Battery life prediction device and operation method thereof

Abstract

A battery life prediction device capable of predicting the remaining life of a battery more quickly and accurately in response to input of a complex battery usage pattern, the battery life prediction device including: a data generation unit that generates remaining life data corresponding to each of a plurality of usage pattern information related to a battery; and a battery life prediction unit that classifies the plurality of usage patterns into at least one class using a trained artificial neural network model, determines a representative usage pattern from among the usage patterns belonging to each class, and predicts the life of the battery based on the remaining life data corresponding to the representative usage pattern.

Description

[Technical field]

[0001] [CROSS REFERENCE TO RELATED APPLICATIONS] The present invention claims the benefit of priority based on Korean Patent Application No. 10-2022-0034315, filed on March 18, 2022, and all contents disclosed in the documents of the Korean patent application are incorporated herein by reference.

[0002] SUMMARY OF THE DISCLOSURE The embodiments disclosed herein relate to a battery life prediction device and method of operation. [Background technology]

[0003] An energy storage system (ESS) is a system that stores electrical energy and manages it so that it can be used when needed. In general, power storage systems installed in power plants that operate large-scale power grids or buildings with large power consumption include hundreds to thousands of battery cells. It is important for users to predict and manage the replacement timing of batteries that make up the ESS, and for manufacturers to ensure quality, to accurately predict the remaining lifespan of the batteries.

[0004] Conventional battery life prediction methods mainly use cell-based modeling, which predicts remaining life using information on individual battery cells (current, voltage, temperature, etc.) measured by sensors and a corresponding estimated life table. This method can predict the life relatively accurately for a single usage pattern, but the accuracy of prediction decreases when multiple usage patterns are input. In addition, it takes a lot of time because the life must be recalculated every time the usage pattern changes, and when a deterioration model with a standard SOH set to 100% is applied, the battery life is already reduced at the time of the change in usage pattern, so the prediction error is bound to become even larger. Summary of the Invention [Problem to be solved by the invention]

[0005] An objective of the embodiments disclosed herein is to provide an apparatus and a prediction method that can more quickly and accurately predict the remaining life of a battery for a complex battery usage pattern input.

[0006] The technical problems of the embodiments disclosed in this document are not limited to the technical problems mentioned above, and other technical problems not mentioned will be clearly understood by those skilled in the art from the following description. [Means for solving the problem]

[0007] A battery life prediction device according to one embodiment includes a data generation unit that generates remaining life data corresponding to each of a plurality of usage pattern information related to a battery, and a battery life prediction unit that classifies the plurality of usage patterns into at least one class using a trained artificial neural network model, determines a representative usage pattern from among the usage patterns belonging to each class, and predicts the battery life based on the remaining life data corresponding to the representative usage pattern.

[0008] The battery life prediction device according to an embodiment may further include an artificial neural network model training unit that trains an artificial neural network model to classify an inputted usage pattern into at least one class.

[0009] In one embodiment of a battery life prediction device, the artificial neural network model can calculate a similarity between an input usage pattern and a representative usage pattern, and determine a class of the usage pattern based on the similarity.

[0010] In an embodiment of the battery life prediction device, the artificial neural network model can calculate a similarity based on the battery life deduction rate in a specific period of an inputted usage pattern and a representative usage pattern.

[0011] In the battery life prediction device according to an embodiment, the battery life prediction unit may determine a usage pattern that has the greatest effect on the remaining life of the battery from among a plurality of usage patterns belonging to each class as a representative usage pattern.

[0012] In one embodiment of a battery life prediction device, a battery life prediction unit can predict the battery life based on remaining life data corresponding to a representative usage pattern of a class to which the actual usage pattern belongs, instead of remaining life data corresponding to an actual usage pattern of the battery.

[0013] In one embodiment of a battery life prediction device, a battery life prediction unit converts first data indicating the remaining life for a usage time of a representative usage pattern into second data indicating the time required to reach the remaining life, and when the battery usage pattern is changed, can calculate the remaining life for the change in the battery usage pattern based on the second data of the representative usage pattern of each class to which the usage pattern before the change and the usage pattern after the change belong.

[0014] In the battery life prediction device according to an embodiment, the usage pattern information may include at least one of the battery's accumulated discharge energy, a State of Charge (SOC), a Depth of Discharge (DOD), a C rate, and a CP rate.

[0015] An operating method of a battery life prediction device according to one embodiment includes the steps of generating remaining life data corresponding to each of a plurality of usage pattern information related to a battery, classifying the plurality of usage patterns into at least one class using a trained artificial neural network model, determining a representative usage pattern among the usage patterns belonging to each class, and predicting the battery life based on the remaining life data corresponding to the representative usage pattern.

[0016] The method of operating a battery life prediction device according to an embodiment may further include training an artificial neural network model to classify input usage patterns into at least one class.

[0017] In an operating method of a battery life prediction device according to one embodiment, the artificial neural network model can calculate a similarity between an input usage pattern and a representative usage pattern, and determine a class of the usage pattern based on the similarity.

[0018] In an operating method of a battery life prediction device according to an embodiment, the artificial neural network model can calculate a similarity based on a battery life deduction rate in a specific period of an inputted usage pattern and a representative usage pattern.

[0019] In an operating method of a battery life prediction device according to one embodiment, the step of determining a representative usage pattern may include a step of determining a usage pattern that has the greatest impact on the remaining life of the battery as the representative usage pattern among a plurality of usage patterns belonging to each class.

[0020] In an operating method of a battery life prediction device according to one embodiment, in a step of predicting the life of a battery, the life of the battery can be predicted based on remaining life data corresponding to a representative usage pattern of a class to which the actual usage pattern belongs, instead of remaining life data corresponding to an actual usage pattern of the battery.

[0021] In an operating method of a battery life prediction device according to one embodiment, the step of predicting the battery life may include a step of converting first data indicating the remaining life with respect to the usage time of a representative usage pattern into second data indicating the time required to reach the remaining life, and when the battery usage pattern is changed, a step of calculating the remaining life with respect to the change in the battery usage pattern based on the second data of the representative usage pattern of each class to which the usage pattern before the change and the usage pattern after the change belong.

[0022] In an operating method of a battery life prediction device according to an embodiment, the usage pattern information may include at least one of the battery's accumulated discharge energy, a State of Charge (SOC), a Depth of Discharge (DOD), a C rate, and a CP rate. Effect of the Invention

[0023] According to one embodiment, an apparatus and method are provided that can more quickly and accurately predict the remaining life of a battery for a complex battery usage pattern input.

[0024] According to the proposed device, the time required for lifespan prediction can be reduced by classifying multiple usage patterns according to their similarity using an artificial neural network model and applying the lifespan deduction rate of the representative usage pattern of each group instead of the actual usage pattern. Also, unlike existing methods in which the accuracy of lifespan prediction decreases as the battery deterioration progresses, the proposed device can accurately predict the lifespan of a battery that has already deteriorated even if the usage pattern changes.

[0025] In addition, various other effects may be provided that can be understood directly or indirectly through this document.

[0026] In order to more clearly describe the embodiments disclosed in this document or the technical solutions of the prior art, the drawings necessary for the description of the embodiments are briefly introduced below. The drawings below should be understood as being for the purpose of describing the embodiments of the present specification only, and not for the purpose of limitation. In addition, for the sake of clarity, the expressions of some components in the drawings may be exaggerated or omitted. [Brief description of the drawings]

[0027] [Figure 1] 1 is a block diagram showing a configuration of a battery life prediction device according to an embodiment; [Diagram 2] 1 is a graph showing a comparison of results obtained by applying a battery life prediction method according to a conventional technology and a life prediction method according to an embodiment. [Diagram 3] 11 is a graph illustrating a result of verifying accuracy of classifying input usage patterns using an artificial neural network model according to an embodiment. [Figure 4] 1 is a graph illustrating the results of verifying the accuracy of battery life prediction using an artificial neural network model according to one embodiment. [Figure 5a] FIG. 13 is a diagram showing a table for predicting battery life using usage time-remaining life data for each class of battery usage pattern. [Figure 5b] FIG. 13 is a diagram showing a table for predicting battery life using usage time-remaining life data for each class of battery usage pattern. [Figure 6] 4 is a flowchart illustrating an operation method of a battery life prediction device according to an embodiment. [Figure 7] FIG. 1 illustrates a computer system that performs a battery life prediction method according to one embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0028] Hereinafter, the embodiments disclosed in this document will be described in detail with reference to the exemplary drawings. When adding reference numerals to components in each drawing, it should be noted that the same components are given the same numerals as much as possible even if they are displayed in different drawings. In addition, when describing the embodiments disclosed in this document, if a detailed description of related known configurations or functions is determined to hinder understanding of the embodiments disclosed in this document, the detailed description will be omitted.

[0029] The terms used in this document are currently selected as widely used as possible while taking into consideration the functions, but these may change depending on the intention or practice of the technical personnel in this field, or the emergence of new technology. In addition, in certain cases, the applicant may arbitrarily select a term, and in this case, the meaning will be described in the relevant description of the specification. Therefore, it is made clear that the terms used in this document should be interpreted based on the substantial meaning of the term and the overall content of this document, rather than simply the name of the term. In addition, the terms used in this document are used merely to describe a specific embodiment, and are not intended to limit the scope of other embodiments. A singular expression may include a plural expression unless the context clearly indicates otherwise.

[0030] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a preferred embodiment of a battery life prediction device and an operating method thereof will be described with reference to the accompanying drawings.

[0031] The components shown in the block diagrams are divided according to their functions and roles, and each block does not necessarily have to be realized as independent hardware or software. For example, the divided components may actually be realized as one device or program, or one component may be realized as a combination of multiple devices and programs. Furthermore, in this document, expressions such as "first" and "second" are used to distinguish components from each other, and do not imply a ranking or hierarchy between the components.

[0032] FIG. 1 is a block diagram showing the configuration of a battery life prediction device according to an embodiment.

[0033] As shown in FIG. 1, the battery life prediction device 10 may include a data generating unit 110, an artificial neural network model learning unit 120 and a battery life prediction unit 130.

[0034] The data generating unit 110 is configured to generate remaining life data corresponding to each of a plurality of pieces of information on battery usage patterns. The remaining life data includes remaining life information after a battery of the same specification is used in a specific usage pattern for a predetermined period (e.g., 24 hours), and can be expressed as a State of Health (SOH) value.

[0035] According to one embodiment, the data generating unit 110 may perform a step of calculating life factors (e.g., SOC, CP rate, module temperature, etc.) that affect the reduction of battery life in order to generate remaining life data for each usage pattern, and a step of normalizing different types of data by calculating appropriate ranges for each life factor (e.g., SOC 1 to 96, CP rate 0.1 to 0.5, module temperature 25 to 40° C., number of cycles 1 to 4, etc.).

[0036] For example, the data generating unit 110 may acquire battery usage pattern information for the past 15 years and generate a table in which more than 20,000 usage patterns and the remaining life (SOH) corresponding to each usage pattern are mapped according to a predetermined criterion. According to an embodiment, the usage pattern information may include at least one of the battery's accumulated discharge energy, SOC (State of Charge), DOD (Depth of Discharge), C rate, and CP rate.

[0037] The artificial neural network model learning unit 120 trains an artificial neural network model to classify the inputted usage pattern into at least one class according to a preset criterion. According to an embodiment, a deep feed-forward network (DFN) algorithm, which is a forward neural network model, may be used as the artificial neural network model, but is not limited thereto. The artificial neural network model learning unit 120 can learn data classified into each class based on the usage pattern information, and the battery life prediction unit 130 can predict the remaining life at a specific point in time according to a change in the usage pattern using the trained artificial neural network model.

[0038] According to an embodiment, the artificial neural network model may calculate a similarity between an inputted usage pattern and a representative usage pattern, and determine a class of the usage pattern based on the similarity. The similarity between the usage patterns may be calculated based on a battery life reduction rate in a specific period. For example, in determining the class of the inputted usage pattern, a class to which a representative usage pattern having a closest life reduction rate (or remaining life after use) belongs may be selected based on a result of analyzing battery usage information in a specific year. That is, unlike the existing method of predicting life based on sensing information and a deterioration model of a battery, the remaining life may be predicted by analyzing the similarity between usage patterns based on a large amount of data and applying the life reduction rate of the most similar representative pattern.

[0039] According to one embodiment, the artificial neural network model learning unit 120 may perform the steps of dividing a dataset generated for machine learning into a learning part, a validation part, and a testing part, estimating parameters with the highest model prediction accuracy based on pre-set hyperparameter criteria, extracting parameters that calculate the highest model accuracy index, and applying the calculated parameters to the artificial neural network model.

[0040] The battery life prediction unit 130 is configured to classify a plurality of usage patterns into at least one class using a trained artificial neural network model, determine a representative usage pattern from among the usage patterns belonging to each class, and predict the battery life based on remaining life data corresponding to the representative usage pattern.

[0041] As described above, an artificial neural network model (e.g., a DFN model) can calculate the similarity between usage patterns using the life deduction rate for a specific period, and based on this, classify the input usage pattern into multiple classes (e.g., class 1 to class 30).

[0042] When the class for each usage pattern is determined, the battery life prediction unit 130 may determine the usage pattern that has the greatest effect on the remaining battery life among the multiple usage patterns belonging to each class as the representative usage pattern. For example, assuming that usage patterns 1 to 5 belong to class 1 and usage patterns 6 to 10 belong to class 2, the usage pattern with the largest battery life deduction rate among usage patterns 1 to 5 may be determined as the representative usage pattern of class 1, and the usage pattern with the largest battery life deduction rate among usage patterns 6 to 10 may be determined as the representative usage pattern of class 2. The battery life deduction rate of each usage pattern may be obtained from a table of remaining life data for each usage pattern generated by the data generation unit 110.

[0043] When the representative usage pattern is determined for each class, the battery life prediction unit 130 can predict the battery life based on the remaining life data corresponding to the representative usage pattern of the class to which the actual usage pattern belongs, instead of the remaining life data corresponding to the actual usage pattern. For example, assuming that usage patterns 1 to 5 belong to class 1 and the representative usage pattern of the class is usage pattern 2, the battery life can be predicted by applying the life deduction rate of representative usage pattern 2 instead of the actual usage pattern 1 of the input battery.

[0044] As the usage patterns of ESS batteries become more diverse, if multiple usage patterns are applied one by one, the remaining life corresponding to tens of thousands of usage patterns must be calculated, which makes the calculation complicated and time-consuming. As in the embodiment, by applying a deduction rate of a representative usage pattern of the class to which the battery belongs instead of an actual usage pattern, the life can be predicted more quickly.

[0045] FIG. 2 is a graph showing a comparison of the results of applying a battery life prediction method according to a conventional technology and the life prediction method according to an embodiment.

[0046] In a simulation to compare the results, the capacity degradation rate over the course of battery usage was calculated for three cases (Case 1, Case 3, and Case 6) each including one or more usage patterns, using a lifespan prediction method (Sizing tool) based on a conventional deterioration model and a lifespan prediction method (IFC tool) of the embodiment. Table 1 below shows the error between the deterioration rate (Rs) calculated by the conventional lifespan prediction method and the deterioration rate (Rf) calculated by the lifespan prediction method of the embodiment over the course of battery usage. [Table 1] As shown in the graph of FIG. 2 and Table 1, although the error may increase slightly as the usage period passes, it can be seen that the life prediction results are generally almost the same for each case (Case 1, Case 3, Case 6). This result shows that although the prediction model of the embodiment using class classification and representative usage patterns has a faster calculation speed than the conventional technology, there is no significant difference in the prediction results. The more complex and numerous usage patterns are applied, the greater the effect of reducing time costs by this technology.

[0047] For example, the battery life can be predicted based on remaining life data corresponding to a representative usage pattern of a class to which the actual usage pattern belongs, instead of remaining life data corresponding to an actual usage pattern of the battery. For example, assuming that usage patterns 1 to 5 belong to class 1 and the representative usage pattern of the class is usage pattern 2, the battery life can be predicted by applying a life deduction rate of representative usage pattern 2 instead of actual usage pattern 1 of the input battery.

[0048] In this way, by applying the deduction rate of the representative usage pattern of the class to which the device belongs instead of the actual usage pattern, it is possible to predict the lifespan more quickly. The comparison of the lifespan prediction results according to the method of the embodiment and the method of the prior art has been described above with reference to Figures 2 to 4.

[0049] According to one embodiment, the step of predicting the battery life (S650) may include a step of converting first data indicating the remaining life with respect to the usage time of a representative usage pattern into second data indicating the time required to reach the remaining life, and, if the usage pattern of the battery is changed, a step of calculating the remaining life with respect to the change in the usage pattern of the battery based on the second data of the representative usage pattern of each class to which the usage pattern before the change and the usage pattern after the change belong.

[0050] According to the present embodiment, unlike the conventional technology that can perform accurate prediction only in the initial state where the battery is not degraded, it is possible to predict a more accurate deterioration curve when there is a change in the usage pattern even if the battery has already degraded. An exemplary method for predicting the battery life using the usage time-remaining life data by class has been described above with reference to FIG.

[0051] The method of operating the battery life prediction device according to the embodiment may be implemented in the form of program instructions that can be executed by various computer components, such as an application, and may be recorded on a computer-readable recording medium. The computer-readable recording medium may include program instructions, data files, data structures, and the like, either alone or in combination.

[0052] FIG. 7 illustrates a computer system that performs a battery life prediction method according to one embodiment.

[0053] As shown in FIG. 7 , a computing device 200 according to an embodiment disclosed in this document may include an MCU 220 , a memory 210 , an input / output I / F 230 and a communication I / F 240 .

[0054] The MCU 220 may be a processor that executes various programs stored in the memory 210, predicts the remaining life of the battery cells based on the usage pattern through such programs, and performs the functions of the battery life prediction device 10 described with reference to Figures 1 to 5, or a processor that performs the operating method of the battery life prediction device described with reference to Figure 6.

[0055] The memory 210 can store various programs related to battery life prediction (e.g., a usage pattern-remaining life data generation program, an artificial neural network model learning program, a battery life prediction program, etc.). The memory 210 can also store various data generated by the data generation unit 110.

[0056] A plurality of such memories 210 may be provided as necessary. The memories 210 may be volatile memories or non-volatile memories. The volatile memories 210 may be RAM, DRAM, SRAM, etc. The non-volatile memories 210 may be ROM, PROM, EAROM, EPROM, EEPROM, flash memory, etc. The listed examples of the memories 210 are merely illustrative and are not limited to these examples.

[0057] The input / output I / F 230 can provide an interface that connects input devices (not shown) such as a keyboard, mouse, or touch panel and output devices such as a display (not shown) to the MCU 220 to transmit and receive data.

[0058] The communication I / F 240 is configured to be capable of transmitting and receiving various data to and from a server, and may be any device capable of supporting wired or wireless communication. For example, a program or various data for predicting the remaining life of a battery or learning an artificial neural network may be transmitted and received from a separately provided external server via the communication I / F 240.

[0059] Thus, a battery management method according to one embodiment disclosed herein may be stored in memory 210 and executed by MCU 220.

[0060] According to the above embodiment, the time required for life prediction can be reduced by classifying a plurality of usage patterns according to their similarity using an artificial neural network model and applying a life deduction rate of a representative usage pattern of each group instead of an actual usage pattern. Also, unlike the existing method in which the accuracy of life prediction decreases as the deterioration of a battery progresses, the life prediction can be accurately performed even when the usage pattern of a battery that has already deteriorated changes.

[0061] In the above, all components constituting the embodiment have been described as being combined or operating in combination, but the present invention is not necessarily limited to such an embodiment, and all components may be selectively combined and operated in one or more combinations within the scope of the purpose. In addition, the terms "include," "constitute," "have," and the like described above mean that the relevant component may be contained therein, unless otherwise specified to the contrary, and therefore should be interpreted as not excluding other components, but as including other components.

[0062] The above description is merely an illustrative example of the technical ideas disclosed in this document, and a person having ordinary skill in the art to which the embodiments disclosed in this document pertains may make various modifications and variations thereto without departing from the essential characteristics of the embodiments disclosed in this document.

[0063] Therefore, the embodiments disclosed in this document are intended to explain, not to limit, the technical ideas disclosed in this document, and such embodiments do not limit the scope of the technical ideas disclosed in this document. The scope of protection of the technical ideas disclosed in this document should be interpreted according to the claims below, and all technical ideas within the scope equivalent thereto should be interpreted as being included in the scope of rights of this document. [Explanation of symbols]

[0064] 10 Battery life prediction device 110 Data Generation Division 120 Artificial Neural Network Model Learning Unit 130 Battery Life Prediction Unit 200 Computer Equipment 220 Memory

Claims

1. a data generating unit that generates remaining life data corresponding to each of a plurality of pieces of information on usage patterns of the battery; and a battery life prediction unit that classifies the multiple usage patterns into at least one class using a trained artificial neural network model, determines a representative usage pattern from among the usage patterns belonging to each class, and predicts the battery life based on remaining life data corresponding to the representative usage pattern.

2. 2. The battery life prediction device according to claim 1, further comprising an artificial neural network model training unit for training the artificial neural network model to classify an inputted usage pattern into the at least one class.

3. 3. The battery life prediction device of claim 2, wherein the artificial neural network model calculates a similarity between the input usage pattern and the representative usage pattern, and determines a class of the usage pattern based on the similarity.

4. The battery life prediction device according to claim 3 , wherein the artificial neural network model calculates the similarity based on a battery life deduction rate in a specific period of the input usage pattern and the representative usage pattern.

5. The battery life prediction device according to claim 1 , wherein the battery life prediction unit determines a usage pattern that has the greatest effect on the remaining life of the battery from among a plurality of usage patterns belonging to each class as a representative usage pattern.

6. The battery life prediction device of claim 5, characterized in that the battery life prediction unit predicts the battery life based on remaining life data corresponding to a representative usage pattern of a class to which the actual usage pattern belongs, instead of remaining life data corresponding to an actual usage pattern of the battery.

7. the battery life prediction unit converts first data indicating a remaining life with respect to a usage time of the representative usage pattern into second data indicating a time required to reach the remaining life, The battery life prediction device of claim 6, characterized in that when the usage pattern of the battery is changed, the remaining life for the change in the usage pattern of the battery is calculated based on second data of the representative usage pattern of each class to which the usage pattern before the change and the usage pattern after the change belong.

8. 2. The battery life prediction device according to claim 1, wherein the information on the usage pattern includes at least one of an accumulated discharge energy, a State of Charge (SOC), a Depth of Discharge (DOD), a C rate, and a CP rate of the battery.

9. generating remaining life data corresponding to each of a plurality of usage pattern information relating to the battery; classifying the plurality of usage patterns into at least one class using a trained artificial neural network model; determining a representative usage pattern among the usage patterns belonging to each class; and A method of operating a battery life prediction apparatus, comprising predicting a life of the battery based on remaining life data corresponding to the representative usage pattern.

10. 10. The method of claim 9, further comprising training the artificial neural network model to classify input usage patterns into at least one class.

11. The method of claim 10, wherein the artificial neural network model calculates a similarity between the input usage pattern and the representative usage pattern, and determines a class of the usage pattern based on the similarity.

12. The method of claim 11, wherein the artificial neural network model calculates the similarity based on a battery life deduction rate in a specific period of the input usage pattern and the representative usage pattern.

13. 10. The method of claim 9, wherein the step of determining the representative usage pattern includes a step of determining, as the representative usage pattern, a usage pattern that has the greatest effect on the remaining life of the battery among a plurality of usage patterns belonging to each class.

14. 14. The method of claim 13, wherein in the step of predicting the life of the battery, the life of the battery is predicted based on remaining life data corresponding to a representative usage pattern of a class to which the actual usage pattern belongs, instead of remaining life data corresponding to an actual usage pattern of the battery.

15. The step of predicting the life of the battery includes: converting first data indicating a remaining life with respect to a usage time of the representative usage pattern into second data indicating a time required to reach the remaining life; and The method for operating a battery life prediction device as described in claim 13, further comprising the step of calculating a remaining life for the change in the usage pattern of the battery based on second data of a representative usage pattern of each class to which the usage pattern before the change and the usage pattern after the change belong when the usage pattern of the battery is changed.

16. 10. The method of claim 9, wherein the information on the usage pattern includes at least one of an accumulated discharge energy, a State of Charge (SOC), a Depth of Discharge (DOD), a C rate, and a CP rate of the battery.

Images