A multi-chamber isolated protective switchgear structure

Abstract

The application relates to the technical field of electrical switch cabinets, in particular to a multi-cavity isolated protective switch cabinet structure. The technical scheme is as follows: a multi-cavity isolated protective switch cabinet structure comprises a cabinet body, a pressure relief chamber, a gas exhaust valve and dry air or nitrogen or sulfur hexafluoride in the closed space of the cabinet body. The closed space is the main wiring chamber of the switch cabinet, and the lower end of the gas exhaust valve is connected with the pressure relief chamber. The main wiring chamber, the pressure relief chamber and the switch operating panel are completely isolated. The switch position is completely isolated from the internal main wiring chamber, and the pressure relief chamber is added. When the pressure in the main wiring chamber rises due to a fault, timely pressure relief and buffering can be realized, the diffusion of toxic gas to the outside is avoided, and the safety of the electrical switch cabinet is greatly improved.

Description

Technical Field

[0001] This invention relates to the field of electrical switchgear technology, and in particular to a protective switchgear structure with multi-chamber isolation. Background Technology

[0002] A switchgear is a type of electrical equipment. External power lines first enter the main control switch inside the cabinet, and then enter the branch control switches. Each branch circuit is configured according to its needs. Examples include instruments, automatic control systems, motor magnetic switches, various AC contactors, etc. Some switchgear also have high-voltage and low-voltage compartments, and are equipped with high-voltage busbars, such as in power plants. Some also have underfrequency load shearing features to protect the main equipment.

[0003] The main function of switchgear is to open, close, control, and protect electrical equipment during the power generation, transmission, distribution, and energy conversion processes of a power system. The main components within switchgear include circuit breakers, disconnect switches, load switches, operating mechanisms, instrument transformers, and various protective devices.

[0004] Existing electrical switchgear commonly uses sulfur hexafluoride (SF6) as the insulating medium, which has a high greenhouse effect potential, making it unfavorable to environmental protection requirements. Furthermore, when a fault occurs in the electrical switchgear, an internal arc fault can occur, causing SF6 to decompose into other harmful gases. Since the various chambers are not completely isolated, this gas can diffuse and pose a toxic risk, resulting in poor safety during use. Summary of the Invention

[0005] This invention proposes a multi-chamber isolated protective switchgear structure, which solves the problem of toxicity risk and poor safety in the prior art when electrical switchgear malfunctions due to the diffusion of toxic gases.

[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A multi-chamber isolated protective switchgear structure includes a cabinet and a pressure relief chamber. The pressure relief chamber is installed at the bottom rear side of the cabinet. A vent valve is installed at the top of the pressure relief chamber. The upper opening of the vent valve connects to the enclosed space of the cabinet. The enclosed space is the main wiring compartment of the switchgear. The cabinet door opens to reveal the switch operation panel. The lower opening of the vent valve connects to the pressure relief chamber. The enclosed space of the cabinet is filled with dry air, nitrogen, or sulfur hexafluoride. The main wiring compartment, the pressure relief chamber, and the switch operation panel are completely isolated. The vent valve has an embedded pressure sensor to monitor the dynamic pressure difference between the enclosed space of the cabinet and the vent chamber.

[0007] The switchgear uses a fault response method, including the following steps: Step S1: Install a temperature sensor array around the busbar connection points, circuit breaker contacts and vent valve inside the cabinet to generate a temperature distribution heat map at a frequency of 1 time / second. Install a wind speed sensor and an ambient temperature and humidity sensor outside the cabinet, with a sampling frequency of 1 time / second. Step S2: Install a humidity sensor on the inner wall of the pressure relief chamber to collect humidity change curves in real time, with a sampling frequency of 10 times / second; Step S3: Obtain the pressure difference fluctuation amplitude between the enclosed space of the cabinet and the pressure relief chamber through the air pressure sensor. ; Step S4: Extract the direction of movement of the region with the largest gradient in the temperature distribution thermogram, and calculate the rate of change of its horizontal distance from the vent valve, i.e., the temperature gradient rate. The unit is cm / s; Step S5: Calculate the second derivative of the humidity change curve to obtain the humidity change intensity. ; Step S6, , , Normalize to the 0-1 range to generate multidimensional feature vectors. ; Step S7, when ≥0.7、 ≥0.5 or When the value is ≥0.6, a graded response mechanism is triggered.

[0008] Furthermore, the fault level classification of the graded response mechanism in step S7 includes the following steps: Level 1 fault diagnosis: When When the value is ≥0.9, high-temperature area movement tracking is activated. If the high-temperature area moves towards the vent valve at a speed of ≥5cm / s, and the moving distance is reduced by more than 30% within 10 seconds, humidity change detection is activated. When the second derivative value of the humidity change curve is ≥0.9, humidity change detection is activated. If three consecutive samples are positive, and the vent valve fails to reach a 90° opening degree within one second after receiving the opening command, it is determined to be a level one fault. Level 2 fault determination: When A level 2 fault is defined as follows: a sawtooth-shaped fluctuation with an amplitude ≥ 0.2 occurs within a period of 2-5 seconds, and the differential pressure fluctuation exceeds ±25 kPa for 1 minute; or the absolute value of humidity exceeds 30% of the ambient humidity but the second derivative is negative. Level 3 fault determination: When When a single characteristic value exceeds the threshold, a 15-minute deterioration trend monitoring is initiated, and the standard deviation of temperature, humidity, and pressure difference fluctuations is calculated. If the standard deviation decrease rate is ≤5% / minute and there are no associated sensor abnormalities, it is judged as a level three fault.

[0009] Furthermore, after fault diagnosis, the following operations are performed respectively: Level 1 Response Operation Chain: The vent valve is forced to fully open to 90°, the cyclone separator at the bottom of the pressure relief chamber is activated to discharge high-temperature particulate matter, a high-frequency carrier signal is sent to the smart meter, and 80% of non-critical loads are cut off; Level 2 response operation chain: Control the vent valve to swing back and forth between 0° and 45° with a period of 2 seconds, simultaneously open the phase change material heat dissipation chamber on the side wall of the cabinet, generate a work order containing temperature hot spot coordinates and suggested maintenance time on the local HMI interface, and broadcast the warning information through the voice synthesis module. Collect the humidity data of the pressure relief chamber every 5 minutes. If it does not decrease for 3 consecutive times, it will be upgraded to Level 1 response. Level 3 response operation chain: Keep the vent valve closed, open the auxiliary ventilation window on the top of the cabinet to 60° opening, and limit the load current to 85% of the rated value through the PWM voltage regulation module. Generate an equipment status report every 5 minutes, including temperature peak, humidity trend and load rate, and send it to the inspection terminal within 500 meters through the LORA wireless network. If no manual confirmation is received within 48 hours, it will automatically upgrade to Level 2 response.

[0010] Furthermore, it also includes response optimization processes: When the temperature gradient rate dT / dt ≥ 8℃ / s, the vent valve remains fully open until the temperature drops below 60℃. The pressure relief efficiency η is calculated as (actual pressure relief amount / theoretical maximum pressure relief amount) × 100%. The theoretical maximum pressure relief amount is obtained by referring to the table based on the vent valve opening-flow curve. When η is below 70% for 3 consecutive seconds, the system switches to the backup pressure relief channel and triggers an audible and visual alarm. The temperature distribution heat map of adjacent switchgear is obtained through power line carrier communication to identify the direction of the high temperature area. If the high temperature area of ​​the adjacent cabinet extends toward this cabinet and the frequency of its vent valve operation increases by ≥2 times within 5 minutes after the fault, the load current limit threshold of this cabinet is dynamically reduced from 85% to 60%, and a pre-depressurization command is sent to the adjacent cabinet to open its vent valve to 30° opening in advance. Based on the fault data of this cabinet over the past 30 days, when the false alarm rate of the first-level fault is greater than 15%, the temperature gradient threshold will be increased from 15℃ / cm to 18℃ / cm. If the second-level response delay is greater than 3 seconds, the pulse cycle of the vent valve will be shortened from 2 seconds to 1.5 seconds, and the reference value of the humidity sensor will be recalibrated. Calculate the ratio of the number of times the vent valve actuates to the number of times it is designed to last. Percentage of worn area of ​​the inner wall coating of the pressure relief chamber Variance of data differences between adjacent sensors Health Index ,when At 0.6, the inspection cycle is shortened to 7 days, and the vent valve seal is forcibly replaced. When the value is ≥0.6, the carbon deposits in the pressure relief chamber should be purged every 24 hours.

[0011] Furthermore, it also includes the fault tracing and iteration process: associating external meteorological data such as temperature, humidity, and wind speed with the real-time load rate of the power grid at the time of the fault, and recording the opening angle, duration, and associated temperature peak of each action of the vent valve; The feature matrix of historical failure cases is extracted using a twin neural network, including temperature gradient distribution, humidity change pattern and pressure difference fluctuation spectrum. The cosine similarity between the current failure and historical cases is calculated, and a treatment plan with a similarity of ≥90% is output. Model optimization involves adding an "inefficient handling" label to fault cases that recur after initial treatment and reducing their weighting by 40%. The neural network is retrained quarterly to update the weighting coefficients of temperature and humidity features. , ; Feedback on the handling results and statistics on the time to resolve the fault. Total pressure relief The ratio of these values ​​is used to generate the objective function for optimization. ,when When the value is greater than 1.2 times the baseline value, the vent valve control strategy version iteration is triggered.

[0012] Furthermore, it also includes virtual simulation and strategy verification processes: importing the three-dimensional model of the switchgear and historical sensor data into the digital twin model to construct a high-precision simulation scenario, and injecting simulated faults such as vent valve jamming opening degree ≤30°, temperature sensor failure deviation ≥10℃, and communication delay ≥500ms; Response performance evaluation, with the following evaluation indicators defined: pressure relief efficiency η≥70%, load shedding accuracy α≥90%, and response delay Δt≤2s; Among them, pressure relief efficiency Load shearing accuracy ×100%, Response Delay Generate a three-dimensional evaluation matrix containing η, α, and Δt, and label the items that do not meet the standards; If any indicator in the evaluation matrix falls below the baseline value by 20% twice consecutively, the parameter calibration procedure is triggered, and the optimized pressure relief angle and pulse frequency are written into the firmware and wirelessly updated to the field equipment via OTA.

[0013] Furthermore, it also includes an adaptive control process for the pressure relief process: the vent valve is equipped with a mass flow meter to monitor the pressure relief flow in real time. In stage 1, 0-5 seconds, the vent valve is fully opened to 90° to prioritize the discharge of high-temperature free metal particles. In stage 2, 5-15 seconds, the opening is controlled to 60% to maintain a constant pressure relief flow of 20m³ / min ±5%. In stage 3, after 15 seconds, the opening is reduced by 10° every 5 seconds, and the pressure decay rate inside the cabinet is monitored synchronously. If the rate is <5kPa / s, the adjustment is paused. Acquire real-time airflow velocity monitoring data in the pressure relief chamber and dynamically select the flow guidance mode based on the velocity value: when the airflow velocity is ≥10m / s, adjust the flow guide plate to the maximum flow guidance angle and prioritize expanding the cross-sectional area of ​​the pressure relief channel to reduce pressure loss; when the airflow velocity is ≤5m / s and <10m / s, periodically swing the flow guide plate angle ±15°; when the airflow velocity is <5m / s, lock the flow guide plate to the minimum angle and start the backup high-pressure air pump for forced pressurization. After each depressurization is completed, the inner wall of the depressurization chamber is scanned with a laser rangefinder to detect local depressions (≥0.5mm) and mark their coordinates. If a depression is detected in the same location three times in a row, a structural reinforcement command is triggered to deploy carbon fiber reinforcement sheets.

[0014] Furthermore, it also includes 3D visualization and trusted evidence preservation processes: Temperature sensor data is mapped onto the cabinet's 3D model to generate a color temperature gradient heat map, where blue represents <50℃ and red represents ≥100℃. Pressure relief airflow vector arrows are superimposed to dynamically display airflow speed and direction, with arrow length indicating airflow speed and arrow direction indicating direction. It allows maintenance personnel to manually adjust the feature threshold within ±10% range or force switch the fault level in the HMI interface, add digital signatures to manual operations, and bind them to the operation time and employee number and store them in the security database; Key operation records, fault determination, vent valve action, and manual intervention are generated to generate SHA-256 hash values ​​and written into the distributed ledger of the consortium blockchain. Each block contains a snapshot of the system state from the preceding 10 minutes.

[0015] Furthermore, it also includes a self-healing and predictive maintenance process: a laser displacement sensor is installed on the top of the pressure relief chamber. The laser displacement sensor scans the inner wall of the pressure relief chamber to detect permanent deformation of ≥1mm. A shape memory alloy bracket is deployed in the deformation area, and it is heated to 80℃ to expand it, fill the gaps and restore more than 90% of the mechanical strength. Construct a temperature-humidity-pressure correlation matrix and use an LSTM network to predict the probability of failure in the next 24 hours. ,when When the leakage rate is ≥30%, pressurize to 110% of the rated value. If the leakage rate is ≤0.5% / min, it is considered qualified. Inject ultrasonic cleaning fluid at a frequency of 28kHz and run for 5 minutes. Based on the Health Index (HI), spare parts replacement suggestions are generated. When HI < 0.6, the system automatically orders vent valve seals and pushes replacement tutorial videos. When HI ≥ 0.6, high-risk cabinets are marked on the GIS map, and the optimal inspection route is planned.

[0016] The positive effects of this invention are as follows: by replacing sulfur hexafluoride in the prior art with dry air or nitrogen, the cabinet is filled with harmless dry air or nitrogen. In the event of a malfunction, if the pressure inside the cabinet exceeds the limit, the gas is discharged from the vent valve and diffuses to the outside without causing pollution, thus greatly improving the safety of the electrical switch cabinet.

[0017] After the high-pressure gas is ejected from the vent valve, it is placed in the pressure relief chamber. The high-pressure gas will not be directly ejected to the outside. The impact force of the high-pressure gas is released through the pressure relief chamber, avoiding impact damage to external personnel and the surrounding environment.

[0018] For certain scenarios with limited production conditions, sulfur hexafluoride can continue to be used. In the event of a malfunction, if the pressure inside the cabinet exceeds the limit, gas will be discharged from the vent valve. Since the front switch operation panel is completely isolated from the main wiring compartment, the gas will not diffuse to the front personnel's position, achieving the protection effect of the first event of a malfunction and greatly improving the safety of the electrical switch cabinet.

[0019] In the event of a malfunction, the high-pressure gas, after being ejected from the vent valve, remains in the pressure relief chamber, preventing it from directly escaping to the outside and thus avoiding the spread of toxic gases. Simultaneously, the switchgear will issue a warning, prompting relevant personnel to quickly evacuate the space where the switchgear is located. The front switch operating panel is completely isolated from the main wiring compartment, preventing the gas from spreading to the personnel in front of it, thus providing them with ample time to leave and greatly improving the safety of the electrical switchgear. In addition, the pressure relief chamber releases the impact force of the high-pressure gas, preventing impact damage to personnel outside and the surrounding environment. Attached Figure Description

[0020] Fig. 1 This is a schematic diagram of the first structure of the electrical switchgear in this invention; Fig. 2 This is a schematic diagram of the second structure of the electrical switchgear in this invention; Fig. 3 This is a schematic diagram of the third structure of the electrical switchgear in this invention; In the diagram: 1. Cabinet; 2. Pressure relief chamber; 3. Air relief valve. Detailed Implementation

[0021] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention. Example 1

[0022] Combination Figs. 1-3 As shown, a multi-chamber isolated protective switchgear structure includes a cabinet 1 and a pressure relief chamber 2. The pressure relief chamber 2 is installed at the rear bottom of the cabinet 1. A vent valve 3 is installed on the top of the pressure relief chamber 2. The upper opening of the vent valve 3 connects to the enclosed space of the cabinet 1. The enclosed space is the main wiring compartment of the switchgear. The cabinet door of the cabinet 1 opens to reveal the switch operation panel. The lower opening of the vent valve 3 connects to the pressure relief chamber 2. The enclosed space of the cabinet 1 is filled with dry air, nitrogen, or sulfur hexafluoride. The main wiring compartment, the pressure relief chamber 2, and the switch operation panel are completely isolated. The vent valve 3 is embedded with a pressure sensor to monitor the dynamic pressure difference between the enclosed space of the cabinet 1 and the pressure relief chamber 2.

[0023] In this embodiment, all wiring and other devices are located in the main wiring room, completely isolated from the switch operation panel on the front side of the cabinet 1. At the same time, the cabinet 1 is filled with harmless dry air or nitrogen. Dry air or nitrogen can be used to replace sulfur hexafluoride in the prior art, achieving greater safety in use.

[0024] In addition, for certain scenarios with limited production conditions, sulfur hexafluoride can continue to be used. In the event of a malfunction, if the pressure inside cabinet 1 exceeds the standard, gas will be discharged from the vent valve 3. Since the front switch operation panel is completely isolated from the main wiring compartment, the gas will not diffuse to the front personnel's position, achieving the protection effect of the first event of a malfunction and greatly improving the safety of the electrical switch cabinet.

[0025] The pressure relief chamber 2 is fixed to the bottom rear side of the cabinet 1. It is made of high-strength steel plate and welded together. The inner wall is coated with a high-temperature resistant ceramic coating, such as Al2O3 ceramic, with a thickness of ≥2mm. A guide channel is set at the bottom to guide the high-temperature gas to diffuse downward. The rear pressure relief chamber structure guides the high-temperature gas to the ground to avoid horizontal spraying and damage to surrounding equipment.

[0026] The vent valve 3 is a butterfly valve with a nickel-based alloy body that can withstand temperatures ≥800℃. It connects the top of the cabinet to the pressure relief chamber via a flange. The valve opening is driven by a stepper motor with a control accuracy of ±0.5°.

[0027] With the vent valve closed, an IP67-level seal is achieved using a fluororubber O-ring (Shore hardness 75±5). An explosion-proof grille (aperture ≤5mm) is installed at the bottom of the pressure relief chamber to prevent backflow of foreign objects.

[0028] Furthermore, in the event of a malfunction, the high-pressure gas ejected from the vent valve 3 is contained within the pressure relief chamber 2, preventing it from directly escaping to the outside and thus avoiding the spread of toxic gases. Simultaneously, the switchgear will issue a warning, prompting relevant personnel to quickly evacuate the switchgear area. The front switch operating panel is completely isolated from the main wiring compartment, preventing the gas from spreading to the personnel's location and providing them with ample time to leave. This significantly improves the safety of the electrical switchgear. Additionally, the pressure relief chamber 2 releases the impact force of the high-pressure gas, preventing any impact damage to external personnel and the surrounding environment. Example 2

[0029] Based on Example 1, the fault response method, applied to the switchgear in Example 1, includes the following steps: Step S1: Install a temperature sensor array around the busbar connection points, circuit breaker contacts and vent valve 3 inside cabinet 1 to generate a temperature distribution heat map at a frequency of 1 time / second. Install a wind speed sensor and an ambient temperature and humidity sensor outside cabinet 1, with a sampling frequency of 1 time / second. Temperature sensor arrays are installed around the busbar connection points, circuit breaker contacts, and vent valves inside the cabinet to form a dense monitoring network.

[0030] An air speed sensor and an ambient temperature and humidity sensor are installed on the outside of the cabinet to compensate for the influence of environmental factors on the internal temperature. Temperature data is collected once per second, and the discrete measurement points are expanded into a continuous two-dimensional temperature distribution map, i.e., a heat map, through a spatial interpolation algorithm.

[0031] The heat map uses color gradients to reflect temperature levels. High-temperature areas (such as arc fault points) are displayed in red, while low-temperature areas are displayed in blue. The heat map dynamically tracks the gradient changes in the temperature field, providing spatial information for subsequent fault location.

[0032] Step S2: Install a humidity sensor on the inner wall of the pressure relief chamber 2 to collect humidity change curves in real time, with a sampling frequency of 10 times / second; A high-sensitivity humidity sensor is installed on the inner wall of the pressure relief chamber to collect humidity data at a frequency of 10Hz, capture humidity abrupt changes caused by leakage of insulating medium or release of vapor, filter the raw humidity data to eliminate noise interference, and generate a smooth humidity change curve.

[0033] Step S3: Obtain the pressure difference fluctuation amplitude between the enclosed space of cabinet 1 and pressure relief chamber 2 using the air pressure sensor. ; The pressure sensor embedded in the vent valve measures the pressure difference ΔP between the enclosed space of the cabinet and the vent chamber in real time, and calculates the standard deviation of ΔP within a 10-second time window as the pressure difference fluctuation amplitude. This reflects the severity of sudden changes in air pressure.

[0034] Step S4: Extract the direction of movement of the region with the largest gradient in the temperature distribution thermogram, and calculate the rate of change of its horizontal distance from the vent valve 3, i.e., the temperature gradient rate. The unit is cm / s; By comparing two adjacent heatmap frames, the region with the largest temperature gradient (such as the path of an electric arc) is identified through an edge detection algorithm.

[0035] The rate of movement is calculated by tracking the change in the position of the centroid of the region and calculating the horizontal distance (in cm / s) it moves per second towards the vent valve. Example: If the centroid moves 15 cm in 1 second, then the temperature gradient rate... =15cm / s.

[0036] Step S5: Calculate the second derivative of the humidity change curve to obtain the humidity change intensity. ; The second derivative of the humidity change curve is calculated to quantify the acceleration of humidity change. The larger the value of the second derivative, the more drastic the change in humidity.

[0037] Step S6, , , Normalize to the 0-1 range to generate multidimensional feature vectors. ; Temperature gradient rate Based on the historical maximum gradient rate (e.g., 20 cm / s), it is linearly mapped to the 0-1 range.

[0038] Intensity of humidity change Normalization after logarithmic transformation eliminates the influence of skewed data distribution.

[0039] Pressure fluctuation Z-score standardization is performed based on the mean and standard deviation, and then truncated to the range of 0-1.

[0040] Vector composition: combining the three normalized eigenvalues ​​into a composite eigenvector. It is used to comprehensively determine the fault level.

[0041] Step S7, when ≥0.7、 ≥0.5 or When the value is ≥0.6, a graded response mechanism is triggered.

[0042] Threshold determination logic: Level 1 fault: When ≥0.7 (temperature gradient rate ≥14cm / s) and ≥0.5 (drastic change in humidity), or When the differential pressure is ≥0.6 (severe pressure fluctuation), the highest priority response is triggered, such as forced pressure relief or cutting off non-critical loads.

[0043] Level 2 fault: When a single characteristic value exceeds the threshold but does not meet the Level 1 condition, dynamic balance control is activated, such as the periodic swing of the vent valve.

[0044] Level 3 fault: When the characteristic value is close to the threshold but there is no trend of deterioration, preventive measures (such as flow restriction, auxiliary ventilation) are implemented.

[0045] Dynamic adaptation mechanism: The threshold is automatically adjusted according to the ambient temperature and humidity to avoid false triggering in high-temperature environments.

[0046] This method achieves precise, tiered response to internal faults in electrical switchgear through multi-dimensional sensor data fusion and dynamic feature analysis. Specifically, it includes the following steps: Multi-source data acquisition: Real-time monitoring of the environment inside the cabinet through temperature, humidity, and air pressure sensors; Dynamic feature extraction: Extracting key fault features from temperature distribution, humidity abrupt changes, and pressure differential fluctuations; Normalization and fusion: Standardizing multiple feature values ​​to generate a comprehensive feature vector; Tiered response triggering: Differentiated control strategies are triggered based on a combination of feature vector thresholds.

[0047] By integrating multi-dimensional data and analyzing dynamic features, the accuracy, response speed, and reliability of electrical switchgear fault detection have been significantly improved. Through dynamic correlation analysis of temperature gradient rate, humidity change intensity, and differential pressure fluctuation, different types of faults such as arc faults, insulation leakage, and mechanical jamming can be accurately distinguished.

[0048] Differentiated control strategies are matched for different fault levels, such as forced pressure relief for level 1 faults, dynamic balancing for level 2 faults, and current limiting for level 3 faults. This avoids power outages caused by "over-response" and prevents chain reactions caused by "insufficient response".

[0049] Frequent acquisition and real-time synchronization of temperature thermograms, humidity curves, and differential pressure data ensure the system's ability to detect transient faults (such as flashover and partial discharge).

[0050] By dynamically adjusting internal monitoring thresholds using external wind speed and ambient temperature and humidity data, for example, the temperature gradient rate threshold is automatically relaxed in hot weather to avoid false triggering caused by environmental heat radiation, thus greatly reducing the false alarm rate.

[0051] Physical quantities of different dimensions (such as temperature gradient cm / s, humidity change intensity, and pressure difference kPa) are uniformly mapped to the 0-1 range to eliminate the influence of occasional errors from a single sensor; for characteristic values ​​close to the threshold (such as... =0.65) Initiate 15-minute deterioration trend monitoring, and combine fluctuation standard deviation analysis to effectively distinguish between instantaneous interference and real faults, and avoid frequent false actions. Example 3

[0052] Based on Example 2, the fault level classification of the graded response mechanism in step S7 includes the following steps: Level 1 fault diagnosis: When When the value is ≥0.9, high-temperature area movement tracking is activated. If the high-temperature area moves towards the vent valve 3 at a speed of ≥5cm / s, and the moving distance is reduced by more than 30% within 10 seconds, humidity change detection is activated. When the second derivative value of the humidity change curve is ≥0.9, humidity change detection is activated. If three consecutive samples are positive, and the vent valve 3 fails to reach a 90° opening degree within 1 second after receiving the opening command, it is determined to be a level one fault. Level 2 fault determination: When A level 2 fault is defined as follows: a sawtooth-shaped fluctuation with an amplitude ≥ 0.2 occurs within a period of 2-5 seconds, and the differential pressure fluctuation exceeds ±25 kPa for 1 minute; or the absolute value of humidity exceeds 30% of the ambient humidity but the second derivative is negative. Level 3 fault determination: When When a single characteristic value exceeds the threshold, a 15-minute deterioration trend monitoring is initiated, and the standard deviation of temperature, humidity, and pressure difference fluctuations is calculated. If the standard deviation decrease rate is ≤5% / minute and there are no associated sensor abnormalities, it is judged as a level three fault.

[0053] The system achieves refined fault determination for Level 3 faults through dynamic tracking, mutation verification, and trend prediction, specifically including: Level 1 Fault (Emergency Pressure Relief Stage): Rapid movement of high-temperature areas and sudden changes in humidity are verified in conjunction with abnormal operation detection of the pressure relief valve; Level 2 Fault (Pressure Imbalance): Correlation Analysis of Periodic Fluctuations with Pressure Difference / Humidity; Level 3 fault (early warning observation level): long-term trend monitoring and interference elimination mechanism.

[0054] The specific implementation method is as follows: Level 1 Fault Determination: Based on frame-by-frame analysis of the temperature distribution heatmap, the movement direction of the high-temperature core region is determined using a centroid tracking algorithm. If the movement direction points towards vent valve 3, it is determined to be a risk of arc propagation.

[0055] The system calculates the change in horizontal distance between the center of mass of the high-temperature area and the vent valve every second. If the moving speed is ≥5cm / s and the total moving distance is reduced by ≥30% within 10 consecutive seconds, for example, if the initial distance of 150cm is reduced to 105cm after 10 seconds, a first-level fault warning is triggered.

[0056] Humidity abrupt change verification was performed by real-time second-order derivative analysis of the humidity curve in the pressure relief chamber, using the following formula: When 3 consecutive sampling values At this time, it indicates that the humidity is rising at an accelerated rate, such as when insulating oil leaks.

[0057] After receiving the opening command, the pressure relief valve uses a Hall sensor to provide real-time feedback on the valve shaft rotation angle. If it does not reach the 90° fully open state within 1 second, it is determined to be mechanical jamming or drive failure, and a level one fault is confirmed.

[0058] A level one fault is only identified when high temperature movement, sudden humidity change, and mechanical abnormality occur simultaneously, with a false alarm rate of less than 1%.

[0059] Rapid response: The entire process from detection to response triggering takes ≤2 seconds to avoid the accumulation of arc energy and the resulting deflagration.

[0060] Level 2 fault diagnosis: Sawtooth fluctuation identification, based on temperature gradient rate. Perform wavelet transform (db4 fundamental) to extract the amplitude component in the 2-5Hz frequency band. If the amplitude is ≥0.2 within a 2-second period (corresponding to intermittent temperature rise caused by partial discharge or poor contact), it is marked as abnormal fluctuation.

[0061] Pressure difference continuously exceeds the standard: Calculate the standard deviation of the pressure difference fluctuation between the cabinet and the pressure relief chamber within 1 minute. If the standard deviation is ≥25kPa and lasts for 60 seconds, it is determined to be a sealing failure or airflow blockage.

[0062] When the humidity in the pressure relief chamber exceeds 30% of the ambient humidity, further analysis of its second derivative is needed. A value <0 (humidity rises slowly) indicates that the leak is a slow release, such as condensation buildup, rather than a sudden liquid leak.

[0063] Differentiate between rapid leakage (Level 1 fault) and slow leakage (Level 2 fault) to guide differentiated handling; Dynamic balancing: The pressure is balanced by periodically swinging the pressure relief valve, such as by 45° opening pulse control, to avoid excessive pressure relief leading to negative pressure inside the cabinet.

[0064] Level 3 fault determination: For a single characteristic value exceeding the threshold, continuous monitoring for 15 minutes is performed, and its standard deviation of fluctuation is calculated. ,in It is a sequence of eigenvalues. The mean, This represents the number of sampling points.

[0065] If the rate of decrease in standard deviation is ≤5% / minute, such as from 0.12 to 0.114, it indicates that the abnormality has not worsened and is judged as an occasional disturbance.

[0066] Check the status of other sensors, such as the correlation between the temperature and humidity sensors. If no abnormal correlation is found, such as the humidity not rising synchronously, then the possibility of a real fault can be ruled out.

[0067] Short-term disturbances, such as temperature rise caused by direct sunlight, are classified as Level 3 faults, reducing unnecessary downtime by 80%; only low-priority responses, such as auxiliary ventilation, are triggered to maintain power supply continuity.

[0068] Improved accuracy in classification: Through multi-condition joint verification and trend prediction, the accuracy of first-level fault identification is greatly improved, and the misclassification rate of second-level faults is significantly reduced. By introducing environmental humidity compensation and fluctuation standard deviation analysis, the system can operate stably in a wide temperature range of -20℃ to 50℃; The three-level fault mechanism reduces a large number of unnecessary maintenance work orders, improves operation and maintenance efficiency, and the abnormal operation detection of pressure relief valves can provide early warning of mechanical failures, avoiding the risk of "operating with defects".

[0069] After fault diagnosis, perform the following operations respectively: Level 1 Response Operation Chain: The vent valve 3 is forced to open fully to 90°, the cyclone separator at the bottom of the pressure relief chamber 2 is activated to discharge high-temperature particulate matter, a high-frequency carrier signal is sent to the smart meter, and 80% of the non-critical load is cut off; Level 2 response operation chain: Control the vent valve 3 to swing back and forth between 0° and 45° with a period of 2 seconds, simultaneously open the phase change material heat dissipation chamber on the side wall of cabinet 1, generate a work order containing temperature hot spot coordinates and suggested maintenance time on the local HMI interface, and broadcast the warning information through the voice synthesis module. Collect the humidity data of pressure relief chamber 2 every 5 minutes. If it does not decrease for 3 consecutive times, it will be upgraded to Level 1 response. Level 3 response operation chain: Keep the vent valve 3 closed, open the auxiliary ventilation window on the top of the cabinet to 60° opening, and limit the load current to 85% of the rated value through the PWM voltage regulation module. Generate an equipment status report every 5 minutes, including temperature peak, humidity trend and load rate, and send it to the inspection terminal within 500 meters through the LORA wireless network. If no manual confirmation is received within 48 hours, it will automatically upgrade to Level 2 response.

[0070] By designing a multi-level response chain and implementing a dynamic escalation mechanism, a complete closed loop for fault handling is constructed, specifically including: Level 1 response chain: rapid pressure relief and load shedding under extreme faults; Level 2 response chain: coordinated control of pressure balancing and active heat dissipation; Level 3 response chain: current limiting protection and status tracking feedback.

[0071] The specific implementation method is as follows: The first-level response operation chain, the emergency pressure relief stage, the pressure relief valve is fully opened to separate from the particulate matter, the vent valve 3 opens to 90° within 0.5 seconds, and the cyclone separator installed at the bottom of the pressure relief chamber 2 starts simultaneously. It uses centrifugal force to separate the high-temperature metal particles from the gas. The particulate matter settles into the bottom collection chamber, and the clean gas is discharged through the explosion-proof grille to avoid the high-temperature gas containing particles from igniting the surrounding equipment. The pressure relief efficiency is significantly improved.

[0072] Non-critical load shedding: High-frequency pulse signals are sent to smart meters via power line carrier communication to shed 80% of non-critical loads, such as landscape lighting and non-production air conditioning, according to preset priorities, while core loads, such as medical equipment and data center power supply, are retained. Based on the load importance tag library, the loads are dynamically sorted to ensure that the outage time of critical loads is ≤0.2 seconds.

[0073] The two-stage response operation chain, pressure imbalance stage, pressure relief valve swing control, the vent valve 3 swings back and forth between 0° and 45° with a period of 2 seconds, maintaining the pressure balance in the cabinet through periodic pressure relief, with a target pressure difference of ±10kPa.

[0074] Oscillation algorithm, opening degree Dynamically adapting to the pressure difference ΔP, the formula is: When ΔP returns to a safe range, the swing amplitude automatically decreases.

[0075] Phase change material heat dissipation: Open the phase change material compartment on the side wall of the cabinet. In this embodiment, it can be a paraffin-based composite material with a melting point of 58°C. It absorbs the heat from the overheated area and solidifies until the temperature drops below 45°C and then re-liquefies and circulates.

[0076] Intelligent work orders and voice alerts: Work orders are generated on the local HMI interface, marking the coordinates of temperature hotspots, such as X=1.2m, Y=0.8m, suggesting the maintenance time, such as "to be handled within 2 hours", and broadcasting it through the TTS engine.

[0077] Upgrade judgment mechanism: Collect humidity data of the pressure relief chamber every 5 minutes. If the humidity does not decrease after 3 consecutive tests, it is judged that the leak is not under control and the response is automatically upgraded to Level 1.

[0078] The three-level response operation chain includes early warning and observation level, ventilation and flow restriction control, and auxiliary ventilation window on the top of the cabinet opening to 60° to introduce external airflow for cooling. At the same time, the load current is limited to 85% of the rated value through the PWM voltage regulation module.

[0079] PWM control logic: The duty cycle is dynamically adjusted according to the temperature peak. For every 1°C increase in temperature, the duty cycle decreases by 2%.

[0080] A status report is generated every 5 minutes, including temperature peaks such as "C phase contact 98℃", humidity trends such as "Relief chamber humidity +3% per hour", and load rates such as "Current load 82%, 70% after limit".

[0081] Data is transmitted via the LoRa wireless network to a handheld inspection terminal within 500 meters, with a packet loss rate of ≤1%.

[0082] Automatic upgrade mechanism: If no manual confirmation is received within 48 hours, in this embodiment, the human can provide feedback by pressing a button or sign for receipt via the APP, and the system will automatically upgrade to a level 2 response, triggering the pressure relief valve to swing and the phase change heat dissipation.

[0083] The first-level response cyclone separator can remove all metal particles, reducing the risk of secondary combustion and explosion. The second-level response pressure relief valve swings to reduce pressure fluctuations inside the cabinet, avoiding structural fatigue damage.

[0084] The intelligent shelving strategy for non-critical loads shortens the outage time of core loads, reduces the user-perceived outage rate, and ensures the normal power supply for users to the greatest extent. The three-level response current limiting control enables the equipment to operate safely under slight overload, improving power supply availability.

[0085] The work order coordinate positioning of the Level 2 response greatly shortens the troubleshooting time, and the LORA status briefing improves the efficiency of on-site inspections and reduces labor costs.

[0086] It also includes the response optimization process: When the temperature gradient rate dT / dt≥8℃ / s, the vent valve 3 remains fully open until the temperature drops below 60℃. The pressure relief efficiency η is calculated as (actual pressure relief amount / theoretical maximum pressure relief amount)×100%. The theoretical maximum pressure relief amount is obtained by referring to the table based on the opening degree-flow curve of the vent valve 3. When η is below 70% for 3 consecutive seconds, the system switches to the backup pressure relief channel and triggers an audible and visual alarm. The temperature distribution heat map of adjacent switchgear is obtained through power line carrier communication to identify the direction of the high temperature area. If the high temperature area of ​​the adjacent cabinet extends toward this cabinet and the frequency of its vent valve operation increases by ≥2 times within 5 minutes after the fault, the load current limit threshold of this cabinet is dynamically reduced from 85% to 60%, and a pre-depressurization command is sent to the adjacent cabinet to open its vent valve to 30° opening in advance. Based on the fault data of this cabinet over the past 30 days, when the false alarm rate of the first-level fault is greater than 15%, the temperature gradient threshold will be increased from 15℃ / cm to 18℃ / cm. If the second-level response delay is greater than 3 seconds, the pulse cycle of the vent valve will be shortened from 2 seconds to 1.5 seconds, and the reference value of the humidity sensor will be recalibrated. Calculate the ratio of the number of times the vent valve actuates to the number of times it is designed to last. Percentage of worn area of ​​the inner wall coating of the pressure relief chamber Variance of data differences between adjacent sensors Health Index ,when At 0.6, the inspection cycle is shortened to 7 days, and the vent valve seal is forcibly replaced. When the value is ≥0.6, the carbon deposits in the pressure relief chamber should be purged every 24 hours.

[0087] Based on graded response, dynamic parameter optimization, multi-cabinet collaborative suppression, and closed-loop health status management are introduced to construct a full life cycle optimization system for fault response. The system dynamically switches channels and issues early warnings based on the pressure relief volume, suppresses fault propagation through data collaboration between adjacent cabinets, and optimizes judgment parameters and maintenance plans based on historical data.

[0088] The specific implementation method is as follows: The pressure relief efficiency is calculated by obtaining the opening-flow characteristic curve of the relief valve 3 through experimental calibration. Based on the opening-flow characteristic curve of the relief valve 3, the theoretical maximum pressure relief amount at the current opening degree is obtained by referring to a table. .

[0089] The actual pressure relief is measured by obtaining the real-time pressure relief amount through the flow sensor at the bottom of pressure relief chamber 2. .

[0090] Calculate the pressure relief efficiency. 00%, when If the pressure relief channel remains blocked for 3 seconds at 0%, it is determined that the pressure relief channel is blocked.

[0091] Channel switching and alarm: Switch to the backup pressure relief channel and trigger an audible and visual alarm to prompt maintenance personnel to clear the main channel, greatly improving pressure relief efficiency and avoiding secondary failures caused by channel blockage.

[0092] Multi-cabinet collaborative fault propagation suppression: Temperature thermal maps of adjacent cabinets are obtained through power line carrier communication to identify the direction of movement of high-temperature areas. If the temperature zone of an adjacent cabinet extends toward this cabinet and the frequency of its vent valve operation increases by ≥2 times within 5 minutes, it is judged as a high risk of fault propagation. The load current limit threshold of this cabinet is dynamically reduced from 85% to 60% to prevent overload from aggravating the fault. A pre-depressurization command is sent to the adjacent cabinet to open its vent valve to 30° in advance to release local pressure, thereby reducing the probability of fault propagation. The load current limiting strategy reduces line loss, and the pre-depressurization operation greatly reduces the risk of explosion in adjacent cabinets.

[0093] Statistically analyze the false alarm rate (number of false alarms / total number of triggers) of Level 1 faults over the past 30 days. If it is >15%, increase the temperature gradient threshold from 15℃ / cm to 18℃ / cm to reduce false triggers caused by environmental temperature rise interference.

[0094] If the secondary response delay is greater than 3 seconds (from the judgment to the completion of the pressure relief valve action), the pressure relief valve pulse cycle will be shortened from 2 seconds to 1.5 seconds to improve the response speed.

[0095] Recalibrate the humidity sensor reference value to eliminate zero-point drift caused by long-term use.

[0096] Health Index Calculate the lifespan of the pressure relief valve compared to The design life is 100,000 cycles, and the coating wear rate is... 00%, variance of adjacent sensor data Health Index ,when At 0.6, the inspection cycle is shortened to 7 days, and the vent valve seal is forcibly replaced. When the value is ≥0.6, the carbon deposits in the pressure relief chamber should be purged every 24 hours, and the backup sensor should be used to replace the abnormal sensor with a difference variance >20%.

[0097] The channel switching mechanism greatly shortens the pressure relief interruption time, ensuring rapid risk mitigation; the multi-cabinet collaborative strategy greatly reduces the incidence of fault chain reactions; the enhanced adaptive capability and threshold self-optimization significantly reduce the false judgment rate and greatly improve the response speed; the extended equipment life and health index management extend the life of key components and reduce maintenance costs.

[0098] It also includes the fault tracing and iteration process: associating external meteorological data such as temperature, humidity, wind speed and real-time load rate of the power grid at the time of the fault, and recording the opening angle, duration and associated temperature peak of each action of the vent valve 3; The feature matrix of historical failure cases is extracted using a twin neural network, including temperature gradient distribution, humidity change pattern and pressure difference fluctuation spectrum. The cosine similarity between the current failure and historical cases is calculated, and a treatment plan with a similarity of ≥90% is output. Model optimization involves adding an "inefficient handling" label to fault cases that recur after initial treatment and reducing their weighting by 40%. The neural network is retrained quarterly to update the weighting coefficients of temperature and humidity features. , ; Feedback on the handling results and statistics on the time to resolve the fault. Total pressure relief The ratio of these values ​​is used to generate the objective function for optimization. ,when When the value is greater than 1.2 times the baseline value, the vent valve control strategy version iteration is triggered.

[0099] Based on the above solutions, a fault lifecycle management system is constructed, covering data tracing, case matching, model iteration and effect feedback. It integrates meteorological, power grid load and equipment status data to accurately locate fault causes; it realizes efficient retrieval of historical handling plans based on twin neural networks; and it continuously improves handling efficiency through effect evaluation and weight optimization.

[0100] The specific implementation method is as follows: External environmental data is fused, meteorological data is synchronized, and external temperature, humidity, and wind speed data at the time of the fault are correlated to establish an Environmental Stress Index (ESI). ,in , , These are the normalized values ​​for ambient temperature, humidity, and wind speed, respectively.

[0101] Record the system load rate (e.g., 85% of peak load) when the fault occurs, analyze the correlation between load fluctuation and cabinet temperature rise, and record the opening angle, duration and associated temperature peak (e.g., when fully open to 90°, the temperature drops from 150° to 80°) for each operation of the vent valve to form an operation efficiency curve.

[0102] By combining the ESI index, it is possible to distinguish between failures caused by environmental factors (such as high temperature exposure) and equipment defects (such as poor contact), which greatly improves the accuracy of root cause location, and the action performance curve provides a quantitative basis for the optimization of pressure relief strategy.

[0103] The historical case matching of the twin neural network is used to extract the shape parameters (aspect ratio, area change rate) of the region with the largest gradient in the heat map; the duration of the mutation, the peak height and the second derivative symbol sequence are statistically analyzed; the 1-5Hz dominant frequency component of the pressure difference fluctuation is extracted by FFT transformation; and the above features are encoded into a 128-dimensional feature matrix M.

[0104] Cosine similarity matching is used to calculate the current fault feature matrix. Output the top 3 historical cases with a similarity of ≥90% and their corresponding handling solutions (such as pressure relief angle adjustment and load switching sequence).

[0105] Model optimization and weight iteration: For fault cases that recur after treatment, an "inefficient treatment" label is added and the weight is reduced in the matching weight. The neural network is retrained quarterly using new data, and the feature weights are dynamically adjusted. The weight of temperature gradient is increased (if the rise in ambient temperature becomes the main cause), and the weight of humidity change is decreased (if leakage is reduced after the sealing process is improved).

[0106] Fault resolution efficiency ratio ,when When the value is 0.2 times the baseline value, the current strategy is considered inefficient, for example, when the baseline value is 0.8 s / m³.

[0107] The pressure relief control strategy was upgraded, the vent valve opening curve was optimized, the full opening time was shortened, the ambient wind speed compensation coefficient was increased, and the pressure relief flow calculation model was adjusted.

[0108] By correlating the ESI index with the load, the contribution of external environment and equipment defects can be distinguished. For example, 70% is caused by high temperature and 30% by contact resistance. Historical case matching greatly shortens the time for developing disposal solutions. Quarterly retraining improves the accuracy of model prediction.

[0109] The EER feedback mechanism improves pressure relief efficiency and reduces the failure recurrence rate; the weighted iteration strategy greatly reduces the probability of matching inefficient cases.

[0110] It also includes virtual simulation and strategy verification process: import the three-dimensional model of the switch cabinet and historical sensor data into the digital twin model, build a high-precision simulation scene, and inject simulated faults such as vent valve jamming opening degree ≤30°, temperature sensor failure deviation ≥10℃, and communication delay ≥500ms; Response performance evaluation, with the following evaluation indicators defined: pressure relief efficiency η≥70%, load shedding accuracy α≥90%, and response delay Δt≤2s; Among them, pressure relief efficiency Load shearing accuracy ×100%, Response Delay Generate a three-dimensional evaluation matrix containing η, α, and Δt, and label the items that do not meet the standards; If any indicator in the evaluation matrix falls below the baseline value by 20% twice consecutively, the parameter calibration procedure is triggered, and the optimized pressure relief angle and pulse frequency are written into the firmware and wirelessly updated to the field equipment via OTA.

[0111] The specific implementation method is as follows: Based on CAD drawings and laser scanning data, a 1:1 three-dimensional digital twin model was constructed, including the cabinet structure, pressure relief chamber, and sensor layout. The physical properties (such as material thermal conductivity and vent valve flow curve) are completely consistent with the actual object, and the model was imported.

[0112] Import historical temperature, humidity, and air pressure data to reproduce the heat conduction, gas flow, and pressure change processes under real-world conditions.

[0113] Simulated fault injection: Air vent valve jamming: Limit the maximum opening of the virtual air vent valve to ≤30% to simulate mechanical wear-induced failure; Sensor failure: Superimpose ±10℃ temperature deviation or humidity drift into the digital model to test the system's fault tolerance; Communication delay: Insert a delay of ≥500ms into the control command transmission link to verify the robustness of the response strategy.

[0114] Response performance assessment and 3D matrix generation; pressure relief efficiency η: the percentage of actual pressure relief to theoretical maximum pressure relief, requiring η ≥ 70%; load shedding accuracy α: the degree of agreement between the actual non-critical load shedding ratio and the target value (e.g., 80%), requiring α ≥ 90%; response delay. ,Require ≤2 seconds.

[0115] Using η, α, and Δt as three-dimensional coordinate axes, the result points of each test are marked, and items that do not meet the standards (such as η=65%, α=85%) are identified.

[0116] Example matrix: If the result of a certain test is (η=68%, α=92%, Δt=1.8s), then mark η as not meeting the standard.

[0117] Parameter calibration and remote upgrade, calibration trigger condition: if any indicator is lower than the benchmark value by 20% twice in a row (e.g., η benchmark 70%, actual measurement 56%), the parameter optimization program will be started: pressure relief angle optimization: search for the optimal opening curve through genetic algorithm to increase η to ≥75%; pulse frequency adjustment: shorten the pressure relief valve swing cycle (e.g., from 2 seconds to 1.5 seconds) to ensure Δt≤2 seconds.

[0118] OTA wireless update: The optimized parameters are packaged into a firmware upgrade package and pushed to the field device through an encrypted channel (AES-256); the device automatically installs the update during low-load periods without manual intervention.

[0119] Testing the vent valve jamming scenario significantly reduced the field failure rate; sensor failure simulation helped optimize the fault-tolerant algorithm and reduced the misjudgment rate; the three-dimensional matrix intuitively reflects the shortcomings of the strategy and improves the clarity of the optimization direction; the quantitative evaluation of the load shedding accuracy α shortened the power outage time of critical loads.

[0120] The parameter optimization cycle is shortened, and the strategy iteration speed is improved; OTA upgrades avoid on-site maintenance and reduce the cost per upgrade.

[0121] It also includes an adaptive control process for the pressure relief process: the vent valve 3 is equipped with a mass flow meter to monitor the pressure relief flow in real time. In stage 1, 0-5 seconds, the vent valve 3 is fully opened to 90°, and the high-temperature free metal particles are quickly discharged by utilizing the initial pressure difference potential energy, so as to reduce the risk of combustion and explosion. The pressure relief flow is monitored in real time by the mass flow meter to ensure that the flow rate in stage 1 is ≥25m³ / min and the diameter of the discharged particles is ≤50μm.

[0122] Phase 2, 5-15 seconds: Adjust the vent valve opening to 60% to maintain a constant pressure relief flow rate of 20m³ / min ±5% to avoid sudden stress changes in the cabinet structure caused by a sudden pressure drop; Calculate the rate of pressure drop inside the cabinet every second. If the rate is ≥10kPa / s, trigger a fine adjustment of the vent valve opening ±5°.

[0123] Phase 3, after 15 seconds: Decrease the opening by 10° every 5 seconds, and monitor the pressure decay rate inside the cabinet simultaneously. If the rate is <5kPa / s, pause the adjustment and maintain the current opening until the pressure stabilizes. Acquire real-time airflow velocity monitoring data in pressure relief chamber 2, and dynamically select the flow guidance mode based on the velocity value: when the airflow velocity is ≥10m / s, adjust the flow guide plate to the maximum flow guidance angle, prioritize expanding the cross-sectional area of ​​the pressure relief channel to reduce pressure loss, expand the cross-sectional area of ​​the pressure relief channel to reduce airflow resistance and reduce pressure loss; when the airflow velocity is ≤5m / s and <10m / s, periodically swing the flow guide plate angle ±15° to avoid particulate matter deposition using periodic turbulence; when the airflow velocity is <5m / s, lock the flow guide plate to the minimum angle and start the backup high-pressure air pump for forced pressurization. After each depressurization is completed, the inner wall of the depressurization chamber 2 is scanned with a laser rangefinder with an accuracy of 0.1mm to detect local depressions ≥0.5mm and mark their coordinates. The depression area is located in three dimensions, such as X=1.2m, Y=0.6m, and recorded in the structural damage database. If a depression is detected at the same location three times in a row, a structural reinforcement command is triggered to deploy carbon fiber reinforcement sheets with a thickness of 0.5mm and a coverage area ≥200% of the depression area.

[0124] It also includes 3D visualization and trusted evidence preservation processes: Temperature sensor data is mapped onto the 3D model of the cabinet to generate a color temperature gradient heat map. Blue represents <50℃, which is a safe zone; red represents ≥100℃, which is a dangerous zone; and yellow represents 50-99℃, which is a warning zone. The discrete temperature points are expanded into a continuous color temperature gradient map using a bilinear interpolation algorithm, which is updated once per second. Overlaying pressure relief airflow vector arrows dynamically displays airflow speed and direction; arrow length indicates airflow speed, and arrow direction indicates direction. Based on the airflow velocity sensor data in the pressure relief chamber, dynamic vector arrows are generated: the arrow length is linearly related to the airflow velocity, such as 10 m / s corresponding to a 1 cm long arrow; the arrow direction is consistent with the actual direction of airflow movement and is corrected in real time through a particle tracking algorithm.

[0125] Example: In the event of an arc fault, the high-temperature area (red) is accompanied by a high-speed airflow arrow (3cm long) pointing towards the vent valve. Operation and maintenance personnel are allowed to manually adjust the feature threshold range of ±10% or force switch the fault level in the HMI interface. The adjustment range is controlled by the hierarchical permission level, such as ±15% for engineers and ±5% for operators. Digital signatures are added to manual operations and bound to the operation time and employee number and stored in the security database. Each manual operation generates a digital signature, including the operation time, employee number, and device serial number, which is encrypted and stored in a secure database. The signature algorithm is based on RSA-2048 asymmetric encryption, and the private key is kept by the operator's smart card.

[0126] When the automatic judgment does not match the actual situation, the fault level can be forcibly switched (e.g., level 2 → level 1). The system records the reason for the switch, such as "on-site observation of electric arc". The key operation records fault judgment, vent valve action, and manual intervention are generated to generate SHA-256 hash values ​​and written into the distributed ledger of the consortium blockchain. Each block contains a snapshot of the system status of the previous 10 minutes.

[0127] The specific implementation method is as follows: Key operation hash generation: SHA-256 hash values ​​are calculated for data such as fault determination time, vent valve operating angle, and manual intervention records. .

[0128] The hash values ​​are written to a consortium blockchain comprised of power grid companies and regulatory agencies. Each block contains a snapshot of the system from the previous 10 minutes, such as temperature peaks and depressurization efficiency.

[0129] The block structure consists of: previous block hash, current operation hash, system snapshot (compressed into a JSON string), and timestamp.

[0130] During the audit, data integrity is verified by comparing the on-chain hash value with the real-time calculation results in the local database.

[0131] Three-dimensional heat maps greatly shorten fault location time and significantly improve troubleshooting efficiency, while manual threshold adjustment mechanisms improve the speed of handling misjudgment scenarios.

[0132] Operational traceability is achieved by binding digital signatures to employee IDs, enabling full-process traceability of "who operated, when operated, and what was operated".

[0133] Collaborative control security: The hierarchical permission mechanism avoids unauthorized operations, illegal operations are completely blocked, and system snapshots on the blockchain provide a complete spatiotemporal context for fault review.

[0134] It also includes a self-healing and predictive maintenance process: A laser displacement sensor is installed on the top of the pressure relief chamber 2. The laser displacement sensor scans the inner wall of the pressure relief chamber 2 to detect permanent deformation of ≥1mm. A shape memory alloy bracket is deployed in the deformation area and heated to 80℃ to expand it, fill the gap and restore more than 90% of the mechanical strength. Construct a temperature-humidity-pressure correlation matrix and use an LSTM network to predict the probability of failure in the next 24 hours. ,when When the leakage rate is ≥30%, pressurize to 110% of the rated value. If the leakage rate is ≤0.5% / min, it is considered qualified. Inject ultrasonic cleaning fluid at a frequency of 28kHz and run for 5 minutes. Based on the Health Index (HI), spare parts replacement suggestions are generated. When HI < 0.6, the system automatically orders vent valve seals and pushes replacement tutorial videos. When HI ≥ 0.6, high-risk cabinets are marked on the GIS map, and the optimal inspection route is planned.

[0135] The specific implementation method is as follows: Laser scanning inspection: A laser displacement sensor is installed on the top of the pressure relief chamber to scan the inner wall surface with an accuracy of 0.1mm and detect permanent deformations of ≥1mm (such as impact dents or corrosion pits).

[0136] Coordinate positioning: The deformed area is marked with three-dimensional coordinates, and the deformation depth and area are recorded.

[0137] A nickel-titanium alloy support is deployed in the deformation area and heated to 80°C to trigger the shape memory effect. The support expands to fill the gaps, restoring the structural strength to more than 90% of the original design. After the power is turned off, the alloy maintains its expanded shape, forming a permanent reinforcement.

[0138] LSTM network modeling, inputting 24-hour historical data of temperature-humidity-pressure correlation matrix, outputting the failure probability for the next 24 hours. ,when When the percentage is ≥30%, it is considered a high-risk state.

[0139] Pressure sealing test: Pressurize the air pressure inside the cabinet to 110% of the rated value and monitor the leakage rate. If the leakage rate is ≤0.5% / minute, the seal is deemed qualified; otherwise, the sealing ring replacement command is triggered.

[0140] Ultrasonic cleaning: Inject a special cleaning solution into the pressure relief chamber and run it with 28kHz ultrasound for 5 minutes to remove deposited particles and oxide layers and restore the smooth flow of air.

[0141] Health Index (HI) Calculation: The HI value (0-1) is generated by combining indicators such as pressure relief valve life, inner wall wear rate, and sensor consistency. HI < 0.6 indicates severe deterioration. If HI < 0.6, the pressure relief valve sealing ring will be ordered automatically and an installation tutorial video will be pushed. If HI ≥ 0.6, high-risk cabinets will be marked on the GIS map, and the optimal inspection route will be planned. The 5 devices with the lowest HI can be selected for coverage.

[0142] The shortest inspection path is calculated based on Dijkstra's algorithm, and high-risk equipment is visited first to reduce inspection time.

[0143] Shape memory alloy repair significantly extends the lifespan of pressure relief chambers; laser detection can provide early warning of structural failure risks; LSTM model prediction of failure probability has a greatly reduced error; health index-driven spare parts ordering reduces inventory costs; and optimal path planning improves the efficiency of manual inspection.

[0144] This technical solution significantly improves the accuracy and timeliness of fault detection through multi-dimensional data fusion and dynamic feature extraction. Based on the collaborative analysis of temperature gradient, humidity change, and differential pressure fluctuation, it achieves refined classification of fault levels, effectively distinguishing different fault types such as arc propagation and insulation leakage, and greatly reducing the risk of misjudgment. The graded response mechanism matches differentiated control strategies to maintain power supply continuity to the maximum extent while ensuring safety. Combined with self-healing repair and predictive maintenance technologies, it can proactively repair structural damage and predict potential faults, extend equipment life, and reduce unplanned downtime.

[0145] Through virtual simulation verification and dynamic parameter optimization, a closed-loop system of self-iterative strategy was constructed to ensure the reliability and adaptability of control logic. Three-dimensional visualization and blockchain evidence storage technologies enhanced the traceability and transparency of the operation and maintenance process, meeting the requirements of power safety audits. The intelligent maintenance decision-making system automatically optimizes spare parts replacement and inspection routes based on health indices, achieving efficient resource allocation. The overall solution upgrades fault handling from passive response to proactive defense and self-optimization, providing systematic support for the reliable operation and efficient maintenance of the smart grid.

[0146] The above-described embodiments are detailed and specific, illustrating preferred embodiments of the present invention. They are only used to illustrate the technical ideas and features of the present invention, with the aim of enabling those skilled in the art to understand the content of the present invention and implement it accordingly. However, they are not limited to the present invention, and the patent scope of the present invention cannot be limited by this embodiment alone. That is, any equivalent changes or modifications made to the spirit disclosed in the present invention, without departing from the structure of the present invention, such as local improvements within the system and modifications or transformations between subsystems, are still within the patent scope of the present invention.

Claims

1. A multi-chamber isolated protective switchgear structure, comprising a cabinet (1), characterized in that, It also includes a pressure relief chamber (2), which is installed at the bottom rear side of the cabinet (1). A vent valve (3) is installed on the top of the pressure relief chamber (2). The upper opening of the vent valve (3) is connected to the enclosed space of the cabinet (1). The enclosed space is the main wiring compartment of the switch cabinet. After the cabinet door of the cabinet (1) is opened, it becomes the switch operation panel. The lower opening of the vent valve (3) is connected to the pressure relief chamber (2). The enclosed space of the cabinet (1) is filled with dry air, nitrogen or sulfur hexafluoride. The main wiring compartment, the pressure relief chamber (2) and the switch operation panel are completely isolated. The vent valve (3) has an embedded pressure sensor for monitoring the dynamic pressure difference between the enclosed space of the cabinet (1) and the pressure relief chamber (2).

2. The multi-chamber isolated protective switchgear structure according to claim 1, characterized in that, The switchgear uses a fault response method, including the following steps: Step S1: Install a temperature sensor array around the busbar connection point, circuit breaker contact and vent valve (3) inside the cabinet (1) to generate a temperature distribution heat map at a frequency of 1 time / second. Install a wind speed sensor and an ambient temperature and humidity sensor outside the cabinet (1) with a sampling frequency of 1 time / second. Step S2: Install a humidity sensor on the inner wall of the pressure relief chamber (2) to collect humidity change curves in real time, with a sampling frequency of 10 times / second; Step S3: Obtain the pressure difference fluctuation amplitude between the enclosed space of the cabinet (1) and the pressure relief chamber (2) through the air pressure sensor. ; Step S4: Extract the direction of movement of the region with the largest gradient in the temperature distribution thermogram, and calculate the rate of change of its horizontal distance from the vent valve (3), i.e., the temperature gradient rate. The unit is cm / s; Step S5: Calculate the second derivative of the humidity change curve to obtain the humidity change intensity. ; Step S6, , , Normalize to the 0-1 range to generate multidimensional feature vectors. ; Step S7, when ≥0.7、 ≥0.5 or When the value is ≥0.6, a graded response mechanism is triggered.

3. The fault response method according to claim 2, characterized in that, The fault level classification of the graded response mechanism in step S7 includes the following steps: Level 1 fault diagnosis: When When the value is ≥0.9, the high-temperature area movement tracking is activated. If the high-temperature area moves towards the vent valve (3) at a speed of ≥5cm / s and the moving distance is shortened by more than 30% within 10 seconds, the humidity change detection is activated. When the second derivative value of the humidity change curve is ≥0.9, the humidity change detection is activated. If three consecutive samples are positive, and the vent valve (3) does not reach a 90° opening degree within 1 second after receiving the opening command, it is judged as a level one fault; Level 2 fault determination: When A level 2 fault is defined as follows: a sawtooth-shaped fluctuation with an amplitude ≥ 0.2 occurs within a period of 2-5 seconds, and the differential pressure fluctuation exceeds ±25 kPa for 1 minute; or the absolute value of humidity exceeds 30% of the ambient humidity but the second derivative is negative. Level 3 fault determination: When When a single characteristic value exceeds the threshold, a 15-minute deterioration trend monitoring is initiated, and the standard deviation of temperature, humidity, and pressure difference fluctuations is calculated. If the standard deviation decrease rate is ≤5% / minute and there are no associated sensor abnormalities, it is judged as a level three fault.

4. The fault response method according to claim 3, characterized in that, After fault diagnosis, perform the following operations respectively: Level 1 response operation chain: The vent valve (3) is forced to open fully to 90°, the cyclone separator at the bottom of the pressure relief chamber (2) is activated, high-temperature particulate matter is discharged, a high-frequency carrier signal is sent to the smart meter, and 80% of the non-critical load is cut off; Level 2 response operation chain: control the vent valve (3) to swing back and forth between 0° and 45° with a period of 2 seconds, simultaneously open the phase change material heat dissipation chamber on the side wall of the cabinet (1), generate a work order containing temperature hot spot coordinates and suggested maintenance time on the local HMI interface, and broadcast the warning information through the voice synthesis module. Collect the humidity data of the pressure relief chamber (2) every 5 minutes. If it does not decrease for 3 consecutive times, it will be upgraded to Level 1 response. Level 3 response operation chain: Keep the vent valve (3) closed, open the auxiliary ventilation window on the top of the cabinet to 60° opening, and limit the load current to 85% of the rated value through the PWM voltage regulation module. Generate a brief equipment status report every 5 minutes, including temperature peak, humidity trend and load rate, and send it to the inspection terminal within 500 meters through the LORA wireless network. If no manual confirmation is received within 48 hours, it will automatically upgrade to Level 2 response.

5. The fault response method according to claim 4, characterized in that, It also includes the response optimization process: When the temperature gradient rate dT / dt≥8℃ / s, the vent valve (3) remains fully open until the temperature drops below 60℃. The pressure relief efficiency η = (actual pressure relief amount / theoretical maximum pressure relief amount) × 100%. The theoretical maximum pressure relief amount is obtained by referring to the table based on the opening degree-flow curve of the vent valve (3). When η is below 70% for 3 consecutive seconds, switch to the backup pressure relief channel and trigger the audible and visual alarm. The temperature distribution heat map of adjacent switchgear is obtained through power line carrier communication to identify the direction of the high temperature area. If the high temperature area of ​​the adjacent cabinet extends toward this cabinet and the frequency of its vent valve operation increases by ≥2 times within 5 minutes after the fault, the load current limit threshold of this cabinet is dynamically reduced from 85% to 60%, and a pre-depressurization command is sent to the adjacent cabinet to open its vent valve to 30° opening in advance. Based on the fault data of this cabinet over the past 30 days, when the false alarm rate of the first-level fault is greater than 15%, the temperature gradient threshold will be increased from 15℃ / cm to 18℃ / cm. If the second-level response delay is greater than 3 seconds, the pulse cycle of the vent valve will be shortened from 2 seconds to 1.5 seconds, and the reference value of the humidity sensor will be recalibrated. Calculate the ratio of the number of times the vent valve actuates to the number of times it is designed to last. Percentage of worn area of ​​the inner wall coating of the pressure relief chamber Variance of data differences between adjacent sensors Health Index ,when At 0.6, the inspection cycle is shortened to 7 days, and the vent valve seal is replaced. When the value is ≥0.6, the carbon deposits in the pressure relief chamber should be purged every 24 hours.

6. The fault response method according to claim 5, characterized in that, It also includes fault tracing and iteration process: associating external meteorological data such as temperature, humidity, wind speed and real-time load rate of the power grid when the fault occurs, and recording the opening angle, duration and associated temperature peak of each action of the vent valve (3); The feature matrix of historical failure cases is extracted using a twin neural network, including temperature gradient distribution, humidity change pattern and pressure difference fluctuation spectrum. The cosine similarity between the current failure and historical cases is calculated, and a treatment plan with a similarity of ≥90% is output. The model was optimized by adding an "inefficient handling" label to cases that recurred after initial treatment and reducing their weight in the matching by 40%. The neural network was retrained quarterly to update the weight coefficients of temperature and humidity features. , ; Feedback on the handling results and statistics on the time to resolve the fault. Total pressure relief The ratio of these values ​​is used to generate the objective function for optimization. ,when When the value is greater than 1.2 times the baseline value, the vent valve control strategy version iteration is triggered.

7. The fault response method according to claim 6, characterized in that, It also includes virtual simulation and strategy verification process: import the three-dimensional model of the switch cabinet and historical sensor data into the digital twin model, build a high-precision simulation scene, and inject simulated faults such as vent valve jamming opening degree ≤30°, temperature sensor failure deviation ≥10℃, and communication delay ≥500ms; Response performance evaluation, with the following evaluation indicators defined: pressure relief efficiency η≥70%, load shedding accuracy α≥90%, and response delay Δt≤2s; Among them, pressure relief efficiency Load shearing accuracy ×100%, Response Delay Generate a three-dimensional evaluation matrix containing η, α, and Δt, and label the items that do not meet the standards; If any indicator in the evaluation matrix falls below the baseline value by 20% twice consecutively, the parameter calibration procedure is triggered, and the optimized pressure relief angle and pulse frequency are written into the firmware and wirelessly updated to the field equipment via OTA.

8. The fault response method according to claim 6, characterized in that, It also includes an adaptive control process for the pressure relief process: the vent valve (3) is equipped with a mass flow meter to monitor the pressure relief flow in real time. In stage 1, 0-5 seconds, the vent valve (3) is fully opened to 90° to preferentially discharge high-temperature free metal particles. In stage 2, 5-15 seconds, the opening is controlled to 60% to maintain a constant pressure relief flow of 20m³ / min±5%. In stage 3, after 15 seconds, the opening is reduced by 10° every 5 seconds, and the pressure decay rate in the cabinet is monitored synchronously. If the rate is <5kPa / s, the adjustment is paused. Obtain real-time airflow velocity monitoring data in the pressure relief chamber (2), and dynamically select the flow guiding mode based on the velocity value: when the airflow velocity is ≥10m / s, adjust the flow guide plate to the maximum flow guiding angle; when 5m / s≤ airflow velocity<10m / s, periodically swing the flow guide plate angle ±15°; when the airflow velocity is <5m / s, lock the flow guide plate to the minimum angle, and start the backup high-pressure air pump for forced pressurization. After each depressurization is completed, the inner wall of the depressurization chamber (2) is scanned by a laser rangefinder to detect local depressions ≥0.5mm and mark the coordinates. If a depression is detected three times in a row at the same location, a structural reinforcement command is triggered and carbon fiber reinforcement sheets are deployed.

9. The fault response method according to claim 8, characterized in that, It also includes 3D visualization and trusted evidence preservation processes: Temperature sensor data is mapped onto the cabinet's 3D model to generate a color temperature gradient heat map, where blue represents <50℃ and red represents ≥100℃. Pressure relief airflow vector arrows are superimposed to dynamically display airflow speed and direction, with arrow length indicating airflow speed and arrow direction indicating direction. It allows maintenance personnel to manually adjust the feature threshold within ±10% range or force switch the fault level in the HMI interface, add digital signatures to manual operations, and bind them to the operation time and employee number and store them in the security database; Key operation records, fault determination, vent valve action, and manual intervention are generated to generate SHA-256 hash values ​​and written into the distributed ledger of the consortium blockchain. Each block contains a snapshot of the system state from the preceding 10 minutes.

10. The fault response method according to claim 9, characterized in that, It also includes self-healing and predictive maintenance processes: a laser displacement sensor is installed on the top of the pressure relief chamber (2), and the inner wall of the pressure relief chamber (2) is scanned by the laser displacement sensor to detect permanent deformation of ≥1mm. A shape memory alloy bracket is deployed in the deformation area, and it is heated to 80℃ to expand it, fill the gap and restore more than 90% of the mechanical strength. Construct a temperature-humidity-pressure correlation matrix and use an LSTM network to predict the probability of failure in the next 24 hours. ,when When the leakage rate is ≥30%, pressurize to 110% of the rated value. If the leakage rate is ≤0.5% / min, it is considered qualified. Inject ultrasonic cleaning fluid at a frequency of 28kHz and run for 5 minutes. Based on the Health Index (HI), spare parts replacement suggestions are generated. When HI < 0.6, the system automatically orders vent valve seals and pushes replacement tutorial videos. When HI ≥ 0.6, high-risk cabinets are marked on the GIS map, and the optimal inspection route is planned.

Images