Method and apparatus for explainable electric vehicle battery diagnosis using ontology rule-based ai inference

Abstract

An explanatory electric vehicle battery diagnosis method using ontology rule-based artificial intelligence inference according to one embodiment comprises the steps of: calculating a current state value from state data of an electric vehicle battery using an artificial intelligence model; mapping the current state value to entities and attributes of an ontology to create an entity instance corresponding to a predefined state class within the ontology and registering the entity instance in an ABox of the ontology; executing an ontology inference rule based on a combination of multiple entity instances registered in the ABox of the ontology to infer whether the electric vehicle battery is abnormal; and generating an explanatory diagnosis report based on the inferred abnormality of the electric vehicle battery. By converting the state data of the electric vehicle battery into a diagnosis result capable of explaining not only simple numerical values ​​but also risk causes and mechanisms through an ontology-based causal inference structure, the method enables the simultaneous early prediction, explanation, and verification of battery abnormalities.

Description

Technology Field

[0001] The present invention relates to a technology for diagnosing the condition of an electric vehicle battery and explaining the cause of an abnormality, and to a method for providing a service that maps electric vehicle battery condition data and AI-based diagnostic results to an ontology rule-based structure to causally infer, and provides the inference result as explanatory diagnostic information, and to an apparatus for performing the same. Background Technology

[0002] As the adoption of electric vehicles expands, ensuring the stability of battery systems, which are directly linked to vehicle safety, is emerging as a critical technological challenge. Lithium-ion batteries undergo internal degradation depending on charging speed, usage environment, driving patterns, and aging conditions; this degradation directly leads to safety issues such as fire, performance degradation, and reduced driving range. Consequently, there is a continuously increasing need for technologies that can accurately diagnose the condition of electric vehicle batteries, detect abnormal signs in advance, and implement preventive measures.

[0003] Conventional electric vehicle battery diagnostic technologies primarily evaluate the condition of the battery by analyzing numerical data such as voltage, current, temperature, and SOC collected from the BMS. Some technologies utilize machine learning or deep learning-based predictive models to calculate SOH, RUL, and the probability of anomalies. While these technologies are useful for quantitatively predicting battery degradation trends, they have limitations in that they cannot explain the causes and pathways through which the diagnostic results were derived.

[0004] Stakeholders such as users, mechanics, and insurance companies require clear explanatory information regarding not only whether a battery is faulty, but also the cause of the fault, the level of risk, and the direction of future actions. However, existing AI-based diagnostic technologies provide only numerical results, leading to problems with low interpretability and reliability of the diagnostic findings.

[0005] In addition, existing technologies have low scalability and maintainability of diagnostic structures because they must rely on model retraining when new degradation mechanisms or operating conditions are added. Accordingly, there is a need for a new diagnostic structure that can structurally interpret AI diagnostic results and provide diagnostic information that can be explained based on causal relationships.

[0006] Against this backdrop, there is an increasing demand for new diagnostic technologies that manage electric vehicle battery status data as a structured knowledge system and provide the causes of anomalies in an explainable form through rule-based causal reasoning.

[0007] Conventional technology refers to technical information that the inventor possessed for the derivation of the present invention or acquired during the process of deriving the present invention, and it cannot necessarily be considered publicly known technology disclosed to the general public prior to the filing of the present invention. Prior art literature

[0008] Registered Patent Publication No. 10-2904823 (Registered Dec. 22, 2025) Registered Patent Publication No. 10-2876216 (Registered Oct. 21, 2025) The problem to be solved

[0009] In resolving the aforementioned problems, the present invention has as its technical objective the implementation of a diagnostic structure that systematizes the components, state, environmental conditions, and degradation phenomena of an electric vehicle battery into an ontology-based knowledge structure, maps real-time or stored battery data and AI-based diagnostic results consistently to said knowledge structure, causally derives the cause of abnormal conditions through a rule-based inference engine, and provides the derived inference results as explanatory diagnostic information understandable to stakeholders such as users, mechanics, and insurance companies.

[0010] The problems that the present invention aims to solve are not limited to those mentioned above, and other problems not mentioned will be clearly understood by a person skilled in the art to which the present invention belongs from the description below. means of solving the problem

[0011] A descriptive electric vehicle battery diagnosis method using ontology rule-based artificial intelligence inference according to one embodiment may include the steps of: calculating a current state value from state data of an electric vehicle battery using an artificial intelligence model; mapping the current state value to entities and attributes of an ontology to create an entity instance corresponding to a predefined state class within the ontology and registering the entity instance in an ABox of the ontology; executing an ontology inference rule based on a combination of multiple entity instances registered in the ABox of the ontology to infer whether the electric vehicle battery is abnormal; and generating a descriptive diagnosis report based on the inferred abnormality of the electric vehicle battery.

[0012] According to one embodiment, the state data of the electric vehicle battery may include at least one of data acquired in real time during the operation of the electric vehicle and battery state data stored as operation history.

[0013] According to one embodiment, the current state value may include at least one of the voltage, current, temperature, C-rate, DOD (Depth of Discharge), impedance, SOC (State of Charge), SOH (State of Health), RUL (Remaining Useful Life), charge / discharge pattern, and anomaly detection result of the electric vehicle battery.

[0014] According to one embodiment, the ontology may be defined as a Web Ontology Language (OWL) structure including a state indicator entity layer representing the current state of the electric vehicle battery, a degradation cause entity layer configured in parallel with the state indicator entity layer to represent the internal degradation mechanism of the battery, and an event entity layer configured in parallel with the state indicator entity layer and the degradation cause entity layer to represent stress events according to the operating conditions, charging and discharging conditions, and environmental conditions of the electric vehicle.

[0015] According to one embodiment, the step of creating an object instance is to create an object instance of a corresponding state class only when the diagnostic result exceeds a preset threshold condition, wherein the object instance may be composed of an object representing at least one current state among an overheated state, an impedance rise state, a cell voltage imbalance state, a rapid charging stress state, and a deep discharge state of an electric vehicle battery.

[0016] According to one embodiment, the object instance may be configured to be activated when the current state value exceeds a preset threshold condition, and to be deactivated or updated when the current state value returns to below the threshold condition, thereby reflecting the history of state changes of the electric vehicle battery in a time-series manner.

[0017] According to one embodiment, the ontology inference rule may be composed of a predefined set of rules to causally derive an abnormal state, a cause of degradation, or a risk type of an electric vehicle battery, based on a combination of state indicators, degradation-related parameters, stress factors, and environmental variables of the electric vehicle battery.

[0018] According to one embodiment, the ontology inference rule can be executed independently without a diagnostic algorithm of the artificial intelligence model.

[0019] According to one embodiment, the ontology inference rule can derive degradation patterns and risk types between vehicle groups by clustering them by simultaneously referencing entity instances for battery status data of multiple electric vehicles.

[0020] According to one embodiment, the rule set may include a multi-layered causal rule structure divided into a first rule layer expressing structural relationships between battery components, a second rule layer expressing causal relationships between degradation patterns, and a third rule layer expressing causal relationships between operating environment variables and risk types.

[0021] According to one embodiment, the rule set may be described in a rule language based on SWRL (Semantic Web Rule Language) or SPARQL and configured to causally infer the cause of an abnormality or the type of risk of an electric vehicle battery based on a combination of conditions between multiple object instances registered in the ABox of the ontology.

[0022] According to one embodiment, the step of registering an object instance in an ABox of an ontology may include the step of mapping an item of a current state value to at least one of a predefined data attribute or object attribute within the ontology, and the step of registering attribute information regarding an electric vehicle battery in the ontology by setting a diagnostic result to the corresponding attribute.

[0023] According to one embodiment, the step of inferring whether an electric vehicle battery is abnormal may include: identifying a combination of object instances that satisfies a preset combination condition among a plurality of object instances registered in an ABox; selecting an ontology inference rule corresponding to the identified combination of object instances; executing the selected ontology inference rule to generate a risk object corresponding to an abnormal type of the electric vehicle battery; and determining whether the electric vehicle battery is abnormal based on the risk level of the generated risk object.

[0024] According to one embodiment, the step of identifying an object instance combination may include determining whether at least two object instances among an overheated state object, an impedance rise state object, a voltage imbalance state object, a rapid charge stress state object, and a deep discharge state object exist simultaneously among object instances registered in ABox, and if the determination result is true, identifying a set containing two or more object instances as an object instance combination.

[0025] According to one embodiment, the step of generating a descriptive diagnostic report may include: selecting a basic descriptive template from a first descriptive template layer defined within an ontology to correspond to a type of risk entity; selecting an output template from a second descriptive template layer defined in an external system to correspond to a user type; and converting the descriptive structure of the basic descriptive template to match the format of the output template to generate a descriptive diagnostic report including the inference basis of the risk entity.

[0026] According to one embodiment, the descriptive diagnostic report can be converted into at least one control command or recommendation command among a charging limit, a change in driving conditions, a maintenance reservation request, or an insurance warning notification, and transmitted to an external system.

[0027] According to one embodiment, different output templates are applied according to the user type and the roles of stakeholders, such as users, mechanics, and insurance companies, to provide customized diagnostic information.

[0028] According to one embodiment, the descriptive diagnostic report may be output in at least one format among JSON, HTML, or API response format.

[0029] According to one embodiment, the step of inferring whether an electric vehicle battery is abnormal may include: obtaining a diagnosis result calculated by an artificial intelligence model and an inference result calculated by an ontology inference rule; comparing the consistency of abnormality types or risk grades between the diagnosis result and the inference result; correcting the warning grade regarding whether the electric vehicle battery is abnormal if, as a result of the comparison, the abnormality type is different or the difference in risk grades exceeds a preset standard; and adding a supplementary explanation or warning message to an explanatory diagnosis report based on the corrected warning grade.

[0030] According to one embodiment, an explanatory electric vehicle battery diagnostic device using ontology rule-based artificial intelligence inference can be provided, comprising a communication module that communicates with the outside, a memory that stores one or more instructions, and a processor that executes one or more instructions stored in the memory, wherein the processor performs the method of the present invention by executing one or more instructions.

[0031] According to one embodiment, a computer program stored on a recording medium readable by a computing device, which is combined with a computing device and includes instructions for executing the method of the present invention, may be provided. Effects of the invention

[0032] The present invention improves the interpretability and reliability of AI-based diagnostic results by providing the causes of abnormalities and risk levels regarding electric vehicle battery diagnostic results in a form that can be explained according to causal relationships.

[0033] The diagnostic structure according to the present invention analyzes battery status data and artificial intelligence diagnostic results using an ontology rule-based inference structure and automatically generates the analysis results into a natural language-based descriptive report, so that even users without specialized knowledge can intuitively understand the battery abnormality status and its cause.

[0034] Accordingly, the present invention enables the direct use of diagnostic results as a basis for decision-making in various industrial fields, such as maintenance, insurance, used car trading, and vehicle management services.

[0035] In addition, the present invention can induce preventive measures, such as pre-maintenance, improvement of usage conditions, or recall decisions, by early detection of multiple abnormal conditions, such as fire risk, rapid deterioration, cell imbalance, and impedance rise, during the operation phase, and by presenting the causal path connecting the said abnormal condition to a specific cause of deterioration.

[0036] Furthermore, the ontology rule-based inference structure significantly improves the scalability and maintainability of the diagnostic system, as it allows the diagnostic system to be easily updated simply by expanding the knowledge structure and rules, even when new degradation mechanisms, diagnostic indicators, or operating conditions are added.

[0037] Furthermore, the present invention provides explainable analysis results to industry entities such as insurers, manufacturers, rental car operators, and transportation companies through a causal inference structure linked to SOH, RUL, stress factors, environmental conditions, etc., thereby providing the effect of systematically improving the reliability of battery warranty policy establishment, risk management, operation policy determination, and service operations.

[0038] The effects of the present invention are not limited to those mentioned above, and other unmentioned effects will be clearly understood by a person skilled in the art to which the present invention pertains from the description below. Brief explanation of the drawing

[0039] FIG. 1 is a diagram illustrating the technical features of providing an explanatory electric vehicle battery diagnostic service based on ontology rules according to one embodiment. FIG. 2 is a diagram illustrating the configuration of an explanatory electric vehicle battery diagnostic device using ontology rule-based artificial intelligence inference according to one embodiment. FIG. 3 is a flowchart illustrating an explanatory electric vehicle battery diagnostic method using ontology rule-based artificial intelligence inference according to one embodiment. FIG. 4 is a diagram illustrating the characteristics of an ontology structure and an application layer referencing the ontology according to one embodiment. FIG. 5 is a diagram illustrating the diagnosis of an electric vehicle battery state as a lithium plating risk according to a state vector-based ontology rule inference flow according to one embodiment. Specific details for implementing the invention

[0040] In the present invention, the attached drawings may be illustrated with exaggerated expressions to distinguish it from the prior art, ensure clarity, and facilitate the understanding of the technology. Furthermore, the terms described below are defined considering their functions in the present invention; since these terms may vary depending on the intentions or conventions of the user or operator, their definitions should be based on the technical content throughout this specification. Meanwhile, the embodiments are merely exemplary details of the components presented in the claims of the present invention and do not limit the scope of the rights of the present invention; the scope of rights should be interpreted based on the technical concept throughout the specification of the present invention.

[0041] Throughout the specification, when a configuration is described as "including" a configuration, this means that, unless specifically stated otherwise, it does not exclude other configurations but may include additional configurations.

[0042] Furthermore, when it is said that one configuration is "connected," "connected," or "combined" with another configuration, this means that it is not only "directly connected," "directly connected," or "directly combined," but also that there may be cases where it is "connected with another configuration interposed," "connected with another configuration interposed," or "combined with another configuration interposed." On the other hand, when it is said that one configuration is "directly connected," "directly connected," or "directly combined" with another configuration, it should be understood that there is no other configuration in between.

[0043] In addition, when directional terms such as "front," "back," "up," "down," "left," "right," "first end," "other end," and "both ends" are used, they are used exemplarily in relation to the orientation of the disclosed drawings and should not be interpreted restrictively, and when terms such as "first" and "second" are used, they are terms used to distinguish each configuration and should not be interpreted restrictively.

[0044] Additionally, the terms 'module' or 'part' as used in the specification refer to software or hardware components, and the 'module' or 'part' performs certain roles. However, the meaning of 'module' or 'part' is not limited to software or hardware. The 'module' or 'part' may be configured to reside in an addressable storage medium or configured to run on one or more processors. Thus, as an example, the 'module' or 'part' may include components such as software components, object-oriented software components, class components, and task components, and at least one of processes, functions, attributes, procedures, subroutines, segments of program code, drivers, firmware, microcode, circuits, data, databases, data structures, tables, arrays, or variables. The components and the functions provided within the 'module' or 'part' may be combined into a smaller number of components and 'modules' or 'parts', or further separated into additional components and 'modules' or 'parts'.

[0045] In order to more clearly explain the features of the embodiments of the present invention, detailed descriptions of matters widely known to those skilled in the art to which the following embodiments pertain are omitted. Additionally, detailed descriptions of parts in the drawings that are unrelated to the description of the embodiments are omitted.

[0046] Hereinafter, embodiments of the present invention will be described in detail with reference to the attached drawings.

[0047] FIG. 1 is a diagram illustrating the technical features of providing an explanatory electric vehicle battery diagnostic service based on ontology rules according to one embodiment.

[0048] The present invention aims to go beyond evaluating the condition of an electric vehicle battery at the level of simple numerical diagnosis; it structures the causal relationships between the battery's structure, degradation mechanism, stress factors, operating environment, and risk types into an ontology, logically infers the causes of battery abnormalities based on this ontology, and provides the results as user-customized natural language explanations, thereby enabling the prior recognition of battery degradation and accident risks and allowing for preemptive response. Accordingly, the present invention possesses the technical feature of providing an explainable diagnostic system that fundamentally improves the reliability of lifespan management and ensures the safety of electric vehicle batteries.

[0049] According to one embodiment, the electric vehicle battery status data obtained from the electric vehicle (20) is collected in the form of real-time or stored historical data during operation from multiple data sources such as OBD2, BMS, VCU, weather API, and GPS, and is refined through preprocessing processes such as outlier removal, missing value correction, normalization, and filtering. This battery status data may include voltage, current, temperature, impedance, C-rate, DOD, SOC (State of Charge), SOH (State of Health), RUL (Remaining Useful Life), charge / discharge patterns, and anomaly detection results.

[0050] According to one embodiment, the preprocessed data is transmitted to an artificial intelligence model (11), and an AI diagnostic model such as an LSTM / GRU-based AutoEncoder, XGBoost, etc., is used to produce a diagnostic result regarding abnormal signs of the battery, potential fire risk, and remaining lifespan.

[0051] According to one embodiment, the computing device (100) automatically maps the calculated diagnostic result according to the object and attribute structure of the ontology (12), creates an object instance corresponding to a predefined state class within the ontology, and registers it in the ABox of the ontology. For example, if the diagnostic result related to battery temperature, current, discharge rate, or discharge depth satisfies a predefined judgment criterion (threshold or interval criterion), the computing device (100) can activate a state object or event object corresponding to an overheating state (High Temperature), a fast discharge state (Fast Discharge), or a deep charge / discharge state (Deep Cycle) and register it in the ABox along with a time tag. The object registered at this time is utilized as an input condition for a subsequent SWRL-based causal inference rule, providing an inference basis for logically deriving the cause of battery degradation and the type of risk.

[0052] According to one embodiment, subsequently, an ontology inference rule based on a combination of multiple entity instances registered in an ABox is executed through a SWRL (Semantic Web Rule Language) or SPARQL-based rule engine, thereby causally inferring the abnormal state of the battery, the cause of degradation, and the type of risk. For example, a combination of HighTemperature and ImpedanceRise can be configured as a rule that activates risk entities such as ThermalRunawayRisk or cathode lithium plating risk.

[0053] According to one embodiment, the system can be configured to mutually complement the explanatory power and accuracy of the diagnostic results by optionally comparing the consistency between the SOH, RUL, and anomaly detection results calculated by an artificial intelligence model and the ontology inference results. In other words, the key point is that the state vector structure is utilized even when AI is not used.

[0054] According to one embodiment, the inferred result is linked to an explanatory template structure and is generated as a natural language-based explanatory diagnostic report (13) customized for roles such as users, mechanics, and insurers by referencing a first explanatory template layer defined within the ontology and a second explanatory template layer defined in an external system. The generated explanatory diagnostic report can be output in JSON, HTML, or API response formats and is configured to intuitively convey the cause of the battery abnormality in a form such as "There is a risk of lithium plating as high temperature conditions of 60°C or higher have been repeated for the past 3 days."

[0055] Accordingly, the present invention combines ontology-based causal reasoning and AI diagnosis to provide a highly reliable, explanatory battery diagnostic service that goes beyond simple detection of electric vehicle battery abnormalities to explain them based on causes and enables customized responses for the user.

[0056] FIG. 2 is a diagram illustrating the configuration of an explanatory electric vehicle battery diagnostic device using ontology rule-based artificial intelligence inference according to one embodiment. Hereinafter, the explanatory electric vehicle battery diagnostic device (100) using ontology rule-based artificial intelligence inference will be described as a computing device (100).

[0057] According to one embodiment, a computing device (100) includes a communication module (130) that communicates with the outside, a memory (110) that stores one or more instructions, and a processor (120) that executes one or more instructions stored in the memory (110), and the processor (120) can perform the method of the present invention by executing one or more instructions.

[0058] According to one embodiment, a computer program (155) stored in a recording medium readable by the computing device (100) and including instructions for executing the method of the present invention may be provided, which is combined with the computing device (100).

[0059] Here, only the components related to the embodiments of the present invention are illustrated in FIG. 2. Therefore, a person skilled in the art to which the present invention pertains will understand that other general-purpose components may be included in addition to the components illustrated in FIG. 2.

[0060] According to one embodiment, the processor (120) controls the overall operation of each component of the computing device (100). The processor (120) may be configured to include a CPU (Central Processing Unit), an MPU (Micro Processor Unit), an MCU (Micro Controller Unit), a GPU (Graphic Processing Unit), or any type of processor well known in the art of the present invention.

[0061] According to one embodiment, the processor (120) can perform operations for at least one application or program for executing a method according to embodiments of the present invention, and the computing device (100) may have one or more processors.

[0062] According to one embodiment, the processor (120) may further include RAM (Random Access Memory, not shown) and ROM (Read-Only Memory, not shown) for temporarily and / or permanently storing signals (or data) processed within the processor (120). Additionally, the processor (120) may be implemented in the form of a system-on-chip (SoC) comprising at least one of a graphics processing unit, RAM, and ROM.

[0063] According to one embodiment, the memory (110) stores various data, instructions and / or information. The memory (110) may load a computer program (155) from a storage (150) to execute a method / operation according to various embodiments of the present invention. When the computer program (155) is loaded into the memory (110), the processor (120) may perform the method / operation by executing one or more instructions constituting the computer program (155). The memory (110) may be implemented as a volatile memory such as RAM, but the technical scope of the present disclosure is not limited thereto.

[0064] According to one embodiment, the input / output interface (140) is connected to a BMS, VCU, OBD, power converter, or external sensor module mounted on a vehicle via wired (CAN, LIN, Ethernet) or wireless (Bluetooth, LTE, 5G) communication method to receive raw sensor signals such as voltage, current, temperature, impedance, charge / discharge power, accumulated DOD, SOC, and SOH in real-time or batch form at the cell, module, and pack levels. Through this, an input data stream for the current state value calculation step (S10) is formed.

[0065] Additionally, the input / output interface (140) is connected to external information systems such as a weather API server, a map-based road information server, and a driver profile server to receive ambient temperature, humidity, road type, driving pattern, driving habit indicators, etc., and supports combining this with battery status data to form an extended status data set.

[0066] According to one embodiment, the input / output interface (140) transmits the diagnostic result in the form of JSON, HTML, or API response generated in the descriptive diagnostic report generation step (S40) to an external system such as a user terminal, a repair shop system, an insurance company server, or a vehicle control system. Furthermore, control commands or recommendation commands, such as charging restrictions, changes in driving conditions, repair reservation requests, and insurance warning notifications, are transmitted to the external system so that the diagnostic result is reflected in actual vehicle operation and management policies.

[0067] Furthermore, the input / output interface (140) receives a response signal or control result from an external system and reflects it in updating the ABox of the ontology, updating the activation state of an object instance, and accumulating the history of state changes, thereby forming a closed-loop diagnostic system in which the diagnostic structure of the present invention is continuously updated in a time-series manner.

[0068] According to one embodiment, the bus (160) provides a communication function between components of the computing device (100). The bus (160) can be implemented as various types of buses, such as an address bus, a data bus, and a control bus.

[0069] According to one embodiment, the communication module (130) supports wired and wireless internet communication of the computing device (100). Additionally, the communication module (130) may support various communication methods other than internet communication. To this end, the communication module (130) may be configured to include a communication module well known in the technical field of the present invention. In some embodiments, the communication module (130) may be omitted.

[0070] According to one embodiment, the storage (150) may store a computer program (155) non-temporarily. According to one embodiment, the storage (150) may be configured to include non-volatile memory such as ROM (Read Only Memory), EPROM (Erasable Programmable ROM), EEPROM (Electrically Erasable Programmable ROM), flash memory, a hard disk, a removable disk, or any form of computer-readable recording medium well known in the art to which the present invention belongs. The computer program (155) may include one or more instructions that cause the processor (120) to perform a method / operation according to various embodiments of the present invention when loaded into memory (110). That is, the processor (120) may perform the method / operation according to various embodiments of the present invention by executing the one or more instructions.

[0071] FIG. 3 is a flowchart illustrating an explanatory electric vehicle battery diagnostic method using ontology rule-based artificial intelligence inference according to one embodiment.

[0072] A descriptive electric vehicle battery diagnosis method using ontology rule-based artificial intelligence inference according to one embodiment may include the steps of: calculating a current state value from state data of an electric vehicle battery using an artificial intelligence model (S10); mapping the current state value to an object and attribute of an ontology to create an object instance corresponding to a predefined state class within the ontology and registering the object instance in an ABox of the ontology (S20); executing an ontology inference rule based on a combination of a plurality of object instances registered in an ABox of the ontology to infer whether the electric vehicle battery is abnormal (S30); and generating a descriptive diagnosis report based on the inferred abnormality of the electric vehicle battery (S40).

[0073] According to one embodiment, the state data of an electric vehicle battery may include at least one of data acquired in real time during the operation of the electric vehicle and battery state data stored as operation history.

[0074] According to one embodiment, the state data of an electric vehicle battery may be acquired in real time during the operation of the electric vehicle or regenerated from a stored driving history after the operation ends. More specifically, the state data may be acquired in real time by receiving sensor signals periodically output from a battery management system (BMS), a vehicle control unit (VCU), an on-board diagnostic device (OBD), and a power converter installed in the electric vehicle. In this case, the sensor signals may include voltage, current, temperature, impedance, charge / discharge power, charge / discharge cycles, average C-rate, cumulative DOD, State of Charge (SOC), and State of Health (SOH) at the cell or module level.

[0075] In addition, status data can be obtained by loading from driving history logs collected during operation and stored in the electric vehicle's internal memory or an external server. For example, the driving history logs may include charge / discharge patterns by driving time period, high-temperature driving sections, rapid charging history, deep discharge history, average driving temperature, road type, driving speed profile, and ambient temperature information; these history logs can be sorted along a time axis and used to reconstruct the history of state changes of the electric vehicle battery.

[0076] Furthermore, the status data can be obtained by combining it with information related to ambient temperature, humidity, driving area characteristics, and driving habits provided by an external weather API, a map-based road information system, or a driver profile server, and thereby can be configured into an extended status data set that reflects the operating environment of the electric vehicle battery.

[0077] According to one embodiment, the technical core of the diagnostic structure of the present invention lies in first forming a State Representation Space capable of ontology-based causal inference and explanation generation, rather than generating a simple judgment result. While diagnostic results generally correspond to conclusive and discrete (label-based) information such as "normal / abnormal" or "risk / non-risk," the "current state value" of the present invention corresponds to a continuous and structural State Vector that quantitatively expresses the instantaneous physical, chemical, and operating environmental states of an electric vehicle battery. This State Vector is composed of numerical parameters such as voltage, current, temperature, impedance, C-rate, DOD, SOC, SOH, RUL, charge / discharge pattern, and anomaly detection signal, and can continuously express the phases of battery degradation progression and stress accumulation on the time axis.

[0078] According to one embodiment, the ontology may be constructed based on OWL (Web Ontology Language). Here, the OWL-based ontology is configured to formally define the relationships between structural entities at the cell, module, and pack levels constituting an electric vehicle battery, battery state indicators, degradation mechanisms, stress events, and risk types as classes, data properties, and object properties, and can logically express inclusion relationships, causal relationships, and hierarchical structures between each entity.

[0079] In addition, based on the relationship between object instances registered in ABox, OWL is configured to automatically derive new risk objects or deterioration cause objects according to predefined rules or constraints through its logical reasoning structure.

[0080] The present invention forms an inference structure capable of explaining why an anomaly occurs, through which degradation paths lead to risk, and what risks may arise in the future, rather than merely producing a judgment result of "anomaly" by mapping the "current state value" configured in this way to entities and attributes of an ontology and inputting it into a multi-layered causal structure leading to state classes, degradation mechanisms, stress events, and risk types.

[0081] According to one embodiment, the step of creating an object instance in S20 creates an object instance of a corresponding state class only when the diagnostic result exceeds a preset threshold condition, wherein the object instance may be composed of an object representing at least one current state among an overheated state, an impedance rise state, a cell voltage imbalance state, a rapid charging stress state, and a deep discharge state of an electric vehicle battery.

[0082] According to one embodiment, an object instance may be configured to be activated when the current state value exceeds a preset threshold condition, and deactivated or updated when the current state value returns to below the threshold condition, thereby reflecting the history of state changes of an electric vehicle battery in a time-series manner.

[0083] According to one embodiment, ABox is a data area that stores actual state information for individual objects in an ontology, and refers to a structure that stores object instances (Individuals) and their attribute values ​​corresponding to predefined concepts (Classes) and attributes (Properties).

[0084] More specifically, ABox functions as an area that registers and manages the actual state of the electric vehicle battery at a specific point in time on an object instance basis for the state class, degradation cause class, and event class defined by TBox, and the registered object instances are used as direct input data for ontology inference rules.

[0085] According to one embodiment, registering an object instance in the ABox of an ontology means that the current state value calculated by an artificial intelligence model or sensor data is not stored as simple numerical data, but is structured and stored as object-unit state information regarding the actual occurring state corresponding to a state class predefined within the ontology.

[0086] More specifically, if the current state value exceeds a critical condition, an object instance corresponding to that state is created and registered in the ABox; this means that the actual state of the electric vehicle battery at a specific point in time is represented as object-unit state information and used as an input target for ontology inference rules.

[0087] According to one embodiment, an object instance registered in ABox functions as structured state information that can be directly referenced as a target for combination, comparison, and causal relationship analysis by ontology inference rules, unlike a simple numeric log, thereby forming a state representation-based inference structure for causally inferring the cause of battery abnormalities, degradation paths, and risk types.

[0088] Additionally, object instances can be configured to be disabled or updated when their current state value returns below a critical condition, thereby enabling the management of the electric vehicle battery's state change history as object-level time-series state information.

[0089] According to one embodiment, the step of registering an object instance in an ABox of an ontology may include the step of mapping an item of a current state value to at least one of a predefined data attribute or object attribute within the ontology, and the step of registering attribute information regarding an electric vehicle battery in the ontology by setting a diagnostic result to the corresponding attribute.

[0090] According to one embodiment, the step of mapping items of current state values ​​to data attributes or object attributes is a step of mapping each item of current state values ​​calculated from an electric vehicle battery to match an attribute structure predefined within an ontology. For example, numerical current state values ​​such as voltage, current, temperature, impedance, SOC, SOH, RUL, C-rate, DOD, and charge / discharge patterns are mapped to and connected to attributes such as "hasVoltage", "hasTemperature", "hasImpedance", "hasSOC", "hasSOH", and "hasRUL", which are defined as data attributes within the ontology. Furthermore, if the current state values ​​express a relationship with a specific degradation mechanism, stress event, or risk type, they are mapped to object attributes such as "inducesDegradation", "triggersStressEvent", and "associatedWithRisk", and structured as relationships between entity instances. Through this, numerical data is formalized not as a simple set of values, but as state information having semantic relationships within the ontology structure.

[0091] According to one embodiment, the step of registering an object instance of an electric vehicle battery in an ontology by setting a diagnostic result in a corresponding attribute is a step of registering an object instance of an electric vehicle battery in an ABox by setting a current state value or a diagnostic result calculated by an artificial intelligence model as an actual value in a data attribute or object attribute determined in the corresponding step. For example, for an object instance corresponding to a specific cell or module, "hasTemperature = 62 Data attribute values ​​such as "hasImpedance = 1.35 mΩ" and "hasSOC = 48%" are set, and at the same time, object attribute relationships such as "inducesDegradation → ImpedanceRise" and "triggersStressEvent → FastDischarge" are created and stored. Accordingly, the state of the electric vehicle battery at different time points, signs of degradation, stress events, and risk associations are accumulated in ABox as specific factual information, which is then used as base data referenced by ontology inference rules.

[0092] According to one embodiment, in step S30, the computing device identifies a set of object instances that satisfy a predefined combination condition among object instances registered in ABox. At this time, the combination condition is defined to determine whether at least two of an overheated state object, an impedance rise state object, a cell voltage imbalance state object, a rapid charge stress state object, and a deep discharge state object are simultaneously active.

[0093] Next, an ontology inference rule corresponding to the identified combination of object instances is selected. The ontology inference rule is described in a rule language based on SWRL or SPARQL, and each rule is predefined to generate a specific risk type object or anomaly cause object based on a logical combination relationship between multiple state objects, deterioration cause objects, and event objects.

[0094] Next, the selected ontology inference rule is executed, and as a result of the execution, a new risk object instance corresponding to the combination of object instances is created in ABox. The risk object instance created at this time may consist of an object corresponding to at least one risk type among fire risk, rapid life reduction risk, or short circuit risk.

[0095] According to one embodiment, the ontology inference rule may be composed of a predefined set of rules to causally derive an abnormal state, a cause of degradation, or a risk type of an electric vehicle battery, based on a combination of state indicators, degradation-related parameters, stress factors, and environmental variables of the electric vehicle battery.

[0096] According to one embodiment, the ontology inference rule can be executed independently without the diagnostic algorithm of the artificial intelligence model. More specifically, since the ontology inference rule generates a risk entity by utilizing the activation status of an entity instance generated from a current state value, relationship attributes between entities, and a predefined set of causal rules, it is possible to independently determine whether the battery is abnormal and the type of risk based on a combination of the electric vehicle battery's state change, accumulated degradation state, and stress events, even when an abnormality determination result by the artificial intelligence model is not provided. Accordingly, the present invention forms a rule-based diagnostic structure capable of continuous operation even in the learning state of the artificial intelligence model, model replacement, or model failure situations.

[0097] According to one embodiment, an ontology inference rule can derive degradation patterns and risk types between vehicle groups by clustering them by simultaneously referencing entity instances regarding battery status data of multiple electric vehicles. According to one embodiment, battery status entity instances corresponding to each of multiple electric vehicles are registered in parallel in the ABox of the ontology, and the ontology inference rule is configured to simultaneously refer to the entity instances of these multiple vehicles.

[0098] Accordingly, the inference rule can derive deterioration trends that are commonly repeated across the entire vehicle group or risk types that occur intensively under specific operating conditions by clustering and analyzing the simultaneous occurrence frequency, combination patterns, and temporal accumulation patterns of overheating state objects, impedance rise state objects, voltage imbalance state objects, rapid charge stress state objects, and deep discharge state objects generated for each vehicle.

[0099] Through this, the present invention forms a structure that goes beyond diagnosis at the single vehicle level and provides basis data for classifying deterioration patterns, clustering risk types, and establishing preventive maintenance policies at the vehicle group level.

[0100] According to one embodiment, the rule set may include a multi-layered causal rule structure divided into a first rule layer expressing structural relationships between battery components, a second rule layer expressing causal relationships between degradation patterns, and a third rule layer expressing causal relationships between operating environment variables and risk types.

[0101] According to one embodiment, the first rule layer is a rule layer that expresses inclusion, connection, and dependency relationships between structural entities at the cell, module, and pack levels constituting an electric vehicle battery, and defines, based on rules, how a state entity or degradation entity occurring at a specific cell level is transmitted and reflected to higher-level structural entities at the module and pack levels. Accordingly, this layer explicitly expresses the structural propagation path through which an abnormal state at the cell level extends to the module or pack level, thereby enabling the inference of the hierarchical transmission structure of the abnormal occurrence location.

[0102] According to one embodiment, the second rule layer is a rule layer that expresses causal relationships between degradation mechanism entities, such as impedance rise, cell voltage imbalance, capacity reduction, internal resistance increase, lithium plating, and electrolyte degradation, and defines the interrelationships and cumulative effects as rules when multiple degradation entities occur simultaneously or sequentially. Accordingly, the layer is configured to structure transition paths, reinforcement relationships, or offset relationships between degradation mechanisms into causal rules to infer the phases of degradation progression step by step.

[0103] According to one embodiment, the third rule layer is a rule layer that expresses the causal relationship between operating environment variables, such as operating temperature, charge / discharge C-rate, cumulative DOD, ambient temperature, road type, and driving habits, and risk type entities, such as fire risk, risk of rapid life reduction, short-circuit risk, and safety degradation risk, and defines in the form of rules conditions in which a specific combination of operating conditions causes a specific risk type. Accordingly, this layer explicitly structures the causal flow in which environmental stress conditions transition into a risk type through a degradation mechanism.

[0104] According to one embodiment, the rule set is described in a SWRL or SPARQL-based rule language and may be configured to include conditions for the simultaneous existence, occurrence order, cumulative count, and temporal duration of multiple object instances registered in an ABox as condition clauses. Accordingly, the present invention forms a rule-based inference structure that goes beyond single-parameter threshold determination and causally derives abnormal causes and risk types based on combinations of complex conditions.

[0105] According to one embodiment, the step (S30) of inferring whether an electric vehicle battery is abnormal may include: identifying a combination of object instances that satisfies a preset combination condition among a plurality of object instances registered in an ABox; selecting an ontology inference rule corresponding to the identified combination of object instances; executing the selected ontology inference rule to generate a risk object corresponding to an abnormal type of the electric vehicle battery; and determining whether the electric vehicle battery is abnormal based on the risk grade of the generated risk object.

[0106] According to one embodiment, the step of identifying an object instance combination may include determining whether at least two object instances among an overheated state object, an impedance rise state object, a voltage imbalance state object, a rapid charge stress state object, and a deep discharge state object exist simultaneously among object instances registered in ABox, and if the determination result is true, identifying a set containing two or more object instances as an object instance combination.

[0107] For example, if an "overheating state object" and an "impedance rise state object" are registered as being simultaneously activated at the same point in time or within a preset allowable time error range within an ABox corresponding to the same electric vehicle battery, the two or more object instances are identified as a single object instance combination. In this case, the object instance combination is used as an input condition representing a state pattern of "overheating + impedance rise," and a corresponding ontology inference rule is automatically selected.

[0108] Furthermore, when the selected ontology inference rule is executed, a risk entity corresponding to the aforementioned combination pattern, such as a "thermal runaway risk entity," is generated, and a risk grade predefined for the corresponding risk type is associated with and set for the risk entity. Subsequently, if the risk grade set for the risk entity exceeds a predefined judgment criterion, the electric vehicle battery is configured to be determined to be in an abnormal state; thus, causal abnormality determination based on complex state patterns is performed, rather than a simple determination of whether a numerical value is exceeded.

[0109] According to one embodiment, the step (S30) of inferring whether an electric vehicle battery is abnormal may include: obtaining a diagnosis result calculated by an artificial intelligence model and an inference result calculated by an ontology inference rule; comparing the consistency of the abnormal type or risk level between the diagnosis result and the inference result; correcting the warning level regarding whether the electric vehicle battery is abnormal if, as a result of the comparison, the abnormal type is different or the difference in the risk level exceeds a preset standard; and adding a supplementary explanation or warning message to an explanatory diagnosis report based on the corrected warning level.

[0110] According to one embodiment, the step of obtaining diagnosis results and inference results comprises obtaining prediction results and anomaly detection results calculated by an artificial intelligence model based on the current state value of an electric vehicle battery, and risk entities, degradation cause entities, or risk type entities causally derived by ontology inference rules based on combinations of entity instances registered in an ABox, respectively. In this step, data-based probabilistic judgment results and knowledge-based logical inference results are simultaneously obtained, thereby forming a dual judgment basis set for subsequent consistency comparison.

[0111] According to one embodiment, the step of comparing consistency involves comparing an anomaly type or risk level included in the diagnosis result of an artificial intelligence model with an anomaly type or risk level calculated by an ontology inference rule by mutual correspondence. In this step, it is determined whether the two results indicate the same risk type or belong to the same risk level range, and consistency is evaluated based on whether the difference between the risk levels is within a preset tolerance range.

[0112] According to one embodiment, the step of correcting the warning level is a step of recalculating the final warning level of an electric vehicle battery by considering at least one of reliability weighting, data recency, history accumulation, or priority by risk type, without relying on any one result, when, as a result of the comparison, the abnormality type between the two results is different or the difference in risk level exceeds a preset standard. In this step, if the result of the artificial intelligence model is underestimated or overestimated due to temporary noise, the warning level may be adjusted upward or downward by reflecting the ontology inference result.

[0113] According to one embodiment, the step of adding supplementary explanations or warning messages is a step of automatically inserting content regarding the reason for correction, the cause of the discrepancy, additional precautions, or recommended actions into the natural language explanation to be included in the descriptive diagnostic report based on the corrected warning level. In this step, the method is configured to generate supplementary explanations or enhanced warning messages in the form of, "Although the AI ​​model determined it to be normal, the risk level was upgraded as an ontology inference result confirmed a cumulative pattern of high-temperature repetition and impedance rise."

[0114] According to one embodiment, the step (S40) of generating a descriptive diagnostic report may include: selecting a basic description template from a first description template layer defined within an ontology to correspond to the type of a risk entity; selecting an output template from a second description template layer defined in an external system to correspond to the user type; and converting the description structure of the basic description template to match the format of the output template to generate a descriptive diagnostic report including the inference basis of the risk entity.

[0115] According to one embodiment, the first description template layer refers to a set of templates that are predefined within the ontology and provide a basic description structure corresponding to the type of a risk entity and its sub-risk classification. The first description template layer is configured to receive inference rules associated with the risk entity, combinations of activated entity instances, and causal relationship links as input, and to define a basic description framework such as "cause-reason-conclusion," key description items to be included, and required sentence slots for each risk type. That is, the first description template layer refers to a standard template layer that defines a common description frame that is maintained regardless of the user type for the same risk entity type, thereby ensuring that the reasoning basis is not omitted.

[0116] According to one embodiment, the second description template layer refers to a set of output templates that are predefined in a system outside the ontology and transform or supplement the basic descriptive structure of the first description template layer according to user type, provision channel, output format, and level of presentation. The second description template layer is configured to perform term substitution, sentence length adjustment, detail adjustment of evidence presentation, intensity adjustment of warning messages, and field structuring suitable for JSON / HTML / API response formats, tailored to the level of understanding of each user (driver, mechanic, insurance company, etc.). That is, the second description template layer refers to a presentation and output layer that determines "how to display" for each user and channel while maintaining the content frame of the first description template layer, which determines "what to explain."

[0117] Accordingly, the present invention can provide customized diagnostic information by applying different output templates according to the user type and the roles of stakeholders, such as users, mechanics, and insurance companies.

[0118] According to one embodiment, the descriptive diagnostic report can be converted into at least one control command or recommendation command among a charging limit, a change in driving conditions, a maintenance reservation request, or an insurance warning notification, and transmitted to an external system.

[0119] According to one embodiment, the descriptive diagnostic report is not limited to simple text guidance but can be converted into and transmitted as a control command or recommendation command that an external system can directly execute based on the risk type, risk level, and reasoning grounds included in the report. This conversion is configured to be performed by referring to a predefined control mapping rule according to the type of risk entity and risk level within the descriptive diagnostic report.

[0120] For example, if a risk entity corresponding to "lithium plating risk" is activated as a result of ontology inference and the risk level is determined to be above a preset threshold, the descriptive diagnostic report is converted into control commands or control recommendation information corresponding to "fast charging restriction" and transmitted to the vehicle control system or charging infrastructure management system. Accordingly, the vehicle is controlled to disable the fast charging mode or to limit the maximum charging power in stages.

[0121] As another example, if the "thermal runaway risk" is activated and the corresponding risk level is determined to be critical, the explanatory diagnostic report is converted into a recommendation command corresponding to a "change in driving conditions" and can be transmitted to the vehicle's instrument panel display or a mobile application. In this case, the system is configured to provide the user with messages recommending driving restrictions in high-temperature zones, high-speed driving restrictions, or immediate maintenance.

[0122] In another embodiment, if "impedance rise and cell voltage imbalance" are simultaneously activated and a risk of early degradation is determined, the explanatory diagnostic report is converted into a recommendation command corresponding to a "maintenance reservation request" and can be transmitted to a vehicle manufacturer server or a designated repair shop reservation system. Accordingly, the user terminal is configured to provide a maintenance reservation screen or suggest available time slots at the nearest repair shop.

[0123] In another embodiment, when a risk entity corresponding to "fire risk" is activated, the explanatory diagnostic report may be converted into risk information or recommendation information corresponding to "insurance warning notification" and transmitted to the insurer server, and accordingly, the insurer is configured to provide preemptive risk management services, including requesting a pre-inspection of the risk vehicle, providing accident prevention notifications, or reviewing insurance conditions.

[0124] Accordingly, the present invention provides an active preventive diagnostic system that converts a descriptive diagnostic report into an executable control command or recommendation command capable of being linked with an external system, thereby going beyond the level of mere recognition of an abnormal condition of an electric vehicle battery and directly linking it to actual operation control, maintenance management, and insurance risk management.

[0125] FIG. 4 is a diagram illustrating the characteristics of an ontology structure and an application layer referencing the ontology according to one embodiment.

[0126] According to one embodiment, the ontology may be defined as a Web Ontology Language (OWL) structure including a state indicator entity layer representing the current state of an electric vehicle battery, a degradation cause entity layer configured in parallel with the state indicator entity layer to represent the internal degradation mechanism of the battery, and an event entity layer configured in parallel with the state indicator entity layer and the degradation cause entity layer to represent stress events according to the operating conditions, charging and discharging conditions, and environmental conditions of the electric vehicle.

[0127] Figure 4 is a diagram showing that the ontology of the present invention is not a simple data dictionary or rule table, but a parallel three-layer semantic structure in which the state, degradation, and environmental stress of an electric vehicle battery are expressed as a single causal model. This structure arranges states, causes, and events in a "non-sequential coordinate system," thereby enabling all diagnoses to be inferred within spatial semantic coordinates.

[0128] According to one embodiment, the State Indicator Layer is a primary state coordinate system that represents the current physical, electrical, and thermal state of an electric vehicle battery. Here, overheating, rapid discharge, capacity reduction, voltage drop, SOH degradation, etc., are not merely warning flags, but coordinate values ​​that represent the current state phase of the battery. In other words, this layer defines "what phase this battery is currently in."

[0129] According to one embodiment, the degradation cause layer is a model of internal physical and chemical degradation causes arranged in parallel with the state layer. Here, lithium plating, impedance rise, cell imbalance, SEI layer growth, electrode degradation, etc., represent degradation paths rather than measured values. In other words, this layer defines "why this state was created."

[0130] According to one embodiment, the Stress / Event Layer is a coordinate system of external stress and environmental events accumulated during the operation of an electric vehicle. Temperature, charging / discharging speed, road type, charging / discharging pattern, driving habits, etc., are defined not as simple logs but as stress coordinates that trigger the causes of degradation. In other words, this layer represents "which usage environment caused the cause."

[0131] According to one embodiment, the inference result hierarchy is a risk result entity such as fire risk, rapid lifespan reduction, and short-circuit risk derived as a result of the causal combination of three layers. This domain does not refer to a simple result, but rather to a risk phase reached by the combined result of the state, cause, and event three-coordinate system.

[0132] According to one embodiment, the first description template layer and the second description template layer are application layers that interpret an ontology structure, and the first description template layer is used as a semantic interpreter that converts the logical structure of a risk entity into a natural language descriptive structure. The second description template layer is an output formatter that converts into a sentence structure according to roles such as user, mechanic, and insurer.

[0133] According to one embodiment, the user-specific descriptive diagnostic report output layer is an interface of the language result of the ontology structure and is an application area outside the ontology.

[0134] FIG. 5 is a diagram illustrating the diagnosis of an electric vehicle battery state as a lithium plating risk according to a state vector-based ontology rule inference flow according to one embodiment.

[0135] According to one embodiment, Table 1 below shows an example of electric vehicle battery anomaly cause inference using state vector-based ontology causation rules.

[0136] Antecedent Consequent Explanation of the cause of deterioration 1 In cases where a high-load charging state is repeated at low temperatures or under adverse environmental conditions Electrochemical stress-based risk indicators Increase in internal stress due to charging conditions 2 When stress events accumulate in low-temperature conditions or environments with large temperature fluctuations Interfacial degradation risk indicators Possibility of reduced interfacial stability 3 When the voltage or temperature deviation between cells expands beyond the standard Cell imbalance-based risk indicator Risk of inter-cell behavioral imbalance 4 In the case where rapid charging or high load events are repeated Risk indicators of accelerated deterioration trends Potential for accelerated degradation 5 When a downward trend in performance-related indicators is observed even under moderate usage conditions Risk indicators of the tendency for dose reduction Signs of long-term performance degradation

[0137] More specifically, Table 1 shows combinations of state entities, event entities, and degradation cause entities generated from multiple state parameters of an electric vehicle battery. As an input condition , A computing device (100) demonstrates an example of a rule set that causally infers the battery internal degradation mechanism and risk type by executing SWRL (Semantic Web Rule Language) based ontology rules. Here, the computing device (100) is configured to determine whether object instances corresponding to the condition combination (Antecedent) of each rule are simultaneously activated, and if the result of the determination is true, to create a corresponding risk type (Consequent) as a risk object and register it in the ABox.

[0138] <Rule 1 - Rule for Inferring Electrochemical Stress Risk Based on High-Load Charging>

[0139] According to one embodiment, the computing device (100) is configured to interpret the combination of conditions as a precursor state of electrochemical instability and to create a corresponding electrochemical stress-based risk object and register it in the ABox when, according to rule 1, the electric vehicle battery is in a charged state under adverse temperature conditions and at the same time a stress event object associated with high-rate charging is activated.

[0140] <Rule 2 - Interfacial Degradation Risk Inference Rule Based on Low Temperature and Abnormal Voltage Behavior>

[0141] According to one embodiment, the computing device (100) is configured to interpret the combination of conditions as a state in which electrode interface stability is reduced, and to create a risk object related to interface degradation corresponding to this, and register it in the ABox, when a low charge state persists for a certain period of time or longer according to rule 2 and a low temperature environment or abnormal cell voltage behavior is simultaneously observed and a related object is activated.

[0142] <Rule 3 - Rule for Inferring Risk of Inter-Cell Behavior Imbalance>

[0143] According to one embodiment, the computing device (100) is configured to interpret the condition as a state in which the behavioral imbalance between cells in the battery pack is intensified when the voltage or temperature deviation indicator between the cells expands beyond a predefined standard according to rule 3 and an imbalance-related object is activated, and to create a cell imbalance-based risk object corresponding thereto and register it in the ABox.

[0144] <Rule 4 - Rule for Inferring Risk of Accelerated Deterioration Based on Repetitive High-Load Events>

[0145] According to one embodiment, the computing device (100) is configured to interpret the combination of conditions as a state in which the rate of battery degradation is likely to accelerate, and to create a risk object related to accelerated degradation corresponding to the same, and register it in the ABox, when a rapid charging or high load event is repeatedly observed according to rule 4 and a storage or operating state under adverse environmental conditions is sustained and a related object is activated.

[0146] <Rule 5 - Rule for Inferring Performance Degradation Trends Based on Long-term Usage Patterns>

[0147] According to one embodiment, the computing device (100) is configured to interpret the condition as a state in which a battery performance degradation trend is in progress when a degradation trend of performance-related indicators accumulates and a related object is activated in a situation where a long-term usage pattern under moderate load conditions continues according to rule 5, and to create a performance degradation trend risk object corresponding to this and register it in the ABox.

[0148] Referring to FIG. 5, an embodiment is described in which the entire process of generating and delivering an explanatory diagnostic report when the user type is 'mechanic' is inferred in a real vehicle scenario to which Rule 1 is applied, and the rule is applied in the form of a time-series pipeline.

[0149] According to one embodiment, during the process of an electric vehicle performing rapid charging after driving, raw data including battery temperature, C-rate, internal impedance, SOC, DOD, ambient temperature, and charge / discharge history data is collected in real time from a battery management system (BMS), a vehicle controller (VCU), and an external environment sensor. A computing device (100) can generate a diagnostic input stream after performing missing value correction, normalization, and outlier removal on the raw data.

[0150] According to one embodiment, a computing device (100) can generate a multidimensional state vector based on a current state value using preprocessed data as input. Here, the state vector includes temperature, C-rate, impedance rise index, cumulative DOD, recent rapid charging frequency, and environmental stress index, and can form a structured state representation space that continuously represents the instantaneous degradation phase and the stress accumulation phase of the battery. According to one embodiment, the computing device (100) maps the state vector to data attributes and object attributes of an ontology, and activates corresponding entities only for items exceeding preset threshold conditions and registers them in an ABox. Specifically, i) if the cell average temperature > 60℃, a <High Temperature State Entity> is created and updated; ii) if the C-rate ≥ 2.5, a <Rapid Charging Event Entity> is created and updated; and iii) if the internal impedance rise is above a reference value, a <Impedance Rise Degradation Cause Entity> is created and updated, respectively. At this time, the entities are registered along with time tags so that the history of state changes can be managed in the form of a timeline.

[0151] According to one embodiment, a computing device (100) scans a set of object instances registered in an ABox to identify object combinations that satisfy predefined combination conditions and selects a corresponding SWRL-based causal rule. Specifically, the computing device (100) refers to a predefined set of rules to match a rule that logically matches the currently active combination of object instances, and executes the matched rule through an ontology inference engine to derive a new risk object or degradation cause object based on the relationship between objects registered in the ABox. For example, if a high temperature state object, a rapid charging event object, and an impedance rise degradation cause object are simultaneously activated at the same time or within an allowed time window, the computing device (100) determines that all preconditions of Rule 1 are satisfied.

[0152] According to one embodiment, the computing device (100) executes Rule 1 to interpret the condition as having a high probability of lithium deposition on the cathode surface, and creates a new risk entity corresponding to the 'lithium plating risk' and registers it in the ABox. A predefined risk level (e.g., Warning / Severe / Critical) is associated with the risk entity.

[0153] According to one embodiment, the computing device (100) executes a consistency verification block between the AI ​​diagnosis result and the ontology inference result. It compares the SOH, RUL, or anomaly probability calculated by the AI ​​model with the ontology-based risk grade, and if the discrepancy is greater than a threshold, it corrects the final warning grade and stores the reason for the correction as meta-information.

[0154] According to one embodiment, the computing device (100) selects a basic explanation structure (cause-reason-conclusion frame) in a first explanation template layer within the ontology corresponding to the generated risk object type. At the same time, as the user type is identified as 'mechanic', a mechanic output format (JSON / HTML / API) is selected in a second explanation template layer of an external system.

[0155] According to one embodiment, a computing device (100) can convert the causal structure of a first description template into a second output template format to automatically generate a maintenance-explanatory diagnostic report including the following: i) a causal path in which a combination of high temperature, high-rate charging, and impedance rise causes a lithium plating risk,

[0156] ii) Explanation of the degradation mechanism in which lithium plating can lead to increased cathode surface resistance and a sharp decrease in cell lifespan,

[0157] iii) the actual status value of the vehicle, the object activation timeline, and whether the rule is satisfied, and

[0158] iv) List of cell / module locations requiring inspection, cooling system inspection items, and priority actions

[0159] According to one embodiment, the computing device (100) transmits the report to a maintenance terminal and, at the same time, links the risk type and grade within the report to a control mapping rule to generate at least one recommendation / control command among a rapid charging restriction, a cooling system inspection request, and a cell module precision diagnosis reservation, and transmits it to a vehicle system or a maintenance server. Through this, a maintenance technician can perform real-time evidence-based measures. As described above, the present invention has been described with reference to the embodiments illustrated in the drawings, but this is merely illustrative, and it should be understood that various modifications and equivalent alternative embodiments are possible based on the ordinary knowledge of the art to which the technology belongs. Accordingly, the true technical scope of protection of the present invention is determined by the claims described below and should be determined based on the specific details of the invention described above. Industrial applicability

[0160] The present invention relates to an explanatory electric vehicle battery diagnostic method and apparatus using ontology rule-based artificial intelligence inference, and can be applied to the field of electric vehicle battery condition diagnosis. Explanation of the symbols

[0161] 100: Explanatory electric vehicle battery diagnostic device (computing device) using ontology rule-based artificial intelligence reasoning 110: Memory 120: Processor 130: Communication module 140: Input / Output Interface 150: Storage 155: Computer program

Claims

Claim 1 A method executed by a computing device comprises: a step of calculating a current state value from state data of an electric vehicle battery using an artificial intelligence model; a step of mapping the current state value to an entity and an attribute of an ontology to create an entity instance corresponding to a state class predefined within the ontology and registering the entity instance in an ABox of the ontology; a step of executing an ontology inference rule based on a combination of a plurality of entity instances registered in the ABox of the ontology to infer whether the electric vehicle battery is abnormal; and a step of generating a descriptive diagnostic report based on the inferred abnormality of the electric vehicle battery; wherein the step of inferring whether the electric vehicle battery is abnormal includes a diagnostic result calculated by the artificial intelligence model and an inference result calculated by the ontology inference rule, and the step of registering the entity instance in the ABox of the ontology includes a step of mapping an item of the current state value to at least one of a data attribute or an object attribute predefined within the ontology. A descriptive electric vehicle battery diagnosis method using ontology rule-based artificial intelligence inference, comprising the step of registering attribute information regarding the electric vehicle battery in the ontology by setting the diagnosis result in the corresponding attribute. Claim 2 The explanatory electric vehicle battery diagnosis method using ontology rule-based artificial intelligence inference according to claim 1, wherein the ontology is defined in an OWL (Web Ontology Language) structure comprising: a state indicator entity layer representing the current state of the electric vehicle battery; a degradation cause entity layer configured in parallel with the state indicator entity layer representing the battery internal degradation mechanism; and an event entity layer configured in parallel with the state indicator entity layer and the degradation cause entity layer representing stress events according to the operating conditions, charging and discharging conditions, and environmental conditions of the electric vehicle. Claim 3 The method for diagnosing an explanatory electric vehicle battery using ontology rule-based artificial intelligence inference according to claim 1, wherein the step of creating an object instance and registering the object instance in the ABox of the ontology is characterized by creating an object instance of a corresponding state class only when the diagnosis result exceeds a preset threshold condition, wherein the object instance is composed of an object representing at least one current state among an overheating state, an impedance rise state, a cell voltage imbalance state, a rapid charging stress state, and a deep discharge state of the electric vehicle battery. Claim 4 In claim 1, the ontology inference rule is composed of a predefined set of rules to causally derive an abnormal state, cause of degradation, or risk type of the electric vehicle battery based on a combination of state indicators, degradation-related parameters, stress factors, and environmental variables of the electric vehicle battery; the set of rules includes a multi-layered causal rule structure divided into a first rule layer expressing structural relationships between battery components, a second rule layer expressing causal relationships between degradation patterns, and a third rule layer expressing causal relationships between driving environment variables and risk types; and the set of rules is described in a rule language based on SWRL (Semantic Web Rule Language) or SPARQL and is configured to causally infer the cause of abnormality or risk type of the electric vehicle battery based on a combination of conditions between multiple entity instances registered in the ABox of the ontology. Claim 5 delete Claim 6 The method for diagnosing an electric vehicle battery using ontology rule-based artificial intelligence inference according to claim 1, wherein the step of inferring whether the electric vehicle battery is abnormal comprises: identifying a combination of object instances that satisfies a preset combination condition among a plurality of object instances registered in the ABox; selecting an ontology inference rule corresponding to the identified combination of object instances; executing the selected ontology inference rule to generate a risk object corresponding to the abnormal type of the electric vehicle battery; and determining whether the electric vehicle battery is abnormal based on the risk grade of the generated risk object. Claim 7 In claim 6, the step of identifying the object instance combination comprises: determining whether at least two object instances among an overheated state object, an impedance rise state object, a voltage imbalance state object, a rapid charge stress state object, and a deep discharge state object exist simultaneously among the object instances registered in the ABox; and if the result of the determination is true, identifying a set including the two or more object instances as the object instance combination; characterized in that it is an explanatory electric vehicle battery diagnosis method using ontology rule-based artificial intelligence inference. Claim 8 In claim 6, the step of generating the explanatory diagnostic report comprises: selecting a basic explanatory template from a first explanatory template layer defined within the ontology to correspond to the type of the risk entity; selecting an output template from a second explanatory template layer defined in an external system to correspond to the type of the electric vehicle user; and converting the explanatory structure of the basic explanatory template to match the format of the output template to generate an explanatory diagnostic report including the reasoning basis of the risk entity; characterized in that it is an explanatory electric vehicle battery diagnostic method using ontology rule-based artificial intelligence inference. Claim 9 The method for diagnosing an explanatory electric vehicle battery using ontology rule-based artificial intelligence inference according to claim 1, wherein the step of inferring whether the electric vehicle battery is abnormal further comprises: a step of comparing the consistency of abnormal type or risk grade between the diagnosis result and the inference result; a step of correcting the warning grade regarding whether the electric vehicle battery is abnormal if, as a result of the comparison, the abnormal type is different or the difference in the risk grade exceeds a preset standard; and a step of adding a supplementary explanation or warning message to the explanatory diagnosis report based on the corrected warning grade. Claim 10 An explanatory electric vehicle battery diagnostic device using ontology rule-based artificial intelligence inference, comprising: a communication module that communicates with the outside; a memory that stores one or more instructions; and a processor that executes one or more instructions stored in the memory, wherein the processor performs the method of claim 1 by executing one or more instructions.

Images