Building electrical fire hazard intelligent monitoring system and monitoring method thereof

Abstract

The invention discloses a building electrical fire hazard intelligent monitoring system and a monitoring method thereof. The system comprises an infrared thermal imaging module, a current harmonic analysis module, an edge computing node, a self-healing protection module and a user interaction terminal. The infrared thermal imaging module collects temperature field data and dynamically adjusts an alarm threshold value. The current harmonic analysis module collects harmonic characteristic data and identifies load abnormity. The edge computing node fuses the multi-dimensional data based on a random forest algorithm to generate a risk level and a fault type; the self-healing protection module adopts a three-stage tripping strategy to realize graded protection, and improves the handling capacity through closed-loop verification and fire-fighting linkage; and the user interaction terminal realizes fault information visualization. Through multi-dimensional data perception, intelligent analysis and hierarchical active protection strategies, a whole-process closed-loop management system from hidden danger detection, risk assessment to self-healing treatment is constructed, and the problems of high false alarm rate, response lag and management fragmentation in the prior art are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of building electrical safety monitoring, in particular to a building electrical fire hazard intelligent monitoring system and a monitoring method thereof. BACKGROUND

[0002] With the continuous improvement of the degree of electrification of buildings, various electrical equipment is increasingly widely used in buildings, and the operation state of electrical nodes such as power distribution cabinets, cable bridges, and busbars is directly related to the safety of building electricity. However, internal overheating and harmonic distortion caused by problems such as poor contact, overload, and insulation aging of these nodes have become the primary cause of electrical fires.

[0003] The current traditional electrical fire monitoring technology has obvious limitations: first, it relies on single temperature or current parameter collection, which is difficult to capture early fault characteristics. For example, the early stage of cable joint oxidation only shows harmonic distortion, and the temperature change is weak, so the traditional monitoring technology is prone to delayed early warning, and environmental temperature interference is prone to false alarms; second, fault location requires manual inspection, and the response cycle is long, often missing the initial disposal window; third, the fixed trip threshold of the circuit breaker cannot be flexibly adjusted according to the risk level, which is prone to excessive power failure affecting normal power consumption or insufficient protection causing safety accidents; fourth, there is a lack of standardized maintenance knowledge base support, and maintenance personnel cannot quickly obtain effective repair guidance, resulting in low repair efficiency.

[0004] Therefore, there is an urgent need for an electrical fire monitoring solution that can perceive multiple dimensions, make edge intelligent decisions, and provide graded self-healing protection to address the shortcomings of traditional monitoring technology and improve the accuracy, timeliness, and reliability of building electrical fire hazard monitoring. SUMMARY

[0005] The present application aims to provide a building electrical fire hazard intelligent monitoring system and a monitoring method thereof, which constructs a full-process closed-loop management system from hazard detection, risk assessment to self-healing disposal through multi-dimensional data perception, intelligent algorithm fusion and proactive protection strategy, solving the problems of high false alarm rate, delayed response and fragmented management in the prior art.

[0006] To achieve the above-mentioned purpose, the present application provides a building electrical fire hazard intelligent monitoring system, comprising:

[0007] an infrared thermal imaging module for collecting temperature field data of the electrical system in the target building and identifying overheating areas, and dynamically adjusting the temperature alarm threshold according to the environmental temperature and the rated temperature of the equipment;

[0008] a current harmonic analysis module for collecting harmonic feature data of the electrical lines in the target building and identifying abnormal operating states of the load;

[0009] The edge computing node is configured to fuse the temperature field data and the harmonic feature data based on a random forest algorithm to generate a risk assessment result, wherein the risk assessment result comprises an electrical fire risk level and an electrical fault type.

[0010] The self-healing protection module is configured to dynamically adjust a trip threshold of a corresponding circuit breaker in the electrical system or trigger the corresponding circuit breaker to open and parallelly control a strong cut device of a fire power supply according to the electrical fire risk level in the risk assessment result and a preset hierarchical early warning strategy, wherein the circuit breaker is an intelligent circuit breaker with an adjustable trip threshold.

[0011] The user interaction terminal is configured to visually display fault positioning information, repair guidance information, and spare part information.

[0012] Preferably, the infrared thermal imaging module comprises:

[0013] A non-cooled focal plane array infrared detector is arranged on a servo gimbal, and the non-cooled focal plane array infrared detector is configured to perform multi-angle scanning on key electrical components of the electrical system according to a preset scanning track to generate a three-dimensional temperature field cloud picture, wherein the key electrical components comprise busbars in a power distribution cabinet and circuit breaker contacts.

[0014] A overheated area detection unit is configured to identify a local overheated area caused by an abnormal state of the electrical system based on a defect detection model trained by a YOLOv5 algorithm from the three-dimensional temperature field cloud picture, wherein the abnormal state of the electrical system comprises cable joint oxidation and screw loosening.

[0015] An alarm threshold dynamic adjustment unit is configured to dynamically adjust the temperature alarm threshold according to an ambient temperature and a device rated temperature rise according to a preset threshold adjustment formula.

[0016] A visible light camera is arranged on the servo gimbal, and the visible light camera is configured to shoot a close-up image of a local part of the electrical system when a local temperature of the electrical system exceeds the temperature alarm threshold and upload the close-up image to the edge computing node.

[0017] Preferably, the threshold adjustment formula is as follows:

[0018] T alarm = T env + 0.8 × ΔT rated

[0019] wherein T alarm is an adjusted temperature alarm threshold, T env is an ambient temperature, and ΔT rated is a device rated temperature rise.

[0020] Preferably, the current harmonic analysis module comprises:

[0021] a current signal acquisition unit configured to acquire a current signal in an electrical line;

[0022] a harmonic feature extraction unit configured to extract harmonic feature data in the current signal by using a fast Fourier transform algorithm, wherein the harmonic feature data comprises 2-40th harmonic content, total harmonic distortion, and inter-harmonic amplitude;

[0023] a load state identification unit configured to establish a harmonic-load correlation database, and identify an abnormal operating state of a load in the electrical system based on the harmonic feature data and the harmonic-load correlation database by a pattern matching algorithm.

[0024] Preferably, the edge computing node comprises:

[0025] an input vector construction unit configured to construct a multi-modal input vector based on the temperature field data and the harmonic feature data, wherein the multi-modal input vector comprises a hotspot maximum temperature, a temperature rise rate, a ratio of 3rd harmonic content to 5th harmonic content, a ratio of 5th harmonic content to 7th harmonic content, and a total harmonic distortion;

[0026] a risk score unit configured to perform risk assessment on the electrical system based on the multi-modal input vector by a pre-trained risk assessment model, and output a risk score and an electrical fault type, wherein the risk assessment model is trained by a forest random algorithm on historical multi-modal input vectors of electrical systems;

[0027] a risk level determination unit configured to determine an electrical fire risk level according to the risk score according to a preset risk grading determination strategy, wherein the electrical fire risk level comprises low risk, medium risk, and high risk.

[0028] Preferably, the self-healing protection module comprises:

[0029] a hierarchical early warning unit configured to control a circuit breaker in a non-critical load circuit in the electrical system by a three-level tripping strategy based on the electrical fire risk level, wherein,

[0030] when the electrical fire risk level is low risk, adjusting the tripping threshold of the circuit breaker of the corresponding load circuit to 90% of the initial tripping threshold, while recording the temperature change trend information of the corresponding load circuit and sending an inspection reminder;

[0031] when the electrical fire risk level is medium risk, adjusting the tripping threshold of the circuit breaker of the corresponding load circuit to 80% of the initial tripping threshold, and starting a preset load transfer control strategy to transfer the load of the corresponding load circuit to a backup circuit;

[0032] When the electrical fire risk level is high risk, the corresponding load circuit breaker is tripped, and the emergency power supply is started to push the repair work order;

[0033] A closed-loop verification unit is used to detect the insulation resistance and temperature rise information of the corresponding load circuit in real time after tripping, and when the insulation resistance of the corresponding load circuit is greater than or equal to the preset resistance threshold and the temperature rise rate is within the preset temperature rise rate threshold, the fault is marked as cleared, and a closing control instruction is sent to the breaker of the corresponding load circuit;

[0034] A linkage control unit is in communication connection with the fire fighting system of the target building, and the linkage control unit is used to control the non-fire fighting power supply to be forcibly cut off and control the emergency lighting system and the smoke exhaust system to be turned on when it is confirmed that an electrical fire has occurred.

[0035] To achieve the above-mentioned purpose, the present application also provides a monitoring method of the building electrical fire hazard intelligent monitoring system as described in any one of the above, comprising the following steps:

[0036] S1, multi-dimensional data acquisition: temperature field data of the electrical system in the target building are collected by an infrared thermal imaging module, and the temperature alarm threshold is dynamically adjusted, and harmonic feature data of the electrical circuit in the target building are collected by a current harmonic analysis module;

[0037] S2, multi-modal data fusion analysis: the temperature field data and the harmonic feature data are fused by an edge computing node based on a random forest algorithm to generate an electrical fire risk level and an electrical fault type;

[0038] S3, hierarchical self-healing protection control: the self-healing protection module dynamically adjusts the tripping threshold of the corresponding circuit breaker according to the electrical fire risk level, or triggers the corresponding circuit breaker to trip and the fire linkage control;

[0039] S4, fault information visualization interaction: the fault positioning, repair guidance and spare parts information are displayed through a user interaction terminal.

[0040] Preferably, step S1 specifically comprises:

[0041] S11, a non-cooled focal plane array infrared detector is used to scan key electrical components according to a preset trajectory to generate a three-dimensional temperature field cloud map;

[0042] S12, a local overheating area in the three-dimensional temperature field cloud map is identified based on a YOLOv5 defect detection model;

[0043] S13, the adjusted temperature alarm threshold is calculated according to the formula T alarm = T env + 0.8 x ΔT rated , wherein T alarmT is the adjusted temperature alarm threshold env T is the ambient temperature, ΔT rated T is the device rated temperature rise;

[0044] S14, when the local temperature exceeds the temperature alarm threshold, a close-up image of the local part is taken by a visible light camera;

[0045] S15, an electrical line current signal is collected, and a fast Fourier transform algorithm is used to extract harmonic feature data in the current signal, wherein the harmonic feature data includes 2-40 harmonic content, total harmonic distortion, and inter-harmonic amplitude;

[0046] S16, based on the harmonic-load correlation database, a pattern matching algorithm is used to identify the abnormal operation state of the load.

[0047] Preferably, step S2 specifically comprises:

[0048] S21, a multi-modal input vector is constructed, which contains the hotspot maximum temperature, temperature rise rate, 3rd and 5th harmonic content ratio, 5th and 7th harmonic content ratio, and total harmonic distortion;

[0049] S22, the multi-modal input vector is input into a pre-trained risk assessment model, and a risk score and an electrical fault type are output, wherein the risk assessment model is trained by a forest random algorithm on historical multi-modal input vectors of the electrical system;

[0050] S23, according to the risk score, a preset risk classification determination strategy is used to determine the electrical fire risk level, wherein the electrical fire risk level includes low risk, medium risk, and high risk.

[0051] Preferably, step S3 specifically comprises:

[0052] S31, if the electrical fire risk level is low, the tripping threshold of the corresponding circuit breaker is adjusted to 90% of the initial tripping threshold, the temperature change trend is recorded, and an inspection reminder is sent;

[0053] S32, if the electrical fire risk level is medium, the tripping threshold of the corresponding circuit breaker is adjusted to 80% of the initial tripping threshold, a load transfer control strategy is started to transfer the load to a backup circuit;

[0054] S33, if the electrical fire risk level is high, the corresponding circuit breaker is controlled to be tripped, an emergency power supply is started, and a repair work order is pushed;

[0055] S34: After the circuit is tripped, the insulation resistance and temperature rise information of the circuit are detected in real time. When the insulation resistance is greater than or equal to the preset resistance threshold and the temperature rise rate is within the preset temperature rise rate threshold, the fault is cleared and a closing control command is sent.

[0056] S35, when an electrical fire is confirmed, control the forced disconnection of non-fire-fighting power supplies and activate emergency lighting and smoke extraction systems.

[0057] This application achieves intelligent monitoring and handling of electrical fire hazards throughout the entire process through the collaborative work of an infrared thermal imaging module, a current harmonic analysis module, an edge computing node, a self-healing protection module, and a user interaction terminal. This avoids the limitations of traditional single-parameter monitoring and improves the comprehensiveness of hazard identification. The edge computing node enables localized real-time analysis and uses a random forest algorithm to fuse multi-dimensional data for risk assessment, improving the accuracy of risk level and fault type judgment. The self-healing protection module's hierarchical early warning strategy and intelligent circuit breaker control enable flexible protection based on the degree of risk, avoiding losses caused by excessive power outages and promptly cutting off dangerous circuits in high-risk situations to ensure safety. The user interaction terminal's visual display facilitates users' quick access to key information and improves fault handling efficiency. Attached Figure Description

[0058] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0059] Figure 1 This is a schematic diagram of the structure of an intelligent monitoring system for electrical fire hazards in buildings, as described in an embodiment of this application.

[0060] Figure 2 This is a flowchart of a detection method for an intelligent monitoring system for electrical fire hazards in buildings, as described in this application. Detailed Implementation

[0061] To enable those skilled in the art to better understand the technical solutions in this application, the technical solutions in the embodiments of this application will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0062] In the embodiments of the present application, it should be understood that the disclosed method and system can be implemented in other manners. The embodiments described below are merely schematic. For example, the division of the units and modules is merely logical function division, and there can be other division manners in actual implementation. For example, a plurality of units or modules can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed coupling, or direct coupling or communication connection between the components can be indirect coupling or communication connection through some interfaces, devices or modules, and can be electrical, mechanical or other forms.

[0063] In addition, each function unit in each embodiment of the present application can be integrated into one processor, or each unit can be a separate device, or two or more units can be integrated into one device. Each function unit in each embodiment of the present application can be implemented in the form of hardware or in the form of hardware plus software function unit.

[0064] Those skilled in the art can understand that all or part of the steps of the following method embodiments can be completed by program instructions and related hardware. The aforementioned program instructions can be stored in a computer readable storage medium, and the program instructions are executed to perform the steps of the method embodiments. The aforementioned storage medium includes mobile storage devices, read-only memories (ROM), magnetic disks or optical disks, and various media that can store program codes.

[0065] In addition, the terms "first", "second" are used only for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first", "second" can explicitly or implicitly include one or more of the features. In the description of the present application, the meaning of "a plurality of" or "several" is two or more, unless otherwise explicitly specified.

[0066] As shown in the Figure 1 The embodiments of the present application provide a building electrical fire hazard intelligent monitoring system, which comprises:

[0067] The infrared thermal imaging module 101 is configured to collect temperature field data of the electrical system in the target building and identify overheated areas, and dynamically adjust a temperature alarm threshold according to an ambient temperature and a device rated temperature;

[0068] The current harmonic analysis module 102 is configured to collect harmonic characteristic data of the electrical line in the target building and identify an abnormal running state of the load;

[0069] The edge computing node 103 is configured to fuse the temperature field data and the harmonic feature data based on a random forest algorithm to generate a risk assessment result, wherein the risk assessment result comprises an electrical fire risk level and an electrical fault type.

[0070] The self-healing protection module 104 is configured to dynamically adjust a trip threshold of a corresponding circuit breaker in the electrical system or trigger the corresponding circuit breaker to trip and parallelly control a fire power strong cutting device according to the electrical fire risk level in the risk assessment result and a preset hierarchical early warning strategy, wherein the circuit breaker is a smart circuit breaker with an adjustable trip threshold.

[0071] The user interaction terminal 105 is configured to visually display fault positioning information, repair guidance information and spare part information.

[0072] In this embodiment, the infrared thermal imaging module 101 is responsible for collecting temperature field data of the electrical system, identifying overheating areas, and dynamically adjusting a temperature alarm threshold in combination with an environmental temperature and a rated temperature rise of a device to ensure the adaptability of the alarm threshold. The current harmonic analysis module 102 is configured to collect harmonic feature data of the electrical circuit and identify abnormal operating states of the load by analyzing the harmonic features. The edge computing node 103 receives data from the infrared thermal imaging module 101 and the current harmonic analysis module 102, fuses the temperature field data and the harmonic feature data by using a random forest algorithm, and generates a risk assessment result comprising an electrical fire risk level and an electrical fault type through analysis and calculation of an algorithm model. The self-healing protection module 104 controls the smart circuit breaker with an adjustable trip threshold in the electrical system according to the electrical fire risk level in the risk assessment result and a preset hierarchical early warning strategy, dynamically adjusts the trip threshold of the smart circuit breaker, or triggers the circuit breaker to trip when the risk is high and parallelly controls the fire power strong cutting device. The user interaction terminal 105 visually displays the fault positioning information, the repair guidance information and the spare part information to the user, so that the user can intuitively understand the relevant situation and take corresponding measures.

[0073] The multiple modules in this embodiment work cooperatively to realize multi-dimensional monitoring of building electrical fire hazards, avoid the limitations of traditional single parameter monitoring, and improve the comprehensiveness of hazard identification. The random forest algorithm of the edge computing node 103 fuses multi-dimensional data for risk assessment, which improves the accuracy of risk level and fault type judgment. The hierarchical early warning strategy of the self-healing protection module 104 and the control of the smart circuit breaker realize flexible protection according to the risk level, which avoids the loss caused by excessive power-off and timely cuts off the dangerous circuit in high risk to ensure safety. The visual display of the user interaction terminal 105 facilitates the user to quickly obtain key information and improves the fault handling efficiency.

[0074] In one embodiment, the infrared thermal imaging module 101 comprises:

[0075] The non-cooled focal plane array infrared detector is arranged on the servo gimbal, and is used for multi-angle scanning of key electrical components of the electrical system according to a preset scanning track to generate a three-dimensional temperature field cloud picture, wherein the key electrical components include busbars and circuit breaker contacts in the power distribution cabinet.

[0076] The overheated area detection unit is used for identifying a local overheated area caused by an abnormal state of the electrical system based on a defect detection model trained by a YOLOv5 algorithm based on the three-dimensional temperature field cloud picture, wherein the abnormal state of the electrical system includes cable joint oxidation and screw loosening.

[0077] The alarm threshold dynamic adjustment unit is used for dynamically adjusting a temperature alarm threshold according to an ambient temperature and a device rated temperature rise according to a preset threshold adjustment formula.

[0078] The visible light camera is arranged on the servo gimbal, and is used for shooting a close-up image of a local area when a local temperature of the electrical system exceeds the temperature alarm threshold and uploading the close-up image to the edge computing node 103.

[0079] In the embodiment, the non-cooled focal plane array infrared detector in the infrared thermal imaging module 101 is installed on the servo gimbal, the servo gimbal drives the detector to perform multi-angle scanning on key electrical components such as busbars and circuit breaker contacts in the power distribution cabinet according to a preset scanning track, the detector converts received infrared radiation signals into electrical signals, and a three-dimensional temperature field cloud picture is generated after processing, which can clearly reflect the temperature distribution of the key components; the overheated area detection unit analyzes the three-dimensional temperature field cloud picture by using a defect detection model trained based on a YOLOv5 algorithm, the model identifies temperature abnormal areas in the cloud picture to determine whether there is a local overheated area caused by an abnormal state of the electrical system such as cable joint oxidation and screw loosening; the alarm threshold dynamic adjustment unit calculates and dynamically adjusts the temperature alarm threshold according to a real-time collected ambient temperature and a device rated temperature rise according to a preset threshold adjustment formula; the visible light camera is also arranged on the servo gimbal, and when the overheated area detection unit detects that a local temperature of the electrical system exceeds the adjusted temperature alarm threshold, the visible light camera immediately shoots the local area to obtain a close-up image and uploads the close-up image to the edge computing node 103, thereby providing a visual basis for further analysis of the edge computing node 103.

[0080] In this embodiment, the infrared thermal imaging module 101 generates a three-dimensional temperature field cloud map through multi-angle scanning using an uncooled focal plane array infrared detector combined with a servo gimbal. This map comprehensively and accurately reflects the temperature status of key electrical components, providing reliable data for overheating area identification. The defect detection model based on the YOLOv5 algorithm has high recognition accuracy and speed, enabling rapid and accurate identification of local overheating areas and corresponding abnormal states, thus improving the timeliness of early hazard detection. The dynamic adjustment of the alarm threshold avoids the impact of ambient temperature changes on alarm accuracy and reduces the false alarm rate. Close-up images captured by the visible light camera provide intuitive evidence for fault analysis of the edge computing node 103 and for users to understand the fault site conditions, which is helpful for subsequent processing.

[0081] In one embodiment, the threshold adjustment formula is as follows:

[0082] T alarm =T env +0.8×ΔT rated

[0083] Among them, T alarm The adjusted temperature alarm threshold, T env For ambient temperature, ΔT rated This is the rated temperature rise of the equipment.

[0084] In this embodiment, the alarm threshold dynamic adjustment unit adjusts according to formula T. alarm =T env +0.8×ΔT rated The adjusted temperature alarm threshold is calculated by adding the ambient temperature to 0.8 times the equipment's rated temperature rise, thus obtaining a temperature alarm threshold suitable for the current environmental conditions. This formula takes into account the impact of ambient temperature on the equipment's operating temperature, while using 0.8 times the equipment's rated temperature rise as a safety margin to ensure timely warnings are issued before the equipment approaches its rated temperature rise limit.

[0085] The threshold adjustment formula is simple and practical, and can quickly calculate alarm thresholds that are suitable for different ambient temperatures, making the alarm thresholds more consistent with the actual operating conditions of the equipment. By introducing the ambient temperature parameter, it avoids the problem of false alarms or missed alarms caused by the fixed alarm threshold due to differences in ambient temperature. The setting of 0.8 times the rated temperature rise provides a certain safety margin for equipment operation, and can trigger the alarm in time before the equipment temperature reaches a dangerous level, thus improving the reliability of the early warning.

[0086] In one embodiment, the current harmonic analysis module 102 includes:

[0087] A current signal acquisition unit, used to acquire current signals in electrical circuits;

[0088] a harmonic feature extraction unit configured to extract harmonic feature data in the current signal using a fast Fourier transform algorithm, wherein the harmonic feature data comprises 2-40th harmonic content ratio, total harmonic distortion, and inter-harmonic amplitude;

[0089] a load state recognition unit configured to establish a harmonic-load correlation database, and recognize abnormal operation state of the load in the electrical system based on the harmonic feature data and the harmonic-load correlation database through a pattern matching algorithm.

[0090] In this embodiment, the current signal acquisition unit of the current harmonic analysis module 102 acquires the current signal in the electrical circuit through a current sensor or the like; the harmonic feature extraction unit processes the acquired current signal using a fast Fourier transform algorithm, converts the time-domain signal into a frequency-domain signal, and thereby extracts the harmonic feature data such as 2-40th harmonic content ratio, total harmonic distortion, and inter-harmonic amplitude; and the load state recognition unit pre-establishes a harmonic-load correlation database, which stores the corresponding harmonic feature data under normal and abnormal operation states of different loads, and judges whether the operation state of the load in the electrical system is abnormal by performing pattern matching between the current harmonic feature data and the data in the database.

[0091] The current harmonic analysis module 102 of this embodiment can efficiently and accurately extract harmonic feature data through a fast Fourier transform algorithm, and provide high-quality data support for load state recognition; the establishment of the harmonic-load correlation database provides a clear basis for judging the load state through harmonic features, and improves the accuracy of abnormal operation state recognition of the load; comprehensive analysis of 2-40th harmonic and inter-harmonic can capture the characteristics of different types of loads, and expand the range of load abnormality monitoring.

[0092] In one embodiment, the edge computing node 103 comprises:

[0093] an input vector construction unit configured to construct a multi-modal input vector based on the temperature field data and the harmonic feature data, wherein the multi-modal input vector comprises hotspot maximum temperature, temperature rise rate, ratio of 3rd harmonic content ratio to 5th harmonic content ratio, ratio of 5th harmonic content ratio to 7th harmonic content ratio, and total harmonic distortion;

[0094] a risk score unit configured to perform risk assessment on the electrical system based on the multi-modal input vector through a pre-trained risk assessment model, and output a risk score and an electrical fault type, wherein the risk assessment model is trained through a forest random algorithm based on historical multi-modal input vectors of the electrical system;

[0095] a risk grade determination unit configured to determine an electrical fire risk grade according to the risk score and a preset risk grading determination policy, wherein the electrical fire risk grade comprises low risk, medium risk and high risk.

[0096] In this embodiment, the input vector construction unit of the edge computing node 103 obtains the hotspot maximum temperature, temperature rise rate and other temperature field key data from the infrared thermal imaging module 101, obtains the ratio of 3rd harmonic content to 5th harmonic content, the ratio of 5th harmonic content to 7th harmonic content and total harmonic distortion rate and other harmonic characteristic key data from the current harmonic analysis module 102, and integrates and constructs the multi-modal input vector from the data; the risk scoring unit inputs the constructed current multi-modal input vector into the risk assessment model obtained by previously training a large number of electrical system historical multi-modal input vectors through a random forest algorithm, and outputs the risk score and the corresponding electrical fault type through weight analysis and comprehensive calculation of each characteristic parameter. The electrical fault type can include poor contact, overload and insulation aging, etc. The risk grade determination unit determines the corresponding electrical fire risk grade according to the preset risk grading determination policy. The risk score is calculated in percentage. The risk score greater than 0 and less than or equal to 30 is classified as low risk, greater than 30 and less than or equal to 70 is classified as medium risk, and greater than 70 and less than or equal to 100 is classified as high risk.

[0097] The edge computing node 103 of this embodiment integrates temperature and harmonic multi-dimensional key data through the construction of the multi-modal input vector, provides comprehensive information for risk assessment, and avoids the limitation of single type of data. The risk assessment model based on the random forest algorithm has strong generalization ability and anti-interference ability. After training with historical data, it can accurately output the risk score and the fault type, and improve the reliability of risk assessment. The clear risk grade division provides a clear basis for the subsequent hierarchical control of the self-healing protection module 104, and facilitates precise risk response.

[0098] In one embodiment, the self-healing protection module 104 comprises:

[0099] a hierarchical early warning unit configured to control the circuit breaker in the non-critical load circuit in the electrical system based on the electrical fire risk grade and a three-level tripping strategy, wherein,

[0100] when the electrical fire risk grade is low risk, the tripping threshold of the circuit breaker of the corresponding load circuit is adjusted to 90% of the initial tripping threshold, and the temperature change trend information of the corresponding load circuit is recorded and a patrol reminder is sent;

[0101] When the electrical fire risk level is medium risk, adjust the trip threshold of the corresponding load circuit breaker to 80% of the initial trip threshold, and start the preset load transfer control strategy to transfer the load of the corresponding load circuit to the backup circuit;

[0102] When the electrical fire risk level is high risk, control the corresponding load circuit breaker to open, and start the emergency power supply to push the repair work order;

[0103] A closed-loop verification unit is used to detect the insulation resistance and temperature rise information of the corresponding load circuit in real time after opening, and when the insulation resistance of the corresponding load circuit is greater than or equal to the preset resistance threshold and the temperature rise rate is within the preset temperature rise rate threshold, the fault is marked as cleared, and a close control instruction is sent to the corresponding load circuit breaker;

[0104] A linkage control unit is in communication connection with the fire fighting system of the target building, and the linkage control unit is used to control the non-fire fighting power supply to be forcibly cut off and control the emergency lighting system and the smoke exhaust system to be turned on when it is confirmed that an electrical fire has occurred.

[0105] In this embodiment, the hierarchical early warning unit of the self-healing protection module 104 controls the intelligent circuit breaker of the non-critical load circuit in the electrical system according to the electrical fire risk level determined by the edge computing node 103. When the risk level is low risk, the corresponding circuit breaker trip threshold is adjusted to 90% of the initial value, and at the same time, the temperature change trend information of the circuit is recorded and a patrol reminder is sent to the maintenance personnel through the user interaction terminal 105, prompting them to pay attention to the state of the circuit; when the risk level is medium risk, the corresponding circuit breaker trip threshold is adjusted to 80% of the initial value, and at the same time, the load transfer control strategy is started, and through the circuit switching mechanism of the electrical system, the load carried by the circuit is transferred to the pre-set backup circuit, avoiding the risk of upgrading due to the continuous operation of the load in the original circuit; when the risk level is high risk, an opening instruction is sent to the corresponding circuit breaker to control it to open the circuit, and at the same time, the emergency power supply is started to ensure the power supply of the critical areas (such as fire passages and emergency command centers) in the building, and a repair work order is generated through the system background and pushed to the mobile terminal of the maintenance personnel.

[0106] The closed-loop verification unit collects the insulation resistance value and temperature rise information in real time through the insulation resistance sensor and the temperature sensor in the circuit after the circuit breaker is opened, compares the collected data with the preset resistance threshold (such as 100MΩ) and temperature rise rate threshold (such as 0.5℃ / min), and when both indicators meet the requirements, it is determined that the fault has been cleared, and a close control instruction is immediately sent to the circuit breaker to restore the power supply of the circuit.

[0107] The linkage control unit establishes a data connection with the building fire protection system through a communication interface (such as RS485, Ethernet). When the edge computing node 103 confirms the occurrence of an electrical fire by combining temperature field data, harmonic feature data, and visible light images, the linkage control unit sends a control signal to the fire protection system, which on the one hand cuts off the non-fire power supply (such as the power supply loop of office equipment and lighting fixtures), and on the other hand starts the emergency lighting system and the smoke exhaust system, providing protection for personnel evacuation and fire rescue.

[0108] The self-healing protection module 104 of the present embodiment realizes differentiated protection according to the risk level through a three-stage tripping strategy. The threshold adjustment at low risk takes into account both safety and power continuity. Load transfer at medium risk avoids unnecessary power loss. Emergency tripping and emergency power supply at high risk ensure safety and critical function operation in extreme situations. The closed-loop verification mechanism realizes automatic power restoration after fault clearance without the need for manual intervention, improving power restoration efficiency. The linkage control with the fire protection system forms a complete closed loop for electrical fire disposal, improving the overall fire safety response capability of the building.

[0109] In one embodiment, the user interaction terminal 105 displays fault location information, repair guidance information, and spare parts information through an iOS / Android mobile terminal and / or a Web terminal visual interface. The fault location information displays the three-dimensional location of the fault point through the AR (Augmented Reality) superposition technology based on the SLAM (Simultaneous Localization and Mapping) algorithm, and is associated with historical maintenance records. The repair guidance information includes a repair scheme knowledge base based on fault types, which can contain repair video guidance. The spare parts information can include detailed parameters of spare parts such as torque standards (such as M6 screw torque 8-10 N·m) and cable replacement specifications (such as BV-2.5 mm 2 ) and the storage location of the spare parts in the spare parts warehouse.

[0110] The user interaction terminal 105 can also generate a device full life cycle file, and statistics maintenance indicators such as the recurrence rate of similar faults and the average fault-free time (MTBF) of components, and support Excel export.

[0111] Next, two specific embodiments are used to explain the working process of the building electrical fire hazard intelligent monitoring system of the present application in detail:

[0112] Embodiment 1: Office building office area electrical fire hazard monitoring

[0113] Scene description: A 15-story office building in an office area, equipped with 5 power distribution cabinets, carrying office loads such as computers, printers, and air conditioners. The electrical system uses the intelligent monitoring system of the present application.

[0114] Monitoring process:

[0115] Multi-dimensional data acquisition: The non-cooled focal plane array infrared detector of the infrared thermal imaging module 101 scans the busbar row in the No. 3 power distribution cabinet according to the preset trajectory, generates a three-dimensional temperature field cloud map, and finds that there is a local overheating area (temperature 45℃) at the busbar joint through the YOLOv5 model identification; at the same time, the alarm threshold dynamic adjustment unit calculates T alarm = 25 + 0.8 * 30 = 49℃ according to the current environmental temperature 25℃ and the equipment rated temperature rise 30℃, and the current temperature does not exceed the threshold. The current harmonic analysis module 102 collects the current signal of the loop, extracts the 3rd harmonic content 8%, the 5th harmonic content 5%, and the total harmonic distortion rate 12% through Fourier transform, and matches with the harmonic-load correlation database to identify the load abnormality caused by the operation of the printer cluster.

[0116] Multi-modal data fusion analysis: The edge computing node 103 constructs a multi-modal input vector containing the hotspot maximum temperature 45℃, the temperature rise rate 0.2℃ / min, the 3rd and 5th harmonic ratio 1.6, the 5th and 7th harmonic ratio 2.0, and the total harmonic distortion rate 12%, and substitutes it into the random forest model to output a risk score of 25 points, which is determined as low risk, and the fault type is "non-critical load abnormality leading to harmonic distortion".

[0117] Hierarchical self-healing protection control: The self-healing protection module 104 adjusts the circuit breaker tripping threshold of the No. 3 power distribution cabinet from the initial value 100A to 90A, records the temperature change trend, and sends a patrol reminder "No. 2 loop of No. 3 power distribution cabinet in the 15th floor has harmonic distortion due to abnormal load of printers, it is recommended to check the operation status of the printers" to the operation and maintenance personnel through the user interaction terminal 105.

[0118] Fault information visualization interaction: The user interaction terminal 105 displays the fault positioning map, harmonic characteristic curve and patrol guide, and the operation and maintenance personnel turn off some idle printers according to the guide, and the system monitors that the harmonic distortion rate decreases to 5%, the risk score returns to 10 points, and the initial tripping threshold of the circuit breaker is restored.

[0119] Example 2: Old residential building electrical fire hazard monitoring

[0120] Scene description: A resident building built 20 years ago has aging problems in electrical lines, and the system of the application is installed to monitor the unit building power distribution cabinet and main line.

[0121] Monitoring process:

[0122] Multi-dimensional data acquisition: When the infrared thermal imaging module 101 scans the 1 unit power distribution cabinet, it finds that the temperature at the cable joint reaches 65℃, the current environmental temperature is 28℃, and the equipment rated temperature rise is 30℃, and T alarm= 28 + 0.8 x 30 = 52 °C, the temperature exceeds the threshold value, the visible light camera captures the presence of oxidation blackening at the joint. The current harmonic analysis module 102 collects the loop current signal and extracts the total harmonic distortion rate of 25%, the 5th harmonic contains 15%, and identifies it as "poor contact caused by cable joint oxidation" through pattern matching.

[0123] Multimodal data fusion analysis: After the edge computing node 103 constructs the input vector, the model outputs a risk score of 75 points, which is determined as high risk, and the fault type is "cable joint oxidation overheating".

[0124] Hierarchical self-healing protection control: The self-healing protection module 104 immediately controls the circuit breaker to trip, starts the emergency lighting system, and pushes the repair work order to the operation and maintenance personnel terminal. The work order includes fault location, required spare parts (copper wiring terminals) and repair steps. After the operation and maintenance personnel replace the joint, the closed-loop verification unit detects that the insulation resistance is 150MΩ and the temperature rise rate is 0.1°C / min, which meets the threshold requirement, and the circuit breaker is automatically controlled to close to restore power supply.

[0125] Fault information visualization interaction: The user interaction terminal 105 shows the whole process of fault handling to the property management personnel of the residential building, including tripping time, repair progress and power restoration time, improving management transparency.

[0126] To achieve the above purpose, the application also provides a monitoring method of the building electrical fire hazard intelligent monitoring system of any one of the above embodiments, comprising the following steps:

[0127] S1, multi-dimensional data acquisition: temperature field data of the electrical system in the target building are collected by an infrared thermal imaging module and the temperature alarm threshold is dynamically adjusted, and harmonic feature data of the electrical circuit in the target building are collected by a current harmonic analysis module;

[0128] S2, multimodal data fusion analysis: the temperature field data and the harmonic feature data are fused by an edge computing node based on a random forest algorithm to generate an electrical fire risk level and an electrical fault type;

[0129] S3, hierarchical self-healing protection control: the self-healing protection module dynamically adjusts the tripping threshold of the corresponding circuit breaker according to the electrical fire risk level, or triggers the corresponding circuit breaker to trip and triggers the fire linkage control;

[0130] S4, fault information visualization interaction: the fault location, repair guide and spare parts information are displayed through the user interaction terminal.

[0131] In this embodiment, in step S1, the infrared thermal imaging module completes temperature field data acquisition, overheating area identification and alarm threshold adjustment through scanning and algorithm analysis, and the current harmonic analysis module synchronously completes current signal acquisition, harmonic feature extraction and load state identification, realizing parallel acquisition of temperature and harmonic two-dimensional data.

[0132] In step S2, the edge computing node converts the collected temperature field data and harmonic feature data into standardized multi-modal input vectors, substitutes them into the pre-trained random forest model, and outputs risk scores, fault types and risk levels through multi-feature fusion calculation of the model.

[0133] In step S3, the self-healing protection module executes corresponding protection strategies according to the determined risk level, including threshold adjustment, load transfer, split power-off and other operations, and realizes whole-process risk management and control through closed-loop verification and fire-fighting linkage.

[0134] In step S4, the user interaction terminal receives the fault positioning (such as "No. 2 busbar row of No. 3 building underground power distribution cabinet"), repair guide (such as "after closing the power distribution cabinet main switch, replace the oxidized joint and tighten the screw") and spare parts information (such as "required spare part model: copper terminal OT-10A") sent by the edge computing node and the self-healing protection module, and displays them in the form of charts, texts and images.

[0135] The flow design of the detection method of the embodiment forms a complete monitoring closed loop from data acquisition, analysis and evaluation to protection control and information display, ensuring the coherence and systematicness of electrical fire hazard monitoring; multi-dimensional data acquisition provides comprehensive support for subsequent analysis, multi-modal fusion analysis improves the accuracy of risk assessment, hierarchical self-healing protection realizes precise risk disposal, and visual interaction reduces the operation threshold of operation and maintenance personnel, significantly improving the efficiency and reliability of electrical fire hazard monitoring and disposal.

[0136] In one embodiment, step S1 specifically includes:

[0137] S11, using a non-cooled focal plane array infrared detector to scan key electrical components according to a preset trajectory to generate a three-dimensional temperature field cloud map;

[0138] S12, identifying local overheating areas in the three-dimensional temperature field cloud map based on a YOLOv5 defect detection model;

[0139] S13, calculating the adjusted temperature alarm threshold according to the formula T alarm = T env + 0.8 x ΔT rated , wherein T alarmT is the adjusted temperature alarm threshold env T is the ambient temperature, ΔT rated T is the device rated temperature rise;

[0140] S14, when the local temperature exceeds the temperature alarm threshold, a close-up image of the local part is taken by the visible light camera;

[0141] S15, the electrical line current signal is collected, and the fast Fourier transform algorithm is used to extract harmonic feature data in the current signal, wherein the harmonic feature data includes 2-40 harmonic content, total harmonic distortion and inter-harmonic amplitude;

[0142] S16, based on the harmonic-load correlation database, the mode matching algorithm is used to identify the abnormal running state of the load.

[0143] In this embodiment, steps S11-S12 complete temperature field visualization and overheating area positioning through infrared detector scanning and YOLOv5 model identification; step S13 realizes dynamic adaptation of the alarm threshold by formula calculation; the visible light shooting of step S14 provides visual evidence for fault confirmation; steps S15-S16 complete harmonic feature extraction and load state judgment through Fourier transform and pattern matching, and the six sub-steps are sequentially connected to build a complete multi-dimensional data collection logic.

[0144] The multi-dimensional data collection stage of this embodiment ensures the professionalism and accuracy of data collection through the detailed division of work of each sub-step, the combination of three-dimensional temperature field cloud map and YOLOv5 model improves the overheating area recognition accuracy, the dynamic threshold formula solves the environmental interference problem, and the comprehensive extraction of harmonic features and pattern matching realizes the early discovery of load abnormalities, providing a high-quality data basis for subsequent risk assessment.

[0145] In one embodiment, step S2 specifically comprises:

[0146] S21, a multi-modal input vector is constructed, which contains the hottest temperature, temperature rise rate, 3rd and 5th harmonic content ratio, 5th and 7th harmonic content ratio, and total harmonic distortion;

[0147] S22, the multi-modal input vector is input into a pre-trained risk assessment model, and a risk score and an electrical fault type are output, wherein the risk assessment model is trained by a forest random algorithm on historical multi-modal input vectors of the electrical system;

[0148] S23, according to the risk score, the electrical fire risk level is determined according to a preset risk classification determination strategy, wherein the electrical fire risk level includes low risk, medium risk and high risk.

[0149] In this embodiment, step S21 selects the highest temperature of hot spots in the temperature field, the temperature rise rate, and the key ratio parameter in the harmonic characteristic to construct a multi-modal input vector, ensuring that the vector can comprehensively reflect the running state of the electrical system; step S22 inputs the vector into the trained random forest model, and the model outputs a risk score and a fault type through the voting mechanism of multiple decision trees; and step S23 divides the risk level according to a preset score interval, realizing the conversion from data to a risk conclusion.

[0150] In the multi-modal data fusion analysis stage of this embodiment, the input vector constructed by the selected feature parameters has strong representativeness and reduces data redundancy; the multi-decision tree fusion of the random forest model improves the anti-interference ability and accuracy of risk assessment; and the clear risk level division standard makes the evaluation result more practical, providing a clear basis for the execution of protection control strategies.

[0151] In one embodiment, step S3 specifically includes:

[0152] S31, if the electrical fire risk level is low risk, adjusting the tripping threshold of the corresponding circuit breaker to 90% of the initial tripping threshold, recording the temperature change trend and sending a patrol reminder;

[0153] S32, if the electrical fire risk level is medium risk, adjusting the tripping threshold of the corresponding circuit breaker to 80% of the initial tripping threshold, and starting the load transfer control strategy to transfer the load to the standby circuit;

[0154] S33, if the electrical fire risk level is high risk, controlling the corresponding circuit breaker to open and starting the emergency power supply, and pushing a repair work order;

[0155] S34, after opening, real-time detection of the loop insulation resistance and temperature rise information, when the insulation resistance is greater than or equal to the preset resistance threshold and the temperature rise rate is within the preset temperature rise rate threshold, marking the fault clearance and sending a closing control instruction;

[0156] S35, when an electrical fire occurs, controlling the non-firefighting power supply to be forcibly cut off, and starting the emergency lighting and smoke exhaust system.

[0157] In this embodiment, steps S31-S33 execute differentiated circuit breaker control and load management strategies for different risk levels; step S34 realizes automatic closing after fault clearance through real-time monitoring; and step S35 starts firefighting linkage after fire confirmation, and the five sub-steps form a complete control logic from hierarchical protection to emergency disposal.

[0158] The hierarchical control strategy of this embodiment balances safety and power demand, the automatic closing mechanism improves power supply recovery efficiency, the firefighting linkage enhances fire emergency disposal capability, and the coordinated work of the sub-steps realizes the whole life cycle management and control of electrical fire risk.

[0159] The various embodiments described in this specification are presented by way of example, and each embodiment is not necessarily composed of all features described with respect to other embodiments. Each embodiment described in this specification can be implemented in combination with one or more other embodiments described in this specification, and each embodiment can be implemented independently of any other embodiments. Features described as being part of one embodiment can also be implemented as part of another embodiment. The scope of the application is not to be determined by the preferred embodiments described in this specification, but only by the claims that follow, along with the full scope of equivalents to which such claims are entitled.

[0160] Those skilled in the art will further appreciate that the units and algorithms described in the examples presented herein can be implemented in electronic hardware, computer software, or any combination thereof. To clearly illustrate this interchangeability of hardware and software, various components will be described herein generally in terms of their functionality, without reference to the particular manner in which they are implemented. Skilled persons will appreciate that the described functionality can be implemented by one or more computer software programs or general purpose computers programmed with one or more such programs. Such programs can be stored on any computer readable medium, such as random access memory (RAM), read only memory (ROM), magnetic disk, optical disk, nonvolatile memory including flash memory, and the like.

[0161] The steps of a method or algorithm described in connection with the embodiments disclosed herein can be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module can reside in RAM, flash memory, ROM, electrically programmable ROM (EPROM or EEPROM), registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor such that the processor can read information from, and write information to, the storage medium. In the alternative, the storage medium can be integral to the processor. The processor and the storage medium can reside in an application-specific integrated circuit (ASIC).

[0162] The above description of disclosed embodiments is meant to be illustrative of the application and not limiting. Many variations of the described embodiments will be readily apparent to those skilled in the art, and the general principles defined herein can be applied to other embodiments without departing from the spirit or scope of the application. Accordingly, the scope of the application is not to be limited to the scope of the embodiments disclosed herein, but is to be accorded the broadest scope consistent with the principles and novel features disclosed herein.

Claims

1. A building electrical fire hazard intelligent monitoring system, characterized in that, The infrared thermal imaging module is configured to collect temperature field data of an electrical system in a target building and identify overheated areas, and dynamically adjust a temperature alarm threshold according to an ambient temperature and a rated temperature rise of equipment. The current harmonic analysis module is configured to collect harmonic characteristic data of an electrical line in the target building and identify abnormal operating states of loads. The edge computing node is configured to fuse the temperature field data and the harmonic characteristic data based on a random forest algorithm to generate a risk assessment result, wherein the risk assessment result includes an electrical fire risk level and an electrical fault type. The self-healing protection module is configured to dynamically adjust a tripping threshold of a corresponding circuit breaker in the electrical system or trigger the corresponding circuit breaker to open and parallelly control a fire power supply forced switching device according to the electrical fire risk level in the risk assessment result and a preset hierarchical warning strategy, wherein the circuit breaker is an intelligent circuit breaker with an adjustable tripping threshold. The user interaction terminal is configured to visually display fault positioning information, repair guidance information, and spare part information. The infrared thermal imaging module includes:

2. The system as claimed in claim 1, wherein, A non-cooled focal plane array infrared detector disposed on a servo gimbal, which is configured to perform multi-angle scanning on key electrical components of the electrical system according to a preset scanning track to generate a three-dimensional temperature field cloud picture, wherein the key electrical components include busbars and circuit breaker contacts in a power distribution cabinet. An overheated area detection unit configured to identify local overheated areas caused by abnormal states of the electrical system based on a defect detection model trained by a YOLOv5 algorithm using the three-dimensional temperature field cloud picture, wherein the abnormal states of the electrical system include cable joint oxidation and screw loosening. An alarm threshold dynamic adjustment unit configured to dynamically adjust the temperature alarm threshold according to the ambient temperature and the rated temperature rise of the equipment and a preset threshold adjustment formula. A visible light camera disposed on the servo gimbal, which is configured to capture close-up images of local areas of the electrical system when the local temperature exceeds the temperature alarm threshold and upload the images to the edge computing node. The threshold adjustment formula is as follows:

3. The system as claimed in claim 2, wherein, The current harmonic analysis module includes: T alarm = T env + 0.8 x ΔT rated where T alarm is the adjusted temperature alarm threshold, T env is the ambient temperature, ΔT rated is the device rated temperature rise.

4. The system as claimed in claim 1, wherein, A current signal acquisition unit configured to acquire current signals in the electrical line. A harmonic feature extraction unit configured to extract harmonic characteristic data from the current signals using a fast Fourier transform algorithm, wherein the harmonic characteristic data includes 2-40 harmonic content rates, total harmonic distortion rates, and inter-harmonic amplitudes. A load state identification unit configured to establish a harmonic-load correlation database and identify abnormal operating states of loads in the electrical system based on the harmonic characteristic data and the harmonic-load correlation database through a pattern matching algorithm. The edge computing node includes:

5. The system as claimed in claim 1, wherein, An input vector construction unit configured to construct a multi-modal input vector based on the temperature field data and the harmonic characteristic data, wherein the multi-modal input vector contains a hotspot maximum temperature, a temperature rise rate, a ratio of 3rd harmonic content to 5th harmonic content, a ratio of 5th harmonic content to 7th harmonic content, and a total harmonic distortion rate. ​ a risk score unit configured to perform risk assessment on the electrical system based on the multi-modal input vector by a pre-trained risk assessment model, and output a risk score and an electrical fault type, wherein the risk assessment model is trained by a forest random algorithm on historical multi-modal input vectors of the electrical system; a risk level determination unit configured to determine the electrical fire risk level according to the risk score and a preset risk grading determination strategy, wherein the electrical fire risk level includes low risk, medium risk and high risk.

6. The system as claimed in claim 5, wherein, The self-healing protection module comprises: a hierarchical early warning unit configured to control circuit breakers in non-critical load circuits in the electrical system based on the electrical fire risk level by a three-level tripping strategy, wherein, when the electrical fire risk level is low risk, the tripping threshold of the circuit breaker of the corresponding load circuit is adjusted to 90% of the initial tripping threshold, while recording the temperature change trend information of the corresponding load circuit and sending an inspection reminder; when the electrical fire risk level is medium risk, the tripping threshold of the circuit breaker of the corresponding load circuit is adjusted to 80% of the initial tripping threshold, and a preset load transfer control strategy is started to transfer the load of the corresponding load circuit to a standby circuit; when the electrical fire risk level is high risk, the circuit breaker of the corresponding load circuit is controlled to be tripped, and an emergency power supply is started, and a repair work order is pushed; a closed-loop verification unit configured to detect the insulation resistance and temperature rise information of the corresponding load circuit in real time after tripping, and when the insulation resistance of the corresponding load circuit is greater than or equal to a preset resistance threshold and the temperature rise rate is within a preset temperature rise rate threshold, the fault is marked as cleared, and a close command is sent to the circuit breaker of the corresponding load circuit; a linkage control unit in communication connection with a fire fighting system of the target building, and configured to control non-fire fighting power supply to be forcibly cut off and control emergency lighting system and smoke exhaust system to be turned on when it is confirmed that an electrical fire occurs.

7. A monitoring method of the intelligent monitoring system for electrical fire hazards in buildings according to any one of claims 1 to 6, characterized in that, The method comprises the following steps: S1, multi-dimensional data acquisition: temperature field data of the electrical system in the target building are acquired by an infrared thermal imaging module, and a temperature alarm threshold is dynamically adjusted, and harmonic feature data of the electrical circuit in the target building are acquired by a current harmonic analysis module; S2, multi-modal data fusion analysis: the temperature field data and the harmonic feature data are fused by an edge computing node based on a random forest algorithm to generate an electrical fire risk level and an electrical fault type; S3, hierarchical self-healing protection control: the self-healing protection module dynamically adjusts the tripping threshold of the corresponding circuit breaker according to the electrical fire risk level, or triggers the corresponding circuit breaker to be tripped and the fire fighting linkage control; S4, fault information visualization interaction: fault positioning, repair guidance and spare part information are displayed on a user interaction terminal.

8. The method for intelligent monitoring of electrical fire hazards in buildings as claimed in claim 7 wherein, Step S1 specifically comprises: S11, a non-cooled focal plane array infrared detector is used to scan key electrical components according to a preset trajectory to generate a three-dimensional temperature field cloud map; S12, a local overheating area in the three-dimensional temperature field cloud map is identified based on a YOLOv5 defect detection model; S13, the adjusted temperature alarm threshold T alarm = T env + 0.8 x ΔT rated is calculated, where T alarm is the adjusted temperature alarm threshold, T env is the ambient temperature, and ΔT rated is the device rated temperature rise; S14, when the local temperature exceeds the temperature alarm threshold, a close-up image of the local part is taken by a visible light camera; S15, an electrical line current signal is collected, and a fast Fourier transform algorithm is used to extract harmonic feature data in the current signal, wherein the harmonic feature data includes 2-40 harmonic content, total harmonic distortion, and inter-harmonic amplitude; S16, based on a harmonic-load correlation database, a mode matching algorithm is used to identify the abnormal running state of the load.

9. The method for intelligent monitoring of electrical fire hazards in buildings as claimed in claim 7 wherein, Step S2 specifically includes: S21, constructing a multi-modal input vector, the multi-modal input vector including the hottest temperature, the temperature rise rate, the ratio of 3rd and 5th harmonic content, the ratio of 5th and 7th harmonic content, and the total harmonic distortion; S22, inputting the multi-modal input vector into a pre-trained risk assessment model to output a risk score and an electrical fault type, the risk assessment model being trained by a forest random algorithm on historical multi-modal input vectors of electrical systems; S23, determining the electrical fire risk level according to the risk score and a pre-set risk classification determination strategy, wherein the electrical fire risk level includes low risk, medium risk, and high risk.

10. The method for intelligent monitoring of electrical fire hazards in buildings as claimed in claim 7 wherein, Step S3 specifically includes: S31, if the electrical fire risk level is low risk, adjusting the tripping threshold of the corresponding circuit breaker to 90% of the initial tripping threshold, recording the temperature change trend and sending an inspection reminder; S32, if the electrical fire risk level is medium risk, adjusting the tripping threshold of the corresponding circuit breaker to 80% of the initial tripping threshold, and starting a load transfer control strategy to transfer the load to a backup circuit; S33, if the electrical fire risk level is high risk, controlling the corresponding circuit breaker to open and starting an emergency power supply, and pushing a repair work order; S34, after opening, real-time detecting the circuit insulation resistance and temperature rise information, when the insulation resistance is greater than or equal to a pre-set resistance threshold and the temperature rise rate is within a pre-set temperature rise rate threshold, marking the fault as cleared and sending a close control instruction; S35, when an electrical fire occurs, controlling the non-firefighting power supply to be forcibly cut off, starting the emergency lighting and smoke exhaust system.

Images