Intelligent management method of power distribution cabinet based on edge computing and intelligent power distribution cabinet

Abstract

The present invention relates to the technical fields of intelligent power distribution system, artificial intelligence, edge computing, etc., and provides an intelligent management method for distribution cabinet based on edge computing and an intelligent distribution cabinet. The method receives a data feature summary uploaded by an edge gateway, the data feature summary is generated by an edge server and transmitted to the edge gateway, and then the user role authority of the user end is preset. The user access behavior is analyzed by an AI model of user access behavior according to the user role authority and user access behavior log of the user end, so as to dynamically manage the access of the user end user, and then the abnormal alarm information in the data feature summary is identified and pushed. When multiple user ends perform remote abnormal control operations at the same time, the operations are given priority control according to the preset priority strategy, thereby improving the data processing efficiency and security in the distribution cabinet management, ensuring the consistency of operations, and improving the diagnostic accuracy and operation and maintenance efficiency.

Description

Technical Field

[0001] The present invention relates to the technical fields of intelligent power distribution systems, artificial intelligence, edge computing, and more particularly to an intelligent management method for power distribution cabinets based on edge computing and an intelligent power distribution cabinet. Background Art

[0002] Smart distribution cabinet systems usually use cloud platforms to achieve multi-point remote monitoring and data collection. Manufacturers provide a unified platform to achieve standardized management of devices for different users, avoiding the need for users to develop additional platforms. Users can access the system through their login accounts to view the real-time data, historical trends, alarm records, etc. of the distribution cabinet, which is convenient for rapid response to faults. For example, after logging in through the smart distribution cabinet App or Web platform, users can remotely view the data status of multiple distribution cabinets without relying on local access devices. Multi-point remote monitoring requires high real-time performance, and network delays or interruptions will cause data to be unable to be uploaded in real time, which may cause monitoring information lags and affect alarm response. Delayed push of alarm information may lead to untimely fault handling, further increasing the risk of equipment damage. In addition, the distribution cabinet system usually allows multiple users to log in, but the existing permission management mode is not flexible enough, and it is easy to have unauthorized access or data leakage. The system is vulnerable to unauthorized operations, resulting in risks such as data tampering. Moreover, when multiple user terminals access the same distribution cabinet at the same time, the synchronization mechanism of the operation is insufficient, and data conflicts or operation failures may occur. The data fed back to the user by the system may be inconsistent with the actual status, causing users to make wrong judgments. In addition, existing intelligent distribution cabinet systems mostly rely on manual analysis and fixed rules to judge abnormalities, lack intelligent diagnostic mechanisms, easily miss potential faults and fail to detect complex faults in a timely manner, affecting the operating safety and operation and maintenance efficiency of the equipment.

[0003] In summary, the existing management technology of intelligent distribution cabinets has technical problems such as low data processing efficiency and low security, insufficient operation synchronization mechanism that easily causes data conflicts or operation failures, easy omission of potential faults and failure to detect complex faults in a timely manner. Summary of the invention

[0004] In view of the shortcomings of the above-mentioned prior art, the present invention provides an intelligent management method for distribution cabinets and an intelligent distribution cabinet based on edge computing, so as to improve data processing efficiency and security, ensure operation consistency, and improve diagnostic accuracy and operation and maintenance efficiency.

[0005] In a first aspect, the present invention provides an intelligent management method for a power distribution cabinet based on edge computing, comprising:

[0006] Receive a data feature summary uploaded by an edge gateway, the data feature summary is generated by an edge server and then transmitted to the edge gateway; the edge server acquires high-frequency sampling data from the power distribution cabinet in real time, filters out redundant information in the high-frequency sampling data, generates de-redundant high-frequency sampling data, and extracts summary features from the de-redundant high-frequency sampling data to generate the data feature summary; the high-frequency sampling data is data recorded at a preset high sampling rate for the electrical signal at the power distribution cabinet, and the high-frequency sampling data is used to capture instantaneous changes and detailed fluctuations of the electrical signal at the power distribution cabinet;

[0007] Preset user role permissions of the user terminal, the user role permissions include administrator permissions, operation and maintenance personnel permissions, and general user permissions with different permission levels, input the user role permissions of the user terminal and the user access behavior log of the user terminal into the trained user access behavior analysis AI model, analyze the user access behavior, and dynamically manage the access of the user of the user terminal;

[0008] The abnormal alarm information in the data feature summary is identified, and the identified abnormal alarm information is pushed to the user terminal. When multiple user terminals with different user role permissions perform remote abnormal control operations at the same time, the remote abnormal control operations are given priority operation control according to a preset priority strategy.

[0009] In a second aspect, the present invention provides an intelligent distribution cabinet, which uses the above-mentioned distribution cabinet intelligent management method based on edge computing for intelligent management.

[0010] Compared with the prior art, the present invention has the following beneficial effects:

[0011] The present invention provides an intelligent management method for a power distribution cabinet based on edge computing and an intelligent power distribution cabinet. By receiving a data feature summary uploaded by an edge gateway, the data feature summary is generated by an edge server and then transmitted to the edge gateway. The edge server obtains high-frequency sampling data at the power distribution cabinet end in real time, filters out redundant information in the high-frequency sampling data, generates de-redundant high-frequency sampling data, and extracts summary features from the de-redundant high-frequency sampling data to generate the data feature summary. The high-frequency sampling data is data recorded at a preset high sampling rate for an electrical signal at the power distribution cabinet end. The high-frequency sampling data is used to capture instantaneous changes and detailed fluctuations of the electrical signal at the power distribution cabinet end. Then, the user role permissions of the user end are preset. The user role permissions include Prepare administrator permissions, operation and maintenance personnel permissions and ordinary user permissions at different permission levels, input the user role permissions of the user terminal and the user access behavior log of the user terminal into the trained user access behavior analysis AI model, analyze the user access behavior to dynamically manage the access of the user of the user terminal, and then identify the abnormal alarm information in the data feature summary, and push the identified abnormal alarm information to the user terminal. When multiple user terminals with different user role permissions perform remote abnormal control operations at the same time, the remote abnormal control operations are given priority operation control according to the preset priority strategy, thereby improving the data processing efficiency and security in the distribution cabinet management, ensuring operation consistency, and improving diagnostic accuracy and operation and maintenance efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The drawings described herein are used to provide further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0013] Figure 1 It is a flow chart of a method for intelligent management of power distribution cabinets based on edge computing according to an embodiment of the present invention;

[0014] Figure 2 It is a schematic diagram of a system architecture when the intelligent distribution cabinet is managed using the distribution cabinet intelligent management method based on edge computing in an embodiment of the present invention. DETAILED DESCRIPTION

[0015] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only an embodiment of a part of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work should fall within the scope of protection of the present invention.

[0016] See also Figure 1-Figure 2 , an embodiment of the present invention provides an intelligent management method for a power distribution cabinet based on edge computing and an intelligent power distribution cabinet. Using the intelligent management method for a power distribution cabinet based on edge computing, the intelligent power distribution cabinet can be intelligently managed. The intelligent management method for a power distribution cabinet based on edge computing may include steps S101, S102, and S103. Steps S101, S102, and part or all of step S103 may be run on a cloud server. The cloud server receives a data feature summary uploaded by an edge gateway. The data feature summary is generated by an edge server and transmitted to the edge gateway. The edge server obtains high-frequency sampling data from the power distribution cabinet in real time, filters out redundant information in the high-frequency sampling data, generates de-redundant high-frequency sampling data, and extracts summary features from the de-redundant high-frequency sampling data to generate the data feature summary. The high-frequency sampling data is data recorded at a preset high sampling rate for the electrical signal at the power distribution cabinet. The high-frequency sampling data is used to capture the instantaneous changes and detail waves of the electrical signal at the power distribution cabinet. The user role permissions of the user end are preset according to the dynamic situation, and the user role permissions of the user end are preset. The user role permissions include administrator permissions, operation and maintenance personnel permissions and ordinary user permissions with different permission levels. The user role permissions of the user end and the user access behavior log of the user end are input into the trained user access behavior analysis AI model to analyze the user access behavior to dynamically manage the access of the user of the user end, and then the abnormal alarm information in the data feature summary is identified, and the identified abnormal alarm information is pushed to the user end. When multiple user ends with different user role permissions perform remote abnormal control operations at the same time, the remote abnormal control operations are given priority operation control according to the preset priority strategy, thereby improving the data processing efficiency and security in the distribution cabinet management, ensuring operation consistency, and improving diagnostic accuracy and operation and maintenance efficiency.

[0017] The intelligent management method of the power distribution cabinet based on edge computing includes:

[0018] S101, receiving a data feature summary uploaded by an edge gateway, the data feature summary is generated by an edge server and then transmitted to the edge gateway; the edge server acquires high-frequency sampling data from a power distribution cabinet in real time, filters out redundant information in the high-frequency sampling data, generates de-redundant high-frequency sampling data, and extracts summary features from the de-redundant high-frequency sampling data to generate the data feature summary; the high-frequency sampling data is data recorded at a preset high sampling rate for an electrical signal at the power distribution cabinet, and the high-frequency sampling data is used to capture instantaneous changes and detailed fluctuations of the electrical signal at the power distribution cabinet;

[0019] S102, presetting user role permissions of the user terminal, wherein the user role permissions include administrator permissions, operation and maintenance personnel permissions, and general user permissions with different permission levels, inputting the user role permissions of the user terminal and the user access behavior log of the user terminal into the trained user access behavior analysis AI model, analyzing the user access behavior, so as to dynamically manage the access of the user of the user terminal;

[0020] S103, identifying the abnormal alarm information in the data feature summary, pushing the identified abnormal alarm information to the user terminal, and when multiple user terminals with different user role permissions perform remote abnormal control operations at the same time, giving priority to the remote abnormal control operations according to a preset priority strategy.

[0021] It should be noted that in this embodiment, by introducing edge computing, key tasks such as data processing of high-frequency sampling data are sunk to the edge server, thereby improving data processing efficiency and the real-time performance of the intelligent distribution cabinet system. At the same time, the use of AI models for user behavior analysis and dynamic authority management can enhance the security of the intelligent distribution cabinet system and prevent unauthorized operations and data leakage. Finally, by managing the remote control operations of multiple users through priority strategies, the problem of insufficient operation synchronization mechanism in the existing intelligent distribution cabinet system can be solved to ensure data consistency and the effectiveness of operations. Among them, the edge server obtains high-frequency sampling data from the distribution cabinet end in real time, filters redundant information, generates de-redundant high-frequency sampling data, and extracts summary features and transmits them to the edge gateway. By locally processing high-frequency sampling data through edge computing, the amount of data transmission can be reduced and the data processing efficiency can be improved. Edge computing can be processed at a location close to the data source (such as the distribution cabinet end), reduce network latency, and ensure real-time upload and processing of data. In addition, different role permissions (administrators, operation and maintenance personnel, and ordinary users) of the user end are preset, and the user role permissions and access behavior logs are input into the trained AI model for analysis to dynamically manage user access. Through detailed role permission management, it is ensured that different users can only access and operate functions within their authority scope to prevent unauthorized access and data leakage. At the same time, using the AI ​​model to dynamically analyze user behavior can achieve more flexible and intelligent permission management and improve the security of the intelligent distribution cabinet system. In addition, the abnormal alarm information in the data feature summary is identified and pushed to the user end; when multiple users perform remote abnormal control operations at the same time, priority control is performed according to the priority strategy. In this embodiment, by identifying abnormal alarms and pushing them to users in a timely manner, it is possible to ensure rapid response to faults and reduce the risk of damage to the distribution cabinet. Setting a priority strategy can ensure that when multiple users operate at the same time, they proceed in an orderly manner, avoid data conflicts and operation failures, and improve the reliability and consistency of the intelligent distribution cabinet system. Furthermore, the preset priority strategy is: the remote abnormal control operation of the administrator takes precedence over the remote abnormal control operation of the operation and maintenance personnel, and the remote abnormal control operation of the operation and maintenance personnel takes precedence over the remote abnormal control operation of ordinary users; when the user roles are the same, the remote abnormal control operation of the user with a high user trust score takes precedence over the remote abnormal control operation of the user with a low user trust score; the user trust score is evaluated by the user access behavior analysis AI model according to the context of the access behavior log. It should be noted that in this embodiment, through the hierarchical division of role permissions and the dynamic evaluation of trust scores, it is ensured that key operations are performed by high-authority and high-trust users first, preventing misoperation or malicious operations by low-authority or potential risk users, and ensuring the stability and security of the system.By clarifying priority allocation, the system's response speed and coordination capabilities in multi-user operation scenarios can be improved, operational conflicts can be reduced, and key exceptions can be handled quickly and efficiently, thereby improving overall operation and maintenance efficiency and user experience.

[0022] It should be noted that in the working environment of the power distribution cabinet, transient phenomena such as instantaneous short circuit, surge, voltage or current mutation, harmonic interference, etc. may occur. High-frequency sampling can capture these transient signals and detailed fluctuations, thereby providing a more accurate basis for fault analysis. Among them, the high-frequency sampling data may include electrical transient waveform data, harmonic component data, transient event data, and high-frequency noise signal data. The electrical transient waveform data can be used to detect waveform distortion, short circuit, surge, voltage or current mutation; the harmonic component data can be used to detect harmonic interference problems caused by nonlinear loads in the power distribution system; the transient event data can be used to analyze the type and occurrence time of abnormal events and determine whether the circuit needs to be cut off; the high-frequency noise signal data can be used to determine whether there is an interference source in the power distribution system. If there is only low-frequency or coarse-grained data, these transient processes may not be perceived in time, resulting in missed reports or false reports, and accurate fault location and diagnosis suggestions cannot be given. It should be noted that the high-frequency sampling data is not a simple conventional voltage and current measurement, but a preset higher sampling rate (such as 10kHz, 20kHz, 50kHz) to continuously record the instantaneous waveform of the electrical signal at the distribution cabinet end, which is used to capture and analyze various rapid or short-term electrical characteristic changes, and provide more adequate data support for fault detection and equipment predictive maintenance. In the face of complex fault scenarios (such as multiple fault superposition, harmonics and short-term surges), high-frequency sampling data can more accurately reflect the characteristic changes before and after the fault occurs, and reduce the misjudgment rate and missed judgment rate. It should also be noted that since the amount of high-frequency sampling data is usually very large, if it is sent to the cloud, the network load will be huge and the delay will increase, which is not conducive to real-time monitoring and rapid response. In this embodiment, the redundant data of the high-frequency sampling data is pre-processed on the edge side (such as feature extraction, compression, denoising, etc.), and de-redundant high-frequency sampling data is generated, and the de-redundant high-frequency sampling data is extracted to generate the data feature summary, and only the data feature summary is transmitted to the cloud, thereby reducing network bandwidth consumption and reducing cloud storage pressure. The edge server can retain the original high-frequency sampling data and provide it to users with higher permissions when necessary. The cloud can manage user permissions more flexibly and call on a dedicated user access behavior analysis AI model to analyze whether user behavior contains malicious operations.

[0023] It should be noted that the user access behavior analysis AI model can adopt supervised learning models such as decision trees and random forests. When training the user access behavior analysis AI model, the model training can be carried out through the user role permissions of the user end and the historical data of the user access behavior log of the user end. The specific training method is a conventional technology and will not be repeated in this embodiment.

[0024] In some preferred embodiments, when the user terminal that obtains the priority operation initiates a request to view the de-redundant high-frequency sampling data, an access path to access the edge server is sent to the user terminal that obtains the priority operation, and the access path is used to access the de-redundant high-frequency sampling data stored by the edge server. It should be noted that in the management architecture of high-frequency sampling data of the distribution cabinet, high-authority or key role users can directly retrieve the original data of the de-redundant high-frequency sampling data. By storing the de-redundant high-frequency sampling data in the edge server and only issuing the access path to the priority operation user, data acquisition with strong real-time performance can be achieved under the premise of ensuring data security. The amount of high-frequency sampling data is large and the transmission bandwidth occupies a high amount. If all users need to upload the data to the cloud before downloading or viewing it, the efficiency is low and delays may also occur. In this embodiment, when a specific user needs to view the data, the access path is sent to the edge server, and the user terminal is directly connected to the edge server to obtain the data locally, which can achieve faster and lower latency viewing and analysis. In addition, when multiple users view or operate the same distribution cabinet data at the same time, if there is no priority strategy, conflicts or resource competition will occur. In this embodiment, the access path is sent only when it is confirmed that a certain user is in a priority operation state, so as to ensure that data access and control operations are in order and avoid unnecessary conflicts or data abuse.

[0025] In some preferred embodiments, after generating the de-redundant high-frequency sampling data, the edge server performs abnormal state analysis on the de-redundant high-frequency sampling data at the distribution cabinet end to obtain the abnormal alarm information, and automatically fills the obtained abnormal alarm information into the data feature summary formation module to form the data feature summary containing the abnormal alarm information. It should be noted that in the distribution cabinet monitoring scenario, real-time performance is very critical for instantaneous events (such as short circuits, surges, rapid mutations, etc.). If it is completely dependent on cloud analysis, it is necessary to upload massive high-frequency data before detection, and network delays and bandwidth limitations may cause alarm lags. In this embodiment, the abnormal state analysis is completed directly on the edge side, the abnormality can be discovered in the first time, and the alarm information is produced locally. The amount of high-frequency sampling data itself is huge. If all analysis work is done in the cloud, a large amount of data must be uploaded to the cloud first. Through local analysis on the edge server, we can first filter out the parts with obvious anomalies or key concerns, and write the corresponding abnormal alarm information into the summary. The cloud only needs to identify the abnormal alarm information based on the summary, and then push the abnormal alarm information to the user end, thereby significantly reducing the bandwidth burden and cloud storage and computing overhead. For example, once the edge server identifies an anomaly, it can insert key information such as the anomaly type, occurrence time, and severity into the data feature summary. The data feature summary is a simplified and structured expression. When receiving the summary, the cloud or other analysis modules can quickly focus on the abnormal event and directly obtain the anomaly that has been identified locally on the edge server without having to analyze all the original data from scratch to avoid repeated calculations. Furthermore, when the de-redundant high-frequency sampling data is subjected to abnormal state analysis at the distribution cabinet end, it includes: performing multi-dimensional feature extraction on the electrical instantaneous waveform data, harmonic component data, transient event data, and high-frequency noise signal data of the high-frequency sampling data to obtain several key indicators for judging the abnormal state of the distribution cabinet end; comparing the key indicators with the preset normal distribution cabinet operation threshold range to screen out potential anomalies that exceed the normal distribution cabinet operation threshold range; performing multi-dimensional cross-validation on the potential anomalies, and if the verification result meets the preset abnormal judgment condition, generating corresponding abnormal alarm information. It should be noted that in this embodiment, through the hierarchical analysis of feature extraction-threshold comparison-multi-dimensional cross-validation, various anomalies such as instantaneous faults and harmonic distortion can be effectively captured. If only a single threshold judgment is simply relied on, omissions or false alarms are prone to occur; the multi-dimensional cross-validation method can improve the accuracy and reliability of the alarm information.In addition, when performing multi-dimensional cross-validation on the potential anomaly, it may include: comparing the similarity of several key indicators of the potential anomaly with the characteristics of equipment failure samples in the historical operation data of the distribution cabinet equipment, if the similarity exceeds the preset threshold, the potential anomaly is marked as a suspicious anomaly; combining the real-time electrical parameters, environmental parameters and equipment switch operation records of the distribution cabinet, the potential anomaly marked as a suspicious anomaly is secondary verified to distinguish between occasional fluctuation anomalies and actual possible anomalies; if the potential anomaly marked as a suspicious anomaly after the secondary verification is an actual possible anomaly, the corresponding abnormal alarm information is generated. It should be noted that if the anomaly is immediately determined to be abnormal only by relying on the parameter exceeding the threshold at a certain moment, a large number of false alarms (such as false anomalies caused by short-term fluctuations in grid voltage and current) or missed alarms (some abnormal characteristics are not obvious and ignored) may occur. In this embodiment, by comparing the similarity with the characteristics of historical fault samples, if the similarity is high, the potential anomaly is marked as a suspicious anomaly, and more data can be combined in the subsequent steps for re-analysis, thereby reducing misjudgment and missed judgment and improving the accuracy of the alarm. The failure of the distribution cabinet usually has a certain historical law or characteristic pattern. Comparing the similarity of several key indicators of potential abnormalities (waveform characteristics, harmonic characteristics, instantaneous current and voltage changes, etc.) with the historical fault sample library can help to quickly locate abnormalities that are similar to specific fault modes. At the same time, abnormal conditions are often not only related to electrical parameters, but may also be affected by ambient temperature, humidity, wind speed or equipment operation conditions (such as the number of switch actions, closing time, etc.). If an alarm is issued based on a single electrical parameter exceeding the limit, it is easy to confuse occasional fluctuations with real equipment failures. In this embodiment, by performing a secondary check on the potential abnormalities marked as suspicious abnormalities with real-time electrical data, environmental data, and switch operation records, it is possible to comprehensively judge whether it is a short-term harmless fluctuation or an actual possible sign of a fault, thereby further improving the reliability of the alarm information.

[0026] Furthermore, when the obtained abnormal alarm information is automatically filled into the formation module of the data feature summary, it includes: structurally extracting the elements in the abnormal alarm information that characterize the fault type, fault occurrence time, fault severity and corresponding electrical parameter indicators; writing the abnormal alarm information after structured extraction into the formation module of the data feature summary, so as to automatically update the abnormal alarm information display part in the formation module of the data feature summary, and obtain the data feature summary containing the abnormal alarm information. It should be noted that if the abnormal alarm information and the data feature summary are stored separately, problems of information dispersion or repeated management will arise. In addition, the abnormal alarm information often contains multiple elements (fault type, occurrence time, severity, electrical parameter limit value, etc.). If it is recorded in the form of text stacking or log files, subsequent retrieval is inconvenient. In this embodiment, by unified integration management of abnormal alarm information and data feature summary, multiple elements of abnormal alarm information are structured and extracted, and written into the data feature summary formation module, the system or operation and maintenance personnel can quickly read or call through a unified data structure (such as JSON, XML, database field, etc.), and automatically display or visualize the abnormality, thereby improving readability and analysis efficiency. In addition, the data feature summary formation module is responsible for collecting, integrating, organizing and storing high-frequency sampling data and related abnormal alarm information collected from the distribution cabinet end, thereby forming a data feature summary. For example, high-frequency sampling data (such as voltage, current waveform, harmonic components, transient events, etc.) from different sensors and data sources at the distribution cabinet end are summarized and integrated to form a comprehensive operation feature summary. For example, after the abnormal alarm information (such as fault type, occurrence time, severity, electrical parameter overrun, etc.) is structured, it is integrated into the data feature summary to ensure the comprehensiveness and consistency of the summary content. The data feature summary formation module can use an efficient data storage mechanism (such as a relational database, a NoSQL database, or an in-memory database) to save the generated feature summary, supporting fast query, retrieval and update operations. The module that forms the data feature summary can provide an API or data interface for other modules (such as user terminals, cloud servers, analysis engines, etc.) to access, query or update the data feature summary.

[0027] In some preferred embodiments, the edge server obtains the real-time electrical parameter data, real-time equipment operating environment data and equipment switch operation records of the distribution cabinet, performs equipment fault prediction based on the real-time electrical parameter data, the real-time equipment operating environment data and the equipment switch operation records, and generates diagnostic suggestions based on the equipment fault prediction results. It should be noted that real-time alarms focus on handling current anomalies, while in this embodiment, fault prediction focuses on potential problems in the future. Fault prediction and real-time alarms jointly ensure the stable operation of the distribution cabinet. High-frequency sampling data is used for real-time monitoring, while real-time electrical parameter data, real-time equipment operating environment data and equipment switch operation records are used for fault prediction, so as to improve the foresight and prevention capabilities of the intelligent distribution cabinet management system, and enhance the reliability and operation and maintenance efficiency of the intelligent distribution cabinet management system as a whole.

[0028] In some preferred embodiments, after generating the diagnostic suggestion, the edge server generates the best maintenance plan according to the diagnostic suggestion and the maintenance knowledge graph of the distribution cabinet end, so as to deal with the predicted equipment failure of the distribution cabinet end in advance. It should be noted that in the management of the distribution cabinet, fault prediction can identify potential problems in advance, but how to take corresponding maintenance measures efficiently and accurately is still a challenge. In this embodiment, by combining the diagnostic suggestion with the maintenance knowledge graph of the distribution cabinet end, the optimal maintenance plan is automatically generated based on the existing maintenance knowledge and experience, reducing the subjective judgment and manual intervention of the operation and maintenance personnel in formulating the maintenance plan, and improving the scientificity and consistency of the decision. It should be noted that when processing high-frequency electrical sampling data, attention is paid to real-time monitoring and abnormal alarms. Based on the fault prediction results and knowledge graph, attention is paid to fault response and maintenance decisions. The combination of real-time alarm and predictive maintenance can enhance the intelligence level of the distribution cabinet management system, which can not only deal with current abnormalities, but also prevent potential faults in the future.

[0029] In some preferred embodiments, the edge server deploys a trained distribution cabinet fault analysis AI model; the distribution cabinet fault analysis AI model predicts equipment faults based on the real-time electrical parameter data, the real-time equipment operating environment data, and the equipment switch operation records, and generates diagnostic suggestions based on the equipment fault prediction results. It should be noted that by deploying the distribution cabinet fault analysis AI model, the distribution cabinet fault analysis AI model can learn and identify complex fault modes from a large amount of real-time electrical parameter data, equipment operating environment data, and switch operation records, thereby improving the accuracy and intelligence level of fault prediction.

[0030] In some preferred embodiments, the distribution cabinet fault analysis AI model adopts a machine learning model. When training the distribution cabinet fault analysis AI model, the machine learning model is trained through the real-time electrical parameter data, the real-time equipment operating environment data, the historical data of the equipment switch operation record, and the distribution cabinet fault state data corresponding to the historical data to obtain the trained distribution cabinet fault analysis AI model. It should be noted that the operating state of the distribution cabinet is affected by many factors, such as electrical parameters, environmental conditions, and operation records, and there are complex nonlinear relationships between these factors. Traditional statistical methods are difficult to effectively capture and model these complex patterns, while machine learning models are good at processing high-dimensional, multivariate data, and can automatically discover potential patterns and associations in the data. By learning from a large amount of historical data, the machine learning model can continuously optimize its own prediction ability and adapt to changes in the operating environment and fault mode of the distribution cabinet. Among them, the machine learning model can use supervised learning models such as decision trees and random forests. In addition, real-time electrical parameter data can include voltage, current, power, etc., which can reflect the electrical operating state of the distribution cabinet in real time and are the basic data for fault prediction. Real-time equipment operating environment data, such as temperature, humidity, vibration and other environmental parameters, can reflect the operating status of the equipment under different environmental conditions, and can help identify potential faults caused by environmental changes. The historical data of the equipment switch operation record, which records the operation history of the equipment, including the number of switches, switch time, etc., can help the model understand the relationship between equipment operation and failure, and identify the risk of failure caused by frequent operation or abnormal operation mode. In this embodiment, by integrating electrical parameters, environmental data and operation records, multi-dimensional information input is provided, so that the machine learning model can analyze and predict faults from multiple angles, improving the accuracy and comprehensiveness of the prediction. Using the distribution cabinet fault status data corresponding to the historical data for training, the model can learn the differences between various parameters under normal and abnormal conditions, and identify potential faults.

[0031] It should be pointed out that the above embodiments are only preferred specific implementation modes of the present invention, and the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention. The protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for intelligent management of power distribution cabinets based on edge computing, characterized in that: include: Receiving a data feature summary uploaded by an edge gateway, wherein the data feature summary is generated by an edge server and then transmitted to the edge gateway; The edge server acquires high-frequency sampling data from the power distribution cabinet in real time, filters out redundant information in the high-frequency sampling data, generates de-redundant high-frequency sampling data, and extracts summary features from the de-redundant high-frequency sampling data to generate the data feature summary; the high-frequency sampling data is data that records the electrical signal at the power distribution cabinet at a preset high sampling rate, and the high-frequency sampling data is used to capture the instantaneous changes and detailed fluctuations of the electrical signal at the power distribution cabinet; the high-frequency sampling data includes electrical instantaneous waveform data, harmonic component data, transient event data, and high-frequency noise signal data; the electrical instantaneous waveform data is used to detect waveform distortion, short circuit, surge, voltage or current mutation; the harmonic component data is used to detect harmonic interference problems caused by nonlinear loads in the power distribution system; the transient event data is used to analyze the type and occurrence time of abnormal events to determine whether the circuit needs to be cut off; The high-frequency noise signal data is used to determine whether there is an interference source in the power distribution system; Preset user role permissions of the user terminal, the user role permissions include administrator permissions, operation and maintenance personnel permissions, and general user permissions with different permission levels, input the user role permissions of the user terminal and the user access behavior log of the user terminal into the trained user access behavior analysis AI model, analyze the user access behavior, and dynamically manage the access of the user of the user terminal; The abnormal alarm information in the data feature summary is identified, and the identified abnormal alarm information is pushed to the user terminal. When multiple user terminals with different user role permissions perform remote abnormal control operations at the same time, the remote abnormal control operations are given priority operation control according to a preset priority strategy; the preset priority strategy is: the remote abnormal control operation of the administrator takes precedence over the remote abnormal control operation of the operation and maintenance personnel, and the remote abnormal control operation of the operation and maintenance personnel takes precedence over the remote abnormal control operation of ordinary users; when the user roles are the same, the remote abnormal control operation of a user with a high user trust score takes precedence over the remote abnormal control operation of a user with a low user trust score; the user trust score is evaluated by the user access behavior analysis AI model according to the context of the access behavior log; After generating the de-redundant high-frequency sampling data, the edge server performs abnormal state analysis of the power distribution cabinet end on the de-redundant high-frequency sampling data to obtain the abnormal alarm information, and automatically fills the obtained abnormal alarm information into the data feature summary formation module to form the data feature summary containing the abnormal alarm information; When the de-redundant high-frequency sampling data is subjected to abnormal state analysis at the power distribution cabinet end, the method includes: performing multi-dimensional feature extraction on the electrical instantaneous waveform data, harmonic component data, transient event data and high-frequency noise signal data of the high-frequency sampling data to obtain several key indicators for judging the abnormal state at the power distribution cabinet end; comparing the key indicators with the preset normal power distribution cabinet operation threshold range to screen out potential abnormalities that exceed the normal power distribution cabinet operation threshold range; The potential anomaly is cross-validated in multiple dimensions, and if the validation result meets the preset anomaly determination condition, a corresponding anomaly alarm message is generated.

2. The method for intelligent management of power distribution cabinets based on edge computing according to claim 1, characterized in that: When the user terminal that obtains the priority operation initiates a request to view the de-redundant high-frequency sampling data, an access path for accessing the edge server is sent to the user terminal that obtains the priority operation, and the access path is used to access the de-redundant high-frequency sampling data stored in the edge server.

3. The intelligent management method of power distribution cabinet based on edge computing according to claim 1, characterized in that: The edge server obtains real-time electrical parameter data, real-time equipment operating environment data and equipment switch operation records of the distribution cabinet, predicts equipment failures based on the real-time electrical parameter data, the real-time equipment operating environment data and the equipment switch operation records, and generates diagnostic suggestions based on the equipment failure prediction results.

4. The method for intelligent management of power distribution cabinets based on edge computing according to claim 3, characterized in that: After generating the diagnostic suggestion, the edge server generates an optimal maintenance plan based on the diagnostic suggestion and the maintenance knowledge graph of the power distribution cabinet to deal with the predicted equipment failure of the power distribution cabinet in advance.

5. The method for intelligent management of power distribution cabinets based on edge computing according to claim 3, characterized in that: The edge server deploys a trained distribution cabinet fault analysis AI model; the distribution cabinet fault analysis AI model predicts equipment failures based on the real-time electrical parameter data, the real-time equipment operating environment data, and the equipment switch operation records, and generates diagnostic suggestions based on the equipment failure prediction results.

6. The method for intelligent management of power distribution cabinets based on edge computing according to claim 5, characterized in that: The distribution cabinet fault analysis AI model adopts a machine learning model. When training the distribution cabinet fault analysis AI model, the machine learning model is trained through the real-time electrical parameter data, the real-time equipment operating environment data, the historical data of the equipment switch operation records, and the distribution cabinet fault status data corresponding to the historical data to obtain the trained distribution cabinet fault analysis AI model.

7. An intelligent power distribution cabinet, characterized in that: The intelligent distribution cabinet is intelligently managed using the distribution cabinet intelligent management method based on edge computing as described in any one of claims 1 to 6.

Images