Power distribution room intelligent operation and maintenance method based on video real-time monitoring and disaster dynamic prediction

Abstract

The invention discloses a power distribution room intelligent operation and maintenance method based on video real-time monitoring and disaster dynamic prediction, and relates to the technical field of power distribution room operation and maintenance, and the method comprises the steps: obtaining a digital space of a power distribution room, and constructing a data Internet of Things; constructing an automatic collector, carrying out front-end staggered collection, returning and updating the data Internet of Things, and obtaining a real-time configuration scene of the power distribution room; performing operation trend prediction and disaster risk assessment of the power distribution room based on configuration data migration, and performing operation and maintenance decision to determine a power distribution room operation and maintenance strategy; and carrying out operation and maintenance management based on the distribution room operation and maintenance strategy. According to the invention, the technical problems of low operation and maintenance intelligence level, insufficient risk prevention and control capability, poor operation and maintenance efficiency and lack of guarantee of operation safety of the power distribution room in the prior art are solved, and the technical effects of realizing intelligent operation and maintenance management and risk prevention and control of the power distribution room, improving the operation and maintenance efficiency and safety and guaranteeing stable operation of the power distribution room are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of power distribution room operation and maintenance, and particularly to a smart operation and maintenance method for power distribution rooms based on real-time video monitoring and dynamic disaster prediction. Background Art

[0002] With the continuous development of the power system and the continuous growth of power demand, as a key link in power distribution, the operation and maintenance management of power distribution rooms has become increasingly prominent. Under the background of traditional technologies, the operation and maintenance of power distribution rooms mainly rely on manual inspections and experience judgments. Maintenance personnel regularly conduct on-site inspections of equipment, record various instrument data and equipment operation status. This method not only consumes a large amount of manpower and time, but also is prone to human negligence and misjudgment. At the same time, the ability to detect and warn of potential equipment failures at an early stage is weak. It is often only possible to detect when a failure has occurred or is about to occur, lacking effective analysis and prediction means for operation trends. In terms of data collection, each front-end monitoring device is usually in an isolated state, and data cannot be efficiently interconnected and shared, making it difficult to grasp the overall operation situation of the power distribution room. Moreover, traditional operation and maintenance technologies do not adequately consider the impact of environmental factors on the operation of power distribution rooms, and fail to fully combine the dynamic changes of internal and external environments for comprehensive risk assessment.

[0003] The prior art has technical problems such as low intelligent level of power distribution room operation and maintenance, insufficient risk prevention and control ability, poor operation and maintenance efficiency, and lack of guarantee for operation safety. Summary of the Invention

[0004] This application provides a smart operation and maintenance method for power distribution rooms based on real-time video monitoring and dynamic disaster prediction, which is used to solve the technical problems of low intelligent level of power distribution room operation and maintenance, insufficient risk prevention and control ability, poor operation and maintenance efficiency, and lack of guarantee for operation safety in the prior art.

[0005] In view of the above problems, this application provides a smart operation and maintenance method for power distribution rooms based on real-time video monitoring and dynamic disaster prediction, and the method includes:

[0006] Obtain the digital space of the power distribution room and construct a data Internet of Things. Among them, the data Internet of Things is constructed by interconnecting data of front-end monitoring devices through a device gateway; construct an automatic collector to control the front-end monitoring devices to perform front-end off-peak data collection, transmit back and update the data Internet of Things, and obtain the actual configuration scenario of the power distribution room; for the actual configuration scenario, combine with the intelligent operation and maintenance module, and perform operation trend prediction and disaster risk assessment of the power distribution room based on configuration data migration. The operation and maintenance decision determines the operation and maintenance strategy of the power distribution room. Among them, the disaster risk assessment includes internal environment compensation and external environment compensation, and the intelligent operation and maintenance module is placed in the centralized control center; perform operation and maintenance management based on the operation and maintenance strategy of the power distribution room. Among them, the operation and maintenance strategy of the power distribution room includes an automatic strategy and an artificial strategy, and the automatic strategy uses power distribution equipment as the response terminal.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0008] Obtain the digital space of the power distribution room and construct a data Internet of Things; construct an automatic collector to control the front-end monitoring devices to perform front-end off-peak data collection, transmit back and update the data Internet of Things, and obtain the actual configuration scenario of the power distribution room; for the actual configuration scenario, combine with the intelligent operation and maintenance module, and perform operation trend prediction and disaster risk assessment of the power distribution room based on configuration data migration. The operation and maintenance decision determines the operation and maintenance strategy of the power distribution room; perform operation and maintenance management based on the operation and maintenance strategy of the power distribution room. It achieves the technical effects of realizing intelligent operation and maintenance management and risk prevention and control of the power distribution room, improving operation and maintenance efficiency and safety, and ensuring the stable operation of the power distribution room. Description of the Drawings

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0010] Figure 1 It is a schematic flowchart of the intelligent operation and maintenance method for the power distribution room based on video real-time monitoring and disaster dynamic prediction provided by the embodiment of this application;

[0011] Figure 2 It is a schematic flowchart of constructing a data Internet of Things in the intelligent operation and maintenance method for the power distribution room based on video real-time monitoring and disaster dynamic prediction provided by the embodiment of this application. Detailed Embodiments

[0012] The present application provides a smart operation and maintenance method for a power distribution room based on real-time video monitoring and dynamic disaster prediction, aiming to solve the technical problems of low intelligence level, insufficient risk prevention and control ability, poor operation and maintenance efficiency, and lack of guarantee for operation safety in the prior art.

[0013] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts shall fall within the protection scope of the present application.

[0014] Embodiment, as Figure 1 shown, the present application provides a smart operation and maintenance method for a power distribution room based on real-time video monitoring and dynamic disaster prediction, and the method includes:

[0015] Step S100: Obtain the digital space of the power distribution room and construct a data Internet of Things, where the data Internet of Things is constructed by interconnecting data of front-end monitoring devices through a device gateway as a medium.

[0016] Specifically, at the initial stage of the smart operation and maintenance of the power distribution room, the first thing to do is to obtain its digital space and construct a data Internet of Things. In this process, a device gateway with strong data transmission and conversion capabilities is used as the key medium to organically connect the front-end monitoring devices distributed at various key positions in the power distribution room to achieve data interconnection and interoperability. These front-end monitoring devices cover various types, such as high-definition video collectors that can accurately capture the appearance and operation status images of devices, and heat map collectors that can sensitively sense the temperature distribution of devices, etc. They are each responsible for specific space monitoring areas and comprehensively collect various operation data of the power distribution room. Through the device gateway, the massive data collected by these front-end monitoring devices are aggregated, integrated, and standardized to construct a complete and real-time updated data Internet of Things, providing comprehensive, accurate, and real-time data support for subsequent operation and maintenance decisions, making its operation status well-known, and thus laying a solid data foundation for realizing smart operation and maintenance.

[0017] Step S200: Construct an automatic collector, control the front-end monitoring devices to perform front-end staggered collection, transmit back and update the data Internet of Things, and obtain the actual configuration scene of the power distribution room.

[0018] Specifically, building an automatic collector becomes a key task. By deeply analyzing the performance characteristics and monitoring requirements of front-end monitoring devices, a set of scientific and reasonable monitoring and collection staggered rules are determined. For monitoring targets of different natures, such as relatively stationary devices or areas, they are set as static targets and configured with a lower first collection frequency to reduce unnecessary data redundancy; while for those devices and areas that are in frequent motion or have rapidly changing states, they are identified as dynamic targets and given a higher second collection frequency to ensure that key information is not missed. At the same time, according to the unique attributes of each monitoring target, an alternating collection sequence of videos and heatmaps is set to achieve the efficiency and comprehensiveness of data collection. Based on these rules, multiple control threads are determined, and then a powerful automatic collector is built, which is placed in the centralized control center and establishes a stable connection with the front-end monitoring devices. The automatic collector accurately controls the front-end monitoring devices to collect data at different time nodes according to the set staggered collection rules, and then quickly transmits the collected rich data back to the data Internet of Things to update the information content in real time. Finally, the real and detailed live configuration scene of the power distribution room is successfully obtained, providing accurate and intuitive on-site basis for subsequent operation and maintenance analysis and decision-making, enabling operation and maintenance personnel to timely grasp the actual operation status of the power distribution room and make more targeted operation and maintenance measures.

[0019] Step S300: For the live configuration scene, in combination with the intelligent operation and maintenance module, based on the configuration data migration, perform the operation trend prediction and disaster risk assessment of the power distribution room, and determine the operation and maintenance strategy of the power distribution room through operation and maintenance decision-making. Among them, the disaster risk assessment includes internal environment compensation and external environment compensation, and the intelligent operation and maintenance module is placed in the centralized control center.

[0020] Specifically, after obtaining the actual configuration scenario of the power distribution room, make full use of the intelligent operation and maintenance module located in the centralized control center and carry out work with the configuration data migration as the core basis. First, accurately identify the actual configuration scenario, and determine the data pairs for updating the configuration from it, that is, clarify the upper node data for updating the configuration and the corresponding real-time node data. These data pairs are the key elements for subsequent analysis. Then, based on these data pairs, deeply explore the data migration characteristics, and use this to detail the local operation trend of the updated configuration, and comprehensively understand the changes in the operation status of specific areas or equipment in the power distribution room. At the same time, taking the relevant parts where the configuration update has not occurred as a quantitative reference, conduct a whole-room operation analysis based on the data migration characteristics covering the entire power distribution room, so as to accurately determine the whole-room operation trend and grasp the overall operation direction of the power distribution room. On this basis, conduct a disaster risk assessment, which includes two important links: internal environment compensation and external environment compensation. By interacting with various element values of the internal environment, such as indoor temperature, humidity, air circulation, etc., optimize and adjust the local operation trend and the whole-room operation trend once to determine the first calibration trend. Further consider the element values of the external environment, such as weather conditions, surrounding electromagnetic field interference, etc., and supplement and improve the first calibration trend again to obtain the second calibration trend. Finally, conduct a scientific and rigorous disaster assessment based on the second calibration trend to obtain an accurate disaster prediction result. According to the results of the operation trend prediction and the disaster risk assessment, formulate a practical power distribution room operation and maintenance decision, and determine the power distribution room operation and maintenance strategy including automation strategy and manual strategy. Among them, the automation strategy directly acts on the response terminal of the power distribution equipment to ensure the safe, stable and efficient operation of the power distribution room, effectively prevent the occurrence of potential disasters, and improve the intelligent level and reliability guarantee of the power distribution room operation and maintenance management.

[0021] Step S400: Perform operation and maintenance management based on the power distribution room operation and maintenance strategy, where the power distribution room operation and maintenance strategy includes an automation strategy and a manual strategy, and the automation strategy takes the power distribution equipment as the response terminal.

[0022] Specifically, after formulating the operation and maintenance strategy for the power distribution room, comprehensive operation and maintenance management work is carried out. The formulated operation and maintenance strategy for the power distribution room integrates automated strategies and manual strategies, which cooperate closely to jointly ensure the stable operation of the power distribution room. The automated strategy mainly acts on power distribution equipment, and its operation mechanism is highly intelligent. For example, when abnormal situations such as overload and short circuit are detected in the circuit, the automated strategy can quickly perform cut-off control based on the circuit breaker. Through precise command control of the execution element, the faulty area can be isolated in a timely manner to prevent the problem from further expanding. The whole process is fast and efficient, without the need for real-time manual intervention, greatly improving the timeliness and accuracy in dealing with sudden electrical problems. At the same time, for some more complex situations that require the coordination of professional judgment and manual operation, such as fine-tuning of equipment parameters and delicate maintenance of specific components, the relevant task information will be sent to the operation and maintenance personnel's terminal, and experienced operation and maintenance personnel will handle it precisely according to the actual situation. Relying on professional knowledge and practical experience, the operation and maintenance personnel conduct detailed inspections and necessary operation adjustments on the equipment to ensure that the equipment is in the best operating state. Through this operation and maintenance management mode combining automation and manual work, the respective advantages are fully utilized to achieve all-round and multi-level efficient operation and maintenance management of the power distribution room, ensuring the reliability and continuity of power supply and laying a solid foundation for the stable operation of the power system.

[0023] In a possible implementation manner, as Figure 2 shown, step S100 further includes:

[0024] Step S110: Configure front-end monitoring devices for the operation and maintenance management elements of the power distribution room, where the front-end monitoring devices include video collectors and thermal image collectors, and each front-end monitoring device corresponds to a spatial monitoring area.

[0025] Step S120: Using the front-end monitoring devices as the data source end, perform data interconnection in the digital space based on the device gateway to generate the data Internet of Things.

[0026] Specifically, in the key link of intelligent operation and maintenance of the power distribution room, in order to comprehensively and accurately grasp the operation status of the power distribution room, it is necessary to reasonably configure front-end monitoring devices according to the core elements of its operation and maintenance management. Among them, the selected front-end monitoring devices mainly include video collectors and thermal image collectors. The video collector can capture intuitive information such as the appearance, operation actions of various devices in the power distribution room, and personnel activities with clear images and smooth video pictures, while the thermal image collector focuses on keenly perceiving the temperature distribution of the equipment through thermal imaging technology to discover potential overheating fault hidden dangers in advance. Moreover, each front-end monitoring device is carefully arranged to correspond to a specific spatial monitoring area, ensuring that every corner and key part of the power distribution room can be effectively covered without monitoring blind spots, providing comprehensive and detailed original materials for subsequent data collection.

[0027] As a data source, the front-end monitoring devices output raw data that follows specific data formats and communication protocols. These devices establish reliable communication connections with the device gateway and use transmission methods such as TCP / IP or industrial Ethernet to transmit information such as video streams and heat map data collected to the gateway. As the core data aggregation and conversion node, the device gateway has powerful computing capabilities and multi-protocol parsing functions. It performs protocol conversion, data cleaning, and preliminary integration on the data from different data sources received. For example, it parses and repackages video data according to the encoding standard, and converts heat map data into a standardized temperature matrix data format, enabling various types of data to interact and fuse in the digital space following a unified data model. By constructing a data routing table and switching logic, it realizes the directional transmission and sharing of data between different front-end monitoring devices, and finally generates a structurally complete, data-real-time-updated, and interoperable data Internet of Things, providing a standardized and normalized data input interface for subsequent operation and maintenance analysis and decision-making systems, ensuring the accuracy, integrity, and timeliness of data to support the efficient operation of the power distribution room intelligent operation and maintenance management process based on data-driven.

[0028] In a possible implementation manner, step S200 further includes:

[0029] Step S210: For the front-end monitoring devices, determine the out-of-time acquisition rules for monitoring and acquisition.

[0030] Step S220: According to the out-of-time acquisition rules, determine multiple control threads, and construct an automatic collector. The automatic collector is placed in the centralized control center and is connected to the front-end monitoring devices.

[0031] Specifically, in the key process of intelligent operation and maintenance of the power distribution room, for the front-end monitoring devices distributed throughout the power distribution room, multiple factors need to be comprehensively considered to determine the staggered acquisition rules for their monitoring and acquisition. This includes the characteristics of the objects monitored by the devices. For relatively static devices or areas with slow-changing operating states, such as the fixed structures in the power distribution room and the basic power facilities with long-term stable operation, a first acquisition frequency with a lower frequency is set according to their stability characteristics to reduce unnecessary data redundancy and avoid excessive resource occupation of the normal operation of the devices. For those key components or areas with frequent dynamic changes, such as the frequent opening and closing actions of power switches and the real-time fluctuations of power loads, a second higher acquisition frequency is set to ensure that the detailed operating states of their rapid changes can be accurately captured without missing any potential abnormal situations. In addition, factors such as the importance of the devices, the historical change trends of the data, and the current key points of operation and maintenance also need to be combined to further optimize the acquisition rules, determine the alternating acquisition sequences of different types of data such as videos and heat maps, so as to achieve refined management of the acquisition behavior of the front-end monitoring devices, ensure that the collected data can comprehensively reflect the operating conditions of the power distribution room, highlight the key points, and improve the effectiveness of the data.

[0032] Determine multiple control threads and construct an automatic collector according to the established staggered acquisition rules. First, analyze the staggered acquisition rules and convert information such as the acquisition time series, frequency parameters, and data types (such as videos or heat maps) of different front-end monitoring devices into machine-readable instruction sets. According to these instructions, independent control threads are allocated to each front-end monitoring device. These threads are created and managed in the kernel space of the operating system and have precise time scheduling capabilities to ensure that each device performs data acquisition according to the predetermined staggered rules, avoiding resource conflicts and chaos in data acquisition. Then, use hardware resources and programming frameworks to construct an automatic collector. As a process running on the centralized control center server, the automatic collector is equipped with functional modules for communicating with the front-end monitoring devices. These modules are based on network communication protocols (such as TCP / IP or industrial Ethernet protocols) to achieve stable connections with the front-end devices. When the automatic collector is initialized, it scans and identifies the connected front-end monitoring devices, and establishes a mapping relationship between the device list and its corresponding control threads. During operation, it regularly sends acquisition instructions to the front-end monitoring devices through the control threads and waits for the devices to respond within the specified time window to receive the collected data. At the same time, the automatic collector is equipped with data caching and preliminary processing functions to perform operations such as integrity verification and format conversion on the received data, and organize it into data blocks suitable for subsequent transmission and storage, preparing for updating the data Internet of Things, thus ensuring the efficient and stable operation of the entire data acquisition process to meet the accuracy and real-time requirements of data acquisition for the intelligent operation and maintenance of the power distribution room.

[0033] In a possible implementation manner, step S210 further includes:

[0034] Step S211: Determine the first monitoring target of the first front-end monitoring device and configure the monitoring and acquisition rules for the first monitoring target.

[0035] Among them, the configuration of the monitoring and acquisition rules includes:

[0036] Obtain the characteristics of the first monitoring target and set an alternating acquisition sequence of video and thermal images.

[0037] If the first monitoring target is a static target, set the first acquisition frequency.

[0038] If the first monitoring target is a dynamic target, set the second acquisition frequency, where the second acquisition frequency is higher than the first acquisition frequency.

[0039] Use the alternating acquisition sequence - the first acquisition frequency, or the alternating acquisition sequence - the second acquisition frequency as the monitoring and acquisition rules of the first front-end monitoring device.

[0040] Specifically, first, through a comprehensive analysis of the distribution room layout, equipment distribution, and operation and maintenance requirements, accurately identify and determine the first monitoring target responsible for by the first front-end monitoring device. This target may be a key device in the distribution room, such as a large transformer, or a specific area, such as the connection part of the power bus.

[0041] Immediately afterwards, start the monitoring and acquisition rule configuration process for this monitoring target, and use advanced sensor data analysis technology to deeply obtain the characteristic information of the first monitoring target. If this target is determined to be a static target, for example, a fixed installation of a distribution cabinet body, its physical position and appearance are relatively fixed, and its operating state changes slowly. At this time, a relatively low first acquisition frequency will be set to reduce unnecessary data acquisition, avoid resource waste and data redundancy, and at the same time ensure that abnormal state changes that may occur to this target can be effectively captured, such as slight heat generation or deformation on the surface of the cabinet body caused by internal faults.

[0042] On the contrary, if the first monitoring target shows dynamic characteristics, such as a running circuit breaker switch, which frequently opens and closes, and its internal mechanical structure and electrical parameters are constantly changing, then a higher second acquisition frequency will be set to ensure that every state detail of its running moment can be completely and carefully recorded, without missing any key running information, providing detailed data support for subsequent fault diagnosis and performance evaluation.

[0043] On this basis, according to the target characteristics, an alternating acquisition sequence of videos and heatmaps will also be set. For static targets, the time interval of video acquisition may be appropriately extended, while the frequency of heatmap acquisition is increased, so as to more accurately monitor its temperature distribution and timely detect potential overheating hazards; for dynamic targets, the alternating acquisition order and time ratio of videos and heatmaps will be flexibly adjusted according to its action cycle and frequency, ensuring that while capturing its dynamic actions, the temperature change trend can also be comprehensively grasped.

[0044] Finally, the carefully set alternating acquisition sequence is combined with the corresponding first acquisition frequency (for static targets) or the second acquisition frequency (for dynamic targets) to form a complete and highly adaptable monitoring and acquisition rule for the first front-end monitoring device, thereby realizing the scientific and refined management of the data acquisition work for the first monitoring target, providing a solid and reliable data foundation guarantee for the intelligent operation and maintenance of the entire power distribution room, and effectively improving the operation and maintenance efficiency and the stability of the power system.

[0045] In a possible implementation manner, step S300 further includes:

[0046] Step S310: Identify the live configuration scenario and determine the data pair for updating the configuration, where the data pair is the upper node data and the real-time node data for updating the configuration.

[0047] Step S320: Conduct a running trend migration analysis on the data pair to determine the disaster prediction result.

[0048] Specifically, first, an object detection and feature extraction algorithm based on deep learning is adopted to identify the live configuration scenario and determine the data pairs for updating the configuration. First, a pre-trained convolutional neural network (CNN) model is used to perform object detection on the video images collected by the front-end monitoring devices, and various devices in the power distribution room and their operating states are identified. For example, for the switchgear, the model can accurately detect the opening and closing positions of the switches, the states of the indicator lights, etc. At the same time, a heat map analysis algorithm is used to process the heat map data to extract the temperature feature information of the devices. When changes in the operating state or temperature and other features of the devices are detected, these changed areas are determined as the areas for updating the configuration. To determine the data pairs for updating the configuration, a data association algorithm is adopted. This algorithm constructs a data association graph based on the electrical connection relationships and logical topological structures of the devices. In the graph, the upper-node data is obtained by querying the control system database of the power distribution room, and this data includes the preset operating parameters, control instructions, and historical operating trend data of the devices, etc. The real-time node data is directly extracted from the real-time collected data of the front-end monitoring devices. For example, the values of current, voltage, temperature, etc. measured by the sensors are read through a data acquisition card. Through the synergistic effect of the deep learning algorithm and the data association algorithm, the data pairs for updating the configuration are accurately determined, providing a reliable data basis for subsequent operation trend migration analysis, thereby achieving an accurate grasp of the operating state of the power distribution room and an effective assessment of the disaster risk, and ensuring the stable operation of the power system.

[0049] To perform a running trend migration analysis on data pairs and determine the disaster prediction results, a machine learning algorithm system combining long short-term memory network (LSTM) and support vector machine (SVM) is adopted. First, the LSTM network is used to process time series data. The upper node data and real-time node data in the data pair are arranged in chronological order and input into the LSTM model. Through its unique gating mechanism, the LSTM network can effectively learn the long-term dependence relationships and short-term change characteristics in the data, capture the dynamic change trends of the device operating state over time, such as the rising or falling trends of the device temperature in different time periods, the fluctuation rules of current and voltage, etc., so as to construct a time series model of the device operating state and accurately extract the feature vectors of the local operating trends. Then, the principal component analysis (PCA) algorithm is used to reduce the dimension of the high-dimensional feature vectors output by the LSTM, remove redundant information, retain the most critical feature information, and at the same time reduce the computational complexity of the data and improve the training and prediction efficiency of the subsequent model. Next, the feature vectors after PCA dimensionality reduction are input into the SVM classifier. The SVM finds an optimal hyperplane to classify data samples with different operating states and trends, and determines whether the current operating trend is in a normal state or tends to a certain disaster risk state. When training the SVM model, normal samples and abnormal samples before various disasters in the historical operating data of the power distribution room are used for training. By adjusting the parameters of the kernel function (such as the radial basis function), the classification performance of the model is optimized so that it can accurately distinguish different operating trend categories. Finally, according to the output results of the SVM classifier and in combination with the pre-set disaster risk level threshold, the possibility of a disaster occurring in the power distribution room under the current operating trend and the corresponding disaster prediction results are determined, providing strong data support and scientific basis for subsequent operation and maintenance decisions and ensuring the safe and stable operation of the power distribution room.

[0050] In a possible implementation manner, step S320 further includes:

[0051] Step S321: Based on the data pair, determine the data migration characteristics and analyze the local operating trends of the updated configuration.

[0052] Step S322: Taking the non-updated configuration as the quantitative basis, perform a whole-room operating analysis based on the data migration characteristics to determine the whole-room operating trend, where the non-updated configuration related to the updated configuration is taken as the analysis target.

[0053] Step S323: Based on the local operating trend and the whole-room operating trend, perform disaster risk prediction.

[0054] Specifically, first, the clustering analysis algorithm in the field of data mining is used to group the data based on its similarity, initially identifying different patterns in the data. Then, through the principal component analysis (PCA) technique, the data dimension is reduced, key information is extracted, highlighting the main direction and characteristics of data changes, thereby determining the data migration characteristics. For example, when monitoring a key device, it is found that its current data shows a gradually increasing trend over a period of time and there is a certain correlation pattern with the change of voltage data, which is one of the data migration characteristics determined through data analysis. Based on these characteristics, the physical model of equipment operation and statistical analysis methods are used to analyze the local operation trend. Considering the normal operation parameter range of the equipment and historical operation data, through time series analysis, an ARIMA model is constructed to predict the short-term operation trend of the equipment. If it is found that the temperature data of a certain device continuously deviates from the normal range and the rising rate accelerates, combined with the change of its load data, it can be judged that there may be operation risks in the local area where the device is located, such as overloading and overheating problems, and then the local operation trend is clarified, providing key local information for subsequent overall analysis.

[0055] Using the system dynamics modeling method, an overall operation model of the power distribution room is constructed, including the relationships of energy flow, signal transmission, etc. between various equipment configurations. For non-updated configurations related to updated configurations, through sensitivity analysis, determine their key role degree and influence range in the whole system. On the basis of considering data migration characteristics, the change situation of the updated configuration area is introduced as an input variable into the system dynamics model. For example, when the power of a device in a certain updated configuration area changes, based on the determined data migration characteristics, analyze how this change affects the operation state of the devices in the related non-updated configuration area through electrical connections and signal feedback, etc., and then use the state space model to estimate and update the operation state of the whole room in real time, finally determining the overall operation trend of the whole room. At the same time, in the face of a large amount of data, instead of comprehensively analyzing all historical data, it is cleverly combined with the operation state of the whole room at the last acquisition node and the change of data migration characteristics brought by the current updated data. Through methods such as comparison and difference calculation, quickly determine the change direction and degree of the overall operation trend of the whole room, effectively improving the analysis efficiency and avoiding resource waste.

[0056] Standardize the changes in key parameters in the local operation trend, such as the rapid temperature rise rate of specific equipment, the current overload multiple, etc., and data such as the overall power fluctuation and voltage stability index in the whole-room operation trend, so that they are within the same dimension range. Then, by constructing a Bayesian network model, use these processed data as input nodes, and set various possible disaster types, such as short circuits, fires, equipment explosions, etc. as output nodes, and determine the conditional probability relationship between nodes using historical data and expert experience. At the same time, use the particle swarm optimization algorithm to optimize and adjust the parameters of the Bayesian network to improve the prediction accuracy and generalization ability of the model. Based on this model, it is possible to calculate the probability values of different disasters occurring according to the current local and overall operation trend data, and combine the pre-set risk thresholds to judge the current disaster risk level faced by the power distribution room, so as to provide accurate decision-making basis for maintenance personnel to take targeted prevention and control measures in advance, effectively reduce potential disaster losses, and ensure the safe and stable operation of the power distribution room.

[0057] In a possible implementation manner, step S323 further includes:

[0058] Step S3231: Interact the element values of the internal environment, perform a first compensation on the local operation trend and the whole-room operation trend, and determine the first calibration trend.

[0059] Step S3232: Interact the element values of the external environment, perform a second compensation on the first calibration trend, and determine the second calibration trend.

[0060] Step S3233: Based on the second calibration trend, conduct a disaster assessment and determine the disaster prediction result.

[0061] Specifically, first, comprehensive and accurate collection and quantification of various element values in the internal environment of the power distribution room are carried out. These elements cover multiple key aspects such as temperature, humidity, air circulation rate, and electromagnetic interference intensity. Through a highly sensitive sensor network deployed indoors, accurate values of these elements are obtained in real-time and converted into a data format that can be directly processed by the system. Subsequently, a trend compensation model based on machine learning is used. This model is trained based on a large amount of historical data and can accurately identify the complex correlation between the internal environment element values and the equipment operation trends. For example, when the indoor temperature is detected to rise, the model will reasonably compensate for the downward trend of the heat dissipation efficiency of specific equipment in the local operation trend according to the empirical data of the impact of past temperatures on equipment performance, by adjusting the predicted curve of the heat dissipation rate to make it more in line with the actual situation; at the same time, for the increasing trend of the resistance loss of power transmission in the overall room operation trend, corresponding compensation and correction are made based on the physical relationship between temperature and resistance. By integrating these compensation operations based on the internal environment elements, the first calibrated trend closer to the actual operation state is finally determined, providing a reliable data basis guarantee for subsequent precise operation and maintenance decisions, and effectively reducing the probability of misjudgment of operation and maintenance risks caused by internal environment changes.

[0062] Extending to the external environment of the power distribution room, various element values such as weather conditions (air temperature, humidity, air pressure, precipitation probability, etc.), surrounding geographical information (altitude, impact of terrain and landform on air flow, etc.), and the operation status of nearby other facilities (start-stop of large electrical equipment in surrounding factories, interference of road traffic flow on electromagnetic fields, etc.) are widely collected. Using big data analysis technology and neural network algorithms, the potential correlation patterns and influence weights between these external elements and the operation of the power distribution room are deeply explored. For example, when the external air temperature suddenly rises or the humidity increases sharply, through the established heat conduction and humidity influence model, the trend curves related to the heat dissipation rate and electrical insulation performance of the equipment in the first calibrated trend are corrected again. Considering that air humidity may accelerate the aging of electrical components and reduce insulation performance, the relevant indicators related to the possible electrical fire risk are optimized and adjusted; another example is that the electromagnetic interference generated by the frequent start-stop of surrounding large equipment may affect the stability of the control signals of the equipment inside the power distribution room, and the operation stability trend of the corresponding equipment will be compensated secondly according to the electromagnetic interference propagation model. Through the comprehensive interaction and accurate compensation calculation of these external environment elements, the second calibrated trend that is more accurate and can better reflect the actual potential risk situation is finally determined, providing highly reliable trend data for subsequent disaster assessment, greatly enhancing the accuracy and forward-looking of the operation risk prediction of the power distribution room, and effectively ensuring the safe and stable operation of the power system.

[0063] The goal of disaster assessment and prediction results is achieved by using the Gated Recurrent Unit (GRU) and Support Vector Machine (SVM) to work together. At the beginning, the operating parameter data of various equipment covered by the second calibration trend, such as the temperature value, current size, and voltage of the equipment, are transmitted to the GRU network one by one according to the time sequence of their generation. With its own special update gate and reset gate structure, GRU can capture the internal connection of these data in a long span and the fluctuation characteristics in a short period of time. It can autonomously explore the trend and regular characteristics of the equipment operation status over time, and then extract the feature information that is extremely critical for subsequent analysis, and integrate this information into a feature vector output with a fixed dimension. Subsequently, the feature vector output from GRU is used as the basis for SVM classification judgment. SVM uses a large amount of historical data accumulated in the past to carry out training in advance, with the aim of constructing a hyperplane that can distinguish different operating states. This hyperplane can clearly divide the feature vectors into two categories, one representing the normal operation of the equipment and the other corresponding to various potential disaster risk states. In the training phase, by using a specific kernel function (such as the radial basis function), the feature vector originally in the low-dimensional space is cleverly mapped to the high-dimensional space. In this way, the distribution characteristics of the data in the high-dimensional space are more significant and easier to be accurately classified. At the same time, the optimization algorithm is used to continuously find and determine the optimal parameters of the hyperplane, thereby significantly improving the accuracy of SVM classification and its wide adaptability to different data. When it is necessary to conduct a disaster assessment on the current equipment operation status, just input the corresponding feature vector at this moment into the trained SVM model. The model will accurately determine which category this feature vector belongs to based on the position of the feature vector on the previously constructed hyperplane, and then clarify whether there is a disaster risk at the moment. If there is a risk, it can further determine the specific type of disaster that may occur and give the corresponding estimated value of the probability of occurrence. Finally, through such a process, accurate and reliable disaster prediction results are obtained, which will provide solid and strong support for the decision-making of the personnel responsible for the operation and maintenance of the distribution room, and effectively ensure that the distribution room can continue to operate safely, stably and continuously.

[0064] In a possible implementation, step S400 further includes:

[0065] Step S410: Monitor all elements of the interactive power distribution room and mine the risk threshold of each element.

[0066] Step S420: Based on the risk threshold, a rigid limiter is configured, and the rigid limiter is pre-placed in the intelligent operation and maintenance module.

[0067] Step S430: Based on the rigid limiter, the returned real-time node data is judged to be out of limit, and risk warning information is generated.

[0068] Specifically, through a high-precision sensor network deployed at various key positions in the power distribution room, comprehensive monitoring data of all elements are collected, including electrical parameters (such as voltage, current, power factor, etc.), equipment operating status (such as switch opening and closing status, equipment temperature, etc.), and environmental conditions (such as temperature, humidity, concentration of harmful gases, etc.). Using the clustering analysis algorithm in big data analysis, historical data are classified and aggregated according to normal operating status and various abnormal conditions, and then the boundary values of each element data under abnormal conditions are determined through statistical analysis methods. These boundary values are repeatedly verified and corrected, and finally recognized as the risk critical values of each element. For example, if the temperature of a certain equipment exceeds 80°C, it may be in a failure risk state, and this 80°C is the risk critical value of the temperature element of this equipment.

[0069] Based on the risk critical values of each element accurately mined in the early stage, the professional configuration process of the rigid limiter is started. First, in terms of electrical parameters, if the risk critical value of the current of an important line is set at 500A, through professional debugging tools, in the current threshold setting module of the rigid limiter, it is accurately set to trigger the warning mechanism when the current reaches 480A (a certain safety margin can be reserved according to the actual situation), and start the circuit breaker protection action when the current reaches 500A to prevent damage to the line and equipment caused by overcurrent. For the equipment temperature, if the risk critical temperature of a key equipment is 70°C, it is set in the temperature sensing unit of the rigid limiter to issue a temperature rise warning when the temperature rises to 65°C and forcibly cut off the power supply of the equipment when it reaches 70°C to avoid equipment failure or even fire caused by overheating. These rigid limiters that have been finely tuned and built-in with the trigger conditions of the risk critical values of each element are reasonably placed before the intelligent operation and maintenance module to build a physical defense line. With its fast response characteristics and accurate threshold judgment ability, it can quickly respond when the real-time node data is about to break through the risk critical value or has exceeded the limit, prevent abnormal data from continuing to be transmitted to the intelligent operation and maintenance module, and avoid the situation of overload or misjudgment of the intelligent operation and maintenance system due to receiving a large amount of abnormal data, thus ensuring the stable operation of the entire power distribution room operation and maintenance system and the accuracy of data, and providing a reliable pre-protection mechanism for subsequent risk control and equipment maintenance.

[0070] The rigid limiter plays a crucial role in real-time monitoring and early warning. When the real-time node data collected from various monitoring devices at the front end of the power distribution room and transmitted back continuously flows into the intelligent operation and maintenance system, the rigid limiter will, according to the pre-set determination rules based on risk critical values, conduct one-by-one comparison and analysis on each piece of data. For example, for voltage data, if the rated voltage of a certain line is 380V, and the upper limit critical value set according to risk assessment is 400V, and the lower limit critical value is 360V, the rigid limiter will accurately detect each real-time transmitted voltage value. Once it is found that the voltage data exceeds this pre-set normal range, whether it is higher than the upper limit or lower than the lower limit, it will immediately activate the over-limit determination mechanism. At this time, the rigid limiter will quickly send a trigger signal to the intelligent operation and maintenance system, and the intelligent operation and maintenance system will, according to the built-in alarm information generation module, combine information such as the type of over-limit data, the specific value, and the current operating conditions of the power distribution room, and quickly generate detailed and accurate risk alarm information. This information includes the time when the alarm occurs, the location of the equipment or line corresponding to the over-limit data, the specific value of the over-limit, and the possible risk consequences inferred based on historical data and expert experience. These alarm messages will be sent in various ways, such as popping up a prominent alarm pop-up window on the monitoring terminal, sending a text message notification to the mobile phone of the operation and maintenance personnel, and triggering an audible and visual alarm at the power distribution room site, etc., to ensure that the operation and maintenance personnel can receive the risk alarm in the first time, so as to take effective countermeasures in time to ensure the safe and stable operation of the power distribution room.

[0071] It should be noted that the above-mentioned sequence of embodiments of the present application is only for description and does not represent the superiority or inferiority of the embodiments. And the above describes specific embodiments of this specification. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0072] The above are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application shall be included within the protection scope of the present application.

[0073] This specification and the drawings are only exemplary descriptions of the present application and are considered to have covered any and all modifications, variations, combinations, or equivalents within the scope of the present application. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application is intended to include these changes and modifications.

Claims

1. A smart operation and maintenance method for a power distribution room based on real-time video monitoring and dynamic disaster prediction, characterized in that, The method includes: Obtain the digital space of the power distribution room and construct a data Internet of Things, where the data Internet of Things is constructed through data interconnection of front-end monitoring devices with the device gateway as the medium; Construct an automatic collector, control the front-end monitoring devices to perform front-end staggered collection, transmit back and update the data Internet of Things, and obtain the actual configuration scenario of the power distribution room; For the actual configuration scenario, combined with the intelligent operation and maintenance module, perform operation trend prediction and disaster risk assessment of the power distribution room based on configuration data migration, and determine the operation and maintenance strategy of the power distribution room through operation and maintenance decision-making. Among them, the disaster risk assessment includes internal environment compensation and external environment compensation, and the intelligent operation and maintenance module is placed in the centralized control center; Perform operation and maintenance management based on the operation and maintenance strategy of the power distribution room. Among them, the operation and maintenance strategy of the power distribution room includes an automatic strategy and a manual strategy, and the automatic strategy uses power distribution equipment as the response terminal.

2. The intelligent operation and maintenance method for a power distribution room based on real-time video monitoring and dynamic disaster prediction according to claim 1, wherein, The construction of the data Internet of Things includes: Configure front-end monitoring devices for the operation and maintenance management elements of the power distribution room. Among them, the front-end monitoring devices include video collectors and heat map collectors, and each front-end monitoring device corresponds to a space monitoring area; Using the front-end monitoring devices as the data source end, perform data interconnection in the digital space based on the device gateway to generate the data Internet of Things.

3. The intelligent operation and maintenance method for a power distribution room based on real-time video monitoring and dynamic disaster prediction according to claim 1, wherein The construction of the automatic collector includes: Determine the staggered collection rules for monitoring and collection for the front-end monitoring devices; According to the staggered collection rules, determine multiple control threads and construct an automatic collector. Among them, the automatic collector is placed in the centralized control center, and the automatic collector is connected to the front-end monitoring devices.

4. The intelligent operation and maintenance method for a power distribution room based on real-time video monitoring and dynamic disaster prediction according to claim 3, characterized in that, Determining the staggered collection rules for monitoring and collection includes: Determine the first monitoring target of the first front-end monitoring device and configure the monitoring and collection rules for the first monitoring target; Among them, the configuration of the monitoring and collection rules includes: Obtain the characteristics of the first monitoring target and set the alternating collection sequence of video and heat map; If the first monitoring target is a static target, set the first collection frequency; If the first monitoring target is a dynamic target, set the second collection frequency, where the second collection frequency is higher than the first collection frequency; Use the alternating collection sequence - the first collection frequency, or the alternating collection sequence - the second collection frequency as the monitoring and collection rules for the first front-end monitoring device.

5. The intelligent operation and maintenance method for a power distribution room based on real-time video monitoring and dynamic disaster prediction according to claim 1, characterized in that, Performing operation trend prediction and disaster risk assessment of the power distribution room based on configuration data migration includes: Identify the actual configuration scenario and determine the data pairs of updated configurations, where the data pairs are the upper node data and real-time node data of the updated configurations; Perform operation trend migration analysis on the data pairs to determine the disaster prediction results.

6. The intelligent operation and maintenance method for a power distribution room based on real-time video monitoring and disaster dynamic prediction according to claim 5, characterized in that, Performing operation trend migration analysis on the data pairs includes: Based on the data pairs, determine the data migration characteristics and analyze the local operation trends of the updated configurations; Taking the non-updated configuration as the quantification, perform whole-room operation analysis based on the data migration characteristics to determine the whole-room operation trends, where the non-updated configuration related to the updated configuration is the analysis target; Based on the local operation trends and the whole-room operation trends, perform disaster risk prediction.

7. The intelligent operation and maintenance method for a power distribution room based on real-time video monitoring and disaster dynamic prediction according to claim 6, characterized in that Based on the local operation trend and the whole room operation trend, disaster risk prediction is performed, including: The element value of the interactive internal environment is used to compensate the local operation trend and the whole room operation trend to determine a first calibration trend; The element value of the interactive external environment is used to perform secondary compensation on the first calibration trend to determine a second calibration trend; Based on the second calibration trend, a disaster assessment is performed to determine a disaster prediction result.

8. The intelligent operation and maintenance method for a power distribution room based on real-time video monitoring and dynamic disaster prediction according to claim 5, characterized in that The method further comprises: Monitor all elements of the interactive power distribution room and explore the risk threshold of each element; Based on the risk threshold, a rigid limiter is configured, and the rigid limiter is placed in front of the intelligent operation and maintenance module; Based on the rigid limiter, the returned real-time node data is judged to be out of limit, and risk warning information is generated.

Images