Fault diagnosis technology based on battery internal parameter monitoring

Abstract

The present invention relates to a fault diagnosis apparatus and method based on battery internal parameter monitoring. The fault diagnosis apparatus based on battery internal parameter monitoring, according to an embodiment of the present invention, comprises: a data collection unit that collects voltage and current data during charging and discharging of an ESS battery module; a parameter extraction unit that extracts, on the basis of the voltage and current data collected by the data collection unit, series resistance (R_i), diffusion resistance (R_diff), and diffusion capacitance (C_diff) that are internal model parameters of a battery; a threshold setting unit that sets a fault threshold on the basis of deviations of the extracted series resistance (R_i), diffusion resistance (R_diff), and diffusion capacitance (C_diff); and a fault determination unit that compares the set fault threshold with the extracted series resistance (R_i), diffusion resistance (R_diff), and diffusion capacitance (C_diff) to determine whether the ESS battery module is normal or faulty.

Description

Battery Internal Parameter Monitoring-Based Fault Diagnosis Technology

[0001] The present invention relates to a fault diagnosis device and method based on monitoring internal battery parameters, and specifically, to a series resistance (R) which is an internal battery model parameter. i ), diffusion resistance (R diff ), and diffusion capacitance (C diff This invention relates to a device and method for ensuring the stability and reliability of an ESS system by analyzing ) in real time.

[0002] The energy storage system (ESS) market is growing rapidly due to policies expanding renewable energy, and ESS is attracting attention as a technology that compensates for the shortcomings of intermittent renewable energy sources such as solar and wind power. ESS can be designed to maintain the stability of the power system by storing electricity produced during periods of low demand and supplying it during periods of high demand.

[0003] Currently, ESS is divided into grid-connected and residential types, and its deployment is progressing through demonstration projects in various countries, contributing to increased efficiency in the utilization of renewable energy. However, as the adoption of ESS expands, various technical issues are emerging. Frequent fire accidents in ESS further highlight the need for technological improvement. While various factors such as thermal management failures, defects in specific components, assembly process issues, and external impacts have been cited as causes of the fires, the clear cause remains unidentified.

[0004] In particular, lithium-ion batteries, a core component of ESS, are highly sensitive to external conditions, and prolonged exposure to high temperatures, overvoltage, or overcurrent can lead to deformation of the battery's internal structure. Such deformation can result in the loss of function of the separator that isolates the positive and negative electrodes, thereby increasing the risk of fire or explosion. While methods such as reducing maximum storage capacity or employing FMEA (Failure Mode and Effects Analysis) techniques have been proposed to mitigate the fire risk in ESS, the fundamental instability of battery materials that trigger fires and the technology to monitor them in real time remain insufficient. This situation underscores the need for precise monitoring technologies and data analysis to ensure the safety of ESS.

[0005] Furthermore, since ESSs are composed of multiple battery cells connected in series and parallel, if some cells operate under abnormal conditions (e.g., high temperature, overvoltage, overcurrent), it is highly likely to affect the entire system. Damage to a single cell can degrade overall system performance, and in severe cases, may render the system inoperable. Imbalances in cell lifespan act as a major cause of reduced stability and shortened lifespan in ESSs, and if these conditions persist, they can lead to serious accidents such as explosions and fires. These issues can lower the reliability of ESSs and act as major factors hindering safe system operation. Therefore, technical solutions capable of precisely monitoring the status of individual cells and detecting imbalances at an early stage may be required.

[0006] Although Battery Management Systems (BMS) possess the capability to store voltage, current, and temperature data for ESS status estimation and fault diagnosis, there is currently a lack of systems to systematically analyze or utilize this data. This demonstrates the inefficiency of underutilizing ESS data, despite the advancement of artificial intelligence and big data technologies. Only through systematic analysis can this data provide useful information to prevent fire and explosion accidents. Predictive models must be established through data-driven monitoring and analysis to detect abnormal conditions in ESS components early and prevent accidents.

[0007] While current ESS technology has made significant advancements in terms of power storage and efficiency maximization, major technical limitations, such as fire accidents and cell imbalances, can act as factors that undermine the safety and reliability of the system. Therefore, to enhance the safety and stability of ESS systems, systematic management utilizing real-time monitoring, precise data analysis, and predictive models is required.

[0008] The present invention aims to provide a fault diagnosis device and method for precisely analyzing the fundamental causes of battery failures and fire accidents that impair the stability of an ESS, and for detecting and preventing them at an early stage.

[0009] In addition, the present invention aims to provide a device and method that effectively monitors lifespan imbalances and abnormal conditions between battery cells to reduce performance degradation and the possibility of accidents in an ESS system.

[0010] However, the technical problem that this embodiment aims to solve is not limited to the technical problem described above, and other technical problems may exist.

[0011] A fault diagnosis device for an ESS battery module according to one embodiment of the present invention comprises: a data collection unit that collects voltage and current data during the charging and discharging of an ESS battery module; and a series resistance (R), which is an internal model parameter of the battery, based on the voltage and current data collected by the data collection unit. i ), diffusion resistance (R diff ), diffusion capacitance (C diff A parameter extraction unit that extracts ), the extracted series resistance (R i ), diffusion resistance (R diff ), diffusion capacitance (C diff A threshold setting unit that sets a fault threshold based on the deviation of ), the set fault threshold and the extracted series resistance (R i ), diffusion resistance (R diff ), diffusion capacitance (C diff It may include a fault determination unit that determines whether the battery module is normal or faulty by comparing ).

[0012] According to one embodiment, the data collection unit can collect voltage and current data in real time.

[0013] According to one embodiment, internal model parameters can be extracted in real time using a Recursive Least Squares (RLS) algorithm.

[0014] According to one embodiment, a recursive least squares algorithm can dynamically estimate the state of a battery module by setting the initial State-of-charge (SOC), Open Circuit Voltage (OCV), and internal resistance values ​​of the battery module, receiving voltage and current data as input, and continuously updating the error covariance and weights.

[0015] According to one embodiment, the fault determination unit may determine that the battery module is normal when the internal model parameter extracted in real time is within a set threshold, and determine that the battery module is abnormal when it deviates from the set threshold.

[0016] According to one embodiment, normal and abnormal states can be determined based on data collected by artificially inducing overcharge and overdischarge states for the battery module.

[0017] A fault diagnosis method for an ESS battery module according to another embodiment of the present invention comprises the steps of: collecting voltage and current data during the charging and discharging of the ESS battery module; and, based on the collected voltage and current data, the series resistance (R), which is an internal model parameter of the battery. i ), diffusion resistance (R diff ), diffusion capacitance (C diff A step of extracting ), series resistance (R) in preset normal and abnormal states i ), diffusion resistance (R diff ), diffusion capacitance (C diff A step of setting a failure threshold based on the deviation of ), the set failure threshold and the extracted series resistance (R i ), diffusion resistance (R diff ), diffusion capacitance (C diff It may include a step of determining whether the battery module is normal or faulty by comparing ).

[0018] According to one embodiment, the step of collecting voltage and current data may include the step of collecting voltage and current data in real time.

[0019] According to one embodiment, the internal model parameters may include a step of being extracted in real time using a Recursive Least Squares (RLS) algorithm.

[0020] According to one embodiment, the step of extracting in real time may include the step of setting the initial State of Charge (SOC), Open Circuit Voltage (OCV), and internal resistance value of the battery module, and the step of receiving voltage and current data and continuously updating the error covariance and weights to dynamically estimate the state of the battery module.

[0021] According to one embodiment, the step of determining whether a battery module is normal or faulty may include determining the battery module as normal if internal model parameters extracted in real time are within a set threshold, and determining it as abnormal if they fall outside the set threshold.

[0022] According to the present invention, the cause of battery failure can be analyzed in real time, and the stability of the ESS system can be enhanced through an early warning system.

[0023] In addition, according to the present invention, by precisely monitoring the imbalance between battery cells, preventive measures for the safe operation of the ESS and accident prevention can be provided.

[0024] FIG. 1 is a block diagram schematically showing the configuration of a fault diagnosis device for an ESS battery module according to one embodiment of the present invention.

[0025] Figure 2 is a diagram showing the Thevenin model for mathematically expressing the electrical operation of a battery.

[0026] FIG. 3a shows the series resistor (R i This is a graph showing the internal behavior of ).

[0027] Figure 3b shows the diffusion resistance (R diff This is a graph showing the internal behavior of ).

[0028] Fig. 3c shows the diffusion capacitance (C diff This is a graph showing the internal behavior of ).

[0029] Figure 4 is a graph showing the results of battery internal model parameter extraction and analysis extracted in real time from the ESS system.

[0030] Figure 5a is a graph showing the process of diagnosing early battery failure based on abnormal signs occurring during the charging period in an ESS system.

[0031] Figure 5b is a graph showing the process of diagnosing early battery failure based on abnormal signs occurring during the discharge period in an ESS system.

[0032] FIG. 6 is a flowchart illustrating a method for diagnosing a fault in an ESS battery according to another embodiment of the present invention.

[0033] Embodiments of the present invention are described in detail below with reference to the attached drawings so that those skilled in the art can easily implement the invention. Since the present invention is susceptible to various modifications and may have various embodiments, specific embodiments are illustrated in the drawings and described in detail in the description. However, this is not intended to limit the present invention to specific embodiments, and it should be understood that it includes all modifications, equivalents, and substitutions that fall within the spirit and scope of the invention.

[0034] To clearly explain the present invention, parts unrelated to the description have been omitted from the drawings, and similar parts throughout the specification have been given similar reference numerals. Furthermore, while describing with reference to the drawings, even components indicated by the same name may have different drawing numbers depending on the drawing, and drawing numbers are provided merely for the convenience of explanation; the concept, feature, function, or effect of each component is not to be interpreted restrictively by the corresponding drawing number.

[0035] Similar reference numerals are used for similar components when describing each drawing. Terms such as "first," "second," etc., may be used to describe various components, but said components should not be limited by said terms. These terms are used solely for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, the first component may be named the second component, and similarly, the second component may be named the first component. The term "and / or" includes a combination of a plurality of related described items or any of a plurality of related described items.

[0036] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as generally understood by those skilled in the art to which this invention pertains.

[0037] Terms such as those defined in commonly used dictionaries should be interpreted as having meanings consistent with their meanings in the context of the relevant technology, and should not be interpreted in an ideal or overly formal sense unless explicitly defined in this application.

[0038] Throughout the specification, when a part is described as being "connected" to another part, this includes not only cases where they are "directly connected" but also cases where they are "electrically connected" with other elements interposed between them. Furthermore, when a part is described as "comprising" a certain component, this means that, unless specifically stated otherwise, it does not exclude other components but rather allows for the inclusion of additional components; it should be understood that this does not preclude the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0039] As used in the present invention, the term “part” refers to software or hardware components such as FPGAs (field-programmable gate arrays) and ASICs (application-specific integrated circuits). However, “part” is not limited to hardware and software. “Part” may be configured to reside in an addressable storage medium or may be configured to run on one or more processors. Thus, by example, “part” includes components such as software components, object-oriented software components, class components, and task components, as well as processes, functions, attributes, procedures, subroutines, segments of program code, drivers, firmware, microcode, circuits, data, databases, data structures, tables, arrays, and variables. Components and functions provided within “part” may be combined into a smaller number of components and “parts” or separated into additional components and “parts.”

[0040] Hereinafter, a device and method for diagnosing a fault in an ESS battery module according to embodiments of the present invention will be described in detail with reference to the drawings.

[0041] FIG. 1 is a block diagram schematically showing the configuration of a fault diagnosis device for an ESS battery module according to one embodiment of the present invention.

[0042] Referring to FIG. 1, the fault diagnosis device (100) of an ESS battery module may include a data collection unit (110), a parameter extraction unit (120), a threshold setting unit (130), and a fault determination unit (140).

[0043] The data collection unit (110) can measure voltage and current data generated during the charging and discharging of the ESS battery module in real time and provide data so that the state of the battery can be analyzed in real time. The data collection unit (110) measures voltage and current data individually from a plurality of battery cells constituting the battery module, and through this, can analyze the difference between the normal state and the abnormal state.

[0044] In some embodiments, the data collection unit (110) collects voltage and current data in real time at the battery module level and can continuously monitor the charging and discharging status of the battery based on this. To this end, it may include voltage and current measuring equipment capable of high-speed sampling.

[0045] In the present invention, as an example, a 6S2P battery module was designed to collect charge / discharge data of an ESS battery module, and data was collected by artificially inducing overcharge and overdischarge states so as to compare and analyze the characteristics between normal cells and abnormal cells. Specifically, voltage and current data according to changes in battery state were obtained by performing a degradation experiment of 100 cycles.

[0046] In some embodiments, charge and discharge data of ESS battery modules are designed to be applicable to the entire ESS system by extending the electrical equivalent circuit model, and to this end, voltage and current data of batteries connected in series and parallel can be measured and stored. Since multiple battery modules are connected in series and parallel in an actual ESS system, the data collection unit must collect not only simple individual cell data but also synchronized data at the module and ESS levels. Through this, the data collection unit can go beyond simple voltage and current measurement and be utilized to analyze battery status in real time and detect abnormal conditions even in a large-scale operating environment of the ESS system.

[0047] In this way, the data collection unit (110) of the present invention measures voltage and current data in real time during the charging and discharging of the ESS battery module and can provide data to distinguish between normal and abnormal states of the battery based on this. A real-time monitoring system for the ESS can be established through the data provided by the data collection unit (110), thereby ensuring the safety of the ESS and increasing operational reliability.

[0048] The parameter extraction unit (120) analyzes voltage and current data collected during the charging and discharging of the ESS battery module and the series resistance (R), which is an internal model parameter representing the internal state of the battery. i ), diffusion resistance (R diff ), and diffusion capacitance (C diff ) can be extracted in real time. Internal model parameters are variables representing the electrical behavior of the battery and can be used to quantitatively analyze the degradation state of the battery cell, the electrochemical reaction rate, the dynamic response characteristics during charging and discharging, and changes in internal impedance.

[0049] The parameter extraction unit (120) uses collected voltage and current data as input values ​​and applies a Recursive Least Squares (RLS) algorithm as follows, and is designed to calculate internal model parameters in real time based on this.

[0050] [Recursive Least Squares and Recursive Least Squares Symbol]

[0051]

[0052] Unlike the conventional Least Squares method, the Recursive Least Squares (RLS) algorithm can estimate parameters by continuously updating weights based on newly input data without storing all data. RLS can be designed to be suitable for real-time monitoring and control systems and can effectively reflect the dynamic characteristics of given data.

[0053] The objective function of RLS aims to minimize the sum of squared errors (Cost Function) between observations and estimates. To achieve this, RLS calculates weights based on errors and updates parameters. During this process, a forgetting factor (λ) is applied to gradually reduce the influence of previous data. The forgetting factor is set to a value between 0 and 1, assigning greater weight to recent data to reduce inefficiencies that may arise from historical data. This approach can be utilized to effectively reflect the dynamic characteristics and current state of the system.

[0054] The key steps of the RLS algorithm are as follows. In the initial stage, initial conditions for the parameter estimates and the error covariance matrix are established. These initial values ​​can be defined based on system design requirements or prior information. Subsequently, the error between the predicted values ​​and the actual observations is calculated using real-time observed data as input. Weights are updated based on the calculated error, thereby updating the parameters. The updated parameters are stored and used in the next iteration. These steps are performed repeatedly to process real-time data and provide estimates that reflect the latest data.

[0055] RLS enables efficient data processing and dynamic state analysis in real-time monitoring systems such as ESS. For example, the series resistance (R) representing the internal state of the ESS batteryi ), diffusion resistance (R diff ), diffusion capacitance (C diff It can be applied to extract and update internal model parameters such as ). In particular, due to its high computational efficiency, it can be used reliably even in systems that need to process large-scale data.

[0056] The parameter extraction unit (120) mathematically defines internal model parameters based on an electrical equivalent circuit model, and in the present invention, the electrical operation of the battery is mathematically expressed using the Thevenin model shown in FIG. 2, and the internal model parameters can be extracted by applying an RLS algorithm based on this.

[0057] In the Thevenin model, the transfer function of Equation 1 below, which represents the internal impedance of the battery, is derived, and the internal model parameters are calculated by applying the inverse z-transform of Equation 2 below to convert it into a relationship in the time domain.

[0058] (Mathematical formula)

[0059]

[0060] (Mathematical Formula 2)

[0061]

[0062] Here, the input vector is equal to Equation 3 below, and

[0063] (Mathematical Formula 3)

[0064]

[0065] Here,

[0066]

[0067] The output vector is as shown in the following mathematical formulas 4 to 6.

[0068] (Mathematical Formula 4)

[0069]

[0070] (Mathematical Formula 5)

[0071]

[0072] (Mathematical Formula 6)

[0073]

[0074] In some embodiments, the parameter extraction process proceeds as follows. First, an initial model is established using an SOC (State-of-charge) based OCV (Open circuit voltage) lookup table and an initial internal resistance value provided by the battery manufacturer. Subsequently, voltage and current data measured during charging and discharging are used as input values ​​to calculate the error between the model's predicted value and the measured value, and an RLS algorithm is applied to correct the internal model parameters based on this. During this process, a weight (Gain) is determined in a direction that minimizes the error, and this is reflected in the Error Covariance to continuously update the internal model parameters. Additionally, the forgetting factor (λ) of the RLS algorithm is applied to reflect the influence of the latest data more significantly and to reduce the influence of past data.

[0075] In some embodiments, as a result of comparing internal model parameters between normal cells and abnormal cells, R between normal cells diff The deviation is 8% or less, C diff The deviation was maintained below 9%, but in the case of abnormal cells, R diff Deviation of 20% or more, C diff It was analyzed that a deviation of more than 30% occurs.

[0076] The parameter extraction unit (120) is designed to be scalable at the actual ESS module and system level, and is configured to calculate internal model parameters at the module level as well as individual battery cells, taking into account that the ESS is a large-scale system composed of multiple battery cells. To this end, the parameter extraction unit (120) is designed to synchronize data at the module and ESS levels so that the RLS algorithm can be applied.

[0077] As some embodiments, an approach that extends electrical equivalent circuit modeling to the ESS system level based on existing experimental data may be applied. Actual ESSs operate with multiple battery cells connected in series and parallel, and beyond simple cell-level analysis, it must be possible to accurately derive internal model parameters even in such large-scale systems.

[0078] In series-connected batteries of an ESS, the voltages of individual cells accumulate so that the total voltage increases by the number of cells, and it can be assumed that all cells have the same capacity. On the other hand, in parallel-connected batteries, the total capacity increases by the number of cells, and it is assumed that all cells have the same voltage. To reflect these structural characteristics, the series resistance (R) derived at the cell level i ), diffusion resistance (R diff ), diffusion capacitance (C diff The data can be expanded into an electrical equivalent circuit model based on serial and parallel connection structures, and an RLS algorithm can be implemented based on this expanded model to analyze the internal model parameters of the entire ESS system in real time.

[0079] Through this, the parameter extraction unit (120) is designed to enable not only parameter analysis at the single cell level but also data synchronization and computation across the entire module and ESS system, and can reflect changes in the internal state of the battery in real time. In addition, by utilizing an extended electrical equivalent circuit model and an RLS algorithm, it is implemented to perform accurate state diagnosis based on internal model parameters even in large-scale ESS systems.

[0080] The series resistance (R) extracted from the parameter extraction unit (120) i ), diffusion resistance (R diff ), diffusion capacitance (C diff It can be used to distinguish between normal and abnormal states of the ESS system and to detect battery degradation and failure at an early stage. In particular, since changes in the internal state of battery cells are difficult to observe directly from the outside, real-time monitoring of internal model parameters is essential for the stable operation of the ESS system.

[0081] The threshold setting unit (130) distinguishes between normal and abnormal states of the ESS battery module, and the series resistor (R i ), diffusion resistance (R diff ), diffusion capacitance (C diff A failure threshold can be set by analyzing deviations in internal model parameters such as ). As charge and discharge cycles are repeated during operation, the electrochemical properties of the battery change, and consequently, internal resistance and diffusion characteristics may change. Since it is difficult to accurately identify abnormal conditions of the battery based on simple voltage and current values ​​alone, it is necessary to set a threshold that can distinguish between normal and failure states by analyzing deviations in internal model parameters that reflect the dynamic characteristics of the battery.

[0082] To clearly distinguish between the normal and abnormal states of a battery cell, deviations of internal model parameters under preset normal and abnormal conditions can be measured, and failure thresholds can be defined based on this.

[0083] According to some experimental examples, R between normal cells diff The deviation is 8% or less, C diff While the deviation is maintained below 9%, in abnormal cells, R diff Deviation of 20% or more, C diff It was confirmed that deviations exceeded 30%. Based on this, threshold values ​​for fault diagnosis can be defined in three categories. First, the safe area is R diff The deviation is 8% or less and C diff It can be set to the case where the deviation is 9% or less. Second, the warning area is R diff The deviation is greater than 8% and less than or equal to 20%, and C diff It can be classified into cases where the deviation exceeds 9% and is 30% or less. Third, the risk area is R diff The deviation is greater than 20% and C diff It can be set to the case where the deviation exceeds 30%. Through such categorization, the battery status can be analyzed in real time, and if a specific threshold is exceeded, it can be configured to immediately determine an abnormal state and activate an early warning system.

[0084] In some embodiments, the threshold setting unit (130) is R derived from each battery cell and module unit. i , R diff , C diff By comparing and analyzing values, the deviation between normal and abnormal states can be mathematically modeled. The process of setting the failure threshold involves establishing initial conditions and the initial R in the normal state. i , R diff , C diffThe process can proceed by establishing a reference database through value measurement, performing charge and discharge tests on normal and abnormal batteries, and continuously recording changes in internal model parameters. Subsequently, R in the normal and abnormal states diff and C diff By comparing the patterns of change, a threshold range can be set, and after distinguishing between normal range, warning range, and danger range to derive the optimal threshold, a real-time fault determination system can be implemented based on this.

[0085] In some embodiments, the threshold setting unit (130) may include an algorithm that can dynamically adjust the threshold by considering the battery's usage environment and charging / discharging conditions, rather than applying a simple fixed threshold. For example, a method of adjusting the threshold in a specific environment by reflecting the battery's State of Charge (SOC) and temperature changes may be applied.

[0086] The fault determination unit (140) determines whether the ESS battery module is normal or faulty using a set fault threshold and R, an internal model parameter extracted in real time. i , R diff , and C diff The condition of the battery can be determined by comparing these factors. During charging and discharging, the internal resistance and diffusion characteristics of the battery change, and failure can be determined by monitoring these changes in real time. Generally, it is difficult to directly detect abnormal battery conditions from the outside, and relying solely on simple voltage and current changes has limitations in accurately analyzing internal degradation or performance degradation.

[0087] In the present invention, to more precisely distinguish between the normal and failure states of a battery, a comparative analysis can be performed between a set failure threshold and internal model parameters measured in real time. In the normal state of the battery, R diff and C diffThe deviation is maintained within a relatively constant range, but in a failure state, a deviation exceeding a specific threshold may occur.

[0088] In the experimental example described above, R between normal cells diff The deviation is 8% or less, C diff The deviation is maintained at 9% or less, and in abnormal cells, R diff Deviation of 20% or more, C diff The deviation may show a tendency to increase by more than 30%. Based on this, the fault determination unit (140) can be designed to evaluate the condition of the battery in real time and classify it as an abnormal state if it exceeds a specific deviation range.

[0089] The fault determination unit (140) can classify the state of the battery into three categories based on a set fault threshold. First, the normal state is R diff The deviation is 8% or less and C diff It can be defined as the case where the deviation is 9% or less. Second, the warning state is R diff The deviation is greater than 8% and less than or equal to 20%, and C diff It can be classified as a case where the deviation exceeds 9% and is 30% or less. Third, the risk state is R diff The deviation is greater than 20% and C diff It can be set to a case where the deviation exceeds 30%, and if this range is exceeded, it can be determined that there is a high probability of serious performance degradation or failure of the battery. By applying this classification method, the failure determination unit (140) can be used to detect an abnormal state of the ESS battery module early and to ensure the stability of the battery.

[0090] In some embodiments, the fault determination unit (140) may continuously analyze internal model parameter data extracted in real time and immediately generate a warning signal if the fault threshold is exceeded. For example, in a specific battery cell R diffIf the deviation exceeds 20%, a warning can be generated for the battery module containing the cell, and information can be provided to enable the system operator to respond immediately. Additionally, if the warning state persists for a specific period, an automated system can be linked to perform replacement or protective measures for the battery module. Through this, an early warning system is activated before the battery system reaches a critical failure state, and this can be utilized to enhance the stability of ESS operations.

[0091] Since the ESS is a large-scale system in which not only individual cells but also multiple cells are connected in series and parallel, the fault determination unit (140) can be designed to extend cell-level fault determination to the entire module and ESS system. For example, cell-level data can be extended to the module and ESS level so that the same criteria can be applied in the large-scale ESS system. In the case of battery modules connected in series and parallel, the fault determination is designed to be performed by reflecting changes in the internal model parameters of individual cells, thereby enabling consistent state diagnosis across the entire module and ESS system.

[0092] FIGS. 3a and 3b show the series resistance (R), which is an internal model parameter of the battery, in the normal state and under overcharge and overdischarge states. i ), diffusion resistance (R diff ), and diffusion capacitance (C diff This is a graph showing the change of ) over time.

[0093] FIG. 3a shows the series resistance (R) of the battery. i The change in ) is expressed as a function of time. The change in Ri is compared between normal cells (Cell 1 Normal, Cell 6 Normal), overcharged (Cell 1 Overcharge, Cell 6 Overcharge), and overdischarged (Cell 1 Overdischarge, Cell 6 Overdischarge) states. In the initial stages of charging and discharging, R iIt shows a tendency to decrease rapidly, and after stabilization, the R of normal cells i remains constant. On the other hand, under overcharge and over-discharge conditions, R over time i An increasing trend is confirmed, indicating that the battery's internal resistance increases due to degradation. Under overcharged conditions, R compared to the normal state i The value of increases significantly, and shows an even more rapid increase in the over-discharged state. This may indicate the effect of electrochemical damage caused by overcharging and over-discharging on the internal resistance of the battery.

[0094] Figure 3b shows the diffusion resistance (R diff It expresses the change in ) as a function of time and can explain how the ion diffusion resistance inside the battery changes according to charge and discharge conditions. For a normal cell, R diff It exhibits a pattern of gradually increasing during initial charging and discharging, then stabilizing at a certain level. However, under overcharging and over-discharging conditions, R diff maintains a significantly higher value compared to the normal state, and in the case of over-discharge, R over time diff It can be observed that R shows a characteristic of rising rapidly. This R diff Changes in may indicate the degradation of the electrolyte and an increase in the resistance of the electrode surface due to overcharging and over-discharging, and this can be used as a useful indicator to distinguish between the normal state and the failure state of the battery.

[0095] Fig. 3c shows the diffusion capacitance (C diff It shows the change in ), which can indicate changes in the electrochemical reaction rate and charging capacity within the battery. In the steady state, C in the initial stage diff After decreasing rapidly, it stabilizes, and the charging and discharging rates maintain a constant value. On the other hand, under overcharging and over-discharging conditions, C diffis maintained at a value significantly lower than the normal state, and in the over-discharge state, C over time diff The value of [the value] shows a phenomenon of rapid decrease. This may suggest that factors such as loss of active material, damage to the electrode surface, and electrolyte degradation under overcharging and over-discharging conditions affect the electrochemical reaction of the battery.

[0096] FIGS. 3a to 3c visually illustrate the process of utilizing internal model parameters to distinguish between the normal state and the overcharged and over-discharged states of an ESS battery cell, thereby providing practical data that can be used to set reference values ​​necessary for determining battery degradation and failure. Each internal model parameter can represent important information regarding the performance and stability of the battery according to charging and discharging conditions, and can be utilized to design a real-time status monitoring and fault diagnosis system for the ESS based on this information.

[0097] Figure 4 is a graph showing the results of battery internal model parameter extraction and analysis extracted in real time from the ESS system.

[0098] Referring to Fig. 4, the first graph shows the series resistance (R) of the battery. i Shows the change of ) as a function of time. R i It remains stable initially but may change after a certain point, and this can react sensitively to charge / discharge conditions and the state of battery degradation. In a steady state, R i The value remains constant, but in overcharged or overdischarged conditions, R i An increasing characteristic can be observed. This change reflects the increase in electrochemical reaction resistance and electrode degradation within the battery, and can be used as an indicator of performance degradation.

[0099] The second graph shows the diffusion resistance (R diff Visually shows the change in ). R diffR represents the resistance generated during ion diffusion inside the battery, and can maintain a constant value under steady conditions. However, under abnormal conditions, R diff It fluctuates significantly after a certain point in time and may show a tendency to decrease sharply, particularly in over-discharged conditions. These changes reflect internal battery damage, such as electrolyte degradation, reduced electrode surface reactivity, and loss of active material, and can be utilized to detect abnormal conditions early.

[0100] The third graph shows the diffusion capacitance (C diff Shows the change of ) as a function of time. C diff is an indicator representing the electrochemical reaction rate inside the battery, and under normal conditions, it can fluctuate within a certain range. However, under overcharge and over-discharge conditions, C diff The value of may change abnormally and exhibit characteristics of rapidly increasing or decreasing after a certain point in time. Such changes reflect changes in the state of the electrolyte, loss of active material, and damage to the electrode surface structure, and can be used to evaluate the degradation state of the battery.

[0101] As described above, FIG. 4 can provide data necessary to distinguish between normal and abnormal states by analyzing internal battery model parameters in real time. In particular, sections in the graph where changes in each parameter appear prominent can be interpreted as abnormal signals of the battery status, and based on this, a fault determination unit can be used to analyze whether the battery is faulty. This analysis can be applied to maintain battery stability and implement an early warning system in an ESS system.

[0102] Figure 5a is a graph showing the process of diagnosing early failure of a battery based on abnormal signs occurring during the charging and discharging phases in an ESS system, and Figure 5b is a graph showing the process of diagnosing early failure of a battery based on abnormal signs occurring during the discharging phases in an ESS system.

[0103] As shown in FIG. 5a, during the initial charging phase, the battery voltage gradually increases, and R diff and C diff The deviation of maintains a very low level, so it can be classified as a steady state. As the charging process proceeds, at a specific point in time, R diff and C diff The deviation of begins to increase, which may indicate that degradation or abnormal electrochemical reactions have occurred within the battery. R diff and C diff If the deviation exceeds a preset threshold, it enters the warning zone (green area), which may indicate that battery performance degradation has begun. Beyond the warning zone, R diff and C diff If the deviation exceeds the risk threshold, it enters the risk zone (red area), where the likelihood of severe degradation or failure within the battery increases. This data can be used to prevent failure by detecting abnormal conditions occurring during charging at an early stage.

[0104] As shown in Fig. 5b, during the initial discharge phase, the battery voltage gradually decreases, and R diff and C diff The deviation of also remains stable at a low level and can be considered a steady state. However, as the discharge process progresses, R diff and C diff The deviation of increases sharply after a certain point in time, which may indicate that an abnormal condition occurred inside the battery during the discharge process. R diff and C diffIf the deviation exceeds the warning threshold, it enters the warning zone, during which battery performance degradation may accelerate. Subsequently, R diff and C diff If the deviation exceeds the risk threshold, it enters a risk zone, at which point battery damage worsens and the likelihood of failure becomes very high. Immediate action is required in this risk zone, and system stability can be ensured by stopping battery operation or implementing protective measures.

[0105] FIG. 6 is a flowchart illustrating a method for diagnosing a fault in an ESS battery according to another embodiment of the present invention.

[0106] Referring to Fig. 6, step S210 is a process of collecting voltage and current data during the charging and discharging of an ESS battery module, which can be utilized to secure basic data for fault diagnosis. Since the ESS system reflects dynamic changes within the battery in real time during charging and discharging, it may be necessary to collect accurate data. In step S210, data is measured not only at the individual battery cell level but also at the module level, and changes in battery status, such as voltage rise during charging or voltage drop during discharging, can be recorded. The collected data can be used as basic data necessary to calculate internal model parameters and determine whether a fault exists in subsequent steps.

[0107] Step S220 is based on the voltage and current data collected in Step S210 and the series resistance (R i ), diffusion resistance (R diff ), diffusion capacitance (C diffInternal model parameters such as ) can be extracted. Internal model parameters numerically represent the electrochemical characteristics of the battery and can be utilized to analyze dynamic changes in the battery state. In step S220, the Recursive Least Squares (RLS) algorithm is applied to receive voltage and current data as input values, and R i , R diff , C diff It can be calculated in real time. The RLS algorithm continuously updates internal model parameters by reflecting the latest data and may be suitable for tracking changes in battery state in real time. The internal model parameters extracted in this step can be used to quantitatively represent the battery's degradation state, electrochemical reaction rate, and dynamic response characteristics during charging and discharging.

[0108] Step S230 uses the series resistance (R) extracted in Step S220. i ), diffusion resistance (R diff ), diffusion capacitance (C diff This is the process of setting a failure threshold by analyzing the deviation of ). The failure threshold can be used as a criterion to distinguish whether the battery condition is normal or abnormal. In step S230, R between the normal state and the abnormal state is experimentally determined. i , R diff , C diff A threshold value can be defined by comparing deviations. For example, R in the steady state diff The deviation is 8% or less and C diff While the deviation is maintained at 9% or less, in abnormal conditions, R diff Deviation of 20% or more, C diff A deviation of 30% or more may occur. Based on this data, threshold ranges classified into normal, warning, and danger sections are set, which can be applied as a criterion for determining failure in subsequent steps.

[0109] Step S240 involves the set fault threshold and the series resistance (Ri) and diffusion resistance (R) extracted in Step S220. diff ), diffusion capacitance (C diff It is a process of determining the normal state and failure of the battery module by comparing ). In step S240, R i , R diff , C diff If the value is within the normal state threshold range, the battery can be determined to be in a normal state, and if it exceeds the warning threshold or reaches the danger threshold, it can be determined to be in an abnormal state. For example, R of a specific battery cell diff If the deviation exceeds 20% or C diff If the deviation exceeds 30%, the battery can be classified as faulty and early warning or protective measures can be performed. This step can be utilized to diagnose faults by analyzing the battery status in real time and to maintain the stability of the ESS system.

[0110] The order in which the steps of FIG. 6 are described should not be interpreted as restrictive, and the described steps may be executed in any order as long as it does not contradict the purpose of the invention, and may be subdivided into a number of sub-operations or a number of steps may be executed in parallel.

[0111] Although the embodiments have been described above with reference to limited examples and drawings, those skilled in the art can make various modifications and variations from the description above. For example, suitable results may be achieved even if the described techniques are performed in a different order than described, and / or if the components of the described system, structure, device, circuit, etc. are combined or assembled in a form different from described, or replaced or substituted by other components or equivalents. Therefore, other implementations, other embodiments, and equivalents to the claims below are also within the scope of the claims.

[0112] This invention was carried out with funding from the government (Ministry of Science and ICT) in 2022 and supported by the Korea Institute of Information and Communications Planning and Evaluation (No. 2022-1711152629, Development of an intelligent SW framework for safe autonomous operation and performance evaluation of large-scale distributed energy storage infrastructure).

[0113] This work was supported by the National Institute of Information and Communications Technology Evaluation and Planning with financial resources from the government (Ministry of Science and ICT) in 2022 (No. 2022- 1711152629, Functionality of Optimization Techniques in Machine Learning for SoC Estimation)

[0114] [Explanation of the symbol]

[0115] 100: Fault diagnosis device

[0116] 110: Data Collection Unit

[0117] 120: Parameter extraction unit

[0118] 130: Threshold setting unit

[0119] 140: Fault determination unit

Claims

1. A data acquisition unit that collects voltage and current data during the charging and discharging of an ESS battery module; Based on the voltage and current data collected above, the series resistance (R), which is an internal model parameter of the battery, is i ), diffusion resistance (R diff ), diffusion capacitance (C diff Parameter extraction unit (120) for extracting ); Series resistance (R) in preset normal and abnormal states i ), diffusion resistance (R diff ), diffusion capacitance (C diff A threshold setting unit that sets a failure threshold based on the deviation of ); and The above-set fault threshold and the above-extracted series resistance (R i ), diffusion resistance (R diff ), diffusion capacitance (C diff A fault determination unit that determines whether the battery module is normal or faulty by comparing ); A fault diagnosis device for an ESS battery module including 2. In Paragraph 1, The above data collection unit is a fault diagnosis device for an ESS battery module that collects voltage and current data in real time.

3. In Paragraph 2, A fault diagnosis device for an ESS battery module, wherein the above internal model parameters are extracted in real time using a Recursive Least Squares (RLS) algorithm.

4. In Paragraph 3, The above recursive least squares algorithm sets the initial SOC (State of charge), OCV (Open circuit voltage), and internal resistance values ​​of the battery module, and receives the voltage and current data as input to continuously update the error covariance and weights to dynamically estimate the state of the battery module, thereby forming a fault diagnosis device for an ESS battery module.

5. In Paragraph 3, A fault diagnosis device for an ESS battery module, wherein the fault determination unit determines the battery module to be normal when internal model parameters extracted in real time are located within a set threshold, and determines the battery module to be abnormal when they deviate from the set threshold.

6. In Paragraph 1, A fault diagnosis device for an ESS battery module, wherein the above preset normal and abnormal states are determined based on data collected by artificially inducing overcharge and overdischarge states for the battery module.

7. A method for diagnosing a fault in an ESS battery module performed by a fault diagnosis device for an ESS battery module, wherein the method comprises: A step of collecting voltage and current data during the charging / discharging of the ESS battery module; Based on the voltage and current data collected above, the series resistance (R), which is an internal model parameter of the battery, is i ), diffusion resistance (R diff ), diffusion capacitance (C diff Step of extracting ); Series resistance (R) in preset normal and abnormal states i ), diffusion resistance (R diff ), diffusion capacitance (C diff Step of setting a failure threshold based on the deviation of ); and The above-set fault threshold and the above-extracted series resistance (R i ), diffusion resistance (R diff ), diffusion capacitance (C diff A step of determining whether the battery module is normal or faulty by comparing ); A method for diagnosing a fault in an ESS battery module, including 8. In Paragraph 7, A method for diagnosing a fault in an ESS battery module, wherein the step of collecting voltage and current data includes the step of collecting voltage and current data in real time.

9. In Paragraph 8, A method for diagnosing a fault in an ESS battery module, comprising the step of extracting the internal model parameters in real time using a Recursive Least Squares (RLS) algorithm.

10. In Paragraph 9, A method for diagnosing a fault in an ESS battery module, comprising the step of extracting in real time, the step of setting the initial SOC, OCV, and internal resistance values ​​of the battery module, and the step of receiving the voltage and current data and continuously updating the error covariance and weights to dynamically estimate the state of the battery module.

11. In Paragraph 9, A method for diagnosing a fault in an ESS battery module, comprising the step of determining whether the battery module is normal or faulty, wherein if internal model parameters extracted in real time are located within a set threshold, the battery module is determined to be normal, and if they deviate from the set threshold, the battery module is determined to be abnormal.

12. In Paragraph 7, A method for diagnosing faults in an ESS battery module, wherein the above-mentioned preset normal and abnormal states are determined based on data collected by artificially inducing overcharge and overdischarge states for the battery module.

Images