WO2024136347A1 - 배터리 상태 예측 장치 및 그것의 동작 방법 - Google Patents
배터리 상태 예측 장치 및 그것의 동작 방법 Download PDFInfo
- Publication number
- WO2024136347A1 WO2024136347A1 PCT/KR2023/020817 KR2023020817W WO2024136347A1 WO 2024136347 A1 WO2024136347 A1 WO 2024136347A1 KR 2023020817 W KR2023020817 W KR 2023020817W WO 2024136347 A1 WO2024136347 A1 WO 2024136347A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- battery
- data
- machine learning
- gas generation
- prediction
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
- G01R31/367—Software therefor, e.g. for battery testing using modelling or look-up tables
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/20—Ensemble learning
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
- G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/047—Probabilistic or stochastic networks
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/09—Supervised learning
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N7/00—Computing arrangements based on specific mathematical models
- G06N7/01—Probabilistic graphical models, e.g. probabilistic networks
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01M—PROCESSES OR MEANS, e.g. BATTERIES, FOR THE DIRECT CONVERSION OF CHEMICAL ENERGY INTO ELECTRICAL ENERGY
- H01M10/00—Secondary cells; Manufacture thereof
- H01M10/42—Methods or arrangements for servicing or maintenance of secondary cells or secondary half-cells
- H01M10/425—Structural combination with electronic components, e.g. electronic circuits integrated to the outside of the casing
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01M—PROCESSES OR MEANS, e.g. BATTERIES, FOR THE DIRECT CONVERSION OF CHEMICAL ENERGY INTO ELECTRICAL ENERGY
- H01M10/00—Secondary cells; Manufacture thereof
- H01M10/42—Methods or arrangements for servicing or maintenance of secondary cells or secondary half-cells
- H01M10/44—Methods for charging or discharging
- H01M10/441—Methods for charging or discharging for several batteries or cells simultaneously or sequentially
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01M—PROCESSES OR MEANS, e.g. BATTERIES, FOR THE DIRECT CONVERSION OF CHEMICAL ENERGY INTO ELECTRICAL ENERGY
- H01M10/00—Secondary cells; Manufacture thereof
- H01M10/42—Methods or arrangements for servicing or maintenance of secondary cells or secondary half-cells
- H01M10/48—Accumulators combined with arrangements for measuring, testing or indicating the condition of cells, e.g. the level or density of the electrolyte
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01M—PROCESSES OR MEANS, e.g. BATTERIES, FOR THE DIRECT CONVERSION OF CHEMICAL ENERGY INTO ELECTRICAL ENERGY
- H01M10/00—Secondary cells; Manufacture thereof
- H01M10/42—Methods or arrangements for servicing or maintenance of secondary cells or secondary half-cells
- H01M10/425—Structural combination with electronic components, e.g. electronic circuits integrated to the outside of the casing
- H01M2010/4271—Battery management systems including electronic circuits, e.g. control of current or voltage to keep battery in healthy state, cell balancing
Definitions
- Embodiments disclosed in this document relate to a battery state prediction device and a method of operating the same.
- Electric vehicles receive electricity from outside, charge the battery, and then use the voltage charged in the battery to drive the motor to obtain power.
- the battery of an electric vehicle may generate heat due to a chemical reaction that occurs in the process of charging and discharging electricity, and this heat may damage the performance and lifespan of the battery and cause gas to be generated inside the battery.
- BMS Battery Management System
- the battery management device can predict the amount of gas generated by the battery by training battery data into an artificial intelligence model that analyzes the state of the battery.
- an artificial intelligence model that analyzes the state of the battery.
- the battery management device inputs the same battery data as input to the artificial intelligence model, it can always derive only the same output data (Output), which causes the actual battery to generate different distributions of gas generation in the same environment. There is a problem that it cannot reflect the actual distribution.
- One purpose of the embodiments disclosed in this document is to provide a battery status prediction device and a method of operating the same that can obtain gas generation prediction data in the form of a probability distribution using a plurality of artificial intelligence models for predicting the battery gas generation amount. There is.
- a battery state prediction device includes a generator that generates a plurality of machine learning models that are learned based on battery data and features included in the battery data to predict the amount of gas generation of the battery; It may include a controller that applies the battery data to the plurality of machine learning models to obtain a plurality of prediction data predicting the gas generation amount of the battery, and predicts the gas generation amount of the battery based on the plurality of prediction data.
- the plurality of machine learning models include a plurality of deep neural network (DNN) models
- the controller applies the battery data to the plurality of deep neural network models to generate the plurality of prediction data. can be obtained.
- DNN deep neural network
- the battery data includes cumulatively measured temperature, SOC, and SOH of the battery, and the characteristics of the battery data include at least one of an electrode type, assembly process, and separator type of the battery. can do.
- the controller weights each of the plurality of prediction data based on a weight assigned based on characteristics of the battery data in the process of training the deep neural network model used to generate each of the plurality of prediction data. and inputting the plurality of weighted prediction data into an ensemble learning model to generate the gas generation prediction data in the form of a probability distribution.
- the controller determines the accuracy of the plurality of machine learning models by calculating an average value of the gas generation prediction data in the form of the probability distribution and comparing the average value with a previously stored gas generation measurement value of the battery. can do.
- a method of operating a battery state prediction device is learned based on battery data and features included in the battery data to generate a plurality of machine learning models that predict the amount of gas generation of the battery.
- obtaining a plurality of prediction data predicting the gas generation amount of the battery by applying the battery data to the plurality of machine learning models, and predicting the gas generation amount of the battery based on the plurality of prediction data. may include.
- the plurality of machine learning models include a plurality of deep neural network (DNN) models
- the battery data is applied to the plurality of machine learning models to predict the amount of gas generation of the battery.
- Obtaining the plurality of prediction data may include applying the battery data to the plurality of deep neural network models to obtain the plurality of prediction data.
- the step of generating a plurality of machine learning models that predict the amount of gas generation of the battery based on the battery data is based on the battery data including the cumulatively measured temperature, SOC, and SOH of the battery. and generating the plurality of machine learning models, and the characteristics of the battery data may include at least one of an electrode type, an assembly process, and a separator type of the battery.
- the step of predicting the gas generation amount of the battery based on the plurality of prediction data is based on the characteristics of the battery data in the process of training the deep neural network model that generates each of the plurality of prediction data. assigning weights to each of the plurality of prediction data based on the assigned weights, and inputting the plurality of prediction data to which the weights have been assigned into an ensemble learning model to generate the gas generation prediction data in the form of a probability distribution. It may include a creation step.
- the step of predicting the gas generation amount of the battery based on the plurality of prediction data includes calculating an average value of the gas generation amount prediction data in the form of the probability distribution, and dividing the average value into a pre-stored gas generation amount of the battery. By comparing the measured values, the accuracy of the plurality of machine learning models can be determined.
- the battery state prediction device and its operating method according to an embodiment disclosed in this document can obtain gas generation prediction data in the form of a probability distribution using a plurality of artificial intelligence models.
- FIG. 1 is a diagram showing a battery pack according to an embodiment disclosed in this document.
- FIG. 2 is a diagram for generally explaining a battery state prediction device according to an embodiment disclosed in this document.
- FIG. 3 is a block diagram showing the configuration of a battery state prediction device according to an embodiment disclosed in this document.
- Figure 4 is a diagram for explaining the operation of the generator according to an embodiment disclosed in this document.
- Figure 5 is a diagram for explaining a method of operating a controller according to an embodiment disclosed in this document.
- Figure 6 is a flowchart showing a method of operating a battery state prediction device according to an embodiment disclosed in this document.
- Figure 7 is a block diagram showing the hardware configuration of a computing system that implements a battery state prediction device according to an embodiment disclosed in this document.
- FIG. 1 is a diagram showing a battery pack according to an embodiment disclosed in this document.
- a battery pack 1000 may include a battery module 100, a battery state prediction device 200, and a relay 300.
- the battery module 100 may include a plurality of battery cells 110, 120, 130, and 140. Although the plurality of battery cells is shown in FIG. 1 as four, the battery module 100 is not limited thereto, and the battery module 100 may be configured to include n (n is a natural number of 2 or more) battery cells.
- the battery module 100 may supply power to a target device (not shown). To this end, the battery module 100 may be electrically connected to the target device.
- the target device may include an electrical, electronic, or mechanical device that operates by receiving power from the battery pack 1000 including a plurality of battery cells 110, 120, 130, and 140, for example.
- the target device may be an electric vehicle (EV) or an energy storage system (ESS), but is not limited thereto.
- EV electric vehicle
- ESS energy storage system
- a plurality of battery cells are the basic units of a battery that can be used by charging and discharging electrical energy, and include a lithium-ion (Li-ion) battery, a lithium-ion polymer (Li-ion polymer) battery, and a nickel battery. It may be a cadmium (Ni-Cd) battery, a nickel hydride (Ni-MH) battery, etc., but is not limited thereto. Meanwhile, in FIG. 1, there is shown a single battery module 100, but depending on the embodiment, the battery module 100 may be comprised of a plurality of battery modules 100.
- the battery state prediction device 200 predicts a plurality of battery cells (110, 120, 130, 140) based on temperature, current, voltage, SOC (State of Charge), and SOH (State of Health) data 110, 120, 130, 140) of gas generation can be predicted.
- the battery state prediction device 200 may predict the amount of gas generation of the plurality of battery cells (110, 120, 130, and 140) based on the battery data (A) of the plurality of battery cells (110, 120, 130, and 140). .
- the battery state prediction device 200 may be implemented in the form of a battery management system (BMS). Additionally, depending on the embodiment, the battery state prediction device 200 may be mounted on a battery management device.
- BMS battery management system
- the battery management device may manage and/or control the status and/or operation of the battery module 100.
- the battery management device may manage and/or control the status and/or operation of the plurality of battery cells 110, 120, 130, and 140 included in the battery module 100.
- the battery management device may manage charging and/or discharging of the battery module 100.
- the battery management device 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. Additionally, for monitoring by the battery management device, sensors or various measurement modules, not shown, may be additionally installed in the battery module 100, the charging/discharging path, or any other location in the battery module 100. The battery management device may calculate parameters indicating the state of the battery module 100, such as SOC or SOH, based on monitored measured values such as voltage, current, and temperature.
- the battery management device can control the operation of the relay 300. For example, the battery management device may short-circuit the relay 300 to supply power to the target device. Additionally, the battery management device may short-circuit the relay 300 when a charging device is connected to the battery pack 1000.
- the battery management device may calculate the cell balancing time of each of the plurality of battery cells 110, 120, 130, and 140.
- the cell balancing time may be defined as the time required to balance battery cells.
- the battery management device 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.
- FIG. 2 is a diagram for generally explaining a battery state prediction device according to an embodiment disclosed in this document.
- the battery state prediction device 200 may extract part of the battery data A and input it into a plurality of machine learning models 211, 212, 213, and 214.
- a plurality of machine learning models are shown as four, but this is not limited to this, and the plurality of machine learning models (211, 212, 213, 214) include n (n is a natural number of 2 or more) machine learning models. It can be configured as follows.
- the battery state prediction device 200 may acquire battery data A of a plurality of battery cells 110, 120, 130, and 140.
- the battery management device operates the battery from a voltage value at which the SOC of the battery is 0% until it reaches a voltage value at which the SOC of the battery is 100%.
- Battery data (A) including measured values can be obtained.
- the battery state prediction device 200 provides battery data (A) including the voltage, current, temperature, SOC, and SOH of the plurality of battery cells 110, 120, 130, and 140 cumulatively measured during the charging and discharging period. can be obtained,
- the battery data (A) may further include at least one characteristic among the features of the battery data (A) including the separator type, assembly process, and electrode type of the battery. That is, the battery data (A) may further include at least one of data related to what type of separator each battery has, what assembly process it was assembled through, and/or what electrode type it has. .
- the battery state prediction device 200 can input battery data (A) into a plurality of machine learning models (211, 212, 213, and 214) to predict the amount of gas generation of the battery.
- machine learning is a technology that trains computers to predict certain results.
- the results using machine learning include the process of preparing training data to train the machine and training it in a way appropriate for the problem, the process of validating the model with test data, It includes the process of predicting results using a model that has passed verification.
- Machine learning is created using training data that is limitedly selected according to certain criteria because it is important that the training data well represent the characteristics that are intended to be generalized through machine learning. If the correlation between the characteristics to be generalized through machine learning and the characteristics of the learning data is low, sampling noise occurs and it is difficult for the machine problem analysis model to find inherent patterns, which is independent of the accuracy of the machine problem analysis model itself. There is a problem that the model's reliability decreases as the model's error increases. Therefore, machine learning technology must invest time in evaluating and processing learning data to select a learning data set.
- the battery state prediction device 200 may generate a learning data set by extracting at least a portion of the battery data A in order to generate and learn a plurality of machine learning models 211, 212, 213, and 214.
- the plurality of machine learning models 211, 212, 213, and 214 may refer to learning models that can predict the state of the battery, including the amount of gas generated by the battery, based on input battery data.
- the battery state prediction device 200 generates a plurality of machine learning models (211, 212, 213, 214) of the same structure based on one learning data set generated by extracting at least a portion of the battery data (A). ) can be created.
- the battery state prediction device 200 is a plurality of machines through a plurality of machine learning models (211, 212, 213, 214) generated based on one learning data set generated by extracting at least part of the battery data (A). Prediction data predicting the amount of battery gas generation for each of the learning models 211, 212, 213, and 214 can be obtained.
- the battery state prediction device 200 combines the battery gas generation prediction data of each of the plurality of machine learning models 211, 212, 213, and 214 to finally obtain one gas generation prediction data (C) in the form of a probability distribution. there is. Therefore, the battery state prediction device 200 obtains output data (Output) of each of a plurality of machine learning models (211, 212, 213, 214) of the same structure based on one input data (Input), and each By combining the output data, one final gas generation prediction data (C) in the form of a probability distribution can be obtained to reduce learning errors and increase reliability.
- Output data generated by artificial intelligence models may include random errors and main effects.
- an artificial intelligence model repeats an experiment multiple times based on the same input data, it can obtain different output data, such as random error or white noise.
- the artificial intelligence model repeatedly runs the same experiment sufficiently many times, the average value of the random error converges to 0, and the artificial intelligence model can only obtain main effect data.
- the battery state prediction device 200 uses the same input data to enter a plurality of machine learning models 211, 212, 213, and 214 to obtain a plurality of output data at the same time, thereby setting up the same experimental environment and performing repeated experiments.
- the effect can be obtained. That is, the battery state prediction device 200 inputs at least some of the same battery data (A) into a plurality of machine learning models (211, 212, 213, 214) of the same structure to generate a battery with high accuracy corresponding to the main effect data.
- the amount of gas generated can be predicted.
- FIG. 3 is a block diagram showing the configuration of a battery state prediction device according to an embodiment disclosed in this document
- FIG. 4 is a diagram for explaining the operation of the generator according to an embodiment disclosed in this document.
- the battery state prediction device 200 may include a generator 210 and a controller 220.
- the generator 210 may collect battery data (A).
- the battery data (A) is a value that records the change in state of the battery from the discharged state to the fully charged state or from the fully charged state to the discharged state of the plurality of battery cells 110, 120, 130, and 140. can be defined.
- battery data (A) may include cumulatively measured voltage, current, temperature, SOC, and SOH of the battery.
- SOH may include the capacity degradation and resistance degradation of the battery.
- the battery data A may further include at least one piece of data related to what type of separator the battery is, what assembly process it was assembled through, and/or what electrode type it has.
- the generator 210 may generate a plurality of machine learning models 211, 212, 213, and 214 by extracting at least a portion of the battery data A as a learning data set. For example, the generator 210 may extract 80% of the battery data A as a learning data set and generate a plurality of machine learning models 211, 212, 213, and 214.
- the plurality of machine learning models 211, 212, 213, and 214 may include a plurality of deep neural network (DNN) models.
- DNN deep neural network
- a deep neural network model is an artificial neural network technology that includes multiple hidden layers between the input layer and the output layer. Deep neural network models can learn a variety of complex nonlinear relationships, including multiple hidden layers.
- the generator 210 may use at least some of the battery data A as a learning data set to generate a plurality of machine learning models 211, 212, 213, and 214 that can predict the amount of gas generated by the battery.
- the amount of internal gas generated by the battery may increase, and depending on what type of separator the battery has, what assembly process the battery was assembled through, and/or what type of electrode the battery has.
- the amount of internal gas generation may vary. Therefore, a plurality of machine learning models (211, 212, 213, 214) calculate the gas inside the battery based on the voltage, current, temperature, SOC, SOH, assembly process, electrode type, and separator type of the battery included in the battery data (A). It is possible to predict whether gas will be generated and the amount of gas generated.
- the generator 210 generates a plurality of machine learning models 211, 212, 213, and 214 learned based on at least one feature of the battery data A. You can. For example, when considering the separator type and electrode type among the features of the battery data (A), the generator 210 generates the separator type and electrode type of the battery in addition to the voltage, current, temperature, SOC, and SOH of the battery. Battery data (A) including can be extracted as a learning data set to generate a plurality of machine learning models (211, 212, 213, 214).
- the generator 210 assigns various weights to the characteristics of the battery data (A), including the separator type, assembly process, and electrode type of the battery, among various characteristics included in the battery data (A), to generate the battery data (A).
- the generator 210 can be trained on a plurality of machine learning models (211, 212, 213, 214).
- the generator 210 minimizes the intervention of relatively unimportant data among the battery separator type, assembly process, and electrode type based on the drop-out technique (a), and performs multiple machine learning.
- a plurality of machine learning models (211, 212, 213, 214) can be created by using a regularization technique to resolve overfitting of the models (211, 212, 213, and 214). That is, the generator 210 trains the battery data (A) with the connections of nodes related to characteristics that do not want intervention or minimize intervention among the separator type, assembly process, and electrode type of the battery to create a plurality of machine learning models. (211, 212, 213, 214) can be generated.
- the generator 210 generates a weight assigned to the characteristics of the battery data (A), including the separator type, assembly process, and electrode type of the battery, among various characteristics included in the battery data (A). ) can be fixed (b) to generate a plurality of machine learning models (211, 212, 213, 214) trained on the battery data (A). For example, if the generator 210 wants to generate a machine learning model that is greatly influenced by the separator type of the battery, it assigns a high weight (e.g., *a) to the separator type among the characteristics of the battery data (A). By assigning it, the battery data (A) can be learned. In addition, the generator 210 assigns a low weight (e.g., *b) to the electrode type among the characteristics of the battery data (A) to generate a machine learning model that is less affected by the electrode type of the battery. (A) can be learned.
- a high weight e.g., *a
- the generator 210 assigns a low weight (e.g.,
- the generator 210 generates a bias on a node connected to the battery data (A) including the battery separator type, assembly process, and electrode type among various characteristics included in the battery data (A).
- +a', +b' can be assigned (c) to generate a plurality of machine learning models (211, 212, 213, 214) trained on the battery data (A).
- the generator 210 generates a plurality of machine learning models (211, 212, 213, 214) that learn the battery data (A) based on weight regulations including L1 regulation (Lasso) and L2 regulation (Ridge). can be generated, and an embodiment disclosed in this document is not limited to this example.
- the generator 210 may perform Min-Max Scaling on at least a portion of the battery data A.
- maximum-minimum scaling is a method of adjusting the range of all variables because if the size or unit of a numeric variable is different for each variable, the impact on the dependent variable is not properly reflected.
- the generator 210 may convert at least a portion of the battery data A into a value between 0 and 1 by maximally scaling at least a portion of the battery data A to the minimum.
- the generation unit 210 performs K-Fold Cross Validation on the plurality of machine learning models (211, 212, 213, 214) to improve the accuracy of the plurality of machine learning models (211, 212, 213, 214). (Accuracy) can be judged.
- K-fold cross-validation divides the preprocessed data set into a training data set and a test set, divides the training data set into 'K' folds, uses one fold for verification, and uses (K-1) folds. is a method used for model training so that all data can be used in the training and verification process.
- the generation unit 210 performs 5-fold cross validation on a plurality of machine learning models (211, 212, 213, 214) to generate a plurality of machine learning models (211, 212, 213, 214) can be judged on accuracy.
- the generator 210 generates a plurality of machine learning models (211, 212, 213, 214) can be evaluated.
- the controller 220 uses at least some of the battery data (A) including voltage, current, temperature, SOC, and SOH changes of the plurality of battery cells 110, 120, 130, and 140 as a test data set to perform a plurality of machine learning tests. Can be input into models (211, 212, 213, 214).
- the controller 220 generates a plurality of machine learning models (211, 212, 213, 214) by extracting 80% of the battery data (A) as a learning data set, and then generates the battery The remaining 20% of the data (A) can be extracted as a test data set and input into multiple machine learning models (211, 212, 213, 214).
- Figure 5 is a diagram for explaining a method of operating a controller according to an embodiment disclosed in this document.
- the controller 220 applies at least a portion of the battery data (A) to a plurality of machine learning models (211, 212, 213, 214) to create a plurality of machine learning models (211, 212, 213, 214).
- a plurality of prediction data (B1, B2, B3, B4) predicting the gas generation amount of the battery can be obtained from each. That is, the controller 220 may obtain a plurality of output data (Output) by inputting at least some of the battery data (A), which is one input data (Input), into separate machine learning models.
- the controller 220 may generate gas generation prediction data C that predicts the gas generation amount of the battery based on a plurality of prediction data B1, B2, B3, and B4.
- the controller 220 may predict the state of the battery based on a plurality of prediction data (B1, B2, B3, B4). Depending on the embodiment, the controller 220 creates a plurality of machine learning models 211, 212, 213, and 214 based on at least one of the battery separator type, assembly process, and electrode type among the characteristics of the battery data A. Weights (x1, x2, x3, x4) can be assigned to each of the generated plurality of prediction data (B1, B2, B3, B4).
- the size of the first prediction data (B1) is amplified to 1
- the proportion of predicted data (B1) in the total predicted data (B1, B2, B3, B4) can be increased.
- the weights (x1, x2, x3, x4) assigned to each of the plurality of prediction data (B1, B2, B3, and B4) are used by the battery when training each of the plurality of machine learning models (211, 212, 213, 214). It may be related to the weight given to the characteristics of data (A).
- the controller 220 makes a plurality of predictions according to weights assigned based on the characteristics of the battery data (A). Weights can be assigned to data (B1, B2, B3, B4). That is, when training the plurality of machine learning models 211, 212, 213, and 214, the weight given to the characteristics of the battery data A and the plurality of prediction data B1, B2, B3, and B4, respectively. Weights are distinct. For example, the controller 220 trains the battery data ( The weight given based on the characteristics of A) can be considered.
- the controller 220 uses the battery data (A) when training a deep neural network model included in the machine learning model 212 that generates the prediction data (B2) to determine the weight to be given to the prediction data (B2). Weights given based on characteristics can be considered, and when training a deep neural network model included in the machine learning model 213 that generates prediction data B3, weights given based on the characteristics of the battery data A can be used. When training a deep neural network model included in the machine learning model 214 that generates the prediction data B4, weights assigned based on the characteristics of the battery data A can be considered.
- the weights given to the plurality of prediction data are based on the characteristics of the battery data (A) in the process of training each machine learning model (211, 212, 213, 214). It may be proportional to the weight assigned as a basis. For example, if characteristics related to the separator type and assembly process are considered among the characteristics of the battery data (A) in the process of learning the machine learning model 211 that generates the prediction data (B1), the controller 220 A weight corresponding to the separator type and a weight corresponding to the assembly process type can be assigned to the predicted data (B1), and the weight corresponding to the separator type and the weight corresponding to the assembly process are used in the process of learning the machine learning model 211.
- a weight may be assigned that is proportional to the weight given to the characteristics of the battery data related to the separator type and the weight given to the characteristics of the battery data related to the assembly process.
- the weight (x1) may be the sum of the weight corresponding to the separator type and the weight corresponding to the assembly process type, but is not limited to this example.
- the controller 220 simply assigns weights (x1, x2, x3, x4) to each of the plurality of prediction data (B1, B2, B3, B4).
- the final gas generation forecast data (C) can be calculated by applying a fuzzy algorithm that can reflect the importance and characteristics of specific variables.
- the controller 220 inputs a plurality of predicted data (B1, B2, B3, B4) with weights (x1, x2, x3, x4) into an ensemble learning model to form a probability distribution.
- Gas generation prediction data (C) can be generated.
- an ensemble learning model is a machine learning technique that combines two or more learning models to perform better than a single learning model.
- the ensemble learning model can calculate a weighted sum by assigning weight to the output data of each model, rather than the average of the output data of each model.
- the weighted sum can be defined as the average value obtained by reflecting the weight corresponding to the importance or influence of the data value when calculating the average of the data.
- the controller 220 may calculate the average value (Mean) of the gas generation prediction data (C) in the form of a probability distribution. Specifically, the controller 220 may calculate a 95% prediction interval (Prediction Interval) of the mean value (Mean) of the gas generation amount prediction data (C) in the form of a probability distribution. For example, the controller 220 calculates the mean value (Mean) and standard deviation of the gas generation prediction data (C) in the form of a probability distribution, and then calculates 'mean value ⁇ standard deviation * 1.96' as the gas generation prediction data in the form of a probability distribution. It can be calculated as the 95% prediction interval of the average value in (C).
- Prediction Interval 95% prediction interval of the mean value (Mean) of the gas generation amount prediction data (C) in the form of a probability distribution.
- the controller 220 calculates the mean value (Mean) and standard deviation of the gas generation prediction data (C) in the form of a probability distribution, and then calculates 'mean value
- the controller 220 determines the accuracy of the plurality of machine learning models (211, 212, 213, 214) by comparing the prediction section of the gas generation prediction data (C) in the form of a probability distribution with the previously stored gas generation measurement value of the battery. You can.
- the battery state prediction device can obtain gas generation prediction data in the form of a probability distribution using a plurality of artificial intelligence models that predict battery gas generation amount.
- a battery status prediction device can calculate final prediction data with high accuracy by inputting a small number of input data into a plurality of artificial intelligence models, and reduce the time and cost required for data collection and management.
- the battery state prediction device can set weights according to the characteristics of the data to reflect the unique characteristics of the actual battery data and the characteristics of the actual battery usage environment.
- Figure 6 is a flowchart showing a method of operating a battery state prediction device according to an embodiment disclosed in this document.
- the battery state prediction device 200 may be substantially the same as the battery state prediction device 200 described with reference to FIGS. 1 to 5, it will be briefly described below to avoid duplication of description.
- the method of operating the battery status prediction device includes generating a plurality of machine learning models that predict the amount of gas generation of the battery based on battery data (S101), and applying the battery data to the plurality of machine learning models to determine the battery It may include obtaining a plurality of prediction data predicting the gas generation amount of (S102) and predicting the gas generation amount of the battery based on the plurality of prediction data (S103).
- the generator 210 may collect battery data (A).
- the battery data (A) is a value that records the change in state of the battery from the discharged state to the fully charged state or from the fully charged state to the discharged state of the plurality of battery cells 110, 120, 130, and 140. can be defined.
- battery data (A) may include cumulatively measured voltage, current, temperature, SOC, and SOH of the battery.
- SOH may include the capacity degradation and resistance degradation of the battery.
- the battery data A may further include at least one piece of data related to what type of separator the battery is, what assembly process it was assembled through, and/or what electrode type it has.
- the generator 210 may generate a plurality of machine learning models 211, 212, 213, and 214 by extracting at least some of the battery data A as a learning data set.
- the generator 210 may extract 80% of the battery data A as a learning data set and generate a plurality of machine learning models 211, 212, 213, and 214.
- the generator 210 may generate a plurality of machine learning models 211, 212, 213, and 214 learned based on at least one feature of the battery data A. .
- the generator 210 when considering the separator type and electrode type among the features of the battery data (A), the generator 210 generates the separator type and electrode type of the battery in addition to the voltage, current, temperature, SOC, and SOH of the battery.
- Battery data (A) including can be extracted as a learning data set to generate a plurality of machine learning models (211, 212, 213, 214).
- the generator 210 assigns various weights to the characteristics of the battery data (A), including the separator type, assembly process, and electrode type of the battery, among various characteristics included in the battery data (A), to generate the battery data (A).
- the generator 210 can be trained on a plurality of machine learning models (211, 212, 213, 214).
- the generator 210 minimizes the intervention of relatively unimportant data among the battery separator type, assembly process, and electrode type based on the drop-out technique (a), and performs multiple machine learning.
- a plurality of machine learning models (211, 212, 213, 214) can be created using a regularization technique to resolve overfitting of the models (211, 212, 213, and 214). That is, the generator 210 trains the battery data (A) with the connections of nodes related to characteristics that do not want intervention or minimize intervention among the separator type, assembly process, and electrode type of the battery to create a plurality of machine learning models. (211, 212, 213, 214) can be generated.
- the generator 210 generates a weight assigned to the characteristics of the battery data (A), including the separator type, assembly process, and electrode type of the battery, among various characteristics included in the battery data (A). ) can be fixed (b) to generate a plurality of machine learning models (211, 212, 213, 214) trained on the battery data (A). For example, if the generator 210 wants to generate a machine learning model that is greatly influenced by the separator type of the battery, it assigns a high weight (e.g., *a) to the separator type among the characteristics of the battery data (A). By assigning it, the battery data (A) can be learned. In addition, the generator 210 assigns a low weight (e.g., *b) to the electrode type among the characteristics of the battery data (A) to generate a machine learning model that is less affected by the electrode type of the battery. (A) can be learned.
- a high weight e.g., *a
- the generator 210 assigns a low weight (e.g.,
- the generator 210 generates a bias on a node connected to the battery data (A) including the battery separator type, assembly process, and electrode type among various characteristics included in the battery data (A).
- +a', +b' can be assigned (c) to generate a plurality of machine learning models (211, 212, 213, 214) trained on the battery data (A).
- the generator 210 generates a plurality of machine learning models (211, 212, 213, 214) that learn the battery data (A) based on weight regulations including L1 regulation (Lasso) and L2 regulation (Ridge). can be generated, and an embodiment disclosed in this document is not limited to this example.
- the plurality of machine learning models 211, 212, 213, and 214 may include a plurality of deep neural network (DNN) models.
- a deep neural network model is an artificial neural network technology that includes multiple hidden layers between the input layer and the output layer. Deep neural network models can learn a variety of complex nonlinear relationships, including multiple hidden layers.
- the generator 210 uses at least some of the battery data (A) as a learning data set to generate a plurality of machine learning models (211, 212, 213, 214) that can predict the amount of gas generation of the battery. You can.
- the generator 210 may perform Min-Max Scaling on at least a portion of the battery data A.
- maximum-minimum scaling is a method of adjusting the range of all variables because if the size or unit of a numeric variable is different for each variable, the impact on the dependent variable is not properly reflected.
- the generator 210 may convert at least a portion of the battery data A into a value between 0 and 1 by maximally scaling at least a portion of the battery data A to the minimum.
- step S101 the generator 210 performs K-Fold Cross Validation on a plurality of machine learning models (211, 212, 213, 214) to generate a plurality of machine learning models (211, 212, 213, 214) can be judged on accuracy.
- K-fold cross-validation divides the preprocessed data set into a training data set and a test set, divides the training data set into 'K' folds, uses one fold for verification, and uses (K-1) folds. is a method used for model training so that all data can be used in the training and verification process.
- the generation unit 210 performs 5-fold cross validation on a plurality of machine learning models (211, 212, 213, 214) to generate a plurality of machine learning models (211, 212, 213, 214) can be judged on accuracy.
- step S101 for example, the generator 210 generates a plurality of machine learning models (211, 212, 213, 214) can be evaluated.
- step S102 the controller 220 stores battery data (A) including the voltage, current, temperature, SOC, SOH change, separator type, electrode type, and assembly process of the plurality of battery cells 110, 120, 130, and 140. ) can be used as a test data set and input into a plurality of machine learning models (211, 212, 213, 214).
- step S102 the controller 220 generates a plurality of machine learning models (211, 212, 213, 214) by extracting 80% of the battery data (A) as a learning data set. After that, the remaining 20% of the battery data (A) can be extracted as a test data set and input into a plurality of machine learning models (211, 212, 213, 214).
- step S102 the controller 220 applies at least a portion of the battery data (A) to a plurality of machine learning models (211, 212, 213, and 214) to obtain a plurality of machine learning models (211, 212, 213, and 214), respectively.
- a plurality of prediction data (B1, B2, B3, B4) predicting the gas generation amount of the battery can be obtained.
- step S102 the controller 220 may acquire a plurality of output data (Output) by inputting at least some of the battery data (A), which is one input data (Input), into separate machine learning models.
- the controller 220 may generate gas generation prediction data C that predicts the gas generation amount of the battery based on a plurality of prediction data B1, B2, B3, and B4.
- the controller 220 may apply at least some of the battery data A to a plurality of deep neural network models (DNNs) included in the plurality of machine learning models 211, 212, 213, and 214.
- DNNs deep neural network models
- the controller 220 may predict the state of the battery based on a plurality of prediction data (B1, B2, B3, B4).
- the controller 220 generates a plurality of prediction data (B1, B2, B3, B4) can be assigned weights (x1, x2, x3, x4) respectively.
- the weights (x1, x2, x3, x4) assigned to each of the plurality of prediction data (B1, B2, B3, and B4) are used by the battery when training each of the plurality of machine learning models (211, 212, 213, 214). It may be related to the weight given to the characteristics of data (A).
- the controller 220 assigns weights based on the characteristics of the battery data (A) in the process of training the deep neural network model used to generate each of the plurality of prediction data (B1, B2, B3, B4). Accordingly, weights can be assigned to a plurality of prediction data (B1, B2, B3, B4). That is, when training the plurality of machine learning models 211, 212, 213, and 214, the weight given to the characteristics of the battery data A and the plurality of prediction data B1, B2, B3, and B4, respectively. Weights are distinct. For example, the controller 220 trains the battery data ( The weight given based on the characteristics of A) can be considered.
- the controller 220 uses the battery data (A) when training a deep neural network model included in the machine learning model 212 that generates the prediction data (B2) to determine the weight to be given to the prediction data (B2). Weights given based on characteristics can be considered, and when training a deep neural network model included in the machine learning model 213 that generates prediction data B3, weights given based on the characteristics of the battery data A can be used. When training a deep neural network model included in the machine learning model 214 that generates the prediction data B4, weights assigned based on the characteristics of the battery data A can be considered.
- step S103 for example, the controller 220 inputs a plurality of prediction data (B1, B2, B3, B4) with weights (x1, x2, x3, x4) into an ensemble learning model.
- Gas generation prediction data (C) in the form of a probability distribution can be generated.
- an ensemble learning model is a machine learning technique that combines two or more learning models to perform better than a single learning model.
- the controller 220 may calculate the average value (Mean) of the gas generation prediction data (C) in the form of a probability distribution.
- the controller 220 may calculate a 95% prediction interval (Prediction Interval) of the mean value (Mean) of the gas generation amount prediction data (C) in the form of a probability distribution.
- the controller 220 calculates the mean value (Mean) and standard deviation of the gas generation prediction data (C) in the form of a probability distribution, and then calculates 'mean value ⁇ standard deviation * 1.96' in the form of a probability distribution. It can be calculated as a 95% prediction interval of the average value of the gas generation prediction data (C).
- step S103 the controller 220 compares the prediction section of the gas generation amount prediction data (C) in the form of a probability distribution with the previously stored gas generation measurement value of the battery to determine the number of machine learning models (211, 212, 213, 214). Accuracy can be judged.
- the computing system 2000 may include an MCU 2100, a memory 2200, an input/output I/F 2300, and a communication I/F 2400. there is.
- the MCU 2100 executes various programs (for example, a battery gas generation prediction program) stored in the memory 2200, processes these programs various data, and operates the battery management device 200 shown in FIG. 1 above. ) may be a processor that performs the functions of.
- the memory 2200 may store various programs related to the operation of the battery state prediction device 200. Additionally, the memory 2200 may store operating data of the battery state prediction device 200.
- the memory 2200 may be a volatile memory or a non-volatile memory.
- the memory 2200 as a volatile memory may use RAM, DRAM, SRAM, etc.
- the memory 2200 as a non-volatile memory may be ROM, PROM, EAROM, EPROM, EEPROM, flash memory, etc.
- the examples of memories 2200 listed above are merely examples and are not limited to these examples.
- the input/output I/F 2300 is an interface that connects input devices such as a keyboard, mouse, and touch panel (not shown) and output devices such as a display (not shown) and the MCU 2100 to transmit and receive data. can be provided.
- the communication I/F 2400 is a component that can transmit and receive various data with a server, and may be various devices that can support wired or wireless communication. For example, programs or various data for resistance measurement and abnormality diagnosis can be transmitted and received from a separately provided external server through the communication I/F 2400.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Probability & Statistics with Applications (AREA)
- Manufacturing & Machinery (AREA)
- Chemical & Material Sciences (AREA)
- Microelectronics & Electronic Packaging (AREA)
- Chemical Kinetics & Catalysis (AREA)
- Algebra (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Electrochemistry (AREA)
- General Chemical & Material Sciences (AREA)
- Secondary Cells (AREA)
- Tests Of Electric Status Of Batteries (AREA)
Abstract
Description
Claims (10)
- 배터리 데이터 및 상기 배터리 데이터에 포함되는 특성(feature)을 기초로 학습되어 상기 배터리의 가스 발생량을 예측하는 복수의 머신 러닝 모델을 생성하는 생성부; 및상기 배터리 데이터를 상기 복수의 머신 러닝 모델에 적용하여 상기 배터리의 가스 발생량을 예측한 복수의 예측 데이터를 획득하고, 상기 복수의 예측 데이터에 기초하여 상기 배터리의 가스 발생량을 예측하는 컨트롤러를 포함하는 배터리 상태 예측 장치.
- 제1 항에 있어서,상기 복수의 머신 러닝 모델은 복수의 심층 신경망(Deep Neural Network, DNN) 모델을 포함하고,상기 컨트롤러는,상기 배터리 데이터를 상기 복수의 심층 신경망 모델에 적용하여 상기 복수의 예측 데이터를 획득하는 배터리 상태 예측 장치.
- 제2 항에 있어서,상기 배터리 데이터는 누적적으로 측정된 상기 배터리의 온도, SOC 및 SOH를 포함하고,상기 배터리 데이터의 특성은 상기 배터리의 전극 타입, 조립 공정 및 분리막 타입 중 적어도 하나 이상을 포함하는 것을 특징으로 하는 배터리 상태 예측 장치.
- 제2 항에 있어서, 상기 컨트롤러는,상기 복수의 예측 데이터 각각을 생성하는 상기 심층 신경망 모델을 학습시키는 과정에서 상기 배터리 데이터의 특성을 기초로 부여된 가중치를 기초로 상기 복수의 예측 데이터 각각에 가중치를 부여하고, 상기 가중치가 부여된 상기 복수의 예측 데이터를 앙상블 학습 (Ensemble Learning) 모델에 입력하여 확률 분포 형태의 상기 가스 발생량 예측 데이터를 생성하는 것을 특징으로 하는 배터리 상태 예측 장치.
- 제4 항에 있어서,상기 컨트롤러는 상기 확률 분포 형태의 상기 가스 발생량 예측 데이터의 평균값을 산출하고,상기 평균값을 기 저장된 상기 배터리의 가스 발생량 측정값과 비교하여 상기 복수의 머신 러닝 모델의 정확도를 판단하는 것을 특징으로 하는 배터리 상태 예측 장치.
- 배터리 데이터 및 상기 배터리 데이터에 포함되는 특성(feature)을 기초로 학습되어 상기 배터리의 가스 발생량을 예측하는 복수의 머신 러닝 모델을 생성하는 단계;상기 배터리 데이터를 상기 복수의 머신 러닝 모델에 적용하여 상기 배터리의 가스 발생량을 예측한 복수의 예측 데이터를 획득하는 단계; 및상기 복수의 예측 데이터에 기초하여 상기 배터리의 가스 발생량을 예측하는 단계를 포함하는 배터리 상태 예측 장치의 동작 방법.
- 제6 항에 있어서,상기 복수의 머신 러닝 모델은 복수의 심층 신경망(Deep Neural Network, DNN) 모델을 포함하고,상기 배터리 데이터를 상기 복수의 머신 러닝 모델에 적용하여 상기 배터리의 가스 발생량을 예측한 복수의 예측 데이터를 획득하는 단계는상기 배터리 데이터를 상기 복수의 심층 신경망 모델에 적용하여 상기 복수의 예측 데이터를 획득하는 단계를 포함하는 배터리 상태 예측 장치의 동작 방법.
- 제6 항에 있어서,상기 배터리 데이터를 기초로 배터리의 가스 발생량을 예측하는 복수의 머신 러닝 모델을 생성하는 단계는누적적으로 측정된 상기 배터리의 온도, SOC 및 SOH를 포함하는 상기 배터리 데이터를 기초로 상기 복수의 머신 러닝 모델을 생성하는 단계를 포함하고,상기 배터리 데이터의 특성은,상기 배터리의 전극 타입, 조립 공정 및 분리막 타입 중 적어도 하나 이상을 포함하는 배터리 상태 예측 장치의 동작 방법.
- 제6 항에 있어서,상기 복수의 예측 데이터에 기초하여 상기 배터리의 가스 발생량을 예측하는 단계는,상기 복수의 예측 데이터 각각을 생성하는 상기 심층 신경망 모델을 학습시키는 과정에서 상기 배터리 데이터의 특성을 기초로 부여된 가중치를 기초로 상기 복수의 예측 데이터 각각에 가중치를 부여하는 단계; 및상기 가중치가 부여된 상기 복수의 예측 데이터를 앙상블 학습 (Ensemble Learning) 모델에 입력하여 확률 분포 형태의 상기 가스 발생량 예측 데이터를 생성하는 단계;를 포함하는 배터리 상태 예측 장치의 동작 방법.
- 제6 항에 있어서,상기 복수의 예측 데이터에 기초하여 상기 배터리의 가스 발생량을 예측하는 단계는 상기 확률 분포 형태의 상기 가스 발생량 예측 데이터의 평균값을 산출하고,상기 평균값을 기 저장된 상기 배터리의 가스 발생량 측정값과 비교하여 상기 복수의 머신 러닝 모델의 정확도를 판단하는 것을 특징으로 하는 배터리 상태 예측 장치의 동작 방법.
Priority Applications (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2025534186A JP2025540363A (ja) | 2022-12-19 | 2023-12-15 | 電池状態予測装置およびその動作方法 |
| EP23907632.6A EP4629395A4 (en) | 2022-12-19 | 2023-12-15 | BATTERY STATE PREDICTION DEVICE AND ITS OPERATING METHOD |
| CN202380086453.8A CN120359647A (zh) | 2022-12-19 | 2023-12-15 | 电池状态预测设备及其操作方法 |
Applications Claiming Priority (4)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| KR10-2022-0178733 | 2022-12-19 | ||
| KR20220178733 | 2022-12-19 | ||
| KR10-2023-0180528 | 2023-12-13 | ||
| KR1020230180528A KR102925815B1 (ko) | 2022-12-19 | 2023-12-13 | 배터리 상태 예측 장치 및 그것의 동작 방법 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2024136347A1 true WO2024136347A1 (ko) | 2024-06-27 |
Family
ID=91589291
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/KR2023/020817 Ceased WO2024136347A1 (ko) | 2022-12-19 | 2023-12-15 | 배터리 상태 예측 장치 및 그것의 동작 방법 |
Country Status (4)
| Country | Link |
|---|---|
| EP (1) | EP4629395A4 (ko) |
| JP (1) | JP2025540363A (ko) |
| CN (1) | CN120359647A (ko) |
| WO (1) | WO2024136347A1 (ko) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118983548A (zh) * | 2024-10-22 | 2024-11-19 | 中海巢(河北)新能源科技有限公司 | 电池组压力智能控制方法、装置、系统及存储介质 |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20170085365A (ko) * | 2016-01-14 | 2017-07-24 | 삼성전자주식회사 | 배터리의 상태 추정 장치 및 방법 |
| KR20180033766A (ko) * | 2016-09-26 | 2018-04-04 | 주식회사 엘지화학 | 인공지능형 연료전지 시스템 |
| JP2021176131A (ja) * | 2020-05-01 | 2021-11-04 | ダイキン工業株式会社 | 学習モデル生成方法、プログラム、記憶媒体、学習済みモデル |
| WO2022034983A1 (ko) * | 2020-08-14 | 2022-02-17 | 주식회사 한국파워셀 | 신경망 기반의 배터리 셀 불량 및 화재 사전 진단 방법 및 장치 |
| KR20220064575A (ko) * | 2020-11-12 | 2022-05-19 | 주식회사 에니트 | Ess 화재 예방 시스템 및 그 방법 |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR102892949B1 (ko) * | 2020-03-17 | 2025-11-27 | 주식회사 엘지에너지솔루션 | 배터리 이상 감지 장치 및 방법, 그 방법을 제공하는 배터리 관리 시스템 |
-
2023
- 2023-12-15 EP EP23907632.6A patent/EP4629395A4/en active Pending
- 2023-12-15 WO PCT/KR2023/020817 patent/WO2024136347A1/ko not_active Ceased
- 2023-12-15 JP JP2025534186A patent/JP2025540363A/ja active Pending
- 2023-12-15 CN CN202380086453.8A patent/CN120359647A/zh active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20170085365A (ko) * | 2016-01-14 | 2017-07-24 | 삼성전자주식회사 | 배터리의 상태 추정 장치 및 방법 |
| KR20180033766A (ko) * | 2016-09-26 | 2018-04-04 | 주식회사 엘지화학 | 인공지능형 연료전지 시스템 |
| JP2021176131A (ja) * | 2020-05-01 | 2021-11-04 | ダイキン工業株式会社 | 学習モデル生成方法、プログラム、記憶媒体、学習済みモデル |
| WO2022034983A1 (ko) * | 2020-08-14 | 2022-02-17 | 주식회사 한국파워셀 | 신경망 기반의 배터리 셀 불량 및 화재 사전 진단 방법 및 장치 |
| KR20220064575A (ko) * | 2020-11-12 | 2022-05-19 | 주식회사 에니트 | Ess 화재 예방 시스템 및 그 방법 |
Non-Patent Citations (1)
| Title |
|---|
| See also references of EP4629395A1 |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118983548A (zh) * | 2024-10-22 | 2024-11-19 | 中海巢(河北)新能源科技有限公司 | 电池组压力智能控制方法、装置、系统及存储介质 |
Also Published As
| Publication number | Publication date |
|---|---|
| EP4629395A4 (en) | 2026-04-15 |
| JP2025540363A (ja) | 2025-12-11 |
| EP4629395A1 (en) | 2025-10-08 |
| CN120359647A (zh) | 2025-07-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2022103185A1 (ko) | 전지의 용량 측정 장치 및 방법, 및 상기 장치를 포함하는 전지 제어 시스템 | |
| WO2021006566A1 (ko) | 배터리 셀 진단 장치 및 방법 | |
| WO2020262787A1 (ko) | 내부 단락 셀 검출 방법 | |
| WO2022149824A1 (ko) | 배터리 관리 장치 및 방법 | |
| WO2022265458A1 (ko) | 이차전지 저전압 불량 예측 장치 및 방법, 상기 장치를 포함하는 전지 제어 시스템 | |
| JP7658464B2 (ja) | 電池管理装置およびそれを含む電池検査システム | |
| WO2023153765A1 (ko) | 배터리 관리 장치 및 그것의 동작 방법 | |
| WO2023132520A1 (ko) | 배터리 용량 예측 장치 및 그것의 동작 방법 | |
| WO2024049172A1 (ko) | 배터리 상태 예측 장치 및 그것의 동작 방법 | |
| WO2023075244A1 (ko) | 배터리 관리 장치 및 그것의 동작 방법 | |
| WO2023136512A1 (ko) | 배터리 충전 심도 산출 장치 및 그것의 동작 방법 | |
| WO2024136347A1 (ko) | 배터리 상태 예측 장치 및 그것의 동작 방법 | |
| KR20240000294A (ko) | 배터리 관리 장치 및 그것의 동작 방법 | |
| KR102925815B1 (ko) | 배터리 상태 예측 장치 및 그것의 동작 방법 | |
| WO2023158195A1 (ko) | 배터리 관리 장치 및 서버 | |
| WO2024185954A1 (ko) | 배터리 데이터 관리 장치 및 그것의 동작 방법 | |
| WO2021025295A1 (ko) | 배터리 관리 장치, 배터리 관리 방법 및 배터리 팩 | |
| WO2024136225A1 (ko) | 배터리 데이터 관리 장치 및 그것의 동작 방법 | |
| WO2024205079A1 (ko) | 배터리 관리 장치 및 그것의 동작 방법 | |
| WO2024101908A1 (ko) | 배터리 팩 관리 장치 및 그것의 동작 방법 | |
| WO2024248267A1 (ko) | Ev(electric vehicle)용 sohc 예측방법 및 시스템 | |
| WO2026024021A1 (en) | Battery management apparatus and method | |
| WO2025221126A1 (ko) | 배터리 진단 장치 및 그것의 동작 방법 | |
| WO2025028836A1 (ko) | 배터리 관리 장치 및 그것의 동작 방법 | |
| WO2025116212A1 (ko) | 배터리 관리 장치 및 그것의 동작 방법 |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 23907632 Country of ref document: EP Kind code of ref document: A1 |
|
| ENP | Entry into the national phase |
Ref document number: 2025534186 Country of ref document: JP Kind code of ref document: A |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 2025534186 Country of ref document: JP |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 202380086453.8 Country of ref document: CN |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 2023907632 Country of ref document: EP |
|
| ENP | Entry into the national phase |
Ref document number: 2023907632 Country of ref document: EP Effective date: 20250703 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| WWP | Wipo information: published in national office |
Ref document number: 202380086453.8 Country of ref document: CN |
|
| WWP | Wipo information: published in national office |
Ref document number: 2023907632 Country of ref document: EP |