Distribution network three-dimensional visualization dynamic operation and maintenance system development method and system

Abstract

The invention relates to the technical field of power industry, and discloses a distribution network three-dimensional visualization dynamic operation and maintenance system development method and system, and the method comprises the steps: scanning power distribution equipment and environment, and obtaining three-dimensional point cloud data; preprocessing and feature extraction are carried out on the point cloud data, and a three-dimensional model of the power distribution network equipment is constructed. And importing the three-dimensional model and the real-time monitoring data of the power distribution network into a visual platform, and evaluating the running state of the power distribution network. Different early warning mechanisms are set, and abnormity can be found in time. According to the invention, the fault detection and response speed is improved. And the accuracy of state evaluation and fault positioning is improved. In combination with machine learning and multi-level state evaluation, the intelligent operation and maintenance of the power distribution network are realized, risks can be automatically identified, early warning can be generated, manual operation errors are reduced, and the stability and safety of the system are improved. Through the three-dimensional visualization platform, operation and maintenance personnel can understand the state and the fault condition of the power distribution network more visually, the decision process is optimized, and the working efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of the electric power industry, and in particular to a method and system for developing a three-dimensional visualized dynamic operation and maintenance system for a distribution network. Background Art

[0002] As modern power systems develop towards intelligence and automation, the operation and maintenance of distribution networks face increasingly complex challenges. Traditional distribution network operation and maintenance mostly rely on plane drawings, manual inspections and simplified status monitoring, which makes fault location, equipment monitoring, resource scheduling and other work inefficient and slow to respond. Driven by information technology and the Internet of Things, the operation and maintenance of distribution networks has gradually shifted to data-driven and intelligent analysis. In particular, the operation and maintenance methods based on three-dimensional modeling and visualization technology can present information such as equipment and environment in a graphical way, greatly improving the intuitiveness, operability and decision-making support capabilities of operation and maintenance. As an important basis for three-dimensional modeling, point cloud data has been widely used in recent years, especially in the modeling, environmental monitoring and status detection of power equipment, showing good application prospects.

[0003] Despite this, the existing operation and maintenance systems based on 3D modeling and visualization still have several limitations. First, most systems in the existing technology have weak processing capabilities for point cloud data, resulting in the accuracy and update speed of the model being unable to meet actual operation and maintenance needs. In particular, for real-time monitoring and response to changes in the state of dynamic distribution networks, traditional methods still rely on manual intervention and lack automated, real-time analysis methods. Secondly, the processing and analysis of point cloud data is relatively complex, and existing systems often face problems of large computational complexity and slow processing speed when processing large-scale data, especially in scenarios with high real-time requirements. In addition, the existing operation and maintenance system lacks sufficient intelligent support and cannot automatically predict equipment status, diagnose faults, and conduct risk assessments based on real-time data, resulting in insufficient guarantees for the operating efficiency and safety of the distribution network.

[0004] The present invention proposes a method for developing a three-dimensional visualization dynamic operation and maintenance system for distribution networks based on point cloud technology. By introducing efficient point cloud data processing and feature extraction algorithms, combined with machine learning and multi-level dynamic evaluation technology, the present invention solves the problems of insufficient processing capacity, poor real-time performance, and low intelligence level in the prior art. Compared with the traditional static, manual operation and maintenance methods, the present invention realizes real-time dynamic monitoring and automated management of distribution networks, and improves the accuracy, efficiency, and response speed of operation and maintenance. Summary of the invention

[0005] In view of the above-mentioned problems, the present invention is proposed.

[0006] Therefore, the technical problem solved by the present invention is that the existing operation and maintenance systems based on three-dimensional modeling and visualization still have several limitations. First, most systems in the prior art have weak processing capabilities for point cloud data, resulting in the accuracy and update speed of the model being unable to meet actual operation and maintenance needs. In particular, for real-time monitoring and response to changes in the state of dynamic distribution networks, traditional methods still rely on manual intervention and lack automated, real-time analysis methods. Secondly, the processing and analysis of point cloud data is relatively complex, and existing systems often face problems of large computational complexity and slow processing speed when processing large-scale data, especially in scenarios with high real-time requirements. In addition, the existing operation and maintenance system lacks sufficient intelligent support and is unable to automatically predict equipment status, diagnose faults, and conduct risk assessments based on real-time data, resulting in the operating efficiency and safety of the distribution network not being fully guaranteed.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions: a method for developing a three-dimensional visual dynamic operation and maintenance system for a distribution network, comprising: scanning distribution equipment and environment to obtain three-dimensional point cloud data; Preprocess and extract features of point cloud data to build a three-dimensional model of distribution network equipment; Import the 3D model and real-time monitoring data of the distribution network into the visualization platform to evaluate the operating status of the distribution network; Set up different early warning mechanisms to detect abnormalities in a timely manner; Preprocessing and feature extraction of point cloud data include data denoising, downsampling and coordinate alignment; Data denoising uses statistical filtering to eliminate isolated points; Downsampling uses the voxel grid downsampling method to divide the point cloud into a fixed-size cubic voxel grid. The centroid of all points in each grid is calculated as the representative point, which reduces the amount of point cloud data and improves processing efficiency. Coordinate alignment uses the rigid body transformation formula to align the coordinate system; Feature extraction includes geometric feature extraction, segmentation and classification; Geometric features extract the geometric features of the device and use surface normals to calculate the detection plane area; Identify the vertical structure and plane distribution area of ​​the tower based on the change of normal vector; Segmentation and classification uses the region growing algorithm to segment the point cloud. Starting from the initial point, points with the same normal direction and a Euclidean distance less than the set threshold are gradually added to form a region. The device type is classified according to the shape characteristics of the region.

[0008] As a preferred solution of the method for developing a three-dimensional visualization dynamic operation and maintenance system for a distribution network described in the present invention, the scanning of the distribution equipment and environment to obtain three-dimensional point cloud data includes using a vehicle-mounted laser radar to achieve large-scale, high-density point cloud data collection, and equipping with a high-precision GPS and inertial measurement unit to ensure the accuracy of the equipment positioning information.

[0009] Use GIS data to build a geographical model of the distribution network area, generate the optimal scanning path, and import the scanning path into the drone and vehicle-mounted lidar navigation system to ensure coverage of all distribution network equipment and environments.

[0010] The steps to obtain 3D point cloud data include vehicle-mounted scanning and drone scanning. Vehicle-mounted scanning installs the laser radar on the top of the vehicle, and the vehicle drives slowly along the planned path to perform a 360° scan of the target area. The laser radar generates millions of points of data per second, and records the spatial coordinates (x, y, z) and reflection intensity in real time. Drone scanning is for the drone to fly along the planned path, start the laser radar during the flight to scan, collect point cloud data, and be equipped with a high-precision RTK module to correct GPS deviation in real time to ensure the accuracy of spatial positioning.

[0011] During the scanning process, the point cloud density is monitored in real time to ensure that the average density of the collected point cloud reaches more than 500 points / square meter.

[0012] Repeat scanning of important equipment to ensure that the point cloud density in key areas reaches 1,000 points per square meter.

[0013] As a preferred solution of the method for developing a three-dimensional visualization dynamic operation and maintenance system for a distribution network described in the present invention, the construction of a three-dimensional model of distribution network equipment includes performing grid reconstruction and using a Delaunay triangulation algorithm to construct a surface model of the equipment: Construct a triangular mesh for the points in the point cloud, ensuring that the maximum radius of the triangle circumscribed circle is less than the set maximum radius threshold.

[0014] Extract textures from high-resolution image data and map them onto mesh surfaces to enhance the realism of your model.

[0015] As a preferred solution of the distribution network three-dimensional visualization dynamic operation and maintenance system development method described in the present invention, wherein: the importing of the three-dimensional model and the real-time monitoring data of the distribution network into the visualization platform includes standardization of the three-dimensional model data format and real-time access to the monitoring data.

[0016] The 3D model data format standardization converts the constructed 3D model into GLTF format for use in the 3D visualization platform, and uses the 3D model nodes to bind one-to-one with the distribution network equipment numbers.

[0017] Real-time access to monitoring data is achieved through edge computing nodes to collect monitoring data, including voltage, current, temperature, load power and frequency. The data is uploaded to the central data server via the MQTT protocol and stored in a standardized JSON format.

[0018] As a preferred solution of the distribution network three-dimensional visualization dynamic operation and maintenance system development method described in the present invention, wherein: the evaluation of the distribution network operation status includes performing multi-level status evaluation, including node-level evaluation, regional-level evaluation and full-network evaluation.

[0019] Node-level assessment is based on the degree of state volatility and the dynamic risk index.

[0020] Regional level assessments are based on the amount of power imbalance and frequency stability.

[0021] The network-wide assessment is based on the risk level of the entire network.

[0022] The calculation of state fluctuation degree includes building a dynamic behavior model using the historical monitoring data of the distribution network nodes, building a state prediction model using the change of the state vector, and predicting the future state. .

[0023] Perform state stability analysis, calculate the state change rate of the node in the last n moments, and measure the degree of state fluctuation, expressed as:

[0024]

[0025] in, Indicates the state change rate of node i. Represents node i at time status. Indicates a time interval. Represents node i at time status.

[0026] The power imbalance is expressed as:

[0027]

[0028] in, represents the power imbalance in region r, Represents the power of load node i. represents the power of the generating node j. represents the regional power loss, Indicates load, Indicates power generation.

[0029] Frequency stability is the frequency of the node minus the reference frequency |F(t)−F0|.

[0030] The risk level of the entire network is expressed as:

[0031]

[0032] in, Represents the comprehensive risk of the entire network. Represents the dynamic risk index of node i. represents the weight of node i. Represents a node.

[0033] As a preferred solution of the distribution network three-dimensional visualization dynamic operation and maintenance system development method described in the present invention, the setting of different early warning mechanisms to timely discover abnormalities includes setting up multi-level early warning mechanisms according to the evaluation results of the node level, regional level and the whole network level, including first-level early warning, second-level early warning, third-level early warning and fourth-level early warning.

[0034] The trigger conditions for the first-level warning include:

[0035] Single node risk: R i ≤0.3.

[0036] Regional risk: ΔP r ≤ΔP max And |F(t)−F0|≤0.05.

[0037] Network-wide risk: R net ≤0.4.

[0038] At this time, the nodes and regions are running smoothly, with no obvious abnormalities, and the network status is in a stable period. No special measures are needed immediately, but continuous monitoring is maintained to ensure the continued stability of the network. Regular maintenance and inspection of node status ensures that the prediction system continues to operate normally.

[0039] The trigger conditions for the second-level warning include:

[0040] Single node risk: 0.3 <R i ≤0.6.

[0041] Regional risk: ΔP r >ΔP max Or |F(t)−F0|>0.05, but not more than 1.1 times the threshold.

[0042] Network-wide risk: 0.4 <R net ≤0.7.

[0043] At this time, there is a certain risk in the node or area, and there will be minor failures or fluctuations, which need to be closely monitored. It is necessary to strengthen the monitoring of key nodes and areas and regularly check the status of electrical equipment. Start the backup power supply and prepare to respond to potential power fluctuations. Track the areas with large frequency fluctuations in real time, analyze the source of the fluctuations, and avoid larger-scale failures. Highlight potential risk nodes and areas in the visualization platform to alert operation and maintenance personnel.

[0044] The trigger conditions for the third-level warning include:

[0045] Single node risk: R i >0.6.

[0046] Regional risk: ΔP r >1.1*ΔP max Or |F(t)−F0|>0.1.

[0047] Network-wide risk: R net >0.7.

[0048] At this time, the system faces a greater risk, which may cause regional or network-wide electrical failures, and requires immediate response. Start the automatic load adjustment system to balance the power load and prevent large-scale failures. Start the rapid fault isolation and recovery mechanism to quickly isolate the risk area from the entire network and reduce the spread of faults. Increase the startup frequency of backup power supplies to ensure stable network operation. Implement mandatory dispatch in areas with large frequency fluctuations to stabilize the grid frequency. Immediately notify the operation and maintenance team to conduct a detailed inspection of the abnormal area, quickly locate the source of the fault and repair it.

[0049] The four-level warning trigger conditions include:

[0050] Multiple nodes R i >0.8, and the regional power imbalance is serious.

[0051] At this time, there is a very high risk that the entire distribution network faces catastrophic failure, which may lead to a long-term outage. A comprehensive emergency recovery mechanism needs to be activated, including regional switching, load shedding, and backup power access, to ensure that the power system is restored to stability as soon as possible. Start the full-network automatic fault detection and recovery system (FDIR) to reduce the time delay of human intervention. Start the full-network load reduction strategy to reduce the overloaded area and gradually stabilize the load of the entire network. Send an emergency notice to all relevant departments through the remote control center to conduct full-network dispatch and repair the faulty nodes. Completely shut down the power supply to the unstable area to ensure that the stable area is supported to prevent a large-scale collapse of the entire network.

[0052] A distribution network three-dimensional visualization dynamic operation and maintenance system development system, characterized by: including:

[0053] The data acquisition module scans the power distribution equipment and environment to obtain three-dimensional point cloud data.

[0054] The data processing module pre-processes and extracts features of point cloud data to build a three-dimensional model of distribution network equipment.

[0055] The evaluation module imports the three-dimensional model and real-time monitoring data of the distribution network into the visualization platform to evaluate the operating status of the distribution network.

[0056] The early warning module sets different early warning mechanisms to detect abnormalities in a timely manner.

[0057] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

[0058] A computer-readable storage medium stores a computer program, which implements the steps of the method described above when executed by a processor.

[0059] Beneficial effects of the present invention: Through real-time point cloud data collection and processing, the dynamic changes of the distribution network can be instantly reflected on the three-dimensional visualization platform, improving the speed of fault detection and response. The high-precision collection and processing of point cloud data ensures the accuracy of the equipment model, thereby improving the accuracy of status assessment and fault location. Combined with machine learning and multi-level status assessment, the present invention realizes the intelligent operation and maintenance of the distribution network, can automatically identify risks and generate early warnings, reduce human operational errors, and improve the stability and safety of the system. Through the three-dimensional visualization platform, operation and maintenance personnel can more intuitively understand the status and fault conditions of the distribution network, optimize the decision-making process, and improve work efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0061] Figure 1 A distribution network three-dimensional visualization dynamic operation and maintenance system development method and an overall flow chart of the system are provided for the first embodiment of the present invention. DETAILED DESCRIPTION

[0062] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0063] Example 1, reference Figure 1 , is an embodiment of the present invention, and provides a method for developing a distribution network three-dimensional visualization dynamic operation and maintenance system, comprising:

[0064] S1: Scan the power distribution equipment and environment to obtain 3D point cloud data.

[0065] The vehicle-mounted laser radar is used to achieve large-scale, high-density point cloud data collection, and is equipped with high-precision GPS and inertial measurement units to ensure the accuracy of equipment positioning information.

[0066] Use GIS data to build a geographical model of the distribution network area, generate the optimal scanning path, and import the scanning path into the drone and vehicle-mounted lidar navigation system to ensure coverage of all distribution network equipment and environments.

[0067] The steps to obtain 3D point cloud data include vehicle-mounted scanning and drone scanning. Vehicle-mounted scanning installs the laser radar on the top of the vehicle, and the vehicle drives slowly along the planned path to perform a 360° scan of the target area. The laser radar generates millions of points of data per second, and records the spatial coordinates (x, y, z) and reflection intensity in real time. Drone scanning is for the drone to fly along the planned path, start the laser radar during the flight to scan, collect point cloud data, and be equipped with a high-precision RTK module to correct GPS deviation in real time to ensure the accuracy of spatial positioning.

[0068] During the scanning process, the point cloud density is monitored in real time to ensure that the average density of the collected point cloud reaches more than 500 points / square meter.

[0069] Repeat scanning of important equipment to ensure that the point cloud density in key areas reaches 1,000 points per square meter.

[0070] S2: Preprocess and extract features of point cloud data to build a three-dimensional model of distribution network equipment.

[0071] Preprocessing includes data denoising, downsampling, and coordinate alignment.

[0072] Data denoising uses statistical filtering to eliminate isolated points, which can be expressed as:

[0073]

[0074] in, Yes The average neighborhood distance. is the target point in the point cloud. is the neighborhood point of the target point in the point cloud. is the number of neighborhood points, which is 50 in the present invention.

[0075] It should be noted that if If the value exceeds the set threshold, the point is considered as a noise point and is removed. The threshold is set as the mean value plus 2 times the standard deviation in the present invention.

[0076] Furthermore, in point cloud data processing, isolated points refer to points in the data set that obviously deviate from the normal data distribution. These points are usually caused by noise, sensor errors, environmental interference, etc. In point cloud data, the existence of isolated points will affect the subsequent data processing and modeling effects, resulting in inaccurate three-dimensional models and erroneous physical analysis results. Therefore, eliminating isolated points is a crucial step in point cloud data preprocessing. Statistical filtering is a method that identifies and removes these isolated points through statistical features. This method uses the local neighborhood information of point cloud data and makes judgments based on the distance distribution between points and neighboring points. When the distance between a point and other points in its neighborhood deviates from the expected statistical characteristics, the point is considered to be an isolated point. In this way, abnormal and irrelevant points can be efficiently removed and normal point cloud data can be retained.

[0077] benefit:

[0078] Improve data quality: The existence of isolated points may lead to inaccurate 3D modeling and affect the construction of distribution network equipment models. By eliminating isolated points, the quality of point cloud data can be significantly improved, making subsequent modeling, analysis and monitoring more accurate.

[0079] Reduce calculation errors: Isolated points may introduce calculation errors and affect the stability of the algorithm. By filtering out isolated points, the robustness of the algorithm can be improved, making subsequent processing (such as feature extraction, modeling, etc.) more stable and efficient.

[0080] Optimize subsequent processing efficiency: The removal of isolated points can reduce the computational complexity of subsequent steps, reduce the burden of invalid calculations, and improve the system processing speed.

[0081] Downsampling uses the voxel grid downsampling method to divide the point cloud into a fixed-size cubic voxel grid. The centroid of all points in each grid is calculated as the representative point, which reduces the amount of point cloud data and improves processing efficiency.

[0082] Coordinate alignment uses the rigid body transformation formula to align the coordinate system, expressed as:

[0083]

[0084] in, is the aligned point. is the rotation matrix (calculated using the attitude angles provided by the IMU). is the translation vector (calculated from the position data provided by GPS). are the coordinates of the original point.

[0085] Feature extraction includes geometric feature extraction, segmentation and classification.

[0086] The geometric features of the device are extracted by calculating the detection plane area using the surface normal, which is expressed as:

[0087]

[0088] in, Indicate point The normal vector of . Indicates the number of neighborhood points. Represents a neighborhood point.

[0089] The vertical structure and plane distribution area of ​​the tower are identified based on the change of the normal vector.

[0090] It should be noted that in the processing of 3D point cloud data, surface normal is an important parameter to describe the local features of the point cloud data surface. By calculating the normal of each point, the plane area and other geometric features in the point cloud can be effectively identified. The detection of plane areas is particularly important in the 3D modeling and condition monitoring of distribution network equipment, because the equipment in the distribution network (such as transformers, switches, etc.) usually has a flat surface or structure.

[0091] Benefits of surface normal calculation:

[0092] Plane area detection: By calculating the normal of each point in the point cloud data, it is possible to determine whether the point is located on a plane. Plane area detection is crucial for 3D modeling of distribution network equipment because the equipment is usually flat (for example, distribution cabinets, panels, etc.). Accurately detecting these plane areas can provide high-quality basic data for equipment modeling.

[0093] Improve model accuracy: Normal calculation can not only help identify plane areas, but also further segment and classify point cloud data. For example, for the equipment area in the distribution network, normal calculation can be used to accurately segment it, ensuring that the 3D model of each device accurately reflects its true form and avoids model distortion due to noise in point cloud data.

[0094] Optimize point cloud processing efficiency: Extraction of planar regions helps reduce the complexity of point cloud data. By identifying planar regions, subsequent point cloud simplification and compression can be performed more efficiently, reducing interference from irrelevant areas and increasing processing speed.

[0095] Segmentation and classification Use the region growing algorithm to segment the point cloud. Starting from the initial point, gradually add points with the same normal direction and Euclidean distance less than the set threshold to form a region. Classify the device type according to the shape characteristics of the region.

[0096] Furthermore, in the present invention, the threshold is set to 5 cm.

[0097] Perform mesh reconstruction and use the Delaunay triangulation algorithm to build the device surface model:

[0098] Construct a triangular mesh for the points in the point cloud, ensuring that the maximum radius of the triangle circumscribed circle is less than the set maximum radius threshold.

[0099] It should be noted that in the present invention, the threshold is set between 0.01 meters and 0.1 meters, and the specific value should be adjusted according to the following standards:

[0100] For high-precision models (such as key equipment in distribution networks and refined modeling), the threshold can be set to 0.01 meters to ensure that the radius of the circumscribed circle is small enough to generate a high-precision triangular mesh that accurately reflects the surface features of the equipment.

[0101] For coarser models (such as regional equipment in distribution networks or low-precision requirements), the threshold can be set to 0.05 meters to 0.1 meters. This can reduce the computational burden and improve processing efficiency while ensuring accuracy.

[0102] Extract textures from high-resolution image data and map them onto mesh surfaces to enhance the realism of your model.

[0103] S3: Import the 3D model and real-time monitoring data of the distribution network into the visualization platform to evaluate the operating status of the distribution network.

[0104] Importing the three-dimensional model and real-time monitoring data of the distribution network into the visualization platform includes standardization of the three-dimensional model data format and real-time access to the monitoring data.

[0105] The 3D model data format standardization converts the constructed 3D model into GLTF format for use in the 3D visualization platform, and uses the 3D model nodes to bind one-to-one with the distribution network equipment numbers.

[0106] Real-time access to monitoring data is achieved through edge computing nodes to collect monitoring data, including voltage, current, temperature, load power and frequency. The data is uploaded to the central data server via the MQTT protocol and stored in a standardized JSON format.

[0107] Conduct multi-level status assessments, including node-level assessments, regional-level assessments, and full-network assessments.

[0108] Node-level evaluation is based on the degree of state fluctuation.

[0109] Regional level assessments are based on the amount of power imbalance and frequency stability.

[0110] The network-wide assessment is based on the risk level of the entire network.

[0111] The calculation of state fluctuation degree includes building a dynamic behavior model using the historical monitoring data of the distribution network nodes, building a state prediction model using the change of the state vector, and predicting the future state. , time series data modeling is expressed as:

[0112]

[0113] in, Represents the state vector of node i at time t. are the node voltage and current. is the node temperature. are the node active power and reactive power. It's the frequency.

[0114] Use the changes in the state vector to build a state prediction model (GRU deep neural network) to predict future states:

[0115]

[0116] Perform state stability analysis, calculate the state change rate of the node in the last n moments, and measure the degree of state fluctuation, expressed as:

[0117]

[0118] in, Indicates the state change rate of node i. Represents node i at time status. Indicates a time interval. Represents node i at time status.

[0119] The power imbalance is expressed as:

[0120]

[0121] in, represents the power imbalance in region r, Represents the power of load node i. represents the power of the generating node j. Represents the regional power loss.

[0122] Frequency stability is the frequency of the node minus the reference frequency |F(t)−F0|.

[0123] The risk level of the entire network is expressed as:

[0124]

[0125] in, Represents the comprehensive risk of the entire network. Represents the dynamic risk index of node i. represents the weight of node i.

[0126] S4: Set up different early warning mechanisms to detect abnormalities in a timely manner.

[0127] Based on the assessment results at the node level, regional level and network level, a multi-level early warning mechanism is set up, including level one, level two, level three and level four warning.

[0128] The trigger conditions for the first-level warning include:

[0129] Single node risk: R i ≤0.3.

[0130] Regional risk: ΔP r ≤ΔP max And |F(t)−F0|≤0.05.

[0131] Network-wide risk: R net ≤0.4.

[0132] At this time, the nodes and regions are running smoothly, with no obvious abnormalities, and the network status is in a stable period. No special measures are needed immediately, but continuous monitoring is maintained to ensure the continued stability of the network. Regular maintenance and inspection of node status ensures that the prediction system continues to operate normally.

[0133] The trigger conditions for the second-level warning include:

[0134] Single node risk: 0.3 <R i ≤0.6.

[0135] Regional risk: ΔP r >ΔP max Or |F(t)−F0|>0.05, but not more than 1.1 times the threshold.

[0136] Network-wide risk: 0.4 <R net ≤0.7.

[0137] At this time, there is a certain risk in the node or area, and there will be minor failures or fluctuations, which need to be closely monitored. It is necessary to strengthen the monitoring of key nodes and areas and regularly check the status of electrical equipment. Start the backup power supply and prepare to respond to potential power fluctuations. Track the areas with large frequency fluctuations in real time, analyze the source of the fluctuations, and avoid larger-scale failures. Highlight potential risk nodes and areas in the visualization platform to alert operation and maintenance personnel.

[0138] The trigger conditions for the third-level warning include:

[0139] Single node risk: R i >0.6.

[0140] Regional risk: ΔP r >1.1*ΔP max Or |F(t)−F0|>0.1.

[0141] Network-wide risk: R net >0.7.

[0142] At this time, the system faces a greater risk, which may cause regional or network-wide electrical failures, and requires immediate response. Start the automatic load adjustment system to balance the power load and prevent large-scale failures. Start the rapid fault isolation and recovery mechanism to quickly isolate the risk area from the entire network and reduce the spread of faults. Increase the startup frequency of backup power supplies to ensure stable network operation. Implement mandatory dispatch in areas with large frequency fluctuations to stabilize the grid frequency. Immediately notify the operation and maintenance team to conduct a detailed inspection of the abnormal area, quickly locate the source of the fault and repair it.

[0143] The four-level warning trigger conditions include:

[0144] Multiple nodes R i >0.8, and the regional power imbalance is serious.

[0145] At this time, there is a very high risk that the entire distribution network faces catastrophic failure, which may lead to a long-term outage. A comprehensive emergency recovery mechanism needs to be activated, including regional switching, load shedding, and backup power access, to ensure that the power system is restored to stability as soon as possible. Start the full-network automatic fault detection and recovery system (FDIR) to reduce the time delay of human intervention. Start the full-network load reduction strategy to reduce the overloaded area and gradually stabilize the load of the entire network. Send an emergency notice to all relevant departments through the remote control center to conduct full-network dispatch and repair the faulty nodes. Completely shut down the power supply to the unstable area to ensure that the stable area is supported to prevent a large-scale collapse of the entire network.

[0146] Furthermore, ΔP max It indicates the maximum permissible deviation of regional power balance in the distribution network.

[0147] It should be noted that multi-level status assessment is an important technology in the present invention, which divides the health status assessment of the distribution network into multiple levels (such as node level, regional level and whole network level), and analyzes and comprehensively judges the operating status of the distribution network layer by layer. The core advantage of this method is that it can more comprehensively and accurately identify potential risks in the system, thereby effectively preventing the occurrence of faults.

[0148] benefit:

[0149] Accurately locate the problem: By refining the assessment to the node level, regional level, and network level, the specific location and scope of the problem can be accurately identified. For example, if the dynamic risk index R of a node is i If the risk is high, the node can be quickly located for key monitoring and maintenance. If the risk increases at the network level, it is necessary to consider whether there are systemic failures such as global load problems or frequency instability.

[0150] Hierarchical management: Multi-level assessment provides a more flexible management method. When the risk of the entire network is high, specific areas can be monitored. When the risk of a single node or area is high, local repairs or optimizations can be carried out. This hierarchical assessment method can help operators make targeted decisions and improve response efficiency.

[0151] Enhance the system's fault tolerance: Through hierarchical risk assessment, potential failures and bottlenecks can be discovered at an early stage and preventive measures can be taken, rather than waiting for failures to occur before responding. This predictive maintenance helps reduce network-wide failures and improve system stability and reliability.

[0152] Improve decision support: Multi-level status assessment provides operators with rich information. Comprehensive assessment results from nodes to the entire network help decision makers take corresponding measures according to different risk levels. This makes decision making more scientific and accurate, helps maximize the use of distribution network resources, and improves the overall operating efficiency of the system.

[0153] Enhanced intelligent operation and maintenance: Multi-level status assessment can be combined with the intelligent operation and maintenance system to dynamically adjust system operating parameters (such as load, power generation, etc.), thereby reducing energy consumption and improving operation and maintenance efficiency. Through real-time monitoring and automatic adjustment, more intelligent and efficient distribution network operation and maintenance can be achieved.

[0154] The above embodiments also include a distribution network three-dimensional visualization dynamic operation and maintenance system development system, specifically:

[0155] The data acquisition module scans the power distribution equipment and environment to obtain three-dimensional point cloud data.

[0156] The data processing module pre-processes and extracts features of point cloud data to build a three-dimensional model of distribution network equipment.

[0157] The evaluation module imports the three-dimensional model and real-time monitoring data of the distribution network into the visualization platform to evaluate the operating status of the distribution network.

[0158] The early warning module sets different early warning mechanisms to detect abnormalities in a timely manner.

[0159] The computer device may be a server. The computer device includes a processor, a memory, an input / output interface (I / O for short) and a communication interface. The processor, the memory and the input / output interface are connected via a system bus, and the communication interface is connected to the system bus via the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data cluster data of the power monitoring system. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a method for developing a three-dimensional visualization dynamic operation and maintenance system for a distribution network is realized.

[0160] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0161] Example 2 is an embodiment of the present invention, which provides a distribution network three-dimensional visualization dynamic operation and maintenance system development method and system. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through simulation experiments.

[0162] In order to verify the advantages and innovation of the present invention, a comparative experiment was conducted to compare the effects of the distribution network three-dimensional visualization dynamic operation and maintenance system of the present invention and the existing technology in the distribution network operation and maintenance. The experiment involved steps such as point cloud data acquisition, three-dimensional model reconstruction, real-time monitoring and dynamic state evaluation, and intelligent early warning system. This experiment used typical equipment of the distribution network (such as transformers, switches, lines, etc.) and their corresponding monitoring data.

[0163] In the existing technology, distribution network operation and maintenance mainly rely on traditional two-dimensional electrical drawings and manual inspection methods. Although some systems are also using point cloud data for modeling, due to limited computing power, there are usually problems such as low data collection accuracy, poor model accuracy, and weak real-time monitoring and early warning mechanisms. The specific implementation process is as follows:

[0164] Point cloud data collection: LiDAR equipment is used to collect point cloud data of the distribution network area, but its accuracy is usually between 0.2m and 0.5m.

[0165] 3D model construction: The traditional grid modeling method is used to reconstruct the 3D model, which has low accuracy and slow processing speed.

[0166] Real-time monitoring and status assessment: Equipment monitoring is performed using manual inspection and simplified calculation methods. Status assessment is mainly based on manual data input and lacks intelligent analysis.

[0167] Early warning mechanism: The system issues early warnings through manually set thresholds, which takes a long time to respond and makes it difficult to detect abnormalities in a timely manner.

[0168] The implementation process of the present invention:

[0169] This invention proposes a three-dimensional visualization dynamic operation and maintenance system for distribution networks based on point cloud technology, which combines high-precision point cloud data acquisition, real-time dynamic monitoring, intelligent evaluation and automatic early warning mechanism. The implementation steps are as follows:

[0170] Point cloud data collection and preprocessing: Use high-precision laser radar (accuracy up to 0.05m) to collect point cloud data, and use statistical filtering to remove isolated points to ensure high data quality.

[0171] 3D model construction and feature extraction: The surface normal calculation method is used to extract the geometric features of the equipment, and the efficient Delaunay triangulation method is used to reconstruct an accurate 3D model, ensuring that the circumscribed circle radius of each triangle is less than the set threshold of 0.05m, thereby obtaining a high-precision 3D model.

[0172] Real-time monitoring and status assessment: By integrating multiple sensors (voltage, current, temperature, etc.), the equipment status is monitored in real time, combined with machine learning models to conduct dynamic health status assessment and automatically generate early warning signals.

[0173] Intelligent early warning and decision support: The system automatically analyzes the risk level of the equipment by comparing real-time data with historical data, and generates automated responses based on the early warning level, reducing manual intervention.

[0174] The experimental results are shown in Table 1.

[0175] Table 1 Experimental results

[0176] From the experimental data table, it can be clearly seen that the present invention has obvious innovation and advantages in many aspects compared with the prior art:

[0177] In the prior art, the point cloud data acquisition accuracy is low and the error is large, and the acquisition error is usually around 0.2 meters. However, the present invention uses high-precision laser radar and combines statistical filtering to remove isolated points, successfully improving the point cloud accuracy to 0.05 meters and reducing the denoising error to 0.05 meters. This significantly improves the accuracy of subsequent three-dimensional models and provides a reliable data foundation for accurate modeling of distribution network equipment.

[0178] In terms of three-dimensional modeling, the model construction time of the prior art is 90 minutes, while the construction time of the present invention is 60 minutes, which reduces the modeling time by 33.3%. At the same time, the model accuracy of the prior art is low, and the maximum radius of the triangle circumscribed circle is 0.2 meters, while the maximum radius of the triangle circumscribed circle is strictly controlled within 0.05 meters in the present invention, ensuring the fineness of the model. This advantage enables the present invention to generate a more accurate and efficient three-dimensional distribution network model, providing more accurate decision support for operation and maintenance work.

[0179] In terms of data transmission delay, the system response time of the present invention is significantly shorter than that of the prior art. The real-time data transmission delay of the prior art is 5.5 seconds, while the delay of the present invention is shortened to 2.5 seconds. This optimization means that the dynamic monitoring system of the distribution network can respond to system status changes more quickly, transmit data in real time and perform status assessment, thereby improving the operational safety of the distribution network.

[0180] In terms of the accuracy of equipment status assessment, the accuracy of the existing technology is only 80%, while the system of the present invention has achieved an accuracy of more than 95%, an improvement of more than 15%. By integrating advanced machine learning algorithms, the present invention can more accurately assess the health status of equipment, thereby identifying faults in advance and generating early warnings. This innovation improves the intelligence level of the distribution network, reduces human operational errors, and improves the efficiency and accuracy of fault detection.

[0181] System response time is one of the key indicators that affect the operation and maintenance efficiency of the distribution network. In the prior art, the system response time is usually around 10 seconds, while the response time of the present invention is shortened to 3 seconds, greatly improving the response speed and operation and maintenance efficiency of the system.

[0182] From the comparison of experimental data, it can be seen that the present invention is significantly superior to the existing technology in many aspects. By improving the acquisition accuracy of point cloud data, reducing modeling time, and optimizing data transmission and evaluation processes, the present invention not only improves the accuracy of three-dimensional modeling of the distribution network, but also makes the operation and maintenance work more efficient, intelligent and accurate through intelligent analysis and rapid response mechanisms. These innovations and advantages give the present invention a clear leading advantage in real-time monitoring, status assessment and fault warning, and have broad application prospects.

[0183] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for developing a three-dimensional visualization dynamic operation and maintenance system for a distribution network, characterized in that: include: Scan power distribution equipment and environment to obtain 3D point cloud data; Preprocess and extract features of point cloud data to build a three-dimensional model of distribution network equipment; Import the 3D model and real-time monitoring data of the distribution network into the visualization platform to evaluate the operating status of the distribution network; Set up different early warning mechanisms to detect abnormalities in a timely manner; Preprocessing and feature extraction of point cloud data include data denoising, downsampling and coordinate alignment; Data denoising uses statistical filtering to eliminate isolated points; Downsampling uses the voxel grid downsampling method to divide the point cloud into a fixed-size cubic voxel grid. The centroid of all points in each grid is calculated as the representative point, which reduces the amount of point cloud data and improves processing efficiency. Coordinate alignment uses the rigid body transformation formula to align the coordinate system; Feature extraction includes geometric feature extraction, segmentation and classification; Geometric features extract the geometric features of the device and use surface normals to calculate the detection plane area; Identify the vertical structure and plane distribution area of ​​the tower based on the change of normal vector; Segmentation and classification uses the region growing algorithm to segment the point cloud. Starting from the initial point, points with the same normal direction and a Euclidean distance less than the set threshold are gradually added to form a region. The device type is classified according to the shape characteristics of the region.

2. The method for developing a three-dimensional visualization dynamic operation and maintenance system for a distribution network according to claim 1, characterized in that: The scanning of the power distribution equipment and environment to obtain three-dimensional point cloud data includes using a vehicle-mounted laser radar to achieve large-scale, high-density point cloud data collection, and equipping with a high-precision GPS and inertial measurement unit to ensure the accuracy of equipment positioning information; Use GIS data to build a geographic model of the distribution network area, generate the optimal scanning path, and import the scanning path into the drone and vehicle-mounted lidar navigation system to ensure coverage of all distribution network equipment and environments; The steps to obtain 3D point cloud data include vehicle-mounted scanning and drone scanning. Vehicle-mounted scanning installs the laser radar on the top of the vehicle, and the vehicle drives slowly along the planned path to perform a 360° scan of the target area. The laser radar generates millions of points of data per second and records the spatial coordinates (x, y, z) and reflection intensity in real time. Drone scanning is for the drone to fly along the planned path, start the laser radar during the flight to scan and collect point cloud data. It is equipped with a high-precision RTK module to correct GPS deviation in real time to ensure the accuracy of spatial positioning. Monitor the point cloud density in real time during the scanning process to ensure that the average density of the collected point cloud reaches more than 500 points / square meter; Repeat scanning of important equipment to ensure that the point cloud density in key areas reaches 1,000 points per square meter.

3. The method for developing a three-dimensional visualization dynamic operation and maintenance system for a distribution network according to claim 2, characterized in that: The construction of the three-dimensional model of the distribution network equipment includes performing mesh reconstruction and using the Delaunay triangulation algorithm to construct the equipment surface model: Construct a triangular mesh for the points in the point cloud, ensuring that the maximum radius of the triangle circumscribed circle is less than the set maximum radius threshold; Extract textures from high-resolution image data and map them onto mesh surfaces to enhance the realism of your model.

4. The method for developing a three-dimensional visualization dynamic operation and maintenance system for a distribution network according to claim 3, characterized in that: The importing of the three-dimensional model and the real-time monitoring data of the power distribution network into the visualization platform includes standardization of the three-dimensional model data format and real-time access to the monitoring data; Standardization of 3D model data format Convert the constructed 3D model into GLTF format for use in the 3D visualization platform, and use the 3D model nodes to bind one-to-one with the distribution network equipment numbers; Real-time access to monitoring data: Monitoring data, including voltage, current, temperature, load power and frequency, is collected through edge computing nodes; data is uploaded to the central data server via the MQTT protocol and stored in a standardized JSON format.

5. The method for developing a three-dimensional visualization dynamic operation and maintenance system for a distribution network according to claim 4, characterized in that: The evaluation of the distribution network operation status includes performing multi-level status evaluation, including node-level evaluation, regional-level evaluation and full-network evaluation; Node-level evaluation is based on the degree of state fluctuation; Regional-level assessments are based on the amount of power imbalance and frequency stability; The network-wide assessment is based on the risk level of the entire network; The calculation of state fluctuation degree includes building a dynamic behavior model using the historical monitoring data of the distribution network nodes, building a state prediction model using the change of the state vector, and predicting the future state. ; Perform state stability analysis, calculate the state change rate of the node in the last n moments, and measure the degree of state fluctuation , expressed as: ,in, represents the state change rate of node i; Represents node i at time Status; Indicates a time interval; Represents node i at time Status; The power imbalance is expressed as: ,in, represents the power imbalance in region r, represents the power of load node i; represents the power of the generating node j; represents the regional power loss, Indicates load, Indicates power generation; Frequency stability is the node frequency minus the reference frequency |F(t)−F0|; The risk level of the entire network is expressed as: ,in, Indicates the comprehensive risk of the entire network; represents the dynamic risk index of node i; represents the weight of node i, Represents a node.

6. The method for developing a three-dimensional visualization dynamic operation and maintenance system for a distribution network according to claim 5, characterized in that: The setting of different early warning mechanisms to timely detect anomalies includes setting a multi-level early warning mechanism according to the evaluation results of the node level, the regional level and the whole network level, including the first level early warning, the second level early warning, the third level early warning and the fourth level early warning; The trigger conditions for the first-level warning include: Single node risk: R i ≤0.3; Regional risk: ΔP r ≤ΔP max and |F(t)−F0|≤0.05; Network-wide risk: R net ≤0.4; At this time, the nodes and regions are running smoothly without obvious abnormalities, and the network status is in a stable period. No special measures need to be taken immediately, but continuous monitoring is maintained to ensure the continued stability of the network; regular maintenance and inspection of node status are carried out to ensure the continued normal operation of the prediction system; The trigger conditions for the second-level warning include: Single node risk: 0.3 <R i ≤0.6; Regional risk: ΔP r >ΔP max Or |F(t)−F0|>0.05, but not more than 1.1 times the threshold; Network-wide risk: 0.4 <R net ≤0.7; At this time, there are certain risks in nodes or areas, and there will be minor failures or fluctuations, which need to be closely monitored; it is necessary to strengthen the monitoring of key nodes and areas, and regularly check the status of electrical equipment; start the backup power supply and prepare to respond to potential power fluctuations; track areas with large frequency fluctuations in real time, analyze the source of the fluctuations, and avoid larger-scale failures; highlight potential risk nodes and areas in the visualization platform to alert operation and maintenance personnel; The trigger conditions for the third-level warning include: Single node risk: R i >0.6; Regional risk: ΔP r >1.1*ΔP max or |F(t)−F0|>0.1; Network-wide risk: R net >0.7; At this time, the system faces a greater risk, which may cause regional or network-wide electrical failures, requiring immediate response; start the automatic load adjustment system to balance the power load and prevent large-scale failures; start the rapid fault isolation and recovery mechanism to quickly isolate the risk area from the entire network and reduce the spread of faults; increase the startup frequency of backup power supplies to ensure stable network operation; implement mandatory dispatching in areas with large frequency fluctuations to stabilize the grid frequency; immediately notify the operation and maintenance team to conduct a detailed inspection of the abnormal area, quickly locate the source of the fault and repair it; The four-level warning trigger conditions include: Multiple nodes R i >0.8, and the regional power imbalance is serious; There is an extremely high risk at this time. The entire distribution network is facing catastrophic failure, which may lead to a long-term shutdown. It is necessary to activate a comprehensive emergency recovery mechanism, including regional switching, load shedding, and backup power access to ensure that the power system returns to stability as soon as possible; start the full-network automatic fault detection and recovery system (FDIR) to reduce the time delay of human intervention; initiate a full-network load reduction strategy to reduce overloaded areas and gradually stabilize the load of the entire network; send emergency notifications to all relevant departments through the remote control center, conduct full-network dispatch, and repair faulty nodes; completely shut down the power supply to unstable areas to ensure that stable areas are supported and prevent large-scale collapse of the entire network.

7. A distribution network three-dimensional visualization dynamic operation and maintenance system development system using the method according to any one of claims 1 to 6, characterized in that: Acquisition data module, scans power distribution equipment and environment, and obtains three-dimensional point cloud data; The data processing module pre-processes and extracts features from point cloud data to build a three-dimensional model of distribution network equipment; The evaluation module imports the 3D model and real-time monitoring data of the distribution network into the visualization platform to evaluate the operating status of the distribution network; The early warning module sets different early warning mechanisms to detect abnormalities in a timely manner.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images