AI-Based Real-Time Diagnostic Monitoring System for Medium- and High-Voltage Switchgear Using LPIT and Infrared Thermography

Abstract

The present invention relates to a smart digital switchgear diagnostic system for extra-high voltage applied with LPIT, and more specifically, to a smart digital switchgear diagnostic system that collects and analyzes voltage, current, partial discharge, and thermal imaging data in real time to determine whether there is an abnormality in electrical equipment. According to the present invention, by utilizing the LPIT and the infrared thermal imaging measuring device to integrate and analyze voltage, current, partial discharge, and thermal image information of the extra-high voltage switchgear in real time, the invention provides the effect of accurately diagnosing the presence or absence of abnormalities in electrical equipment and responding early, by including a measuring unit composed of a low-power measuring current transformer (LPCT) and a low-power measuring voltage transformer (LPVT) and installed in an extra-high voltage switchgear to measure current and voltage; a signal converter that converts signals collected from the LPCT and LPVT into measurable voltage and current values; a partial discharge converter that receives a current waveform signal provided from the LPCT and analyzes a partial discharge pattern; an infrared thermal imaging measuring device that measures real-time degradation distribution from electrical equipment and outputs thermal image data; and a diagnostic monitoring unit equipped with artificial intelligence software (AI S / W) that receives real-time voltage, current, partial discharge analysis values, and thermal image data provided respectively from the signal converter, the partial discharge converter, and the infrared thermal imaging measuring device, and analyzes and determines the presence or absence of abnormalities in electrical equipment based thereon. there is.

Description

Technology Field

[0001] The present invention relates to a diagnostic system for an extra-high voltage smart digital switchgear with LPIT applied, and more specifically, to a diagnostic system for an extra-high voltage smart digital switchgear with LPIT applied that collects and analyzes voltage, current, partial discharge, and thermal imaging data in real time to determine whether there is an abnormality in electrical equipment. Background Technology

[0002] Conventional extra-high voltage switchgear is equipped with large and expensive Potential Transformers (PTs) and Current Transformers (CTs) to measure voltage and current for instrumentation and protective relays. To measure the waveforms of voltage and current, separate, expensive power quality meters must be installed additionally.

[0003] Furthermore, to measure partial discharge in electrical equipment installed in switchboards, separate partial discharge measurement sensors, transducers, and analysis devices are configured and operated. For thermal imaging measurements as well, sensors, transducers, and analysis devices must be installed separately for each piece of electrical equipment, which results in a complex device configuration and increased costs.

[0004] Furthermore, this approach fails to provide an integrated diagnostic system capable of determining equipment abnormalities by comprehensively analyzing voltage, current, partial discharge, and thermal imaging data measured for each electrical facility. As a result, there are limitations in accurately diagnosing the overall equipment status in real time and responding early.

[0005] Referring to FIG. 1, a conventional system for measuring voltage and current in an extra-high voltage switchgear is shown.

[0006] Voltage measurement is performed via a high-voltage potential transformer (PT) located on the upper left, which converts high voltage into a low voltage suitable for instrument measurement and transmits it to the digital relay. Current measurement is performed via three high-voltage current transformers (CTs) located at the bottom, which measure three-phase currents (IA, IB, IC) and supply them to the digital relay. Based on the received voltage and current data, the digital relay performs measurement and protection functions (e.g., overcurrent, ground fault, etc.). While this system is suitable for basic power measurement and protection due to its simple components, it does not include advanced analysis functions such as partial discharge, thermal imaging, or AI-based diagnostics. Furthermore, since each sensor and protection device operates independently, it is difficult to expand into an integrated analysis system.

[0007] Therefore, conventional technology has problems such as high cost, complex system configuration, lack of integrated analysis, and difficulty in real-time diagnosis. Prior art literature

[0008] Registered Patent No. 1421588 (July 15, 2014) The problem to be solved

[0009] The present invention aims to solve the aforementioned problems by utilizing LPIT and infrared thermal imaging measurement devices to integrate and analyze voltage, current, partial discharge, and thermal imaging information of extra-high voltage switchgear in real time, thereby enabling accurate diagnosis of abnormalities in electrical equipment and early response. means of solving the problem

[0010] The present invention relates to a diagnostic system for a high-voltage smart digital switchgear with LPIT applied, comprising: a measuring unit installed in a high-voltage switchgear to measure current and voltage, composed of a low-power measuring current transformer (LPCT) and a low-power measuring voltage transformer (LPVT); a signal converter that converts signals collected from the LPCT and LPVT into measurable voltage and current values; a partial discharge converter that receives a current waveform signal provided from the LPCT and analyzes a partial discharge pattern; an infrared thermal imaging measuring device that measures a real-time degradation distribution from electrical equipment and outputs thermal image data; and a diagnostic monitoring unit equipped with artificial intelligence software (AI S / W) that receives real-time voltage, current, partial discharge analysis values, and thermal image data provided by the signal converter, partial discharge converter, and infrared thermal imaging measuring device, respectively, and analyzes and determines whether there is an abnormality in the electrical equipment based thereon.

[0011] The infrared thermal imaging measurement device collects individual thermal image information for multiple electrical equipment components, identifies the point exhibiting the highest temperature, and provides the coordinates and temperature information of that point to the diagnostic monitoring unit.

[0012] The signal converter converts the current and voltage signals measured from the LPCT and LPVT into measurable voltage and current values ​​and provides them in real time to the digital protection relay, and the digital protection relay uses the data to simultaneously perform power measurement and protection relay functions.

[0013] The artificial intelligence software installed in the diagnostic monitoring unit determines the presence of abnormalities by comparing voltage, current, partial discharge patterns, and maximum thermal temperature values ​​input from signal converters, partial discharge converters, and infrared thermal imaging measurement devices with reference data, and performs the function of notifying the relevant personnel in real time if an abnormality is detected.

[0014] The artificial intelligence software includes a machine learning-based anomaly detection algorithm that pre-trains or stores reference data regarding the normal operating state of electrical equipment and determines whether there are abnormal patterns by comparing it with data received in real time.

[0015] The diagnostic monitoring unit (500) includes a reference value setting module (510) that sets the average value, standard deviation, and upper / lower limit value of the voltage, current, partial discharge pattern, and thermal image temperature value collected under normal operating conditions of the electrical equipment as reference values, a machine learning or deep learning-based judgment model unit (520) that is trained to classify whether there is an abnormality using a plurality of labeled learning data including voltage, current, partial discharge, and thermal image data in normal and faulty states, and an alarm level classification module (530) that classifies alarm levels into multiple stages according to the severity of the abnormal state and outputs response messages such as warning, inspection recommendation, and immediate action request according to each alarm level.

[0016] The diagnostic monitoring unit (500) further includes a standard comparison module (540) that stores standard data for normal operation of electrical equipment in advance and determines an abnormal state by comparing it with received real-time data.

[0017] The diagnostic monitoring unit (500), low-power measuring current transformer (LPCT), low-power measuring voltage transformer (LPVT), and infrared thermal imaging measuring device (400) are configured with a modular structure that can be additionally installed on an existing high-voltage switchgear, thereby allowing the existing switchgear to be functionally expanded into a smart digital switchgear without any structural changes. Effects of the invention

[0018] According to the present invention, by utilizing LPIT and an infrared thermal imaging measuring device to integrate and analyze voltage, current, partial discharge, and thermal imaging information of an extra-high voltage switchgear in real time, it is possible to accurately diagnose whether there is an abnormality in the electrical equipment and respond early.

[0019] In addition, by replacing the PT and CT of the extra-high voltage switchgear with LPCT and LPVT, the size of the switchgear can be reduced and installation costs can be lowered.

[0020] In addition, partial discharge of electrical equipment can be measured in real time using LPCT.

[0021] In addition, a real-time diagnostic system capable of accurate diagnosis can be established by measuring four factors—partial discharge, thermal imaging, voltage, and current—in real time for electrical equipment installed in extra-high voltage switchgear and analyzing them through AI software.

[0022] Furthermore, by installing LPIT, infrared thermal imaging measuring devices, and AI software-based monitoring systems on existing high-voltage switchgear, existing facilities can be easily converted into smart digital switchgear. Brief explanation of the drawing

[0023] Figure 1 is a configuration diagram of a conventional digital relay system. FIG. 2 is a block diagram of an extra-high voltage smart digital switchgear diagnostic system with LPIT applied according to one embodiment of the present invention. FIG. 3 is a configuration diagram of an extra-high voltage smart digital switchgear diagnostic system with LPIT applied according to one embodiment of the present invention. FIG. 4 shows the detailed configuration of the diagnostic monitoring unit of an extra-high voltage smart digital switchgear diagnostic system with LPIT applied according to one embodiment of the present invention. FIG. 5 is a configuration diagram of an extra-high voltage smart digital switchgear diagnostic system with LPIT applied according to another embodiment of the present invention. Specific details for implementing the invention

[0024] The present invention relates to a smart digital switchgear diagnostic system that applies a Low Power Instrument Transformer (LPIT) and a low-cost infrared thermal imaging device to an extra-high voltage switchgear to measure and analyze voltage, current, partial discharge, and thermal imaging data of electrical equipment in real time, and diagnoses whether there is an abnormality in electrical equipment through AI-based software.

[0025] FIG. 2 is a block diagram of a smart digital switchgear diagnostic system according to the present invention, and FIG. 3 is an example configuration diagram illustrating the configuration flow and connection relationships of a system according to the present invention.

[0026] As illustrated in FIG. 2, the smart digital switchgear diagnostic system for diagnosing abnormal conditions of electrical equipment in real time according to the present invention includes a measuring unit (100), a signal converter (200), a partial discharge converter (300), an infrared thermal imaging measuring device (400), and a diagnostic monitoring unit (500).

[0027] The measuring unit (100) is installed in an extra-high voltage switchgear and includes a low-power measuring current transformer (LPCT) and a low-power measuring voltage transformer (LPVT) to safely measure current and voltage at low power.

[0028] LPCT measures the phase current (IA, IB, IC) and LPVT measures the phase voltage (VA, VB, VC) and outputs it in a form that can be transmitted to the signal converter (200).

[0029] That is, LPCT measures the phase current (IA, IB, IC) and LPVT measures the phase voltage (VA, VB, VC) in real time, respectively, converts them into a standardized signal form that can be analyzed by the AI-based diagnostic monitoring unit, and outputs them to the signal converter (200).

[0030] The signal converter (200) processes low-power signals received from the LPCT and LPVT and converts them into measurable real-time voltage and current values ​​that can be used in AI diagnosis and protection relays.

[0031] This conversion is necessary because the output of the LPCT / LPVT is not suitable for direct measurement and needs to be converted into a standardized value. The converted voltage and current values ​​are provided to the diagnostic monitoring unit (500) for use in AI-based abnormal diagnosis, and at the same time, are supplied to protective relay devices such as digital relays to enable the performance of protective relay functions.

[0032] That is, the converted voltage and current values ​​are provided to the diagnostic monitoring unit (500) and can also be supplied to protective relay devices such as digital relays at the same time.

[0033] The signal converter (200) according to an embodiment of the present invention can convert current and voltage signals measured from the LPCT and LPVT into measurable voltage and current values ​​and provide them in real time to the digital protection relay (600) of FIG. 5.

[0034] The digital protection relay (600) performs power measurement and protection relay functions simultaneously using provided voltage and current data, and is responsible for interrupting and protecting the switchboard when abnormal situations such as overcurrent or ground fault occur.

[0035] That is, the signal converter converts the current and voltage signals measured from the LPCT and LPVT into measurable voltage and current values ​​and provides them in real time to the digital protection relay. The digital protection relay receives the voltage and current values ​​converted by the signal converter as input, performs power measurement functions during normal operation, and when an abnormal condition including overcurrent, ground fault, or short circuit is detected, simultaneously performs circuit breaker trip control, alarm output, and protection relay functions according to the set protection logic. This digital protection relay (600) is configured to implement stable operation of the equipment and immediate protection action in the event of a fault by processing power measurement data and protection relay events in parallel.

[0036] This configuration is intended to enable operation in parallel with an AI-based diagnostic system while maintaining existing power measurement and protection functions by converting low-power signals of LPCT and LPVT into standardized measurement values ​​and providing them directly to the protective relay.

[0037] The voltage and current values ​​converted by the signal converter (200) are supplied to the digital protection relay in real time and linked with the protection relay logic, enabling stable operation of the switchgear and rapid response to abnormalities.

[0038] The partial discharge converter (300) analyzes the current waveform signal of the LPCT to detect whether partial discharge occurs and pattern data, which are early signs of insulation breakdown within the electrical equipment.

[0039] The detected partial discharge occurrence status and pattern data are transmitted to the diagnostic monitoring unit (500) in real time and are used as a standard to detect the insulation deterioration state early and to determine in advance the possibility of failure due to local discharge.

[0040] The infrared thermal imaging measuring device (400) captures the thermal distribution of the electrical equipment in real time, generates thermal image data including the highest temperature value among them, and provides it to the diagnostic monitoring unit (500).

[0041] This device acquires the temperature distribution of the electrical equipment surface as pixel-unit thermal image data, analyzes the temperature value of each pixel to automatically detect the pixel or area exhibiting the highest temperature, and extracts the coordinates (X, Y) and the highest temperature value of the corresponding point.

[0042] This treatment is intended to detect localized overheating or insulation degradation in electrical equipment at an early stage and to quickly diagnose abnormal conditions.

[0043] The internal algorithm may use a histogram-based thresholding technique to find the maximum value among pixel values ​​or to find a value that is significantly higher than the average of a certain area, and the generated data is transmitted to the diagnostic monitoring unit (500) in real time.

[0044] For reference, the internal algorithm refers to an image analysis algorithm that processes thermal image data within an infrared thermal imaging measuring device (400) to extract the highest temperature value and the corresponding coordinates (X,Y).

[0045] In other words, it comprehensively refers to thermal image analysis logic that detects anomalies in pixel-based temperature data and determines the highest temperature coordinates, and in this embodiment, maximum value detection and histogram-based thresholding can be seen as representative examples.

[0046] Additionally, this algorithm can independently calculate the maximum temperature and coordinates for each piece of equipment by sequentially analyzing or parallel processing thermal images of multiple electrical facilities, and can include a function to detect rapid temperature increases early by tracking the rate of change of the temperature data map to ensure real-time performance.

[0047] In addition, correction logic can be embedded to minimize false alarms by analyzing the temperature change pattern over a certain period of time to distinguish between temporary temperature spikes and continuous overheating.

[0048] This infrared thermal imaging measuring device (400) acquires the surface temperature distribution of the electrical equipment as pixel-unit thermal imaging data, analyzes the temperature value of each pixel to automatically search for a pixel or area showing the highest temperature, extracts the coordinates (X, Y) of the point and the highest temperature value, and provides them to the diagnostic monitoring unit (500).

[0049] This device is capable of independent thermal imaging measurements for multiple electrical facilities and can identify the location of high-temperature anomalies in the form of coordinates.

[0050] For reference, the infrared thermal imaging measuring device according to the present embodiment captures the temperature distribution of the electrical equipment surface in real time and generates a pixel-unit temperature data map. Individual images (ROI, Region of Interest) can be collected for each electrical equipment component.

[0051] Since the generated thermal image data includes the temperature value of each pixel, the internal algorithm can use histogram-based thresholding to search for the maximum value among pixel values ​​or to find a value that is significantly higher than the average of a certain area.

[0052] A pixel or area having the highest temperature is identified, and the coordinates (X, Y) of the pixel and the highest temperature value are extracted and transmitted to the diagnostic monitoring unit (500).

[0053] The infrared thermal imaging measuring device (400) in this embodiment is a device that collects pixel-unit thermal imaging information by individually capturing the surface temperature distribution of a plurality of electrical equipment components.

[0054] The above measuring device automatically identifies the pixel or area showing the highest temperature in the thermal image data for each electrical facility, extracts the coordinates (X, Y) of the corresponding point and the highest temperature value, and provides them to the diagnostic monitoring unit (500).

[0055] This treatment is intended to enable early detection of localized overheating or insulation degradation in electrical equipment, and to accurately identify the location of the abnormality to enable rapid inspection and response.

[0056] The above measuring device is configured to derive the highest temperature point using a histogram-based thresholding algorithm that searches for the maximum value among pixel values ​​or finds a value that is significantly higher than the average of a certain area, and to transmit the data to the diagnostic monitoring unit (500) in real time.

[0057] The diagnostic monitoring unit (500) is an analysis device equipped with AI S / W, and receives real-time data provided from a signal converter (200), a partial discharge converter (300), and a thermal imaging measuring device (400), and determines whether there is an abnormality by comparing it with a normal operation reference value of the electrical equipment or through a machine learning-based abnormality detection algorithm.

[0058] When an abnormality is detected, the alarm level is classified, and response messages such as warnings, inspection recommendations, and requests for immediate action are generated and notified to the person in charge via a smart device.

[0059] In addition, the diagnostic monitoring unit (500) includes a reference value setting module, a judgment model based on labeled learning data, and an alarm level classification module, and can learn and determine various failure patterns on its own.

[0060] A diagnostic monitoring unit (500) according to an embodiment of the present invention includes a reference value setting module (510) that sets the average value, standard deviation, and upper / lower limit values ​​of voltage, current, partial discharge pattern, and thermal image temperature value collected under normal operating conditions of the electrical equipment as reference values.

[0061] This threshold value is used as a reference indicator to quickly identify abnormal conditions by comparing it with real-time input data.

[0062] Additionally, the diagnostic monitoring unit (500) includes a machine learning or deep learning-based judgment model unit (520) trained to classify abnormalities using a plurality of labeled learning data including voltage, current, partial discharge, and thermal image data in normal and faulty states.

[0063] This judgment model unit is configured to perform more precise anomaly detection than simple threshold comparison by learning various operating conditions and failure patterns in advance.

[0064] In addition, it includes an alarm level classification module (530) that classifies alarm levels into multiple stages according to the severity of abnormal conditions and outputs response messages such as warnings, inspection recommendations, and requests for immediate action according to each level, thereby enabling early response to failures and preventive maintenance.

[0065] A diagnostic monitoring unit (500) according to an embodiment of the present invention further includes a reference comparison module (540) that pre-stores voltage, current, partial discharge, and thermal image data measured in the normal operating state of the electrical equipment as reference data.

[0066] The above module directly compares data received in real time with reference data stored in advance to simply and quickly identify abnormal conditions, and by implementing a dual diagnostic system in parallel with a machine learning model, it ensures both the reliability and real-time nature of the diagnosis.

[0067] The diagnostic monitoring unit (500), LPCT, LPVT, and infrared thermal imaging measuring device (400) according to an embodiment of the present invention are implemented in a modular structure that can be additionally installed on an existing extra-high voltage switchgear.

[0068] This configuration enables the addition of real-time AI diagnostic and thermal imaging measurement functions without modifying the hardware of existing switchgear, thereby providing the effect of easily expanding the functionality of existing analog or simple protection relay systems into smart digital switchgear.

[0069] The LPCT, LPVT, thermal imaging measuring device, and diagnostic monitoring unit, which are components of the present invention, can be additionally installed in existing high-voltage switchgear systems, and the functionality can be expanded into a smart digital switchgear without changing the structure of the existing equipment.

[0070] Here, the diagnostic monitoring unit (500), the low-power measuring current transformer (LPCT) (100), the low-power measuring voltage transformer (LPVT) (100), and the infrared thermal imaging measuring device (400) are designed with a modular structure so that they can be additionally installed in an existing extra-high voltage switchgear.

[0071] The aforementioned 'modular structure' refers to a form in which each component is equipped with an independent housing and standard interfaces (power, communication, and signal input / output ports), allowing it to be attached or connected to an external or internal empty space without modifying the main circuit of an existing switchgear.

[0072] LPCT and LPVT can be attached to a three-phase power line where existing CT and PT are installed via a clamp or an insulating adapter for low-power measurement, and an infrared thermal imaging measuring device (400) is additionally installed in a non-contact manner on parts of the switchgear where heat generation is a concern, such as circuit breakers and busbars. The diagnostic monitoring unit (500) is mounted as an independent module in a transformer room or a protective relay and is connected in parallel with the existing protective relay system via a communication cable.

[0073] This modular design enables the implementation of additional measurement and AI-based diagnostic functions without altering the switchgear's original wiring and protection circuits, and offers the technical advantage of easily expanding functionality into a smart digital switchgear without modifying the structure of existing facilities.

[0074] The artificial intelligence software installed in the diagnostic monitoring unit (500) according to the present embodiment receives voltage, current, partial discharge pattern, and maximum thermal image temperature values ​​input from the signal converter (200), partial discharge converter (300), and infrared thermal image measuring device (400), and performs the function of determining in real time whether there is an abnormality in the electrical equipment by comparing it with predefined reference data.

[0075] The above artificial intelligence software enables early response by notifying the person in charge in real time of information in the form of an alarm message when an abnormal condition is detected.

[0076] In addition, the artificial intelligence software according to an embodiment of the present invention includes a machine learning-based anomaly detection algorithm that pre-learns or stores voltage, current, partial discharge, and thermal image data collected during the normal operating state of electrical equipment as reference data, and automatically determines whether there is an abnormal pattern by comparing real-time data received during operation with the reference data.

[0077] The above machine learning algorithm constructs an anomaly classification model based on training data labeled as normal and faulty states, and enables precise detection of complex anomaly patterns in electrical equipment under various operating conditions.

[0078] FIG. 3 is a system configuration diagram illustrating the signal flow and connection structure between each component in a smart digital switchgear diagnostic system based on LPIT (Low Power Instrument Transformer) according to the present invention.

[0079] This system is largely composed of a signal measurement unit (LPCT, LPVT), a signal processing and conversion unit, a protection relay unit (digital relay), a degradation analysis unit (infrared measuring instrument), and an AI-based diagnostic monitoring unit.

[0080] Regarding LPCT and LPVT, the LPCT (Low Power Current Transformer) measures the three-phase current in the switchgear in real time.

[0081] The LPVT (Low Power Voltage Transformer) measures the voltage of each phase in real time.

[0082] These two sensors detect high voltage and high current signals in a non-contact or low-power manner and output them as low-power signals.

[0083] In relation to the signal converter and partial discharge converter, the voltage and current signals transmitted from the LPCT and LPVT are transmitted to the signal converter and converted into measurable normal values ​​(e.g., RMS value, waveform data).

[0084] The high-speed current waveform received from the LPCT is internally input into a partial discharge converter, which analyzes whether partial discharge occurs and derives characteristic patterns (e.g., period, magnitude, peak value, etc.).

[0085] This device processes voltage, current, and partial discharge data simultaneously, and is configured and represented as a single integrated module in Fig. 3.

[0086] In relation to digital relays, the voltage and current signals output from the signal and partial discharge converters are provided to the digital protection relay.

[0087] Digital relays perform protection and control functions such as power measurement, overcurrent and ground fault protection, and trip determination.

[0088] In the present invention, the digital relay function maintains its existing function and is operated in a parallel or auxiliary relationship with the diagnostic system.

[0089] In relation to infrared thermal imaging measuring instruments, the surface temperature distribution of electrical equipment (e.g., circuit breakers, busbars, switches, etc.) within a switchboard is measured in a non-contact manner.

[0090] This device derives thermal images and maximum temperature values ​​in real time and transmits the corresponding data to an AI diagnostic module.

[0091] In some configurations, multiple units may be installed near the measurement unit or per electrical facility, and may also include coordinate location and time information.

[0092] Regarding AI S / W-equipped diagnostic monitoring, voltage, current, and partial discharge pattern data received from signal converters and partial discharge converters, and thermal image peak temperature and distribution data received from infrared measuring instruments are collected.

[0093] AI-based software (AI S / W) determines the presence or absence of abnormalities by comparing with predefined normal reference values ​​or based on a trained model.

[0094] Based on the judgment result, alarm levels are classified into No Abnormalities, Warning, Inspection Required, or Request for Immediate Action, and corresponding alarm messages are transmitted in real-time to the responsible person's smart device.

[0095] The diagnostic results are recorded and can be utilized in the long term for predicting the status of individual facilities and for preventive maintenance.

[0096] To elaborate, the diagnostic monitoring unit equipped with AI software is a device that diagnoses the presence or absence of abnormalities in electrical equipment in real time by integrating and analyzing data collected from measurement devices throughout the system.

[0097] This device collects voltage, current, and partial discharge pattern data received from a signal converter and a partial discharge converter, and also collects thermal image peak temperature and temperature distribution data received from an infrared thermal imaging detector to utilize as basic data for analysis.

[0098] This collected data is compared with predefined normal threshold values ​​within the AI ​​software or analyzed according to a machine learning-based anomaly detection algorithm using a trained classification model.

[0099] Based on the analysis results, the AI ​​software classifies the equipment status into four alert levels: 'no abnormalities', 'warning', 'inspection recommended', and 'immediate action required', and automatically generates alert messages according to each level.

[0100] The generated alert message is transmitted to the person in charge's smart device in real time, and as the alert level increases, the intensity of the response message and the notification method become more diverse.

[0101] For example, if the maximum thermal image temperature of the circuit breaker is 88°C and partial discharge signals are repeatedly detected, the AI ​​determines that it is in a warning state and sends a message recommending inspection.

[0102] In addition, if the thermal imaging temperature rises rapidly above 94℃ and the partial discharge signal increases abnormally, it is classified as an abnormal condition requiring immediate action and an immediate alarm is sent to the manager.

[0103] Diagnostic results are automatically recorded in an internal storage device, and the accumulated diagnostic history can be utilized for long-term analysis of electrical equipment condition trends, lifespan prediction, and preventive maintenance.

[0104] This configuration goes beyond a simple measurement system to realize a smart digital switchgear diagnostic system that enables advanced AI-based real-time diagnosis, alarms, and maintenance optimization.

[0105] The artificial intelligence software according to the present embodiment is software that includes a function to pre-learn or store reference data regarding the normal operating state of electrical equipment in order to determine the abnormal state of electrical equipment in real time.

[0106] This reference data is constructed by collecting voltage, current, partial discharge, and thermal image data measured during normal equipment operation, and is used for comparison with reference values ​​or as training data for machine learning models.

[0107] Reference values ​​include, for example, the rated ranges of voltage and current, the permissible limit of partial discharge energy, and the safety standard for the maximum thermal imaging temperature; these reference values ​​are stored within the algorithm and used as criteria for determining abnormalities.

[0108] Artificial intelligence software automatically determines whether the equipment status is normal by comparing measurement data input in real-time during operation with reference data or inputting it into a trained machine learning-based judgment model.

[0109] The above judgment is performed by considering complex factors, including not only whether a simple threshold is exceeded, but also abnormal changes in current and voltage waveforms, repeatability and irregularity of partial discharge signals, and whether there is a rapid rise in thermal imaging temperature.

[0110] The machine learning-based anomaly detection algorithm in the present invention takes various features, such as real-time voltage, current, PD patterns, and thermal image temperatures, as input values ​​and determines whether an anomaly exists by applying a classification algorithm based on supervised or unsupervised learning.

[0111] This algorithm is trained on historical driving data labeled as 'normal' and 'abnormal' to precisely classify whether the input data has a potential for failure.

[0112] If an abnormality is detected, an alarm level is set according to the severity, and the diagnostic monitoring unit generates messages such as warnings, inspection recommendations, or requests for immediate action to notify the person in charge.

[0113] This method provides higher diagnostic precision than a simple reference value comparison method and serves as a foundation for early detection of equipment failures and the realization of a preventive maintenance system.

[0114] Accordingly, the artificial intelligence software of the present invention enables advanced diagnosis of electrical equipment abnormalities by detecting abnormal patterns in real time based on reference data of a normal operating state and performing complex state analysis through a machine learning algorithm.

[0115] FIG. 4 shows the detailed configuration of the diagnostic monitoring unit of an extra-high voltage smart digital switchgear diagnostic system with LPIT applied according to one embodiment of the present invention.

[0116] As illustrated in FIG. 4, the diagnostic monitoring unit (510) includes a reference value setting module (510), a judgment model unit (520), an alarm level classification module (530), and a reference comparison module (540).

[0117] The reference value setting module (510) sets the average value, standard deviation, and upper / lower limit value of the voltage, current, partial discharge pattern, and thermal image temperature value collected under normal operating conditions of the electrical equipment as reference values.

[0118] The reference value setting module (510) in this embodiment is a module that performs the function of deriving the average value, standard deviation, upper limit, and lower limit of voltage, current, partial discharge pattern, and thermal image temperature value based on data collected while the electrical equipment is operating normally, and setting them as normal reference values.

[0119] These reference values ​​serve as a baseline for determining abnormalities by comparing them with real-time measurement data.

[0120] For example, when the maximum thermal image temperature during normal operation of circuit breaker A is an average of 76°C and a standard deviation of 4.2°C, and 95% of past measurements are 85°C or lower, the reference values ​​are set to an upper limit of 85°C, a lower limit of 60°C, etc.

[0121] Afterwards, if the temperature of the same circuit breaker exceeds 90℃, it is determined that the threshold value has been exceeded and an abnormal signal is generated.

[0122] The judgment model unit (520) is a machine learning or deep learning-based judgment model trained to classify abnormalities using a plurality of labeled training data including voltage, current, partial discharge, and thermal image data in normal and fault states.

[0123] This judgment model unit (520) is an artificial intelligence-based judgment module that performs the role of classifying abnormal states by training a machine learning or deep learning algorithm based on multiple training data labeled as 'normal state' and 'failure state'.

[0124] This configuration includes pattern recognition, data correlation analysis, and failure probability calculation, and enables more precise anomaly detection than the reference value method.

[0125] For example, if sufficient cases have been learned from past data where the combination of 'partial discharge occurrence + increased current fluctuation + sudden spike in thermal image' led to actual failures, when the same pattern is detected in real-time data, the judgment model classifies it as an abnormal state and delivers a 'recommendation for inspection' or 'immediate action.'

[0126] The alarm level classification module (530) classifies alarm levels into multiple stages according to the severity of the abnormal condition and outputs response messages such as warnings, inspection recommendations, and requests for immediate action according to each alarm level.

[0127] This alarm level classification module (530) performs the role of classifying alarm levels into multiple stages according to the severity of the determined abnormal state and generating a response message according to each alarm level.

[0128] This module includes various response levels ranging from simple anomalies to emergency situations (e.g., warnings, inspection recommendations, requests for immediate action).

[0129] For example, if the PD energy value is repeatedly slightly higher than the threshold, it is classified as a 'Warning' grade, and if the maximum thermal temperature rises above 93℃ while the PD increases rapidly at the same time, it is classified as a 'Request for Immediate Action' grade and a warning message is sent in real time to the relevant personnel.

[0130] The standard comparison module (540) stores standard data for normal operation of electrical equipment in advance and determines an abnormal state by comparing it with the received real-time data.

[0131] This standard comparison module (540) is responsible for the function of determining whether there is an anomaly by directly comparing the standard value set in the standard value setting module with the received real-time data.

[0132] This module has the advantage of being able to determine abnormal conditions simply and quickly by checking whether a threshold is exceeded, without complex learning.

[0133] For example, if the current exceeds 130A, which is above the normal reference value of 100A ± 20%, the reference comparison module determines this as an abnormal state and transmits a warning signal to the alarm level classification module.

[0134] As such, the internal configuration of the diagnostic monitoring unit according to Fig. 4 is structured to enhance the real-time diagnosis and preventive maintenance system of the equipment by including not only a simple determination of exceeding a reference value, but also precise state classification linked to a learned judgment model and an automated response message delivery function.

[0135] The data flow is as follows.

[0136] First, the current and voltage signals measured by the LPCT and LPVT are transmitted to a signal converter and a partial discharge converter, converted into measurable electrical signals and partial discharge patterns, and then simultaneously provided to a digital relay and a diagnostic monitoring unit.

[0137] Meanwhile, thermal image data measured by the infrared thermal imaging instrument is transmitted to the diagnostic monitoring unit and used to analyze whether there are any temperature abnormalities in the electrical equipment.

[0138] Subsequently, the diagnostic monitoring unit determines whether there are any abnormalities based on the received data, and if an abnormality is detected, notifies the relevant information in real time to the person in charge's smart device.

[0139] According to the present invention, technical effects such as expanded measurement range (including thermal imaging and partial discharge), improved diagnostic accuracy, AI-based self-judgment, and cost reduction are provided compared to existing switchgear systems.

[0140] The form shown in Fig. 2 can be seen as a form in which a diagnostic system is optionally added to an existing switchgear.

[0141] FIG. 5 is a configuration diagram of an extra-high voltage smart digital switchgear diagnostic system with LPIT applied according to another embodiment of the present invention. FIG. 5 is an exemplary embodiment based on the configuration of FIG. 3.

[0142] The measuring unit (100) includes a low-power measuring current transformer (LPCT) and a low-power measuring voltage transformer (LPVT) installed in an extra-high voltage switchgear to measure current and voltage, and the measured signal is transmitted to a signal converter (200).

[0143] The signal converter (200) converts the signals input from the LPCT and LPVT into measurable voltage and current values ​​and provides them simultaneously to the diagnostic monitoring unit (500) and the digital protection relay (600).

[0144] The digital protection relay (600) performs power measurement and protection relay functions using the above voltage and current data, and is responsible for interrupting and protecting the switchboard when abnormal situations such as overcurrent or ground fault occur.

[0145] The partial discharge converter (300) analyzes the partial discharge pattern from the current waveform of the LPCT and provides it to the diagnostic monitoring unit (500), and the infrared thermal imaging measuring device (400) measures the degradation distribution of the electrical equipment in real time and provides thermal imaging data to the diagnostic monitoring unit (500).

[0146] The diagnostic monitoring unit (500) integrates and analyzes data provided by the signal converter (200), partial discharge converter (300), and infrared thermal imaging measuring device (400), determines in real time whether there is an abnormality in the electrical equipment through AI software, and outputs the result in the form of an alarm message.

[0147] In this way, the digital protection relay (600) of FIG. 5 performs existing power measurement and protection functions between the signal converter (200) and the diagnostic monitoring unit (500), while the diagnostic monitoring unit (500) processes data in parallel with the digital protection relay (600) to simultaneously implement protection relay and AI-based diagnosis. As a result, existing protection functions and AI-based diagnosis functions can be integrated and implemented in a single system, and real-time safety of power equipment and early response to abnormalities can be achieved.

Claims

Claim 1 A smart digital switchgear diagnostic system for diagnosing abnormal conditions of electrical equipment in real time comprises: a measuring unit (100) composed of a low-power measuring current transformer (LPCT) and a low-power measuring voltage transformer (LPVT), installed in an extra-high voltage switchgear to measure current and voltage; a signal converter (200) that converts signals collected from the LPCT and LPVT into measurable voltage and current values; a partial discharge converter (300) that receives a current waveform signal provided from the LPCT and analyzes a partial discharge pattern; an infrared thermal imaging measuring device (400) that measures the real-time degradation distribution from electrical equipment and outputs thermal image data, wherein the temperature distribution of the electrical equipment surface is acquired as pixel-unit thermal image data, analyzes the temperature value of each pixel to automatically search for a pixel or area showing the highest temperature, extracts the coordinates (X, Y) and the highest temperature value of the corresponding point, identifies the point showing the highest temperature, and provides the coordinates and temperature information of the corresponding point to a diagnostic monitoring unit; and the signal converter, the partial discharge converter, and the infrared thermal image A diagnostic system for an extra-high voltage smart digital switchgear with LPIT applied, characterized by including a diagnostic monitoring unit (500) equipped with artificial intelligence software (AI S / W) that determines an abnormal state by analyzing the maximum temperature value extracted using a plurality of labeled learning data including voltage, current, partial discharge, and thermal image data in normal and fault states and voltage, current, and pattern data measured in real time using learned artificial intelligence S / W to integrate and analyze real-time voltage, current, partial discharge analysis values, and thermal image data provided from each measuring device, and compares the occurrence of the maximum temperature value with reference data. Claim 2 delete Claim 3 An extra-high voltage smart digital switchgear diagnostic system with LPIT applied according to claim 1, wherein the signal converter converts current and voltage signals measured from the LPCT and LPVT into measurable voltage and current values ​​and provides them in real time to a digital protection relay, and the digital protection relay receives the voltage and current values ​​converted by the signal converter as input, performs a power measurement function during normal operation, and when an abnormal condition including overcurrent, ground fault, or short circuit is detected, simultaneously performs circuit breaker trip control, alarm output, and protection relay functions according to a set protection logic. Claim 4 A high-voltage smart digital switchgear diagnostic system with LPIT applied, characterized in that, in claim 1, the artificial intelligence software mounted on the diagnostic monitoring unit determines whether there is an abnormality by comparing voltage, current, partial discharge pattern, and maximum thermal temperature value input from a signal converter, a partial discharge converter, and an infrared thermal imaging measuring device with reference data, and if an abnormality is determined, notifies the person in charge of the relevant information in real time. Claim 5 A high-voltage smart digital switchgear diagnostic system with LPIT applied according to claim 1, characterized in that the artificial intelligence software includes a machine learning-based anomaly detection algorithm that pre-learns or stores reference data regarding the normal operating state of electrical equipment and determines whether there is an abnormal pattern by comparing it with data received in real time. Claim 6 In claim 1, the diagnostic monitoring unit (500) is characterized by comprising: a reference value setting module (510) that sets the average value, standard deviation, and upper / lower limit value of voltage, current, partial discharge pattern, and thermal image temperature value collected under normal operating conditions of the electrical equipment as reference values; a machine learning or deep learning-based judgment model unit (520) trained to classify abnormalities using a plurality of labeled learning data including voltage, current, partial discharge, and thermal image data in normal and fault states; and an alarm level classification module (530) that classifies alarm levels into multiple stages according to the severity of the abnormal state and outputs response messages such as warning, inspection recommendation, and immediate action request according to each alarm level. Claim 7 A high-voltage smart digital switchgear diagnostic system with LPIT applied, characterized in that, in claim 1, the diagnostic monitoring unit (500) further includes a standard comparison module (540) that stores standard data for normal operation of electrical equipment in advance and determines an abnormal state by comparing it with received real-time data. Claim 8 A high-voltage smart digital switchgear diagnostic system with LPIT applied, characterized in that, in claim 1, the diagnostic monitoring unit, LPCT, LPVT, and infrared thermal imaging measuring device can be additionally installed on an existing high-voltage switchgear, thereby enabling the existing switchgear to be functionally expanded into a smart digital switchgear.

Images