Battery state prediction device and operation method thereof

Abstract

A battery state prediction apparatus according to one embodiment disclosed herein may include: a data management unit for extracting first battery data acquired within a predetermined time after completion of battery charging and second battery data acquired within a predetermined time after completion of battery discharging; and a controller for obtaining first state data for predicting the state of the battery by applying the first battery data to a first deep learning model, obtaining second state data for predicting the state of the battery by applying the second battery data to a second deep learning model, and predicting a state of the battery based on the first state data and the second state data.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims priority to and the benefit of Korean Patent Application No. 10-2022-0109524, filed on August 30, 2022, which is hereby incorporated by reference in its entirety. Technical Field

[0003] The embodiments disclosed herein relate to a battery state prediction device and an operation method thereof. Background Art

[0004] Electric vehicles are supplied with electricity from the outside to charge batteries, and then the motor is driven by the voltage charged in the batteries to obtain power. The batteries of electric vehicles may have heat generated therein by chemical reactions occurring during charging and discharging, and this heat may impair the performance and life of the batteries. Therefore, a battery management system (BMS) that monitors the temperature, voltage, and current of the batteries is driven to predict the life (state of health (SOH)) of the batteries.

[0005] The battery management system can analyze the state of the battery by training a large amount of battery data in an artificial intelligence model that analyzes the state of the battery. However, the battery management system suffers from efficiency degradation in the process of managing and analyzing a large amount of battery data, which reduces the performance and speed of the artificial intelligence model and causes errors. Therefore, it is necessary to analyze the state of the battery by establishing a suitable training data set to be input into the artificial intelligence model. Summary of the invention

[0006] [Technical issues]

[0007] The embodiments disclosed herein aim to provide a battery status prediction device and an operating method thereof, wherein a portion of the battery data is extracted as training data to train an artificial intelligence model for analyzing the battery status, thereby improving the performance of the artificial intelligence model.

[0008] Technical problems of the embodiments disclosed herein are not limited to the above-mentioned technical problems, and other unmentioned technical problems will be clearly understood by ordinary technicians in the art from the following description.

[0009] [Technical solution]

[0010] According to an embodiment disclosed in this document, a battery state prediction device includes: a data management unit, which is configured to extract first battery data obtained within a predetermined time after battery charging is completed and second battery data obtained within a predetermined time after battery discharging is completed; and a controller, which is configured to: obtain first state data for predicting the state of the battery by applying the first battery data to a first deep learning model, obtain second state data for predicting the state of the battery by applying the second battery data to a second deep learning model, and predict the state of the battery based on the first state data and the second state data.

[0011] According to an embodiment, the controller may be further configured to generate third state data by combining the first state data with the second state data, and predict the state of the battery based on the third state data.

[0012] According to an embodiment, the controller may also be configured to extract features of the first battery data by applying the first battery data to a first convolutional neural network (CNN) model, and to extract features of the second battery data by applying the second battery data to a second CNN model.

[0013] According to an embodiment, the controller can also be configured to: generate a third value based on a weighted sum of a first value and a second value, the first value being obtained by converting a feature of the first battery data into an embedding vector, and the second value being obtained by converting a feature of the second battery data into an embedding vector; and predict the state of the battery based on the third value.

[0014] According to an embodiment, the first battery data and the second battery data may include cumulatively measured voltage, current, and temperature of the battery, and the first state data, the second state data, and the third state data may include a state of health (SOH) of the battery calculated based on the first battery data and the second battery data.

[0015] The operating method of the battery state prediction device according to the embodiment disclosed in this article includes the following operations: extracting first battery data obtained within a predetermined time after the battery charging is completed from the battery data, extracting second battery data obtained within a predetermined time after the battery discharging is completed from the battery data, obtaining first state data for predicting the state of the battery by applying the first battery data to a first deep learning model, obtaining second state data for predicting the state of the battery by applying the second battery data to a second deep learning model, and predicting the state of the battery based on the first state data and the second state data.

[0016] According to an embodiment, predicting the state of the battery based on the first state data and the second state data may include generating third state data by combining the first state data with the second state data, and predicting the state of the battery based on the third state data.

[0017] According to an embodiment, the operation of obtaining first status data for predicting a status of a battery by applying first battery data to a first deep learning model may include: extracting features of the first battery data by applying the first battery data to a first convolutional neural network (CNN) model, and the operation of obtaining second status data for predicting a status of a battery by applying second battery data to a second deep learning model may include: extracting features of the second battery data by applying the second battery data to a second CNN model.

[0018] According to an embodiment, the operation of predicting the state of a battery based on first state data and second state data may include: generating a third value based on a weighted sum of a first value and a second value, the first value being obtained by converting a feature of the first battery data into an embedding vector, the second value being obtained by converting a feature of the second battery data into an embedding vector; and predicting the state of the battery based on the third value.

[0019] According to an embodiment, the operation of predicting the state of the battery based on the first state data and the second state data may include predicting a state of health (SOH) of the battery. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 A battery pack according to embodiments disclosed herein is shown.

[0021] Figure 2 is a diagram for generally describing a battery state prediction apparatus according to an embodiment disclosed herein.

[0022] Figure 3 is a block diagram showing a configuration of a battery state prediction apparatus according to an embodiment disclosed herein.

[0023] Figure 4a is a graph showing voltage changes of battery data according to an embodiment disclosed herein.

[0024] Figure 4b is a graph showing voltage changes of first battery data according to an embodiment disclosed herein.

[0025] Figure 4c is a graph showing voltage changes of second battery data according to an embodiment disclosed herein.

[0026] Figure 5is a flowchart of an operating method of a battery state prediction device according to an embodiment disclosed herein.

[0027] Figure 6 is a block diagram showing a hardware configuration of a computing system for implementing a battery state prediction apparatus according to an embodiment disclosed herein. DETAILED DESCRIPTION

[0028] Hereinafter, some embodiments disclosed herein will be described in detail with reference to the exemplary drawings. When adding reference numerals to the components in each drawing, it should be noted that the same components are given the same reference numerals even if they are indicated on different drawings. In addition, when describing the embodiments disclosed herein, when it is determined that the detailed description of the related known configuration or function interferes with the understanding of the embodiments disclosed herein, its detailed description will be omitted.

[0029] In order to describe the components of the embodiments disclosed herein, terms such as first, second, A, B, (a), (b), etc. may be used. These terms are only used to distinguish one component from another component and do not limit the components to the nature, sequence, order, etc. of the components. In addition, the terms used herein, including technical terms and scientific terms, have the same meanings as those commonly understood by those skilled in the art, as long as these terms are not defined differently. In general, terms defined in general dictionaries should be interpreted as having the same meanings as the contextual meanings of the relevant technology, and should not be interpreted as having ideal or exaggerated meanings unless they are clearly defined herein.

[0030] Figure 1 A battery pack according to embodiments disclosed herein is shown.

[0031] Reference Figure 1 , a battery pack 1000 according to an embodiment disclosed herein may include a battery module 100 , a battery state prediction device 200 , and a relay 300 .

[0032] The battery module 100 may include a plurality of battery cells 110, 120, 130, and 140. Figure 1 Although shown as four, the present disclosure is not limited thereto, and the battery module 100 may include n battery cells (n is a natural number equal to or greater than 2).

[0033] The battery module 100 may supply power to the target device. To this end, the battery module 100 may be electrically connected to the target device. Herein, the target device may include an electrical, electronic or mechanical device that operates by receiving power from the battery pack 1000 including the plurality of battery cells 110, 120, 130 and 140, and the target device may be, for example, an electric vehicle (EV) or an energy storage system (ESS), but is not limited thereto.

[0034] The plurality of battery cells 110, 120, 130, and 140, each of which is a basic unit of a battery usable by charging and discharging electric energy, may be a lithium ion (Li-ion) battery, a lithium ion polymer battery, a nickel cadmium (Ni-Cd) battery, a nickel metal hydride (Ni-MH) battery, etc., and are not limited thereto. Figure 1 One battery module 100 is shown in FIG. 1 , but according to an embodiment, the battery module 100 may be configured in plurality.

[0035] The battery state prediction device 200 can predict the state of the battery cells based on the temperature, current and voltage data of the plurality of battery cells 110, 120, 130 and 140, which state includes the lifespan (state of health (SOH)) of the plurality of battery cells 110, 120, 130 and 140. The battery state prediction device 200 can predict the state of the plurality of battery cells 110, 120, 130 and 140 for each of the temperature, current and voltage of the battery based on the battery data of the plurality of battery cells 110, 120, 130 and 140.

[0036] According to an embodiment, the battery state prediction device 200 may be implemented in the form of a battery management system (BMS). In addition, according to an embodiment, the battery state prediction device 200 may be installed on the battery management system.

[0037] Herein, the battery management system may manage and / or control the state and / or operation of the battery module 100. For example, the battery management system may 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 system may manage charging and / or discharging of the battery module 100.

[0038] In addition, the battery management system may 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. Sensors or various measurement modules for monitoring performed by the battery management system may be additionally installed in the battery module 100, a charge / discharge path, any position of the battery module 100, etc. The battery management system may calculate parameters indicating the state of the battery module 100 such as the state of charge (SOC), SOH, etc. based on measurement values ​​such as the monitored voltage, current, temperature, etc.

[0039] The battery management system may control the operation of the relay 300. For example, the battery management system may short-circuit the relay 300 to supply power to the target device. The battery management system may short-circuit the relay 300 when a charging device is connected to the battery pack 1000.

[0040] The battery management system may calculate a cell balancing time for each of the plurality of battery cells 110, 120, 130, and 140. Herein, the cell balancing time may be defined as a time required to balance the battery cells. For example, the battery management system may calculate the cell balancing time based on the SOC, battery capacity, and balancing efficiency of each of the plurality of battery cells 110, 120, 130, and 140.

[0041] Figure 2 is a diagram for generally describing a battery state prediction apparatus according to an embodiment disclosed herein.

[0042] Reference Figure 2 , the battery state prediction device 200 can extract a portion of the battery data A.

[0043] The battery state prediction device 200 can obtain battery data A of the plurality of battery cells 110, 120, 130, and 140. In order to identify the degradation performance, i.e., SOH, of the plurality of battery cells 110, 120, 130, and 140, the battery management system can obtain battery data A, which includes the measured values ​​of the battery from the voltage value at which the SOC of the battery is 0% to the voltage value at which the SOC of the battery is 100%. Therefore, the battery state prediction device 200 can obtain battery data A including the voltage, current, and temperature of the plurality of battery cells 110, 120, 130, and 140 cumulatively measured during the charge / discharge period.

[0044] The battery state prediction device 200 can establish a training data set for the deep learning model of the battery state prediction device 200 based on the battery data A. Deep learning, which is one of the types of machine learning algorithms, may refer to a technique for connecting artificial neural networks in a multi-layer form. Machine learning may refer to a technique for training a computer to predict a specific result. Typically, using the results of machine learning includes a process of preparing training data for training a machine and training the computer in a manner suitable for the problem, a process of validating the model using test data, and a process of predicting a result using a model that has been validated.

[0045] For machine learning, it is important that the training data well represents the features to be generalized by machine learning, and therefore the training data is generated using training data that is limitedly selected according to specific criteria. For low correlation between the features to be generalized by machine learning and the features of the training data, sampling noise may occur, and it is difficult to find the pattern in which the machine problem analysis model is embedded, which is irrelevant to the accuracy of the machine problem analysis model, increasing the error of the model, thereby reducing the reliability of the model. Therefore, machine learning technology requires time to invest in evaluating training data and processing data to select a training data set.

[0046] The battery state prediction device 200 can extract a portion of the battery data A and generate a training data set for training the first deep learning model 221 and the second deep learning model 222. Herein, the first deep learning model 221 and the second deep learning model 222 may refer to learning models capable of predicting the state of the battery (including the life (SOH) of the battery) based on the input battery data.

[0047] That is, the battery state prediction device 200 can generate a training data set for machine learning of the first deep learning model 221 and the second deep learning model 222 to reduce the error of the machine learning of the first deep learning model 221 and the second deep learning model 222 and improve the reliability of the machine learning. The battery state prediction device 200 can input the training data set into the first deep learning model 221 and the second deep learning model 222 to predict the states of the plurality of battery cells 110, 120, 130, and 140.

[0048] Figure 3 is a block diagram showing a configuration of a battery state prediction apparatus according to an embodiment disclosed herein.

[0049] In the following, reference is made to Figure 3 , the configuration and operation of the battery state prediction device 200 will be described in detail.

[0050] Reference Figure 3 , the battery state prediction device 200 may include a data management unit 210 and a controller 220 .

[0051] The data management unit 210 may collect the battery data A. For example, the battery data A may be defined as a value recording a change in charge from a discharge state of the plurality of battery cells 110 , 120 , 130 , and 140 to a fully charged state thereof or from a fully charged state to a discharge state.

[0052] The data management unit 210 may extract first battery data A1 obtained within a predetermined time after battery charging is completed and second battery data A2 obtained within a predetermined time after battery discharging is completed. Herein, the first battery data A1 and the second battery data A2 may include voltage, current, and temperature of the battery measured cumulatively.

[0053] Figure 4a is a graph showing voltage changes of battery data according to an embodiment disclosed herein.

[0054] Reference Figure 4a , the interval in which the voltage of the battery changes sharply can be extracted from the battery data A. The data management unit 210 can set the interval in which the voltage of the battery in the battery data A changes sharply during the idle period after the battery is charged as the first battery data A1. In addition, the data management unit 210 can set the interval in which the voltage of the battery in the battery data A changes sharply during the discharge period after the battery is idle as the second battery data A2.

[0055] Conventionally, the behavior of the voltage or current of the entire battery data A is used to train the deep learning model, but when the entire battery data A is used as training data, a very long analysis time is required. The data management unit 210 can extract data required for battery life prediction in which the voltage or current of the battery changes sharply in the battery data A as the first battery data A1 and the second battery data A2.

[0056] Figure 4b is a graph showing voltage changes of first battery data according to an embodiment disclosed herein.

[0057] Figure 4b The first battery data A1 shown in may include voltage change information of the battery during an idle period after the battery is charged. When the charge / discharge cycle is repeated, the capacity of the battery may continue to deteriorate as the resistance of the battery increases. Therefore, even after the battery is charged, the voltage or open circuit voltage (OCV) of the battery may continue to decrease. The data management unit 210 may extract the first battery data A1 corresponding to the voltage change during the idle period after the battery is charged to use it for training the deep learning model.

[0058] For example, the first battery data A1 may include data of the last 1 second of the charging period of the battery and data of 60 seconds of the idle period immediately after the battery is charged. That is, the first battery data A1 may include battery data A corresponding to a period of 61 seconds in total.

[0059] Figure 4c is a graph showing voltage changes of second battery data according to an embodiment disclosed herein.

[0060] Figure 4c The second battery data A2 shown may include voltage change information of the battery when entering a discharge period after the battery is idle. The battery may be affected by ohmic resistance, which is caused by the internal resistance of the battery when current is applied to the battery at the beginning of discharge after being idle. Therefore, the second battery data A2 including the voltage change information of the battery during the discharge period after the battery is idle may include a voltage drop (IR drop) phenomenon that occurs due to the internal resistance of the battery. In this article, the voltage drop phenomenon may refer to the potential difference between two points in the conductive phase when current flows. The data management unit 210 can extract the second battery data A2 corresponding to the voltage change during the discharge period after the battery is idle to use it for training a deep learning model.

[0061] For example, the second battery data A2 may include data for the last 1 second of an idle period after battery charging and data for 40 seconds of a discharge period immediately after battery idling. That is, the second battery data A2 may include battery data A corresponding to a total period of 41 seconds.

[0062] Return to reference Figure 3 , the data management unit 110 can collect the first battery data A1 and the second battery data A2 as a training data set to be input into the deep learning model. The data management unit 210 can manage the first battery data A1 and the second battery data A2 as respective time series data. In this article, time series data may refer to data arranged at specific time intervals over time.

[0063] The controller 220 may predict the state of the battery by applying the first battery data A1 and the second battery data A2 including temperature, current, and voltage changes of the plurality of battery cells 110 , 120 , 130 , and 140 to a deep learning model of a time series analysis structure.

[0064] The controller 220 may obtain first state data for predicting the state of the battery by applying the first battery data A1 to the first deep learning model 221. In addition, the controller 220 may obtain second state data for predicting the state of the battery by applying the second battery data A2 to the second deep learning model 222. That is, the controller 220 may obtain separate output data by inputting the first battery data A1 and the second battery data A2 into separate deep learning models, respectively.

[0065] Herein, the first state data and the second state data may include battery life (SOH) data calculated using the first battery data A1 as training data for the first deep learning model 221 and battery SOH data calculated using the second battery data A2 as training data for the second deep learning model 222 .

[0066] According to an embodiment, the first deep learning model 221 and the second deep learning model 222 may include, for example, a one-dimensional (1D) convolutional neural network (CNN) model. In this article, the CNN model may mainly extract features of matrix data or image data. In this article, 1D may refer to a kernel for convolution and a data sequence to be applied having a 1D shape. 1D CNN may analyze features of time series data or text.

[0067] According to an embodiment, the controller 220 may extract features of the first battery data A1 by applying the first battery data A1 to the first CNN model. The controller 220 may also extract features of the second battery data A2 by applying the second battery data A2 to the second CNN model.

[0068] The controller 220 may predict the state of the battery based on the first state data and the second state data. More specifically, the controller 220 may generate third state data by combining the first state data with the second state data. The controller 220 may predict the state of the battery based on the third state data. In this article, the third state data may include the life data of the battery.

[0069] According to an embodiment, the controller 220 can generate a third value by combining a first value obtained by converting a feature of the first battery data A1 into an embedding vector and a second value obtained by converting a feature of the second battery data A2 into an embedding vector. In this article, embedding can be defined as a process of converting a natural language used by humans into a digital array that can be understood by a machine. More specifically, the controller 220 can combine the first value and the second value by inputting them into a sequential function (CONCATENATE).

[0070] The controller 220 may generate a third value by inputting data generated by combining the first value with the second value into a single-layer perceptron model. In this article, the single-layer perceptron model may receive multiple signals as input and output one signal. The single-layer perceptron model may generate an output signal based on a weighted summation that assigns a weight value to each of the multiple input signals and sums the values ​​obtained by multiplying each signal by its unique weight value. In this article, the weighted summation may be defined as an average value obtained by reflecting a weight value corresponding to the importance or influence of the data value when obtaining the average value of the data.

[0071] That is, the controller 220 can generate the third value based on the weighted sum of the first value and the second value by inputting the first value and the second value into the single-layer perceptron model. The third value can be generated by multiplying the first value and the second value by the weight value and summing the multiplication results. In this article, the weight values ​​of the first value and the second value can be determined by testing using a test data set to maximize accuracy. Therefore, the controller 220 can obtain the third value, which is the predicted data of the state of the battery. In this article, the third value may include the life data of the battery.

[0072] The embodiments disclosed herein aim to provide a battery status prediction device and an operating method thereof, wherein a portion of the battery data is extracted as training data to train an artificial intelligence model for analyzing the battery status, thereby improving the performance of the artificial intelligence model.

[0073] As described above, the battery state prediction device according to the embodiment disclosed herein can extract a portion of the battery data from the training data set, and then use the portion of the battery data to train multiple deep learning models, thereby improving the accuracy of the deep learning models.

[0074] In addition, the battery state prediction device can simplify and lightweight training data, thereby reducing the calculation time of the deep learning model used to predict the state of the battery in real time, and reducing the time required to collect and manage training data.

[0075] Figure 5 is a flowchart of an operating method of a battery state prediction device according to an embodiment disclosed herein.

[0076] The battery state prediction device 200 can be used with reference to Figures 1 to 4c The described battery state prediction apparatus 200 is substantially the same and thus will be briefly described to avoid redundant description.

[0077] Reference Figure 5, the operating method of the battery state prediction device may include: operation S101, extracting first battery data A1 obtained during a predetermined time after battery charging is completed from the battery data A; operation S102, extracting second battery data A2 obtained during a predetermined time after battery discharging is completed from the battery data; operation S103, obtaining first state data for predicting the state of the battery by applying the first battery data A1 to the first deep learning model 221; operation S104, obtaining second state data for predicting the state of the battery by applying the second battery data A2 to the second deep learning model 222, and operation S105, predicting the state of the battery based on the first state data and the second state data.

[0078] Hereinafter, operations S101 to S105 will be described in detail.

[0079] In operation S101, the data management unit 210 may collect battery data A of the plurality of battery cells 110, 120, 130, and 140. For example, the battery data A may be defined as a value recording a change in the amount of electricity from a discharge state of the plurality of battery cells 110, 120, 130, and 140 to a fully charged state thereof or from a fully charged state to a discharged state. Thus, the data management unit 210 may obtain the battery data A including the voltage, current, and temperature of the plurality of battery cells 110, 120, 130, and 140 cumulatively measured during a charge / discharge period.

[0080] In operation S101, the data management unit 210 may extract first battery data A1 obtained during a predetermined time after battery charging is completed from the battery data A. In operation S101, the data management unit 210 may extract the first battery data A1 corresponding to a voltage change in an idle period after battery charging to use it for training a deep learning model.

[0081] Herein, the first battery data A1 may include the voltage, current, and temperature of the battery measured cumulatively. For example, the first battery data A1 may include data of the last 1 second of the charging period of the battery and data of 60 seconds of the idle period immediately after the battery is charged. That is, the first battery data A1 may include battery data A corresponding to a period of 61 seconds in total.

[0082] In operation S102, the data management unit 210 may extract second battery data A2 obtained during a predetermined time after the battery discharge is completed from the battery data A. In operation S102, the data management unit 210 may extract the second battery data A2 corresponding to the voltage change of the discharge period after the battery is idle to use it for training the deep learning model.

[0083] Herein, the second battery data A2 may include the voltage, current, and temperature of the battery measured cumulatively. For example, the second battery data A2 may include data of the last 1 second of the idle period after the battery is charged and data of 40 seconds of the discharge period immediately after the battery is idle. That is, the second battery data A2 may include battery data A corresponding to a total period of 41 seconds.

[0084] In operation S103 , the controller 220 may obtain first state data for predicting the state of the battery by applying the first battery data A1 to the first deep learning model 221 .

[0085] In operation S103 , more specifically, the controller 220 may predict the state of the battery by applying the first battery data A1 including temperature, current, and voltage changes of the plurality of battery cells 110 , 120 , 130 , and 140 to a deep learning model of a time series analysis structure.

[0086] Herein, the first state data may include the life (SOH) data of the battery calculated using the first battery data A1 as the training data of the first deep learning model 221. In addition, in this article, the first deep learning model 221 may include, for example, a 1D CNN model. The CNN model may mainly extract features of matrix data or image data. In addition, in this article, 1D may refer to that the kernel used for convolution and the data sequence to be applied have a 1D shape. 1D CNN can analyze the features of time series data or text.

[0087] In operation S103, the controller 220 may also extract features of the first battery data A1 by applying the first battery data A1 to the first CNN model.

[0088] In operation S104 , the controller 220 may obtain second state data for predicting the state of the battery by applying the second battery data A2 to the second deep learning model 222 .

[0089] In operation S104, more specifically, the controller 220 may predict the state of the battery by applying the second battery data A2 including temperature, current, and voltage changes of the plurality of battery cells 110, 120, 130, and 140 to a deep learning model of a time series analysis structure. Herein, the second state data may include life span (SOH) data of the battery calculated using the second battery data A2 as training data of the second deep learning model 222. Herein, the second deep learning model 222 may include, for example, a 1D CNN model.

[0090] In operation S104, the controller 220 may extract features of the second battery data A2 by applying the second battery data A2 to the second CNN model.

[0091] In operation S105, the controller 220 may predict the state of the battery based on the first state data and the second state data. In operation S105, more specifically, the controller 220 may generate third state data by combining the first state data with the second state data. In operation S105, the controller 220 may predict the state of the battery based on the third state data. In this article, the third state data may include life data of the battery.

[0092] In operation S105, according to an embodiment, the controller 220 may generate a third value by combining a first value obtained by converting a feature of the first battery data A1 into an embedding vector and a second value obtained by converting a feature of the second battery data A2 into an embedding vector. In this article, embedding may be defined as a process of converting a natural language used by humans into a digital array that can be understood by a machine.

[0093] In operation S105, more specifically, the controller 220 may combine the first value and the second value by inputting them into a sequential function. In operation S105, the controller 220 may generate a third value by inputting data generated by combining the first value and the second value into a single-layer perceptron model. In this article, the single-layer perceptron model may receive multiple signals as input and output one signal. The single-layer perceptron model may generate an output signal based on a weighted summation that assigns a weight value to each of the multiple input signals and sums the values ​​obtained by multiplying each signal by its unique weight value.

[0094] In operation S105, the controller 220 may generate a third value based on a weighted sum of the first value and the second value by inputting the first value and the second value into a single layer perceptron model. The third value may be generated by multiplying the first value and the second value by a weight value and summing the multiplication results.

[0095] In operation S105, the controller 220 may obtain a third value, which is prediction data of the state of the battery. Herein, the third value may include life data of the battery.

[0096] Figure 6 is a block diagram showing a hardware configuration of a computing system for implementing a battery state prediction apparatus according to an embodiment disclosed herein.

[0097] Reference Figure 6 , the computing system 2000 according to the embodiment disclosed herein may include an MCU 2100 , a memory 2200 , an input / output I / F 2300 , and a communication I / F 2400 .

[0098] The MCU 2100 may be a processor that executes various programs (eg, a battery state prediction program, etc.) stored in the memory 2200, processes various data through these programs, and performs Figure 1 The above functions of the battery management device 200 are shown.

[0099] The memory 2200 may store various programs regarding the operation of the battery state prediction device 200. In addition, the memory 2200 may store operation data of the battery state prediction device 200.

[0100] Such a memory 2200 may be provided in plurality as required. The memory 2200 may be a volatile memory or a non-volatile memory. For the memory 2200 as a volatile memory, a random access memory (RAM), a dynamic RAM (DRAM), a static RAM (SRAM), etc. may be used. For the memory 2200 as a non-volatile memory, a read-only memory (ROM), a programmable ROM (PROM), an electrically variable ROM (EAROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), a flash memory, etc. may be used. The examples of the memory 2200 listed above are only examples and are not limited thereto.

[0101] The input / output I / F 2300 may provide an interface for transmitting and receiving data by connecting an input device such as a keyboard, a mouse, a touch panel, etc. and an output device such as a display, etc. to the MCU 2100 .

[0102] The communication I / F 2400, which is a component capable of sending and receiving various data to and from a server, may be various devices capable of supporting wired or wireless communication. For example, a program or various data for resistance measurement and abnormality diagnosis of a battery cell may be sent to and received from a separately provided external server through the communication I / F 2400.

[0103] The above description is only used to illustrate the technical idea of ​​the present disclosure, and those skilled in the art to which the present disclosure belongs may make various modifications and variations without departing from the basic features of the present disclosure.

[0104] Therefore, the embodiments disclosed in the present disclosure are intended to describe rather than limit the technical spirit of the present disclosure, and the scope of the technical spirit of the present disclosure is not limited by these embodiments. The protection scope of the present disclosure should be interpreted by the attached claims, and all technical spirits within the same scope should be understood to be included in the scope of the present disclosure.

[0105] [Description of Reference Signs]

[0106] 1000: Battery Pack

[0107] 100: Battery module

[0108] 200: Battery status prediction device

[0109] 210: Data Management Unit

[0110] 220: Controller

[0111] 221: The first deep learning model

[0112] 222: Second Deep Learning Model

[0113] 300: Relay

[0114] A: Battery data

[0115] A1: First battery data

[0116] A2: Second battery data

[0117] 2000: Computing Systems

[0118] 2100: MCU

[0119] 2200: Memory

[0120] 2300: Input / Output I / F

[0121] 2400: Communication I / F

Claims

1. A battery status prediction device, comprising: a data management unit configured to extract first battery data obtained within a predetermined time after completion of battery charging and second battery data obtained within the predetermined time after completion of battery discharging; as well as A controller configured to obtain first state data for predicting a state of the battery by applying the first battery data to a first deep learning model, obtain second state data for predicting a state of the battery by applying the second battery data to a second deep learning model, and predict 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 is further configured to generate third status data by combining the first status data with the second status data, and predict a status of the battery based on the third status data.

3. The battery state prediction device according to claim 2, wherein: The controller is also configured to extract features of the first battery data by applying the first battery data to a first convolutional neural network (CNN) model, and to extract features of the second battery data by applying the second battery data to a second CNN model.

4. The battery state prediction device according to claim 3, wherein: The controller is also configured to: generate a third value based on a weighted sum of a first value and a second value, the first value being obtained by converting a feature of the first battery data into an embedding vector, and the second value being obtained by converting a feature of the second battery data into an embedding vector; and predict the state of the battery based on the third value.

5. The battery state prediction device according to claim 1, wherein: The first battery data and the second battery data include cumulatively measured voltage, current, and temperature of the battery, and the first state data, the second state data, and the third state data include a state of health (SOH) of the battery calculated based on the first battery data and the second battery data.

6. A method for operating a battery status prediction device, the method comprising the following operations: extracting, from the battery data, first battery data obtained within a predetermined time after battery charging is completed; extracting, from the battery data, second battery data obtained within the predetermined time after the battery discharge is completed; Obtaining first state data for predicting a state of the battery by applying the first battery data to a first deep learning model; Obtaining second state data for predicting a state of the battery by applying the second battery data to a second deep learning model; as well as The state of the battery is predicted based on the first state data and the second state data.

7. The method for operating the battery state prediction device according to claim 6, wherein: The operation of predicting the state of the battery based on the first state data and the second state data includes generating third state data by combining the first state data with the second state data, and predicting the state of the battery based on the third state data.

8. The method for operating the battery state prediction device according to claim 7, wherein: The operation of obtaining first state data for predicting the state of the battery by applying the first battery data to the first deep learning model includes: extracting features of the first battery data by applying the first battery data to a first convolutional neural network (CNN) model, and The operation of obtaining second state data for predicting the state of the battery by applying the second battery data to the second deep learning model includes: extracting features of the second battery data by applying the second battery data to a second CNN model.

9. The method for operating the battery state prediction device according to claim 8, wherein: The operation of predicting the state of the battery based on the first state data and the second state data includes: generating a third value based on a weighted sum of a first value and a second value, the first value being obtained by converting a feature of the first battery data into an embedding vector, and the second value being obtained by converting a feature of the second battery data into an embedding vector; and predicting the state of the battery based on the third value.

10. The method for operating the battery state prediction device according to claim 6, wherein: The operation of predicting the state of the battery based on the first state data and the second state data includes predicting a state of health (SOH) of the battery.

Images