Battery state prediction device and operation method thereof

Abstract

A battery state prediction device according to one embodiment of the present disclosure may include a data management unit that extracts first battery data acquired within a predetermined time period after charging of a battery is completed and second battery data acquired within a predetermined time period after discharging of the battery is completed, and a controller that applies the first battery data to a first deep learning model to acquire first state data that predicts the state of the battery, applies the second battery data to a second deep learning model to acquire second state data that predicts the state of the battery, and predicts the state of the battery based on the first state data and the second state data.

Description

[Technical Field]

[0001] This application claims the benefit of priority based on Korean Patent Application No. 10-2022-0109524, filed on August 30, 2022, the entire contents of which are incorporated herein by reference.

[0002] SUMMARY OF THE INVENTION Embodiments of the present disclosure relate to a battery state prediction device and a method of operation thereof. [Background technology]

[0003] Electric vehicles receive electricity from an external source to charge their batteries, and then use the voltage stored in the batteries to drive the motor to generate power. Electric vehicle batteries can generate heat due to chemical reactions during charging and discharging, which can damage the battery's performance and lifespan. Therefore, a Battery Management System (BMS) is used to monitor the battery's temperature, voltage, and current, allowing the battery's State of Health (SOH) to be predicted.

[0004] A battery management device can analyze the battery state by training an artificial intelligence model that analyzes the battery state using large amounts of battery data. However, the battery management device has problems such as reduced efficiency, reduced performance and speed of the artificial intelligence model, and errors in the process of managing and analyzing large amounts of battery data. Therefore, it is necessary to construct a learning dataset suitable for input into the artificial intelligence model to analyze the battery state. Summary of the Invention [Problem to be solved by the invention]

[0005] One object of an embodiment of the present disclosure is to provide a battery state prediction device and an operating method thereof that can extract a portion of battery data as learning data to train an artificial intelligence model that analyzes the battery state, thereby improving the performance of the artificial intelligence model.

[0006] The technical problems of the embodiments of the present disclosure 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 following description. [Means for solving the problem]

[0007] A battery state prediction device according to one embodiment of the present disclosure may include a data management unit that extracts first battery data acquired within a predetermined time period after charging of a battery is completed and second battery data acquired within a predetermined time period after discharging of the battery is completed, and a controller that applies the first battery data to a first deep learning model to acquire first state data that predicts the state of the battery, applies the second battery data to a second deep learning model to acquire second state data that predicts the state of the battery, and predicts the state of the battery based on the first state data and the second state data.

[0008] According to one embodiment, the controller may combine the first status data and the second status data to generate third status data, and predict the status of the battery based on the third status data.

[0009] According to one embodiment, the controller may apply the first battery data to a first convolutional neural network (CNN) model to extract a feature of the first battery data, and apply the second battery data to a second convolutional neural network model to extract a feature of the second battery data.

[0010] According to one embodiment, the controller generates a third value based on a weighted sum of a first value obtained by converting characteristics of the first battery data into an embedding vector and a second value obtained by converting characteristics of the second battery data into an embedding vector, and can predict the state of the battery based on the third value.

[0011] According to one embodiment, the first battery data and the second battery data may include cumulatively measured voltage, current, and temperature of the battery, and the first status data, the second status data, and the third status data may include a State of Health (SOH) of the battery calculated based on the first battery data and the second battery data.

[0012] An operating method of a battery state prediction device according to one embodiment of the present disclosure may include the steps of: extracting first battery data from battery data that is acquired within a predetermined time period after charging of the battery is completed; extracting second battery data from the battery data that is acquired within a predetermined time period after discharging of the battery is completed; applying the first battery data to a first deep learning model to acquire first state data that predicts the state of the battery; applying the second battery data to a second deep learning model to acquire second state data that predicts the state of the battery; and predicting the state of the battery based on the first state data and the second state data.

[0013] According to one embodiment, the step of predicting the state of the battery based on the first state data and the second state data may include combining the first state data and the second state data to generate third state data, and predicting the state of the battery based on the third state data.

[0014] According to one embodiment, the step of applying the first battery data to a first deep learning model to obtain first state data predicting the state of the battery may apply the first battery data to a first convolutional neural network model to extract characteristics of the first battery data, and the step of applying the second battery data to a second deep learning model to obtain second state data predicting the state of the battery may apply the second battery data to a second convolutional neural network model to extract characteristics of the second battery data.

[0015] According to one embodiment, the step of predicting the state of the battery based on the first state data and the second state data may generate a third value based on a weighted sum of a first value obtained by converting characteristics of the first battery data into an embedding vector and a second value obtained by converting characteristics of the second battery data into an embedding vector, and predict the state of the battery based on the third value.

[0016] According to one embodiment, predicting the state of the battery based on the first state data and the second state data may predict a SOH of the battery. [Brief explanation of the drawings]

[0017] [Figure 1] FIG. 1 is a diagram illustrating a battery pack according to an embodiment of the present disclosure. [Figure 2] 1 is a diagram for generally explaining a battery state prediction device according to an embodiment of the present disclosure; [Figure 3] 1 is a block diagram showing a configuration of a battery state prediction device according to an embodiment of the present disclosure. [Figure 4a] 1 is a graph showing voltage changes of battery data according to an embodiment of the present disclosure. [Figure 4b] 10 is a graph showing voltage changes of first battery data according to an embodiment of the present disclosure. [Figure 4c] 10 is a graph showing voltage changes of second battery data according to an embodiment of the present disclosure. [Figure 5] 1 is a flowchart illustrating a method of operating a battery state prediction device according to an embodiment of the present disclosure. [Figure 6] 1 is a block diagram showing a hardware configuration of a computing system that realizes a battery state prediction device according to an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE INVENTION

[0018] Some embodiments of the present disclosure will be described in detail below with reference to exemplary drawings. It should be noted that when assigning reference numerals to components in each drawing, the same components are assigned the same numerals as much as possible when they appear in other drawings. Furthermore, when describing the embodiments of the present disclosure, if a detailed description of related known structures or functions is deemed to hinder understanding of the embodiments of the present disclosure, the detailed description will be omitted.

[0019] When describing components of the embodiments of the present disclosure, terms such as "first," "second," "A," "B," "(a)," and "(b)" may be used. Such terms are merely used to distinguish the component from other components and do not limit the nature, order, or sequence of the components. Furthermore, unless otherwise defined, all terms used herein, 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 of the present disclosure belong. Terms defined in commonly used dictionaries are to be interpreted as having a meaning consistent with the context of the relevant art, and are not to be interpreted in an idealized or overly formal sense unless expressly defined in this disclosure.

[0020] FIG. 1 is a diagram illustrating a battery pack according to an embodiment of the present disclosure. Referring to FIG. 1 , a battery pack 1000 according to an embodiment of the present disclosure may include a battery module 100 , a battery state prediction device 200 , and a relay 300 .

[0021] The battery module 100 may include a plurality of battery cells 110, 120, 130, and 140. Although the number of battery cells is shown as four in FIG. 1, the number is not limited thereto, and the battery module 100 may include n (n is a natural number equal to or greater than 2) battery cells.

[0022] The battery module 100 can supply power to a target device (not shown). To this end, the battery module 100 can be electrically connected to the target device. Here, the target device can include an electrical, electronic, or mechanical device that operates by receiving power from a battery pack 1000 including a plurality of battery cells 110, 120, 130, and 140. For example, the target device can be, but is not limited to, an electric vehicle (EV) or an energy storage system (ESS).

[0023] The plurality of battery cells 110, 120, 130, 140 are basic units of a battery that can be used by charging and discharging electrical energy, and may be, but are not limited to, lithium ion (Li-ion) batteries, lithium ion polymer (Li-ion polymer) batteries, nickel cadmium (Ni-Cd) batteries, nickel metal hydride (Ni-MH) batteries, etc. Meanwhile, although Fig. 1 shows one battery module 100, according to an embodiment, a plurality of battery modules 100 may be included.

[0024] The battery state prediction device 200 can predict the states of the battery cells 110, 120, 130, 140, including the lifespan (SOH) of the battery cells 110, 120, 130, 140, based on the temperature, current, and voltage data of the battery cells 110, 120, 130, 140. The battery state prediction device 200 can predict the states of the battery cells 110, 120, 130, 140 by battery temperature, current, and voltage, based on the battery data of the battery cells 110, 120, 130, 140.

[0025] According to an embodiment, the battery state prediction device 200 may be implemented in the form of a battery management system (BMS). Also, according to an embodiment, the battery state prediction device 200 may be mounted in the battery management system.

[0026] Here, the battery management unit can manage and / or control the state and / or operation of the battery module 100. For example, the battery management unit can manage and / or control the state and / or operation of the plurality of battery cells 110, 120, 130, and 140 included in the battery module 100. The battery management unit can manage charging and / or discharging of the battery module 100.

[0027] The battery management device can also monitor the voltage, current, temperature, etc. of the battery module 100 and / or each of the plurality of battery cells 110, 120, 130, and 140 included in the battery module 100. For monitoring via the battery management device, sensors and various measurement modules (not shown) can be further provided in the battery module 100, a charge / discharge path, or at any position in the battery module 100. The battery management device can calculate parameters indicating the state of the battery module 100, such as SOC (State of Charge) and SOH (State of Health), based on the measured values ​​of the monitored voltage, current, temperature, etc.

[0028] The battery management unit can control the operation of the relay 300. For example, the battery management unit can turn on the relay 300 to supply power to a target device. Also, the battery management unit can turn on the relay 300 when a charging device is connected to the battery pack 1000.

[0029] The battery management unit can calculate the cell balancing time for each of the multiple battery cells 110, 120, 130, and 140. Here, the cell balancing time can be defined as the time required for balancing the battery cells. For example, the battery management unit can calculate the cell balancing time based on the SOC (State of Charge), battery capacity, and balancing efficiency of each of the multiple battery cells 110, 120, 130, and 140.

[0030] FIG. 2 is a diagram for generally explaining a battery state prediction device according to an embodiment of the present disclosure. 2, the battery state prediction device 200 extracts a portion of the battery data (A), and the battery state prediction device 200 can acquire the battery data (A) of the multiple battery cells 110, 120, 130, and 140. The battery management device can acquire the battery data (A) including battery measurement values ​​from a voltage value at which the SOC of the battery is 0% to a voltage value at which the SOC of the battery reaches 100%, in order to check the state of health (SOH) of the multiple battery cells 110, 120, 130, and 140. Therefore, the battery state prediction device 200 can acquire the battery data (A) including the voltages, currents, and temperatures of the multiple battery cells 110, 120, 130, and 140 measured cumulatively during a charge / discharge period.

[0031] The battery state prediction device 200 can construct a training data set to be used in the deep learning model of the battery state prediction device 200 based on the battery data (A). Deep learning is a type of machine learning algorithm and refers to a technology in which artificial neural networks are connected in a multi-layered form. Machine learning is a technology that enables a computer to learn and predict a certain result. In general, the results of machine learning include a process of preparing training data for machine learning, training the machine in a manner appropriate for the problem, validating the model with test data, and predicting a result using the model that has passed the validation.

[0032] In machine learning, it is important that the training data closely represents the features that are being generalized through machine learning, so training data is generated using limited training data selected according to certain criteria. If the features to be generalized through machine learning are poorly related to the characteristics of the training data, sampling noise occurs, making it difficult for the machine problem analysis model to find inherent patterns. This increases model error and reduces the model's reliability, regardless of the accuracy of the machine problem analysis model itself. Therefore, machine learning technology requires time-consuming evaluation and data processing of the training data to select a training dataset.

[0033] The battery state prediction device 200 can extract a portion of the battery data (A) to generate a learning data set in order to train the first deep learning model 221 and the second deep learning model 222. Here, the first deep learning model 221 and the second deep learning model 222 may refer to learning models that can predict the state of a battery, including the battery's state of health (SOH), based on input battery data.

[0034] That is, the battery state prediction device 200 generates a training data set for machine learning the first deep learning model 221 and the second deep learning model 222, thereby reducing errors in the machine learning of the first deep learning model 221 and the second deep learning model 222 and increasing reliability. The battery state prediction device 200 inputs the training data set into the first deep learning model 221 and the second deep learning model 222, and can predict the states of the multiple battery cells 110, 120, 130, and 140.

[0035] FIG. 3 is a block diagram showing a configuration of a battery state prediction device according to an embodiment of the present disclosure. The configuration and operation of the battery state prediction device 200 will be specifically described below with reference to FIG.

[0036] Referring to FIG. 3, the battery state prediction device 200 may include a data management unit 210 and a controller 220. The data management unit 210 can collect battery data (A). For example, the battery data (A) can be defined as a value that records the change in the amount of electricity of the plurality of battery cells 110, 120, 130, and 140 from a discharged state to a fully charged state or from a fully charged state to a discharged state.

[0037] The data management unit 210 can extract first battery data (A1) acquired during a predetermined time period after the battery charging is completed, and second battery data (A2) acquired during a predetermined time period after the battery discharging is completed. Here, the first battery data (A1) and the second battery data (A2) can include cumulatively measured battery voltage, current, and temperature.

[0038] FIG. 4a is a graph showing voltage variation of battery data according to one embodiment of the present disclosure. 4a, a section in which the battery voltage suddenly changes can be extracted from the battery data (A). The data management unit 210 can set the section in which the battery voltage suddenly changes during a rest period after the battery is charged from the battery data (A) as first battery data (A1). Furthermore, the data management unit 210 can set the section in which the battery voltage suddenly changes during a discharge period after the battery is rested from the battery data (A) as second battery data (A2).

[0039] Conventionally, deep learning models have been trained using all of the voltage or current behavior of all battery data (A), but when the entire battery data (A) is used as training data, there is a problem that the analysis time is extremely long. The data management unit 210 can extract, from the battery data (A), data that is necessary for predicting the battery life due to a sudden change in the battery voltage or current, as first battery data (A1) and second battery data (A2).

[0040] FIG. 4b is a graph showing voltage changes of the first battery data according to one embodiment of the present disclosure. The first battery data (A1) shown in FIG. 4b includes information about the change in the battery voltage during the rest period after the battery is charged. As the battery undergoes repeated charge and discharge cycles, the battery resistance increases and the battery capacity may steadily deteriorate. Therefore, even after the battery is charged, the battery voltage or open circuit voltage (OCV) may steadily decrease. The data management unit 210 extracts the first battery data (A1) corresponding to the change in the voltage during the rest period after the battery is charged and can use it for training the deep learning model.

[0041] For example, the first battery data (A1) may include data for the last second of the battery charging period and 60 seconds of data for the rest period immediately after the battery is charged, i.e., the first battery data (A1) may include battery data (A) corresponding to a total period of 61 seconds.

[0042] FIG. 4c is a graph showing voltage changes of second battery data according to one embodiment of the present disclosure. The second battery data (A2) shown in FIG. 4c includes information about the voltage change of the battery when the battery enters the discharge period after rest. When a current is applied to the battery at the start of discharge after rest, the battery may be affected by ohmic resistance generated by the battery's internal resistance. Therefore, the second battery data (A2) including information about the voltage change of the battery during the discharge period after rest may include a voltage drop (IR drop) generated by the battery's internal resistance. Here, the voltage drop refers to the potential difference between two points in the conducting phase while current is flowing. The data management unit 210 extracts the second battery data (A2) corresponding to the voltage change during the discharge period after rest and uses it for training the deep learning model.

[0043] For example, the second battery data (A2) may include data for the last second of the rest period after the battery is charged and the 40 seconds of the discharge period immediately after the battery is rested, i.e., the second battery data (A2) may include battery data (A) corresponding to a total of 41 seconds.

[0044] 3, the data management unit 210 may collect first battery data (A1) and second battery data (A2) as learning data sets to be input to the deep learning model. The data management unit 210 may manage the first battery data (A1) and the second battery data (A2) by classifying them into time series data. Here, time series data refers to data arranged at regular time intervals along the flow of time.

[0045] The controller 220 can predict the state of the battery by applying the first battery data (A1) and the second battery data (A2), including the temperature, current, and voltage changes of the multiple battery cells 110, 120, 130, 140, to a deep learning model with a time series analysis structure.

[0046] The controller 220 can obtain first status data that predicts the status of the battery by applying the first battery data (A1) to the first deep learning model 221. The controller 220 can also obtain second status data that predicts the status of the battery by applying the second battery data (A2) to the second deep learning model 222. That is, the controller 220 can input the first battery data (A1) and the second battery data (A2) into separate deep learning models and obtain separate output data.

[0047] Here, the first status data and the second status data may include battery life (SOH) data calculated using the first battery data (A1) as learning data for the first deep learning model 221, and battery life (SOH) data calculated using the second battery data (A2) as learning data for the second deep learning model 222.

[0048] According to an embodiment, the first deep learning model 221 and the second deep learning model 222 may include, for example, a one-dimensional convolutional neural network (1DCNN) model. Here, the convolutional neural network model can mainly extract features of matrix data or image data. Here, "one-dimensional" means that the kernel for convolution and the data sequence to which it is applied have a one-dimensional shape. The one-dimensional convolutional neural network can analyze the features of time series data or text.

[0049] According to the embodiment, the controller 220 may apply the first battery data (A1) to a first convolutional neural network model to extract a feature of the first battery data (A1), and may apply the second battery data (A2) to a second convolutional neural network model to extract a feature of the second battery data (A2).

[0050] The controller 220 can predict the state of the battery based on the first state data and the second state data. Specifically, the controller 220 can generate third state data by combining the first state data and the second state data. The controller 220 can predict the state of the battery based on the third state data. Here, the third state data can include battery life data.

[0051] According to an embodiment, the controller 220 may generate a third value by combining a first value obtained by converting the characteristics of the first battery data (A1) into an embedding vector and a second value obtained by converting the characteristics of the second battery data (A2) into an embedding vector. Here, embedding may be defined as a process of converting a natural language used by humans into a number sequence that can be understood by a machine. Specifically, the controller 220 may combine the first value and the second value by inputting them into a concatenation function.

[0052] The controller 220 may input data generated by combining the first value and the second value to a single-layer perceptron model to generate a third value. Here, the single-layer perceptron model may receive multiple signals as input and output one signal. The single-layer perceptron model may assign weights to the multiple input signals and generate an output signal based on a weighted sum obtained by adding up values ​​obtained by multiplying each signal by its own weight. Here, the weighted sum may be defined as an average value obtained by reflecting weights corresponding to the importance or degree of influence of data values ​​when calculating an average of data.

[0053] That is, the controller 220 can input the first value and the second value into a single-layer perceptron model and generate a third value based on a weighted sum of the first value and the second value. The third value can be generated by multiplying the first value and the second value by weights and adding up the values ​​multiplied by the weights. Here, the weight between the first value and the second value can be determined to maximize accuracy through testing using a test data set. The controller 220 can obtain a third value that is predicted data of the battery state. Here, the third value data can include battery life data.

[0054] One object of an embodiment of the present disclosure is to provide a battery state prediction device and an operating method thereof that can extract a portion of battery data as learning data to train an artificial intelligence model that analyzes the battery state, thereby improving the performance of the artificial intelligence model.

[0055] As described above, a battery state prediction device according to one embodiment of the present disclosure can improve the accuracy of deep learning models by extracting a portion of battery data as a learning dataset and then training the data in multiple deep learning models.

[0056] In addition, the battery state prediction device can simplify and reduce the size of the learning data, shorten the computing time for the deep learning model that predicts the battery state in real time, and also shorten the time required to collect and manage the learning data.

[0057] FIG. 5 is a flowchart illustrating a method of operation of a battery state prediction device according to an embodiment of the present disclosure. The battery state predicting device 200 is substantially similar to the battery state predicting device 200 described with reference to FIGS. 1 to 4c, and therefore will be described briefly below to avoid duplication of description.

[0058] Referring to FIG. 5, the operating method of the battery state prediction device includes a step (S101) of extracting first battery data (A1) from the battery data (A) that is acquired within a predetermined time period after the battery charging is completed, a step (S102) of extracting second battery data (A2) from the battery data that is acquired within a predetermined time period after the battery discharging is completed, a step (S103) of applying the first battery data (A1) to a first deep learning model 221 to acquire first state data that predicts the battery state, a step (S104) of applying the second battery data (A2) to a second deep learning model 222 to acquire second state data that predicts the battery state, and a step (S105) of predicting the battery state based on the first state data and the second state data.

[0059] Steps S101 to S105 will be specifically described below. In step S101, the data management unit 210 can collect battery data (A) of the multiple battery cells 110, 120, 130, and 140. For example, the battery data (A) can be defined as a value that records the change in the amount of electricity of the multiple battery cells 110, 120, 130, and 140 from a discharged state to a fully charged state or from a fully charged state to a discharged state. Therefore, the data management unit 210 can acquire battery data (A) that includes the voltage, current, and temperature of the multiple battery cells 110, 120, 130, and 140 that are cumulatively measured during the charge and discharge period.

[0060] In step S101, the data management unit 210 can extract first battery data (A1) from the battery data (A) that has been acquired during a predetermined time period after the battery charging is completed. In step S101, the data management unit 210 extracts the first battery data (A1) that corresponds to a voltage change in a pause period after the battery charging, and can use the extracted first battery data (A1) for training a deep learning model.

[0061] Here, the first battery data (A1) may include cumulatively measured battery voltage, current, and temperature. For example, the first battery data (A1) may include data for the last second of the battery charging period and 60 seconds of rest period immediately after the battery charging. That is, the first battery data (A1) may include battery data (A) corresponding to a total period of 61 seconds.

[0062] In step S102, the data management unit 210 can extract second battery data (A2) from the battery data (A) that has been acquired during a predetermined time period after the battery has been discharged. In step S102, the data management unit 210 extracts the second battery data (A2) that corresponds to the voltage change in the discharge period after the battery has been put to sleep, and can use the second battery data (A2) for training the deep learning model.

[0063] Here, the second battery data (A2) may include cumulatively measured battery voltage, current, and temperature. For example, the second battery data (A2) may include data from the last second of the rest period after the battery is charged and the 40 seconds of the discharge period immediately after the battery is rested. That is, the second battery data (A2) may include battery data (A) corresponding to a total of 41 seconds.

[0064] In step S103, the controller 220 can apply the first battery data (A1) to the first deep learning model 221 to obtain first state data that predicts the state of the battery.

[0065] Specifically, in step S103, the controller 220 can apply the first battery data (A1) including the temperature, current, and voltage changes of the multiple battery cells 110, 120, 130, and 140 to a deep learning model with a time series analysis structure to predict the battery state.

[0066] Here, the first status data may include battery life (SOH) data calculated using the first battery data (A1) as learning data for the first deep learning model 221. Here, the first deep learning model 221 may include, for example, a one-dimensional convolutional neural network model. The convolutional neural network model can mainly extract features of matrix data or image data. Here, "one-dimensional" means that a kernel for convolution and a sequence of data to be applied have a one-dimensional shape. The one-dimensional convolutional neural network can analyze the features of time series data or text.

[0067] In step S103, the controller 220 may apply the first battery data (A1) to a first convolutional neural network model to extract characteristics of the first battery data (A1).

[0068] In step S104, the controller 220 can apply the second battery data (A2) to the second deep learning model 222 to obtain second state data that predicts the state of the battery.

[0069] Specifically, in step S104, the controller 220 can predict the battery state by applying second battery data (A2) including temperature, current, and voltage changes of the multiple battery cells 110, 120, 130, and 140 to a deep learning model with a time series analysis structure. Here, the second state data can include battery life (SOH) data calculated using the second battery data (A2) as learning data for the second deep learning model 222. Here, the second deep learning model 222 can include, for example, a one-dimensional convolutional neural network model.

[0070] In step S104, the controller 220 may apply the second battery data (A2) to a second convolutional neural network model to extract features of the second battery data (A2).

[0071] In step S105, the controller 220 can predict the battery state based on the first state data and the second state data. Specifically, in step S105, the controller 220 can combine the first state data and the second state data to generate third state data. In step S105, the controller 220 can predict the battery state based on the third state data. Here, the third state data can include battery life data.

[0072] In step S105, according to an embodiment, the controller 220 may generate a third value by combining a first value obtained by converting the characteristics of the first battery data (A1) into an embedding vector and a second value obtained by converting the characteristics of the second battery data (A2) into an embedding vector. Here, embedding can be defined as a process of converting natural language used by humans into a number sequence that can be understood by machines.

[0073] Specifically, in step S105, the controller 220 can sequentially input the first value and the second value into a function to combine them. In step S105, the controller 220 can input data generated by combining the first value and the second value into a single-layer perceptron model to generate a third value. Here, the single-layer perceptron model can receive multiple signals as input and output one signal. The single-layer perceptron model can assign weights to the multiple input signals, and generate an output signal based on a weighted sum of values ​​obtained by multiplying each signal by its own weight.

[0074] In step S105, the controller 220 inputs the first value and the second value into the single-layer perceptron model, and generates a third value based on a weighted sum of the first value and the second value. The third value can be generated by multiplying the first value and the second value by weights and adding up the values ​​multiplied by the weights.

[0075] In step S105, the controller 220 can obtain a third value, which is predicted data of the battery state, where the third value data can include battery life data.

[0076] FIG. 6 is a block diagram showing the hardware configuration of a computing system that realizes a battery state prediction device according to an embodiment of the present disclosure. Referring to FIG. 6, a computing system 2000 according to one embodiment of the present disclosure may include a microcontroller unit (MCU) 2100, a memory 2200, an input / output interface (I / F) 2300, and a communication interface (I / F) 2400.

[0077] The MCU 2100 may be a processor that executes various programs (e.g., a battery state prediction program) stored in the memory 2200, processes various data through such programs, and performs the functions of the battery state prediction device 200 shown in Figure 1 described above.

[0078] The memory 2200 can store various programs related to the operation of the battery state prediction device 200. The memory 2200 can also store operation data of the battery state prediction device 200.

[0079] A plurality of such memories 2200 may be provided as necessary. The memories 2200 may be volatile memories or nonvolatile memories. The volatile memories 2200 may be RAM, DRAM, SRAM, etc. The nonvolatile memories 2200 may be ROM, PROM, EAROM, EPROM, EEPROM, flash memory, etc. The examples of the memories 2200 listed above are merely illustrative and are not limited to these examples.

[0080] The input / output I / F 2300 can provide an interface that connects input devices (not shown) such as a keyboard, mouse, or touch panel, and output devices such as a display (not shown), to the MCU 2100, enabling data to be sent and received.

[0081] The communication I / F 2400 is configured to be able to send and receive various data to and from a server, and may be any device that supports wired or wireless communication. For example, programs for resistance measurement and abnormality diagnosis, various data, and the like can be sent and received from a separately provided external server via the communication I / F 2400.

[0082] The above description merely exemplifies the technical concept of the present disclosure, and various modifications and variations are possible by a person having ordinary knowledge in the technical field to which the present disclosure pertains, without departing from the essential characteristics of the present disclosure.

[0083] Therefore, the embodiments disclosed in this disclosure are intended to illustrate, not limit, the technical idea of ​​the disclosure, and the scope of the technical idea of ​​the disclosure is not limited by such embodiments. The scope of protection of the disclosure is interpreted by the claims below, and all technical ideas within the equivalent range thereof are interpreted as being included in the scope of rights of the disclosure. [Explanation of symbols]

[0084] 1000: Battery pack 100: Battery module 200: Battery state prediction device 210: Data Management Department 220: Controller 221: First deep learning model 222: Second deep learning model 300: Relay A: Battery data A1: First battery data A2: Second battery data 2000: Computing Systems 2100:MCU 2200:Memory 2300: Input / output interface 2400:Communication I / F

Claims

1. a data management unit that extracts first battery data acquired during a predetermined time period after charging of the battery is completed and second battery data acquired during a predetermined time period after discharging of the battery is completed; a controller that applies the first battery data to a first deep learning model to obtain first state data that predicts the state of the battery, applies the second battery data to a second deep learning model to obtain second state data that predicts the state of the battery, and predicts the state of the battery based on the first state data and the second state data.

2. The battery state prediction device according to claim 1 , wherein the controller generates third state data by combining the first state data and the second state data, and predicts the state of the battery based on the third state data.

3. 3. The battery state prediction device of claim 2, wherein the controller applies the first battery data to a first convolutional neural network (CNN) model to extract characteristics of the first battery data, and applies the second battery data to a second convolutional neural network model to extract characteristics of the second battery data.

4. 4. The battery state prediction device of claim 3, wherein the controller generates a third value based on a weighted sum of a first value obtained by converting the characteristics of the first battery data into an embedded vector and a second value obtained by converting the characteristics of the second battery data into an embedded vector, and predicts the state of the battery based on the third value.

5. 3. The battery state prediction device according to claim 2, wherein the first battery data and the second battery data include cumulatively measured voltage, current, and temperature of the battery, and the first status data, the second status data, and the third status data include a SOH (State of Health) of the battery calculated based on the first battery data and the second battery data.

6. extracting first battery data acquired during a predetermined time period after charging of the battery is completed from the battery data; extracting second battery data from the battery data, the second battery data being acquired during a predetermined time period after the completion of discharging of the battery; Applying the first battery data to a first deep learning model to obtain first state data that predicts a state of the battery; Applying the second battery data to a second deep learning model to obtain second state data that predicts the state of the battery; and predicting the state of the battery based on the first state data and the second state data.

7. 7. The method for operating a battery state prediction device according to claim 6, wherein the step of predicting the state of the battery based on the first state data and the second state data comprises combining the first state data and the second state data to generate third state data, and predicting the state of the battery based on the third state data.

8. The step of applying the first battery data to a first deep learning model to obtain first state data predicting the state of the battery includes applying the first battery data to a first convolutional neural network (CNN) model to extract characteristics of the first battery data; 8. The method of claim 7, wherein the step of applying the second battery data to a second deep learning model to obtain second state data that predicts the state of the battery comprises applying the second battery data to a second convolutional neural network model to extract characteristics of the second battery data.

9. 9. The method of claim 8, wherein the step of predicting the state of the battery based on the first state data and the second state data comprises generating a third value based on a weighted sum of a first value obtained by converting a characteristic of the first battery data into an embedded vector and a second value obtained by converting a characteristic of the second battery data into an embedded vector, and predicting the state of the battery based on the third value.

10. 7. The method of claim 6, wherein the step of predicting the state of the battery based on the first state data and the second state data predicts a State of Health (SOH) of the battery.

Images