Data processing device and method for analyzing battery manufacturing process

Abstract

A data processing apparatus for analyzing a battery manufacturing process according to one embodiment of the present invention may comprise: at least one processor; and a memory configured to store at least one instruction executed by the at least one processor. Herein, the at least one instruction may include: an instruction for collecting process data for each process factor of the plurality of batteries; instructions for constructing a machine learning-based performance prediction model using the process data for each process factor, the performance prediction model for predicting battery performance; instructions to generate analysis information indicative of an impact of one or more process factors on a performance predictor of the performance prediction model; and instructions for outputting the generated analysis information through a predefined graphical user interface (GUI).

Description

Technical Field

[0001] This application claims priority and benefit to Korean Patent Application No. 10-2023-0190095, filed on December 22, 2023, with the Korean Intellectual Property Office, the entire contents of which are incorporated herein by reference.

[0002] This invention relates to data processing apparatus and methods, and more particularly to data processing apparatus and methods for analyzing battery manufacturing processes. Background Technology

[0003] Rechargeable and reusable secondary batteries can be used as energy for small devices such as mobile phones, tablet PCs and vacuum cleaners, as well as for medium and large devices such as automobiles and energy storage systems (ESS) for smart grids.

[0004] Batteries can be manufactured by sequentially performing electrode manufacturing, assembly, and activation processes. During and after the battery manufacturing process, various performance parameters of the battery—such as resistance, capacity, and charging time—are reviewed.

[0005] Because various process factors associated with each cell process affect battery performance, and these process factors have very complex correlations, it is difficult to identify which process factor has a significant impact on battery performance, and how much the corresponding process factor affects battery performance.

[0006] As a technology to solve the above problems, an appropriate data processing procedure is needed, which can identify process factors that affect battery performance by using data related to the battery manufacturing process, and analyze the impact of the corresponding process factors. Summary of the Invention

[0007] Technical issues

[0008] To eliminate one or more problems in related technologies, embodiments of this disclosure provide a data processing apparatus for analyzing the battery manufacturing process.

[0009] To eliminate one or more problems of the related technologies, embodiments of this disclosure also provide a data processing method performed by a data processing apparatus.

[0010] Technical solution

[0011] To achieve the purposes of this disclosure, a data processing apparatus for analyzing a battery manufacturing process may include: at least one processor; and a memory configured to store at least one instruction executed by the at least one processor.

[0012] Here, at least one instruction may include: instructions for collecting process data for each process factor of multiple batteries; instructions for using the process data for each process factor to construct a machine learning-based performance prediction model for predicting battery performance; instructions for generating analytical information indicating the impact of one or more process factors on the performance prediction value of the performance prediction model; and instructions for outputting the generated analytical information through a predefined graphical user interface (GUI).

[0013] Here, process data may include data related to one or more unit processes among electrode coating process, electrode rolling process, assembly process, activation process and end-of-line (EOL) process.

[0014] Instructions for building machine learning-based performance prediction models may include instructions for removing process data for one or more process factors among a plurality of process factors based on one of the relevance and importance of those process factors.

[0015] Instructions for building a machine learning-based performance prediction model may include instructions for building a performance prediction model that outputs at least one predicted value for one or more performance factors, including battery capacity and resistance, by using process data for each process factor as learning data.

[0016] Instructions for building machine learning-based performance prediction models may include: instructions for generating multiple performance prediction models by applying different learning algorithms; and instructions for selecting the performance prediction model that exhibits the best performance among the performance prediction models based on the difference between the performance prediction value and the performance measurement value of each performance prediction model.

[0017] Instructions for generating analytical information may include instructions for calculating the influence index of each process factor among the process factors based on changes in performance prediction values ​​according to changes in individual process factors.

[0018] Instructions for outputting analytical information via a predefined GUI may include instructions for visualizing and outputting one or more of the length, size, orientation, and color of objects corresponding to process factors so as to correspond to the influence index of the corresponding process factor.

[0019] Instructions for outputting analytical information via a predefined GUI may include instructions for selecting the top N process factors with high influence indices (N is a natural number greater than or equal to 2); and instructions for visualizing and outputting objects corresponding to each of the N process factors so as to correspond to the respective influence index of the objects.

[0020] Instructions for outputting analytical information via a predefined GUI may include: instructions for grouping multiple process factors based on unit processes; instructions for summing the influence indices of the process factors included in each unit process; and instructions for visualizing and outputting objects corresponding to each unit process such that the objects correspond to the influence indices of each unit process.

[0021] Instructions for outputting analytical information via a predefined GUI may include instructions for outputting changes in performance predictions based on variations in process factors as a two-dimensional graph.

[0022] According to another embodiment of this disclosure, a data processing method for analyzing a battery manufacturing process may include: collecting process data for each process factor of a plurality of batteries; using the process data for each process factor to construct a machine learning-based performance prediction model for predicting battery performance; generating analysis information indicating the impact of one or more process factors on the performance prediction value of the performance prediction model; and outputting the generated analysis information through a predefined graphical user interface (GUI).

[0023] Here, process data may include data related to one or more unit processes among electrode coating process, electrode rolling process, assembly process, activation process and end-of-line (EOL) process.

[0024] Building a machine learning-based performance prediction model can include removing process data for one or more process factors from a set of process factors based on either the relevance or importance of those factors.

[0025] Building a machine learning-based performance prediction model can include constructing a performance prediction model that outputs at least one predicted value for one or more performance factors, including battery capacity and resistance, by using process data for each process factor as learning data.

[0026] Building a machine learning-based performance prediction model can include generating multiple performance prediction models by applying different learning algorithms; and selecting the performance prediction model that exhibits the best performance among the performance prediction models based on the difference between the performance prediction value and the performance measurement value of each performance prediction model.

[0027] Generating analytical information can include calculating the impact index of each process factor among the process factors based on changes in performance predictions according to changes in individual process factors.

[0028] Outputting analytical information via a predefined GUI can include visualizing and outputting one or more of the length, size, orientation, and color of objects corresponding to process factors, so as to correspond to the influence index of the corresponding process factor.

[0029] Outputting analytical information via a predefined GUI can include selecting the top N process factors with high impact indices (N is a natural number greater than or equal to 2); and visualizing and outputting objects corresponding to each of the N process factors so as to correspond to the object's respective impact index.

[0030] Outputting analytical information via a predefined GUI can include grouping multiple process factors based on unit processes; summing the influence indices of the process factors included in each unit process; and visualizing and outputting objects corresponding to each unit process such that the objects correspond to the influence index of each unit process.

[0031] Outputting analytical information through a predefined GUI can include displaying changes in performance predictions based on variations in process factors as a two-dimensional graph.

[0032] Beneficial effects

[0033] According to embodiments of this disclosure, analytical information on process factors affecting battery performance can be provided, thereby supporting improvements in the battery manufacturing process. Attached Figure Description

[0034] Figure 1 The general battery manufacturing process is illustrated.

[0035] Figure 2 This is an operation flowchart of a data processing method according to an embodiment of the present invention.

[0036] Figure 3 This is an example of process data according to an embodiment of the present invention.

[0037] Figure 4 This is an operation flowchart of a method for constructing a performance prediction model according to an embodiment of the present invention.

[0038] Figure 5 This is a reference diagram for explaining a data preprocessing method according to an embodiment of the present invention.

[0039] Figure 6 This is an example screen of a user terminal used to interpret a GUI according to an embodiment of the present invention.

[0040] Figures 7 to 10 This is an example of a user terminal screen used for interpreting and analyzing information according to an embodiment of the present invention.

[0041] Figure 11 This is a block diagram of a data processing apparatus according to an embodiment of the present invention.

[0042] 610: Output Item Selection Window

[0043] 620: Analysis Information Output Window

[0044] 1100: Data processing device Detailed Implementation

[0045] This invention can be modified in various forms and has various embodiments, and specific embodiments thereof are shown by way of example in the accompanying drawings and will be described in detail below. However, it should be understood that the invention is not intended to be limited to the specific embodiments; rather, the invention should cover all modifications, equivalents, and alternatives falling within the spirit and scope of the invention. Throughout the description of the drawings, the same reference numerals refer to the same elements.

[0046] It will be understood that although terms such as first, second, A, B may be used herein to describe various elements, these elements should not be limited by these terms. These terms are used only to distinguish one element from another. For example, a first element may be referred to as a second element without departing from the scope of the invention, and similarly, a second element may be referred to as a first element. As used herein, the term "and / or" includes a combination of a plurality of associated listed items or any one of a plurality of associated listed items.

[0047] It will be understood that when a component is referred to as "coupled" or "connected" to another component, it can be directly coupled or connected to that other component, or there can be intermediate components. Conversely, when a component is referred to as "directly coupled" or "directly connected" to another component, there are no intermediate components.

[0048] The terminology used herein is for describing particular embodiments only and is not intended to limit the invention. Unless the context clearly indicates otherwise, the singular forms “a,” “an n,” and “the” used herein are intended to also include the plural forms. It will be further understood that the terms “comprising,” “including,” “containing,” “covering,” and / or “having,” as used herein, specify the presence of stated features, integers, steps, operations, constituent elements, components, and / or combinations thereof, but do not preclude the presence or addition of one or more other features, integers, steps, operations, constituent elements, components, and / or combinations thereof.

[0049] Unless otherwise defined, all terms used herein, including technical and scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. It will be further understood that terms, such as those defined in common dictionaries, should be interpreted as having the same meaning as they have in the context of the relevant art, and will not be interpreted in an idealized or overly formal sense unless expressly defined herein.

[0050] Figure 1 The general battery manufacturing process is illustrated.

[0051] Batteries can be manufactured by sequentially executing multiple unit processes. More specifically, the battery manufacturing process can be classified into N unit processes, and batteries can be manufactured by sequentially executing the first through the Nth process.

[0052] For example, cell batteries can be manufactured by sequentially executing unit processes that can be classified as electrode coating process (first process), electrode rolling process (second process), assembly process (third process), activation process (fourth process), and end-of-line (EOL) process (fifth process).

[0053] Performance tests can be performed during the individual cell manufacturing process or after battery manufacturing is complete to confirm whether the battery exhibits the expected performance. The performance metrics to be tested may include the battery's discharge capacity, charging resistance, and discharging resistance.

[0054] If the performance test determines that the battery does not meet the expected performance, then changes to the battery design or adjustments to manufacturing process variables are required.

[0055] However, because various process factors associated with each cell process affect battery performance, and these process factors have highly complex interrelationships, it is difficult to deduce the process factors that have a significant impact on battery performance. Furthermore, even if the process factors affecting battery performance are identified, it is difficult to determine the extent of their influence on battery performance.

[0056] This invention relates to a technique for solving such problems, and to a data processing apparatus and method that can identify process factors affecting battery performance by using data related to the battery manufacturing process, and analyze the impact of these process factors.

[0057] In the following, the operation of the data processing apparatus according to the present invention and various embodiments thereof, and the data processing methods performed by the data processing apparatus will be described in detail with reference to the accompanying drawings.

[0058] Figure 2 This is an operation flowchart of a data processing method according to an embodiment of the present invention.

[0059] The data processing device can collect process data (S210). Here, the process data may include process data for each process factor of each of the multiple batteries.

[0060] Process data may include data associated with each of the multiple unit processes. Here, a unit process may include one or more of the electrode coating process, electrode rolling process, assembly process, activation process, and EOL process.

[0061] Figure 3 This is an example of process data according to an embodiment of the present invention.

[0062] refer to Figure 3 The data processing device can collect process data that is categorized into multiple data instances. Here, a data instance can refer to the process data of each process factor of an individual battery. For example, the data processing device can collect process data of each process factor of each of 100,000 battery cells.

[0063] The process factors included in the electrode coating process (first unit process) may include one or more of the following: slurry temperature, slurry flow rate, coating gap, coating speed, and coating thickness.

[0064] The process factors included in the electrode rolling process (second unit process) may include one or more of the following: rolling roll gap, rolling pressure, rolling speed, and rolling thickness.

[0065] The process factors included in the assembly process (the third unit process) may include one or more of the following: laminate temperature, lamination pressure, lamination roller temperature, and lamination speed.

[0066] The process factors included in the activation process (Unit 4 process) may include one or more of the following: fixture forming temperature, fixture forming pressure, and waiting time between processes.

[0067] The process factors included in the EOL process (the fifth unit process) may include one or more of the following: electrolyte quantity, performance measurement temperature, and monomer thickness.

[0068] The data processing device can use the collected process data to generate process data for specific process factors. For example, the data processing device can use the collected process data to generate process data such as rolling rate, thickness of activated cathode, and porosity of activated cathode.

[0069] at the same time, Figure 3 Each process factor shown in the unit process is an example for clearly explaining the invention, and therefore the scope of the invention is not limited to the type of process factor.

[0070] Return to reference Figure 2 The data processing device can use process data to construct a performance prediction model (S220). Here, the performance prediction model may correspond to a machine learning-based artificial intelligence model, which outputs one or more predicted values ​​of one or more battery performance factors when specific process data is input.

[0071] Specifically, the data processing device can use process data for each process factor to construct a performance prediction model, which outputs one or more predicted values ​​for one or more performance factors. Here, the data processing device can use process data for each process factor as training data to train the performance prediction model. The performance prediction model can be defined as outputting at least one predicted value for a battery performance factor as output data when specific process data is input as input data. Here, the performance prediction value output by the performance prediction model can correspond to at least one predicted value for one or more performance factors, including battery capacity and resistance.

[0072] The data processing device can use the performance prediction model constructed in S220 to generate analytical information indicating the influence of one or more process factors on the performance prediction value of the performance prediction model (S230). Here, the analytical information may include an influence index of one or more process factors. The influence index may refer to a numerical value indicating the degree of influence of an individual process factor on the performance prediction value of the performance prediction model.

[0073] In one embodiment, the data processing device can calculate the impact index of individual process factors based on the change in the performance prediction value according to the change in individual process factors.

[0074] For example, a data processing device can calculate the impact index of a corresponding process factor based on the difference between the performance prediction value of a performance prediction model and the average of the performance prediction values ​​calculated by randomly changing the characteristic values ​​of a specific process factor.

[0075] In another example, the data processing device can calculate the impact index of a corresponding process factor based on the difference between the performance prediction value of the performance prediction model and the performance prediction value calculated by removing the feature values ​​of a specific process factor.

[0076] In another embodiment, the data processing apparatus may calculate the influence index of a specific process factor based on the contribution of the specific process factor calculated by backpropagating the performance prediction value to the performance prediction model.

[0077] The impact index can be calculated as a positive or negative number. Here, if the impact index is positive, it means that it has a positive impact on the performance prediction (i.e., it helps to increase the performance prediction), while if it is negative, it means that it has a negative impact on the performance prediction (i.e., it helps to decrease the performance prediction).

[0078] The data processing device can output the analysis information generated in S230 (S240) through a predefined graphical user interface (GUI).

[0079] The data processing device can select the top N process factors with high impact indices (N is a natural number greater than or equal to 2, which is predefined or input by the user) and output the selected process factors and the impact index of each process factor through the GUI.

[0080] The data processing device can generate visual content corresponding to output items input by the user and output the visual content through a GUI. Here, the visual content can include a combination of objects that the user can visually recognize (e.g., lines, arrows, shapes, N-dimensional graphs, etc.). For example, when a user inputs a specific output item (e.g., instance-specific impact, process factor-specific impact, unit process-specific impact), the data processing device can collect and process analytical information related to the output item to generate visual content and output the visual content through a GUI.

[0081] The data processing device can visualize and output one or more of the length, size, orientation, and color of objects corresponding to process factors to correspond to the influence index of the corresponding process factor. For example, when receiving an output request for a specific influence of a process factor, the data processing device can visualize and output the influence index of each of multiple process factors in the form of a horizontal bar chart. Here, the data processing device can group process factors by unit process and output each group with a different color.

[0082] According to embodiments of the present invention, users can identify process factors affecting battery performance through analysis information output via the GUI, or intuitively understand the degree of influence of process factors on battery performance.

[0083] Figure 4 This is an operation flowchart of a method for constructing a performance prediction model according to an embodiment of the present invention. Furthermore, Figure 5 This is a reference diagram for explaining a data preprocessing method according to an embodiment of the present invention.

[0084] The data processing device can execute a pre-defined preprocessing procedure (S410) for the process data collected in S210.

[0085] The preprocessing process may include data engineering processes for processing data instances and feature engineering processes for processing the process factors included in each data instance.

[0086] First, during the data engineering process, one or more data instances can be removed based on predefined criteria.

[0087] For example, a data processing device can remove data instances from 100,000 data instances that are specific to batteries manufactured in a predefined coated LOT and those with feature values ​​that deviate from predefined domain knowledge.

[0088] Next, during feature engineering, process data for one or more process factors can be removed based on predefined criteria. Specifically, the data processing device can remove process data for one or more process factors from a plurality of process factors based on either their relevance or importance.

[0089] For example, such as Figure 5 As shown, the data processing device can calculate the correlation coefficients among process factors (#1 to #30) and remove process data for process factors with correlation coefficients exceeding a predefined threshold. Here, the correlation coefficients may include the Pearson correlation coefficient. Furthermore, the data processing device can calculate the importance index for each process factor and remove process data for process factors with importance indices below a predefined threshold. Here, the importance index may include the permutation importance score.

[0090] Return to reference Figure 4 The data processing unit can use preprocessed process data to build a machine learning-based performance prediction model (S420). Here, the data processing unit can generate multiple performance prediction models by applying different learning algorithms.

[0091] For example, the data processing device can generate performance prediction models based on random forests, XGBoost, and deep neural networks, and use process data for each process factor as learning data to train each of the performance prediction models.

[0092] Subsequently, the data processing device can calculate the model evaluation index for each of the performance prediction models (S430), and determine the performance prediction model that exhibits the best performance among the performance prediction models based on the model evaluation index (S440). Here, the model evaluation index can be calculated based on the difference between the performance prediction value and the performance measurement value.

[0093] In this embodiment, the data processing device can calculate the root mean square error (RMSE) and the R-squared fraction (R^2) based on the difference between the performance prediction value and the performance measurement value of each of the performance prediction models based on random forests, XGBoost, and deep neural networks. Here, the data processing device can determine the performance prediction model with the lowest RMSE or the R^2 closest to 1 as the optimal performance prediction model.

[0094] For example, if the RMSE of the random forest-based charging resistance prediction model is 0.441 and R^2 is 0.625, the RMSE of the XGBoost-based charging resistance prediction model is 0.311 and R^2 is 0.814, and the RMSE of the deep neural network-based charging resistance prediction model is 0.395 and R^2 is 0.669, then the data processing device can determine the XGBoost-based charging resistance prediction model with the lowest RMSE and R^2 closest to 1 as the best charging resistance prediction model.

[0095] In this embodiment, the data processing device can perform an optimization process on the optimal performance prediction model determined in S440. Here, the data processing device can derive the optimal combination of multiple hyperparameters included in the performance prediction model by using a Bayesian optimization algorithm.

[0096] Subsequently, the data processing device can use the optimized performance prediction model to generate analytical information for identifying process factors affecting battery performance (S230), and output the analytical information through the GUI (S240).

[0097] Figure 6 This is an example screen of a user terminal used to interpret a GUI according to an embodiment of the present invention.

[0098] refer to Figure 6 The system can visualize and output analysis information corresponding to the output items input by the user through a GUI. Here, the GUI may include an output item selection window 610 and an analysis information output window 620.

[0099] The data processing device can receive a selection signal for an output item input by a user through the output item selection window 610. Here, the output item may include one or more of instance-specific effects, process factor-specific effects, and unit process-specific effects.

[0100] When a user inputs a specific output item, the data processing device can collect and process analysis information related to the output item to generate visual content, and output the visual content through the analysis information output window 620.

[0101] If instance-specific impact is selected as the output and a specific data instance (e.g., a battery identifier) ​​is input, the data processing device can visualize and output the impact index of each process factor for the corresponding data instance (corresponding battery).

[0102] If a specific impact of a process factor is selected as an output, the data processing device can visualize and output the impact index of each of the multiple process factors.

[0103] If a specific process factor is selected as the output and a specific process factor (e.g., coating speed) is input, the data processing apparatus can visualize and output the changes in performance predictions based on the changes in the corresponding process factor.

[0104] If a unit process-specific impact is selected as an output item, the data processing device can group process factors by unit process and visualize and output the impact index of each unit process.

[0105] Figures 7 to 10 This is an example of a user terminal screen used for interpreting and analyzing information according to an embodiment of the present invention.

[0106] refer to Figure 7 When a user selects instance-specific impact as an output item and inputs a specific data instance (e.g., bat#266), the data processing device can visualize and output the impact index of the corresponding data instance by process factor (bat #266).

[0107] Here, the data processing device can select the top N process factors with high impact indices (N is a natural number greater than or equal to 2, which is preset or input by the user) and output them by visualizing the objects corresponding to each of the selected process factors, so that they correspond to each impact index.

[0108] For example, such as Figure 7 As shown, the data processing device can select the top 7 process factors with high influence indices (#10, #15, #7, #4, #19, #27, #21) and output the objects (arrows) corresponding to each process factor by sorting them in descending order. Here, each object (arrow) has a length corresponding to the magnitude of the influence index and can indicate the direction corresponding to the sign of the influence index.

[0109] Furthermore, the data processing device can define the average predicted value of the performance prediction model as the starting point based on the x-axis, and calculate the starting and ending points for each object based on the influence index of process factors, and visualize them, such as... Figure 7 As shown in the diagram. Therefore, users can visually identify the main process factors affecting the performance of a specific selected battery, as well as the magnitude and direction of each process factor's impact on battery performance.

[0110] refer to Figure 8 When a user selects a specific impact of a process factor as an output item, the data processing device can output the impact index of each process factor among multiple process factors by visualizing the impact index of each process factor among multiple process factors.

[0111] Here, the data processing device can select the top N process factors with high impact indices and output these objects by visualizing the objects corresponding to each of the selected process factors, so that these objects correspond to each impact index.

[0112] For example, such as Figure 8 As shown, the data processing device can select the top 7 process factors with high impact indices (#10, #15, #7, #4, #19, #27, #21) and output the objects (bars) corresponding to each process factor by arranging them in descending order. Here, each object (bar) can be represented as a length corresponding to the absolute value of the impact index. Therefore, users can intuitively identify the main process factors affecting battery performance and the level of influence of each process factor on battery performance.

[0113] refer to Figure 9 When a user selects a specific process factor impact as an output and inputs a specific process factor (e.g., #4), the data processing device can visualize and output the changes in the performance prediction value based on the changes in the corresponding process factor.

[0114] For example, such as Figure 9 As shown, the data processing device can visualize and output the changes in performance prediction values ​​based on variations in the characteristic values ​​of the selected process factor (#4) in the form of a two-dimensional graph. Here, the data processing device can simultaneously output a graph showing the relationship between the characteristic value and the performance prediction value for each data instance, and a graph showing the relationship between the average value of the data instances. Therefore, users can intuitively identify the trends in battery performance changes based on variations in specific process factors and the locations of rapidly changing characteristic values ​​in battery performance.

[0115] refer to Figure 10 When a user selects a specific impact of a unit process as an output item, the data processing device can output the impact index of each unit process by visualizing the impact index of each unit process.

[0116] Here, the data processing device can group multiple process factors based on unit processes, sum up the influence indices of the process factors included in each unit process, and output the objects corresponding to each unit process by visualizing the objects corresponding to each unit process, so that they correspond to the influence index of each unit process.

[0117] For example, such as Figure 10As shown, the data processing device can categorize process factors into electrode coating, electrode rolling, assembly, activation, and EOL processes. It sums the influence indices of the process factors included in each unit process and then outputs the corresponding objects (bars) for each unit process by visualizing them, ensuring they correspond to the influence index of each unit process. Here, each object (bar) can be represented as the length corresponding to the absolute value of the influence index of each unit process. Therefore, users can intuitively identify the impact of each unit process on battery performance.

[0118] Figure 11 This is a block diagram of a data processing apparatus according to an embodiment of the present invention.

[0119] A data processing apparatus 1100 for analyzing a battery manufacturing process according to an embodiment of the present invention may include at least one processor 1110, a memory 1120 storing at least one instruction executed by the processor, and a transceiver 1130 connected to a network to perform communication.

[0120] At least one instruction may include: instructions for collecting process data for each process factor of multiple batteries; instructions for using the process data for each process factor to construct a machine learning-based performance prediction model for predicting battery performance; instructions for generating analytical information indicating the impact of one or more process factors on the performance prediction values ​​of the performance prediction model; and instructions for outputting the generated analytical information via a predefined graphical user interface (GUI).

[0121] Here, process data may include data related to one or more unit processes among electrode coating process, electrode rolling process, assembly process, activation process and end-of-line (EOL) process.

[0122] Instructions for building machine learning-based performance prediction models may include instructions for removing process data for one or more process factors among a plurality of process factors based on one of the relevance and importance of those process factors.

[0123] Instructions for building a machine learning-based performance prediction model may include instructions for building a performance prediction model that outputs at least one predicted value for one or more performance factors, including battery capacity and resistance, by using process data for each process factor as learning data.

[0124] Instructions for building machine learning-based performance prediction models may include: instructions for generating multiple performance prediction models by applying different learning algorithms; and instructions for selecting the performance prediction model that exhibits the best performance among the performance prediction models based on the difference between the performance prediction value and the performance measurement value of each performance prediction model.

[0125] Instructions for generating analytical information may include instructions for calculating the influence index of each process factor among the process factors based on changes in performance prediction values ​​according to changes in individual process factors.

[0126] Instructions for outputting analytical information via a predefined GUI may include instructions for visualizing and outputting one or more of the length, size, orientation, and color of objects corresponding to process factors so as to correspond to the influence index of the corresponding process factor.

[0127] Instructions for outputting analytical information via a predefined GUI may include instructions for selecting the top N process factors with high influence indices (N is a natural number greater than or equal to 2); and instructions for visualizing and outputting objects corresponding to each of the N process factors so as to correspond to the respective influence index of the objects.

[0128] Instructions for outputting analytical information via a predefined GUI may include: instructions for grouping multiple process factors based on unit processes; instructions for summing the influence indices of the process factors included in each unit process; and instructions for visualizing and outputting objects corresponding to each unit process such that the objects correspond to the influence indices of each unit process.

[0129] Instructions for outputting analytical information via a predefined GUI may include instructions for outputting changes in performance predictions based on variations in process factors as a two-dimensional graph.

[0130] The data processing device 1100 may also include an input interface 1140, an output interface 1150, a storage device 1160, etc. The various components included in the data processing device 1100 can be connected to each other via a bus 1170 to communicate with each other.

[0131] Here, processor 1110 may refer to a central processing unit (CPU), graphics processing unit (GPU), or dedicated processor, on which the method according to an example of the invention is executed. The memory (or storage device) may include at least one of volatile storage media and non-volatile recording media. For example, the memory may include at least one of read-only memory (ROM) and random access memory (RAM).

[0132] The operation of the method according to embodiments of the present invention can be implemented as a computer-readable program or code on a computer-readable recording medium. A computer-readable recording medium includes all types of recording devices in which computer systems store data readable by the computer. Furthermore, the computer-readable recording medium can be distributed across network-connected computer systems to store and execute computer-readable programs or code in a distributed manner.

[0133] While some aspects of the invention have been described in the context of apparatus, they may also refer to, according to the description of the corresponding method, a block or apparatus corresponding to a method step or feature of a method step. Similarly, aspects described in the context of a method may also refer to features of a corresponding block or item or corresponding apparatus. Some or all of the method steps may be performed by (or using) hardware devices, such as, for example, microprocessors, programmable computers, or electronic circuits. In some embodiments, one or more of the most important method steps may be performed by such apparatus.

[0134] The invention has been described above with reference to exemplary embodiments thereof; however, those skilled in the art will appreciate that various corrections and modifications may be made to the invention within the scope thereof without departing from the spirit and scope of the invention as described in the appended claims.

Claims

1. A data processing apparatus for analyzing a battery manufacturing process, comprising: at least one processor; and a memory configured to store at least one instruction executed by the at least one processor, wherein the at least one instruction comprises: an instruction for collecting process data of each process factor of a plurality of batteries; an instruction for constructing a machine learning-based performance prediction model using the process data of each process factor, the performance prediction model being used to predict a battery performance; an instruction for generating analysis information indicating an effect of one or more process factors on a performance prediction value of the performance prediction model; and an instruction for outputting the generated analysis information through a predefined graphical user interface (GUI). The process data comprises data related to one or more unit processes among an electrode coating process, an electrode rolling process, an assembly process, an activation process, and an end-of-line (EOL) process.

2. The apparatus of claim 1, wherein, The instruction for constructing the machine learning-based performance prediction model comprises:

3. The apparatus of claim 2, wherein, an instruction for removing process data on one or more process factors among the plurality of process factors based on one of a degree of correlation and a degree of importance of the plurality of process factors. The instruction for constructing the machine learning-based performance prediction model comprises:

4. The apparatus of claim 2, wherein, an instruction for constructing a performance prediction model that outputs at least one prediction value of one or more performance factors among a capacity and a resistance of the battery by using the process data of each process factor as learning data. The instruction for constructing the machine learning-based performance prediction model comprises:

5. The apparatus of claim 4, wherein, an instruction for generating a plurality of performance prediction models by applying different learning algorithms; and an instruction for selecting a performance prediction model that exhibits the best performance among the performance prediction models based on a difference between a performance prediction value and a performance measurement value of each of the performance prediction models. The instruction for generating the analysis information comprises:

6. The apparatus of claim 1, wherein, an instruction for calculating an influence index of each of the process factors based on a change in a performance prediction value according to a change in an individual process factor. The instruction for outputting the analysis information through the predefined GUI comprises:

7. The apparatus of claim 6, wherein, an instruction for visualizing and outputting one or more of a length, a size, a direction, and a color of an object corresponding to a process factor so as to correspond to an influence index of the corresponding process factor. The instruction for outputting the analysis information through the predefined GUI comprises:

8. The apparatus of claim 6, wherein, an instruction for selecting top N process factors having a high influence index, N being a natural number greater than or equal to 2; and an instruction for visualizing and outputting an object corresponding to each of the N process factors so as to correspond to a respective influence index of the object. The instruction for outputting the analysis information through the predefined GUI comprises:

9. The apparatus of claim 6, wherein, an instruction for grouping the plurality of process factors based on the unit processes; an instruction for summing up the influence indexes of the process factors included in each of the unit processes; and an instruction for outputting the sum of the influence indexes of the process factors included in each of the unit processes. instructions for visualizing and outputting an object corresponding to each of the unit processes such that the object corresponds to an influence index of each unit process.

10. The apparatus of claim 6, wherein, The instructions for outputting the analysis information through the predefined GUI include: instructions for outputting a change in the performance prediction value according to a change in the process factor in the form of a two-dimensional graph. 11.A data processing method for analyzing a battery manufacturing process, comprising: collecting process data of each process factor of a plurality of batteries; constructing a machine learning-based performance prediction model using the process data of each process factor, the performance prediction model for predicting a battery performance; generating analysis information indicating an influence of one or more process factors on a performance prediction value of the performance prediction model; and outputting the generated analysis information through a predefined graphical user interface (GUI). The process data includes data related to one or more unit processes among an electrode coating process, an electrode rolling process, an assembly process, an activation process, and an end-of-line (EOL) process.

12. The method of claim 11, wherein, Constructing the machine learning-based performance prediction model includes:

13. The method of claim 12, wherein, removing process data on one or more of the plurality of process factors based on one of a degree of correlation and a degree of importance of the plurality of process factors. Constructing the machine learning-based performance prediction model includes:

14. The method of claim 12, wherein, constructing a performance prediction model that outputs at least one prediction value of one or more performance factors among a capacity and a resistance of the battery by using the process data of each process factor as learning data. Constructing the machine learning-based performance prediction model includes:

15. The method of claim 14, wherein, generating a plurality of performance prediction models by applying different learning algorithms; and selecting a performance prediction model that exhibits the best performance among the performance prediction models based on a difference between a performance prediction value and a performance measurement value of each of the performance prediction models. Generating the analysis information includes:

16. The method of claim 11, wherein, calculating an influence index of each of the process factors based on a change in a performance prediction value according to a change in an individual process factor. Outputting the analysis information through the predefined GUI includes:

17. The method of claim 16, wherein, visualizing and outputting one or more of a length, a size, a direction, and a color of an object corresponding to a process factor so as to correspond to an influence index of the corresponding process factor. Outputting the analysis information through the predefined GUI includes:

18. The method of claim 16, wherein, selecting top N process factors having a high influence index, N being a natural number greater than or equal to 2; and visualizing and outputting an object corresponding to each of the N process factors so as to correspond to a respective influence index of the object. Outputting the analysis information through the predefined GUI includes:

19. The method of claim 16, wherein, grouping the plurality of process factors based on the unit processes; summing up influence indexes of the process factors included in each of the unit processes; and visualizing and outputting an object corresponding to each of the unit processes such that the object corresponds to an influence index of each unit process. Outputting the analysis information through the predefined GUI includes:

20. The method of claim 16, wherein, ​ outputting the change in the performance prediction value according to the change in the process factor in the form of a two-dimensional graph.

Images