Cloud-based battery management system and method therefor

Abstract

The present invention relates to a cloud-based battery management system and a method therefor, and relates to a method for efficiently managing each battery mounted on an electric vehicle or an energy storage system by constructing a cloud BMS using measurement data for each battery received from an OEM server and charging data for each battery received from a charging device.

Description

Cloud-based battery management system and method thereof

[0001] The present invention relates to a cloud-based battery management system and method thereof, and more particularly, to a method for efficiently managing the batteries of each product by constructing a cloud BMS using measurement data for each battery received from an OEM server and charging data for each battery received from a charging device for each battery mounted on an electric vehicle or an energy storage system.

[0002] The importance of batteries is increasing day by day as various products that run on batteries or store energy in batteries, including portable electronic devices, electric vehicles, robots, and energy storage devices, are developed and advanced.

[0003] These batteries are secondary batteries that can be recharged and used repeatedly after being discharged, and have the advantage of allowing the product to be operated for a long time without a separate commercial power source.

[0004] Most batteries are manufactured using the so-called OEM (Original Equipment Manufacturing) method and are used in various products, including electric vehicles.

[0005] Batteries manufactured using this OEM method can be managed by the manufacturer through a separate OEM server. Since there are numerous manufacturers of these batteries, the OEM servers operated by various manufacturers can be integrated and managed through a third-party server. This third-party server can provide an open API for the BMS.

[0006] Therefore, third-party servers can collect and manage battery status information, including measurement data and SOC, of ​​the batteries installed in each product through the OEM server via open API.

[0007] Typically, BMS is configured to estimate battery status information using OCV (open circuit voltage) or current accumulation method.

[0008] Estimating battery status information using the above OCV is impossible in real time because it is a method of measuring the open circuit voltage (OCV) of the battery and estimating it based on a battery status information table (e.g., SOC table).

[0009] Estimating the battery's status information using the above current accumulation method has the problem that the error increases as the current accumulation accumulates because it estimates the SOC (state of charge) using the amount of charge and current.

[0010] These BMSs only estimate the SOC among the battery status information, and have limitations in predicting SOH (state of health), SOE (state of energy), and RUL (remaining useful lifetime) to manage the actual battery.

[0011] To solve these problems, various methods for estimating battery status information using artificial intelligence technology have been recently proposed. However, BMS is an embedded controller that has limitations in resources (e.g., CPU, memory) for running learning or learning models for artificial intelligence.

[0012] Meanwhile, in the case of a charging device that charges a battery, there is a problem that the occurrence rate of fire may increase during battery charging because it is configured to operate passively without estimating the status information of the battery or detecting abnormal conditions during charging.

[0013] Accordingly, the present invention proposes a method for safely charging batteries by configuring a cloud-based battery management system (i.e., cloud BMS) to manage batteries installed in each product in an integrated manner, detect abnormal conditions, and provide the corresponding status information or detection information to the user of the charging device or battery.

[0014] Next, we will briefly explain the prior art existing in the technical field of the present invention, and then describe the technical details that the present invention seeks to achieve differently from the prior art.

[0015] First, Korean Patent Publication No. 2024-0050757 (April 19, 2024) relates to a battery management device, a BMS data storage system, and a BMS data storage method, wherein when an event requiring emergency storage of BMS data for a battery is detected, the BMS data is transmitted to a first controller within the device and stored.

[0016] In other words, Korean Patent Publication No. 2024-0050757 only enables data from a BMS that manages a battery installed in a product (device) to be transmitted to a controller installed in the product and stored in memory when there is an urgent need to store the data.

[0017] On the other hand, the present invention configures a battery management system for each battery on the cloud and utilizes the hardware resources of the cloud server to estimate the status information of the battery or detect an abnormal condition, thereby efficiently managing the battery and stably charging the battery. Korean Patent Publication No. 2024-0050757 does not describe, suggest, or imply any technical features of the present invention.

[0018] In addition, Korean Patent Publication No. 2023.0040808 (March 23, 2023) relates to a battery management device, a battery system, and an operating method thereof, and relates to a battery management device, a battery system, and an operating method thereof that provides battery status information including at least one of voltage, current, and temperature of a battery pack and diagnosis result information related to abnormality diagnosis of a battery pack based on the battery status information to a higher controller or another battery management device.

[0019] The above Korean Patent Publication No. 2023.0040808 simply provides a diagnosis result based on one of the current, voltage, and temperature of the battery, and does not configure a battery management system on the cloud for battery management by OEM or manufacturer as proposed in the present invention. It also does not describe a method for estimating battery status information or detecting abnormal conditions by simultaneously receiving battery measurement data and battery charging data. Therefore, the two inventions have significant differences in their technical structure, purpose, and effect.

[0020] The present invention was created to solve the above problems, and its purpose is to provide a cloud-based battery management system and method that implements a battery management system for managing batteries on the cloud to enable integrated management of multiple batteries.

[0021] In addition, the present invention aims to provide a cloud-based battery management system and method that can precisely estimate battery status information by receiving measurement data measured from a battery and charging data measured from a charging device when charging a battery from the cloud.

[0022] In addition, the present invention provides a battery management system and method that can stably charge a battery by providing estimated battery status information, received battery measurement data, or a combination thereof to a charging device, thereby using the battery status information, battery measurement data, charging data, or a combination thereof.

[0023] In addition, the present invention aims to provide a cloud-based battery management system and method that detects an abnormal condition during charging using battery measurement data and charging data, and, when an abnormal condition is detected, provides information to a charging device, a user terminal, or a combination thereof to immediately stop charging.

[0024] In addition, the present invention provides a battery management system and method that have the advantage of enabling decomposition of ASIL (Automotive Safely Integrity Level) by enabling hardware, including BMS and charging devices mounted on a product, to be separated and configured by implementing a battery management system on the cloud, thereby enabling configuration with a lower ASIL.

[0025] A cloud-based battery management method according to one embodiment of the present invention includes a battery measurement data reception step of receiving measurement data of a battery from an OEM (Original Equipment Manufacturing) server of a battery mounted on a product, a charging data reception step of receiving charging data of the battery from a charging device, and a cloud BMS configuration step of configuring a BMS on the cloud by storing and managing the received measurement data and charging data in a database for each battery, and is characterized in that a plurality of batteries mounted on a plurality of products or charged in a charging device are integrated and managed through the configured cloud BMS.

[0026] In addition, the measurement data includes voltage, current, temperature or a combination thereof of the battery, and the charging data includes charging voltage, charging current, charging temperature or a combination thereof.

[0027] In addition, the battery management method further includes a battery state information estimation step for estimating state information of the battery, including SOC (state of charge), SOH (state of health), RUL (remaining useful lifetime) or a combination thereof, using the received measurement data and charging data, and is characterized in that the estimated battery state information is stored and managed in the database for each battery through the cloud BMS configuration step.

[0028] In addition, the battery status information estimation step further includes a first input data configuring step of supplementing the received battery measurement data using the received charging data to fit the input of the product encoder to configure first input data, and a status information encoding step of estimating the battery status information by inputting the first input data to the product encoder and encoding the supplemented measurement data into the battery status information, and the product encoder is characterized in that it is configured by learning first learning data that labels the battery status information on the voltage, current, temperature, or a combination thereof of the battery collected from the OEM server of the battery mounted on the product.

[0029] In addition, the battery management method further includes an abnormality detection step for detecting an abnormality using the received measurement data and charging data, and is characterized in that the results of detecting the abnormality are stored and managed in the database for each battery through the cloud BMS configuration step.

[0030] In addition, the above-described abnormal state detection step comprises a second input data configuration step of configuring second input data suitable for an abnormal state detection model including a first abnormal state detection model including a product encoder and a charging device decoder, a second abnormal state detection model including a charging device encoder and a product decoder, a third abnormal state detection model including a product encoder and a product decoder, a fourth abnormal state detection model including a charging device encoder and a charging device decoder, or a combination thereof, and an error calculation step of inputting each of the second input data into a corresponding abnormal state detection model and calculating an error between each of the second input data and output data of each abnormal state detection model, and detecting that an abnormal state has occurred when an average result of each of the calculated errors exceeds a predetermined threshold error or when any one of the calculated errors exceeds a predetermined threshold error.

[0031] In addition, the above-described abnormal condition detection step is characterized in that the product encoder is configured by learning first learning data in which battery status information is labeled with the voltage, current, temperature, or a combination thereof of the battery collected from the OEM server of the battery mounted on the product, the product decoder is configured by learning second learning data in which the collected battery status information is labeled with the voltage, current, temperature, or a combination thereof, the charging device encoder is configured by learning third learning data in which the battery status information is labeled with the charging voltage, charging current, charging temperature, or a combination thereof collected from the charging device, and the charging device decoder is configured by learning fourth learning data in which the collected charging voltage, charging current, charging temperature, or a combination thereof is labeled with the battery status information.

[0032] In addition, the above-described abnormal state detection step is characterized by further including a third input data configuration step of configuring third input data suitable for the input of the fifth abnormal state detection model and the sixth abnormal state detection model using the received measurement data and charging data, and an error calculation step of inputting each of the configured third input data into each of the corresponding abnormal state detection models and calculating an error between each of the third input data and the output data of each of the abnormal state detection models.

[0033] In addition, the fifth abnormal state detection model is configured by learning fifth learning data that labels the voltage, current, temperature, or a combination thereof of the battery collected from the OEM server of the battery mounted on the product, and the charging voltage, charging current, charging temperature, or a combination thereof collected from the charging device, and the sixth abnormal state detection model is configured by learning sixth learning data that labels the voltage, current, temperature, or a combination thereof of the battery collected from the OEM server of the battery mounted on the product, and the charging voltage, current, temperature, or a combination thereof collected from the charging device.

[0034] In addition, a cloud-based battery management system according to one embodiment of the present invention is characterized by including a memory that stores a program code implementing the battery management method and a processor configured to load and execute the program code stored in the memory.

[0035] As described above, the present invention implements a battery management system on the cloud, receives measurement data of a battery and charging data of a charging device, performs the role of a BMS in the cloud, and provides the data to each product or charging device, thereby lowering the ASIL of the BMS or charging device of the battery installed in each product and enabling integrated management of individual batteries.

[0036] In addition, the present invention has the effect of enabling stable use or charging of a battery by accurately estimating battery status information using battery measurement data and charging data in a cloud BMS and then providing the result to each battery user and charging device.

[0037] In addition, the present invention has the effect of preventing dangers caused by abnormal conditions such as fire in a charging device by detecting the occurrence of an abnormal condition during charging in the charging device and immediately stopping charging.

[0038] FIG. 1 is a diagram illustrating a cloud-based battery management system and method according to one embodiment of the present invention.

[0039] FIG. 2 is a drawing illustrating in detail a cloud-based battery management system and method according to one embodiment of the present invention.

[0040] FIG. 3 is a diagram illustrating the operation of a cloud-based battery management system according to one embodiment of the present invention.

[0041] FIG. 4 is a diagram illustrating an encoder according to one embodiment of the present invention.

[0042] FIG. 5 is a diagram illustrating a decoder according to one embodiment of the present invention.

[0043] FIG. 6 is a diagram illustrating an abnormal state detection model according to one embodiment of the present invention.

[0044] FIG. 7 is a diagram illustrating an abnormal condition detection model according to another embodiment of the present invention.

[0045] FIG. 8 is a block diagram showing the configuration of a cloud-based battery management system according to one embodiment of the present invention.

[0046] FIG. 9 is a flowchart illustrating a procedure for managing a battery based on a cloud according to one embodiment of the present invention.

[0047] [Description of symbols] 100: Battery management system; 110: Battery measurement data receiving unit; 120: Charging data receiving unit; 130: Abnormal condition detection unit; 140: Battery status information estimation unit; 150: Notification providing unit; 160: Battery status information transmitting unit; 170: User information management unit; 200: Product; 300: OEM server; 400: Charging device; 500: User terminal; 600: Database; 700: Third-party server.

[0048] Hereinafter, preferred embodiments of a cloud-based battery management system and method of the present invention will be described in detail with reference to the attached drawings. The same reference numerals in each drawing represent the same components. In addition, specific structural and functional descriptions of embodiments of the present invention are merely illustrative for the purpose of explaining embodiments according to the present invention, and unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as generally understood by those of ordinary skill in the art to which the present invention pertains. Terms defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning they have in the context of the relevant technology, and are preferably not interpreted in an ideal or excessively formal sense unless explicitly defined herein.

[0049] FIG. 1 is a diagram illustrating a cloud-based battery management system and method according to one embodiment of the present invention.

[0050] As illustrated in FIG. 1, a cloud-based battery management system (100) (hereinafter referred to as a cloud BMS (Battery Management System)) according to one embodiment of the present invention is configured on the cloud and receives battery measurement data from an OEM server (300) of a battery mounted on a product (200), and receives charging data when charging the battery from a charging device (400), thereby providing a function for efficiently managing the battery, such as estimating battery status information and detecting abnormal conditions.

[0051] Since the above cloud BMS (100) manages each battery individually, it is possible to manage multiple batteries in an integrated manner.

[0052] In Fig. 1, it is preferable that the cloud BMS (100) is configured on a third-party server (700), but it may also be configured on an actual OEM server (300).

[0053] Here, the OEM server (300) may be provided by a manufacturer that manufactured the battery in an OEM (Original Equipment Manufacturing) manner, and may be configured to receive battery measurement data from a measurement data provision device mounted on a product (200) and transmit it to a cloud BMS (100) through an open API.

[0054] Additionally, the third-party server (700) refers to an integrated server on the cloud that collects and comprehensively manages information about batteries managed by each OEM server (300).

[0055] Here, the product (300) refers to a finished product including a battery provided by an OEM, including an electric vehicle (EV), an energy storage system (ESS), etc. In this case, if the product (300) is an electric vehicle, the measurement data providing device may be an on-board diagnostics (OBD).

[0056] The above charging device (400) is for charging a battery mounted on a product (200) and can be configured to transmit charging data when charging the battery to a cloud BMS (100).

[0057] The user terminal (500) above refers to a wired or wireless communication terminal including a smartphone, etc., provided by a user using the product (200). However, the present invention is not limited thereto, and the user terminal (500) may be mounted on the product (200), such as a navigation device including a display.

[0058] The above database (600) is configured to store various data or information for integrated management of the battery, including battery measurement data, charging data, estimated battery status information, abnormal condition detection results, etc.

[0059] The measurement data, charging data, estimated battery status information, and abnormal condition detection results of the above batteries are stored in a database (600) by categorizing each battery through each cloud BMS (100).

[0060] FIG. 2 is a drawing illustrating in detail a cloud-based battery management system and method according to one embodiment of the present invention.

[0061] As illustrated in FIG. 2, a cloud BMS (100) according to one embodiment of the present invention receives battery measurement data from an OEM server (300) and receives charging data when charging a battery from a charging device (300).

[0062] The measurement data of the above battery is configured to include voltage, current, temperature, or a combination thereof inside the battery measured by the BMS of the battery mounted on the product (200).

[0063] The above product (200) is configured to transmit battery measurement data to the OEM server (300) through a measurement data providing device when discharging the battery (i.e., when operating the product) and also when charging the battery.

[0064] The above product (200) can also transmit battery measurement data to a user terminal (500) to enable the status of the battery to be checked.

[0065] At this time, the measurement data of the battery may further include user information including product (200) identification information (e.g., car number, battery number, etc.), user identification information (e.g., user ID), or a combination thereof.

[0066] The OEM server (300) transmits specific data of the battery to the cloud BMS (100), and the charging device (400) transmits charging data to the cloud BMS (100) when charging the battery mounted on the product (200).

[0067] At this time, the charging device (400) transmits charging data to the user terminal (500) so that the charging status can be checked.

[0068] The above charging data comprises the measured charging voltage, charging current, charging temperature, or a combination thereof when charging the battery. The charging voltage refers to the voltage applied to the battery during charging, the charging current refers to the current flowing into the battery during charging, and the charging temperature refers to the temperature during charging.

[0069] The above cloud BMS (100) detects an abnormal condition during battery charging by using the battery measurement data and charging data received from the OEM server (300).

[0070] The above abnormal condition may occur in the charging device or the battery, and when the above cloud BMS (300) detects the abnormal condition, it transmits a notification to the charging device (400), the user terminal (500), or a combination thereof, so that battery charging can be stopped immediately.

[0071] Detecting the above abnormal condition is performed by simply comparing the battery measurement data and charging data or by using a learning model constructed (created) through machine learning, and will be described in detail with reference to Fig. 3.

[0072] If an abnormal condition is not detected, the cloud BMS (100) estimates the battery status information using the battery measurement data and charging data.

[0073] The status information of the above battery is configured to include SOC (state of charging), SOH (state of health), SOE (state of energy), RUL (remaining useful life), SOP (state of power), or a combination thereof.

[0074] The above SOC refers to the state of charge of the battery, SOE refers to the amount of usable energy, SOP, SOH, and RUL refer to the lifespan of the battery. SOP indicates how powerful the battery is, SOH indicates the performance of the battery, and RUL indicates the remaining cycles of the battery.

[0075] Estimating the above battery status is performed using a learning model created through machine learning, and will be described in detail with reference to Fig. 3.

[0076] Additionally, the cloud BMS (100) transmits the estimated battery status information to the user terminal (500) so that the battery status information can be recognized.

[0077] Additionally, the cloud BMS (100) transmits the estimated battery status information and the battery measurement data received from the OEM server (300) to the charging device (400).

[0078] At this time, the charging device (400) can directly receive the measurement data of the battery from the measurement data providing device of the product (200).

[0079] The above charging device (400) is configured to charge the battery stably and efficiently by using battery status information, battery measurement data, charging data, or a combination thereof.

[0080] For example, a low SOH means a low battery life, so the charging device (400) charges the battery by lowering the charging current or c-rate in proportion to the lower SOH, thereby extending the battery life and allowing stable charging.

[0081] As another example, the charging device (400) may be configured to charge the battery by varying the charging current based on the battery temperature measured from the battery's measurement data. For example, since optimal charging of the battery is possible between 15 and 40 degrees Celsius, the battery can be charged using a preset optimal charging current.

[0082] As another example, the charging device (400) can limit the maximum chargeable SOC and charge voltage according to the user's charging pattern. For example, if the user frequently charges below a predetermined SOC (e.g., almost discharged) or frequently drives long distances, the charging can be performed by increasing the maximum charge SOC and charge voltage. On the other hand, if the user frequently charges above another predetermined SOC (e.g., 30% or more) or frequently drives short distances, the charging can be performed by lowering the maximum charge SOC and charge voltage. In other words, by limiting the maximum chargeable SOC rather than setting it to 100%, the life of the battery can be extended.

[0083] FIG. 3 is a diagram illustrating the operation of a cloud-based battery management system according to one embodiment of the present invention.

[0084] As illustrated in FIG. 3, a cloud BMS (100) according to one embodiment of the present invention receives battery measurement data and charging data from an OEM server (300) and a charging device (400), respectively.

[0085] The above cloud BMS (100) detects an abnormal condition during charging by using the battery's measurement data and charging data.

[0086] The above cloud BMS (100) calculates the error between the received battery measurement data and the received charging data, and detects that an abnormal condition has occurred if the calculated error exceeds a predetermined threshold error.

[0087] At this time, the cloud BMS (100) calculates the error between the received battery measurement data and the received charging data corresponding to the received battery measurement data. For example, the voltage, current, and temperature of the battery measurement data are calculated to calculate the error for each of the charging voltage, charging current, and temperature of the corresponding charging data, and if any one of the errors exceeds the critical error or if the average of the errors exceeds the critical error, it can be detected that an abnormal condition has occurred.

[0088] The above error is calculated by calculating the absolute value of the difference between the battery's measurement data and charging data.

[0089] Additionally, the cloud BMS (100) can be configured to detect the occurrence of an abnormal state using an abnormal state detection model. The abnormal state detection model has an encoder-decoder structure.

[0090] The structure of the above encoder and decoder will be described in detail with reference to FIGS. 4 and 5, and detecting the occurrence of an abnormal state using an abnormal state detection model will be described in detail with reference to FIGS. 6 and 7.

[0091] The above cloud BMS (100) detects the occurrence of an abnormal condition and transmits it to the charging device (400) so that charging can be stopped immediately.

[0092] Additionally, if the occurrence of an abnormal condition is not detected, the cloud BMS (100) estimates the status information of the battery using the battery measurement data and charging data.

[0093] Estimating the status information of the above battery is performed using a product encoder, which is a learning model for estimating battery status information.

[0094] Estimating the status information of the above battery and the product encoder will be described in detail with reference to Fig. 4.

[0095] The above cloud BMS (100) transmits the estimated battery status information and the received battery measurement data to the charging device (400) to control the charging of the battery.

[0096] FIG. 4 is a diagram illustrating an encoder according to one embodiment of the present invention.

[0097] As illustrated in FIG. 4, an encoder according to one embodiment of the present invention is configured to input voltage, current, temperature, or a combination thereof, and encode and output battery status information.

[0098] The above encoder is constructed by learning through a deep neural network. In other words, a deep neural network that has completed training is constructed as an encoder.

[0099] A deep neural network for constructing an encoder through the above learning is configured to include an input layer into which voltage, current, temperature or a combination thereof constituting learning data are input, at least one hidden layer, and an output layer that outputs battery status information encoded according to the input.

[0100] The above learning data is composed by labeling the battery status information with voltage, current, temperature, or a combination thereof.

[0101] The above input layer, each hidden layer, and output layer are configured to include at least one input node, one hidden node, and one output node, and are each connected through a link having a predetermined weight.

[0102] The above output node outputs the learning result according to the input, and since it already knows the label (i.e., battery status information) according to the input input to the input layer of the encoder during the learning process, it can reduce the error of the learning result output during the learning process by updating (adjusting) the weight of the link through the backpropagation method.

[0103] That is, the input of the encoder configured by completing learning is voltage, current, temperature, or a combination thereof, and the output is the status information of the encoded battery.

[0104] The encoder of the present invention comprises a product encoder and a charging device encoder. The product encoder can be used to estimate battery status information and as a component of an abnormality detection model.

[0105] The above product encoder is configured by learning first learning data that labels the battery status information on the battery voltage, current, temperature, or a combination thereof collected from the OEM server (300).

[0106] Accordingly, the cloud BMS (100) configures the third input data according to the input of the product encoder from the received measurement data of the battery, inputs the third input data into the product encoder, and encodes the third input data into the battery status information, thereby estimating the battery status information.

[0107] Configuring the third input data refers to arranging the measurement data such that voltage, current, temperature, or a combination thereof are input to the input nodes of the product encoder. For example, if voltage, current, and temperature are to be input to the first, second, and third input nodes, the cloud BMS (100) configures the third input data by arranging the measurement data in the order of voltage, current, and temperature.

[0108] At this time, the cloud BMS (100) can supplement the battery measurement data using the received charging data. For example, if the battery measurement data does not include current, the charging current in the charging data can be substituted to supplement the battery measurement data, thereby estimating the battery status information.

[0109] Meanwhile, the charger encoder is configured by learning third learning data that labels the battery status information with charging voltage, charging current, charging temperature, or a combination thereof collected from the charger.

[0110] That is, the input of the charging device encoder is charging data, and the output is battery status information encoded according to the charging data.

[0111] The above charging device encoder can be used to estimate battery status information using charging data, but it is preferable to use it to construct an abnormality detection model for abnormality detection.

[0112] FIG. 5 is a diagram illustrating a decoder according to one embodiment of the present invention.

[0113] As illustrated in FIG. 5, a decoder according to one embodiment of the present invention is configured to input battery status information and reconstruct the input into voltage, current, temperature, or a combination thereof and output it.

[0114] The above decoder is constructed by training data labeled with battery status information such as voltage, current, temperature, or a combination thereof using a deep neural network. The deep neural network that has completed training becomes the decoder.

[0115] The deep neural network for constructing a decoder through the above learning is configured to include an input layer into which battery status information constituting the learning data is input, at least one hidden layer, and an output layer that outputs the reconstructed voltage, current, temperature, or a combination thereof.

[0116] The above input layer, each hidden layer, and output layer are configured to include at least one input node, one hidden node, and one output node, and are each connected through a link having a predetermined weight.

[0117] The above output node outputs the learning result according to the input, and since it already knows the label according to the input data input to the input layer of the decoder during the learning process, it can reduce the error of the learning result output during the learning process by updating (adjusting) the weight of the link through the back propagation method.

[0118] That is, the input of the decoder constructed by completing learning becomes the battery status information encoded through the encoder, and the output is composed of the reconstructed voltage, current, temperature, or a combination thereof.

[0119] The decoder in the present invention comprises a product decoder and a charging device decoder.

[0120] The above product decoder is configured by learning second learning data that labels the voltage, current, temperature or a combination thereof of the battery collected from the OEM server (300) in the battery status information.

[0121] The above-mentioned charging device decoder is configured by learning the fourth learning data that labels the charging voltage, charging current, charging temperature, or a combination thereof collected from the charging device (400) in the battery status information.

[0122] FIG. 6 is a diagram illustrating an abnormal state detection model according to one embodiment of the present invention.

[0123] As illustrated in FIG. 6, an abnormal condition detection model according to one embodiment of the present invention is composed of an encoder-decoder.

[0124] The above abnormal condition detection model is configured to take voltage, current, temperature, or a combination thereof as input and reconstruct the voltage, current, temperature, or a combination thereof as output according to the decoder.

[0125] The above abnormal condition detection model is configured with a decoder connected to the rear end of the encoder.

[0126] The above encoder may be configured as a product encoder or a charging device encoder, and the decoder may be configured as a product decoder or a charging device decoder.

[0127] That is, the abnormal state detection model can be composed of a total of four structures, including a first abnormal state detection model composed of a product encoder and a charging device decoder, a second abnormal state detection model composed of a charging device encoder and a product decoder, a third abnormal state detection model composed of a product encoder and a product decoder, and a fourth abnormal state detection model composed of a charging device encoder and a charging device decoder.

[0128] Additionally, the cloud BMS (100) configures second input data corresponding to the input of each abnormal condition detection model. The configuration of the second input data is identical to the configuration of the first input data described with reference to FIG. 4, with the only difference being whether battery measurement data or charging data is input.

[0129] The above cloud BMS (100) inputs each configured second input data into the corresponding abnormal state detection model, and calculates the error between the second input data and the output data of each abnormal state detection model.

[0130] The above error is calculated by calculating the absolute value of the difference between the second input data and the output data corresponding to the second input data. In other words, the error is calculated for each of the voltage, current, and temperature of the output data that is reconstructed by reconstructing the voltage, current, and temperature of the second input data, and the average is calculated.

[0131] In addition, the cloud BMS (100) detects that an abnormal state has occurred when the average result of each error calculated through each abnormal state detection model exceeds a predetermined threshold error, or when any one of the above-described errors exceeds a predetermined threshold error.

[0132] FIG. 7 is a diagram illustrating an abnormal condition detection model according to another embodiment of the present invention.

[0133] As illustrated in FIG. 7, an abnormal state detection model according to another embodiment of the present invention is configured to take voltage, current, temperature, or a combination thereof as input, and reconfigure the voltage, current, temperature, or a combination thereof as output based on learning.

[0134] The above abnormal condition detection model is composed of a deep neural network including an input layer that comprises voltage, current, temperature or a combination thereof that constitutes learning data, at least one hidden layer, and an output layer that outputs reconstructed voltage, current, temperature or a combination thereof.

[0135] The above input layer, each hidden layer, and output layer are configured to include at least one input node, one hidden node, and one output node, and are each connected through a link having a predetermined weight.

[0136] The above output node outputs the learning result according to the input, and since it already knows the label according to the input data input to the input layer during the learning process, it can reduce the error of the learning result output during the learning process by updating (adjusting) the weight of the link through the back propagation method.

[0137] The above abnormal state detection model can be configured as a fifth abnormal state detection model and a sixth abnormal state detection model depending on the learning data.

[0138] The above fifth abnormal condition detection model is configured by learning fifth learning data labeled with the voltage, current, temperature or a combination thereof of the battery (i.e., battery measurement data) collected from the OEM server (300) of the battery mounted on the product, and the charging voltage, charging current, charging temperature or a combination thereof collected from the charging device (400).

[0139] The above-mentioned sixth abnormal condition detection model is configured by learning the sixth learning data labeled with the voltage, current, temperature or combination thereof of the battery collected from the OEM server (300) and the charging voltage, charging current, charging temperature or combination thereof collected from the charging device (400) during charging.

[0140] That is, the input of the 5th abnormal state detection model is the battery measurement data, and the output is the battery measurement data reconstructed into charging data, and the input of the 6th abnormal state detection model is the charging data, and the output is the battery charging data reconstructed into the battery measurement data.

[0141] The above cloud BMS (100) uses the measurement data and charging data of the received battery to configure third input data corresponding to the input of the fifth abnormal state detection model and the sixth abnormal state detection model.

[0142] The above cloud BMS (100) inputs each configured third input data into each abnormal state detection model corresponding to the third input data and calculates the error between each third input data and the output data of each abnormal state detection model.

[0143] Here, the error is calculated as the average of the absolute value of the difference between the voltage, current, and temperature that constitute the third input data and the voltage, current, and temperature that constitute the output data.

[0144] In addition, the cloud BMS (100) detects that an abnormal condition has occurred if the average result of each of the above-described errors exceeds a predetermined threshold error or if any one of the above-described errors exceeds a predetermined threshold error.

[0145] As described with reference to FIGS. 4 to 7, learning for configuring an encoder, decoder, and abnormal condition detection model can be performed in a cloud BMS (100). However, it is not limited thereto and can be performed through a separate learning server.

[0146] FIG. 8 is a block diagram showing the configuration of a cloud-based battery management system according to one embodiment of the present invention.

[0147] As illustrated in FIG. 8, a cloud BMS (100) according to one embodiment of the present invention is configured to include a battery measurement data receiving unit (110), a charging data receiving unit (120), an abnormality detection unit (130), a battery status information estimation unit (140), a notification providing unit (150), a battery status information transmitting unit (160), and a user information management unit (170).

[0148] The above battery measurement data receiving unit (110) receives battery measurement data from the OEM server (300) of the battery mounted on the product (200) and stores the received battery measurement data in a database (600).

[0149] The above charging data receiving unit (120) receives charging data from the charging device (400) when charging the battery and stores it in the database (600).

[0150] The above abnormal condition detection unit (130) detects an abnormal condition using the measurement data and charging data of the received battery and stores it in the database (600).

[0151] Detecting the above abnormal condition is done by calculating the error between the measured data of the received battery and the charging data or by using an abnormal condition detection model.

[0152] When using the above abnormal state detection model, the abnormal state detection unit (130) configures input data suitable for the input of the abnormal state detection model using the measurement data and charging data of the battery received above, inputs the configured input data into the abnormal state detection model, and detects the abnormal state by calculating the error between the input data and the output data of the abnormal state detection model.

[0153] Detecting the above abnormal condition is described with reference to FIGS. 6 and 7, so it is omitted here.

[0154] Additionally, the battery status information estimation unit (140) estimates the battery status information using the received battery measurement data, charging data, or a combination thereof, and stores it in the database (600).

[0155] The state information of the above battery is estimated using a product encoder, and the battery state information estimation unit (140) configures input data (fourth input data) according to the input of the product encoder using the measurement data and charging data of the battery, and inputs the configured fourth input data into the product encoder to encode the fourth input data into the state information of the battery, thereby estimating the state information.

[0156] Of course, when the product (200) is operated (i.e., battery discharged), charging data is not received. Therefore, the battery status information estimation unit (140) estimates the battery status information by configuring the fourth input data according to the input of the product encoder using the received battery measurement data.

[0157] Estimating the status information of the above battery is described with reference to Fig. 4, so it is omitted here.

[0158] In addition, the notification provision unit (150) detects an abnormal condition and, when it detects that an abnormal condition has occurred, provides a notification to the charging device (400), user terminal (500), or a combination thereof.

[0159] The above battery status information transmitter (160) transmits the estimated battery status information, including the received battery measurement data, to the charging device (400). At this time, the battery status information transmitter (160) can transmit the estimated battery status information to the user terminal (500).

[0160] The above user information management unit (170) performs the function of registering and managing user information for battery management.

[0161] The above user information may be configured to include an identifier of the battery, an identifier of the user, an identifier of the product, or a combination thereof.

[0162] FIG. 9 is a flowchart illustrating a procedure for managing a battery based on a cloud according to one embodiment of the present invention.

[0163] As illustrated in FIG. 9, a procedure for managing a cloud-based battery according to one embodiment of the present invention first performs a battery measurement data reception step (S110) in which a cloud BMS (100) receives measurement data of a battery and a charging data reception step (S110) in which charging data for the battery is received.

[0164] The measurement data of the above battery is received from the OEM server (300) of the battery, and the charging data is received from the charging device (400) that charges the battery.

[0165] Next, the cloud BMS (100) performs a cloud BMS configuration step of configuring a BMS (i.e., cloud BMS) on the cloud by storing and managing the measurement data and charging data of the battery in a database (600) (S120).

[0166] That is, the cloud BMS (100) is configured for each battery. Accordingly, the measurement data and charging data of the battery are stored and managed in a database (600) for each battery.

[0167] Through this, the cloud BMS (100) can be configured for each battery, thereby integrating and managing multiple batteries equipped in multiple products (200) or charged in a charging device (400).

[0168] Next, the cloud BMS (100) performs an abnormal condition detection step to detect an abnormal condition using the measurement data and charging data of the received battery (S130).

[0169] The above abnormal condition detection step calculates the error between the measurement data and charging data of the received battery, or detects the abnormal condition using an abnormal condition detection model.

[0170] The above abnormal state detection step further includes a second input data configuration step of configuring second input data to suit the input of the first to fourth abnormal state detection models using the received battery measurement data and charging data, when the above abnormal state detection model is configured with the first to fourth abnormal state detection models described with reference to FIG. 6.

[0171] Meanwhile, in the case where the abnormal state detection model is composed of the fifth and sixth abnormal state detection models described with reference to FIG. 7, a third input data configuration step is further included to configure third input data to match the input of the fifth and sixth abnormal state detection models using the measurement data and charging data of the received battery.

[0172] In addition, the abnormal state detection step inputs each configured input data into the corresponding abnormal state detection model and performs an error calculation step of calculating the error between the input data and the output data of each abnormal state detection model, and detects the abnormal state using the calculated error and a predetermined threshold error.

[0173] The results of detecting the above abnormal condition are stored and managed in a database (600) for each battery through the cloud BMS configuration step.

[0174] Detecting the above abnormal condition is described with reference to FIGS. 6 and 7, so it is omitted here.

[0175] At this time, if an abnormal condition is detected (S140), a notification step is performed to send a notification to the charging device (400) and the user terminal (500) (S151).

[0176] This allows charging to be stopped immediately before any damage occurs due to an abnormal condition.

[0177] Additionally, if an abnormal condition is not detected (S140), a battery status information estimation step is performed to estimate battery status information using the received battery measurement data and charging data (S150).

[0178] At this time, the battery status information estimation step further includes a first input data configuration step of configuring first input data to match the input of the product encoder using the battery measurement data and charging data, and since estimating the battery status information has been described with reference to FIG. 4, further detailed description will be omitted.

[0179] In addition, the estimated battery status information is stored and managed in the database (600) for each battery through the cloud BMS configuration step.

[0180] Thereafter, the cloud BMS (100) performs a battery status information transmission step of transmitting the estimated battery status information, including the measurement data of the received battery, to the charging device (400) (S160).

[0181] In addition, a cloud-based battery management system (100) according to one embodiment of the present invention is configured to include a memory that stores a program code implementing a method for managing a battery in a cloud-based manner, as described with reference to FIG. 9, and a processor configured to load and execute the program code stored in the memory.

[0182] As described above, the present invention implements a battery management system on the cloud, receives measurement data of a battery and charging data of a charging device, performs the role of a BMS in the cloud, and provides the data to each product or charging device, thereby lowering the ASIL of the BMS or charging device of the battery installed in each product and enabling integrated management of individual batteries.

[0183] In addition, the present invention has the effect of enabling stable use or charging of a battery by accurately estimating battery status information using battery measurement data and charging data in a cloud BMS and then providing the result to each battery user and charging device.

[0184] In addition, the present invention has the effect of preventing dangers caused by abnormal conditions such as fire in a charging device by detecting the occurrence of an abnormal condition during charging in the charging device and immediately stopping charging.

[0185] In addition, although the preferred embodiments of the present invention have been illustrated and described above, the present invention is not limited to the specific embodiments described above, and various modifications can be implemented by a person having ordinary skill in the art to which the invention pertains without departing from the gist of the present invention claimed in the claims, and such modifications should not be understood individually from the technical idea or prospect of the present invention.

[0186] The present invention, configured as described above, implements a battery management system on the cloud, receives measurement data of a battery and charging data of a charging device, performs the role of a BMS in the cloud, and provides the data to each product or charging device, thereby lowering the ASIL of the BMS or charging device of the battery installed in each product, managing individual batteries in an integrated manner, and accurately estimating the status information of the battery using the measurement data and charging data of the battery in the cloud BMS and then providing the result to each battery user and charging device, thereby enabling the stable use or charging of the battery, and preventing the risk of a fire or other abnormal condition in the charging device by detecting the occurrence of an abnormal condition during charging in the charging device and immediately stopping the charging, thereby having industrial applicability.

Claims

1. Battery measurement data reception step for receiving battery measurement data from the OEM (Original Equipment Manufacturing) server of the battery installed in the product; A charging data receiving step for receiving charging data of the battery from a charging device; and It includes a cloud BMS configuration step for configuring a BMS on the cloud by storing and managing the above-described measurement data and charging data for each battery in a database; A battery management method characterized in that it manages multiple batteries equipped in multiple products or charged in a charging device through the above-described cloud BMS.

2. In claim 1, The above measurement data includes voltage, current, temperature or a combination thereof of the battery, A battery management method, characterized in that the charging data includes charging voltage, charging current, charging temperature, or a combination thereof.

3. In claim 1, The above battery management method is, It further includes a battery state information estimation step for estimating the state information of the battery, including SOC (state of charge), SOH (state of health), RUL (remaining useful lifetime) or a combination thereof, using the received measurement data and charging data; A battery management method characterized in that the estimated battery status information is stored and managed in the database for each battery through the cloud BMS configuration step.

4. In claim 3, The above battery status information estimation step is: A first input data configuration step for configuring first input data by supplementing the measurement data of the battery received using the received charging data in accordance with the input of the product encoder; and It further includes a state information encoding step for estimating the state information of the battery by inputting the first input data into the product encoder and encoding the supplemented measurement data into the state information of the battery; A battery management method characterized in that the above product encoder is configured by learning first learning data that labels battery status information such as voltage, current, temperature, or a combination thereof of the battery collected from the OEM server of the battery mounted on the product.

5. In claim 1, The above battery management method is, It further includes an abnormality detection step for detecting an abnormality using the above-mentioned received measurement data and charging data; A battery management method characterized in that the results of detecting the above abnormal condition are stored and managed in the database for each battery through the cloud BMS configuration step.

6. In claim 5, The above abnormal condition detection step is: A second input data configuration step of configuring second input data suitable for a first abnormal state detection model including a product encoder and a charger decoder, a second abnormal state detection model including a charger encoder and a product decoder, a third abnormal state detection model including a product encoder and a product decoder, a fourth abnormal state detection model including a charger encoder and a charger decoder, or a combination thereof; and An error calculation step of inputting each of the second input data into a corresponding abnormal state detection model and calculating an error between each of the second input data and the output data of each abnormal state detection model; A battery management method characterized in that an abnormal condition is detected when the average result of each of the above-described errors exceeds a predetermined threshold error or when any one of the above-described errors exceeds a predetermined threshold error.

7. In claim 6, The above abnormal condition detection step is: The above product encoder is configured by learning the first learning data that labels the battery status information such as the voltage, current, temperature, or a combination thereof collected from the OEM server of the battery mounted on the product. The above product decoder is configured by learning the second learning data that labels the voltage, current, temperature or a combination thereof of the collected battery status information. The above charging device encoder is configured by learning third learning data that labels battery status information on charging voltage, charging current, charging temperature, or a combination thereof collected from the charging device. A battery management method characterized in that the above charging device decoder is configured by learning fourth learning data that labels the collected charging voltage, charging current, charging temperature, or a combination thereof in the battery status information.

8. In claim 6, The above abnormal condition detection step is: A third input data configuration step of configuring third input data suitable for the input of the fifth abnormal state detection model and the sixth abnormal state detection model using the above-described received measurement data and charging data; and It further includes an error calculation step of inputting each of the configured third input data into each of the corresponding abnormal state detection models and calculating an error between each of the third input data and the output data of each of the abnormal state detection models; A battery management method characterized in that an abnormal condition is detected when the average result of each of the above-described errors exceeds a predetermined threshold error or when any one of the above-described errors exceeds a predetermined threshold error.

9. In claim 8, The above fifth abnormal condition detection model is configured by learning fifth learning data labeled with the voltage, current, temperature or a combination thereof of the battery collected from the OEM server of the battery installed in the product and the charging voltage, charging current, charging temperature or a combination thereof collected from the charging device. A battery management method characterized in that the sixth abnormal condition detection model is configured by learning sixth learning data labeled with voltage, current, temperature or a combination thereof of a battery collected from an OEM server of a battery equipped with a product, and charging voltage, charging current, charging temperature or a combination thereof collected from a charging device.

10. A memory storing a program code implementing the battery management method according to any one of claims 1 to 9; and A battery management system comprising a processor configured to load and execute program code stored in the memory.

Images