficient matrix from each feature and apply a classification algorithm to the correlation coefficient matrix to extract patterns from each of the data sets.

[0100] At step S400, the controller (120) can classify multiple data sets based on patterns. For example, the controller (120) can classify data sets having the same pattern into the same cluster.

[0101] At step S500, the controller (120) can analyze the abnormal occurrence history related to the data set included in each of the plurality of clusters and estimate the parameters related to the abnormal occurrence history.

[0102] Figure 8 is a flowchart illustrating a process for extracting a pattern according to one embodiment disclosed in this document.

[0103] Referring to FIG. 8, the controller (120) can produce a correlation coefficient matrix from multiple data sets and extract a pattern from the correlation coefficient matrix.

[0104] In step S310, the controller (120) can generate a correlation coefficient matrix composed of correlation coefficients by calculating correlation coefficients between features of multiple data sets.

[0105] At step S320, the controller (120) can extract patterns by applying a classification algorithm to the correlation coefficient matrix. The classification algorithm may include, for example, the K-means algorithm. The controller (120) can apply the K-means algorithm to the correlation coefficient matrices by changing the K value, and accumulate the classification results. The controller (120) can extract patterns for each data set from the accumulated classification results.

[0106] FIG. 9 is a block diagram showing the hardware configuration of a computing system for performing an operation method of a data analysis device according to one embodiment disclosed in this document.

[0107] Referring to FIG. 9, a computing system (1000) according to one embodiment disclosed in the present document may include an MCU (1010), a memory (1020), an input / output I / F (1030), and a communication I / F (1040).

[0108] The MCU (1010) may be a processor that executes various programs (e.g., a battery data acquisition program, a feature extraction program, a correlation coefficient matrix generation program, a pattern extraction program, a classification program, etc.) stored in the memory (1020), processes various information including time series data of a data set through these programs, and performs the functions of the controller included in the data analysis device shown in FIG. 2 described above.

[0109] The memory (1020) can store various programs, such as a battery data acquisition program, a feature extraction program, a correlation coefficient matrix generation program, a pattern extraction program, and a classification program. In addition, the memory (1020) can store various information, including classification results of a data set and abnormality occurrence history.

[0110] Such memories (1020) may be provided in multiple numbers as needed. The memories (1020) may be volatile memories or non-volatile memories. As volatile memories (1020), RAM, DRAM, SRAM, etc. may be used. As non-volatile memories (1020), ROM, PROM, EAROM, EPROM, EEPROM, flash memories, etc. may be used. The examples of the memories (1020) listed above are merely examples and are not limited to these examples.

[0111] The input / output I / F (1030) can provide an interface that enables data transmission and reception between an input device (not shown) such as a keyboard, mouse, or touch panel, and an output device (not shown) such as a display and the MCU (1010).

[0112] The communication I / F (1040) is a component capable of transmitting and receiving various data with the server, and may be any device capable of supporting wired or wireless communication. For example, a data analysis device can transmit and receive various information, including abnormality occurrence details in a data set, from a separately provided external server via the communication I / F (1040).

[0113] In this way, a computer program according to one embodiment disclosed in this document may be implemented as a module that is recorded in a memory (1020) and processed by an MCU (1010) to perform each function illustrated in FIG. 2, for example.

[0114]

[0115] Although all components constituting the embodiments disclosed in this document have been described as being combined or operating in combination as one, the embodiments disclosed in this document are not necessarily limited to such embodiments. That is, within the scope of the purpose of the embodiments disclosed in this document, all of the components may be selectively combined and operated one or more times.

[0116] In addition, terms such as "include," "comprise," or "have" described above, unless specifically stated otherwise, mean that the corresponding component can be included, and therefore should be interpreted to include other components rather than excluding other components. All terms, including technical or scientific terms, have the same meaning as commonly understood by a person of ordinary skill in the art to which the embodiments disclosed in this document belong, unless otherwise defined. Commonly used terms, such as terms defined in a dictionary, should be interpreted to be consistent with the contextual meaning of the relevant technology, and shall not be interpreted in an idealized or overly formal sense, unless explicitly defined in this document.

[0117] The above description is merely an example of the technical idea disclosed in this document, and those skilled in the art to which the embodiments disclosed in this document pertain may make various modifications and variations without departing from the essential characteristics of the embodiments disclosed in this document. Therefore, the embodiments disclosed in this document are not intended to limit the technical idea of ​​the embodiments disclosed in this document, but to explain it, and the scope of the technical idea disclosed in this document is not limited by these embodiments. The scope of protection of the technical idea disclosed in this document should be interpreted by the claims below, and all technical ideas within a scope equivalent thereto should be interpreted as being included in the scope of rights of this document.

[0118] [Explanation of symbols]

[0119] 10, 20, 30: Battery management device

[0120] 100: Data Analysis Device

[0121] 110: Information Acquisition Department

[0122] 120: Controller

[0123] 121: Feature Extraction Module

[0124] 123: Classification module

[0125] 125: Cause Identification Module

[0126] 1000: Computing Systems

[0127] 1010: MCU

[0128] 1020: Memory

[0129] 1030: Input / Output I / F

[0130] 1040: Communication I / F

Claims

1. An information acquisition unit that acquires multiple data sets, each of which includes time series data for multiple parameters; and Extracting features of each of the plurality of data sets based on the correlation between the plurality of parameters, Extracting a pattern of each of the plurality of data sets based on the similarity between the extracted features, A data analysis device comprising a controller that classifies the plurality of data sets into a plurality of clusters based on the above pattern.

2. In paragraph 1, The above controller, A data analysis device that extracts the features as a matrix representing the correlation between the plurality of parameters.

3. In paragraph 2, A data analysis device in which the correlation between the above plurality of parameters is expressed as the Pearson correlation coefficient of each pair of parameters.

4. In paragraph 1, The above controller, A data analysis device that calculates correlation coefficients between features of the plurality of data sets and generates a correlation coefficient matrix composed of the correlation coefficients.

5. In paragraph 4, The above controller, A data analysis device that extracts the pattern by applying a classification algorithm to the correlation coefficient matrix.

6. In paragraph 5, The above classification algorithm is a data analysis device that is a K-means algorithm.

7. In paragraph 6, The above controller, A data analysis device that extracts the pattern based on the result of applying the classification algorithm while changing the K value in the correlation coefficient matrix.

8. In paragraph 1, The above controller, A data analysis device that analyzes the characteristics of each of the above multiple clusters.

9. In paragraph 8, The above controller, A data analysis device that analyzes anomaly occurrence history related to a data set included in each of the above plurality of clusters and estimates a parameter related to the anomaly occurrence history.

10. In paragraph 1, The above controller, A data analysis device that extracts features by setting a window section for each of the above multiple data sets.

11. A step of obtaining a plurality of data sets, each of which includes time series data for a plurality of parameters; A step of extracting features of each of the plurality of data sets based on the correlation between the plurality of parameters; A step of extracting a pattern of each of the plurality of data sets based on the similarity between the extracted features; and A method of operating a data analysis device, comprising a step of classifying the plurality of data sets into a plurality of clusters based on the above pattern.

12. In paragraph 11, The step of extracting features of each of the above multiple data sets is: A method of operating a data analysis device, characterized in that the feature is extracted as a matrix representing the correlation between the plurality of parameters.

13. In paragraph 11, The step of extracting the pattern of each of the above multiple data sets is: A step of generating a correlation coefficient matrix composed of correlation coefficients by calculating correlation coefficients between features of the plurality of data sets; and A method of operating a data analysis device, comprising a step of extracting the pattern by applying a classification algorithm to the correlation coefficient matrix.

14. In paragraph 13, The above classification algorithm is the K-means algorithm, The step of extracting the pattern by applying the above classification algorithm is: A method of operating a data analysis device, characterized in that the pattern is extracted based on the result of applying the classification algorithm while changing the K value in the correlation coefficient matrix.

15. In paragraph 11, A method of operating a data analysis device, further comprising a step of analyzing anomaly occurrence history related to a data set included in each of the plurality of clusters and estimating a parameter related to the anomaly occurrence history.

Images