Power grid reactive dynamic compensation real-time visual tracking system based on optical flow field analysis

Abstract

The invention discloses a power grid reactive dynamic compensation real-time visual tracking system based on optical flow field analysis, and relates to the technical field of power grid compensation visualization, and a multi-module cooperative system including image acquisition, electric power parameter acquisition, optical flow field calculation, comprehensive analysis, dynamic compensation control and visual interaction is constructed. Accurate recognition and response control of dynamic changes of reactive power of a power grid are achieved, the system intelligently discriminates a disturbance area by fusing optical flow vectors and electric power parameter characteristics, compensation equipment is guided to adjust in real time, and meanwhile the operation state of the power grid is displayed in a graphical mode. The accuracy and real-time performance of reactive compensation and the stability of system operation are effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid compensation visualization, and particularly to a real-time visualization tracking system for dynamic reactive power compensation of power grids based on optical flow field analysis. Background Art

[0002] Real-time visualization tracking of dynamic reactive power compensation in power grids refers to achieving visual management of the operating state of power grids by monitoring and displaying the dynamic compensation process of reactive power in power grids in real time. Specifically, it collects data on changes in reactive power through sensors and control systems, and uses visualization technology to present key parameters such as the working state of compensation devices, compensation amounts, and grid voltages in a graphical form on a monitoring platform, helping operation and maintenance personnel to grasp the power balance of the power grid in real time, adjust strategies in a timely manner, and improve power quality and system stability.

[0003] The following deficiencies exist in the prior art: When deploying cameras in outdoor substations for optical flow analysis in the prior art, limited by the rapid changes in environmental lighting conditions (such as heavy rain, haze, or day-night alternation), the image quality is prone to fluctuations. Since optical flow algorithms usually rely on the premise of image brightness continuity, this assumption no longer holds under drastic lighting changes, resulting in distorted calculation of motion vectors, which may cause the system to misjudge a significant fluctuation in reactive power and wrongly trigger compensation devices, interfering with the normal voltage regulation of the power grid and affecting system stability and control accuracy. Summary of the Invention

[0004] The purpose of the present invention is to provide a real-time visualization tracking system for dynamic reactive power compensation of power grids based on optical flow field analysis to solve the deficiencies in the background art.

[0005] To achieve the above purpose, the present invention provides the following technical solution: A real-time visualization tracking system for dynamic reactive power compensation of power grids based on optical flow field analysis, including an image acquisition module, a power parameter acquisition module, an optical flow field calculation module, a comprehensive analysis module, a dynamic compensation control module, and a visualization interaction module; The image acquisition module is used to deploy camera devices in key areas of the power grid to obtain a continuous sequence of monitoring images; The power parameter acquisition module is used to collect real-time power grid operation parameters corresponding to the image area, including voltage, current, and reactive power; The optical flow field calculation module is used to perform optical flow analysis based on the monitoring image sequence and calculate the motion vector field of each pixel point; The comprehensive analysis module is used to perform spatio-temporal corresponding fusion of the optical flow vector field and the power operation parameters to identify the dynamic change trend and spatial distribution of reactive power in the power grid; A dynamic compensation control module, which is used to generate a compensation control instruction according to the comprehensive analysis result and control the static var compensator to implement real-time reactive power compensation for the target area; A visualization interaction module, which is used to display and interact with the power grid state, optical flow change and compensation control result in real time in the form of a graphical interface.

[0006] Preferably, the key areas include the main transformer area, the bus access point and the reactive power compensation device.

[0007] Preferably, it receives continuous frame images from the image acquisition module, uniformly adjusts them to grayscale images, preprocesses the input images, dynamically switches the enhancement strategy according to different scene conditions, and dynamically selects the dense or sparse optical flow algorithm according to the computing resources and accuracy requirements; performs pixel-level matching and motion estimation on two consecutive frame images, and outputs the horizontal and vertical displacements of each pixel; generates a two-dimensional optical flow vector field, that is, each pixel corresponds to a motion vector, indicating its relative displacement from the previous frame to the current frame.

[0008] Preferably, it extracts the motion features in the optical flow vector field, including the local amplitude mutation value, and the power parameter changes in the area, including the reactive power sudden increase value; Among them, the method for obtaining the local amplitude mutation value is: perform optical flow calculation on consecutive image frames, obtain the horizontal displacement u(x,y) and vertical displacement v(x,y) of each pixel point, and calculate the optical flow amplitude: ; where: M(x,y) is the optical flow amplitude of the pixel point, that is, its motion intensity; divide the region of interest in the image and calculate the average amplitude of each region: ; then calculate the local amplitude change between consecutive frames: ; where: is the average amplitude of the ROI region R at time t; ΔM(R) is the local amplitude mutation value; N is the number of pixel points in the region; when ΔM(R) exceeds the set threshold Tm, it is regarded as the local amplitude mutation value.

[0009] Preferably, among them, the method for obtaining the reactive power sudden increase value is: let the reactive power of the power node corresponding to a certain region at time t be , the sampling period is Δt, and the reactive power sudden increase calculation between two consecutive sampling periods is: ; where: ΔQ is the reactive power sudden increase value.

[0010] Preferably, the local amplitude mutation value and the reactive power mutation value are converted into a comprehensive feature vector, and the comprehensive feature vector is used as the input of the machine learning model. The machine learning model takes the evaluation value label corresponding to the optical flow change and the actual power disturbance predicted by each group of comprehensive feature vectors as the prediction target, and takes minimizing the sum of the prediction errors of all the evaluation value labels corresponding to the optical flow change and the actual power disturbance as the training target to train the machine learning model until the sum of the prediction errors reaches convergence and then stops the model training. The evaluation value corresponding to the optical flow change and the actual power disturbance is determined according to the model output result, wherein the machine learning model is a polynomial regression model.

[0011] Preferably, the evaluation value corresponding to the optical flow change and the actual power disturbance is compared with a predetermined threshold. If the evaluation value corresponding to the optical flow change and the actual power disturbance is greater than the predetermined threshold, it is determined that the optical flow change height is highly correlated with the actual power disturbance event and is determined as the area requiring compensation response; if the evaluation value corresponding to the optical flow change and the actual power disturbance is less than or equal to the predetermined threshold, it is considered that there is no obvious disturbance in the area and there is no need to immediately trigger the compensation control.

[0012] Preferably, based on the characteristics of each monitoring area in the current period: the average value of the optical flow amplitude, the reactive power mutation value, and the duration of the optical flow change; a regional feature vector is constructed: all areas are clustered by using the K-means clustering algorithm, and several reactive power disturbance hot spots are automatically divided; each type of area is sorted according to the characteristics of its center point, and different levels of compensation priorities are assigned. The high-priority areas are preferentially matched with compensation devices with fast dynamic response capabilities; the disturbance recognition results in each time period are integrated to generate a dynamic change distribution map; the horizontal axis is the time axis, and the vertical axis is the area number or coordinate; the color depth or symbol represents the optical flow disturbance intensity or the reactive power mutation degree; the evolution path of the reactive power disturbance in the spatial and temporal dimensions is visually displayed.

[0013] In the above technical solution, the technical effects and advantages provided by the present invention are as follows: 1. By constructing a multi-module system integrating image analysis and power data processing, the present invention realizes high-precision recognition and real-time response control of the dynamic change of the reactive power of the power grid. Compared with the problem of easy distortion of the prior art under complex lighting conditions, the present invention introduces a comprehensive feature extraction mechanism combining the detection of optical flow amplitude mutation and the analysis of reactive power sudden increase, and realizes the intelligent corresponding evaluation of image disturbance and power disturbance by means of a polynomial regression model, fundamentally improving the recognition accuracy of the system for real disturbances and effectively avoiding the risk of triggering unnecessary compensation due to misjudgment.

[0014] 2. The present invention uses a clustering analysis algorithm to identify hot spots of reactive power disturbances, and combines a dynamic change distribution map to achieve the visualization of evolution in the spatial and temporal dimensions, providing a clear decision-making basis for compensation strategies. The system outputs control suggestions based on disturbance intensity, duration, and priority, dynamically schedules static var compensators, and achieves fast, accurate, and hierarchical compensation responses. In addition, the system also supports an artificial review mechanism to ensure safety and operation flexibility, significantly improving the intelligent level and operation stability of power grid reactive power control as a whole. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention, and those of ordinary skill in the art can also obtain other drawings based on these drawings.

[0016] Figure 1 It is a system module mind map of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0018] For the embodiments, please refer to Figure 1 As shown, the real-time visualization tracking system for power grid reactive power dynamic compensation based on optical flow field analysis in this embodiment includes an image acquisition module, a power parameter acquisition module, an optical flow field calculation module, a comprehensive analysis module, a dynamic compensation control module, and a visualization interaction module; The image acquisition module is used to deploy camera devices in key areas of the power grid to obtain continuous monitoring image sequences; The power parameter acquisition module is used to collect real-time power grid operation parameters corresponding to the image area, including voltage, current, and reactive power; The optical flow field calculation module is used to perform optical flow analysis based on the monitoring image sequences and calculate the motion vector field of each pixel point; The comprehensive analysis module is used to perform spatio-temporal correspondence fusion of the optical flow vector field and the power operation parameters to identify the dynamic change trend and spatial distribution of reactive power in the power grid; A dynamic compensation control module, which is used to generate a compensation control instruction according to the comprehensive analysis result and control a static var compensator to implement real-time reactive power compensation for the target area; A visualization interaction module, which is used to display and interact with the power grid status, optical flow change and compensation control result in real time in the form of a graphical interface.

[0019] The image acquisition module is the front-end basic component for the system to realize optical flow field analysis and power grid status visualization, and is mainly responsible for acquiring and preliminarily processing continuous image data of key areas of the power grid.

[0020] Camera device deployment: Install high-definition industrial cameras at key nodes of power grid operation (such as main transformer areas, bus access points, and near reactive power compensation devices), and preferably select industrial-grade devices with wide dynamic range (WDR), infrared night vision function, and resistance to harsh environments (IP66 or above).

[0021] The deployment angle of the camera is optimized according to the layout of electrical equipment to ensure that the coverage area has no obvious occlusion and has a depth perspective effect, so as to more accurately capture the tiny movements of objects on the two-dimensional image plane.

[0022] Frame rate and resolution control: The image acquisition control unit uniformly manages the camera device, and preferably sets the image frame rate to ≥30fps to meet the requirements of the optical flow algorithm for time continuity; The image resolution is preferably 1080p or higher to enhance the recognition ability of equipment movement and vibration and improve the accuracy of optical flow calculation.

[0023] Image synchronization and timestamp management: The collected image frames are automatically marked with a unified timestamp and are clock-synchronized with the power data acquisition module (such as SCADA system or PMU system) to ensure that the image information and power parameters can be accurately matched during the analysis stage.

[0024] Image caching and frame loss protection: Configure an edge caching unit (such as an embedded memory or a circular buffer) to temporarily store continuous image sequences; a frame loss monitoring mechanism is provided to ensure image continuity and avoid analysis interruption caused by communication delay or frame loss.

[0025] Image preprocessing unit: Perform real-time denoising, sharpening, brightness equalization, and grayscale processing on the image to improve image quality and reduce the computational load of the optical flow algorithm; In scenes with drastic changes in illumination, the preprocessing unit has functions of adaptive exposure adjustment and gamma correction to enhance the dynamic adaptability of the image and reduce errors caused by illumination changes.

[0026] Interface and Data Output: The preprocessed image data is transmitted in real time to the backend optical flow analysis module through a high-speed communication interface (such as Gigabit Ethernet, USB 3.0, or wireless 5G / 6G); it supports the output of standard image formats (such as YUV, RGB, Gray), facilitating the calling and unified processing by algorithm modules.

[0027] Through the design of the above image acquisition module, this system can achieve high-quality and stable image acquisition in complex outdoor substation environments, meeting the stringent requirements of subsequent optical flow field calculations for image continuity, clarity, and time accuracy, and significantly enhancing the robustness and practicality of the overall system.

[0028] The power parameter acquisition module is the core component in this system to realize the digital perception of grid state variables and the fusion of image analysis information. Its main function is to obtain in real time the key parameters of grid operation corresponding to the image acquisition area, including but not limited to voltage, current, and reactive power, etc., to support subsequent dynamic compensation control decisions.

[0029] The data acquisition unit is installed at the power equipment access points corresponding to the camera monitoring area, such as the feeder outgoing end, the secondary side of the transformer, the capacitor bank, or the busbar end of the static var compensator; high-precision sensors are used to collect the real-time voltage (U) and current (I) waveforms of the three-phase AC system; the data sampling frequency is preferably ≥10kHz to meet the needs of grid dynamic change analysis and retain information characteristics such as harmonics and transients; it supports the simultaneous acquisition of derived parameters such as active power, reactive power, power factor, and frequency.

[0030] The Phasor Measurement Unit (PMU) supports GPS or IEEE 1588 time synchronization protocols to ensure the synchronization of power parameters and image frames, so as to achieve the time consistency of multi-source data; it has phasor measurement functions and can obtain the instantaneous voltage and current amplitudes and phase information of each phase; it marks the sampling data timestamps in real time and supports data collaboration with the SCADA system or the edge control unit.

[0031] Perform preprocessing operations such as filtering, denoising, normalization, and Short-Time Fourier Transform (STFT) on the original power data locally; Realize the preliminary identification of the dynamic trend of reactive power, such as short-period reactive power fluctuations, load mutations, or response delays of compensation devices, etc.; It can calculate event characteristics such as changes in regional power flow direction, reactive power mutation points, and frequency offsets, providing parameter support for subsequent optical flow fusion analysis.

[0032] Upload the real-time power parameters to the comprehensive analysis module through standard industrial communication protocols (such as Modbus-TCP, IEC 61850, DNP3.0); Supports data breakpoint resumption and exception retransmission mechanisms to ensure data integrity under edge network fluctuations; Works in coordination with the timestamp synchronization mechanism of the image acquisition module to ensure a one-to-one correspondence between images and power parameters, forming a spatio-temporal linked data basis.

[0033] Through the above-mentioned structured design of the power parameter acquisition module, this system can achieve high-precision, high-timeliness, and high-reliability multi-parameter power grid operation state perception, support the accurate positioning and rapid response to reactive power changes, and provide key support for the integrated control of power grid dynamic compensation and image intelligent tracking.

[0034] The optical flow field calculation module is used to perform pixel-level motion estimation on the collected continuous image sequence to generate an optical flow vector field reflecting the dynamic changes of the scene. This module is a key component for the system to achieve dynamic vision analysis and reactive power compensation tracking.

[0035] Receives continuous frame images (such as the t-th frame and the (t + 1)-th frame) from the image acquisition module and uniformly adjusts them to grayscale images or a unified resolution; Performs denoising, edge enhancement, and brightness normalization on the input images to improve the robustness of subsequent optical flow calculations; Dynamically switches enhancement strategies (such as gamma correction or adaptive histogram equalization) for different scene conditions (night, shadows, etc.).

[0036] The system dynamically selects dense or sparse optical flow algorithms according to computing resources and accuracy requirements; Dense optical flow (such as the Gunnar Farneback method): used to generate motion vectors for each pixel in the entire image area; Sparse optical flow (such as the Lucas-Kanade pyramid method): used for local motion estimation of selected feature points to improve computing efficiency; Supports deep learning-assisted optical flow (such as PWC-Net or RAFT) to improve robustness and accuracy in complex environments.

[0037] Performs pixel-level matching and motion estimation on two consecutive frames of images, and outputs the horizontal and vertical displacements (u, v) of each pixel; Generates a two-dimensional optical flow vector field, that is, each pixel corresponds to a motion vector representing its relative displacement from the previous frame to the current frame; Can further calculate the optical flow magnitude (|V|) and direction angle (θ) to form the visualization basis of the vector field.

[0038] Uses the velocity threshold method or the motion consistency method to eliminate abnormal optical flow vectors in the calculation results; Applies median filtering or Gaussian smoothing to improve the smoothness and continuity of the overall vector field; If there is occlusion or image perturbation (such as sudden illumination change), the system can trigger a redundant detection mechanism for inter-frame correction.

[0039] Divide the key equipment areas of the power grid into ROIs (Regions of Interest), and calculate the average motion intensity and main direction of each area; Combined with the spatial position mapping of the power parameter acquisition module, mark the areas where corresponding reactive power changes may occur; Extract typical features such as: local violently moving areas, areas with sudden changes in motion direction, for subsequent comprehensive judgment and compensation control decisions.

[0040] Encode the final optical flow vector field into formats such as color maps, arrow maps or heat maps; Output to the comprehensive analysis module and the visualization interface to provide graphical support for the evolution of the power grid state; Support structured data export, such as vector field matrices, regional motion feature tables, etc., for further use by algorithm modules.

[0041] Through the above step-by-step processing, the optical flow field calculation module can achieve high-precision perception of fine-grained dynamic changes in images, especially suitable for identifying small displacements, vibrations, abnormal fluctuations, etc. caused by changes in the power grid state, and providing intuitive and quantifiable visual input for reactive power dynamic compensation.

[0042] The comprehensive analysis module is used to perform spatio-temporal correspondence and fusion calculation on the optical flow vector field and power operation parameters, so as to accurately identify the dynamic change trend and spatial distribution area of reactive power in the power grid.

[0043] Based on the unified timestamp mechanism, align the power parameter data with the image frames at the corresponding time points in the time domain; Combined with the power grid GIS information and the camera deployment angle, establish the correspondence between the image pixel space and the physical space of electrical equipment; Map each monitoring area (ROI) in the image to the corresponding power grid node or device one by one to form an image-power grid area correspondence table.

[0044] Extract the motion features in the optical flow vector field, including the local amplitude mutation value and the power parameter changes in this area, including the sudden rise value of reactive power, for joint analysis; Among them, the method for obtaining the local amplitude mutation value is: perform optical flow calculation on consecutive image frames (the t-th frame and the t+1-th frame), obtain the horizontal displacement u(x,y) and vertical displacement v(x,y) of each pixel point, and calculate the optical flow amplitude: ; where: M(x,y) is the optical flow amplitude of the pixel point, that is, its motion intensity.

[0045] Divide the region of interest (ROI) in the image, such as a 10x10 pixel grid, and calculate the average amplitude of each region: ; Then calculate the local amplitude change between consecutive frames: ; Where: is the average amplitude of ROI region R at time t; ΔM(R) is the local amplitude mutation value; N is the number of pixel points in the region.

[0046] When ΔM(R) exceeds the set threshold Tm (empirical value or set based on historical statistics), it is regarded as the local amplitude mutation value. The larger the local amplitude mutation value, the more significant the dynamic change in this region.

[0047] Among them, the method for obtaining the sudden change value of reactive power is: Let the reactive power of the power node corresponding to a certain region at time t be , and the sampling period is Δt. The sudden rise of reactive power between two consecutive sampling periods is calculated as: ; Where: ΔQ is the sudden change value of reactive power; The larger the sudden change value of reactive power, the more likely a power disturbance event is considered to exist.

[0048] Convert the local amplitude mutation value and the sudden change value of reactive power into a comprehensive feature vector, and use the comprehensive feature vector as the input of the machine learning model. The machine learning model takes the evaluation value label corresponding to the predicted optical flow change and the actual power disturbance for each group of comprehensive feature vectors as the prediction target, and takes minimizing the sum of the prediction errors for all evaluation value labels corresponding to the optical flow change and the actual power disturbance as the training target to train the machine learning model until the sum of the prediction errors reaches convergence and then stops the model training. Determine the evaluation value corresponding to the optical flow change and the actual power disturbance according to the model output result. Among them, the machine learning model is a polynomial regression model.

[0049] Compare the evaluation value corresponding to the optical flow change and the actual power disturbance with a predetermined threshold. If the evaluation value corresponding to the optical flow change and the actual power disturbance is greater than the predetermined threshold, it is determined that the optical flow change is highly likely to be related to the actual power disturbance event and is judged as a region requiring compensation response; If the evaluation value corresponding to the optical flow change and the actual power disturbance is less than or equal to the predetermined threshold, it is considered that the change in this region is small or there is no obvious disturbance, and there is no need to immediately trigger compensation control.

[0050] The system is based on the following characteristics of each monitoring region in the current period: The average value of the optical flow amplitude, the sudden change value of reactive power, and the duration of the optical flow change; Construct its regional feature vector: Use the K-means clustering algorithm to cluster all regions and automatically divide several reactive power disturbance hot spots; Each type of area is sorted according to the characteristics of its center point and assigned different levels of compensation priority, such as level one (high), level two (medium), level three (low), etc.

[0051] The system integrates the disturbance identification results in each time period and generates a dynamic change distribution map; The horizontal axis is the time axis, and the vertical axis is the area number or coordinate; The color depth or symbol represents the intensity of optical flow disturbance or the degree of reactive mutation; Visualizing the evolution path of reactive power disturbances in space and time dimensions helps to observe the propagation trend and source of disturbances.

[0052] High-priority areas are given priority to matching compensation devices with fast dynamic response capabilities (such as STATCOM).

[0053] Evaluate the required reactive power compensation amount of each area based on its comprehensive characteristics (including optical flux intensity, reactive power mutation amplitude, and disturbance duration); Determine the following for each area: target compensation capacity (such as 300 kVar, 500 kVar, etc.); response time level (such as fast <1s, medium <5s); action duration (such as maintaining for more than 10 seconds or dynamically adjusting according to changing trends).

[0054] The control recommendation value is converted into a structured control criterion, which clearly includes: compensation target area number; required reactive power capacity; response level (such as "fast", "standard"); action priority; the control criterion is transmitted to the dynamic compensation execution module to trigger the physical device action.

[0055] According to the signal content, the matching SVC or STATCOM equipment is controlled to adjust the output: change the capacitor / inductor access ratio; adjust the inverter output current; dynamically adjust the reactive power injection into the grid to achieve voltage stability and reactive power balance; the response strategy is adjusted in real time according to the priority to ensure priority response in key areas.

[0056] After the compensation equipment is executed, the voltage and reactive power changes in the target area are monitored in real time. If the disturbance has not been eliminated, the system can provide continuous or additional compensation. If the disturbance has disappeared or the change has shifted, the response is terminated in time. All compensation response data are recorded and transmitted back to the system database for subsequent model training and strategy optimization.

[0057] If the system identifies a high-risk or repeatedly disturbed area, it will automatically send a warning message to the operation and maintenance monitoring platform; the operation and maintenance personnel can review the control criteria and manually intervene to determine whether to perform compensation actions; the system supports manual switching to "manual confirmation mode" to ensure the safety and controllability of the compensation response.

[0058] In this embodiment, through the collaborative work of modules such as image acquisition, power parameter perception, optical flow calculation, comprehensive analysis, intelligent discrimination, and dynamic control, accurate identification and rapid response to dynamic reactive power disturbances in the power grid are achieved. The system integrates the motion features in the image and the power operation parameters, uses machine learning and clustering algorithms to identify the compensation hot spots, generates compensation control strategies and drives the reactive power compensation equipment to adjust in real time, and at the same time provides a visual monitoring interface, effectively improving the intelligence, stability, and compensation efficiency of the power grid operation.

[0059] The above formulas are all dimensionless and take their numerical calculations. The formula is a formula obtained by software simulation through collecting a large amount of data to approximate the real situation. The preset parameters in the formula are set by those skilled in the art according to the actual situation.

[0060] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wired (such as infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that the computer can access or a data storage device such as a server or data center that contains a collection of one or more available media. The available medium can be a magnetic medium (such as a floppy disk, hard disk, magnetic tape), an optical medium (such as a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0061] The above is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in the present application, and all should be covered by the protection scope of the present application.

Claims

1. A real-time visualization tracking system for reactive power dynamic compensation of power grid based on optical flow field analysis, characterized by: It includes image acquisition module, power parameter acquisition module, optical flow field calculation module, comprehensive analysis module, dynamic compensation control module and visual interaction module; Image acquisition module, used to deploy camera equipment in key areas of the power grid to obtain continuous monitoring image sequences; The power parameter acquisition module is used to collect real-time power grid operation parameters corresponding to the image area, including voltage, current and reactive power; An optical flow field calculation module, used to perform optical flow analysis based on the monitoring image sequence and calculate the motion vector field of each pixel; A comprehensive analysis module, used for fusing the optical flow vector field with the power operation parameters in time and space, and identifying the dynamic change trend and spatial distribution of reactive power in the power grid; The dynamic compensation control module is used to generate compensation control instructions based on the comprehensive analysis results and control the static reactive power compensation equipment to implement real-time reactive power compensation for the target area; The visualization interaction module is used to display and interact with the grid status, optical flow changes and compensation control results in real time in the form of a graphical interface.

2. The real-time visualization tracking system for reactive power dynamic compensation of power grid based on optical flow field analysis according to claim 1 is characterized in that: The key areas include the main transformer area, busbar access point and reactive power compensation device.

3. The real-time visualization tracking system for reactive power dynamic compensation of power grid based on optical flow field analysis according to claim 1 is characterized in that: Receive continuous frame images from the image acquisition module and uniformly adjust them to grayscale images, preprocess the input images, dynamically switch enhancement strategies for different scene conditions, and dynamically select dense or sparse optical flow algorithms based on computing resources and accuracy requirements; Perform pixel-level matching and motion estimation on two consecutive frames of images, and output the horizontal and vertical displacement of each pixel; generate a two-dimensional optical flow vector field, that is, each pixel corresponds to a motion vector, indicating its relative displacement from the previous frame to the current frame.

4. The real-time visualization tracking system for power grid reactive power dynamic compensation based on optical flow field analysis according to claim 3 is characterized in that: Extract motion features in the optical flow vector field, including local amplitude mutation values, and regional power parameter changes, including reactive power surge values; The method for obtaining the local amplitude mutation value is as follows: perform optical flow calculation on continuous image frames, obtain the horizontal displacement u(x, y) and vertical displacement v(x, y) of each pixel, and calculate the optical flow amplitude: ; Where: M(x,y) is the optical flow amplitude of the pixel, that is, its motion intensity; divide the region of interest in the image and calculate the average amplitude of each region: ; Then calculate the local amplitude change between consecutive frames: ;in: is the average amplitude of the ROI region R at time t; ΔM(R) is the local amplitude mutation value; N is the number of pixels in the region; when ΔM(R) exceeds the set threshold Tm, it is regarded as a local amplitude mutation value.

5. The real-time visual tracking system for reactive power dynamic compensation of power grid based on optical flow field analysis according to claim 4 is characterized in that: in, The method for obtaining the reactive power surge value is as follows: Assume that the reactive power of the power node corresponding to a certain area at time t is , the sampling period is Δt, and the reactive power surge between two consecutive sampling periods is calculated as: ; Where: ΔQ is the reactive power surge value.

6. The real-time visualization tracking system for power grid reactive power dynamic compensation based on optical flow field analysis according to claim 5 is characterized in that: The local amplitude mutation value and reactive power surge value are converted into a comprehensive feature vector, and the comprehensive feature vector is used as the input of the machine learning model. The machine learning model takes the evaluation value label corresponding to the optical flow change and the actual power disturbance predicted by each group of comprehensive feature vectors as the prediction target, and takes minimizing the sum of prediction errors of all evaluation value labels corresponding to the optical flow changes and the actual power disturbance as the training target. The machine learning model is trained until the sum of prediction errors converges and the model training is stopped. The evaluation value corresponding to the optical flow change and the actual power disturbance is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

7. The real-time visual tracking system for power grid reactive power dynamic compensation based on optical flow field analysis according to claim 6 is characterized in that: The evaluation value corresponding to the optical flow change and the actual power disturbance is compared with the predetermined threshold. If the evaluation value corresponding to the optical flow change and the actual power disturbance is greater than the predetermined threshold, it is determined that the optical flow change is highly related to the actual power disturbance event and is determined to be an area requiring compensation response; if the evaluation value corresponding to the optical flow change and the actual power disturbance is less than or equal to the predetermined threshold, it is considered that there is no obvious disturbance in the area and there is no need to trigger compensation control immediately.

8. The real-time visualization tracking system for power grid reactive power dynamic compensation based on optical flow field analysis according to claim 7 is characterized in that: Based on the characteristics of each monitoring area in the current time period: the mean optical flow amplitude, the reactive power mutation value and the duration of the optical flow change; construct the regional feature vector: use the K-means clustering algorithm to cluster all areas and automatically divide them into several reactive disturbance hotspot areas; sort each type of area according to its central point characteristics and assign different levels of compensation priority. High-priority areas are preferentially matched with compensation equipment with fast dynamic response capabilities; integrate the disturbance identification results in each time period to generate a dynamic change distribution map; the horizontal axis is the time axis and the vertical axis is the area number or coordinate; the color depth or symbol represents the intensity of optical flow disturbance or the degree of reactive mutation; visualize the evolution path of reactive power disturbance in space and time dimensions.

Images