BMS monitoring apparatus and method with multi-safety architecture based on AI predictive diagnosis

Abstract

The objective of the present invention, which aims to solve the aforementioned conventional problems, is to provide a BMS monitoring device and method with a multi-safety structure based on AI prediction diagnosis that can prevent accidents caused by thermal runaway of batteries applied to industrial or national public facilities even in the event of a functional error in the BMS monitoring system. To achieve the above objective, the BMS monitoring device of a multi-safety structure based on AI prediction diagnosis according to the present invention comprises: a battery unit applied for industrial or national public facilities; a BMS monitoring unit that measures the voltage, current, temperature, and charge / discharge status of the battery unit and generates control information; an AI prediction diagnosis / integrated monitoring unit that predicts the risk of thermal runaway by analyzing the correlation of operating factors including overcharging, over-discharging, overcurrent, and temperature abnormalities based on information collected from the BMS monitoring unit, and comprehensively monitors the operating status of the battery unit according to the prediction result; and an input / output cutoff unit that prevents accidents caused by thermal runaway by hardware-blocking the power input / output of the battery unit when an emergency control signal is received from the BMS monitoring unit or the AI ​​prediction diagnosis / integrated monitoring unit.

Description

Technology Field

[0001] The present invention relates to a BMS monitoring device and method with a multi-safety structure based on AI predictive diagnosis, and more specifically, to a BMS monitoring device and method with a multi-safety structure based on AI predictive diagnosis for preventing accidents caused by thermal runaway of batteries applied to industrial or national public facilities. Background Technology

[0002] Battery Management Systems (BMS) are widely utilized as a technology to ensure safety by comprehensively monitoring and controlling the charge state, voltage, current, temperature, and cell balance of secondary batteries, particularly high-capacity lithium-based batteries.

[0003] Conventional BMS monitoring systems are based on a structure that measures the electrical or thermal status of each cell and module through sensors, analyzes the measured data according to a control algorithm, and issues a cutoff command when overcharging, over-discharging, overcurrent, or overheating occurs. Such systems have been used in various application fields, ranging from energy storage systems (ESS), uninterruptible power supply (UPS), ships, and national infrastructure.

[0004] However, conventional BMS monitoring systems have structural limitations and inherent problems. First, if the BMS itself malfunctions or errors occur in its control functions, timely shutdown control for battery cells or modules may not be performed, potentially leading to serious safety accidents such as thermal runaway. In particular, due to the chemical characteristics of high-capacity lithium-based batteries, overcharging causes internal heat generation, electrolyte decomposition, and electrode swelling, which pose a high risk of leading to a chain reaction of fires or explosions.

[0005] Second, conventional BMSs fundamentally rely on reactive shutdown control and lack preventive diagnosis capabilities. In other words, because they operate based on simple threshold-based control, they fail to recognize subtle voltage, current, and temperature change patterns immediately before actual thermal runaway occurs, resulting in a limitation where the response timing is delayed.

[0006] Third, the I / O cutoff device, which operates based on control signals output by the BMS, also has a problem where it does not function properly in the event of a BMS error; this disables the entire battery system, increasing the likelihood of a large-scale accident.

[0007] Safety is recognized as an even more critical issue, particularly for high-capacity batteries used in industrial or public facilities. Lithium Iron Phosphate (LFP) batteries, for instance, offer relatively high stability and have a low risk of thermal runaway due to their chemical structure. For these reasons, LFP batteries are widely adopted in industrial and public infrastructure sectors both domestically and internationally, including large-scale energy storage systems (ESS), storage systems integrated with solar and wind power generation, port power supply systems, subway emergency power systems, national backbone network UPS, and military and marine facilities. However, LFP batteries are not entirely safe. In fact, under conditions of overcharging or prolonged high-temperature operation, there is a risk of ignition due to electrolyte degradation, internal short circuits, or electrode damage. Furthermore, in large-scale packs where multiple cells are connected in series or parallel, even minor abnormalities can lead to the explosion of the entire system.

[0008] In conventional technology, supplementary devices such as redundant sensors, multi-stage protection circuits, and fire detection sensors have been introduced to mitigate these risks; however, they have fundamentally failed to overcome the limitation of relying on the reliability of the BMS. Since a single accident involving batteries for industrial and public facilities can lead to casualties and massive property damage, it is difficult to fully ensure safety relying solely on simple BMS control functions. In particular, as confirmed by cases of ESS fire accidents in Korea, BMS malfunction, control signal transmission errors, and failure to detect thermal runaway have been identified as major causes. Therefore, there is an urgent need for a new safety management system that addresses the shortcomings of conventional technology and can operate independently to cut off battery input and output even in the event of a BMS failure. Prior art literature

[0009] Korean Patent Publication No. 10-2714330 (October 2, 2024) The problem to be solved

[0010] The objective of the present invention, which aims to solve the aforementioned conventional problems, is to provide a BMS monitoring device and method with a multi-safety structure based on AI prediction diagnosis that can prevent accidents caused by thermal runaway of batteries applied to industrial or national public facilities even in the event of a functional error in the BMS monitoring system.

[0011] In addition, the objective of the present invention is to provide a BMS monitoring device and method with a multi-safety structure based on AI predictive diagnosis that implements a preemptive response function that existing BMSs could not provide by associatively diagnosing various operating factors including overcharging, over-discharging, overcurrent, and temperature through AI analysis, and performing multi-stage safety measures by predicting the possibility of risk occurrence in advance.

[0012] Furthermore, the objective of the present invention is to provide a BMS monitoring device and method with a multi-safety structure based on AI predictive diagnosis that can eliminate a single point of failure that may occur in industrial and national public facility battery systems and dramatically improve the overall reliability and safety of the system through hardware redundancy of an input / output blocking device and a triple-redundancy safety structure with an AI monitoring system. means of solving the problem

[0013] To achieve the above objective, the BMS monitoring device of a multi-safety structure based on AI prediction diagnosis according to the present invention comprises: a battery unit applied for industrial or national public facilities; a BMS monitoring unit that measures the voltage, current, temperature, and charge / discharge status of the battery unit and generates control information; an AI prediction diagnosis / integrated monitoring unit that predicts the risk of thermal runaway by analyzing the correlation of operating factors including overcharging, over-discharging, overcurrent, and temperature abnormalities based on information collected from the BMS monitoring unit, and comprehensively monitors the operating status of the battery unit according to the prediction result; and an input / output cutoff unit that prevents accidents caused by thermal runaway by hardware-blocking the power input / output of the battery unit when an emergency control signal is received from the BMS monitoring unit or the AI ​​prediction diagnosis / integrated monitoring unit.

[0014] In addition, in the BMS monitoring device of a multi-safety structure based on AI prediction diagnosis according to the present invention, the battery unit may be characterized as being equipped with a lithium iron phosphate battery.

[0015] In addition, in the BMS monitoring device of a multi-safety structure based on AI prediction diagnosis according to the present invention, the AI ​​prediction diagnosis / integrated monitoring unit may be characterized by including a machine learning or deep learning-based prediction algorithm to multivariately analyze data on overcharging, over-discharging, overcurrent, and temperature abnormalities and classify the possibility of thermal runaway occurrence into good, caution, and danger levels.

[0016] In addition, in the BMS monitoring device of a multi-safety structure based on AI prediction diagnosis according to the present invention, the input / output cutoff unit may be characterized by including a first cutoff circuit and a second cutoff circuit in a redundant structure so that even when the first cutoff circuit fails, the second cutoff circuit operates independently to cut off the power input / output of the battery unit.

[0017] In addition, in the BMS monitoring device of a multi-safety structure based on AI prediction diagnosis according to the present invention, the AI ​​prediction diagnosis / integrated monitoring unit may be characterized by verifying the communication status between the BMS monitoring unit and the input / output blocking unit in real time, and independently controlling the input / output blocking unit even when the communication error is detected.

[0018] In addition, in the BMS monitoring device of a multi-safety structure based on AI prediction diagnosis according to the present invention, the AI ​​prediction diagnosis / integrated monitoring unit may be characterized by visually displaying the result of predicting the possibility of battery thermal runaway occurrence on a user terminal or monitoring screen so that the operator can immediately recognize it.

[0019] In addition, in the BMS monitoring device of a multi-safety structure based on AI prediction diagnosis according to the present invention, the AI ​​prediction diagnosis / integrated monitoring unit may be characterized by being linked with a cloud server or a local server to integrate and learn driving data collected from a plurality of battery units and applying an updated prediction model to each battery unit.

[0020] In addition, in the BMS monitoring device of a multi-safety structure based on AI prediction diagnosis according to the present invention, the input / output blocking unit may be characterized by having a mechanical blocking element and an electronic blocking element in parallel or in series to ensure hardware dual safety.

[0021] In addition, in the BMS monitoring device of a multi-safety structure based on AI prediction diagnosis according to the present invention, the BMS monitoring unit and the AI ​​prediction diagnosis / integration monitoring unit may include a network communication module and may be characterized by exchanging data with an external management server via TCP / IP, Ethernet, or a wireless communication protocol.

[0022] In addition, in the BMS monitoring device of a multi-safety structure based on AI prediction diagnosis according to the present invention, the battery unit, the BMS monitoring unit, and the input / output blocking unit are divided into a single set, and the AI ​​prediction diagnosis / integrated monitoring unit is characterized by operating in conjunction with a plurality of single sets to perform AI prediction diagnosis and the generation of control signals for each single set.

[0023] In addition, to achieve the above objective, the BMS monitoring device of a multi-safety structure based on another AI prediction diagnosis according to the present invention includes a battery unit applied for industrial or national public facilities, an AI prediction diagnosis / integrated monitoring unit that predicts the risk of thermal runaway by analyzing the correlation of operating factors including overcharging, over-discharging, overcurrent, and temperature abnormalities based on collected information measuring the voltage, current, temperature, and charge / discharge status of the battery unit, and integrally monitors the operating status of the battery according to the prediction result, and an input / output cutoff unit that prevents accidents caused by thermal runaway by hardware-blocking the power input / output of the battery unit when an emergency control signal is received from the AI ​​prediction diagnosis / integrated monitoring unit.

[0024] In addition, in the BMS monitoring device of a multi-safety structure based on AI prediction diagnosis according to the present invention, the battery unit may be characterized as being equipped with a lithium iron phosphate battery.

[0025] In addition, to achieve the above objective, the AI ​​predictive diagnosis-based multi-safety structure BMS monitoring method according to the present invention comprises: a first monitoring step of measuring the voltage, current, temperature, and charge / discharge status of a battery unit applied to industrial or national public facilities and generating control information in a multi-safety structure BMS monitoring device; a second monitoring step of predicting the risk of thermal runaway by analyzing the correlation of operating factors including overcharging, over-discharging, overcurrent, and temperature abnormalities based on the collected measurement information, and comprehensively monitoring the operating status of the battery according to the prediction result; and a battery input / output cutoff step of preventing an accident caused by thermal runaway by hardware-cutting off the power input / output of the battery unit when an emergency control signal is generated as a result of the first monitoring or the second monitoring.

[0026] In addition, the BMS monitoring method of a multi-safety structure based on AI prediction diagnosis according to the present invention may further include a step of multivariately analyzing data on overcharging, over-discharging, overcurrent, and temperature abnormalities by including a machine learning or deep learning-based prediction algorithm and classifying the possibility of thermal runaway occurrence into good, caution, and danger levels.

[0027] And, to achieve the above objective, the BMS monitoring method of a multi-safety structure based on another AI prediction diagnosis according to the present invention includes a monitoring step of predicting the risk of thermal runaway by analyzing the correlation of operating factors including overcharging, over-discharging, overcurrent, and temperature abnormalities based on collected information measuring the voltage, current, temperature, and charge / discharge status of a battery part applied to industrial or national public facilities in a BMS monitoring device of a multi-safety structure, and comprehensively monitoring the operating status of the battery according to the prediction result; and a battery input / output cutoff step of preventing an accident caused by thermal runaway by hardware-cutting off the power input / output of the battery part when an emergency control signal is generated as a result of the monitoring.

[0028] In addition, the BMS monitoring method of a multi-safety structure based on AI prediction diagnosis according to the present invention may further include a step of multivariately analyzing data on overcharging, over-discharging, overcurrent, and temperature abnormalities by including a machine learning or deep learning-based prediction algorithm and classifying the possibility of thermal runaway occurrence into good, caution, and danger levels.

[0029] Specific details of other embodiments are included in "Specific details for implementing the invention" and the attached "drawings".

[0030] The advantages and / or features of the present invention and the methods for achieving them will become clear by referring to the various embodiments described below in detail together with the accompanying drawings.

[0031] However, it should be understood that the present invention is not limited to the configurations of each embodiment disclosed below, but may be implemented in various different forms, and that each embodiment disclosed in this specification is provided merely to make the disclosure of the present invention complete and to fully inform those skilled in the art of the scope of the present invention, and that the present invention is defined only by the scope of each claim of the claims. Effects of the invention

[0032] The present invention can achieve the following effects through the combination and usage relationships of the embodiments of the present invention described above and the configuration described below.

[0033] According to the present invention, even in the event of a functional error in the BMS monitoring system, there is an effect of preventing accidents caused by thermal runaway of batteries applied to industrial or national public facilities.

[0034] In addition, according to the present invention, the AI ​​prediction diagnosis / integrated monitoring unit analyzes overcharging, over-discharging, overcurrent, and temperature abnormalities in a multivariate manner and diagnoses the possibility of thermal runaway occurrence by classifying it into stages (good, caution, danger) in advance, thereby having the effect of securing a higher level of preventive safety than the existing simple threshold-based cutoff method.

[0035] In addition, according to the present invention, the input / output blocking unit is hardware-redundant, or the BMS monitoring unit, AI prediction diagnosis / integration monitoring unit, and input / output blocking unit form a triple-redundancy safety structure, thereby eliminating a single point of failure and improving the reliability of the entire system.

[0036] In addition, according to the present invention, there is an effect of providing a safety structure for preventing battery explosion accidents through an AI prediction diagnosis / integrated monitoring unit excluding a BMS monitoring unit and an input / output blocking unit.

[0037] In addition, according to the present invention, in a large-capacity battery system such as a lithium iron phosphate battery applied to national infrastructure or large-scale industrial facilities, secondary damage such as large-scale power outages, communication network paralysis, and traffic disruptions can be prevented in the event of an accident, thereby ensuring public safety and the continuity of industrial operations.

[0038] In addition, according to the present invention, since the AI ​​predictive diagnosis results are visualized and displayed on a user terminal or monitoring screen, the operator can intuitively recognize dangerous situations and respond immediately.

[0039] Furthermore, according to the present invention, by linking with a cloud server or a local server to collect and learn multiple battery operation data and applying an updated prediction model to each system, there is an effect of gradually improving prediction accuracy over time. Brief explanation of the drawing

[0040] FIG. 1 is a configuration diagram showing a BMS monitoring device according to one embodiment of the present invention. Figure 2 is an example diagram showing a redundancy structure for the input / output blocking section of Figure 1. FIG. 3 is a configuration diagram showing a BMS monitoring device according to another embodiment of the present invention. FIG. 4 is a flowchart illustrating a BMS monitoring method according to one embodiment of the present invention. And, FIG. 5 is a flowchart illustrating a BMS monitoring method according to another embodiment of the present invention. Specific details for implementing the invention

[0041] Before describing the present invention in detail, it should be understood that the terms and words used in this specification should not be interpreted as being limited to their ordinary or dictionary meanings, and that the inventor of the present invention may appropriately define and use the concepts of various terms to best describe their invention, and furthermore, that these terms and words should be interpreted in a meaning and concept consistent with the technical spirit of the present invention.

[0042] In other words, it should be understood that the terms used in this specification are used merely to describe preferred embodiments of the present invention and are not intended to specifically limit the content of the present invention, and that these terms are defined in consideration of various possibilities of the present invention.

[0043] In addition, it should be noted that in this specification, singular expressions may include plural expressions unless the context clearly indicates a different meaning, and that even if they are expressed in a similarly plural form, they may include the meaning of the singular.

[0044] Throughout this specification, where it is stated that a component "includes" another component, unless specifically stated otherwise, this may mean that it does not exclude any other component but may include any other component.

[0045] Furthermore, it should be noted that where it is stated that a component "exists inside or is installed in connection with" another component, this component may be installed in direct connection or contact with the other component, or it may be installed at a certain distance apart; in the case where it is installed at a certain distance apart, a third component or means for fixing or connecting the component to the other component may exist, and a description of this third component or means may be omitted.

[0046] On the other hand, where it is stated that one component is "directly connected" or "directly connected" to another component, it should be understood that a third component or means does not exist.

[0047] Likewise, other expressions describing the relationship between each component, such as “between” and “right between”, or “adjacent to” and “directly adjacent to”, should be interpreted as having the same intent.

[0048] In addition, it should be understood that in this specification, terms such as "one side," "other side," "one side," "other side," "first," and "second," if used, are intended to clearly distinguish one component from another component, and that the meaning of the component is not restricted by such terms.

[0049] In addition, position-related terms such as "up," "down," "left," and "right" used in this specification should be understood as indicating the relative position of the corresponding component in the drawing, and unless an absolute position is specified, these position-related terms should not be understood as referring to an absolute position.

[0050] Furthermore, in specifying the reference numerals for each component of each drawing in this specification, the same component has the same reference numeral even if it is shown in different drawings; that is, the same reference numeral throughout the specification indicates the same component.

[0051] In the drawings attached to this specification, the size, position, connection relationships, etc., of each component constituting the present invention may be described in a partially exaggerated, reduced, or omitted manner for the convenience of explanation or to sufficiently clearly convey the concept of the present invention, and therefore, the proportions or scale may not be strictly accurate.

[0052] In addition, in describing the present invention below, detailed descriptions of components, such as prior art and known technology, that are deemed to unnecessarily obscure the essence of the invention may be omitted.

[0054] Preferred embodiments of the present invention will be described in detail below with reference to the drawings.

[0055] FIG. 1 is a configuration diagram showing a BMS monitoring device according to one embodiment of the present invention.

[0056] As illustrated in FIG. 1, the BMS monitoring device (100) of a multi-safety structure based on AI prediction diagnosis comprises: a battery unit (110) applied for industrial or national public facilities; a BMS monitoring unit (120) that measures the voltage, current, temperature, and charge / discharge status of the battery unit (110) and generates control information; an AI prediction diagnosis / integrated monitoring unit (130) that predicts the risk of thermal runaway by analyzing the correlation of operating factors including overcharging, over-discharging, overcurrent, and temperature abnormalities based on information collected from the BMS monitoring unit (120), and comprehensively monitors the operating status of the battery unit (110) according to the prediction result; and an input / output cutoff unit (140) that prevents accidents caused by thermal runaway by hardware-blocking the power input / output of the battery unit (110) when an emergency control signal is received from the BMS monitoring unit (120) or the AI ​​prediction diagnosis / integrated monitoring unit (130).

[0057] Here, the battery unit (110) applied for industrial or national public facilities may be equipped with a lithium iron phosphate battery. More specifically, the battery unit (110) may be hierarchically configured in units of cells, modules, and packs, and each cell may be equipped with a voltage sensor and a temperature sensor to be linked in real-time with a BMS monitoring unit (120). Lithium iron phosphate batteries have high thermal stability and low risk of fire and explosion due to their chemical properties, making them suitable for application in industrial energy storage systems (ESS) requiring large-capacity power storage, emergency power supply units for national backbone power grids, emergency power supply units for railways and subways, marine and port power supply systems, and uninterruptible power supply units (UPS) for public infrastructure such as hospitals or data centers.

[0058] According to the embodiment, the battery unit (110) may be mixed with lithium-ion (Li-ion), lithium-manganese (LiMn), lithium-nickel (LiNi), and lithium-cobalt (LiCo) batteries in addition to lithium iron phosphate batteries, and the analysis algorithm of the AI ​​prediction diagnosis / integrated monitoring unit (130) may be adjusted according to the characteristics of each battery type. In addition, the battery unit (110) may be designed to expand the number of modules to meet various capacity and output requirements, and may be configured with a structure that can be expanded from hundreds of kWh to tens of MWh, for example.

[0059] As an additional embodiment, the battery unit (110) may be combined with a liquid cooling method or an air cooling device to enable stable operation even in high or low temperature environments, and may be integrated with a thermal management system to minimize performance degradation due to temperature changes. In addition, the battery unit (110) may be equipped with a waterproof and dustproof rating (IP65 or higher) and a mechanical protection structure capable of withstanding shock and vibration so as to be applicable in harsh environments such as marine and port facilities or underground facilities.

[0060] Accordingly, the battery unit (110) can be configured to be expandable to meet various requirements for industrial and national public facilities by combining a lithium iron phosphate battery as a basic example, with various lithium-based batteries, cooling and protection systems, and an expandable modular structure.

[0061] The BMS monitoring unit (120) includes a basic sensor module and a control circuit, and can adjust the voltage imbalance of each cell through a cell balancing circuit. According to an embodiment, the BMS monitoring unit (120) includes a network communication module and can exchange data with an upper management server via TCP / IP, Ethernet, CAN communication, or wireless communication. Additionally, the BMS monitoring unit (120) can calculate diagnostic indicators such as the battery's SOC (State of Charge), SOH (State of Health), and SOP (State of Power) in real time and transmit them to the AI ​​predictive diagnosis / integrated monitoring unit (130).

[0062] The AI ​​prediction diagnosis / integrated monitoring unit (130) is equipped with a machine learning or deep learning-based prediction algorithm to analyze history such as current, voltage, temperature, and charge / discharge cycles, and can classify the possibility of thermal runaway occurrence into good, caution, and danger stages. According to an embodiment, the AI ​​prediction diagnosis / integrated monitoring unit (130) is linked with a cloud server or a local server to learn data collected from multiple battery units (110) and can continuously update the prediction model. In addition, the module visually displays the danger status on a user terminal or control monitoring screen, allowing the operator to intuitively recognize it and respond immediately.

[0063] More specifically, the AI ​​prediction diagnosis / integrated monitoring unit (130) is equipped with a machine learning or deep learning-based prediction algorithm to analyze various histories such as current, voltage, temperature, charge / discharge cycle, cell balance status, internal resistance value, and charge / discharge efficiency, and based on this, can classify the possibility of thermal runaway occurrence into good, caution, and danger stages. Unlike a simple threshold-based cutoff method, this multivariate analysis function enables more precise risk diagnosis by reflecting the correlation of complex factors.

[0064] According to an embodiment, the AI ​​prediction diagnosis / integrated monitoring unit (130) is linked with a cloud server or a local server to integrate data collected in real time from a plurality of battery units (110) and perform large-scale learning, thereby continuously improving the accuracy of the prediction algorithm. At this time, the cloud server linkage method has the advantage of being able to generate a global-level prediction model by utilizing a large-scale dataset, while the local server linkage method has the advantage of being able to perform learning and diagnosis independently even in the event of a network failure.

[0065] Additionally, the AI ​​predictive diagnosis / integrated monitoring unit (130) can visually display the danger status on a user terminal, a central control monitoring screen, a mobile application, or an AR (augmented reality) based interface. For example, a good level is displayed in green, a caution level in yellow, and a danger level in red, and when entering a danger stage, the operator can immediately recognize it through multiple channels such as an alarm sound, vibration notification, and voice guidance.

[0066] Additionally, the AI ​​predictive diagnosis / integrated monitoring unit (130) can visually display the risk status on a user terminal, a central control monitoring screen, a mobile application, or an AR (augmented reality)-based interface. More specifically, the AI ​​predictive diagnosis / integrated monitoring unit (130) can configure a digital twin screen that maps the battery unit (110) into a hierarchical structure of cell-module-pack units, and can express the risk level for each layer by combining colors and numerical indicators. For example, a good level is displayed in green, a caution level in yellow, and a risk level in red, and by displaying auxiliary indicators such as the predicted risk score, confidence, and rate of increase (e.g., dT / dt, dV / dt) together on the same screen, the operator can intuitively recognize the magnitude of the risk and the trend of progress.

[0067] In mobile applications and wearable terminals, vibration patterns, short alarm sounds, and voice guidance are provided in parallel along with the same color scheme, and in the central control monitoring screen, a layout for a large board is applied so that the status of battery units (110) at multiple sites can be viewed simultaneously in a tile format. In an embodiment of an AR-based interface, when a field worker points a camera at the exterior of a battery pack, the location of a risk cell in module units is projected onto a real-world image as an outline and color overlay, and when entering a risk level, a flashing red border, a direction arrow, and voice instructions (e.g., “Rack 3, 2nd floor module needs to be blocked”) can be output sequentially. This multi-channel alarm operates based on a three-level system of Good, Caution, and Danger, and when entering a risk level, an alarm sound, vibration, and voice guidance are simultaneously activated to ensure the operator's immediate response.

[0068] The judgment logic of the AI ​​prediction diagnosis / integrated monitoring unit (130) can be implemented as a dual path combining a prediction model and safety rules. The first path is a risk calculation using machine learning / deep learning, and can use multivariate time series features as input values, such as current, voltage, temperature, and charge / discharge cycle, as well as cell internal resistance estimates, cell balance correction amounts, recent charge / discharge efficiency degradation rates, and pattern-based outlier indicators (e.g., reconstruction error). These are preprocessed into moving averages, exponentially weighted moving averages (EWMA), first and second differences (Δ, Δ²), temperature rise rates (dT / dt), voltage fluctuation ranges (peak-to-peak), and column counting residuals (accumulated charge-voltage consistency error), and noise can be mitigated by a low-pass filter or a Kalman filter.

[0069] Examples of models related to this include an ensemble structure trained in parallel with a 1D-CNN / Temporal Convolutional Network sensitive to short-term rapid changes, an LSTM / GRU reflecting medium-to-long-term trends, and a Gradient-Boosting (XGBoost / LightGBM) dealing with non-time series derived features.

[0070] When the output probability of each model is provided with a 'prediction and rule' combined structure, it can be designed to prioritize safety measures even if the information reliability of the BMS monitoring unit (120) is reduced or there is a communication error.

[0071] As an example of display and notification, the user terminal may present a circular risk gauge, stage color, top-level features based on the prediction, and recommended actions (load reduction / forced cooling / preparation for shutdown). On the central control monitoring screen, the status of racks / modules at multiple sites is updated in real-time in a grid format, and racks entering a risk stage can be automatically magnified with priority sorting, flashing effects, and an alarm history timeline. The mobile application sends push notifications only to personnel within the site radius via geofencing, and the vibration patterns of the wearable terminal are set differently for each stage (e.g., Caution = short 2 consecutive, Danger = long-short-long 3 consecutive). The AR interface displays a red outline and arrow overlaid on the front of the module, overlays a shutdown procedure checklist step by step, and voice guidance can support multiple languages ​​and auditory assistance modes (subtitles / high contrast). All alarm events are transmitted simultaneously via visual, auditory, and tactile channels, and to suppress false alarms, cumulative delays and grouping rules can be applied to repeated notifications of the same event. The results of notification reception and actions taken are recorded in an immutable audit log and can be used as a basis for subsequent cause analysis and model retraining.

[0072] The AI ​​predictive diagnosis / integrated monitoring unit (130) can be operated as either a field edge device or a cloud server, or a combination of both. In the edge operation embodiment, a completely independent inference and rule engine is operated considering the public facility environment where the network is unstable, and the model can be updated only periodically or conditionally (nighttime / low load time). In the cloud combination embodiment, operation data from multiple sites is aggregated so that the model is periodically retrained, and after verification, it can be distributed to each field equipment through gradual rollout.

[0073] When a communication failure is detected, the AI ​​predictive diagnosis / integrated monitoring unit (130) switches to a safe mode to control the input / output blocking unit (140) using only local rule-based interlocks, and synchronization with the central server can be reflected with a delay upon recovery. Additionally, periodic self-diagnosis items (sensor reliability check, time synchronization, communication delay, and prediction of relay lifespan of the input / output blocking unit (140)) are included to provide advance warning of potential defects and automatically suggest a maintenance schedule.

[0074] Consequently, the above configuration consistently implements the three-stage risk visualization and pre-emptive blocking system, which is the core of the present invention, across all stages of UI / notification, inference, and control, and can be combined with a redundant / triple redundant safety structure that executes control through an input / output blocking device even in the event of a BMS operation error or failure.

[0075] As an extended embodiment, the AI ​​predictive diagnosis / integrated monitoring unit (130) may include a real-time anomaly detection function. For example, if the voltage of a specific cell changes rapidly or the charge / discharge efficiency decreases abnormally, the system may recognize this as a pre-thermal runaway signal rather than a simple warning and execute an early shutdown command. Additionally, the AI ​​predictive diagnosis / integrated monitoring unit (130) can support battery replacement cycle prediction and maintenance optimization by tracking the battery's condition over the long term and calculating the State of Health (SOH), Remaining Useful Life (RUL), etc.

[0076] Furthermore, the AI ​​prediction diagnosis / integrated monitoring unit (130) can implement a network-type safety management system by linking with a central server to comprehensively monitor the status of each site even when industrial or national public facility battery systems are distributed across multiple regions, and by transmitting abnormal patterns occurring in specific regions as advance warning signals to other regional systems. For example, if a specific pattern detected in a port power supply system can act as a risk factor for a subway emergency power supply device, this can be propagated in real time to enable a preemptive response.

[0077] That is, when the battery unit (110), BMS monitoring unit (120), and input / output blocking unit (140) are combined into a single set, it means that the AI ​​prediction diagnosis / integrated monitoring unit (130) can be linked with multiple single sets to perform AI prediction diagnosis and generate control signals for each single set.

[0078] The AI ​​prediction diagnosis / integrated monitoring unit (130) can be expanded beyond simply monitoring battery status to include prediction-based pre-diagnosis, continuous model learning, multiple visualizations and notifications, anomaly detection, remaining life management, and distributed network-type safety management, thereby maximizing the safety and operational efficiency of industrial and national public facility battery systems.

[0079] Additionally, the AI ​​prediction diagnosis / integrated monitoring unit (130) can verify the communication status between the BMS monitoring unit (120) and the input / output blocking unit (140) in real time, and can independently control the input / output blocking unit (140) even when a communication error is detected, and can also be implemented to independently control the input / output blocking unit (140) even when the communication status between the BMS monitoring unit (120) and the input / output blocking unit (140) is normal.

[0080] Accordingly, the BMS monitoring device (100) according to the present invention can prevent thermal runaway of the battery unit (110) through independent control even in the event of a functional error or communication failure of a single module, and by combining a multiplexed safety structure and an AI-based predictive diagnostic function, it can secure enhanced safety and reliability compared to the existing simple monitoring and blocking method.

[0081] Figure 2 is an example diagram showing a redundancy structure for the input / output blocking section of Figure 1.

[0082] As illustrated in FIG. 2, the input / output blocking unit (140) may be equipped with a first blocking circuit and a second blocking circuit in a redundant manner, and may provide hardware-based dual safety by combining an electronic blocking element (e.g., MOSFET, IGBT) and a mechanical blocking element (e.g., relay, circuit breaker) in parallel or in series. According to an embodiment, the input / output blocking unit (140) receives control commands independently from the BMS monitoring unit (120) and the AI ​​prediction diagnosis / integration monitoring unit (130), respectively, so as to be able to reliably block the input / output of the battery unit (110) even in the event of an error in a single module. In addition, it may be designed to independently perform an emergency blocking function even in the event of a communication error or power abnormality.

[0083] More specifically, the first circuit breaker is composed of an electronic circuit breaker capable of high-speed switching, allowing it to operate quickly in situations requiring an immediate response, such as overcurrent, overcharging, or overdischarging, while the second circuit breaker is composed of a mechanical circuit breaker, enabling it to physically isolate the circuit independently even in the event of malfunction, short circuit, or burnout of the electronic circuit breaker. When these two circuits are arranged in a series combination, safety can be maximized through double interruption, and when arranged in a parallel combination, the other circuit can bypass and interrupt the current in the event of a failure in one circuit, thereby ensuring the continuity of the system.

[0084] According to an embodiment, the input / output blocking unit (140) may be designed to operate independently even when there is a system power failure by providing a separate power supply module for each blocking circuit. Additionally, each blocking circuit may include a self-diagnostic function to periodically check for component degradation, contact welding, and response delay, and transmit a warning signal to the AI ​​prediction diagnosis / integrated monitoring unit (130) when an abnormal condition is detected.

[0085] In an extended embodiment, the input / output cutoff unit (140) may be configured to include a surge protection circuit, a fuse, and a thermistor (NTC) to mitigate current spikes and voltage surges that may occur during cutoff operation. Additionally, it may be combined with a cooling device (heat sink, fan, liquid cooling, etc.) to maintain stable cutoff performance even in high-power environments.

[0086] Furthermore, in addition to the structure in which the input / output blocking unit (140) independently receives control commands from the BMS monitoring unit (120) and the AI ​​prediction diagnosis / integration monitoring unit (130), it can also automatically detect a physical trigger (e.g., temperature fuse, pressure switch) in an emergency situation and perform a blocking operation. Through this, the input / output of the battery unit (110) can be safely blocked even in unexpected situations such as communication errors, control signal delays, or software failures.

[0087] Accordingly, the input / output blocking unit (140) according to the present invention can satisfy the high reliability and high safety required for industrial and national public facility battery systems through extended embodiments such as a redundancy structure, self-diagnosis, multiple power independent driving, surge protection and cooling, and physical trigger-based emergency shutdown, going beyond a simple switching function.

[0088] FIG. 3 is a configuration diagram showing a BMS monitoring device according to another embodiment of the present invention.

[0089] As illustrated in FIG. 3, the BMS monitoring device (200) of a multi-safety structure based on AI prediction diagnosis includes a battery unit (210) applied for industrial or national public facilities, an AI prediction diagnosis / integrated monitoring unit (220) that predicts the risk of thermal runaway by analyzing the correlation of operating factors including overcharging, over-discharging, overcurrent and temperature abnormalities based on collected information measuring the voltage, current, temperature and charge / discharge status of the battery unit (210), and comprehensively monitors the operating status of the battery unit (210) according to the prediction result, and an input / output cutoff unit (230) that prevents accidents caused by thermal runaway by hardware-blocking the power input / output of the battery unit (210) when an emergency control signal is received from the AI ​​prediction diagnosis / integrated monitoring unit (220).

[0090] More specifically, the battery unit (210) can be based on a lithium iron phosphate battery and can be expandedly applied to public infrastructure and industrial sites, such as large-capacity energy storage systems (ESS), emergency power sources for railways and subways, port power supply devices, data centers, and uninterruptible power supply devices for hospitals. According to an embodiment, the battery unit (210) may be provided in a pack structure in which multiple modules are connected in parallel or series rather than as a single module, and each module may be expanded to a structure that includes independent voltage and temperature sensors and transmits to an AI prediction diagnosis / integrated monitoring unit (220).

[0091] The AI ​​prediction diagnosis / integrated monitoring unit (220) can analyze collected data by applying machine learning or deep learning-based algorithms and classify the risk of thermal runaway into good, caution, and danger levels. As an example, the AI ​​prediction diagnosis / integrated monitoring unit (220) can calculate the State of Health (SOH), State of Charge (SOC), and Remaining Useful Life (RUL) of the battery unit (210) to predict not only short-term risks but also long-term maintenance and replacement times. In addition, this module can be linked with a cloud server or a local server to collect and learn data from multiple battery systems, learn new thermal runaway patterns or abnormal operation patterns, and apply the updated model to each device. Through this, a battery safety network can be established nationwide or at the level of multiple facilities by linking with a central control system at the public infrastructure level.

[0092] The input / output blocking unit (230) receives a direct control command from the AI ​​prediction diagnosis / integrated monitoring unit (220) to block the power input / output of the battery unit (210), and can ensure hardware safety by configuring a mixture of electronic components (e.g., MOSFET, IGBT) and mechanical components (e.g., relay, circuit breaker). Additionally, according to the embodiment, the input / output blocking unit (230) is equipped with its own independent driving power, so that it can perform an emergency blocking function even in the event of a power failure of the entire system. Furthermore, the blocking unit (230) includes a self-diagnosis function to constantly check for contact deterioration, response speed delay, and communication failure, and can transmit an alarm signal to the AI ​​prediction diagnosis / integrated monitoring unit (220) if an abnormality is detected.

[0093] As an extended embodiment, the BMS monitoring device (200) of FIG. 3 allows an operator to manually check danger signals and issue a direct shutdown command through remote communication with a central control room, and can be linked with mobile terminals, wearable devices, and AR (augmented reality)-based equipment to allow field workers to intuitively check the danger status and perform an immediate response. In addition, the BMS monitoring device (200) can customize the shutdown logic and threshold values ​​according to the characteristics of each facility, so that, for example, short-circuit protection can be strengthened in port power supply devices and uninterruptible switching performance can be guaranteed in hospital UPSs.

[0094] Accordingly, the BMS monitoring device (200) according to another embodiment of the present invention illustrated in FIG. 3 can significantly improve the safety and reliability of industrial and national public facility battery systems through not only basic BMS monitoring functions but also AI-based predictive diagnosis, cloud-linked learning, central control and field response support, self-diagnosis and independent driving functions of the input / output blocking unit.

[0095] FIG. 4 is a flowchart illustrating a BMS monitoring method according to one embodiment of the present invention.

[0096] As illustrated in FIG. 4, the AI ​​predictive diagnosis-based multi-safety structure BMS monitoring method proceeds by performing primary monitoring in the BMS monitoring unit (110) of the multi-safety structure BMS monitoring device (100) to measure the voltage, current, temperature, and charge / discharge status of the battery unit applied for industrial or national public facilities and to generate control information (S100).

[0097] The information measured in step S100 is transmitted to the AI ​​predictive diagnosis / integrated monitoring unit (130) of the BMS monitoring device (100) of the multi-safety structure (S102), and the AI ​​predictive diagnosis / integrated monitoring unit (130) analyzes the correlation of operating factors including overcharging, over-discharging, overcurrent and temperature abnormalities based on the collected measurement information to predict the risk of thermal runaway, and performs secondary monitoring to comprehensively monitor the operating state of the battery according to the prediction result (S104).

[0098] In the BMS monitoring device (100), when an emergency control signal is generated, the generated emergency control signal is transmitted to the input / output blocking unit (140), causing the input / output blocking unit (140) to hardware-block the power input / output of the battery unit (110) (S106 to S110).

[0099] In addition, when an emergency control signal is generated in the AI ​​prediction diagnosis / integrated monitoring unit (130), the generated emergency control signal is transmitted to the input / output blocking unit (140), causing the input / output blocking unit (140) to hardware-block the power input / output of the battery unit (110) (S112 to S116).

[0100] Here, as steps S106 to S110 and S112 to S116 are executed independently at the time of generating the emergency control signal, the order of the steps may be changed in correspondence with the time of generating the emergency control signal.

[0101] The detailed description of steps S100 through S116 described above and the description of additional possible steps shall be in accordance with FIGS. 1 and FIGS. 2 and the description of these figures.

[0102] FIG. 5 is a flowchart illustrating a BMS monitoring method according to another embodiment of the present invention.

[0103] As illustrated in FIG. 5, the AI ​​predictive diagnosis-based multi-safety structure BMS monitoring method proceeds by performing primary monitoring in the BMS monitoring unit (110) of the multi-safety structure BMS monitoring device (100) to measure the voltage, current, temperature, and charge / discharge status of the battery unit applied for industrial or national public facilities and to generate control information (S100).

[0104] As illustrated in FIG. 5, the AI ​​predictive diagnosis-based multi-safety structure BMS monitoring method proceeds by analyzing the correlation of operating factors including overcharging, over-discharging, overcurrent and temperature abnormalities based on information measured from the voltage, current, temperature and charge / discharge status of the battery part applied to industrial or national public facilities in the AI ​​predictive diagnosis / integrated monitoring unit (230) of the multi-safety structure BMS monitoring device (200), predicting the risk of thermal runaway, and integrally monitoring the operating status of the battery according to the prediction result (S200).

[0105] Afterwards, when an emergency control signal is generated, the generated emergency control signal is transmitted to the input / output blocking unit (240), causing the input / output blocking unit (240) to hardware-block the power input / output of the battery unit (210) (S202 to S206).

[0106] The detailed description of steps S200 through S206 described above and the description of additional possible steps shall be in accordance with the description of FIG. 3 and this figure.

[0107] Although various preferred embodiments of the present invention have been described above with some examples, the descriptions of various embodiments described in the "Specific details for carrying out the invention" section are merely illustrative, and those skilled in the art to which the present invention pertains will understand that the present invention can be modified in various ways or equivalent embodiments can be carried out based on the above description.

[0108] In addition, since the present invention can be implemented in various other forms, the present invention is not limited by the description above. The above description is provided merely to make the disclosure of the present invention complete and to fully inform those skilled in the art of the scope of the present invention, and it should be understood that the present invention is defined only by each claim of the claims. Explanation of the symbols

[0109] 100, 200: BMS monitoring device 110, 210: Battery section 120 : BMS Monitoring Section 130, 230: AI Predictive Diagnosis / Integrated Monitoring Department 140, 240: I / O blocking section

Claims

Claim 1 A battery unit applied for industrial or national public facilities; a BMS monitoring unit that measures the voltage, current, temperature, and charge / discharge status of the battery unit and generates control information; and an AI prediction diagnosis / integrated monitoring unit that analyzes the correlation of multiple operating factors, including overcharging, over-discharging, overcurrent, and temperature abnormalities, based on information collected from the BMS monitoring unit using artificial intelligence to predict the abnormal state or thermal runaway risk of the battery unit, and outputs the thermal runaway risk classified into multiple stages. A BMS monitoring device with an AI prediction diagnosis-based multi-safety structure, comprising an input / output cutoff unit configured to cut off power input / output of the battery unit according to an emergency control signal generated from the BMS monitoring unit or the AI ​​prediction diagnosis / integrated monitoring unit, wherein the input / output cutoff unit is configured to hardware cut off power input / output of the battery unit based on an emergency control signal generated from the BMS monitoring unit or the AI ​​prediction diagnosis / integrated monitoring unit, and is configured to physically cut off power input / output of the battery unit separately from cutoff control based on control information generated by the BMS monitoring unit when the thermal runaway risk predicted by the AI ​​prediction diagnosis / integrated monitoring unit is determined to be above a predetermined level, at a stage prior to the actual occurrence of overcharging, and wherein the input / output cutoff unit is configured to hardware cut off power input / output of the battery unit even in the event of a malfunction of the BMS monitoring unit or the AI ​​prediction diagnosis / integrated monitoring unit, an abnormality in measurement information regarding the voltage, current, temperature, or charge / discharge status of the battery unit, a communication abnormality between the BMS monitoring unit and the input / output cutoff unit, or a control error of the BMS monitoring unit. Claim 2 A BMS monitoring device with an AI predictive diagnosis-based multi-safety structure, characterized in that, in claim 1, the battery unit is equipped with a lithium iron phosphate battery. Claim 3 A BMS monitoring device of a multi-safety structure based on AI prediction diagnosis according to claim 1 or 2, wherein the AI ​​prediction diagnosis / integrated monitoring unit includes a machine learning or deep learning-based prediction algorithm to multivariately analyze data on overcharging, over-discharging, overcurrent, and temperature abnormalities, and classify the possibility of thermal runaway occurrence into good, caution, and danger levels. Claim 4 A BMS monitoring device with a multi-safety structure based on AI prediction diagnosis, characterized in that, in claim 1 or 2, the AI ​​prediction diagnosis / integrated monitoring unit verifies the communication status between the BMS monitoring unit and the input / output blocking unit in real time, and independently controls the input / output blocking unit even when communication between the BMS monitoring unit and the input / output blocking unit is not performed normally. Claim 5 A BMS monitoring device with a multi-safety structure based on AI prediction diagnosis, characterized in that, in claim 1 or 2, the input / output blocking unit is equipped with a mechanical blocking element and an electronic blocking element in parallel or in series to ensure hardware dual safety. Claim 6 A BMS monitoring device with a multi-safety structure based on AI prediction diagnosis, characterized in that, in claim 1 or 2, the battery unit, the BMS monitoring unit, and the input / output blocking unit are divided into a single set, and the AI ​​prediction diagnosis / integrated monitoring unit is linked with a plurality of single sets to execute the generation of AI prediction diagnosis and control signals for each single set. Claim 7 A battery unit applied for industrial or national public facilities; an AI predictive diagnosis / integrated monitoring unit that predicts the risk of thermal runaway by analyzing the correlation of operating factors including overcharging, overdischarging, overcurrent, and temperature abnormalities based on collected information measuring the voltage, current, temperature, and charge / discharge status of the battery unit, and comprehensively monitors the operating status of the battery unit according to the prediction result; and includes an input / output cutoff unit configured to cut off the power input / output of the battery unit according to an emergency control signal generated from the AI ​​prediction diagnosis / integrated monitoring unit, wherein the AI ​​prediction diagnosis / integrated monitoring unit analyzes the correlation of a plurality of operating factors, including overcharging, over-discharging, overcurrent, and temperature abnormalities, based on artificial intelligence to predict the abnormal state or thermal runaway risk of the battery unit and is configured to output the thermal runaway risk classified into a plurality of stages; the input / output cutoff unit is configured to hardware-cut off the power input / output of the battery unit based on the emergency control signal generated from the AI ​​prediction diagnosis / integrated monitoring unit; and when the thermal runaway risk predicted by the AI ​​prediction diagnosis / integrated monitoring unit is determined to be above a predetermined stage, the power input / output of the battery unit is configured to physically cut off the power input / output of the battery unit separately from the monitoring or alarm output operation performed by the AI ​​prediction diagnosis / integrated monitoring unit at a stage prior to the actual occurrence of overcharging, and the malfunction of the AI ​​prediction diagnosis / integrated monitoring unit, an abnormality in measurement information regarding the voltage, current, temperature, or charge / discharge status of the battery unit, and the AI ​​prediction diagnosis / integrated A BMS monitoring device with a multi-safety structure based on AI prediction diagnosis, characterized in that the input / output blocking unit is configured to hardware-block the power input / output of the battery unit even when a communication abnormality occurs between the monitoring unit and the input / output blocking unit, or when a control error occurs in the AI ​​prediction diagnosis / integrated monitoring unit. Claim 8 A BMS monitoring device with an AI predictive diagnosis-based multi-safety structure, characterized in that, in claim 7, the battery unit is equipped with a lithium iron phosphate battery. Claim 9 A BMS monitoring device with a multi-safety structure comprising: a first monitoring step for measuring the voltage, current, temperature, and charge / discharge status of a battery unit applied to industrial or national public facilities and generating control information; a second monitoring step for comprehensively monitoring the operating status of the battery unit by analyzing the correlation of multiple operating factors, including overcharging, over-discharging, overcurrent, and temperature abnormalities, based on the collected measurement information using artificial intelligence to predict the abnormal state or thermal runaway risk of the battery unit, and outputting the thermal runaway risk classified into multiple stages. A BMS monitoring method with a multi-safety structure based on AI prediction diagnosis, characterized by preventing accidents caused by thermal runaway of the battery unit even when an emergency control signal is generated according to the results of the first monitoring or the second monitoring, or when the risk of thermal runaway predicted in the second monitoring step is determined to be above a predetermined level, by performing a battery input / output cutoff step to preemptively hardware-cut off the power input / output of the battery unit separately from the cutoff control based on the control information generated in the first monitoring step, prior to the actual occurrence of overcharging, thereby preventing accidents caused by thermal runaway of the battery unit even when there is a malfunction of the first monitoring step or the second monitoring step, an abnormality in measurement information regarding the voltage, current, temperature, or charge / discharge status of the battery unit, a communication abnormality between the BMS monitoring unit performing the first monitoring step and the input / output cutoff unit, or a cutoff control error based on the control information generated in the first monitoring step. Claim 10 A BMS monitoring method of an AI predictive diagnosis-based multi-safety structure, further comprising the step of multivariately analyzing data on overcharging, over-discharging, overcurrent, and temperature abnormalities, including a machine learning or deep learning-based prediction algorithm, and classifying the probability of thermal runaway occurrence into good, caution, and danger levels. Claim 11 A monitoring step in which, in a BMS monitoring device of a multi-safety structure, the correlation of multiple operating factors including overcharging, over-discharging, overcurrent, and temperature abnormalities is analyzed based on artificial intelligence according to collected information measuring the voltage, current, temperature, and charge / discharge status of a battery part applied to industrial or national public facilities, thereby predicting the abnormal state or thermal runaway risk of said battery part, and outputting the thermal runaway risk classified into multiple stages to comprehensively monitor the operating status of said battery part; A BMS monitoring method of an AI prediction diagnosis-based multi-safety structure comprising a battery input / output cutoff step that prevents accidents caused by thermal runaway of the battery unit even when an emergency control signal is generated as a result of monitoring or when the predicted thermal runaway risk level is determined to be above a predetermined level, by preemptively hardware-cutting the power input / output of the battery unit at a stage prior to actual occurrence of overcharging, separately from the monitoring or alarm output operation performed in the monitoring step, thereby preventing accidents caused by thermal runaway of the battery unit even when there is a malfunction of the monitoring step, an abnormality in measurement information regarding the voltage, current, temperature, or charge / discharge status of the battery unit, a communication abnormality between the AI ​​prediction diagnosis / integrated monitoring unit performing the monitoring step and the input / output cutoff unit, or a control error of the AI ​​prediction diagnosis / integrated monitoring unit. Claim 12 A BMS monitoring method of an AI predictive diagnosis-based multi-safety structure, further comprising the step of multivariately analyzing data on overcharging, over-discharging, overcurrent, and temperature abnormalities, including a machine learning or deep learning-based prediction algorithm, and classifying the probability of thermal runaway occurrence into good, caution, and danger levels.

Images