A detection method for electrical switch cabinet based on Internet of Things

Abstract

The present invention discloses an electrical switch cabinet detection method based on the Internet of Things, and specifically relates to the technical field of electrical switch cabinets. Sensors are installed inside or outside the switch cabinet to collect temperature, humidity, current, voltage and gas concentration, and wireless communication technology is used to transmit the data to an edge computing node for preprocessing. The preprocessed data is uploaded to a cloud platform for storage and trend analysis in the cloud. Based on big data analysis and artificial intelligence algorithms, potential faults are predicted. Users can view the equipment operation status and prediction results in real time through the cloud platform, and remotely execute emergency shutdown or equipment scheduling operations according to early warning information. The reliability problems caused by communication interruption or data delay in the existing Internet of Things system are effectively overcome, and the safety and intelligence level of switch cabinet monitoring are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electrical switch cabinets, and in particular to an electrical switch cabinet detection method based on the Internet of Things. Background Art

[0002] Electrical switchgear detection based on the Internet of Things is a way to use the Internet of Things technology to intelligently monitor and manage electrical switchgear. By installing various sensors (such as temperature, humidity, current, voltage and gas concentration sensors) inside or outside the switchgear, the operating status and environmental parameters of the equipment are collected in real time, and these data are transmitted to the cloud or centralized management system for processing and analysis through wireless communication technology (such as Wi-Fi, NB-IoT or ZigBee). This method can help operation and maintenance personnel to grasp the operating status of the equipment in real time, quickly discover potential faults and take targeted measures, thereby improving the operating safety and stability of the equipment.

[0003] In addition, IoT-based detection technology also supports remote monitoring and intelligent early warning functions, displaying historical data and trend analysis of equipment operation through cloud platforms or mobile terminals. Using big data and artificial intelligence algorithms, the collected data can be deeply mined and predicted, and early warning information can be issued in a timely manner to help users avoid power outages or safety accidents caused by equipment failures. This intelligent detection method not only improves the management efficiency of electrical switchgear, but also provides important guarantees for the stable operation of the power system.

[0004] The prior art has the following deficiencies:

[0005] IoT systems rely heavily on network connections. If communication interruptions occur at critical moments in the operation of the equipment (such as wireless signal interference or base station failure), the system may not be able to obtain the operating status of the switch cabinet in real time. For example, when the temperature of the equipment rises sharply, if the data cannot be uploaded to the cloud platform in time, the system may not be able to issue an emergency shutdown command, causing the equipment to overheat and burn or even cause a fire. In addition, system crashes on the cloud platform or the server itself also pose risks, especially during peak periods when multiple devices upload data at the same time, data delays or losses may occur. This problem may further affect the real-time and reliability of the system and delay the handling of critical faults. Summary of the invention

[0006] The purpose of the present invention is to provide an electrical switch cabinet detection method based on the Internet of Things to solve the shortcomings of the background technology.

[0007] In order to achieve the above object, the present invention provides the following technical solution: an electrical switch cabinet detection method based on the Internet of Things, comprising the following steps:

[0008] S1: Install several sensors inside or outside the electrical switch cabinet to collect operating status data and environmental parameters including temperature, humidity, current, voltage and gas concentration;

[0009] S2: Use wireless communication technology to transmit the collected data to the edge computing node, pre-process the data, and reduce the transmission load;

[0010] S3: The pre-processed data is uploaded to the cloud platform through the communication network. The cloud platform stores and archives the data to form the historical data and trend analysis model of the electrical switchgear operation;

[0011] S4: Based on the big data analysis and artificial intelligence algorithms of the cloud platform, the collected data is deeply mined to predict potential faults;

[0012] S5: In the event of communication interruption, the edge computing node operates independently, continues to monitor the device status through local storage and analysis functions, and uploads the data to the cloud platform after communication is restored;

[0013] S6: Users can view the operating status of the electrical switchgear and potential fault prediction results in real time through the cloud platform, and remotely perform emergency shutdown or equipment scheduling operations based on the early warning information.

[0014] Preferably, in S4, the local discharge intensity in the high-voltage switch cabinet is analyzed to generate a local discharge intensity abnormality index, and the method for obtaining the local discharge intensity abnormality index is:

[0015] Collect the partial discharge signal x(t) and record it as a discrete time series x[n], where n=1,2,…,N; represents the sampling point, N is the total number of sampling points of the signal, the sampling frequency is fs, the number of decomposition layers is J, and the wavelet components are calculated: the signal is decomposed into approximate components by wavelet transform and some details ; Low frequency approximate component: ;in, is the scaled and translated wavelet basis function; high-frequency detail component: ;in, is the corresponding wavelet function;

[0016] For each wavelet component and Calculate its energy , the expression is: Total energy The calculation expression is: ; Calculate the proportion of each frequency band energy in the total energy: ;in and Reflect the contribution of signals in different frequency bands and extract the energy distribution under normal conditions from historical operation data ;

[0017] By comparing the current energy distribution with the normal state distribution, the partial discharge intensity abnormality index is defined, and the expression is: ;in is the importance weight of each frequency band, which can be set based on the discharge characteristics experience. KML is the local discharge intensity abnormality index.

[0018] Preferably, in S4, the particle concentration variation index is generated after analyzing the abnormal changes in the particle concentration inside and outside the switch cabinet. The particle concentration variation index is obtained by:

[0019] The monitoring data of particle concentration is set to include multiple dimensions, which are defined as a multidimensional data matrix X: ; Where p is the number of samples, d is the feature dimension, and the covariance matrix of the data matrix X is calculated , the expression is: ; In the formula, is the eigenvector of the i-th sample, Σ is a d×d symmetric positive definite matrix, T is the matrix transpose, is the mean of each feature dimension;

[0020] For each sample Calculate the Mahalanobis distance from the center of the distribution , the expression is: ; If Σ is a positive definite matrix, then directly calculate its inverse matrix , Is a scalar, indicating the degree to which the i-th sample deviates from the center of the distribution;

[0021] The Mahalanobis distance is normalized to the particle concentration variation index, and the expression is: ; Where MKN is the particle concentration variation index.

[0022] Preferably, in S4, the particle concentration variation index and the local discharge intensity anomaly index are converted into a comprehensive feature vector, and the comprehensive feature vector is used as the input of the machine learning model. The machine learning model predicts the stability value label of the switch cabinet operating state with each group of comprehensive feature vectors as the prediction target, and minimizes the sum of prediction errors of the stability value labels of all switch cabinet operating states as the training target. The machine learning model is trained until the sum of prediction errors converges, and the model training is stopped. The stability value of the switch cabinet operating state is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

[0023] Preferably, the acquired stability value of the switch cabinet operating status is compared with a stability reference threshold value pre-set based on historical data. If the stability value of the switch cabinet operating status is greater than or equal to the stability reference threshold value, it means that the stability of the switch cabinet operating status is high and the probability of failure is low; if the stability value of the switch cabinet operating status is less than the stability reference threshold value, it means that the stability of the switch cabinet operating status is low and the probability of failure is high.

[0024] Preferably, in S6, the user views the operating status of the electrical switch cabinet in real time through the cloud platform, and the operating status data includes temperature, humidity, current, voltage and gas concentration;

[0025] The operating status data and the stability value of the switchgear operating status are used as the input items of the fuzzy logic, and the warning level is used as the output item of the fuzzy logic;

[0026] The operating status data and stability values ​​are converted into fuzzy sets and expressed by linguistic variables;

[0027] Use triangular or trapezoidal membership functions to describe the fuzzy sets of each variable; define fuzzy rules based on domain knowledge or historical data to map input variables to output warning levels;

[0028] The fuzzy rule base needs to cover all input combinations, and the number of rules is the Cartesian product of the number of fuzzy sets of input variables;

[0029] According to the input data and rule base, the output fuzzy set is calculated using fuzzy reasoning method;

[0030] Calculate the membership value of the input variable for each rule;

[0031] Use the minimum method to calculate the composite membership of the input variables of each rule;

[0032] Aggregate the outputs of all rules to obtain the output fuzzy set;

[0033] The output fuzzy set is converted into a clear warning level value, and the centroid of the fuzzy set is calculated as the output value. The expression is: ;in, is the membership function of the output fuzzy set, The warning level value.

[0034] Preferably, if 0< ΅ά0.4 is a low risk level, no action is required, regular inspection is sufficient; if 0.4< <0.8, it is a medium risk level, it is recommended to arrange maintenance and check the equipment status; if 0.8< <1.0 indicates a high risk level, which immediately triggers an emergency shutdown or dispatches backup equipment.

[0035] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0036] 1. The present invention proposes an efficient and intelligent electrical switch cabinet detection method by integrating Internet of Things technology, edge computing, big data analysis, artificial intelligence algorithm and fuzzy logic analysis. This method collects equipment operation status data through sensors, uses edge computing nodes to preprocess and locally store the data, and uploads the data to the cloud platform for storage and analysis. Even when the communication is interrupted, the edge node can still operate independently to ensure the continuity of monitoring and the integrity of the data. At the same time, based on the feature extraction and comprehensive analysis of the partial discharge intensity anomaly index and the particle concentration variation index, the polynomial regression model is used to predict the stability of the equipment operation status, and combined with fuzzy logic analysis, the real-time operation status and stability value are mapped to an intuitive warning level, providing users with accurate fault prediction and decision support.

[0037] 2. The present invention effectively solves the high dependence of traditional Internet of Things monitoring systems on network connections and the risk of system crashes, greatly improving the reliability and intelligence level of electrical switchgear monitoring. Through real-time data visualization, historical trend analysis and intelligent early warning functions, users can grasp the operating status of the equipment in real time, and take maintenance or emergency shutdown measures in time according to the early warning level, reduce the equipment failure rate, and ensure the safe and stable operation of the power system. In addition, this method has significant advantages in data transmission efficiency, system fault tolerance, and fault prediction accuracy, providing comprehensive technical support for the intelligent management of electrical equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0039] Figure 1 The present invention is a flow chart of the method. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0041] For examples, see Figure 1As shown, the electrical switch cabinet detection method based on the Internet of Things described in this embodiment includes the following steps:

[0042] S1: Install several sensors inside or outside the electrical switch cabinet to collect operating status data and environmental parameters including temperature, humidity, current, voltage and gas concentration;

[0043] S2: Use wireless communication technology to transmit the collected data to the edge computing node, pre-process the data, and reduce the transmission load;

[0044] S3: The pre-processed data is uploaded to the cloud platform through the communication network. The cloud platform stores and archives the data to form the historical data and trend analysis model of the electrical switchgear operation;

[0045] S4: Based on the big data analysis and artificial intelligence algorithms of the cloud platform, the collected data is deeply mined to predict potential faults;

[0046] S5: In the event of communication interruption, the edge computing node operates independently, continues to monitor the device status through local storage and analysis functions, and uploads the data to the cloud platform after communication is restored;

[0047] S6: Users can view the operating status of the electrical switchgear and potential fault prediction results in real time through the cloud platform, and remotely perform emergency shutdown or equipment scheduling operations based on the early warning information.

[0048] Several sensors are installed inside or outside the electrical switch cabinet to comprehensively monitor the equipment operating status and environmental conditions to ensure the safety and stability of the system.

[0049] Temperature sensor: monitors the temperature of key components in the switch cabinet (such as contact points, busbars, etc.), and promptly detects abnormal temperature rise caused by overload, aging or poor contact to avoid fire or equipment damage.

[0050] Humidity sensor: Detects humidity changes inside the switch cabinet to prevent problems such as insulation degradation, electrical equipment corrosion or short circuit caused by humid environment. It is especially suitable for humid or coastal areas.

[0051] Current sensor: measures operating current in real time, monitors circuit load conditions, identifies abnormal phenomena such as overload, leakage or unbalanced current, and ensures power safety.

[0052] Voltage sensor: measures input and output voltages, ensures voltage stability, detects voltage fluctuations or mutations (such as overvoltage or undervoltage), and protects equipment and load operation.

[0053] Gas concentration sensor: monitors changes in gas composition in the switch cabinet, detects whether there are harmful gases (such as SF6 gas leakage) or combustion gases generated by arc faults, and identifies potential arc discharges or equipment abnormalities in advance.

[0054] Temperature, current and voltage sensors are placed near key components (such as terminals, circuit breakers, busbars) to accurately collect operating data; gas concentration sensors are usually installed on the top of the switch cabinet or at locations where gas is likely to accumulate to detect leaks or burning gases in a timely manner.

[0055] Humidity sensors and environmental monitoring devices can be placed outside the switch cabinet to sense the impact of the external environment on equipment operation, especially in humid or high temperature environments.

[0056] The sensor needs to have high accuracy to ensure that the measurement error of key parameters (such as temperature and current) is within an acceptable range, while maintaining reliability and stability during long-term operation.

[0057] The sensor needs to meet the stringent requirements of the switch cabinet operating environment, including dust and water resistance (such as IP65 or higher), high temperature resistance and anti-electromagnetic interference ability to ensure normal operation.

[0058] The sensor should have the ability to respond quickly, especially when the temperature or current changes drastically, and be able to output data in time to trigger the protection mechanism.

[0059] Sensors such as temperature and current can output standardized analog signals (such as 4-20mA or 0-10V) or digital signals (such as RS485, I2C interface), which is convenient for connection with subsequent data processing modules.

[0060] The multi-channel data acquisition module enables synchronous acquisition of multiple sensors to ensure the integrity of operating status data and environmental parameters.

[0061] According to the application scenario, select the appropriate wireless communication technology for data transmission, such as:

[0062] Wi-Fi: Suitable for short-distance, high-bandwidth demand scenarios, fast data transmission speed, and suitable for places with dense equipment.

[0063] NB-IoT: Low-power wide-area network technology suitable for scenarios requiring long-distance communication and limited power consumption, such as remote or power-constrained areas.

[0064] ZigBee: A low-power, low-cost short-range wireless communication technology suitable for monitoring scenarios with multiple devices in a network.

[0065] LoRa: Suitable for long-distance, low-data-volume transmission scenarios and suitable for regional power monitoring.

[0066] Data transmission adopts a hierarchical mechanism, setting high-priority data (such as abnormal values ​​of temperature and current) as real-time transmission, and low-priority data using batch transmission or periodic reporting to optimize the use of communication resources.

[0067] Edge computing nodes are usually deployed on-site near electrical switch cabinets to reduce the transmission delay from sensor data to computing nodes and improve the real-time performance of the system. Edge computing nodes are equipped with necessary hardware facilities, such as microprocessors, storage units, and communication interfaces, to ensure that they can operate independently and process multi-source sensor data.

[0068] Perform preliminary analysis on the collected data to identify and mark abnormal values. For example, by setting safety thresholds for parameters such as temperature and current, you can quickly find and filter out-of-limit data. Apply signal processing algorithms (such as Kalman filtering or median filtering) to denoise the collected data, reduce errors caused by environmental interference, and ensure the accuracy of uploaded data.

[0069] In the case of large amounts of raw data, the data is compressed to extract key information (such as mean, maximum, and minimum values) to reduce the transmission bandwidth requirements. For example, the temperature fluctuation data over a period of time is compressed into statistical characteristic values ​​for transmission. The data from different sensors is integrated to form a more comprehensive description of the equipment's operating status. For example, the possibility of equipment overload can be assessed by combining temperature and current data.

[0070] After the data is pre-processed at the edge node, the uploaded content is more concise, thereby reducing the data storage and processing pressure on the cloud platform and improving the overall system efficiency.

[0071] Edge nodes can quickly respond to changes in key data and trigger early warning signals when data is abnormal, without relying on subsequent processing on the cloud platform.

[0072] Data compression and priority management significantly reduce the size of uploaded data, thereby reducing communication resource usage and operating costs.

[0073] Edge computing nodes need to have a fault-tolerant mechanism. In the event of transmission interruption, they can temporarily save the collected data through local storage and continue uploading after communication is restored to ensure data integrity.

[0074] According to the scenario requirements, select the appropriate communication network for data upload, such as cellular network (4G / 5G), fiber broadband, or dedicated IoT protocol (NB-IoT, LoRa, etc.). These communication methods ensure stable data transmission over long distances and between multiple nodes.

[0075] High-priority data (such as anomaly detection results) is immediately sent to the cloud platform through a real-time upload mechanism to ensure timely response and processing. For low-priority data (such as regularly monitored environmental parameters), batch upload is used to reduce bandwidth usage. If an interruption occurs during the upload process, the system can record the upload progress and continue uploading after recovery to ensure data integrity.

[0076] The cloud platform uses a distributed storage system (such as HDFS or NoSQL database) to manage large amounts of data, supporting high concurrent access and large-scale storage requirements. Real-time data storage area: used to store the latest uploaded data, supporting real-time analysis and fast query. Historical data archiving area: regularly migrate real-time data to the historical data area for long-term storage and trend analysis.

[0077] The cloud platform stores data by category (such as temperature, current, voltage, humidity, etc.), and adds timestamps and device identifiers to facilitate subsequent retrieval and correlation analysis.

[0078] The cloud platform processes the stored data through big data processing frameworks (such as Hadoop and Spark), including data cleaning, aggregation and analysis, to ensure data quality and consistency.

[0079] Time series analysis: A time series model is constructed based on historical data to capture the time-varying patterns of the operating status of electrical switchgear, such as temperature trends, current load fluctuations, etc.

[0080] Anomaly detection model: Use machine learning algorithms (such as support vector machines or deep learning models) to train fault prediction models to identify potential operational anomalies and issue alerts in advance.

[0081] Correlation analysis model: Combines data from different sensors and discovers implicit relationships through multi-dimensional analysis, such as the correlation between increased humidity and decreased insulation performance.

[0082] The cloud platform provides visualization interfaces in the form of charts, heat maps, trend lines, etc., so that users can intuitively understand the operating status and historical trends of electrical switchgear. Real-time curve display of data such as temperature, current, humidity, etc. Trend change chart of historical data is used for analysis of long-term operating rules. Users can make precise inquiries based on time range, equipment number or parameter type, and quickly obtain the operating records and historical status of specific equipment.

[0083] Partial discharge is a micro-discharge phenomenon caused by the deterioration or defects of the insulation materials in the high-voltage switchgear. Although it does not directly lead to breakdown, it is an early signal of insulation degradation and potential failure. Monitoring the intensity of partial discharge can evaluate the insulation health of the switchgear and detect hidden dangers in advance.

[0084] Ultrasonic sensors or high-frequency current sensors are used to collect partial discharge signals in the switch cabinet and convert them into numerical parameters (such as discharge intensity or pulse frequency).

[0085] Intensity change: The gradual increase in the intensity of partial discharge indicates that the insulation material is deteriorating, which may lead to flashover or breakdown failure. Frequency distribution: High-frequency discharge signals often indicate greater risks, and the location and nature of the fault point can be determined through spectrum analysis.

[0086] Particles in the environment (such as dust, metal particles or salt spray) may accumulate inside the switch cabinet, negatively affecting the ventilation and electrical contact of the equipment, and even causing arc discharge. Monitoring particle concentration, especially in industrial areas or coastal areas, is an important indicator for evaluating the impact of the operating environment on equipment stability.

[0087] Use air quality sensors (such as laser particle sensors) to monitor the particle concentration inside and outside the switch cabinet (in μg / m³) and calibrate the data based on the status of the equipment ventilation system.

[0088] Particle Accumulation Trend: Continuously increasing particle concentrations indicate a deteriorating internal environment, which may result in reduced insulation or reduced heat dissipation capabilities of the equipment.

[0089] Particle type analysis: Detection of metallic particles or salt spray can indicate more serious risks of equipment damage, such as arcing or accelerated corrosion.

[0090] After analyzing the local discharge intensity in the high-voltage switchgear, a local discharge intensity abnormality index is generated. The method for obtaining the local discharge intensity abnormality index is as follows:

[0091] Collect the partial discharge signal x(t) and record it as a discrete time series x[n], where n=1,2,…,N; represents the sampling point, N is the total number of sampling points of the signal, and the sampling frequency is fs: Make sure that the sampling frequency meets the signal bandwidth requirements (according to the sampling theorem), such as fs>2×f highest frequency.

[0092] Commonly used wavelet bases include Daubechies (db), Symlets (sym), Coiflets (coif), etc. For partial discharge signals, db4 or sym4 usually works better.

[0093] The number of decomposition layers is J: The number of decomposition layers is selected according to the characteristics of the signal, satisfying ;

[0094] Calculate wavelet components: decompose the signal into approximate components (low-frequency parts) through wavelet transform and several detail components (high frequency parts) ;

[0095] Low frequency approximation component: ;in, is the scaled and translated wavelet basis function; high-frequency detail component: ;in, is the corresponding wavelet function.

[0096] For each wavelet component and Calculate its energy , the expression is: Total energy The calculation expression is: ; Calculate the proportion of each frequency band energy in the total energy: ;in and Reflect the contribution of signals in different frequency bands and extract the energy distribution under normal conditions from historical operation data .

[0097] By comparing the current energy distribution with the normal state distribution, the partial discharge intensity abnormality index is defined, and the expression is: ;in is the importance weight of each frequency band, which can be set based on the discharge characteristics experience. KML is the local discharge intensity abnormality index.

[0098] Based on normal For statistical distribution, the 3σ principle is usually adopted to set the abnormal threshold, that is, the abnormal threshold = μ + 3σ, where μ and σ are the mean and standard deviation of the abnormal index under normal conditions.

[0099] like If the value is less than or equal to the abnormal threshold, the partial discharge signal is considered to be in a normal state. If it is greater than the abnormal threshold, the partial discharge signal is judged to be abnormal, indicating possible insulation degradation or equipment defects.

[0100] After analyzing the abnormal changes in particle concentration inside and outside the switch cabinet, the particle concentration variation index is generated. The method for obtaining the particle concentration variation index is as follows:

[0101] The monitoring data of particle concentration is assumed to include multiple dimensions (such as time series, different measurement points, particle types, etc.), which is defined as a multidimensional data matrix X: ; where p is the number of samples and d is the feature dimension (such as internal concentration, external concentration, distribution of particle types, etc.). Calculate the covariance matrix of the data matrix X , the expression is: ; In the formula, is the eigenvector of the i-th sample, Σ is a d×d symmetric positive definite matrix, T is the matrix transpose, is the mean of each feature dimension.

[0102] For each sample Calculate the Mahalanobis distance from the center of the distribution , the expression is: ; If Σ is a positive definite matrix, its inverse matrix can be directly calculated , Is a scalar that indicates how far the ith sample deviates from the center of the distribution.

[0103] The Mahalanobis distance is normalized to the particle concentration variation index, and the expression is: ; Where MKN is the particle concentration variation index.

[0104] The particle concentration variation index and the local discharge intensity anomaly index are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as the input of the machine learning model. The machine learning model predicts the stability value label of the switchgear operating state with each group of comprehensive feature vectors as the prediction target, and minimizes the sum of prediction errors of the stability value labels of all switchgear operating states as the training target. The machine learning model is trained until the sum of prediction errors converges, and the model training is stopped. The stability value of the switchgear operating state is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

[0105] The obtained stability value of the switch cabinet operating status is compared with the stability reference threshold value pre-set according to historical data. If the stability value of the switch cabinet operating status is greater than or equal to the stability reference threshold value, it means that the stability of the switch cabinet operating status is high and the probability of failure is low; if the stability value of the switch cabinet operating status is less than the stability reference threshold value, it means that the stability of the switch cabinet operating status is low and the probability of failure is high.

[0106] S5: Edge computing nodes do not need to rely on the cloud platform when communication is interrupted. They can independently complete data collection, storage and analysis to ensure uninterrupted monitoring of the operating status of the electrical switchgear.

[0107] The sensor continues to transmit the collected temperature, humidity, current, voltage and gas concentration data to the edge node to ensure real-time perception of the equipment's operating status. The edge node uses built-in algorithms (such as filtering, denoising and anomaly detection) to pre-process the collected data, filter key information, and reduce storage and computing burdens. For key parameters (such as temperature or current anomalies), the edge node can independently trigger alarms or control instructions (such as emergency power off) to reduce the risk of communication interruption.

[0108] Edge computing nodes need to be equipped with local storage units (such as solid-state drives, SD cards) to temporarily store collected data during communication interruptions. Use a timestamp-marked data storage structure, including sampling time, parameter type, and sensor location. To save storage space, compress the data, for example, only record changes in key parameters or abnormal values. Use a circular buffer to store data, and overwrite the earliest data when the storage space is full to ensure that the latest data is retained first. Abnormal values ​​and alarm events are stored first, and protection marks are set to prevent them from being overwritten.

[0109] During communication interruptions, edge nodes use built-in algorithms to analyze and evaluate data, supporting the following functions:

[0110] Monitor key parameters (such as temperature and current) to see if they exceed set thresholds and generate alarms based on the results. Predict possible risks (such as continued temperature increases) through short-term trend analysis. When a serious anomaly is detected, the edge node can perform the following actions:

[0111] Local alarm: Notify operation and maintenance personnel through sound and light signals or text messages.

[0112] Execution control: Sending instructions to the execution unit of the switch cabinet (such as powering off or switching to backup power supply).

[0113] After communication is restored, the edge node will upload the data stored during the communication interruption to the cloud platform in order to ensure data integrity. Use markers to record the location of uploaded data to avoid duplicate uploads or data loss.

[0114] Priority management: High priority: abnormal event data and key parameter data are uploaded first. Low priority: normal operation data is uploaded in batches to reduce bandwidth usage.

[0115] After receiving the data, the cloud platform confirms the data integrity through a verification mechanism (such as hash verification or timestamp comparison).

[0116] After communication is restored, the edge node will upload the analysis results during the communication interruption to the cloud platform, including anomaly detection records, trend prediction results, and alarm event logs.

[0117] After receiving the data, the cloud platform can combine historical data and more complex analysis algorithms to review the equipment status during the outage and further confirm the abnormal causes and trends.

[0118] S6: Users can view the operating status of the electrical switchgear and potential fault prediction results in real time through the cloud platform, and remotely perform emergency shutdown or equipment scheduling operations based on the early warning information.

[0119] Users can view the operating status of electrical switchgear in real time through the cloud platform. The operating status data includes temperature, humidity, current, voltage and gas concentration;

[0120] The operating status data and the stability value of the switchgear operating status are used as the input items of the fuzzy logic, and the warning level is used as the output item of the fuzzy logic;

[0121] The operating status data and stability values ​​are converted into fuzzy sets and expressed by linguistic variables.

[0122] Input items and their language variables: Temperature (T): low, medium, high. Humidity (H): dry, normal, wet. Current (I): normal, slight overload, severe overload. Voltage (V): undervoltage, normal, overvoltage. Gas concentration (G): non-hazardous, slight leakage, severe leakage. Stability value (S): stable, metastable, unstable.

[0123] Use triangular or trapezoidal membership functions to describe the fuzzy sets for each variable. For example: For temperature T: low: ; Medium and High are defined in a similar manner.

[0124] Based on domain knowledge or historical data, fuzzy rules are defined to map input variables to output warning levels.

[0125] Fuzzy rule example:

[0126] If T is high, I is severely overloaded, and G is severely leaking, the warning level is high. If H is wet, V is overpressured, and S is unstable, the warning level is medium. If T is low and S is stable, the warning level is low.

[0127] The fuzzy rule base needs to cover all possible input combinations, and usually the number of rules is the Cartesian product of the number of fuzzy sets of input variables.

[0128] Based on the input data and rule base, the output fuzzy set is calculated using fuzzy reasoning methods (such as Mamdani reasoning method).

[0129] Reasoning process: Calculate the membership value of the input variable for each rule.

[0130] The minimum method is used to calculate the composite membership of the input variables of each rule.

[0131] Aggregate the outputs of all rules to obtain the output fuzzy set.

[0132] Convert the output fuzzy set into a clear warning level value. Common methods include:

[0133] Calculate the centroid of the fuzzy set as the output value, the expression is: ;in, is the membership function of the output fuzzy set, The warning level value.

[0134] If 0< ΅ά0.4 is a low risk level, no action is required, regular inspection is sufficient; if 0.4< <0.8, it is a medium risk level, it is recommended to arrange maintenance and check the equipment status; if 0.8< <1.0 indicates a high risk level, which immediately triggers an emergency shutdown or dispatches backup equipment.

[0135] Based on the warning level, users send remote commands through the cloud platform to execute equipment scheduling or shutdown operations.

[0136] In this embodiment, by installing temperature, humidity, current, voltage and gas concentration sensors inside or outside the switch cabinet, the equipment operation status data and environmental parameters are collected in real time. The collected data is transmitted to the edge computing node through wireless communication technology for preprocessing to reduce the communication load, and then uploaded to the cloud platform through the communication network, stored and archived in the cloud, and historical data and trend analysis models are formed. Based on the big data analysis and artificial intelligence algorithms of the cloud platform, the data features are deeply mined to predict potential equipment failures. In the event of communication interruption, the edge computing node runs independently, continues to store and analyze data locally, ensures uninterrupted monitoring, and synchronizes data to the cloud platform after communication is restored. Users can view the equipment operation status and fault prediction results in real time through the cloud platform, remotely perform emergency shutdown or equipment scheduling operations based on early warning information, and realize intelligent management of equipment operation.

[0137] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0138] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0139] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there may be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone, where A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the objects associated with each other are in an "or" relationship, but it may also indicate an "and / or" relationship, which can be understood by referring to the context. A person of ordinary skill in the art can appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in this article can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0140] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. An electrical switch cabinet detection method based on the Internet of Things, characterized in that: The following steps are involved: S1: Install several sensors inside or outside the electrical switch cabinet to collect operating status data and environmental parameters including temperature, humidity, current, voltage and gas concentration; S2: Use wireless communication technology to transmit the collected data to the edge computing node, pre-process the data, and reduce the transmission load; S3: The pre-processed data is uploaded to the cloud platform through the communication network. The cloud platform stores and archives the data to form the historical data and trend analysis model of the electrical switchgear operation; S4: Based on the big data analysis and artificial intelligence algorithms of the cloud platform, the collected data is deeply mined to predict potential faults; After analyzing the local discharge intensity in the high-voltage switchgear, a local discharge intensity abnormality index is generated. The method for obtaining the local discharge intensity abnormality index is as follows: Collect the partial discharge signal x(t) and record it as a discrete time series x[n], where n=1,2,…,N; represents the sampling point, N is the total number of sampling points of the signal, the sampling frequency is fs, the number of decomposition layers is J, and the wavelet components are calculated: the signal is decomposed into approximate components by wavelet transform and some details ; Low frequency approximation component: ;in, is the scaled and translated wavelet basis function; high-frequency detail component: ;in, is the corresponding wavelet function; For each wavelet component and Calculate its energy , , the expression is: Total energy The calculation expression is: ; Calculate the proportion of each frequency band energy in the total energy: ;in and Reflect the contribution of signals in different frequency bands and extract the energy distribution under normal conditions from historical operation data ; By comparing the current energy distribution with the normal state distribution, the partial discharge intensity abnormality index is defined, and the expression is: ;in is the importance weight of each frequency band, which can be set based on the discharge characteristics experience, and KML is the abnormal index of local discharge intensity; S5: In the event of communication interruption, the edge computing node operates independently, continues to monitor the device status through local storage and analysis functions, and uploads the data to the cloud platform after communication is restored; S6: Users can view the operating status of the electrical switchgear and potential fault prediction results in real time through the cloud platform, and remotely perform emergency shutdown or equipment scheduling operations based on the early warning information.

2. The electrical switch cabinet detection method based on the Internet of Things according to claim 1 is characterized in that: In S4, the particle concentration variation index is generated after analyzing the abnormal changes in the particle concentration inside and outside the switch cabinet. The particle concentration variation index is obtained by: The monitoring data of particle concentration is set to include multiple dimensions, which are defined as a multidimensional data matrix X: ; Where p is the number of samples, d is the feature dimension, and the covariance matrix of the data matrix X is calculated , the expression is: ; In the formula, is the eigenvector of the i-th sample, Σ is a d×d symmetric positive definite matrix, T is the matrix transpose, is the mean of each feature dimension; For each sample Calculate the Mahalanobis distance from the center of the distribution , the expression is: ; If Σ is a positive definite matrix, then directly calculate its inverse matrix , Is a scalar, indicating the degree to which the i-th sample deviates from the center of the distribution; The Mahalanobis distance is normalized to the particle concentration variation index, and the expression is: ; Where MKN is the particle concentration variation index.

3. The electrical switch cabinet detection method based on the Internet of Things according to claim 2 is characterized in that: In S4, the particle concentration variation index and the local discharge intensity anomaly index are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as inputs of the machine learning model. The machine learning model predicts the stability value label of the switch cabinet operating state with each group of comprehensive feature vectors as the prediction target, and minimizes the sum of prediction errors of the stability value labels of all switch cabinet operating states as the training target. The machine learning model is trained until the sum of prediction errors converges, and the model training is stopped. The stability value of the switch cabinet operating state is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

4. The electrical switch cabinet detection method based on the Internet of Things according to claim 3 is characterized in that: The obtained stability value of the switch cabinet operation state is compared with the stability reference threshold value pre-set according to historical data. If the stability value of the switch cabinet operation state is greater than or equal to the stability reference threshold value, it means that the stability of the switch cabinet operation state is high and the probability of failure is low. If the stability value of the switch cabinet operation state is less than the stability reference threshold, it means that the stability of the switch cabinet operation state is low and the probability of failure is high.

5. The electrical switch cabinet detection method based on the Internet of Things according to claim 4 is characterized in that: In S6, users can view the operating status of the electrical switchgear in real time through the cloud platform. The operating status data includes temperature, humidity, current, voltage and gas concentration; The operating status data and the stability value of the switchgear operating status are used as the input items of the fuzzy logic, and the warning level is used as the output item of the fuzzy logic; The operating status data and stability values ​​are converted into fuzzy sets and expressed by linguistic variables; Use triangular or trapezoidal membership functions to describe the fuzzy sets of each variable; define fuzzy rules based on domain knowledge or historical data to map input variables to output warning levels; The fuzzy rule base needs to cover all input combinations, and the number of rules is the Cartesian product of the number of fuzzy sets of the input variables; According to the input data and rule base, the output fuzzy set is calculated using fuzzy reasoning method; Calculate the membership value of the input variable for each rule; Use the minimum method to calculate the composite membership of the input variables of each rule; Aggregate the outputs of all rules to obtain the output fuzzy set; The output fuzzy set is converted into a clear warning level value, and the center of gravity of the fuzzy set is calculated as the output value. The expression is: ;in, is the membership function of the output fuzzy set, The warning level value.

6. The electrical switch cabinet detection method based on the Internet of Things according to claim 5 is characterized in that: If 0< ά0.4 is a low risk level, no action is required, regular inspection is sufficient; if 0.4< <0.8, it is a medium risk level, it is recommended to arrange maintenance and check the equipment status; if 0.8< <1.0 indicates a high risk level, which immediately triggers an emergency shutdown or dispatches backup equipment.

Images