A power multi-dimensional situation data visualization method processing method

Abstract

The application is suitable for the field of power system monitoring, and provides a power multi-dimensional situation data visualization method processing method, which comprises the following steps: collecting multi-source data and generating visualization rendering data after processing; a main screen displays a core situation and responds to scene switching and click events; a side screen continuously monitors the main screen events, analyzes user intentions in combination with event content, real-time data context and special topic scene knowledge graph, dynamically decides detailed data to be displayed based on the intentions, requests corresponding visualization rendering data to be rendered, realizes dynamic self-adaption linkage of the main screen and the side screen based on context perception, and continuously updates the above steps. Therefore, the method solves the defects of the existing main screen and side screen linkage which depends on static rules and lacks context perception, realizes dynamic self-adaption cooperation based on user intentions and real-time data context, significantly improves the coherence and depth of information display, and effectively supports optimization of dispatching decision efficiency.

Description

Technical Field

[0001] This invention is applicable to the field of power system monitoring and provides a method for visualizing and processing multi-dimensional power situation data. Background Technology

[0002] Power multidimensional situational data visualization refers to the integration, rendering, and presentation of multidimensional data (such as generation, transmission, load, market, and risks) in power operations through graphical and interactive technologies to support dispatching, operation monitoring, analysis, and decision-making. With the advancement of new power system construction and the deepening of energy digital transformation, power data is becoming increasingly massive and complex in scale. Visualization technology has become a key means to improve the intelligence level and operational efficiency of the power grid. Its development prospects are broad; it can not only enhance the ability to perceive the real-time status of the power system but also support precise decision-making through intuitive displays, promoting the safe, economical, and efficient operation of the power grid. It is a core component of future intelligent dispatching and the energy internet.

[0003] Currently, multi-dimensional power situation data visualization typically relies on dispatch operation visualization systems, built upon existing data resources, hardware facilities, and visualization platforms, and employing a unified UI design style. For example, in dispatch halls such as the Southern Dispatch Center, the system often uses a mature 3D scene visualization rendering engine to display key operational indicators, such as load curves, power generation output, and market transactions, on the main screen, enhancing real-time insight capabilities through data visualization charts. Simultaneously, to broaden information dimensions, the system expands to side screens to supplement detailed data and customizes content around thematic scenarios such as "new power systems," "power markets," and "power supply." Regarding the linkage mechanism, existing methods use predefined business logic to associate the existing system on the main screen with the side screens. When the focus of the main screen switches, the side screens passively update relevant indicator content, attempting to achieve information synergy between the main and side screens to meet the needs of power monitoring.

[0004] However, existing methods for visualizing multi-dimensional power situational data still present significant challenges in implementation. Specifically, the main-side screen linkage mechanism relies on static business logic mapping. When the main screen focus dynamically switches between multiple thematic scenarios, the side screen content updates lack intelligent perception and adaptation to real-time data context and user operational intentions. This results in information displayed on the side screen often being disconnected from the main screen focus, exhibiting fragmented and weakly correlated characteristics. Decision-makers struggle to quickly obtain a coherent and in-depth analytical view, thereby reducing information flow efficiency and decision support effectiveness. Furthermore, there are currently technological gaps in addressing the dynamic adaptive linkage between thematic scenario context and real-time user intentions. Summary of the Invention

[0005] To address the aforementioned shortcomings, the present invention aims to provide a method for visualizing multidimensional power situation data, thereby resolving the problems mentioned in the background art. The method includes the following steps:

[0006] S1. Multi-source data acquisition and aggregation;

[0007] S2, hierarchical data storage;

[0008] S3, Multi-dimensional data fusion analysis;

[0009] S4. Real-time detection and alarm;

[0010] S5. Preparation of visualization rendering data;

[0011] S6, Home Screen Display:

[0012] S6.1 The main visualization display is based on a 3D scene visualization rendering engine. It obtains the required "visual rendering data" through a single-sided visualization data extraction unit and renders it to show the core situation of the default theme scene.

[0013] S6.2 When the dispatcher switches between thematic scenes through the main screen interface, the main display captures the switching event, the main screen requests the core data view of the new scene from the single-sided visualization data extraction unit, and re-renders it;

[0014] S6.3 When the dispatcher clicks on a device icon or area on the main screen, the main display captures the "user click event," which triggers a synchronized update of the visual secondary display.

[0015] S7, Side Screen Dynamic Adaptive Linkage:

[0016] S7.1 The secondary display continuously listens for user focus switching events and user click events from the primary display. Upon receiving an event, the secondary display starts the context parsing module, which combines the current event content, real-time data context, and a preset topic-specific knowledge graph to analyze the user's potential operational intent.

[0017] S7.2 Dynamic Content Decision-Making;

[0018] Based on the parsed user intent and scene context, the secondary display calls the intelligent decision engine to dynamically determine the detailed data content that needs to be displayed;

[0019] S7.3, Data Request and Rendering;

[0020] Based on the decision, the secondary display sends specific data requests to the single-sided visualization data extraction unit to obtain the corresponding "visualization rendering data". After the data is returned, the secondary display renders and displays the data according to the preset layout template, presenting a deep analysis view that is seamlessly connected with the focus of the main screen.

[0021] S7.4 The secondary display repeats steps S7.1 to S7.3 to achieve dynamic adaptive linkage.

[0022] Furthermore, the decision engine in step S7.2 comprehensively considers the following factors:

[0023] Intent matching: Based on the user's actions, analyze the action intent and select the corresponding data view template;

[0024] Contextualization: By combining real-time data context, such as whether the current device is in an alarm state or whether the load has exceeded the limit, the most relevant data dimensions are automatically filtered;

[0025] Multi-screen collaboration: If there are multiple secondary displays, the decision engine allocates display tasks based on the relevance of the content to ensure that the left and right screens display complementary information.

[0026] Furthermore, step S1 specifically includes:

[0027] S1.1 The multi-data node collector is deployed in power plants, substations, line towers, distribution stations and user-side nodes to collect various raw monitoring data in real time according to the preset frequency;

[0028] S1.2 The multi-node data aggregation processor receives raw data streams from each collector and logically groups them according to region or data type. At the same time, the processor uses the network time protocol to ensure time synchronization of each collector, eliminate clock deviation, and downsample high-frequency data.

[0029] Furthermore, step S2 specifically includes:

[0030] S2.1, Multiple data management repositories automatically distribute storage based on data type:

[0031] S2.2 The system automatically adjusts the storage strategy based on the popularity of the data and establishes a standardized data query interface for the analysis module and visualization unit to retrieve data as needed.

[0032] Furthermore, the specific method for automatically sorting and storing data according to data type in step S2.1 is as follows:

[0033] High-frequency time-series data are stored in a time-series database in the format of time series points. Each data point contains a timestamp, device identifier, measurement point code, and value.

[0034] Structured parameters are stored in a relational database as table rows;

[0035] Unstructured data is stored in a distributed file system as file objects, with the storage path associated with the device identifier.

[0036] Furthermore, step S3 involves deep analysis performed by the corresponding multi-data fusion analysis processor on each of the four business segments—generation, transmission, distribution, and utilization—to generate situational data; specifically, it includes the following sub-steps:

[0037] S3.1 The generator-side analysis processor combines historical data retrieval to construct an equipment health assessment model; through analysis, it outputs generator-side equipment assessment data including unit health status score, fault warning signal, and predicted power generation capacity.

[0038] S3.2 The transmission side analysis processor combines line current, conductor temperature, micro-meteorological data and image data to perform dynamic capacity expansion assessment and overheating risk analysis, and outputs the dynamic limit value of the line and the overheating risk level.

[0039] S3.3 The distribution side analysis processor locates fault sections through topology analysis, identifies three-phase imbalance, assesses equipment aging risks, and outputs information including the location of the fault section, power supply reliability indicators, and a list of high-risk equipment.

[0040] S3.4 The power consumption analysis processor uses a deep learning model to predict load, identify abnormal power consumption patterns, conduct energy efficiency benchmarking analysis, and output load prediction curves, abnormal power consumption alarms, and energy efficiency benchmarking results.

[0041] Furthermore, step S4 includes the following sub-steps:

[0042] S4.1 The real-time detector continuously monitors the data stream from each analysis processor, and the built-in rule engine performs real-time streaming computation on the data and generates real-time alarm data based on the results.

[0043] S4.2 The alarm system classifies alarms according to their severity and pushes them to the main and secondary visual displays.

[0044] Furthermore, step S5 specifically includes:

[0045] Based on the request content, the system obtains real-time situational data from the corresponding multi-data fusion analysis processor; the extracted data undergoes secondary aggregation and format encapsulation to generate visualization rendering data, which is encapsulated in JSON format and includes metadata such as numerical values, timestamps, units, and warning levels.

[0046] Furthermore, the steps also include:

[0047] S8. The system periodically extracts user interaction logs, side-screen display records, and alarm handling records from multiple data management repositories, and trains and optimizes the context parsing model and content decision engine through machine learning algorithms.

[0048] Therefore, the beneficial effects of the present invention are as follows:

[0049] 1. It solves the defects of rigid main screen and side screen linkage mechanism and lack of context awareness, and realizes dynamic adaptive collaboration based on user intent and real-time data context.

[0050] In existing technologies, the linkage between the main and side screens relies on static business logic mapping and predefined rules. When the main screen focus dynamically switches between multiple thematic scenarios, the content updates on the side screen cannot perceive the user's operational intent and the current real-time data context, resulting in a disconnect between the information displayed on the side screen and the main screen focus, exhibiting fragmented and weakly correlated characteristics. This invention, through steps S7.1 to S7.4, introduces a context parsing module and an intelligent decision engine into the visual secondary display: When the main screen focus switches or the user clicks, the secondary display first combines the event content, real-time data context (such as device alarm status, overload conditions), and a pre-defined thematic scenario knowledge graph to parse the user's potential operational intent (such as viewing details, fault tracing, resource scheduling, etc.); then, based on the parsed intent, it dynamically decides the detailed data content to be displayed; finally, it requests the corresponding "visual rendering data" from the single-side visual data extraction unit for rendering. This presents a deep analysis view that seamlessly connects with the main screen focus and has logical coherence, significantly improving the coherence of information display and the effectiveness of decision support.

[0051] 2. Through steps S1 to S3, a complete data link was established, from multi-data node collectors and multi-node data aggregation processors to multi-data fusion analysis processors: the power generation side analysis processor outputs "power generation side equipment evaluation data," the transmission side analysis processor outputs "transmission side dynamic evaluation data," the distribution side analysis processor outputs "distribution side fault location data," and the power consumption side analysis processor outputs "power consumption side load forecast data." Through cross-stage, multi-dimensional analysis results, standardized definitions and hierarchical storage, this data forms a knowledge base for dynamic decision-making on the side screen. When the secondary display performs intent parsing and content decision-making, it can accurately retrieve the required information based on a unified data naming system, ensuring that the content displayed on the side screen is not only relevant to the main screen's focus but also possesses depth and breadth across different stages.

[0052] 3. Through the real-time detection and alarm system, proactive perception and visual linkage of abnormal events are achieved, enhancing the emergency response capabilities of dispatch operations. Through the S4 step, the real-time detector continuously monitors the data flow of each analysis processor. When alarm rules such as voltage exceeding limits, abnormal vibration trends, or sudden load drops are triggered, structured "real-time alarm data" is immediately generated. When an alarm occurs, the main screen focus may automatically switch to the relevant device or area, while the secondary display parses the user's intent based on the alarm context, dynamically displaying in-depth views such as fault waveforms, associated device status, and historical comparative analysis, significantly improving the efficiency of dispatch personnel in emergency scenarios.

[0053] 4. Through a closed-loop optimization mechanism, the linkage model achieves continuous learning and self-evolution, ensuring that the system's intelligence continuously improves over time. Through the S8 steps, the system periodically extracts user interaction logs, side-screen display records, and alarm handling records from multiple data management repositories, and uses machine learning algorithms to train and optimize the context parsing model and content decision engine. Detailed Implementation

[0054] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0055] The multidimensional power situation data targeted by this invention refers to a comprehensive information set that comprehensively reflects the power grid's operating status, equipment health level, market transaction dynamics, and potential risks by collecting raw monitoring data from various business systems throughout the entire process of power system generation, transmission, distribution, and consumption, and then cleaning, integrating, and deeply analyzing the data.

[0056] For the power generation side data in the multidimensional power data, the data mainly comes from the distributed control systems, monitoring and data acquisition systems, and unit performance monitoring systems of various types of power plants.

[0057] For example, in the central control room of a thermal or hydropower plant, sensors within form a neural network covering the generating units. The collected data includes real-time operating parameters such as the active power, reactive power, and speed of the generator sets, as well as the vibration and sway of the turbine generator sets. For renewable energy plants, the focus is on collecting data such as wind speed and direction in wind farms, solar irradiance and ambient temperature in photovoltaic arrays, and DC-side voltage and current of inverters. On a visualization screen, the data from the generating side is presented as a three-dimensional digital twin model, modularly replicating the physical units on the screen and demonstrating the equipment status through modularly arranged symbols. Key operating parameters are updated in real-time as trend charts. Based on in-depth analysis of the raw data, the dynamic operating status of the units can be constructed, enabling status assessment and fault early warning of the generating equipment.

[0058] For the transmission side data in the multidimensional power data, the data mainly comes from online monitoring devices that are distributed throughout the transmission lines, drone inspection systems, and various measurement and control devices in substations.

[0059] The specific data collected can be divided into the following levels. The core data includes: three-phase current, voltage, and active / reactive power of the transmission line; conductor temperature, sag, and vibration; leakage current of insulators; and environmental data from micro-meteorological monitoring devices and video and image devices deployed on towers. In addition, it may include monitoring data such as infrared thermal images, photos, and videos collected by drones or robots.

[0060] On the visualization screen, power transmission data is typically presented comprehensively based on a GIS map. For example, the map dynamically displays the load rate of the lines and uses icons to mark real-time alarm information from various monitoring devices. Clicking on the icon of a node will bring up its real-time data information. Simultaneously, based on the fusion analysis of transmission-side data, multi-dimensional power grid safety assessments are achieved. For instance, combining line load rate, conductor temperature, and micro-meteorological data can assess dynamic capacity expansion capabilities or line overheating risks, preventing discharge accidents.

[0061] For the distribution side data in multidimensional power data, this includes distribution automation systems, distribution transformer terminals, user electricity information collection systems, and various low-voltage monitoring units.

[0062] The collected data includes three-phase voltage, current, active / reactive power, power factor, etc. of distribution transformers; switch status and fault indicator operation signals of transmission lines; ambient temperature and humidity, cable joint temperature, etc. in switch stations and ring main units.

[0063] On the visualization screen, distribution-side data is typically presented in the form of a power grid diagram, displaying the distribution network topology, real-time load of each node, and power supply range. Currently, digital twin technology is widely used on the distribution side to construct 3D models of substations or switching stations, intuitively displaying the internal structure of equipment, real-time operating conditions, and allowing for linked viewing of on-site video feeds. Deep analysis based on distribution network data is crucial for assessing power supply reliability. For example, when a tripping fault occurs, the system can automatically retrieve information from multiple systems, combine it with topology analysis, accurately locate the faulty section, and automatically generate an inspection report to push to repair personnel. Analysis of real-time operating data can also identify high-risk lines and equipment.

[0064] For electricity consumption-side data in multidimensional power data, the data mainly comes from electricity consumption information collection systems, load management systems, and various user-side intelligent monitoring terminals. Specifically, this includes: active / reactive power consumption and voltage / current curves in residential areas; production load and operating status of major electrical equipment in key enterprises; and various energy consumption components of electricity consumption in public institutions.

[0065] On the visualization screen, load density in different areas of the city is typically displayed visually through load heat maps; time-series curves show the load characteristics on the user side and compare them with historical or predicted values; for public institutions, their energy savings, energy saving rates, and energy efficiency benchmarking against similar institutions can be displayed. Simultaneously, the system integrates in-depth analysis of electricity consumption data to achieve load characteristic assessment. For example, a deep learning-based load forecasting model can predict load changes for the next few hours. When the system detects an abnormal deviation of a user's load curve from the predicted value, it can proactively issue an alert to assist the user in optimizing their energy consumption strategies.

[0066] Based on the above, the purpose of this invention is to provide a method for visualizing multidimensional power situation data. This method, based on a processing system, sequentially includes the following components according to the data flow from acquisition to presentation:

[0067] The system comprises a multi-data node acquisition unit, a multi-node data aggregation processor, a multi-data management and storage unit, a multi-data fusion and analysis processor, a real-time detector, an alarm system, a single-sided visualization data extraction unit, a main visualization display, and a secondary visualization display. These modules support the end-to-end processing and visualization of multi-dimensional power situation data.

[0068] Multi-data-node data acquisition units are deployed at various physical nodes in the power system, such as power plants, substations, transmission line towers, distribution stations, and user-side power consumption nodes. They are responsible for collecting raw monitoring data in real time from various business systems. Multi-data-node data acquisition units support multi-protocol adaptation, such as substation IEC 61850, remote control IEC 104, and industrial control Modbus, converting heterogeneous data into a unified standard format.

[0069] The multi-node data acquisition unit collects parameters such as active power, reactive power, speed, vibration, wind speed, and solar radiation from the generator units on the power generation side; it collects three-phase current and voltage, conductor temperature, sag, micro-meteorological data, and video images from the transmission side; it collects three-phase data of the distribution transformer, switch status, cable joint temperature, and ambient temperature and humidity from the distribution side; and it collects user electricity consumption, voltage and current curves, sub-item energy consumption, and production load data from the user side. The preliminarily processed data is then pushed to the multi-node data aggregation processor in real time.

[0070] The multi-node data aggregation processor, acting as a regional data aggregation unit, receives data streams from multiple data collectors, performs aggregation and verification. It logically groups the data according to region or data type, and utilizes a network time protocol to ensure time synchronization among the data collectors, eliminating clock skew. The processed data is simultaneously sent to a multi-data management repository for archiving and to a multi-data fusion analysis processor for real-time analysis.

[0071] The multi-data management repository adopts a hybrid storage architecture, consisting of a time-series database, a relational database, and a distributed file system. The time-series database stores high-frequency real-time data such as current, voltage, and power, supporting fast time-range queries. The relational database stores structured data such as equipment ledgers, user information, alarm records, and model parameters. The distributed file system stores unstructured data such as drone inspection photos, infrared thermal images, and surveillance videos. The system automatically manages the storage period of data based on its type and provides standardized data query interfaces for analysis modules and visualization units to retrieve data as needed.

[0072] The multi-data fusion analysis processor is configured with independent analysis processors for each of the four business segments: generation, transmission, distribution, and consumption. Each processor focuses on the analysis and situation generation of data on its respective side. The generation-side analysis processor receives unit operating parameters, constructs equipment health assessment models, predicts faults by analyzing vibration trends, integrates weather forecasts to assess the accuracy of new energy predictions, and outputs unit health status scores, fault warning signals, and power generation capacity predictions. The transmission-side analysis processor combines line current, conductor temperature, micro-meteorological data, and image data to conduct dynamic capacity expansion assessments and outputs dynamic line limit values. The distribution-side analysis processor uses transformer data, switch status, and cable temperature to locate fault sections through topology analysis, identify three-phase imbalances, assess equipment aging risks, and outputs fault location results, power supply reliability indicators, and a list of high-risk equipment. The consumption-side analysis processor uses deep learning models to perform load forecasting based on user power consumption curves and sub-item energy consumption, identifies abnormal power consumption patterns, conducts energy efficiency benchmarking analysis, and outputs load forecast curves and abnormal power consumption alarms. The results from each analysis processor are sent to the real-time detector and also to the single-side visualization data extraction unit for front-end access.

[0073] The real-time detector continuously monitors the data stream from each analysis processor, and the built-in rule engine performs real-time streaming computation on the data. When alarm rules are triggered by overvoltage, continuous increase in vibration trend, sudden load drop, or combined conditions such as line overload accompanied by severe weather, an alarm event is immediately generated, alarm suppression is performed to avoid repeated storms, and then the alarm event is sent to the alarm system.

[0074] After receiving an alarm event, the alarm system classifies the alarms according to their severity. Emergency alarms are presented on the main and secondary displays using pop-up windows and audio-visual prompts. The system records the entire process of alarm confirmation, processing, and cancellation, and finally writes the alarm records to a multi-data management repository for subsequent traceability and analysis.

[0075] The single-sided visualization data extraction unit, acting as the data adaptation layer of the visualization front-end, is responsible for responding to data requests from the main and secondary displays. It retrieves real-time situational data from the corresponding multi-data fusion analysis processor or queries historical data from the multi-data management repository based on the request content. The extracted data undergoes secondary aggregation and format encapsulation, converting it into a JSON format suitable for front-end chart rendering, containing information such as numerical values, timestamps, units, and alert levels. For frequently requested static data, the unit implements internal caching optimization to reduce the pressure of repeated queries.

[0076] The main visualization display is located in the center of the dispatch hall. Based on a 3D scene visualization rendering engine, it showcases the core status of power operations, including macroscopic information such as GIS maps, grid topology, power flow at key sections, and overall load curves. The main visualization display supports switching between multiple thematic scenes. Each time it switches, it requests the core data view of that theme from the single-side visualization data extraction unit for rendering. When a dispatcher clicks on a device or area on the main screen, the main display captures the interaction event and sends the focus information to the secondary visualization display in real time, triggering a synchronized update of the side screens.

[0077] The secondary visualization displays are located on either side or multiple sides of the main screen to display detailed data and in-depth analysis views related to the focus of the main screen. They listen for focus switching events from the main display and request corresponding detailed data from the single-side visualization data extraction unit based on the event content. This data may include a list of areas affected by the fault, distribution of emergency resources, interruptible load resource pools, relevant parameter curves, fault diagnosis details, and historical trend comparisons. The left and right side screens can display related information from different dimensions, forming complementary information and broadening the decision-maker's information horizons.

[0078] The data transmission and linkage of the entire hardware system forms a complete closed loop: the multi-data node collectors acquire raw data and upload it to the multi-node data aggregation processor; the aggregated data is stored in the multi-data management repository and sent to the multi-data fusion analysis processor; the analysis results are monitored by the real-time detector, and alarm events are distributed by the alarm system; the main visualization display acquires data through the single-sided visualization data extraction unit and transmits the focus event to the secondary display, which also acquires the relevant data through the extraction unit; finally, all modules work together to achieve a seamless flow from data acquisition, processing, analysis to visualization, providing comprehensive and intelligent situational awareness and decision support for power dispatching and operation.

[0079] Therefore, the raw data collected by the multi-data-node collector is defined according to its source as "raw monitoring data from the power generation side," "raw monitoring data from the transmission side," "raw monitoring data from the distribution side," and "raw monitoring data from the power consumption side." After protocol conversion and preliminary cleaning within the collector, the data is pushed to the multi-node data aggregation processor in real time in a standard format.

[0080] The multi-node data aggregation processor receives the raw data stream and, after aggregation, verification, and time alignment, generates "generation-side aggregated data," "transmission-side aggregated data," "distribution-side aggregated data," and "consumption-side aggregated data." The aggregated data is simultaneously sent to a multi-data management repository for archiving and to a corresponding multi-data fusion analysis processor for real-time analysis.

[0081] When data is sent to the repository, it is routed according to its type:

[0082] High-frequency time-series data, such as current, voltage, and power, are stored in a time-series database in the format of time series points. Each data point contains a timestamp, device identifier, measurement point code, and value.

[0083] Structured parameters, such as equipment ledgers and user information, are stored in a relational database in the form of table rows.

[0084] Unstructured data, such as drone inspection photos, infrared thermal images, and surveillance videos, are stored in a distributed file system as file objects, with the storage path associated with the device identifier.

[0085] The multi-data fusion analysis processor processes data separately for the four business segments, generating four types of analysis results.

[0086] The generator-side analysis processor outputs "generator-side equipment assessment data", including unit health status score, fault warning signal, and power generation capacity prediction value. Among them, trend data that needs to be tracked for a long time is stored in time series database as time series points, and warning records are stored in relational database.

[0087] The transmission side analysis processor outputs "transmission side dynamic evaluation data", which includes line dynamic limit values, overheating risk levels, etc., and is also stored in time series points and relationship tables respectively.

[0088] The distribution side analysis processor outputs "distribution side fault location data", which includes the location of the fault section, power supply reliability indicators, and a list of high-risk equipment. The fault events and indicators are stored in a relational database, and the real-time location information can be temporarily stored in memory for quick retrieval.

[0089] The "electricity-side load forecast data" output by the electricity-side analysis processor includes load forecast curves, abnormal electricity consumption alarms, and energy efficiency benchmarking results. The forecast curves are stored as time-series points, and the alarms and energy efficiency data are stored in a relational database.

[0090] After all analysis results are generated, they are sent to a real-time detector for monitoring in real time, and stored in a multi-data management repository for subsequent queries.

[0091] The real-time detector continuously monitors the data stream from various analytics processors. When an alarm rule is triggered, it generates "real-time alarm data," with each alarm event containing fields such as event ID, alarm level, occurrence time, device identifier, and alarm description. These alarm events are immediately sent to the alarm system and simultaneously written to the alarm record table in the relational database to form a closed-loop management system. The alarm system processes alarms according to their levels and pushes the alarms to a visual display.

[0092] The single-sided visualization data extraction unit responds to data requests from the main and secondary displays, extracting the required data from the multi-data fusion analysis processor or multi-data management repository. After secondary aggregation and format encapsulation, it generates "visualization rendering data." This "visualization rendering data" is encapsulated in JSON format and includes metadata such as values, timestamps, units, and alert levels, allowing for direct parsing by the front-end charts. For frequently requested static data, the extraction unit maintains a copy in a local cache to reduce the pressure of repeated queries.

[0093] User click events generated during the operation of the main visualization display include information such as the current focus scene identifier and the ID of the clicked device or area. These events are sent to the secondary visualization display in real time via an internal communication mechanism, triggering a synchronized update of the side screens. Based on the received events, the secondary display requests the corresponding "visualization rendering data" from the extraction unit for display.

[0094] Based on the above, this invention proposes a method for visualizing and processing multi-dimensional power situation data to achieve dynamic adaptive linkage between the main and side screens, solving the problems of rigid linkage mechanisms and lack of context awareness in existing technologies. The method specifically includes the following steps:

[0095] S1. Multi-source data acquisition and aggregation; including the following steps:

[0096] S1.1 Raw data acquisition;

[0097] Multi-node data acquisition units are deployed at power plants, substations, transmission line towers, distribution stations, and user-side nodes, collecting various types of raw monitoring data in real time at preset frequencies. Based on the source, the collected data are defined as "generator-side raw monitoring data," "transmission-side raw monitoring data," "distribution-side raw monitoring data," and "consumer-side raw monitoring data." The acquisition units support multi-protocol adaptation, uniformly converting heterogeneous data into a standard format.

[0098] S1.2 Data aggregation and alignment;

[0099] The multi-node data aggregation processor receives raw data streams from various data acquisition units, logically groups them according to region or data type, and generates "generation-side aggregation data," "transmission-side aggregation data," "distribution-side aggregation data," and "consumption-side aggregation data." The processor uses network time protocols to ensure time synchronization among the data acquisition units, eliminate clock skew, and downsamples high-frequency data to reduce transmission pressure in subsequent stages.

[0100] The aggregated data is simultaneously sent to a multi-data management repository and a corresponding multi-data fusion analysis processor.

[0101] S2. Tiered data storage; including the following steps:

[0102] S2.1, Data is classified and stored in the database;

[0103] The multi-data management repository employs a hybrid storage architecture, automatically allocating storage based on data type:

[0104] For high-frequency time-series data (such as current, voltage, power, etc.), they are stored in the time-series database in the format of time series points. Each data point contains a timestamp, device identifier, measurement point code and value.

[0105] Structured parameters (such as equipment ledgers and user information) are stored in a relational database in the form of table rows;

[0106] Unstructured data (such as drone inspection photos, infrared thermal images, and surveillance videos) are stored in a distributed file system as file objects, with the storage path associated with the device identifier.

[0107] S2.2 Data Lifecycle Management;

[0108] The system automatically adjusts its storage strategy based on data popularity and establishes a standardized data query interface for analysis modules and visualization units to retrieve data as needed. Specifically, data is categorized into real-time data, recent data, and historical data based on popularity.

[0109] S3, Multi-dimensional data fusion analysis;

[0110] For the four business segments of generation, transmission, distribution, and consumption, the corresponding multi-data fusion analysis processors perform in-depth analysis to generate situational data.

[0111] S3.1 Analysis of the power generation side;

[0112] The generator-side analysis processor receives "generator-side aggregated data," combines it with historical data, and constructs an equipment health assessment model. By analyzing unit vibration trends, temperature changes, and other factors, it predicts potential faults and integrates weather forecasts to assess the accuracy of new energy predictions. It outputs "generator-side equipment assessment data," including unit health status scores, fault warning signals, and predicted power generation capacity. Trend data requiring long-term tracking is stored in a time-series database, while warning records are stored in a relational database.

[0113] S3.2 Analysis of the transmission side;

[0114] The transmission-side analysis processor processes the "transmission-side aggregated data," combining line current, conductor temperature, micro-meteorological data, and image data to perform dynamic capacity expansion assessment and overheating risk analysis, outputting "transmission-side dynamic assessment data," which includes dynamic limit values ​​for lines, overheating risk levels, etc. The data is stored in time series points and relationship tables respectively.

[0115] S3.3 Analysis of the power distribution side;

[0116] The distribution-side analysis processor utilizes "distribution-side aggregated data" to locate faulty sections through topology analysis, identify three-phase imbalances, assess equipment aging risks, and output "distribution-side fault location data," including the location of the faulty section, power supply reliability indicators, and a list of high-risk equipment. Fault events and indicators are stored in a relational database, while real-time location information is temporarily stored in memory for quick retrieval.

[0117] S3.4 Analysis of the power consumption side;

[0118] The power consumption analysis processor, based on "power consumption aggregated data," uses a deep learning model to perform load forecasting, identify abnormal power consumption patterns, conduct energy efficiency benchmarking analysis, and output "power consumption load forecast data," including load forecast curves, abnormal power consumption alarms, and energy efficiency benchmarking results. The forecast curves are stored as time-series points, while alarms and energy efficiency data are stored in a relational database.

[0119] After all analysis results are generated, they are sent to a real-time detector for monitoring in real time, and stored in a multi-data management repository for subsequent queries.

[0120] S4. Real-time detection and alarm; specifically including the following steps:

[0121] S4.1 Alarm rule matching;

[0122] The real-time detector continuously monitors the data stream from various analysis processors, and the built-in rule engine performs real-time streaming computation on the data. When alarm rules such as voltage exceeding limits, continuous increase in vibration trend, sudden load drop, or combination of conditions are triggered, "real-time alarm data" is immediately generated. Each alarm includes fields such as event ID, alarm level, occurrence time, device identifier, and alarm description.

[0123] S4.2 Alarm Distribution;

[0124] The alarm system receives "real-time alarm data" and processes it according to its severity. Emergency alarms are pushed to the main and secondary visual displays, presented as pop-ups and audio-visual prompts; at the same time, alarm records are written to the alarm record table in the relational database, forming a closed-loop management system for subsequent traceability and analysis.

[0125] S5. Preparation of visualization rendering data;

[0126] The single-sided visualization data extraction unit, acting as the front-end data adaptation layer, is responsible for responding to data requests from the main and secondary displays. Based on the request content, the system retrieves real-time situational data from the corresponding multi-data fusion analysis processor or queries historical data from the multi-data management repository. The extracted data undergoes secondary aggregation and format encapsulation to generate "visualization rendering data," packaged in JSON format and containing metadata such as values, timestamps, units, and alert levels. For frequently requested static data, the unit implements internal caching optimization to reduce the pressure of repeated queries.

[0127] S6, Home screen display, includes the following steps:

[0128] S6.1, Main screen initialization rendering;

[0129] After the main visualization display is launched, it uses a 3D scene visualization rendering engine to show the core status of the default thematic scene, including macroscopic information such as GIS maps, power grid topology, power flow at key sections, and overall load curves. The main screen uses a single-sided visualization data extraction unit to obtain the required "visualization rendering data" for rendering.

[0130] S6.2, Thematic Scene Switching;

[0131] When the dispatcher switches between thematic scenes via the main screen interface, the main display captures the switching event and generates a "user focus switching event," which includes the identifier of the currently focused scene. The main screen then requests the core data view of the new scene from the single-sided visualization data extraction unit and re-renders it.

[0132] S6.3 User click interaction;

[0133] When a dispatcher clicks on a device icon or area on the main screen, the main display captures the "user click event," which includes information such as the clicked device ID or area ID and the current focus scene identifier. This event is sent to the secondary display in real time via an internal communication mechanism, triggering a synchronized update on the side screen.

[0134] S7, dynamic adaptive linkage of the side screen;

[0135] S7.1 Event Reception and Context Resolution;

[0136] The secondary display continuously listens for user focus switching events and user click events from the primary display. Upon receiving an event, the secondary display activates its context parsing module, combining the current event content, real-time data context, and a pre-defined thematic scenario knowledge graph to analyze the user's potential operational intent; for example: viewing details, comparative analysis, fault tracing, resource scheduling, etc.

[0137] S7.2 Dynamic Content Decision-Making;

[0138] Based on the parsed user intent and scene context, the secondary display invokes the intelligent decision engine to dynamically determine the detailed data content to be displayed. The decision engine considers the following factors:

[0139] Intent matching: Based on the user's actions, analyze the action intent and select the corresponding data view template.

[0140] Contextual association: Combine real-time data context, such as whether the current device is in an alarm state or whether the load exceeds the limit, to automatically filter the most relevant data dimensions (such as the waveform at the time of the alarm and the operating status of related devices).

[0141] Multi-screen collaboration: If there are multiple secondary displays, the decision engine allocates display tasks based on the relevance of the content to ensure that the left and right screens display complementary information.

[0142] S7.3, Data Request and Rendering;

[0143] Based on the decision, the secondary display sends specific data requests to the single-sided visualization data extraction unit to obtain the corresponding "visualization rendering data". After the data is returned, the secondary display renders and displays the data according to the preset layout template, presenting a depth analysis view that seamlessly connects with the focus of the main screen.

[0144] S7.4, Real-time updates;

[0145] When the main screen focus switches again or the user performs a new click operation, the secondary display repeats steps S7.1 to S7.3 to achieve dynamic adaptive linkage. Simultaneously, the secondary display allows users to manually adjust the displayed content; the system records these adjustments and uses them to optimize the subsequent intent recognition model.

[0146] S8, Closed-loop optimization process;

[0147] The system regularly extracts user interaction logs, side-screen display records, and alarm handling records from multiple data management repositories. Through machine learning algorithms, it trains and optimizes the context parsing model and content decision engine to continuously improve the intelligence and accuracy of the linkage mechanism, achieving continuous closed-loop optimization from data collection to visual decision-making.

[0148] Through the above steps, the present invention realizes dynamic adaptive linkage between the main screen and the side screen based on user intent and real-time data context, which solves the problems of disconnection between the content of the side screen and the focus of the main screen and information fragmentation in traditional methods, and significantly improves the information flow efficiency and decision support effect of the scheduling and operation visualization system.

[0149] Of course, the present invention may have other various embodiments. Without departing from the spirit and essence of the present invention, those skilled in the art can make various corresponding changes and modifications according to the present invention, but these corresponding changes and modifications should all fall within the protection scope of the appended claims.

Claims

1. A method for visualizing multidimensional power situation data, characterized in that, Includes the following steps: S1. Multi-source data acquisition and aggregation; S2, hierarchical data storage; S3, Multi-dimensional data fusion analysis; S4. Real-time detection and alarm; S5. Preparation of visualization rendering data; S6, Home Screen Display: S6.1 The main visualization display is based on a 3D scene visualization rendering engine. It obtains the required "visual rendering data" through a single-sided visualization data extraction unit and renders it to show the core situation of the default theme scene. S6.2 When the dispatcher switches between thematic scenes through the main screen interface, the main display captures the switching event, the main screen requests the core data view of the new scene from the single-sided visualization data extraction unit, and re-renders it; S6.3 When the dispatcher clicks on a device icon or area on the main screen, the main display captures the "user click event," which triggers a synchronized update of the visual secondary display. S7, Side Screen Dynamic Adaptive Linkage: S7.1 The secondary display continuously listens for user focus switching events and user click events from the primary display; Upon receiving an event, the secondary display activates the context parsing module, which combines the current event content, real-time data context, and a pre-defined thematic scenario knowledge graph to analyze the user's potential operational intent. S7.2 Dynamic Content Decision-Making; Based on the parsed user intent and scene context, the secondary display calls the intelligent decision engine to dynamically determine the detailed data content that needs to be displayed; S7.3, Data Request and Rendering; Based on the decision results, the secondary display sends specific data requests to the single-sided visualization data extraction unit to obtain the corresponding "visualization rendering data"; After the data is returned, the secondary display renders and displays the data according to the preset layout template, presenting a deep analysis view that seamlessly connects with the focus of the main screen; S7.4 The secondary display repeats steps S7.1 to S7.3 to achieve dynamic adaptive linkage.

2. The power multi-dimensional situation data visualization processing method according to claim 1, characterized in that, The decision engine in step S7.2 takes into account the following factors: Intent matching: Based on the user's actions, analyze the action intent and select the corresponding data view template; Contextualization: By combining real-time data context, such as whether the current device is in an alarm state or whether the load has exceeded the limit, the most relevant data dimensions are automatically filtered; Multi-screen collaboration: If there are multiple secondary displays, the decision engine allocates display tasks based on the relevance of the content to ensure that the left and right screens display complementary information.

3. The power multi-dimensional situation data visualization method according to claim 1, characterized in that, Step S1 specifically includes: S1.1 The multi-data node collector is deployed in power plants, substations, line towers, distribution stations and user-side nodes to collect various raw monitoring data in real time according to the preset frequency; S1.2 The multi-node data aggregation processor receives raw data streams from each collector and logically groups them according to region or data type. At the same time, the processor uses the network time protocol to ensure time synchronization of each collector, eliminate clock deviation, and downsample high-frequency data.

4. The power multi-dimensional situation data visualization method according to claim 1, characterized in that, Step S2 specifically includes: S2.1, Multiple data management repositories automatically distribute storage based on data type: S2.2 The system automatically adjusts the storage strategy based on the popularity of the data and establishes a standardized data query interface for the analysis module and visualization unit to retrieve data as needed.

5. The power multi-dimensional situation data visualization processing method according to claim 4, characterized in that, The specific method for automatic data splitting and storage in step S2.1 is as follows: High-frequency time-series data are stored in a time-series database in the format of time series points. Each data point contains a timestamp, device identifier, measurement point code, and value. Structured parameters are stored in a relational database as table rows; Unstructured data is stored in a distributed file system as file objects, with the storage path associated with the device identifier.

6. The power multi-dimensional situation data visualization method according to claim 1, characterized in that, Step S3 involves deep analysis of the four business segments—generation, transmission, distribution, and utilization—by the corresponding multi-data fusion analysis processor to generate situational data; specifically... It includes the following steps: S3.1 The generator-side analysis processor combines historical data retrieval to construct an equipment health assessment model; through analysis, it outputs generator-side equipment assessment data including unit health status score, fault warning signal, and predicted power generation capacity. S3.2 The transmission side analysis processor combines line current, conductor temperature, micro-meteorological data and image data to perform dynamic capacity expansion assessment and overheating risk analysis, and outputs the dynamic limit value of the line and the overheating risk level. S3.3 The distribution side analysis processor locates fault sections through topology analysis, identifies three-phase imbalance, assesses equipment aging risks, and outputs information including the location of the fault section, power supply reliability indicators, and a list of high-risk equipment. S3.4 The power consumption analysis processor uses a deep learning model to predict load, identify abnormal power consumption patterns, conduct energy efficiency benchmarking analysis, and output load prediction curves, abnormal power consumption alarms, and energy efficiency benchmarking results.

7. The power multi-dimensional situation data visualization method according to claim 1, characterized in that, Step S4 includes the following sub-steps: S4.1 The real-time detector continuously monitors the data stream from each analysis processor, and the built-in rule engine performs real-time streaming computation on the data and generates real-time alarm data based on the results. S4.2 The alarm system classifies alarms according to their severity and pushes them to the main and secondary visual displays.

8. The power multi-dimensional situation data visualization method according to claim 1, characterized in that, Step S5 specifically includes: Based on the request content, the system obtains real-time situational data from the corresponding multi-data fusion analysis processor; the extracted data undergoes secondary aggregation and format encapsulation to generate visualization rendering data, which is encapsulated in JSON format and includes metadata such as numerical values, timestamps, units, and warning levels.

9. The power multi-dimensional situation data visualization method according to claim 1, characterized in that, The steps also include: S8. The system periodically extracts user interaction logs, side-screen display records, and alarm handling records from multiple data management repositories, and trains and optimizes the context parsing model and content decision engine through machine learning algorithms.