Low-voltage cabinet fault remote monitoring method based on industrial internet of things

Abstract

The application relates to the technical field of industrial Internet of Things, and provides a low-voltage cabinet fault remote monitoring method based on industrial Internet of Things.The method comprises the following steps: based on industrial Internet of Things, a plurality of low-voltage distribution cabinets of a power system are monitored in real time to obtain a monitoring matrix; electrical, thermal runaway and aging risk triggering analysis is carried out according to the monitoring matrix to construct an electrical, thermal runaway and aging risk feature matrix; multi-dimensional fault analysis is carried out on the low-voltage distribution cabinet according to the feature matrix to generate a fault first report; and topological linkage fault prediction is carried out according to the fault first report to generate a fault second report.The application solves the technical problem that the traditional low-voltage cabinet fault monitoring method usually analyzes single risk factors such as electrical faults, thermal runaway or equipment aging in isolation, lacks comprehensive correlation analysis of multi-dimensional risk data, and results in incomplete fault diagnosis and poor early warning timeliness, and achieves the technical effects of improving the comprehensiveness of fault diagnosis and the timeliness of prediction.

Description

Technical Field

[0001] This application relates to the field of industrial Internet of Things (IoT) technology, specifically to a method for remote monitoring of low-voltage switchgear faults based on industrial IoT. Background Technology

[0002] In power systems, low-voltage switchgear, as the core equipment for power distribution and control, directly affects the safety and efficiency of the entire power supply network. With the rapid development of Industrial Internet of Things (IIoT) technology, equipment intelligence and data interconnection have become important trends in the industrial sector. However, traditional low-voltage switchgear fault monitoring methods still have significant limitations: on the one hand, most monitoring methods rely on manual periodic inspections or single parameter threshold alarms, making it difficult to capture the dynamic correlation of multi-dimensional risk factors such as electrical parameter fluctuations, equipment temperature changes, and insulation aging in real time; on the other hand, existing methods often analyze single risk types in isolation, such as electrical faults (e.g., short circuits, overloads), thermal runaway risks (e.g., contact point overheating, poor heat dissipation), or aging failures (e.g., insulation material deterioration, mechanical component wear), lacking the ability to comprehensively integrate and correlate multi-source risk data, resulting in incomplete fault diagnosis coverage and poor early warning timeliness. Furthermore, low-voltage switchgear typically forms a power supply network through interconnected topologies; local faults may trigger chain reactions through electrical connections or thermal conduction, but traditional monitoring does not consider the topological linkages between devices, making it difficult to predict fault propagation paths and potential impact ranges in advance. Summary of the Invention

[0003] This application provides a remote monitoring method for low-voltage switchgear faults based on the Industrial Internet of Things (IIoT). It aims to solve the technical problems of traditional low-voltage switchgear fault monitoring methods, which typically analyze single risk factors such as electrical faults, thermal runaway, or equipment aging in isolation, lacking comprehensive correlation analysis of multi-dimensional risk data. This results in incomplete fault diagnosis coverage and poor early warning timeliness. The method achieves the technical effect of improving the comprehensiveness of fault diagnosis and the timeliness of prediction by integrating multi-source monitoring data in real time through the IIoT, constructing a three-dimensional risk feature matrix of electrical, thermal runaway, and aging, and then predicting fault chain effects based on topological relationships.

[0004] This application provides a method for remote monitoring of low-voltage switchgear faults based on the Industrial Internet of Things (IIoT). The method includes: real-time monitoring of multiple low-voltage switchgear in a power system based on the IIoT to obtain multiple low-voltage switchgear monitoring matrices; electrical risk trigger analysis based on the multiple low-voltage switchgear monitoring matrices to construct multiple electrical risk feature matrices; thermal runaway risk trigger analysis based on the multiple low-voltage switchgear monitoring matrices to construct multiple thermal runaway risk feature matrices; aging risk trigger analysis based on the multiple low-voltage switchgear monitoring matrices to construct multiple aging risk feature matrices; multi-dimensional fault analysis of the multiple low-voltage switchgear based on the multiple electrical risk feature matrices, the multiple thermal runaway risk feature matrices, and the multiple aging risk feature matrices to generate a first low-voltage switchgear fault report; and topology linkage fault prediction based on the first low-voltage switchgear fault report to generate a second low-voltage switchgear fault report.

[0005] One or more technical solutions provided in this application have at least the following technical effects or advantages: The aforementioned remote fault monitoring method for low-voltage switchgear based on the Industrial Internet of Things (IIoT) first collects and generates a monitoring data matrix by real-time monitoring of multiple low-voltage switchgear in the power system, leveraging IIoT technology. Then, electrical risk, thermal runaway risk, and aging risk are analyzed from these data matrices to generate corresponding risk feature matrices. Next, multi-dimensional fault analysis is performed using these risk feature matrices to generate a first fault report. Finally, topology-linked fault prediction is performed based on this report to generate a second fault report, thereby achieving comprehensive fault early warning and risk management for low-voltage switchgear.

[0006] The above description is merely an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, specific embodiments of this application are given below. Attached Figure Description

[0007] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0008] Figure 1 This is a flowchart illustrating a method for remote monitoring of low-voltage switchgear faults based on the Industrial Internet of Things (IIoT) in one embodiment.

[0009] Figure 2This is a flowchart illustrating the process of obtaining the m-th electrical risk feature matrix in a remote monitoring method for low-voltage switchgear faults based on the Industrial Internet of Things in one embodiment. Detailed Implementation

[0010] This application provides a remote monitoring method for low-voltage switchgear faults based on the Industrial Internet of Things (IIoT). This addresses the technical problem that traditional low-voltage switchgear fault monitoring methods typically analyze single risk factors such as electrical faults, thermal runaway, or equipment aging in isolation, lacking comprehensive correlation analysis of multi-dimensional risk data. This results in incomplete fault diagnosis coverage and poor early warning timeliness. The method achieves the technical effect of improving the comprehensiveness of fault diagnosis and the timeliness of prediction by integrating multi-source monitoring data in real time through the IIoT, constructing a three-dimensional risk feature matrix of electrical, thermal runaway, and aging, and then predicting fault chain effects based on topological relationships.

[0011] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0012] It should be noted that the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to such process, method, product, or device.

[0013] Examples, such as Figure 1 As shown, this application provides a method for remote monitoring of low-voltage switchgear faults based on the Industrial Internet of Things (IIoT), the method comprising: Based on the Industrial Internet of Things (IIoT), multiple low-voltage switchgear cabinets in the power system are monitored in real time to obtain a monitoring matrix of multiple low-voltage switchgear cabinets.

[0014] In this embodiment, various operational data within low-voltage switchgear are first collected in real time using a monitoring matrix deployed in the power system. This monitoring matrix includes multiple sensors and intelligent devices, and the collected data includes parameters such as current, voltage, temperature, humidity, power, and frequency. Subsequently, the collected operational data is transmitted to a data processing unit via the real-time transmission function of the Industrial Internet of Things (IIoT). The data is then arranged and integrated according to the configuration of the monitoring matrix, forming multiple low-voltage switchgear monitoring matrices. During the monitoring process, the data processing unit continuously updates these matrices based on the real-time operational data transmitted from the IIoT, ensuring data real-time performance and accuracy, and providing detailed foundational information for subsequent risk analysis and fault prediction.

[0015] Furthermore, based on the Industrial Internet of Things (IIoT), multiple low-voltage switchgear cabinets in the power system are monitored in real time to obtain multiple low-voltage switchgear monitoring matrices, including: Based on real-time monitoring of the multiple low-voltage distribution cabinets using the Industrial Internet of Things, multiple low-voltage cabinet monitoring sets are obtained; the multiple low-voltage cabinet monitoring sets are cleaned and organized to generate the multiple low-voltage cabinet monitoring matrix.

[0016] Preferably, a monitoring matrix composed of various sensors and intelligent devices is deployed within each low-voltage distribution cabinet. Through Industrial Internet of Things (IIoT) technology, the sensors in the monitoring matrix monitor various operating parameters of the low-voltage distribution cabinet in real time and transmit the collected monitoring data for each cabinet to the data processing unit via a wireless network. Subsequently, in the data processing unit, each low-voltage cabinet monitoring set is cleaned by subtracting multiple standard deviations (e.g., three) from the mean of each operating parameter, removing outlier data points exceeding or falling below this range. Missing data is then filled in using the mean of multiple adjacent data points, either through interpolation or by using algorithms to calculate the missing values. Afterward, the cleaned low-voltage cabinet monitoring sets are organized and structured according to the arrangement of monitoring nodes in the monitoring matrix, forming multiple low-voltage cabinet monitoring matrices. Each low-voltage cabinet monitoring matrix represents a complete monitoring dataset for a low-voltage distribution cabinet, containing real-time data of all key monitoring parameters for that cabinet. These matrices are periodically updated on the data processing unit to reflect the dynamic operating status of the distribution cabinet, providing basic data for subsequent fault diagnosis.

[0017] Furthermore, this application provides a method based on the Industrial Internet of Things (IIoT) to retrieve the status data of each sensor and to perform adaptive sensor anomaly optimization on the multiple low-voltage switchgear monitoring sets based on the status data of each sensor.

[0018] Optionally, to ensure the accuracy and reliability of multiple low-voltage cabinet monitoring sets collected through the Industrial Internet of Things, the status data of each sensor is first retrieved, including electrical parameters (such as sensor operating voltage and ambient temperature) and communication parameters (such as packet loss rate and acquisition delay). These status data are then compared with the corresponding thresholds. Typically, the threshold for sensor operating voltage is ±15% of the sensor's rated operating voltage, the threshold for ambient temperature is the upper limit of the sensor specifications, the packet loss rate is 5%, and the acquisition delay is 200ms. If any status data does not meet the corresponding threshold, it indicates that the current sensor is abnormal. At this time, it will be determined whether there is a redundant monitoring matrix. If so, the redundant monitoring matrix will be used to re-acquire data, and the original low-voltage switchgear monitoring set will be replaced with the re-acquired low-voltage switchgear monitoring set, thereby optimizing the previously acquired sensor data. Conversely, if there is no usable redundant monitoring matrix, Kalman filtering or moving average algorithms will be used to suppress noise and compensate for bias in the original monitoring data. For example, if the ambient temperature exceeds the threshold, the data curve will be smoothed through filtering algorithms to retain the true temperature change trend, thereby forming a new low-voltage switchgear monitoring set. This avoids false fault judgments or warning failures caused by sensor failures or data anomalies, thereby improving the safety and reliability of the power system.

[0019] Based on the multiple low-voltage switchgear monitoring matrices, electrical risk triggering analysis is performed to construct multiple electrical risk feature matrices.

[0020] In one embodiment, after obtaining multiple low-voltage switchgear monitoring matrices, an electrical risk trigger detection tree constructed based on electrical risk trigger records is used to analyze the low-voltage switchgear monitoring matrices layer by layer to determine which data points meet the conditions for triggering electrical risks, thereby identifying multiple electrical risk trigger coefficients. Based on these trigger coefficients, electrical risk characteristics are identified in the monitoring matrices of the low-voltage switchgear, generating an electrical risk characteristic matrix. This electrical risk characteristic matrix reflects the electrical risk status of each low-voltage switchgear, providing key data support for subsequent fault diagnosis.

[0021] Furthermore, such as Figure 2 As shown, this application provides an electrical risk trigger analysis based on the multiple low-voltage switchgear monitoring matrices, constructing multiple electrical risk feature matrices, including: Based on the multiple low-voltage switchgear monitoring matrices, extract the m-th low-voltage switchgear monitoring matrix corresponding to the m-th low-voltage switchgear, where m is a positive integer; use the specification and model information of the m-th low-voltage switchgear as the m-th specification and model feature, and perform electrical risk trigger record retrieval based on the m-th specification and model feature to obtain the m-th electrical risk trigger record set; construct the m-th electrical risk trigger detection tree based on the m-th electrical risk trigger record set; input each monitoring parameter in the m-th low-voltage switchgear monitoring matrix into the m-th electrical risk trigger detection tree to obtain multiple electrical risk trigger coefficients; based on the multiple electrical risk trigger coefficients, perform electrical risk feature identification on the m-th low-voltage switchgear monitoring matrix according to the electrical risk trigger threshold to obtain the m-th electrical risk feature matrix.

[0022] Preferably, firstly, the monitoring matrix of the m-th low-voltage switchgear is extracted from multiple low-voltage switchgear monitoring matrices, where m is a positive integer representing the specific low-voltage switchgear serial number. This m-th low-voltage switchgear monitoring matrix contains the sensor data of each low-voltage switchgear. Then, the specification and model information of the m-th low-voltage switchgear is used as the m-th specification and model feature, such as the hardware characteristics of the low-voltage switchgear (rated current, rated voltage, rated power, etc.). This specification and model feature is then used to find and obtain the electrical risk trigger record set related to the m-th low-voltage switchgear. This m-th electrical risk trigger record set contains previously occurring electrical risk events and their corresponding triggering conditions, providing historical data support for subsequent risk assessment and analysis. Next, based on the obtained m-th electrical risk trigger record set, an electrical risk trigger detection tree is constructed. This detection tree is a decision structure used to perform layer-by-layer judgment and analysis of each electrical risk based on the monitoring data. Each node corresponds to an electrical risk triggering condition. By calculating the electrical risk confidence value, potential electrical risks can be identified. Then, each monitoring parameter in the m-th low-voltage switchgear monitoring matrix is ​​input into the m-th electrical risk trigger detection tree for analysis. Each monitoring parameter generates a corresponding electrical risk confidence value based on the detection tree. These electrical risk confidence values ​​serve as electrical risk trigger coefficients, characterizing the likelihood that the parameter will trigger an electrical risk. The larger the electrical risk trigger coefficient, the higher the probability that the parameter's anomaly may lead to an electrical risk. Finally, multiple electrical risk trigger coefficients are compared with preset electrical risk trigger thresholds. If the electrical risk trigger coefficient of a monitoring parameter is greater than or equal to the set threshold, the parameter is considered to have a high electrical risk. In this case, the parameter is added to the m-th electrical risk feature matrix. Thus, the generated m-th electrical risk feature matrix will contain all monitoring parameters with a high electrical risk. These parameters will serve as important bases for further fault diagnosis, risk prediction, and maintenance decisions, helping to predict and prevent potential electrical faults and ensure the stable operation of the power system.

[0023] Furthermore, this application provides a method for constructing an electrical risk triggering detection tree based on the m-th electrical risk triggering record set, including: The m-th electrical risk trigger record set includes a monitoring sample set and an electrical risk trigger sample set; based on the monitoring sample set, the electrical risk trigger sample set is clustered to obtain each electrical risk trigger cluster corresponding to each monitoring sample; the mean of each electrical risk trigger cluster is calculated to obtain each electrical risk confidence value; the m-th electrical risk trigger detection tree is constructed using each monitoring sample as multiple input nodes and each electrical risk confidence value as multiple output nodes.

[0024] Optionally, the m-th electrical risk triggering record set matched based on the specifications and model information of the m-th low-voltage distribution cabinet includes a monitoring sample set and an electrical risk triggering sample set. Each monitoring sample set consists of monitoring data collected by all sensors of the distribution cabinet within a certain time period. This data typically includes parameters such as current, voltage, temperature, and frequency. Each monitoring sample set corresponds to an electrical risk triggering sample set, which records electrical risk triggering events within that time period, including whether there are electrical fault events such as overload, short circuit, or abnormal voltage. Subsequently, cluster analysis is performed on the monitoring sample set and the electrical risk triggering sample set. In this process, the elbow method is used to plot the sum of squared errors corresponding to different K values, and the K value at the elbow is selected as the optimal number of clusters. Then, K monitoring samples are randomly selected as the initial cluster centers. Next, for each monitoring sample, its Euclidean distance to the K cluster centers is calculated, and it is assigned to the cluster with the smallest distance. At this point, each cluster contains several monitoring samples. Once all monitoring samples have been assigned to their respective clusters, the mean of each cluster is calculated, and this mean is used as the new cluster center. The distance between the new cluster center and the monitoring samples is then recalculated, and the samples are reassigned to the nearest cluster, updating the cluster centers. This process iterates until the cluster centers no longer change significantly or the preset number of iterations is reached. After clustering convergence, all monitoring samples are divided into K clusters. The monitoring samples within each cluster have similar characteristics, reflecting a certain common operating mode of the distribution cabinet. Then, according to the correspondence between monitoring samples and electrical risk triggering samples, the electrical risk triggering samples corresponding to the monitoring samples belonging to the same cluster are aggregated to form electrical risk triggering clusters corresponding to each monitoring sample. Then, the mean of each electrical risk triggering cluster is calculated to obtain the electrical risk confidence value of each cluster. This electrical risk confidence value reflects the probability of electrical risk occurring within the cluster; the larger the value, the higher the electrical risk corresponding to the cluster. Finally, using the monitored samples as input nodes and each electrical risk confidence value as an output node, the m-th electrical risk trigger detection tree is constructed. This electrical risk trigger detection tree is a decision tree structure used to assess the electrical risk of each monitored sample. Through this tree, risk identification can be effectively performed, improving the fault early warning capability of the power system and ensuring its safe and stable operation.

[0025] Based on the multiple low-voltage switchgear monitoring matrices, thermal runaway risk triggering analysis is performed, and multiple thermal runaway risk feature matrices are constructed.

[0026] In one embodiment, a thermal runaway risk triggering detection tree, constructed based on thermal runaway risk triggering records, is used to analyze the monitoring matrix of each low-voltage switchgear layer by layer. By analyzing temperature, load, and other relevant parameters in the low-voltage switchgear monitoring matrix, it is possible to identify which data points meet the conditions for triggering thermal runaway risk, thereby generating multiple thermal runaway risk triggering coefficients. Each thermal runaway risk triggering coefficient characterizes the probability that the parameter will trigger thermal runaway risk. Based on these thermal runaway risk triggering coefficients, thermal runaway risk characteristics are identified in the monitoring matrix of the low-voltage switchgear, identifying which monitoring parameters require special attention in the event of thermal runaway. Ultimately, the generated thermal runaway risk feature matrix reflects the thermal runaway risk status of each low-voltage switchgear, providing data support for subsequent fault diagnosis, helping to monitor and predict potential thermal runaway problems in real time, thereby ensuring the safe and stable operation of the power system.

[0027] Furthermore, this application provides a method for analyzing thermal runaway risk triggering based on the aforementioned multiple low-voltage switchgear monitoring matrices, and constructing multiple thermal runaway risk feature matrices, including: Based on the characteristics of the m-th specification model, thermal runaway risk trigger records are retrieved to obtain the m-th thermal runaway risk trigger record set; based on the m-th thermal runaway risk trigger record set, the m-th thermal runaway risk trigger detection tree is constructed; each monitoring parameter in the m-th low-voltage switchgear monitoring matrix is ​​input into the m-th thermal runaway risk trigger detection tree to obtain multiple thermal runaway risk trigger coefficients; based on the multiple thermal runaway risk trigger coefficients, thermal runaway risk features are identified in the m-th low-voltage switchgear monitoring matrix according to the thermal runaway risk trigger threshold to obtain the m-th thermal runaway risk feature matrix.

[0028] Preferably, for the thermal runaway risk triggering analysis process of each low-voltage switchgear, the process first retrieves thermal runaway risk triggering records based on the m-th specification and model feature, obtaining the thermal runaway risk triggering record set for the m-th low-voltage switchgear. These records reflect potential thermal runaway events that may be triggered by the switchgear under similar configurations and conditions, including overheating, poor heat dissipation, and other issues. Then, using the same method as constructing the m-th electrical risk triggering detection tree, a m-th thermal runaway risk triggering detection tree is constructed based on the m-th thermal runaway risk triggering record set. This tree is used to perform layer-by-layer judgment on each monitoring parameter to confirm whether it meets the conditions for triggering thermal runaway. Next, each monitoring parameter in the m-th low-voltage switchgear monitoring matrix is ​​input into the m-th thermal runaway risk triggering detection tree for analysis. Each monitoring parameter passes through each node of the tree, and a thermal runaway risk triggering coefficient is calculated according to the set rules. The thermal runaway risk triggering coefficient represents the probability or severity of the monitoring parameter triggering thermal runaway risk; the larger the coefficient value, the higher the risk of thermal runaway for that parameter. Finally, based on multiple thermal runaway risk triggering coefficients, a comparison is made with the thermal runaway risk triggering threshold. If the triggering coefficient of a certain monitoring parameter is greater than or equal to the preset triggering threshold, it indicates that the parameter has a high risk of thermal runaway. The parameter will then be added to the m-th thermal runaway risk feature matrix. This m-th thermal runaway risk feature matrix reflects the risk status of each low-voltage switchgear in terms of thermal runaway, providing an important basis for subsequent fault diagnosis and helping to predict and prevent thermal runaway problems that may occur in the operation of low-voltage switchgear.

[0029] Based on the multiple low-voltage switchgear monitoring matrices, aging risk trigger analysis is performed to construct multiple aging risk feature matrices.

[0030] In one embodiment, an aging risk trigger detection tree, constructed based on aging risk trigger records, is used to analyze the monitoring matrix of each low-voltage switchgear layer by layer. By analyzing temperature, load cycle, operating duration, and other relevant parameters in the low-voltage switchgear monitoring matrix, it is possible to identify which data points meet the conditions for triggering aging risks, thereby generating multiple aging risk trigger coefficients. Each aging risk trigger coefficient represents the probability that the parameter will trigger an aging risk. Based on these aging risk trigger coefficients, aging risk characteristics are identified in the monitoring matrix of the low-voltage switchgear, identifying which monitoring parameters require special attention under aging conditions. Ultimately, the generated aging risk feature matrix reflects the risk status of each low-voltage switchgear in terms of thermal runaway, providing data support for subsequent fault diagnosis, helping to monitor and predict potential aging problems in equipment in real time, thereby ensuring the safe and stable operation of the power system.

[0031] Based on the multiple electrical risk feature matrices, the multiple thermal runaway risk feature matrices, and the multiple aging risk feature matrices, a multi-dimensional fault analysis is performed on the multiple low-voltage switchgear to generate a first fault report for the low-voltage switchgear.

[0032] In one embodiment, after obtaining multiple electrical risk feature matrices, multiple thermal runaway risk feature matrices, and multiple aging risk feature matrices, the corresponding fault analysis model is used to analyze these risk feature matrices to identify potential fault problems in each low-voltage switchgear, thereby generating a first low-voltage switchgear fault report. This report details the potential fault risks that each switchgear may face, helping maintenance personnel to accurately identify and locate problems, and providing key support for the stable operation of the power system.

[0033] Furthermore, this application provides a method for performing multi-dimensional fault analysis on the multiple low-voltage switchgear based on the multiple electrical risk feature matrices, the multiple thermal runaway risk feature matrices, and the multiple aging risk feature matrices, generating a first low-voltage switchgear fault report, including: Based on the multiple electrical risk feature matrices, electrical fault analysis is performed to obtain multiple electrical fault analysis results; based on the multiple thermal runaway risk feature matrices, thermal runaway fault analysis is performed to obtain multiple thermal runaway fault analysis results; based on the multiple aging risk feature matrices, aging fault analysis is performed to obtain multiple aging fault analysis results; based on the multiple low-voltage switchgear, the multiple electrical fault analysis results, the multiple thermal runaway fault analysis results, and the multiple aging fault analysis results are clustered and sorted to obtain the first fault report of the low-voltage switchgear.

[0034] Preferably, during multi-dimensional fault analysis, multiple electrical risk feature matrices are sequentially input into a pre-built electrical fault analysis model. The model then analyzes the received data, generating multiple fault analysis results. These results describe the type, probability of occurrence, and severity of the electrical fault, providing foundational data for fault investigation and risk assessment. Furthermore, multiple thermal runaway risk feature matrices and multiple aging risk feature matrices are input into their respective thermal runaway fault analysis models and aging fault analysis models, generating multiple thermal runaway fault analysis results and multiple aging fault analysis results. Finally, a clustering method similar to the one described above is used to cluster and organize the electrical fault analysis results, thermal runaway fault analysis results, and aging fault analysis results of multiple low-voltage switchgear. Cluster analysis can be used to categorize these fault analysis results into different categories or types based on their similarity, thereby compiling a comprehensive first report on low-voltage switchgear faults. This report details the electrical, thermal runaway, and aging-related fault risks that each low-voltage switchgear may face, providing a clear basis for subsequent fault diagnosis, maintenance decisions, and risk management. It helps maintenance personnel take timely measures to ensure the safe and stable operation of the power system.

[0035] Furthermore, this application provides methods for analyzing electrical faults based on the aforementioned multiple electrical risk feature matrices to obtain multiple electrical fault analysis results, including: Based on the characteristics of the m-th specification model, retrieve the electrical fault record set of the m-th low-voltage switchgear; perform supervised learning on Q learners based on the electrical fault record set of the m-th low-voltage switchgear to obtain Q electrical fault parsers, where Q is a positive integer greater than 1; perform distillation training on the Q electrical fault parsers, and obtain the distillation loss coefficient after each predetermined number of training iterations; if the distillation loss coefficient is less than the distillation loss threshold, generate the m-th electrical fault parsing model; input the monitoring matrix of the m-th low-voltage switchgear into the m-th electrical fault parsing model to obtain the m-th electrical fault parsing result.

[0036] Optionally, firstly, based on the characteristics of the m-th specification model, a set of electrical fault records related to the m-th low-voltage distribution cabinet is retrieved. This set of records contains historical electrical fault data, including fault type, occurrence time, and fault parameters. This historical data provides the foundation for subsequent training of the electrical fault analysis model. Then, based on the electrical fault record set of the m-th low-voltage cabinet, supervised learning is performed on Q learners to obtain Q electrical fault parsers, where Q is a positive integer greater than 1. Each learner is constructed from a machine learning model, specifically a feedforward neural network, recurrent neural network, multilayer perceptron, decision tree, random forest, etc. By randomly sampling the electrical fault record set of the m-th low-voltage cabinet a predetermined number of times, Q sets of training data are obtained. These Q sets of training data are then sequentially input into the Q learners for supervised training. Taking a learner composed of a neural network as an example, the learner will iteratively train based on the received training data through steps such as forward propagation, loss calculation, backpropagation, and parameter optimization, thereby transforming the learner into an electrical fault parser. After training all Q learners, Q electrical fault parsers are obtained. These parsers are then subjected to distillation training. The purpose of distillation training is to improve the overall model performance by fusing the knowledge from multiple learners. Specifically, the outputs of the Q electrical fault parsers are used as teacher signals and input into the student model for training. Each parser provides predictions based on historical data, and the student model learns how to improve its prediction capabilities using these teacher signals. Through training, the student model not only learns the predictions of the teacher model but also attempts to mimic its behavior under different inputs. During distillation training, the loss value between the student and teacher models is calculated. The distillation loss coefficient consists of two main parts: a soft-objective loss, which is the difference between the probability distributions of the student and teacher model outputs; and a hard-objective loss, which is the error between the student model and the true label. After a predetermined number of training iterations, the distillation loss coefficient is calculated, and the parameters of the student model are updated using backpropagation, with the goal of minimizing this loss coefficient. If the distillation loss coefficient is less than a preset distillation loss threshold, it indicates that the student model has learned enough information to generate the final m-th electrical fault parsing model. If the distillation loss coefficient does not reach the threshold, training will continue until the loss converges. After constructing the m-th electrical fault analysis model, the monitoring matrix of the m-th low-voltage distribution cabinet will be input into the generated m-th electrical fault analysis model to obtain the electrical fault analysis result of the m-th low-voltage distribution cabinet. This analysis result helps operation and maintenance personnel quickly identify potential electrical fault problems and take corresponding measures to handle them, ensuring the safety and reliability of the power system.

[0037] Based on the first low-voltage switchgear fault report, a topology-linked fault prediction is performed to generate a second low-voltage switchgear fault report.

[0038] In one embodiment, after receiving the first report of a low-voltage switchgear fault, the topology data of the low-voltage switchgear in the power system is combined with the first report to analyze the linkage relationships and fault propagation paths between switchgears. This simulates the propagation process of the fault between different switchgears and identifies potential cascading fault risks. Subsequently, a cascading fault risk analysis model is used to analyze the simulation results. The analysis results are then merged with the first report of the low-voltage switchgear fault and the simulation results to generate a second report of the low-voltage switchgear fault. This second report not only includes fault information for individual switchgears but also considers the mutual influence between devices, providing maintenance personnel with more comprehensive and accurate fault warnings. This helps in taking preventative measures to avoid fault spread and ensure the safe and efficient operation of the power system.

[0039] Furthermore, this application provides a method for predicting topology-linked faults based on the first low-voltage switchgear fault report and generating a second low-voltage switchgear fault report, including: Based on the topology dataset of the multiple low-voltage distribution cabinets, a low-voltage cabinet topology linkage model is constructed. Based on the first low-voltage cabinet fault report, fault propagation simulation is performed on the multiple low-voltage distribution cabinets according to the low-voltage cabinet topology linkage model to obtain the fault propagation simulation results for each low-voltage cabinet. Based on a Bayesian network, a linkage fault risk analysis model is trained. The fault propagation simulation results of each low-voltage cabinet are input into the linkage fault risk analysis model to obtain the risk analysis results for each linkage fault. The first low-voltage cabinet fault report, the fault propagation simulation results of each low-voltage cabinet, and the risk analysis results for each linkage fault are combined to obtain the second low-voltage cabinet fault report.

[0040] Preferably, the first step is to acquire a topology dataset of multiple low-voltage distribution cabinets. This dataset includes the power connection relationships and mutual power supply paths between the cabinets. By inputting this topology data into simulation software for modeling, a low-voltage cabinet topology linkage model can be obtained. This model reflects the connection relationships and interdependencies of the low-voltage distribution cabinets in the power system, including the power transmission paths and mutual power supply relationships between the cabinets, providing a foundation for subsequent fault propagation simulation. Subsequently, based on the first fault report of the low-voltage cabinets, the fault propagation simulation of multiple low-voltage distribution cabinets is performed using the low-voltage cabinet topology linkage model. The first fault report already includes the analysis results of the electrical risks, thermal runaway risks, and aging risks of each low-voltage distribution cabinet. By simulating the propagation paths of different fault types in the power system, the potential impact of a fault in one distribution cabinet on other distribution cabinets can be identified, and fault propagation simulation results for each low-voltage cabinet can be generated. These simulation results reveal how faults gradually spread through linkage paths in the power system, helping to identify potential multi-point fault risks. Next, a Bayesian network is used to train a linked fault risk analysis model. A Bayesian network is a probability-based model that describes the dependencies between variables by constructing relationships between nodes and edges. By using fault propagation simulation results as input, a probabilistic relationship network between distribution cabinets is built. In this network, each node represents a low-voltage distribution cabinet, and edges represent fault propagation paths between cabinets. By training the Bayesian network using historical data, it gradually learns to calculate the probability of fault occurrence in each distribution cabinet and its impact on other cabinets. The trained Bayesian network serves as the linked fault risk analysis model. This model combines historical fault data and existing fault propagation patterns to infer the likelihood of multi-point faults and assess the risk of faults in each distribution cabinet. Then, the fault propagation simulation results of each low-voltage cabinet are input into the trained linked fault risk analysis model. The model processes this input data to obtain the linked fault risk analysis results for each distribution cabinet. These results reflect the impact on each distribution cabinet during fault propagation and the probability and severity of fault occurrence. Finally, the first report on low-voltage switchgear faults, the simulation results of fault propagation in each low-voltage switchgear, and the analysis results of linkage fault risks are compiled to generate the second report on low-voltage switchgear faults. This report not only includes single-point fault information of each switchgear, but also considers fault propagation and linkage risks, providing maintenance personnel with more comprehensive fault warnings and risk assessments to ensure the stable operation of the power system.

[0041] Furthermore, this application provides a method for outputting a low-voltage switchgear early warning signal based on the second low-voltage switchgear fault report.

[0042] Optionally, after receiving a second report of a low-voltage switchgear fault, the second report will be encapsulated into a low-voltage switchgear warning signal and sent to the UI interface for display. In addition, relevant maintenance personnel can be notified via email, SMS or other communication methods to ensure that the warning information can be quickly conveyed and timely and effective response measures can be taken.

[0043] In summary, the embodiments of this application have at least the following technical effects: This application embodiment first uses the Industrial Internet of Things (IIoT) to monitor multiple low-voltage switchgear in the power system in real time, obtaining multiple low-voltage switchgear monitoring matrices. Then, based on these monitoring matrices, electrical risk triggering analysis is performed to construct multiple electrical risk feature matrices. Next, based on these monitoring matrices, thermal runaway risk triggering analysis is performed to construct multiple thermal runaway risk feature matrices. Further, based on these monitoring matrices, aging risk triggering analysis is performed to construct multiple aging risk feature matrices. Then, based on these electrical risk feature matrices, thermal runaway risk feature matrices, and aging risk feature matrices, multi-dimensional fault analysis is performed on the multiple low-voltage switchgear, generating a first low-voltage switchgear fault report. Finally, based on the first low-voltage switchgear fault report, topology linkage fault prediction is performed to generate a second low-voltage switchgear fault report. These technologies collectively address the technical problems of traditional low-voltage switchgear fault monitoring methods, which typically analyze single risk factors such as electrical faults, thermal runaway, or equipment aging in isolation, lacking comprehensive correlation analysis of multi-dimensional risk data. This results in incomplete fault diagnosis coverage and poor early warning timeliness. The technologies achieve the technical effect of improving the comprehensiveness of fault diagnosis and the timeliness of prediction by integrating multi-source monitoring data in real time through the Industrial Internet of Things, constructing a three-dimensional risk feature matrix of electrical, thermal runaway, and aging, and then predicting fault chain effects based on topological relationships.

[0044] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0045] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0046] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A method for remote monitoring of low-voltage switchgear faults based on the Industrial Internet of Things, characterized in that, The method includes: Based on the Industrial Internet of Things, multiple low-voltage switchgear in the power system are monitored in real time to obtain multiple low-voltage switchgear monitoring matrices. Based on the multiple low-voltage switchgear monitoring matrices, electrical risk triggering analysis is performed to construct multiple electrical risk feature matrices; Based on the multiple low-voltage switchgear monitoring matrices, thermal runaway risk triggering analysis is performed to construct multiple thermal runaway risk feature matrices. Based on the multiple low-voltage switchgear monitoring matrices, aging risk trigger analysis is performed to construct multiple aging risk feature matrices. Based on the multiple electrical risk feature matrices, the multiple thermal runaway risk feature matrices, and the multiple aging risk feature matrices, a multi-dimensional fault analysis is performed on the multiple low-voltage switchgear to generate a first low-voltage switchgear fault report; Based on the first low-voltage switchgear fault report, a topology-linked fault prediction is performed to generate a second low-voltage switchgear fault report. Based on the multiple electrical risk feature matrices, the multiple thermal runaway risk feature matrices, and the multiple aging risk feature matrices, a multi-dimensional fault analysis is performed on the multiple low-voltage switchgear, generating a first low-voltage switchgear fault report, including: Based on the multiple electrical risk feature matrices, electrical fault analysis is performed to obtain multiple electrical fault analysis results; Based on the multiple thermal runaway risk feature matrices, thermal runaway fault analysis is performed to obtain multiple thermal runaway fault analysis results; Based on the multiple aging risk feature matrices, aging fault analysis is performed to obtain multiple aging fault analysis results; Based on the multiple low-voltage distribution cabinets, the analysis results of the multiple electrical faults, the multiple thermal runaway faults, and the multiple aging faults are clustered and sorted to obtain the first fault report of the low-voltage cabinet; Electrical fault analysis is performed based on the multiple electrical risk feature matrices to obtain multiple electrical fault analysis results, including: Based on the characteristics of the m-th specification model, retrieve the electrical fault record set of the m-th low-voltage switchgear, where m is a positive integer; Based on the electrical fault record set of the m-th low-voltage switchgear, supervised learning is performed on Q learners to obtain Q electrical fault parsers, where Q is a positive integer greater than 1; Distillation training is performed based on the Q electrical fault analyzers, and the distillation loss coefficient is obtained after each predetermined number of training iterations. If the distillation loss coefficient is less than the distillation loss threshold, generate the m-th electrical fault analysis model; Input the m-th low-voltage switchgear monitoring matrix into the m-th electrical fault analysis model to obtain the m-th electrical fault analysis result; Based on the first low-voltage switchgear fault report, a topology-linked fault prediction is performed to generate a second low-voltage switchgear fault report, including: Based on the topology dataset of the multiple low-voltage distribution cabinets, a low-voltage cabinet topology linkage model is constructed. Based on the first low-voltage switchgear fault report, the fault propagation simulation of the multiple low-voltage switchgear is performed according to the low-voltage switchgear topology linkage model to obtain the fault propagation simulation results of each low-voltage switchgear. A linkage fault risk analysis model was trained based on Bayesian networks. The simulation results of the fault propagation of each low-voltage switchgear are input into the linkage fault risk analysis model to obtain the risk analysis results of each linkage fault. By compiling the first report of low-voltage switchgear failure, the simulation results of each low-voltage switchgear failure propagation, and the risk analysis results of each linkage failure, a second report of low-voltage switchgear failure is obtained.

2. The method for remote monitoring of low-voltage switchgear faults based on the Industrial Internet of Things as described in claim 1, characterized in that, Based on the multiple low-voltage switchgear monitoring matrices, electrical risk triggering analysis is performed to construct multiple electrical risk feature matrices, including: Based on the multiple low-voltage switchgear monitoring matrices, extract the monitoring matrix of the m-th low-voltage switchgear corresponding to the m-th low-voltage switchgear; Using the specification and model information of the m-th low-voltage distribution cabinet as the m-th specification and model feature, and performing electrical risk triggering record retrieval based on the m-th specification and model feature, the m-th electrical risk triggering record set is obtained; Based on the m-th electrical risk trigger record set, construct the m-th electrical risk trigger detection tree; Each monitoring parameter in the m-th low-voltage switchgear monitoring matrix is ​​input into the m-th electrical risk triggering detection tree to obtain multiple electrical risk triggering coefficients; Based on the multiple electrical risk triggering coefficients, the electrical risk characteristics of the m-th low-voltage switchgear monitoring matrix are identified according to the electrical risk triggering threshold, thereby obtaining the m-th electrical risk characteristic matrix.

3. The method for remote monitoring of low-voltage switchgear faults based on the Industrial Internet of Things as described in claim 2, characterized in that, Based on the m-th electrical risk trigger record set, construct the m-th electrical risk trigger detection tree, including: The m-th electrical risk trigger record set includes a monitoring sample set and an electrical risk trigger sample set; Based on the monitoring sample set, the electrical risk triggering sample set is clustered to obtain each electrical risk triggering cluster corresponding to each monitoring sample; The mean value of each electrical risk triggering cluster is calculated to obtain the confidence value of each electrical risk. Using each monitoring sample as multiple input nodes and each electrical risk confidence value as multiple output nodes, the m-th electrical risk trigger detection tree is constructed.

4. The method for remote monitoring of low-voltage switchgear faults based on the Industrial Internet of Things as described in claim 1, characterized in that, Based on the multiple low-voltage switchgear monitoring matrices, thermal runaway risk triggering analysis is performed, and multiple thermal runaway risk feature matrices are constructed, including: Based on the characteristics of the m-th specification model, thermal runaway risk trigger records are retrieved to obtain the m-th thermal runaway risk trigger record set; Based on the m-th thermal runaway risk trigger record set, construct the m-th thermal runaway risk trigger detection tree; Input each monitoring parameter in the mth low-voltage switch monitoring matrix into the mth thermal runaway risk triggering detection tree to obtain multiple thermal runaway risk triggering coefficients; Based on the multiple thermal runaway risk triggering coefficients, thermal runaway risk characteristics are identified in the m-th low-voltage switchgear monitoring matrix according to the thermal runaway risk triggering threshold, thus obtaining the m-th thermal runaway risk characteristic matrix.

5. The method for remote monitoring of low-voltage switchgear faults based on the Industrial Internet of Things as described in claim 1, characterized in that, Based on the Industrial Internet of Things (IIoT), multiple low-voltage switchgear cabinets in the power system are monitored in real time, resulting in multiple low-voltage switchgear monitoring matrices, including: Based on real-time monitoring of the multiple low-voltage distribution cabinets using the Industrial Internet of Things, multiple low-voltage cabinet monitoring sets are obtained. Clean and organize the multiple low-voltage switchgear monitoring sets to generate the multiple low-voltage switchgear monitoring matrices.

6. The method for remote monitoring of low-voltage switchgear faults based on the Industrial Internet of Things as described in claim 5, characterized in that, Based on the Industrial Internet of Things (IIoT), the status data of each sensor is retrieved, and adaptive sensor anomaly optimization is performed on the multiple low-voltage cabinet monitoring sets according to the status data of each sensor.

7. The method for remote monitoring of low-voltage switchgear faults based on the Industrial Internet of Things as described in claim 1, characterized in that, Based on the second low-voltage switchgear fault report, a low-voltage switchgear warning signal is output.

Images