WO2023063587A1 - 배터리 진단 방법 및 그 장치 - Google Patents
배터리 진단 방법 및 그 장치 Download PDFInfo
- Publication number
- WO2023063587A1 WO2023063587A1 PCT/KR2022/013472 KR2022013472W WO2023063587A1 WO 2023063587 A1 WO2023063587 A1 WO 2023063587A1 KR 2022013472 W KR2022013472 W KR 2022013472W WO 2023063587 A1 WO2023063587 A1 WO 2023063587A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- battery
- time
- series data
- prediction model
- impedance
- 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
- 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
- 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]
- G01R31/374—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC] with means for correcting the measurement for temperature or ageing
-
- 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]
- G01R31/382—Arrangements for monitoring battery or accumulator variables, e.g. SoC
- G01R31/3842—Arrangements for monitoring battery or accumulator variables, e.g. SoC combining voltage and current measurements
-
- 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]
- G01R31/389—Measuring internal impedance, internal conductance or related variables
-
- 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]
- G01R31/392—Determining battery ageing or deterioration, e.g. state of health
-
- 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
- G06N3/08—Learning methods
- G06N3/086—Learning methods using evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- 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
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/0985—Hyperparameter optimisation; Meta-learning; Learning-to-learn
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/02—Knowledge representation; Symbolic representation
- G06N5/022—Knowledge engineering; Knowledge acquisition
Definitions
- Embodiments of the present invention relate to a method and device for diagnosing states such as life span, capacity, and internal resistance of a battery, and more particularly, to a method and device for diagnosing a battery state using artificial intelligence.
- a technical problem to be achieved by an embodiment of the present invention is a battery diagnosis capable of accurately grasping the state of a battery exhibiting nonlinearity by considering both time-series data such as current, voltage, and temperature of the battery and non-time-series data including the impedance of the battery. It is to provide a method and an apparatus therefor.
- An example of a battery diagnosis method for achieving the above technical problem is a battery diagnosis method performed by a computing device driving a battery prediction model learned to predict battery state information, receiving time-series data including at least one of voltage, current, and temperature of the battery measured during a period of time; Receiving non-time series data including battery impedance measured at a certain point in time; and predicting battery state information by inputting the time-series data and the non-time-series data to the battery prediction model.
- an example of a battery diagnosis apparatus is a battery diagnosis method performed by a computing device driving a battery prediction model learned to predict battery state information, a first input unit that receives time-series data including at least one of voltage, current, and temperature of the battery measured during the period; a second input unit that receives non-time-series data including battery impedance measured at a certain point in time; and a prediction unit inputting the time-series data and the non-time-series data to the battery prediction model to predict battery state information.
- FIG. 1 is a diagram showing an example of a schematic structure of a battery diagnosis method according to an embodiment of the present invention
- FIG. 2 is a diagram showing an example of time-series data for learning a battery diagnostic model according to an embodiment of the present invention
- FIG. 3 is a diagram showing an example of a battery diagnosis model according to an embodiment of the present invention.
- FIG. 4 is a diagram showing an example of implementing the artificial neural network of the battery diagnosis model of FIG. 3 as an RNN;
- FIG. 5 is a diagram showing another example of a battery diagnosis model according to an embodiment of the present invention.
- FIG. 6 is a diagram showing an example of implementing the artificial neural network of the battery diagnosis model of FIG. 5 with a CNN;
- FIG. 7 is a diagram showing an example of a method for optimizing hyperparameters of a battery diagnostic model according to an embodiment of the present invention.
- FIG. 8 is a diagram showing the configuration of an example of a battery diagnostic device according to an embodiment of the present invention.
- FIG. 9 is a diagram showing the configuration of an example of an impedance measuring device according to an embodiment of the present invention.
- FIG. 10 is a diagram showing an implementation example of an impedance measuring device according to an embodiment of the present invention.
- FIG. 11 is a diagram showing an example of supplying power to an impedance measuring device according to an embodiment of the present invention.
- FIG. 12 is a diagram showing an example of an impedance measuring method according to an embodiment of the present invention.
- FIG. 13 is a diagram showing an example of a sinusoidal wave signal generated for impedance measurement according to an embodiment of the present invention.
- FIG. 14 is a diagram illustrating an example of a method for changing the resistance value of a resistor unit according to an embodiment of the present invention.
- FIG. 1 is a diagram illustrating an example of a schematic structure of a method for diagnosing a battery according to an embodiment of the present invention.
- the battery diagnosis apparatus 100 includes an artificial intelligence model (or deep learning model) that predicts a battery state based on battery measurement data.
- an artificial intelligence model for predicting a battery state is referred to as a battery diagnosis model.
- the battery diagnosis model may be classified as a regression model. The structure of the battery diagnosis model will be reviewed again with reference to FIGS. 3 to 6 .
- Battery measurement data is largely divided into time series data 110 and non-time series data 120 .
- the time-series data 110 may include at least one of current, voltage, and temperature of a battery measured for a certain period of time (eg, several months or several hours).
- the time-series data 110 is data sequentially measured and accumulated over time, and the present embodiment presents current, voltage, temperature, etc. of a battery as an example, but is not necessarily limited thereto.
- An example of time series data 110 in a 3D form is shown in FIG. 2 .
- the non-time series data 120 includes impedance data of a battery measured at a certain point in time.
- the non-time series data 120 may be spectral data for battery impedance measured when predicting the state information 130 of the battery.
- the non-time-series data 120 is data measured at a specific point in time rather than information accumulated over time, and impedance information is presented as an example of the non-time-series data 120 in this embodiment, but is not necessarily limited thereto.
- Impedance information of the battery included in the non-time series data 120 may be information measured using a plurality of different sinusoidal wave signals.
- impedance information (frequency (f), real value ( Re), imaginary side value (Im)
- frequency information may be omitted from the non-time-series data 120.
- the battery condition information 130 predicted by the battery diagnosis apparatus 100 using the battery diagnosis model may include at least one of battery life, capacity, and internal resistance.
- the battery state information 130 may be other various types of information according to embodiments, but is not necessarily limited thereto.
- the battery diagnosis model which is a battery state inference model, may be trained by supervised learning, and training data for this includes target data.
- the learning data may include time-series data 110 for at least one battery, non-time-series data 120, and battery state information 130 used as target data.
- the battery diagnostic model compares the prediction information predicted using the time series data 110 and non-time series data 120 included in the learning data with the battery condition information, which is the target data (ie, the correct answer) included in the learning data, to determine the internal parameters. It can be created through a supervised learning process that adjusts the etc.
- the present invention presents a new internal structure of a battery diagnostic model to accurately predict a battery condition by efficiently combining time-series data 110 and non-time-series data 120, examples of which are shown in FIGS. 3 to 6.
- FIG. 2 is a diagram illustrating an example of time-series data for learning a battery diagnostic model according to an embodiment of the present invention.
- time-series data 110 is composed of three features of battery voltage, current, and temperature.
- the time series data 110 includes information on a plurality of batteries and is measured for a certain period of time. Accordingly, the time series data 110 may be three-dimensional data composed of three axes: a plurality of features, a plurality of batteries (ie, the number of samples), and a measurement period (ie, sequence length).
- FIG. 3 is a diagram illustrating an example of a battery diagnosis model according to an embodiment of the present invention.
- the battery diagnosis model includes an artificial neural network 300 and a fully connected layer 310 .
- the artificial neural network 300 may be composed of a recurrent neural network such as a recurrent neural network (RNN) or a convolutional neural network (CNN).
- RNN recurrent neural network
- CNN convolutional neural network
- An example of implementing the artificial neural network 300 as an RNN is shown in FIG.
- various types of conventional artificial neural networks may be applied to this embodiment.
- the fully connected layer 310 is one of the structures of an artificial neural network, and is a layer configured in a state in which all neurons of the previous layer are connected to all neurons of the next layer.
- the fully connected layer 310 may be used as a role of flattening feature values of a 2D array into feature values of a 1D array and deriving a final predicted value.
- the fully connected layer 310 is used to predict the result of the battery state by integrating two data of different properties, non-time series data and time series data.
- Various types of conventional fully connected layers may be used in this embodiment.
- time series data 350 and non-time series data 370 are used as input data of the battery diagnosis model.
- the input data of the artificial neural network 300 is time series data 350
- the input data of the fully connected layer 310 is the output value of the artificial neural network 300 (ie, the first feature value 360) and non-time series data 370. ) is the combined data.
- the battery diagnostic model may perform a preprocessing process (eg, normalization, outlier removal, etc.) on time series data, convert the time series data into a vector form, and input the time series data to the artificial neural network 300 .
- a preprocessing process eg, normalization, outlier removal, etc.
- the battery diagnostic model may introduce a batch normalization layer for normalizing input data to an input layer of the artificial neural network 300 in order to prevent instability of learning.
- the batch normalization layer can increase the stability of learning by processing the average value and standard deviation of the input data to be distributed around 0 and 1.
- the first feature value 360 output by the artificial neural network 300 may include at least one or more features, which may be implemented in various ways according to embodiments.
- the battery diagnosis model generates a second feature value by concatenating the non-time series data 370 and the first feature value 360, and inputs the second feature value to the fully connected layer 310 to obtain battery state information. predicts (380).
- FIG. 4 is a diagram illustrating an example in which the artificial neural network of the battery diagnosis model of FIG. 3 is implemented as an RNN.
- the artificial neural network 300 of the battery diagnosis model of FIG. 3 is implemented as a recurrent neural network RNN cell (eg, long short term memory (LSTM), gated recurrent unit (GRU), etc.) can Since RNN repeatedly inputs time series data (400, 402, 404) up to a certain point in the past, it is effective in predictive models for time series data.
- RNN cell eg, long short term memory (LSTM), gated recurrent unit (GRU), etc.
- RNN representative values are generated for each RNN cell corresponding to each time step.
- the network structure in which the RNN feature value of the last RNN cell corresponding to the most recent type step is combined with impedance data is implemented as That is, non-time-series data (ie, impedance data) 420 is combined with feature values 410 generated through RNN rather than directly combined with time-series data.
- a feature value obtained by combining the output value 410 of the RNN and the non-time series data 420 is input to the fully connected layer 430 to predict battery state information.
- this embodiment shows an example in which RNN cells and stacked cells are each shown as one layer and three time series data are input, but this is only one example and is not limited thereto.
- FIG. 5 is a diagram illustrating another example of a battery diagnosis model according to an embodiment of the present invention.
- the battery diagnosis model includes a plurality of artificial neural networks 500 and 502 and a fully connected layer 510 .
- the plurality of artificial neural networks 500 and 520 may be implemented as CNNs, and an example thereof is shown in FIG. 6 .
- Each of the artificial neural networks 500 and 502 receives time-series data 550 and non-time-series data 560 respectively and outputs respective first feature values. Then, the battery diagnostic model generates a second feature value by connecting each first feature value, and inputs the second feature value to the fully connected layer 510 to predict the battery state information 570 .
- FIG. 6 is a diagram illustrating an example in which the artificial neural network of the battery diagnosis model of FIG. 5 is implemented as a CNN.
- the battery diagnosis model is composed of a one-dimensional CNN.
- Local connectivity between features of input data including time-series data (600, 610, and 620) and non-time-series data (620) may be captured by using a CNN.
- a CNN For example, in order to grasp the local connectivity with time of voltage 600, current 610, and temperature 620, which are characteristics of time-series data, and to capture local connectivity of frequency with respect to impedance 630, which is non-time-series data.
- the battery diagnosis model of this embodiment includes a 1D convolutional layer (1D Conv.) and a pooling layer.
- the battery diagnostic model integrates and flattens each feature value obtained by inputting each data of time series data (600, 610, 620) and non-time series data (630) into a one-dimensional convolution layer through a concatenation layer (640). ), and finally predict battery state information through the fully connected layer 650.
- FIG. 7 is a diagram illustrating an example of a method of optimizing hyperparameters of a battery diagnosis model according to an embodiment of the present invention.
- the performance of the battery diagnosis model may vary according to hyperparameters. For example, since there are many hyperparameters such as sequence length of time series data, network structure (stach size of RNN, etc.), learning rate, and drop-out ratio, optimal combination of these Through this, the performance of the battery diagnosis model can be improved.
- this embodiment suggests a method of deriving optimal hyperparameters through a genetic algorithm.
- the battery diagnosis device learns a battery diagnosis model based on a plurality of candidate hyperparameters 700 (710). It is assumed that the plurality of candidate hyperparameters 700 are predefined in various ways. For example, the plurality of candidate hyperparameters 700 may be configured with arbitrary values.
- the battery diagnosis device evaluates fitness values of a plurality of trained battery diagnosis models using a plurality of candidate hyperparameters 700 (710).
- Various conventional methods for calculating the degree of fitness for the learning result of the battery diagnostic model may be applied to this embodiment.
- the battery diagnosis device selects at least one candidate hyperparameter as a parent hyperparameter based on the suitability. For example, candidate hyperparameters used in the top two battery diagnosis models in the order of highest suitability may be selected as parent hyperparameters.
- the battery diagnosis device adds at least one child hyperparameter generated through cross-over of part of parent hyperparameters or mutation of subbodi hyperparameters as new candidate hyperparameters (720).
- the battery diagnosis device trains a battery diagnosis model using at least one new candidate hyperparameter, evaluates the fitness, selects the candidate hyperparameter of the battery diagnosis model with high fitness as a parent hyperparameter, and selects the parent hyperparameter.
- the process of generating child hyperparameters from is repeated under a predefined condition or up to a predefined number of times.
- Optimal hyperparameters can be derived through this iterative process.
- the battery diagnosis device may define a point at which a fitness curve obtained by connecting the fitness obtained each time is saturated (that is, a point at which the fitness increase rate in each repetition process is less than a certain level) as a stopping point.
- FIG. 8 is a diagram illustrating an example configuration of a battery diagnosis device according to an embodiment of the present invention.
- the battery diagnosis apparatus 800 includes a first input unit 810 , a second input unit 820 and a prediction unit 830 .
- the battery diagnosis device 800 may be implemented as a computing device including a memory, a processor, an input/output device, and the like. In this case, each configuration may be implemented as software, loaded into a memory, and executed by a processor.
- the first input unit 810 receives time-series data such as voltage, current, or temperature of the battery measured for a certain period of time.
- the second input unit 820 receives non-time series data such as the impedance of the battery measured at a certain point in time.
- the prediction unit 830 predicts battery state information by inputting time series data and non-time series data to a battery prediction model that has been fully trained.
- the prediction unit 830 generates a first feature value by inputting time-series data to the artificial neural network of the battery prediction model as shown in FIG. 3 or 5, and includes the first output value of the artificial neural network and the impedance of the battery.
- a second feature value obtained by combining series data may be generated, and battery state information may be output by inputting the second feature value to a fully connected layer of the battery prediction model.
- FIG. 9 is a diagram showing the configuration of an example of an impedance measuring device according to an embodiment of the present invention.
- the impedance measurement device 900 includes a sine wave generator 910, a signal application unit 920, a current measurement unit 930, a voltage measurement unit 940, and an impedance calculation unit 950. .
- the impedance measurement device 900 may further include a control unit 960, a DC measurement unit 970, and a temperature measurement unit 980.
- this embodiment assumes a case including a control unit 960, a direct current measurement unit 970, and a temperature measurement unit 980, but the corresponding components may be omitted.
- the sine wave generator 910 generates a sine wave signal within a certain frequency range.
- the sinusoidal wave generator 910 may generate a plurality of sinusoidal signals having different frequencies in various frequency ranges from 0.1 Hz to 5 kHz.
- the frequency range for generating the sinusoidal wave signal may be variously modified according to embodiments.
- the magnitude of the sinusoidal wave signal may vary depending on embodiments, such as several mV to several tens of V.
- An example of a sine wave signal generated by the sine wave generator 110 within a certain frequency range is shown in FIG. 13 .
- the sine wave generator 910 may be implemented as an IC chip that generates a sine wave as shown in FIG. 10 .
- the sine wave generator 910 may be implemented by applying various conventional techniques for generating sine waves.
- the signal applying unit 920 applies the sinusoidal wave signal to the gate electrode of a MOSFET (Metal Oxide Semiconductor Field Effect Transistor) 925 connected to the object to be measured 990 . That is, the gate electrode of the MOSFET 925 is connected to the signal applying unit 920 and the drain electrode is connected to the target 990 to be measured.
- the subject to be measured 990 is a target device whose impedance is to be determined, and an example is a battery (ie, a secondary battery). In this embodiment, for convenience of description, it is assumed that the subject to be measured 990 is a battery, but the subject to be measured 990 may be various other than the battery, and is not limited to the battery.
- the signal applying unit 920 has an amplification unit that amplifies the small sine wave signal output from the sine wave generator 910 to a gate voltage suitable for driving the MOSFET 925, and the output of the amplification unit is constant. It may include a feedback circuit that allows it to be within the range. Since the configuration of the signal applying unit 920 including the amplifying unit and the feedback circuit is shown in FIG. 10 , this will be reviewed again in FIG. 10 .
- the current measuring unit 930 measures the current value of the measurement target 990 output from the source electrode of the MOSFET 925 when a sinusoidal wave signal is applied to the gate electrode of the MOSFET 925 .
- a DC voltage eg, battery voltage when the subject to be measured is a battery
- a certain magnitude eg, tens to hundreds of V
- the MOSFET 925 outputs a current according to the magnitude of the sinusoidal signal through the source electrode in a linear operating region.
- the voltage measurement unit 940 measures the voltage value of the measurement target 990 according to the application of the sine wave signal. As a sinusoidal signal is applied to the gate electrode of the MOSFET 925, current flows through the MOSFET 925, and the voltage of the target 990 to be measured changes accordingly.
- the voltage measurement unit 940 may be connected in series with the measurement target 990 to measure the voltage of the measurement target.
- the current value measured by the current measuring unit 930 and the voltage value measured by the voltage measuring unit 940 may be AC component values from which the DC component is removed.
- the current measurement unit 930 and the voltage measurement unit 940 may further include a component for removing the DC component (eg, the AC coupling unit in FIG. 10). there is.
- various conventional methods of determining the AC component by removing the DC component from the voltage value and the current value may be applied to this embodiment.
- the impedance calculation unit 950 calculates the impedance of the object to be measured based on the current value determined by the current measurement unit 930 and the voltage value determined by the voltage measurement unit 940 .
- the impedance calculation unit 950 applies the Hamming window algorithm and Discrete Fourier Transform (DFT) to finite current and voltage values identified at regular intervals, and calculates a Nyquist Plot. Impedance can be determined using Since a method of obtaining impedance using a finite number of current and voltage values that change over time is a well-known technique, a detailed description thereof will be omitted.
- DFT Discrete Fourier Transform
- the impedance calculation unit 950 may calculate the impedance of a target to be measured with respect to sinusoidal signals of various frequencies. For example, the impedance calculation unit 950 may calculate first impedance when a sine wave signal of a first frequency is applied, and calculate second impedance when a sinusoidal wave signal of a second frequency is applied. .
- the interval between the first frequency and the second frequency may be a logarithmic interval as shown in FIG. 5 .
- the temperature measuring unit 980 measures the temperature of the MOSFET 925 or the like. A region where the temperature measurement unit 980 measures the temperature may be implemented in various ways according to embodiments.
- the controller 960 may stop the operation of the impedance measurement target 900 in order to prevent a fire or malfunction when the temperature measured by the temperature measurement unit 980 exceeds a predetermined temperature.
- the DC measuring unit 970 measures the DC voltage or the like of the measurement target 990 .
- the control unit 960 controls each component such as the sine wave generator 910 and the like. For example, it is possible to control the sine wave generator 910 to generate sine wave signals of various frequencies.
- the controller 960 may be connected to each control target, such as the sine wave generator 910, through Serial Peripheral Interface (SPI) communication.
- SPI Serial Peripheral Interface
- the controller 960 may transmit the frequency range and the number of measurement points of the sine wave signal to the sine wave generator 910 .
- FIG. 10 is a diagram showing an implementation example of an impedance measuring device according to an embodiment of the present invention.
- the impedance measuring device 900 includes an IC chip 1000 that generates a sine wave signal, an amplifier 1020 that amplifies the sine wave signal, and a MOSFET for measuring current and voltage changes according to the application of the sine wave signal.
- a resistance unit 940 including a plurality of resistors connected in parallel with 1030, a switch unit 1050 that turns on/off each resistance connection, and the like.
- the present embodiment includes all components of the control unit 960, temperature measurement unit 980, and direct current measurement unit 990, but these components may be omitted depending on the embodiment.
- the IC chip 1000 generates a sinusoidal signal. That is, the salpin sine wave generator 910 in FIG. 9 may be implemented as an IC chip. Various conventional IC chips that generate sinusoidal signals may be applied to this embodiment.
- the sine wave signal generated by the IC chip 1000 is input to the gate electrode of the MOSFET (1030, corresponding to 925 in FIG. 9) through the amplifier 1020, and the output of the source electrode of the MOSFET 1030 is the amplifier 1020 ) is fed back. That is, the signal applying unit 920 of FIG. 9 may be composed of the amplifying unit 1020 and the feedback circuit of this embodiment.
- the IC chip 1000 registers the frequency range of the sinusoidal signal received from the control unit 960 and the number of different frequencies to be generated within the range in a register, and generates a sinusoidal signal based on this and can measure impedance.
- a sinusoidal wave signal generated within the frequency range may be generated by being divided into logarithmic intervals as shown in the example of FIG. 13 .
- the IC chip 1000 may store the measured voltage values and current values in a first in first out (FIFO) queue, etc., and calculate impedance values for each frequency of the sinusoidal wave signal based on the stored values.
- FIFO first in first out
- the source electrode of the MOSFET 1030 is connected to the target to be measured 990, and the gate electrode is connected to the amplifier 1020.
- the resistance unit 1040 connected in parallel with the object to be measured 990 is used to measure the current output from the source electrode of the MOSFET 1030.
- the IC chip 1000 measures the voltage value of the resistance unit 1040 through the third line 1074 and the fourth line 1076 connected to the resistance unit 1040, and the measured voltage value and the resistance unit 1040 A current value output from the source electrode of the MOSFET 1030 is calculated using the resistance value of . That is, the current measurement unit 930 of FIG. 9 may be implemented with the resistor unit 1040 and the IC chip 1000 according to the present embodiment.
- An AC coupling unit 1010 including capacitors C2 and C4 may be positioned in the third line 1074 and the fourth line 1076 so that the IC chip 1000 can measure only the current value of the AC component. .
- the resistance values of the plurality of resistors connected in parallel of the resistor unit 1040 are known to the IC chip 1000 or the controller 960 in advance.
- the resistor unit 1040 may be implemented as a shunt resistor.
- This embodiment includes a switch unit 1050 capable of controlling on/off of the connection of each resistance constituting the resistance unit 1040 so that the voltage range of the measurement target 990 that can be measured can vary. do.
- This embodiment shows an example in which the switch unit 1050 is implemented as a MOSFET connected in series with each resistor, but is not necessarily limited thereto. Since the resistance value of each resistor of the resistor unit 1040 is known in advance, the resistance value of the resistor unit 1040 can be obtained according to the operating state (ie, on/off state) of each switch of the switch unit 1050.
- Control of the switch unit 1050 may be performed by the control unit 960 .
- the IC chip 1000 can control each switch.
- the control unit 960 controls the switch unit 1050.
- a method in which the control unit 960 adjusts the resistance value of the resistance unit 1040 by controlling the switch unit 1050 according to the voltage level of the measurement target 990 will be reviewed again in FIG. 14 .
- the IC chip 1000 measures the voltage of the measurement target 990 according to the application of the sine wave signal.
- the IC chip 200 measures the voltage value of the target to be measured 990 according to the application of the sine wave signal through the first line 1070 and the fifth line 1078 connected in series with the target to be measured 990. Measure.
- the AC coupling unit including capacitors C1 and C4 in the first line 1070 and the fifth line 1078 so that the IC chip 1000 can measure the voltage value of the AC component from which the DC component has been removed ( 1010) may exist.
- a diode 1060 may be further included between the MOSFET 1030 and the target to be measured 990 to prevent counter-electromotive force and serve as a rectifier.
- FIG. 11 is a diagram showing an example of supplying power to an impedance measuring device according to an embodiment of the present invention.
- the power supply unit 1110 provides the power supplied through the USB port 1100 to the impedance measuring device 900 .
- the power supply unit 1110 isolates the first voltage (eg, 4.2 to 5.5V) supplied through the USB port 1100 through a transformer, and uses the isolated voltage as an impedance measuring device.
- the voltage may be boosted (eg, a DC-DC converter) to a second voltage (eg, a constant voltage such as 5V and 15V) required by the 900 and supplied to the impedance measuring device 900 .
- a first voltage eg, 4.2 to 5.5V
- the voltage may be boosted (eg, a DC-DC converter) to a second voltage (eg, a constant voltage such as 5V and 15V) required by the 900 and supplied to the impedance measuring device 900 .
- the impedance measurement device 900 only needs to apply a small sine wave signal of several m to several V to the gate voltage of the MOSFETs 925 and 1030, so it does not require large power, so the USB port ( 1100) can be driven with low power.
- the impedance measurement device 900 can be driven by being connected to the USB port 1100, it is connected to various electronic devices (eg, computers, laptops, tablet PCs, smartphones, etc.) in which the USB port 1100 exists. Also, the impedance measurement device 900 may output the measured impedance value to the outside in real time through the USB port 1100. If the electronic device connected to the USB port 1100 is a device capable of wired/wireless communication, the impedance measurement device 900 may transmit the measured impedance value to an external device in real time through wired/wireless communication.
- various electronic devices eg, computers, laptops, tablet PCs, smartphones, etc.
- the impedance measurement device 900 may transmit the measured impedance value to an external device in real time through wired/wireless communication.
- the impedance measurement device 900 can be connected to various electronic devices through a USB port, the electronic devices connected to the USB port 1100 do not have a separate display unit for outputting impedance measurement results through a screen or the like. can be displayed through Of course, a display unit may be present in the impedance measuring device 900 .
- FIG. 12 is a diagram showing an example of an impedance measurement method according to an embodiment of the present invention.
- the impedance measuring device 900 generates a sine wave signal within a certain frequency range (S1200).
- the impedance measuring device 900 applies the sine wave signal to the gate voltage of the MOSFET 925 connected to the target to be measured (S1210).
- the impedance measurement device 900 measures the current value output from the source electrode of the MOSFET 925 and the voltage value output from the measurement target 990 according to the application of the sine wave signal (S1220). Then, the impedance measuring device 900 calculates the impedance of the target to be measured 990 based on the current value and the voltage value (S1230).
- FIG. 13 is a diagram illustrating an example of a sinusoidal wave signal generated for impedance measurement according to an embodiment of the present invention.
- the impedance measuring device 900 may generate a plurality of sinusoidal signals at logarithmic intervals in a predetermined frequency range 1300 .
- the impedance measuring device 900 outputs a sinusoidal signal having a frequency 1310 of f 0 for a certain period of time to measure the impedance of a target to be measured, and then obtains a sinusoidal signal 10 times larger than f 0 (i.e., 2f 0 ( log)) 1320 is output for a certain period of time to measure the impedance of the object to be measured.
- sinusoidal signals having frequencies 1310, 1320, and 1330 at logarithmic intervals are generated within a certain frequency range to measure the impedance of the target to be measured.
- the frequency range for generating sinusoidal signals and the number of sinusoidal signals of different frequencies may be set in various ways according to embodiments, and do not necessarily have to be logarithmic intervals.
- FIG. 14 is a diagram illustrating an example of a method for changing the resistance value of a resistor unit according to an embodiment of the present invention.
- the control unit 960 of the impedance measuring device 900 controls each switch of the switch unit 1050 to be in an on state so that a required current flows through each resistor.
- the control unit 960 determines the voltage value of the resistance unit 1040 (S1400).
- the control unit 960 changes at least one connection among a plurality of connected switches to an off state (S1420), so that the current flowing through each resistance to increase the size of That is, when the magnitude of the current increases, the voltage value of the resistance part increases.
- the control unit 960 repeatedly changes the state of the switch until the voltage value of the resistance unit 1040 is greater than the threshold value.
- each resistor connected in parallel of the resistor unit 1040 may have the same resistance value or different resistance values.
- the control unit 960 may control the switch unit 1050 to be sequentially turned off from the connection of resistors having a large resistance value so that the current flowing through each resistor of the resistor unit 1040 sequentially increases from a small value.
- the size of the voltage or the like of the target to be measured 990 is not set in advance, current is distributed to each resistance to prevent damage to the impedance measurement device due to excessive current and to be suitable for measuring the impedance of the target to be measured 990.
- the size of the resistance value of the resistor unit 1040 may be adjusted.
- Each embodiment of the present invention can also be implemented as computer readable codes on a computer readable recording medium.
- a computer-readable recording medium includes all types of recording devices in which data that can be read by a computer system is stored. Examples of computer-readable recording media include ROM, RAM, CD-ROM, SSD, and optical data storage devices.
- the computer-readable recording medium may be distributed to computer systems connected through a network to store and execute computer-readable codes in a distributed manner.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Physiology (AREA)
- Measurement Of Resistance Or Impedance (AREA)
- Tests Of Electric Status Of Batteries (AREA)
- Apparatuses And Processes For Manufacturing Resistors (AREA)
Abstract
Description
Claims (9)
- 배터리 상태정보를 예측하도록 학습된 배터리예측모델을 구동하는 컴퓨팅 장치가 수행하는 배터리 진단 방법에 있어서,일정 기간 동안 측정된 배터리의 전압, 전류 및 온도 중 적어도 하나 이상을 포함하는 시계열 데이터를 입력받는 단계;일정 시점에 측정된 배터리 임피던스를 포함하는 비시계열 데이터를 입력받는 단계; 및상기 시계열 데이터와 상기 비시계열 데이터를 상기 배터리예측모델에 입력하여 배터리 상태정보를 예측하는 단계;를 포함하는 것을 특징으로 하는 배터리 진단 방법.
- 제 1항에 있어서, 상기 예측하는 단계는,상기 시계열 데이터를 상기 배터리예측모델의 인공신경망에 입력하여 제1 특징값을 생성하는 단계;상기 인공신경망의 제1 특징값과 상기 비시계열 데이터를 결합한 제2 특징값을 생성하는 단계; 및상기 제2 특징값을 상기 배터리예측모델의 완전연결계층에 입력하여 배터리 상태정보를 출력하는 단계;를 포함하는 것을 특징으로 하는 배터리 진단 방법.
- 제 2항에 있어서,상기 인공신경망은 순환신경망인 것을 특징으로 하는 배터리 진단 방법.
- 제1 항에 있어서,상기 배터리예측모델을 위한 학습데이터는, 복수 개의 배터리 샘플에 대해 일정 기간 동안 측정한 전압, 전류 및 온도 중 적어도 하나 이상의 시계열 학습데이터와, 복수 개의 배터리 샘플의 임피던스를 포함하는 비시계열 학습데이터와, 배터리 상태정보를 포함하는 검증데이터를 포함하고,상기 배터리예측모델은 상기 시계열학습데이터와 상기 비시계열학습데이터를 기반으로 한 예측데이터와 상기 검증데이터를 비교하여 이루어지는 지도학습모델인 것을 특징으로 하는 배터리 진단 방법.
- 제 1항에 있어서, 상기 비시계열 데이터는,적어도 둘 이상의 서로 다른 주파수의 정현파 신호를 상기 배터리에 인가하여 측정한 임피던스 정보인 것을 특징으로 하는 배터리 진단 방법.
- 제 1항에 있어서,상기 배터리예측모델의 하이퍼파라미터를 최적화하는 단계;를 더 포함하고,상기 최적화하는 단계는,기 정의된 학습데이터와 기 정의된 복수의 후보 하이퍼파라미터를 이용하여 상기 배터리예측모델을 학습시키는 단계;각각의 후보 하이퍼파라미터를 기반으로 학습된 배터리예측모델의 적합도평가를 기초로 상위 적어도 둘 이상의 후보 하이퍼파라미터를 부모 하이퍼파라미터로 선택하는 단계;상기 부모 하이퍼파라미터 사이의 교차 또는 변형을 통해 새로운 후보 하이퍼파라미터를 적어도 하나 이상 생성하는 단계;새로운 후보 하이퍼파라미터를 기반으로 학습된 배터리예측모델의 적합도평가를 기초로 다시 부모 하이퍼파라미터를 선택하고 새로운 후보 하이퍼파라미터를 생성하는 과정을 기 정의된 조건 또는 횟수까지 반복수행하는 단계;를 포함하는 것을 특징으로 하는 배터리 진단 방법.
- 배터리 상태정보를 예측하도록 학습된 배터리예측모델을 구동하는 컴퓨팅 장치가 수행하는 배터리 진단 방법에 있어서,일정 기간 동안 측정된 배터리의 전압, 전류 및 온도 중 적어도 하나 이상을 포함하는 시계열 데이터를 입력받는 제1 입력부;일정 시점에 측정한 배터리 임피던스를 포함하는 비시계열 데이터를 입력받는 제2 입력부; 및상기 시계열 데이터와 상기 비시계열 데이터를 상기 배터리예측모델에 입력하여 배터리 상태정보를 예측하는 예측부;를 포함하는 것을 특징으로 하는 배터리 진단 방법.
- 제 7항에 있어서, 상기 예측부는,상기 시계열 데이터를 상기 배터리예측모델의 인공신경망에 입력하여 제1 특징값을 생성하고, 상기 인공신경망의 제1 출력값과 상기 배터리의 임피던스를 포함하는 비시계열 데이터를 결합한 제2 특징값을 생성하고, 상기 제2 특징값을 상기 배터리예측모델의 완전연결계층에 입력하여 배터리 상태정보를 출력하는 것을 특징으로 하는 배터리진단장치.
- 제 1항에 기재된 방법을 수행하기 위한 컴퓨터 프로그램을 기록한 컴퓨터로 읽을 수 있는 기록매체.
Priority Applications (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| EP22843990.7A EP4417986A4 (en) | 2021-10-14 | 2022-09-07 | BATTERY DIAGNOSTIC METHOD AND ASSOCIATED DEVICE |
| CN202280006218.0A CN116457674A (zh) | 2021-10-14 | 2022-09-07 | 电池诊断方法以及其装置 |
| US18/161,903 US11879945B2 (en) | 2021-10-14 | 2023-01-31 | Battery diagnosis method and battery diagnosis apparatus |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| KR1020210136823A KR102424916B1 (ko) | 2021-10-14 | 2021-10-14 | 배터리 진단 방법 및 그 장치 |
| KR10-2021-0136823 | 2021-10-14 |
Related Child Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| US18/161,903 Continuation US11879945B2 (en) | 2021-10-14 | 2023-01-31 | Battery diagnosis method and battery diagnosis apparatus |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2023063587A1 true WO2023063587A1 (ko) | 2023-04-20 |
Family
ID=82606038
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/KR2022/013472 Ceased WO2023063587A1 (ko) | 2021-10-14 | 2022-09-07 | 배터리 진단 방법 및 그 장치 |
Country Status (5)
| Country | Link |
|---|---|
| US (1) | US11879945B2 (ko) |
| EP (1) | EP4417986A4 (ko) |
| KR (1) | KR102424916B1 (ko) |
| CN (1) | CN116457674A (ko) |
| WO (1) | WO2023063587A1 (ko) |
Families Citing this family (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR102424916B1 (ko) | 2021-10-14 | 2022-07-22 | 모나일렉트릭 주식회사 | 배터리 진단 방법 및 그 장치 |
| DE102022120558A1 (de) * | 2022-08-16 | 2024-02-22 | Bayerische Motoren Werke Aktiengesellschaft | Verfahren zum Überwachen von Batteriezellen einer Batterie eines Kraftfahrzeugs, Computerprogramm, Datenverarbeitungsvorrichtung und Kraftfahrzeug |
| KR102522683B1 (ko) * | 2022-08-16 | 2023-04-20 | 모나 주식회사 | 배터리진단방법 및 그 장치 |
| CN118112432A (zh) * | 2022-11-30 | 2024-05-31 | 通用汽车环球科技运作有限责任公司 | 根据电池组浸泡行为进行的预测性电池健康状况检测 |
| KR102868169B1 (ko) * | 2023-01-03 | 2025-10-01 | 주식회사 스칼라웍스 | 딥러닝 모델을 활용한 설비 장애 예측 장치 |
| KR102682629B1 (ko) * | 2023-07-19 | 2024-07-08 | 주식회사 엘와이솔루션 | 배터리의 상태를 실시간으로 모니터링하는 방법 및 장치 |
| KR102635946B1 (ko) * | 2023-08-25 | 2024-02-13 | 서울시립대학교 산학협력단 | 이상 데이터 보정 장치 및 방법 |
| KR102881214B1 (ko) * | 2024-01-22 | 2025-11-05 | 주식회사 배러머신 | 다변형 입력벡터를 수용하는 배터리 상태 추정용 dnn 모델 생성 방법 |
| KR20250132228A (ko) * | 2024-02-28 | 2025-09-04 | 삼성에스디아이 주식회사 | 전지 진단 장치 및 그 방법 |
| CN117878360B (zh) * | 2024-03-12 | 2024-05-28 | 长春师范大学 | 一种燃料电池效率预测方法及系统 |
| WO2026079714A1 (ko) * | 2024-10-08 | 2026-04-16 | 주식회사 엘지에너지솔루션 | 배터리 관리 장치 및 방법 |
| CN119805235B (zh) * | 2024-12-19 | 2026-03-27 | 上海科技大学 | 基于图像化单循环采样量的电池寿命预测方法和相关装置 |
| CN120044340B (zh) * | 2025-04-23 | 2025-08-15 | 广东森旭通用设备科技有限公司 | 一种基于人工智能的低压柔直调压器故障诊断方法 |
| CN121350544A (zh) * | 2025-12-19 | 2026-01-16 | 深圳市仕威新能源有限公司 | 一种新能源汽车锂电池状态在线预测方法、设备及介质 |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20190022905A (ko) * | 2016-07-22 | 2019-03-06 | 이오에스 에너지 스토리지 엘엘씨 | 배터리 관리 시스템 |
| EP3044601B1 (fr) * | 2013-09-11 | 2019-12-11 | Commissariat à l'Énergie Atomique et aux Énergies Alternatives | Procede, dispositif et systeme d'estimation de l'etat de charge d'une batterie |
| KR20200119383A (ko) * | 2019-03-26 | 2020-10-20 | 서강대학교산학협력단 | 인공 지능에 기반하여 배터리의 상태를 추정하는 장치 및 방법 |
| JP2021108246A (ja) * | 2019-12-27 | 2021-07-29 | 東洋システム株式会社 | 模擬電池構築方法および模擬電池構築装置 |
| KR102424916B1 (ko) * | 2021-10-14 | 2022-07-22 | 모나일렉트릭 주식회사 | 배터리 진단 방법 및 그 장치 |
Family Cites Families (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE10012964A1 (de) * | 2000-03-16 | 2001-10-04 | Implex Hear Tech Ag | Vorrichtung und Verfahren zum Betreiben eines wiederaufladbaren Speichers für elektrische Energie |
| US20030184307A1 (en) * | 2002-02-19 | 2003-10-02 | Kozlowski James D. | Model-based predictive diagnostic tool for primary and secondary batteries |
| JP3935099B2 (ja) * | 2003-04-15 | 2007-06-20 | 株式会社デンソー | 車両用蓄電装置の内部状態検出システム |
| GB2444511B (en) * | 2006-12-06 | 2008-10-22 | Iti Scotland Ltd | Battery Management System |
| US9177466B2 (en) * | 2011-01-20 | 2015-11-03 | Indiana University Research And Technology Corporation | Advanced battery early warning and monitoring system |
| US20220188700A1 (en) * | 2014-09-26 | 2022-06-16 | Bombora, Inc. | Distributed machine learning hyperparameter optimization |
| US11691518B2 (en) * | 2017-07-21 | 2023-07-04 | Quantumscape Battery, Inc. | Predictive model for estimating battery states |
| JP6973334B2 (ja) | 2018-08-30 | 2021-11-24 | トヨタ自動車株式会社 | 二次電池の劣化状態推定方法および二次電池システム |
| KR102171199B1 (ko) | 2018-12-27 | 2020-10-28 | 광운대학교 산학협력단 | 배터리의 충전 상태값을 추정하는 방법 및 배터리 관리 장치 |
| US11119494B2 (en) * | 2019-01-07 | 2021-09-14 | Wing Aviation Llc | Using machine learning techniques to estimate available energy for vehicles |
| KR20210077948A (ko) | 2019-12-18 | 2021-06-28 | 주식회사 어스텍 | 배터리의 soh 진단 방법 및 장치 |
| KR102140632B1 (ko) | 2020-04-21 | 2020-08-03 | (주) 한일프로텍 | 배터리의 상태를 진단하기 위한 장치 |
| JP7414637B2 (ja) * | 2020-05-19 | 2024-01-16 | 本田技研工業株式会社 | 制御装置、制御システム、制御方法、及びプログラム |
| US11360155B2 (en) * | 2020-09-03 | 2022-06-14 | Volkswagen Aktiengesellschaft | Battery state of health estimator |
| CN112883637A (zh) * | 2021-01-29 | 2021-06-01 | 南昌师范学院 | 锂离子电池rul的混沌时序非线性组合预测方法 |
| CN113093014B (zh) * | 2021-03-31 | 2022-05-27 | 山东建筑大学 | 一种基于阻抗参数的soh与soc的在线协同估计方法及系统 |
-
2021
- 2021-10-14 KR KR1020210136823A patent/KR102424916B1/ko active Active
-
2022
- 2022-09-07 CN CN202280006218.0A patent/CN116457674A/zh active Pending
- 2022-09-07 EP EP22843990.7A patent/EP4417986A4/en active Pending
- 2022-09-07 WO PCT/KR2022/013472 patent/WO2023063587A1/ko not_active Ceased
-
2023
- 2023-01-31 US US18/161,903 patent/US11879945B2/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP3044601B1 (fr) * | 2013-09-11 | 2019-12-11 | Commissariat à l'Énergie Atomique et aux Énergies Alternatives | Procede, dispositif et systeme d'estimation de l'etat de charge d'une batterie |
| KR20190022905A (ko) * | 2016-07-22 | 2019-03-06 | 이오에스 에너지 스토리지 엘엘씨 | 배터리 관리 시스템 |
| KR20200119383A (ko) * | 2019-03-26 | 2020-10-20 | 서강대학교산학협력단 | 인공 지능에 기반하여 배터리의 상태를 추정하는 장치 및 방법 |
| JP2021108246A (ja) * | 2019-12-27 | 2021-07-29 | 東洋システム株式会社 | 模擬電池構築方法および模擬電池構築装置 |
| KR102424916B1 (ko) * | 2021-10-14 | 2022-07-22 | 모나일렉트릭 주식회사 | 배터리 진단 방법 및 그 장치 |
Non-Patent Citations (1)
| Title |
|---|
| See also references of EP4417986A4 * |
Also Published As
| Publication number | Publication date |
|---|---|
| EP4417986A1 (en) | 2024-08-21 |
| US20230168306A1 (en) | 2023-06-01 |
| EP4417986A4 (en) | 2025-08-06 |
| CN116457674A (zh) | 2023-07-18 |
| US11879945B2 (en) | 2024-01-23 |
| KR102424916B1 (ko) | 2022-07-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2023063587A1 (ko) | 배터리 진단 방법 및 그 장치 | |
| WO2022034983A1 (ko) | 신경망 기반의 배터리 셀 불량 및 화재 사전 진단 방법 및 장치 | |
| WO2022035151A1 (ko) | 배터리 이상 진단 장치 및 방법 | |
| WO2022163996A1 (ko) | 자기주의 기반 심층 신경망 모델을 이용한 약물-표적 상호작용 예측 장치 및 그 방법 | |
| WO2021015501A1 (ko) | 배터리 저항 진단 장치 및 방법 | |
| KR102870576B1 (ko) | 배터리 진단 장치, 배터리 시스템 및 배터리 진단 방법 | |
| WO2021201387A1 (ko) | 신경망 기반의 배터리 용량 추정 방법 및 장치 | |
| WO2022035149A1 (ko) | 배터리 이상 진단 장치 및 방법 | |
| WO2019182166A1 (ko) | 변형된 mfcc를 이용한 전기부하 판별 장치 및 방법 | |
| WO2023120777A1 (ko) | 심전도 전역 특징 벡터를 이용한 다중 박동 검출 방법 및 장치 | |
| CN116540167A (zh) | 基于变比测试仪的电流互感器故障监测方法及系统 | |
| Wang et al. | Real-time diagnosis based on signal convolution-pooling processing and shared filter learning for transistor open-circuit faults in a T-type inverter | |
| CN118626990A (zh) | 一种硬盘故障预测方法、装置、设备及存储介质 | |
| WO2024144017A1 (ko) | 임피던스 기반 배터리 불량 진단 장치 및 방법 | |
| CN113138894A (zh) | 基于电力参数监测与屏幕信息识别的实验设备监控方法 | |
| KR102708169B1 (ko) | 신경망 기반의 배터리 사용량 예측 방법 및 장치 | |
| CN112731184A (zh) | 一种电池寿命检测方法和系统 | |
| CN120123872B (zh) | 一种基于改进迁移轻梯度提升机的充电模块开路故障快速诊断方法及系统 | |
| CN118839234A (zh) | 全信息耦合负荷特征图像与轻量化FBNet分类方法及系统 | |
| WO2021025295A1 (ko) | 배터리 관리 장치, 배터리 관리 방법 및 배터리 팩 | |
| CN118196921A (zh) | 巡检采集系统、巡检质量评价方法、装置及服务器 | |
| CN117650626A (zh) | 一种基于双lstm网络的直流微电网大信号稳定性预测与控制方法 | |
| WO2023075317A1 (ko) | 자돈 상태 분석 방법 및 장치 | |
| WO2021100968A1 (ko) | 인공지능을 이용한 태아 박동 진단방법, 장치 및 시스템 | |
| Sonawane | Abetalipoproteinemia |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| WWE | Wipo information: entry into national phase |
Ref document number: 202280006218.0 Country of ref document: CN |
|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 22843990 Country of ref document: EP Kind code of ref document: A1 |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 2022843990 Country of ref document: EP |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| ENP | Entry into the national phase |
Ref document number: 2022843990 Country of ref document: EP Effective date: 20240514 |