An intelligent monitoring method and system based on a power distribution room

Abstract

The application provides an intelligent monitoring method and system based on a power distribution room, which realizes identification and early warning of abnormal conditions of the power distribution room by using a monitoring strategy of sub-regional and hierarchical levels. First, the power distribution room is divided into multiple power distribution regions through power distribution room spatial distribution data, environmental data of each region is collected in a preset sampling time window, and a region abnormality identification model is pre-trained to generate a region abnormality confidence. When the region abnormality confidence exceeds a preset threshold, the region is marked as an abnormal region, and a list of power distribution equipment in the abnormal region is determined. Subsequently, each power distribution equipment in the abnormal region is monitored, a device abnormality confidence is generated through a device abnormality identification model, and finally an abnormal power distribution equipment is identified according to a preset confidence threshold. The data processing pressure of full-time and full-equipment monitoring is avoided, the monitoring efficiency is improved, and intelligent early warning protection is provided for safe operation of the power distribution room.

Description

Technical Field

[0001] This application relates to the field of power distribution, and in particular to an intelligent monitoring method and system based on a power distribution room. Background Technology

[0002] As a critical node in the power system, the substation bears the important responsibility of power distribution and conversion. With the expansion of the power grid and the growth of electricity demand, the safe and stable operation of the substation has a decisive impact on the reliability of the entire power system. Currently, substation monitoring mainly employs methods such as manual inspection, traditional online monitoring, and intelligent monitoring systems.

[0003] Manual inspection is the most traditional monitoring method, involving professional personnel periodically checking the operating status of power distribution equipment. This method is intuitive and based on extensive experience, allowing personnel to identify potential problems. However, manual inspection is inefficient, cannot achieve 24 / 7 monitoring, and suffers from long inspection cycles, discontinuous monitoring, and a tendency to miss early signs of equipment failure. Furthermore, human error in observation and recording can also affect monitoring quality. Working in high-voltage, hazardous environments poses significant safety risks to personnel, and labor costs continue to rise with increasing labor costs, resulting in persistently high maintenance and management costs.

[0004] Traditional online monitoring systems achieve continuous monitoring of power distribution rooms by installing sensors for temperature, humidity, current, and voltage. While this method achieves continuous monitoring compared to manual inspections, it has significant drawbacks. Traditional systems often suffer from severe data silos, with each subsystem operating independently, hindering data sharing and comprehensive analysis, resulting in fragmented monitoring information. Equipment condition assessment relies too heavily on single parameters, lacking multi-dimensional comprehensive judgment capabilities, leading to a high false alarm rate. Summary of the Invention

[0005] This application provides an intelligent monitoring method based on a power distribution room, including the following steps:

[0006] A1, determine the corresponding power distribution room spatial distribution data in the preset power distribution room;

[0007] A2, generates multiple power distribution area data by segmenting the spatial distribution data of the power distribution room;

[0008] A3, collect corresponding power distribution area environmental time window data at corresponding locations in the power distribution room according to the preset sampling time window and the data of each power distribution area;

[0009] A4. Based on the time window data of the power distribution area environment, the corresponding regional anomaly confidence level is generated through a pre-trained regional anomaly identification model.

[0010] A5. If the regional anomaly confidence level is greater than the preset regional anomaly confidence level threshold, then define the corresponding power distribution area data as abnormal power distribution area data, and define the corresponding sampling time window as the abnormal sampling time window.

[0011] A6. Determine the corresponding list of power distribution equipment in the abnormal area based on the abnormal power distribution area data and the preset power distribution equipment installation and positioning data;

[0012] A7, collect the corresponding power distribution equipment time window data of each power distribution equipment according to the abnormal sampling time window and the power distribution equipment list data of the abnormal area;

[0013] A8 generates the corresponding equipment anomaly confidence level based on the time window data of each power distribution equipment through a preset equipment anomaly identification model;

[0014] A9. If the confidence level of the equipment abnormality is greater than the preset confidence level threshold of the equipment abnormality, then the corresponding power distribution equipment is defined as abnormal power distribution equipment.

[0015] By adopting the above technical solution, the intelligent monitoring method based on the power distribution room can first divide the power distribution room into multiple power distribution areas using spatial distribution data, then quickly locate abnormal areas using a pre-trained regional anomaly identification model, and then accurately monitor the equipment within the abnormal areas to finally identify the specific abnormal power distribution equipment. While realizing the monitoring of abnormal power distribution equipment, it avoids monitoring all equipment around the clock, reduces the amount of monitoring data processing, and improves monitoring efficiency. At the same time, the use of a dual threshold judgment mechanism effectively reduces the false alarm rate, providing intelligent early warning protection for the safe operation of the power distribution room.

[0016] Optionally, step A3 includes the following steps:

[0017] A301, based on the power distribution area data, continuously collects corresponding power distribution area temperature data, power distribution area humidity data, power distribution area magnetic field strength data, power distribution area noise intensity data, power distribution area ozone concentration data, and power distribution area nitrogen oxide concentration data at the corresponding locations in the power distribution room;

[0018] A302, based on the sampling time window, obtains the corresponding regional temperature time window data, regional humidity time window data, regional magnetic field strength time window data, regional noise intensity time window data, regional ozone concentration time window data, and regional nitrogen oxide concentration time window data from the power distribution area temperature data, power distribution area humidity data, power distribution area magnetic field strength data, and power distribution area noise intensity data, respectively.

[0019] A303 combines regional temperature time window data, regional humidity time window data, regional magnetic field strength time window data, regional noise intensity time window data, regional ozone concentration time window data, and regional nitrogen oxide concentration time window data to generate environmental time window data for the power distribution area.

[0020] By adopting the above technical solution, the intelligent monitoring method based on the power distribution room can simultaneously collect multi-dimensional environmental data such as temperature, humidity, magnetic field strength, noise intensity, ozone concentration, and nitrogen oxide concentration, and organize them into a structured environmental time window dataset according to time windows. This provides rich feature information for the regional anomaly identification model. Multi-parameter fusion can improve the sensitivity and accuracy of anomaly identification and can capture complex fault modes that are difficult to reflect by a single parameter. At the same time, the time window design realizes the time-series analysis of environmental data, which can effectively identify gradual anomalies and provide data basis for the preventive maintenance of power distribution equipment.

[0021] Optionally, step A4 includes the following steps:

[0022] A401 generates corresponding environmental feature data of the power distribution area based on the environmental time window data of the power distribution area through a preset environmental feature extraction algorithm.

[0023] A402, collect the power distribution external environment time window data corresponding to the external environment of the power distribution room according to the sampling time window;

[0024] A403, based on the power distribution external environment time window data, generates corresponding power distribution external environment feature data through the environmental feature extraction algorithm;

[0025] A404 generates comprehensive distribution environment characteristic data by combining distribution area environmental characteristic data and distribution external environment characteristic data;

[0026] A405, based on the comprehensive characteristic data of the power distribution environment, the corresponding regional anomaly confidence level is generated through the regional anomaly identification model.

[0027] By adopting the above technical solution, the intelligent monitoring method based on the power distribution room can simultaneously extract the internal environmental features of the power distribution area and the external environmental features of the power distribution room, and combine them into comprehensive feature data and input them into the regional anomaly identification model. This effectively eliminates the interference of external environmental factors on monitoring and analysis, enabling the system to accurately distinguish between environmental changes caused by equipment anomalies and environmental changes caused by external weather and other natural factors. This effectively improves the accuracy and reliability of anomaly detection. At the same time, by performing dimensionality reduction and feature enhancement on the original environmental data through environmental feature extraction algorithms, not only is the computational complexity reduced, but the model's sensitivity to minor anomalies is also enhanced, providing data support for early fault warning.

[0028] Optionally, step A7 includes the following steps:

[0029] A701 continuously collects the equipment current data, equipment voltage data, equipment power factor data, and equipment phase angle data corresponding to each power distribution equipment in the power distribution equipment list data of abnormal areas;

[0030] A702 obtains the corresponding device current time window data, device voltage time window data, device power factor time window data, and device phase angle time window data from the device current data, device voltage data, device power factor data, and device phase angle data respectively, based on the abnormal sampling time window.

[0031] A703 generates power distribution equipment time window data by combining equipment current time window data, equipment voltage time window data, equipment power factor time window data, and equipment phase angle time window data.

[0032] By adopting the above technical solution, the intelligent monitoring method based on the power distribution room can simultaneously collect multi-dimensional electrical data such as current, voltage, power factor, and phase angle, and integrate them into structured power distribution equipment time window data according to time windows. This provides comprehensive electrical status data for equipment anomaly identification. This multi-parameter fusion monitoring strategy can comprehensively reflect the operating status of the equipment and effectively capture various types of electrical anomalies, including overload, short circuit, insulation aging, and poor contact. It can not only detect sudden faults, but also identify early signs of gradual degradation of equipment performance, providing data basis for predictive maintenance and life assessment of power distribution equipment.

[0033] Optionally, step A8 includes the following steps:

[0034] A801 generates corresponding frequency domain data of equipment current, equipment voltage, equipment power factor, and equipment phase angle based on the equipment current time window data, equipment voltage time window data, equipment power factor time window data, and equipment phase angle time window data in the power distribution equipment time window data through a preset discrete Fourier transform algorithm.

[0035] A802 generates frequency domain data for power distribution equipment by combining frequency domain data of equipment current, frequency domain data of equipment voltage, frequency domain data of equipment power factor, and frequency domain data of equipment phase angle.

[0036] A803 generates corresponding power distribution equipment frequency domain feature data based on the power distribution equipment frequency domain data using a preset equipment feature extraction algorithm;

[0037] A804, based on the frequency domain characteristic data of the power distribution equipment, the corresponding equipment anomaly confidence level is generated by the equipment anomaly identification model.

[0038] By adopting the above technical solution, the intelligent monitoring method based on the power distribution room can convert the time-domain time window data of the power distribution equipment into corresponding frequency-domain data through the discrete Fourier transform algorithm, and use the equipment feature extraction algorithm to generate frequency-domain feature data as input to the equipment anomaly identification model. This can capture complex anomaly patterns such as harmonic distortion, subsynchronous oscillation, and intermittent faults that are difficult to detect by time-domain analysis, effectively improving the sensitivity and reliability of anomaly detection. At the same time, by analyzing the spectral characteristics of current, voltage, power factor, and phase angle, the system can not only accurately determine whether the equipment is abnormal, but also provide important data basis for fault type identification and root cause analysis.

[0039] Optionally, the intelligent monitoring method based on the power distribution room further includes the following steps:

[0040] B1, obtains historical power distribution area environmental data, historical power distribution external environment data, historical power distribution equipment data, and abnormal times and corresponding abnormal cause classification information of each historical power distribution equipment from the preset historical data collection database;

[0041] B2, determine the corresponding historical anomaly time window based on the anomaly time and sampling time window of each historical power distribution equipment;

[0042] B3, based on each historical anomaly time window, obtain the corresponding historical anomaly distribution area environmental time window data from the historical distribution area environmental data;

[0043] B4. Based on each historical anomaly time window, obtain the corresponding historical anomaly power distribution external environment data from the historical power distribution external environment data.

[0044] B5. Based on each historical anomaly time window, obtain the corresponding historical anomaly power distribution equipment time window data from the historical power distribution equipment data, and generate the corresponding historical anomaly power distribution equipment frequency domain data based on the historical anomaly power distribution equipment time window data through the discrete Fourier transform algorithm.

[0045] B6. Generate comprehensive historical abnormal power distribution time window data by combining historical abnormal power distribution area environmental time window data, historical abnormal power distribution external environment data, and historical abnormal power distribution equipment frequency domain data corresponding to each historical abnormal time window.

[0046] B7. Based on the comprehensive time window data of each historical abnormal power distribution, the corresponding comprehensive feature data of historical abnormal power distribution is generated through the preset comprehensive feature extraction algorithm and transformed into the corresponding comprehensive feature vector of historical abnormal power distribution.

[0047] B8. Based on the classification information of abnormal causes, combine the historical abnormal power distribution comprehensive feature vectors of the same type to generate the corresponding historical abnormal power distribution comprehensive feature vector set of the same type.

[0048] B9. Calculate the corresponding center vector based on the same set of vectors of the comprehensive feature vectors of each historical abnormal power distribution and define the typical feature vectors of the historical equipment abnormality category.

[0049] B10 generates a dataset for comparing the features of equipment anomaly causes by combining the typical feature vectors of all historical equipment anomaly categories with the corresponding anomaly cause classification information.

[0050] By adopting the above technical solution, the intelligent monitoring method based on the power distribution room can construct a dataset for comparing the characteristics of equipment abnormalities by acquiring historical abnormal data and extracting typical feature vectors. This can enhance the intelligent diagnostic capability of equipment abnormalities, not only accurately identify whether the equipment is abnormal, but also predict the type of abnormality through feature vector comparison. At the same time, through clustering of similar abnormal data and calculation of center vectors, the system can continuously accumulate and optimize the abnormal feature library, which can further improve the efficiency and accuracy of fault handling.

[0051] Optionally, the intelligent monitoring method based on the power distribution room further includes the following steps:

[0052] C1 generates current abnormal power distribution comprehensive time window data by combining the power distribution area environmental time window data, the power distribution external environment time window data, and the power distribution equipment frequency domain data corresponding to the abnormal power distribution equipment.

[0053] C2, Based on the current abnormal power distribution comprehensive time window data, the corresponding current abnormal power distribution comprehensive feature data is generated through the comprehensive feature extraction algorithm and transformed into the corresponding current abnormal power distribution comprehensive feature vector;

[0054] C3. Calculate the corresponding Euclidean distance based on the current abnormal power distribution comprehensive feature vector and the typical feature vector of each historical equipment abnormality category in the equipment abnormality cause feature comparison dataset, and define it as the feature category distance.

[0055] C4, determine the minimum value among all feature category distances and define the anomaly cause classification information corresponding to the typical feature vector of the historical device anomaly category as the first estimated anomaly cause information;

[0056] C5. Based on the current abnormal power distribution comprehensive feature vector and the equipment abnormal cause feature comparison dataset, calculate the corresponding cosine similarity of each historical equipment abnormality category and define it as feature cosine similarity.

[0057] C6, determine the maximum value among all feature cosine similarities and define the anomaly cause classification information corresponding to the typical feature vector of the historical device anomaly category as the second estimated anomaly cause information;

[0058] C7 generates a list of estimated anomaly causes for the current device based on a combination of the first and second estimated anomaly cause information.

[0059] By adopting the above technical solution, the intelligent monitoring method based on the power distribution room can compare the current abnormal power distribution comprehensive feature vector with the typical feature vector of historical equipment abnormality categories by using two different measurement methods, Euclidean distance and cosine similarity. This achieves dual matching judgment of the cause of the abnormality. Euclidean distance can assess the difference in the overall amplitude of the feature vectors, while cosine similarity can assess the similarity of the feature vectors in the direction. The combination of the two can comprehensively capture the subtle differences of abnormal features and effectively reduce the risk of misjudgment that may be caused by a single algorithm, thereby improving the accuracy and reliability of intelligent diagnosis of power distribution equipment faults.

[0060] This application also provides an intelligent monitoring system based on a power distribution room, including:

[0061] Power distribution environment data acquisition module;

[0062] Data acquisition module for power distribution equipment;

[0063] Data processing module;

[0064] Power distribution anomaly identification model;

[0065] The power distribution environment data acquisition module, the power distribution equipment data acquisition module, and the power distribution anomaly identification model are respectively connected to the data processing module.

[0066] The power distribution environment data acquisition module includes multiple regional environment data acquisition modules and an external environment data acquisition module, and each of the regional environment data acquisition modules and the external environment data acquisition module is connected to the data processing module.

[0067] The power distribution anomaly identification model includes a regional anomaly identification module and an equipment anomaly identification module, which are respectively connected to the data processing module.

[0068] The intelligent monitoring system based on the power distribution room further includes a power distribution anomaly monitoring strategy, comprising the following steps:

[0069] D1, determine the corresponding power distribution room spatial distribution data in the preset power distribution room;

[0070] D2, generates multiple power distribution area data by segmenting the spatial distribution data of the power distribution room;

[0071] D3, according to the preset sampling time window and the data of each power distribution area, collects the corresponding power distribution area environmental time window data at the corresponding location in the power distribution room through the power distribution environment data acquisition module;

[0072] D4. Based on the time window data of the power distribution area environment, the power distribution anomaly identification module generates the corresponding regional anomaly confidence level through the regional anomaly identification module.

[0073] D5. If the regional anomaly confidence level is greater than the preset regional anomaly confidence level threshold, then define the corresponding power distribution area data as abnormal power distribution area data, and define the corresponding sampling time window as the abnormal sampling time window.

[0074] D6. Determine the corresponding list of power distribution equipment in the abnormal area based on the abnormal power distribution area data and the preset power distribution equipment installation and positioning data;

[0075] D7, based on the abnormal sampling time window and the abnormal area power distribution equipment list data, the corresponding power distribution equipment time window data of each power distribution equipment is collected through the power distribution equipment data acquisition module;

[0076] D8, based on the time window data of each power distribution device, the power distribution anomaly identification module generates the corresponding device anomaly confidence level through the device anomaly identification module.

[0077] D9. If the confidence level of the equipment abnormality is greater than the preset confidence level threshold of the equipment abnormality, then the corresponding power distribution equipment is defined as abnormal power distribution equipment.

[0078] By adopting the above technical solution, the intelligent monitoring system based on the power distribution room can use a regional and hierarchical monitoring strategy. First, the power distribution room is divided into multiple power distribution areas using spatial distribution data. Then, a pre-trained regional anomaly identification model is used to quickly locate abnormal areas. Next, the equipment in the abnormal areas is precisely monitored, and finally, the specific abnormal power distribution equipment is identified. While realizing the monitoring of abnormal power distribution equipment, it avoids monitoring all equipment around the clock, reduces the amount of monitoring data processing, and improves monitoring efficiency. At the same time, the use of a dual threshold judgment mechanism effectively reduces the false alarm rate, providing intelligent early warning protection for the safe operation of the power distribution room.

[0079] In summary, this application includes at least one of the following beneficial technical effects:

[0080] 1. The intelligent monitoring method based on the power distribution room can adopt a regional and hierarchical monitoring strategy. First, the power distribution room is divided into multiple power distribution areas using spatial distribution data. Then, a pre-trained regional anomaly identification model is used to quickly locate the abnormal areas. Next, the equipment in the abnormal areas is precisely monitored, and finally, the specific abnormal power distribution equipment is identified. While realizing the monitoring of abnormal power distribution equipment, it avoids monitoring all equipment at all times, reduces the amount of monitoring data processing, and improves monitoring efficiency. At the same time, the dual threshold judgment mechanism effectively reduces the false alarm rate, providing intelligent early warning protection for the safe operation of the power distribution room.

[0081] 2. The intelligent monitoring method based on the power distribution room can simultaneously collect multi-dimensional environmental data such as temperature, humidity, magnetic field strength, noise intensity, ozone concentration, and nitrogen oxide concentration, and organize them into a structured environmental time window dataset according to time windows. This provides rich feature information for the regional anomaly identification model. Multi-parameter fusion can improve the sensitivity and accuracy of anomaly identification and can capture complex fault modes that are difficult to reflect by a single parameter. At the same time, the time window design realizes the time-series analysis of environmental data, which can effectively identify gradual anomalies and provide data basis for the preventive maintenance of power distribution equipment.

[0082] 3. The intelligent monitoring method based on the power distribution room can simultaneously extract the internal environmental features of the power distribution area and the external environmental features of the power distribution room, and combine them into comprehensive feature data, which is then input into the regional anomaly identification model. This effectively eliminates the interference of external environmental factors on monitoring and analysis, enabling the system to accurately distinguish between environmental changes caused by equipment anomalies and those caused by external weather and other natural factors. This effectively improves the accuracy and reliability of anomaly detection. At the same time, by performing dimensionality reduction and feature enhancement on the original environmental data through environmental feature extraction algorithms, not only is the computational complexity reduced, but the model's sensitivity to minor anomalies is also enhanced, providing data support for early fault warning. Attached Figure Description

[0083] Figure 1 This is a schematic diagram of the process of an intelligent monitoring method based on a power distribution room according to the present invention.

[0084] Figure 2 This is a schematic diagram of the principle of an intelligent monitoring system based on a power distribution room according to the present invention. Detailed Implementation

[0085] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0086] The embodiments of this application will now be described in further detail with reference to the accompanying drawings.

[0087] refer to Figure 1 This invention provides an intelligent monitoring method based on a power distribution room, used to monitor and detect abnormal equipment conditions in the power distribution room, including the following steps:

[0088] A1, determine the corresponding power distribution room spatial distribution data in the preset power distribution room;

[0089] The spatial distribution data of the power distribution room refers to the structural data of the indoor space of the power distribution room, which can be determined through measurement.

[0090] A2, generates multiple power distribution area data by segmenting the spatial distribution data of the power distribution room;

[0091] The power distribution area data is the block location data of the spatial area corresponding to the spatial distribution data of the power distribution room, used to represent each area in the power distribution room;

[0092] Data for each power distribution area can be generated manually or automatically using a specific algorithm.

[0093] A3, collect corresponding power distribution area environmental time window data at corresponding locations in the power distribution room according to the preset sampling time window and the data of each power distribution area;

[0094] The sampling time window is a preset time window used to determine the sampling time range of the data, which can be set according to needs or experience;

[0095] The environmental time window data of the power distribution area is the environmental data collected in the area corresponding to the location of the power distribution area data within the sampling time window. It can include various environmental variable data such as temperature data, humidity data, noise intensity data, and magnetic field intensity data.

[0096] A4. Based on the time window data of the power distribution area environment, the corresponding regional anomaly confidence level is generated through a pre-trained regional anomaly identification model.

[0097] The regional anomaly identification model is a pre-trained identification model used to identify whether there are anomalies in the environment of the power distribution area based on the environmental time window data of the power distribution area.

[0098] The regional anomaly confidence level is the probability that an anomaly exists in the power distribution area, which is identified by the regional anomaly identification model based on the environmental time window data of the power distribution area.

[0099] When power distribution equipment malfunctions, it usually affects the surrounding environment, such as causing the temperature to rise, generating significant noise, or abnormal magnetic field changes. Therefore, by collecting environmental data of the corresponding area, it is possible to identify and predict whether there are any abnormalities in the area, and then further investigate the power distribution equipment in the abnormal area.

[0100] A5. If the regional anomaly confidence level is greater than the preset regional anomaly confidence level threshold, then define the corresponding power distribution area data as abnormal power distribution area data, and define the corresponding sampling time window as the abnormal sampling time window.

[0101] The regional anomaly confidence threshold is a pre-set reference value used to determine the magnitude of regional anomaly confidence.

[0102] Abnormal power distribution area data refers to power distribution area data where the area abnormality confidence level is greater than the area abnormality confidence threshold, that is, there may be abnormal power distribution equipment in the corresponding area of ​​the power distribution room.

[0103] The anomaly sampling window is the sampling time window corresponding to the environmental time window data of the power distribution area corresponding to the regional anomaly confidence level that is greater than the regional anomaly confidence level threshold.

[0104] A6. Determine the corresponding list of power distribution equipment in the abnormal area based on the abnormal power distribution area data and the preset power distribution equipment installation and positioning data;

[0105] The installation and positioning data of the power distribution equipment are the pre-set installation and positioning data of each power distribution equipment in the power distribution room;

[0106] The abnormal area power distribution equipment list data is a list of information on all power distribution equipment located within the abnormal power distribution area data.

[0107] A7, collect the corresponding power distribution equipment time window data of each power distribution equipment according to the abnormal sampling time window and the power distribution equipment list data of the abnormal area;

[0108] The power distribution equipment time window data consists of the equipment parameter data of each power distribution equipment in the abnormal area power distribution equipment list data within the abnormal sampling time window. For example, the current data, voltage data, power factor data, phase angle data, apparent power data, active power data, and reactive power data of the power distribution equipment, and other various equipment electrical parameter data.

[0109] A8 generates the corresponding equipment anomaly confidence level based on the time window data of each power distribution equipment through a preset equipment anomaly identification model;

[0110] The equipment anomaly identification model is a pre-trained identification model used to identify whether there are any malfunctions in the power distribution equipment.

[0111] The equipment anomaly confidence level is the degree of probability that the power distribution equipment identified by the equipment anomaly identification model based on the power distribution equipment time window data is abnormal.

[0112] A9. If the confidence level of the equipment abnormality is greater than the preset confidence level threshold of the equipment abnormality, then the corresponding power distribution equipment is defined as abnormal power distribution equipment.

[0113] The equipment anomaly confidence threshold is a pre-set reference value used to determine the degree of confidence in the equipment anomaly.

[0114] Abnormal power distribution equipment refers to power distribution equipment whose corresponding abnormality confidence level is greater than the abnormality confidence threshold.

[0115] Through the above steps, the intelligent monitoring method based on the power distribution room can adopt a regional and hierarchical monitoring strategy. First, the power distribution room is divided into multiple power distribution areas using spatial distribution data. Then, a pre-trained regional anomaly identification model is used to quickly locate the abnormal areas. Next, the equipment in the abnormal areas is precisely monitored, and finally, the specific abnormal power distribution equipment is identified. While realizing the monitoring of abnormal power distribution equipment, it avoids monitoring all equipment around the clock, reduces the amount of monitoring data processing, and improves monitoring efficiency. At the same time, the use of a dual threshold judgment mechanism effectively reduces the false alarm rate, providing intelligent early warning protection for the safe operation of the power distribution room.

[0116] Further, step A3 includes the following steps:

[0117] A301, based on the power distribution area data, continuously collects corresponding power distribution area temperature data, power distribution area humidity data, power distribution area magnetic field strength data, power distribution area noise intensity data, power distribution area ozone concentration data, and power distribution area nitrogen oxide concentration data at the corresponding locations in the power distribution room;

[0118] The temperature data for the power distribution area is the temperature data continuously collected at the corresponding locations in the power distribution area.

[0119] Humidity data in the power distribution area is continuously collected from the corresponding locations in the power distribution area.

[0120] The magnetic field strength data of the power distribution area is obtained by continuously collecting magnetic field strength data at the corresponding locations in the power distribution area.

[0121] When power distribution equipment operates abnormally under load, it may cause abnormal changes in the surrounding temperature, humidity and magnetic field;

[0122] The noise intensity data of the power distribution area is the noise intensity data obtained by continuously collecting data at the corresponding locations in the power distribution area.

[0123] When a structural abnormality occurs in the power distribution equipment, it may cause the equipment to emit abnormal noise to the surrounding environment.

[0124] The ozone concentration data in the power distribution area is obtained by continuously collecting ozone concentration data at the corresponding locations in the power distribution area.

[0125] Nitrogen oxide concentration data in the power distribution area is obtained by continuously collecting nitrogen oxide concentration data at the corresponding locations in the power distribution area.

[0126] When power distribution equipment experiences abnormal discharge, it may cause an abnormal increase in the concentration of ozone or nitrogen oxides in its surroundings.

[0127] A302, based on the sampling time window, obtains the corresponding regional temperature time window data, regional humidity time window data, regional magnetic field strength time window data, regional noise intensity time window data, regional ozone concentration time window data, and regional nitrogen oxide concentration time window data from the power distribution area temperature data, power distribution area humidity data, power distribution area magnetic field strength data, and power distribution area noise intensity data, respectively.

[0128] The area temperature time window data refers to the temperature data of the power distribution area within the sampling time window.

[0129] The regional humidity time window data refers to the humidity data of the power distribution area within the sampling time window.

[0130] The regional magnetic field strength time window data refers to the magnetic field strength data of the power distribution area within the sampling time window.

[0131] The area noise intensity time window data refers to the noise intensity data of the power distribution area within the sampling time window.

[0132] The regional ozone concentration time window data is the ozone concentration data within the sampling time window of the power distribution area noise intensity data.

[0133] The regional nitrogen oxide concentration time window data refers to the nitrogen oxide concentration data within the sampling time window of the power distribution area noise intensity data.

[0134] A303 combines regional temperature time window data, regional humidity time window data, regional magnetic field strength time window data, regional noise intensity time window data, regional ozone concentration time window data, and regional nitrogen oxide concentration time window data to generate environmental time window data for the power distribution area;

[0135] The environmental time window data for the power distribution area is a collection of data including regional temperature time window data, regional humidity time window data, regional magnetic field strength time window data, regional noise intensity time window data, regional ozone concentration time window data, and regional nitrogen oxide concentration time window data.

[0136] Through the above steps, the intelligent monitoring method based on the power distribution room can simultaneously collect multi-dimensional environmental data such as temperature, humidity, magnetic field strength, noise intensity, ozone concentration, and nitrogen oxide concentration, and organize them into a structured environmental time window dataset according to time windows. This provides rich feature information for the regional anomaly identification model. Multi-parameter fusion can improve the sensitivity and accuracy of anomaly identification and can capture complex fault modes that are difficult to reflect by a single parameter. At the same time, the time window design realizes the time-series analysis of environmental data, which can effectively identify gradual anomalies and provide data basis for the preventive maintenance of power distribution equipment.

[0137] Further, step A4 includes the following steps:

[0138] A401 generates corresponding environmental feature data of the power distribution area based on the environmental time window data of the power distribution area through a preset environmental feature extraction algorithm.

[0139] The environmental feature extraction algorithm is a pre-defined algorithm used to extract feature data from the environmental time window data of the power distribution area;

[0140] The environmental characteristic data of the power distribution area are the characteristic data corresponding to the time window data of the power distribution area environment.

[0141] A402, collect the power distribution external environment time window data corresponding to the external environment of the power distribution room according to the sampling time window;

[0142] The external environment time window data of the power distribution room refers to the environmental data of the external environment of the power distribution room, such as temperature data, humidity data, wind speed data and other environmental data outside the power distribution room. Since the environmental data outside the power distribution room may indirectly affect the environmental data inside the power distribution room, it is necessary to collect the corresponding data for processing and analysis.

[0143] A403, based on the power distribution external environment time window data, generates corresponding power distribution external environment feature data through the environmental feature extraction algorithm;

[0144] The power distribution external environment characteristic data are the characteristic data corresponding to the power distribution external environment time window data.

[0145] A404 generates comprehensive distribution environment characteristic data by combining distribution area environmental characteristic data and distribution external environment characteristic data;

[0146] The comprehensive characteristic data of the power distribution environment is a collection of data on the environmental characteristics of the power distribution area and the external environmental characteristics of the power distribution.

[0147] A405, Based on the comprehensive characteristic data of the power distribution environment, the corresponding regional anomaly confidence level is generated through the regional anomaly identification model;

[0148] By inputting comprehensive characteristic data of the power distribution environment into the regional anomaly identification model to obtain the output regional anomaly confidence score, and combining the regional environmental characteristic data of the power distribution area with the external environmental characteristic data of the power distribution, the regional anomaly confidence score output by the regional anomaly identification model can have better accuracy.

[0149] Through the above steps, the intelligent monitoring method based on the power distribution room can simultaneously extract the internal environmental features of the power distribution area and the external environmental features of the power distribution room, and combine them into comprehensive feature data, which is then input into the regional anomaly identification model. This effectively eliminates the interference of external environmental factors on monitoring and analysis, enabling the system to accurately distinguish between environmental changes caused by equipment anomalies and those caused by external weather and other natural factors. This effectively improves the accuracy and reliability of anomaly detection. At the same time, by performing dimensionality reduction and feature enhancement on the original environmental data through environmental feature extraction algorithms, not only is the computational complexity reduced, but the model's sensitivity to minor anomalies is also enhanced, providing data support for early fault warning.

[0150] Further, step A7 includes the following steps:

[0151] A701 continuously collects the equipment current data, equipment voltage data, equipment power factor data, and equipment phase angle data corresponding to each power distribution equipment in the power distribution equipment list data of abnormal areas;

[0152] The equipment current data refers to the current data of the power distribution equipment;

[0153] The equipment voltage data refers to the voltage data of the power distribution equipment;

[0154] The equipment power factor data refers to the power factor data of the power distribution equipment;

[0155] The equipment phase angle data refers to the phase angle data of the power distribution equipment.

[0156] A702 obtains the corresponding device current time window data, device voltage time window data, device power factor time window data, and device phase angle time window data from the device current data, device voltage data, device power factor data, and device phase angle data respectively, based on the abnormal sampling time window.

[0157] The device current time window data is the corresponding data in the device current data for the abnormal sampling time window;

[0158] The device voltage time window data is the corresponding data in the device voltage data for the abnormal sampling time window;

[0159] The device power factor time window data is the corresponding data in the device power factor data for the abnormal sampling time window;

[0160] The device phase angle time window data is the corresponding data of the abnormal sampling time window in the device phase angle time window data.

[0161] A703 generates power distribution equipment time window data by combining equipment current time window data, equipment voltage time window data, equipment power factor time window data, and equipment phase angle time window data;

[0162] The power distribution equipment time window data is a collection of equipment current time window data, equipment voltage time window data, equipment power factor time window data, and equipment phase angle time window data.

[0163] Through the above steps, the intelligent monitoring method based on the power distribution room can simultaneously collect multi-dimensional electrical data such as current, voltage, power factor, and phase angle, and integrate them into structured power distribution equipment time window data according to time windows. This provides comprehensive electrical status data for equipment anomaly identification. This multi-parameter fusion monitoring strategy can comprehensively reflect the operating status of the equipment and effectively capture various types of electrical anomalies, including overload, short circuit, insulation aging, and poor contact. It can not only detect sudden faults, but also identify early signs of gradual degradation of equipment performance, providing data basis for predictive maintenance and life assessment of power distribution equipment.

[0164] Further, step A8 includes the following steps:

[0165] A801 generates corresponding frequency domain data of equipment current, equipment voltage, equipment power factor, and equipment phase angle based on the equipment current time window data, equipment voltage time window data, equipment power factor time window data, and equipment phase angle time window data in the power distribution equipment time window data through a preset discrete Fourier transform algorithm.

[0166] The Discrete Fourier Transform algorithm is a pre-defined algorithm used to convert time-domain data into frequency-domain data;

[0167] The frequency domain data of the equipment current is the frequency domain data corresponding to the time window data of the equipment current.

[0168] The device voltage frequency domain data is the frequency domain data corresponding to the device voltage time window data;

[0169] The frequency domain data of the device power factor is the frequency domain data corresponding to the time window data of the device power factor;

[0170] The device phase angle frequency domain data is the frequency domain data corresponding to the device phase angle time window data.

[0171] A802 generates frequency domain data for power distribution equipment by combining frequency domain data of equipment current, frequency domain data of equipment voltage, frequency domain data of equipment power factor, and frequency domain data of equipment phase angle.

[0172] The frequency domain data of power distribution equipment is a collection of equipment current frequency domain data, equipment voltage frequency domain data, equipment power factor frequency domain data, and equipment phase angle frequency domain data.

[0173] A803 generates corresponding power distribution equipment frequency domain feature data based on the power distribution equipment frequency domain data using a preset equipment feature extraction algorithm;

[0174] The equipment feature extraction algorithm is a pre-defined algorithm used to extract feature data from the frequency domain data of power distribution equipment;

[0175] The frequency domain characteristic data of power distribution equipment is the characteristic data corresponding to the frequency domain data of power distribution equipment.

[0176] A804, based on the frequency domain characteristic data of the power distribution equipment, the corresponding equipment anomaly confidence level is generated by the equipment anomaly identification model;

[0177] The frequency domain feature data of the power distribution equipment is input into the equipment anomaly identification model to generate the output equipment anomaly confidence level of the power distribution equipment.

[0178] Through the above steps, the intelligent monitoring method based on the power distribution room can convert the time-domain time window data of the power distribution equipment into corresponding frequency-domain data using the discrete Fourier transform algorithm. It then uses a device feature extraction algorithm to generate frequency-domain feature data as input to the device anomaly identification model. This method can capture complex anomaly patterns such as harmonic distortion, subsynchronous oscillations, and intermittent faults that are difficult to detect through time-domain analysis, effectively improving the sensitivity and reliability of anomaly detection. Simultaneously, by analyzing the spectral characteristics of current, voltage, power factor, and phase angle, the system can not only accurately determine whether the equipment is abnormal but also provide important data for fault type identification and root cause analysis.

[0179] Furthermore, the intelligent monitoring method based on the power distribution room further includes the following steps:

[0180] B1, obtains historical power distribution area environmental data, historical power distribution external environment data, historical power distribution equipment data, and abnormal times and corresponding abnormal cause classification information of each historical power distribution equipment from the preset historical data collection database;

[0181] The historical data collection database is a pre-defined database of data related to the power distribution room. It stores various data collected in the past, including historical power distribution environment data, historical external environment data, various electrical data of each power distribution device, and historical abnormal conditions of the power distribution device.

[0182] Historical distribution area environmental data refers to historical data on the distribution area environment that has been continuously collected over the past.

[0183] Historical power distribution external environment data refers to historical data on the power distribution external environment that has been continuously collected over the past.

[0184] Historical power distribution equipment data refers to the historical electrical status data of power distribution equipment collected in the past;

[0185] Historical abnormal moments of power distribution equipment refer to the moments in the past when the power distribution equipment experienced abnormal operating conditions, determined manually or by algorithms.

[0186] The abnormal cause classification information is the classification of abnormal causes corresponding to the abnormal working conditions of power distribution equipment identified in the past investigations.

[0187] B2, determine the corresponding historical anomaly time window based on the anomaly time and sampling time window of each historical power distribution equipment;

[0188] The historical anomaly time window is the corresponding time window determined by combining the abnormal time of each historical power distribution equipment with the sampling time window. For example, if the abnormal time of a certain historical power distribution equipment is 17:23 on March 5, and the sampling time window is set to ±5 minutes, then the corresponding historical anomaly time window is from 17:18 on March 5 to 17:28 on March 5.

[0189] B3, based on each historical anomaly time window, obtain the corresponding historical anomaly distribution area environmental time window data from the historical distribution area environmental data;

[0190] The historical abnormal power distribution area environmental time window data is the data corresponding to the historical abnormal time window in the historical power distribution area environmental data.

[0191] B4. Based on each historical anomaly time window, obtain the corresponding historical anomaly power distribution external environment data from the historical power distribution external environment data.

[0192] Historical abnormal power distribution external environment data refers to the data corresponding to the time period of the historical abnormal time window in the historical power distribution external environment data.

[0193] B5. Based on each historical anomaly time window, obtain the corresponding historical anomaly power distribution equipment time window data from the historical power distribution equipment data, and generate the corresponding historical anomaly power distribution equipment frequency domain data based on the historical anomaly power distribution equipment time window data through the discrete Fourier transform algorithm.

[0194] Historical abnormal power distribution equipment time window data refers to the electrical data of various aspects of power distribution during the corresponding time period of the historical abnormal power distribution equipment data, including current data, voltage data, power factor data, and phase angle data.

[0195] The frequency domain data of historical abnormal power distribution equipment is the frequency domain data corresponding to the time window data of historical abnormal power distribution equipment.

[0196] B6. Generate comprehensive historical abnormal power distribution time window data by combining historical abnormal power distribution area environmental time window data, historical abnormal power distribution external environment data, and historical abnormal power distribution equipment frequency domain data corresponding to each historical abnormal time window.

[0197] The historical abnormal power distribution comprehensive time window data is a collection of historical abnormal power distribution area environmental time window data, historical abnormal power distribution external environment data, and historical abnormal power distribution equipment frequency domain data.

[0198] B7. Based on the comprehensive time window data of each historical abnormal power distribution, the corresponding comprehensive feature data of historical abnormal power distribution is generated through the preset comprehensive feature extraction algorithm and transformed into the corresponding comprehensive feature vector of historical abnormal power distribution.

[0199] The comprehensive feature extraction algorithm is a pre-defined feature extraction algorithm used to extract feature data from the abnormal power distribution comprehensive time window data that combines data from multiple sources;

[0200] The comprehensive feature data of historical abnormal power distribution is the feature data corresponding to the comprehensive time window data of historical abnormal power distribution.

[0201] The historical abnormal power distribution comprehensive feature vector is the feature vector corresponding to the historical abnormal power distribution comprehensive feature data.

[0202] B8. Based on the classification information of abnormal causes, combine the historical abnormal power distribution comprehensive feature vectors of the same type to generate the corresponding historical abnormal power distribution comprehensive feature vector set of the same type.

[0203] The historical abnormal power distribution comprehensive feature vector class set is a set of historical abnormal power distribution comprehensive feature vectors that have the same abnormal cause classification information. In other words, the historical abnormal power distribution comprehensive feature vectors corresponding to the historical abnormal time windows with the same abnormal cause classification information are classified into one set.

[0204] B9. Calculate the corresponding center vector based on the same set of vectors of the comprehensive feature vectors of each historical abnormal power distribution and define the typical feature vectors of the historical equipment abnormality category.

[0205] The typical feature vector of historical equipment anomaly categories is the central vector of all feature vectors in the same category of historical abnormal power distribution comprehensive feature vector.

[0206] B10 generates a device anomaly cause feature comparison dataset by combining typical feature vectors of all historical device anomaly categories and corresponding anomaly cause classification information.

[0207] The equipment anomaly cause feature comparison dataset is a paired dataset of typical feature vectors of each historical equipment anomaly category and the corresponding anomaly cause classification information.

[0208] Through the above steps, the intelligent monitoring method based on the power distribution room can construct a dataset for comparing the characteristics of equipment anomalies by acquiring historical abnormal data and extracting typical feature vectors. This can enhance the intelligent diagnostic capability of equipment anomalies, not only accurately identifying whether the equipment is abnormal, but also predicting the type of anomaly through feature vector comparison. At the same time, through clustering of similar abnormal data and calculation of center vectors, the system can continuously accumulate and optimize the anomaly feature library, further improving the efficiency and accuracy of fault handling.

[0209] Furthermore, the intelligent monitoring method based on the power distribution room further includes the following steps:

[0210] C1 generates current abnormal power distribution comprehensive time window data by combining the power distribution area environmental time window data, the power distribution external environment time window data, and the power distribution equipment frequency domain data corresponding to the abnormal power distribution equipment.

[0211] The current abnormal power distribution comprehensive time window data is a collection of data including the power distribution area environment time window data, the power distribution external environment time window data, and the power distribution equipment frequency domain data corresponding to the abnormal power distribution equipment.

[0212] C2, Based on the current abnormal power distribution comprehensive time window data, the corresponding current abnormal power distribution comprehensive feature data is generated through the comprehensive feature extraction algorithm and transformed into the corresponding current abnormal power distribution comprehensive feature vector;

[0213] The current abnormal power distribution comprehensive feature data is the feature data corresponding to the current abnormal power distribution comprehensive time window data;

[0214] The current abnormal power distribution comprehensive feature vector is the feature vector corresponding to the current abnormal power distribution comprehensive feature data.

[0215] C3. Calculate the corresponding Euclidean distance based on the current abnormal power distribution comprehensive feature vector and the typical feature vector of each historical equipment abnormality category in the equipment abnormality cause feature comparison dataset, and define it as the feature category distance.

[0216] The feature category distance is the Euclidean distance between the current abnormal power distribution comprehensive feature vector and the typical feature vectors of each historical equipment abnormal category. The smaller the feature category distance, the higher the similarity between the vectors.

[0217] C4, determine the minimum value among all feature category distances and define the anomaly cause classification information corresponding to the typical feature vector of the historical device anomaly category as the first estimated anomaly cause information;

[0218] The first estimated anomaly cause information is the possible anomaly cause information of the power distribution equipment determined by matching the minimum value among all feature category distances.

[0219] C5. Based on the current abnormal power distribution comprehensive feature vector and the equipment abnormal cause feature comparison dataset, calculate the corresponding cosine similarity of each historical equipment abnormality category and define it as feature cosine similarity.

[0220] The cosine similarity is the cosine similarity between the current abnormal power distribution integrated feature vector and the typical feature vectors of each historical equipment abnormality category. The larger the cosine similarity, the higher the similarity between the vectors.

[0221] C6, determine the maximum value among all feature cosine similarities and define the anomaly cause classification information corresponding to the typical feature vector of the historical device anomaly category as the second estimated anomaly cause information;

[0222] The second estimated anomaly cause information is the possible anomaly cause information of the power distribution equipment determined by matching the maximum value among all feature cosine similarity.

[0223] C7, Generate a list of estimated anomalies for the current device based on the combination of the first and second estimated anomaly cause information;

[0224] The current list of predicted causes of equipment anomalies is a combination of two aspects of predicted anomaly information: the first predicted cause information and the second predicted cause information.

[0225] Through the above steps, the intelligent monitoring method based on the power distribution room can compare the current abnormal power distribution comprehensive feature vector with the typical feature vector of historical equipment abnormality categories by simultaneously using two different measurement methods: Euclidean distance and cosine similarity. This achieves dual matching judgment of the cause of the abnormality. Euclidean distance can assess the difference in the overall amplitude of the feature vectors, while cosine similarity can assess the similarity of the feature vectors in the direction. The combination of the two can comprehensively capture the subtle differences in abnormal features and effectively reduce the risk of misjudgment that may be caused by a single algorithm, thereby improving the accuracy and reliability of intelligent diagnosis of power distribution equipment faults.

[0226] refer to Figure 2 The present invention also provides an intelligent monitoring system based on a power distribution room, comprising:

[0227] Power distribution environment data acquisition module 10;

[0228] Power distribution equipment data acquisition module 20;

[0229] Data processing module 30;

[0230] Power distribution anomaly identification module 40;

[0231] The power distribution environment data acquisition module 10, the power distribution equipment data acquisition module 20, and the power distribution anomaly identification module 40 are respectively connected to the data processing module 30.

[0232] The power distribution environment data acquisition module 10 includes multiple regional environment data acquisition modules 11 and external environment data acquisition modules 12, and each of the regional environment data acquisition modules 11 and the external environment data acquisition modules 12 is connected to the data processing module 30.

[0233] The power distribution anomaly identification module 40 includes a regional anomaly identification module 41 and an equipment anomaly identification module 42, which are respectively connected to the data processing module 30.

[0234] The power distribution environment data acquisition module 10 is mainly used to collect environmental data inside and outside the power distribution room;

[0235] The regional environmental data acquisition module 11 is mainly used to collect environmental data of each area in the power distribution room;

[0236] The external environment data acquisition module 12 is mainly used to collect environmental data outside the power distribution room.

[0237] The power distribution equipment data acquisition module 20 is mainly used to collect the electrical status data of each power distribution equipment.

[0238] The data processing module 30 is mainly used for data collection and analysis.

[0239] The power distribution anomaly identification module 40 is mainly used for anomaly identification based on the data processed by the data processing module 30.

[0240] The regional anomaly identification module 41 is mainly used to identify and determine whether there is an anomaly in the corresponding area based on the environmental data collected by the power distribution environment data acquisition module 10.

[0241] The equipment anomaly identification module 42 is mainly used to determine whether there is an anomaly in the power distribution equipment based on the electrical data collected by the power distribution equipment data acquisition module 20.

[0242] The intelligent monitoring system based on the power distribution room further includes a power distribution anomaly monitoring strategy, comprising the following steps:

[0243] D1, determine the corresponding power distribution room spatial distribution data in the preset power distribution room;

[0244] D2, generates multiple power distribution area data by segmenting the spatial distribution data of the power distribution room;

[0245] D3, according to the preset sampling time window and the data of each power distribution area, the corresponding power distribution area environmental time window data is collected at the corresponding position in the power distribution room through the power distribution environment data acquisition module 10;

[0246] D4. Based on the distribution area environmental time window data, the distribution anomaly identification module generates the corresponding area anomaly confidence level through the area anomaly identification module 41.

[0247] D5. If the regional anomaly confidence level is greater than the preset regional anomaly confidence level threshold, then define the corresponding power distribution area data as abnormal power distribution area data, and define the corresponding sampling time window as the abnormal sampling time window.

[0248] D6. Determine the corresponding list of power distribution equipment in the abnormal area based on the abnormal power distribution area data and the preset power distribution equipment installation and positioning data;

[0249] D7, based on the abnormal sampling time window and the abnormal area power distribution equipment list data, the power distribution equipment data acquisition module 20 collects the corresponding power distribution equipment time window data of each power distribution equipment;

[0250] D8, based on the time window data of each power distribution device, the power distribution anomaly identification module generates the corresponding device anomaly confidence level through the device anomaly identification module 42;

[0251] D9. If the confidence level of the equipment abnormality is greater than the preset confidence level threshold of the equipment abnormality, then the corresponding power distribution equipment is defined as abnormal power distribution equipment.

[0252] Through the above technical solutions, the intelligent monitoring system based on the power distribution room can adopt a regional and hierarchical monitoring strategy. First, the power distribution room is divided into multiple power distribution areas using spatial distribution data. Then, a pre-trained regional anomaly identification model is used to quickly locate abnormal areas. Next, the equipment within the abnormal areas is precisely monitored, and finally, the specific abnormal power distribution equipment is identified. While realizing the monitoring of abnormal power distribution equipment, it avoids monitoring all equipment around the clock, reduces the amount of monitoring data processing, and improves monitoring efficiency. At the same time, the dual threshold judgment mechanism effectively reduces the false alarm rate, providing intelligent early warning protection for the safe operation of the power distribution room.

[0253] The above are all preferred embodiments of this application and are not intended to limit the scope of protection of this application. Any feature disclosed in this specification (including the abstract and drawings) may be replaced by other equivalent or similar features unless specifically stated otherwise. That is, unless specifically stated otherwise, each feature is only one example of a series of equivalent or similar features.

Claims

1. A smart monitoring method based on a power distribution room, characterized in that, Includes the following steps: A1, determine the corresponding power distribution room spatial distribution data in the preset power distribution room; A2, generates multiple power distribution area data by segmenting the spatial distribution data of the power distribution room; A3, according to the preset sampling time window and the data of each power distribution area, collect the corresponding power distribution area environmental time window data at the corresponding location in the power distribution room; A4. Based on the time window data of the power distribution area environment, the corresponding regional anomaly confidence level is generated through a pre-trained regional anomaly identification model. A5. If the regional anomaly confidence level is greater than the preset regional anomaly confidence level threshold, then define the corresponding power distribution area data as abnormal power distribution area data, and define the corresponding sampling time window as the abnormal sampling time window. A6. Determine the corresponding list of power distribution equipment in the abnormal area based on the abnormal power distribution area data and the preset power distribution equipment installation and positioning data; A7, collect the corresponding power distribution equipment time window data of each power distribution equipment according to the abnormal sampling time window and the power distribution equipment list data of the abnormal area; A8 generates the corresponding equipment anomaly confidence level based on the time window data of each power distribution equipment through a preset equipment anomaly identification model; A9. If the confidence level of the equipment abnormality is greater than the preset confidence level threshold of the equipment abnormality, then the corresponding power distribution equipment is defined as abnormal power distribution equipment.

2. The intelligent monitoring method based on a power distribution room according to claim 1, characterized in that, Step A3 includes the following steps: A301, based on the power distribution area data, continuously collects corresponding power distribution area temperature data, power distribution area humidity data, power distribution area magnetic field strength data, power distribution area noise intensity data, power distribution area ozone concentration data, and power distribution area nitrogen oxide concentration data at the corresponding locations in the power distribution room; A302, based on the sampling time window, obtains the corresponding regional temperature time window data, regional humidity time window data, regional magnetic field strength time window data, regional noise intensity time window data, regional ozone concentration time window data, and regional nitrogen oxide concentration time window data from the power distribution area temperature data, power distribution area humidity data, power distribution area magnetic field strength data, and power distribution area noise intensity data, respectively. A303 combines regional temperature time window data, regional humidity time window data, regional magnetic field strength time window data, regional noise intensity time window data, regional ozone concentration time window data, and regional nitrogen oxide concentration time window data to generate environmental time window data for the power distribution area.

3. The intelligent monitoring method based on a power distribution room according to claim 2, characterized in that, Step A4 includes the following steps: A401 generates corresponding environmental feature data of the power distribution area based on the environmental time window data of the power distribution area through a preset environmental feature extraction algorithm. A402, collect the power distribution external environment time window data corresponding to the external environment of the power distribution room according to the sampling time window; A403, based on the power distribution external environment time window data, generates corresponding power distribution external environment feature data through the environmental feature extraction algorithm; A404 generates comprehensive distribution environment characteristic data by combining distribution area environmental characteristic data and distribution external environment characteristic data; A405, based on the comprehensive characteristic data of the power distribution environment, the corresponding regional anomaly confidence level is generated through the regional anomaly identification model.

4. The intelligent monitoring method based on a power distribution room according to claim 3, characterized in that, Step A7 includes the following steps: A701 continuously collects the equipment current data, equipment voltage data, equipment power factor data, and equipment phase angle data corresponding to each power distribution equipment in the power distribution equipment list data of abnormal areas; A702 obtains the corresponding device current time window data, device voltage time window data, device power factor time window data, and device phase angle time window data from the device current data, device voltage data, device power factor data, and device phase angle data respectively, based on the abnormal sampling time window. A703 generates power distribution equipment time window data by combining equipment current time window data, equipment voltage time window data, equipment power factor time window data, and equipment phase angle time window data.

5. The intelligent monitoring method based on a power distribution room according to claim 4, characterized in that, Step A8 includes the following steps: A801 generates corresponding frequency domain data of equipment current, equipment voltage, equipment power factor, and equipment phase angle based on the equipment current time window data, equipment voltage time window data, equipment power factor time window data, and equipment phase angle time window data in the power distribution equipment time window data through a preset discrete Fourier transform algorithm. A802 generates frequency domain data for power distribution equipment by combining frequency domain data of equipment current, frequency domain data of equipment voltage, frequency domain data of equipment power factor, and frequency domain data of equipment phase angle. A803 generates corresponding power distribution equipment frequency domain feature data based on the power distribution equipment frequency domain data using a preset equipment feature extraction algorithm; A804, based on the frequency domain characteristic data of the power distribution equipment, the corresponding equipment anomaly confidence level is generated by the equipment anomaly identification model.

6. The intelligent monitoring method based on a power distribution room according to claim 5, characterized in that, Further steps include: B1, obtains historical power distribution area environmental data, historical power distribution external environment data, historical power distribution equipment data, and abnormal times and corresponding abnormal cause classification information of each historical power distribution equipment from the preset historical data collection database; B2, determine the corresponding historical anomaly time window based on the anomaly time and sampling time window of each historical power distribution equipment; B3, based on each historical anomaly time window, obtain the corresponding historical anomaly distribution area environmental time window data from the historical distribution area environmental data; B4. Obtain the corresponding historical abnormal power distribution external environment data based on the historical power distribution external environment data for each historical abnormal time window. B5. Based on each historical anomaly time window, obtain the corresponding historical anomaly power distribution equipment time window data from the historical power distribution equipment data, and generate the corresponding historical anomaly power distribution equipment frequency domain data based on the historical anomaly power distribution equipment time window data through the discrete Fourier transform algorithm. B6. Generate comprehensive historical abnormal power distribution time window data by combining historical abnormal power distribution area environmental time window data, historical abnormal power distribution external environment data, and historical abnormal power distribution equipment frequency domain data corresponding to each historical abnormal time window. B7. Based on the comprehensive time window data of each historical abnormal power distribution, the corresponding comprehensive feature data of historical abnormal power distribution is generated through the preset comprehensive feature extraction algorithm and transformed into the corresponding comprehensive feature vector of historical abnormal power distribution. B8. Based on the classification information of abnormal causes, combine the historical abnormal power distribution comprehensive feature vectors of the same type to generate the corresponding historical abnormal power distribution comprehensive feature vector set of the same type. B9. Calculate the corresponding center vector based on the same set of vectors of the comprehensive feature vectors of each historical abnormal power distribution and define the typical feature vectors of the historical equipment abnormality category. B10 generates a dataset for comparing the features of equipment anomaly causes by combining the typical feature vectors of all historical equipment anomaly categories with the corresponding anomaly cause classification information.

7. The intelligent monitoring method based on a power distribution room according to claim 6, characterized in that, Further steps include: C1 generates current abnormal power distribution comprehensive time window data by combining the power distribution area environmental time window data, the power distribution external environment time window data, and the power distribution equipment frequency domain data corresponding to the abnormal power distribution equipment. C2, Based on the current abnormal power distribution comprehensive time window data, the corresponding current abnormal power distribution comprehensive feature data is generated through the comprehensive feature extraction algorithm and transformed into the corresponding current abnormal power distribution comprehensive feature vector; C3. Calculate the corresponding Euclidean distance based on the current abnormal power distribution comprehensive feature vector and the typical feature vector of each historical equipment abnormality category in the equipment abnormality cause feature comparison dataset, and define it as the feature category distance. C4, determine the minimum value among all feature category distances and define the anomaly cause classification information corresponding to the typical feature vector of the historical device anomaly category as the first estimated anomaly cause information; C5. Based on the current abnormal power distribution comprehensive feature vector and the equipment abnormal cause feature comparison dataset, calculate the corresponding cosine similarity of each historical equipment abnormality category and define it as feature cosine similarity. C6, determine the maximum value among all feature cosine similarities and define the anomaly cause classification information corresponding to the typical feature vector of the historical device anomaly category as the second estimated anomaly cause information; C7 generates a list of estimated anomaly causes for the current device based on a combination of the first and second estimated anomaly cause information.

8. An intelligent monitoring system based on a power distribution room, characterized in that, include: Power distribution environment data acquisition module; Data acquisition module for power distribution equipment; Data processing module; Power distribution anomaly identification module; The power distribution environment data acquisition module, the power distribution equipment data acquisition module, and the power distribution anomaly identification module are respectively connected to the data processing module. The power distribution environment data acquisition module includes multiple regional environment data acquisition modules and an external environment data acquisition module, and each of the regional environment data acquisition modules and the external environment data acquisition module is connected to the data processing module. The power distribution anomaly identification module includes a regional anomaly identification module and an equipment anomaly identification module, and the regional anomaly identification module and the equipment anomaly identification module are respectively connected to the data processing module. The intelligent monitoring system based on the power distribution room further includes a power distribution anomaly monitoring strategy, comprising the following steps: D1, determine the corresponding power distribution room spatial distribution data in the preset power distribution room; D2, generates multiple power distribution area data by segmenting the spatial distribution data of the power distribution room; D3, according to the preset sampling time window and the data of each power distribution area, collects the corresponding power distribution area environmental time window data at the corresponding location in the power distribution room through the power distribution environment data acquisition module; D4. Based on the time window data of the power distribution area environment, the power distribution anomaly identification module generates the corresponding regional anomaly confidence level through the regional anomaly identification module. D5. If the regional anomaly confidence level is greater than the preset regional anomaly confidence level threshold, then define the corresponding power distribution area data as abnormal power distribution area data, and define the corresponding sampling time window as the abnormal sampling time window. D6. Determine the corresponding list of power distribution equipment in the abnormal area based on the abnormal power distribution area data and the preset power distribution equipment installation and positioning data; D7, based on the abnormal sampling time window and the abnormal area power distribution equipment list data, the corresponding power distribution equipment time window data of each power distribution equipment is collected through the power distribution equipment data acquisition module; D8, based on the time window data of each power distribution device, the power distribution anomaly identification module generates the corresponding device anomaly confidence level through the device anomaly identification module. D9. If the confidence level of the equipment abnormality is greater than the preset confidence level threshold of the equipment abnormality, then the corresponding power distribution equipment is defined as abnormal power distribution equipment.

Images