A smart grid real-time risk panoramic visualization monitoring system and method

Abstract

The application discloses a kind of intelligent induction type power grid real-time risk panorama visualization monitoring system and method, belong to power grid data processing and visualization technical field.It includes: access second-level real-time operation data and generate risk quantification index;Based on the index self-adapting selection monitoring display mode and handle risk panorama distribution atlas;Extract multidimensional feature to establish dynamic guidance monitoring queue;According to order execution penetration type traceability analysis and generate deep drilling instruction;According to instruction trigger frequency iterative adjustment global monitoring sensitivity threshold value.The application adopts risk quantification driven mode switching, multidimensional feature fusion dynamic guidance sequencing and parameter self-optimization means based on analysis behavior feedback, can realize from macroscopic situation awareness to microcosmic risk traceability closed-loop intelligent analysis, significantly improve the perception efficiency and depth of power grid operation risk, effectively solve the problem of false alarm caused by alarm storm and parameter solidification in traditional monitoring, comprehensively guarantee the safe and stable operation of power grid.

Description

Technical Field

[0001] This invention relates to the field of power grid data processing and visualization technology, and in particular to a smart power grid real-time risk panoramic visualization monitoring system and method. Background Technology

[0002] The real-time risk panoramic visualization monitoring system for power grids is a core component of modern power dispatch and control centers. Through computer data processing technology, it processes, analyzes, and graphically displays massive amounts of operational data collected from various parts of the power grid, such as voltage, current, and power. It aims to provide dispatch and operation personnel with an intuitive and comprehensive view of the power grid's operational status and is a key information support platform for ensuring the safe and stable operation of the power grid.

[0003] In existing technologies, power grid monitoring systems typically employ data processing methods based on fixed thresholds. The system receives real-time data from systems such as Supervisory Control and Data Acquisition (SCADA) and compares it with pre-set safety limits in a database. Once the real-time data at a monitoring point exceeds its set upper or lower limits, the system generates an alarm message and notifies operators via a change in component color or a pop-up alarm window on a graphical user interface such as a main wiring diagram or geographic information map. The alarm list is usually simply arranged chronologically according to the events that occurred.

[0004] However, the aforementioned existing technical solutions have significant drawbacks in practical applications. First, alarm logic based on fixed thresholds is too rigid and cannot adapt to the dynamic changes in power grid operation. It is prone to generating a large number of irrelevant disturbance alarms under certain normal operating conditions, or missing alarms in certain potential risk scenarios. Second, when the power grid experiences disturbances, the system often generates a massive amount of alarm information without priority, forming an alarm storm, which greatly increases the difficulty for operators in information filtering and judgment. Finally, existing alarm information only presents the phenomena in a simple way, lacking in-depth analysis of the causes of risks and intelligent analysis of the correlations between different risk points, and cannot provide operators with effective decision support and analytical guidance. Summary of the Invention

[0005] To address the aforementioned issues, this invention provides an intelligent power grid real-time risk panoramic visualization monitoring system and method. It employs risk quantification-driven adaptive mode selection, multi-dimensional feature fusion-based dynamic guidance and sorting, and global parameter self-optimization based on analytical behavior feedback, enabling proactive guidance, thorough source tracing, and intelligent monitoring of power grid risks.

[0006] The above objectives can be achieved through the following approach:

[0007] A smart power grid real-time risk panoramic visualization monitoring system and method includes: accessing second-level real-time power grid operation data, and using a preset drilling operator to perform a preliminary risk assessment on the real-time power grid operation data at an initial drilling depth to generate risk quantification indicators; adaptively selecting a monitoring display mode to process the power grid risk distribution based on the risk quantification indicators to generate a risk panoramic distribution map; performing multi-dimensional feature extraction and topological association on each risk point in the risk panoramic distribution map to generate a dynamic guidance monitoring queue; performing penetrating source tracing analysis according to the order of the dynamic guidance monitoring queue based on an active guidance feedback mechanism, outputting real-time risk monitoring results and generating deep drilling instructions; and iteratively adjusting the global monitoring sensitivity threshold according to the trigger frequency of the deep drilling instructions to complete the iterative optimization of the system's visualization perception parameters.

[0008] Optionally, the generation of risk quantification indicators includes: dividing the power grid panoramic data, which is the real-time operation data set of the power grid, into county-level power grid regions in the spatial dimension and voltage level sub-bands in the hierarchical dimension; calculating the risk distribution variance in the spatial dimension and the load power fluctuation rate in the time dimension for each county-level power grid region; and generating risk quantification indicators by weighted fusion of the risk distribution variance and the load power fluctuation rate of all regions.

[0009] Optionally, the adaptive selection of a monitoring display mode includes: comparing the risk quantification index with a preset first system threshold; if the risk quantification index is better than the first system threshold, then selecting a standard panoramic monitoring mode to display the risk distribution globally in a regular manner; if the risk quantification index is worse than the first system threshold, then selecting a high-precision intelligent penetration monitoring mode to highlight and analyze the risk distribution locally.

[0010] Optionally, generating the dynamic guidance monitoring queue includes: extracting the risk assessment level, associated user priority weight, and real-time device load rate of each risk point in the risk panorama distribution map set as the multidimensional features; calculating the dynamic priority score of each risk point based on the multidimensional features; and sorting all risk points in descending order according to the dynamic priority scores to generate the dynamic guidance monitoring queue.

[0011] Optionally, the step of performing penetrating source tracing analysis based on the active guidance feedback mechanism includes: performing active guidance prompts on risk points according to the queue order to obtain preliminary source tracing conclusions and their corresponding identification confidence levels; comparing the identification confidence levels with preset second system thresholds and third system thresholds, wherein the second system threshold is higher than the third system threshold; if the identification confidence level is higher than the second system threshold, then adopting the preliminary source tracing conclusion as the final risk monitoring result; if the identification confidence level is between the third system threshold and the second system threshold, then generating a deep drilling command, wherein the deep drilling command is used to trigger an increase in the initial drilling depth of the local topology where the risk point is located, and trigger the input of additional multi-source heterogeneous data for re-association analysis, wherein the multi-source heterogeneous data includes the second-level real-time power grid operation data and preset risk knowledge data.

[0012] Optionally, the high-precision intelligent penetration monitoring mode includes: constructing a multi-level association path based on data drilling technology for the real-time operation data of the power grid to generate a dynamic risk association curve; generating an adaptive risk perception threshold based on the dynamic risk association curve; and comparing the real-time operation data of the power grid with the adaptive risk perception threshold to dynamically update the risk panoramic distribution map.

[0013] Optionally, the construction of multi-level correlation paths includes: using a first drill operator in the drill operator that serves as the data drill technology execution tool to perform hierarchical processing on the power grid topology corresponding to the voltage level sub-band, generating a main grid topology envelope and a distribution network topology envelope; performing a weighted average calculation on the main grid topology envelope and the distribution network topology envelope to obtain a hierarchical average envelope; and using a second drill operator in the drill operator to perform a topology opening operation on the hierarchical average envelope to generate a dynamic risk correlation curve from the 110kV voltage level to the power user end.

[0014] Optionally, the method further includes the process of generating preliminary classification results of causes: calculating the power grid impact range and abnormal amplitude change rate of each risk point in the risk panoramic distribution map set to obtain the spatiotemporal characteristics of the risk; comparing the spatiotemporal characteristics of the risk with the risk type feature template preset by the system to generate preliminary classification results of causes including faults, heavy overloads, load mutations or mode adjustments.

[0015] Optionally, the iterative adjustment of the global monitoring sensitivity threshold includes: counting the trigger frequency of the deep drilling command within a preset monitoring window; comparing the trigger frequency with a preset frequency threshold; if the trigger frequency is consistently higher than the frequency threshold, reducing the value of the first system threshold to increase the system's tendency to actively enter a high-precision intelligent penetration monitoring mode.

[0016] Based on the same inventive concept, this invention also provides a smart guidance-type real-time risk panoramic visualization monitoring system for power grids. The system includes: a real-time perception and assessment module, used to access second-level real-time power grid operation data and perform preliminary risk assessment using a preset drilling operator to generate risk quantification indicators; a smart guidance-type linkage analysis module, used to adaptively select a monitoring display mode to process the power grid risk distribution based on the risk quantification indicators, generating a risk panoramic distribution map; a risk feature and ranking module, used to perform multi-dimensional feature extraction and topological association on each risk point in the risk panoramic distribution map, generating a dynamic guidance monitoring queue; an active guidance and penetration analysis module, used to perform penetrating source tracing analysis according to the order of the dynamic guidance monitoring queue based on an active guidance feedback mechanism, outputting real-time risk monitoring results and generating deep drilling commands; and a global perception optimization module, used to iteratively adjust the global monitoring sensitivity threshold based on the trigger frequency of the deep drilling commands, completing iterative optimization of system visualization perception parameters, including the risk quantification indicators.

[0017] Compared with the prior art, the present invention has the following advantages:

[0018] 1. This invention fundamentally revolutionizes the monitoring of power grid operation risks by constructing a closed-loop intelligent analysis process from macro-level assessment to micro-level tracing. The system adaptively switches monitoring modes based on global risk quantification indicators, intelligently allocating computing resources and personnel attention. It provides a concise global view when the power grid is operating smoothly, and automatically enters a high-precision penetrating analysis mode when the risk increases, effectively solving the deficiency of traditional monitoring systems in balancing information presentation and analytical depth.

[0019] 2. This invention proposes a dynamic guided risk analysis mechanism, which overturns the passive and scattered alarm presentation method of traditional monitoring systems. By extracting multi-dimensional features and prioritizing risk points, the system integrates disordered risk information into a clear and orderly dynamic guided monitoring queue, proactively guiding maintenance personnel or automated systems to conduct in-depth analysis step by step according to the actual severity of the risk, greatly improving the efficiency of emergency response and the accuracy of decision-making in complex risk scenarios.

[0020] 3. This invention introduces an adaptive optimization closed loop for globally perceived parameters, endowing the monitoring system with the ability to learn and evolve on its own. By continuously statistically analyzing the trigger frequency of deep drilling commands, the system can reflect on and evaluate the suitability of its monitoring sensitivity, and iteratively adjust key system thresholds accordingly. This self-optimization mechanism ensures that the system can adapt to changes in power grid operating characteristics over the long term, avoiding system performance degradation caused by parameter fixation, and guaranteeing the continuous effectiveness and robustness of monitoring.

[0021] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures pointed out in the description, claims and drawings. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a flowchart illustrating a method for real-time panoramic visualization monitoring of power grid risks according to an embodiment of the present invention.

[0024] Figure 2 This is a schematic diagram of the structure of a smart grid real-time risk panoramic visualization monitoring system according to an embodiment of the present invention.

[0025] Figure 3 This is a cumulative contribution diagram of each county-level power grid area to the overall risk quantification index Q, according to an embodiment of the present invention.

[0026] Figure 4 This is an embodiment of the invention based on risk level. With load rate Dynamic priority isosurface landscape map of dimensions. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] Reference Figure 1 One embodiment of the present invention proposes a smart-guided real-time risk panoramic visualization monitoring method for power grids. It adopts adaptive switching of monitoring and display modes driven by risk quantification indicators, dynamic guidance and sorting based on multi-dimensional feature fusion, and global parameter iterative optimization based on the trigger frequency of deep drilling commands. It can realize closed-loop intelligent analysis of the entire process from macro-situation perception to micro-risk penetration and tracing, and improve the perception efficiency, analysis depth and adaptive evolution capability of the system to complex operating conditions of power grid operation risks.

[0029] The method described in this embodiment specifically includes:

[0030] By introducing drill-down operators to quickly reduce the dimensionality and conduct preliminary assessments of massive amounts of real-time data at the second level, the complex power grid operation status is abstracted into a unified risk quantification index.

[0031] Based on this metric, an adaptive decision-making mechanism is activated to dynamically switch between two monitoring modes: global overview and detailed investigation.

[0032] After identifying specific risk points, the system does not present them in a disordered manner, but integrates multi-dimensional features such as risk level, topological association, and degree of impact to generate a dynamic priority queue, which serves as a roadmap for proactively guiding the analysis.

[0033] Along this queue, the system initiates penetrating tracing one by one and introduces a feedback mechanism to judge the depth of analysis. For difficult and risky cases, it automatically triggers deeper data drilling.

[0034] By monitoring the activation frequency of deep drilling commands, the system can reflect on its own monitoring sensitivity and make iterative self-adjustments, forming a monitoring logic that can adaptively evolve with changes in power grid conditions.

[0035] Optionally, in generating the risk quantification index, the power grid panoramic data, which serves as the real-time operation data set of the power grid, is divided into county-level power grid regions in the spatial dimension and voltage level sub-bands in the hierarchical dimension; for each county-level power grid region, the risk distribution variance in the spatial dimension and the load power fluctuation rate in the time dimension are calculated; the risk quantification index is generated by weighted fusion of the risk distribution variance and the load power fluctuation rate of all regions.

[0036] Specifically, the engineering objective of this step is to accurately condense the massive and multi-dimensional real-time operational status data from the power grid's panoramic data into a single numerical value that can macroscopically represent the overall risk level of the power grid—that is, a risk quantification index. The generation process of this index serves as the basis for subsequent adaptive selection of monitoring modes. First, the input data undergoes structured preprocessing. Then, risk characteristics representing the spatial and temporal dimensions are calculated. Finally, the risk quantification index is calculated using a normalization and weighted fusion algorithm.

[0037] The first step is the gridding of the power grid panoramic data. The system receives and processes the power grid panoramic data, which is a collection of real-time power grid operation data. This data typically originates from the SCADA (Supervisory Control and Data Acquisition) system and the EMS (Energy Management System), with a data refresh cycle on the order of seconds. Based on preset geographical information and topology rules, the system deconstructs the power grid panoramic data in two dimensions. Spatially, the vast power grid topology is divided into several independent analysis units, i.e., county-level power grid regions, according to administrative divisions. Hierarchically, based on the voltage standards specified in the equipment asset ledger, the data within each county-level power grid region is further divided into different voltage level sub-bands, such as the 500kV and above main grid layer, the 110-220kV transmission layer, and the 10-35kV distribution layer. This operation transforms continuous power grid data into discrete, independently analyzable regional and hierarchical data blocks.

[0038] The second step is to calculate the risk characteristics by region. For each county-level power grid region, the system calculates two core characteristic values ​​in parallel: the spatial dimension of risk distribution variance and the temporal dimension of load power volatility.

[0039] Risk distribution variance measures the evenness of risk distribution within a region. A high variance value indicates that risk is highly concentrated in a few critical pieces of equipment, easily leading to single-point-of-failure risks, while a low variance value indicates that risk is distributed more evenly. The calculation process involves first obtaining data within the region... Preliminary risk values ​​for key equipment or nodes This risk value is generated from the upstream preliminary risk assessment step. The variance is then calculated using the following formula. :

[0040]

[0041] In this formula, This represents the variance of the risk distribution in the power grid area of ​​that county. It is the first in the region Real-time risk value of each node. It is the arithmetic mean of the risk values ​​of all nodes in the region. It represents the total number of nodes within the region.

[0042] Load power fluctuation rate is used to measure the stability of electricity load within a region. Drastic load fluctuations may indicate potential equipment shocks or supply-demand imbalances. Extracting a specified time window... Within, for example, the total active power of the county's power grid area over the past 15 minutes. The time series data is used to calculate its volatility using a formula. :

[0043]

[0044] In this formula, Represents load power fluctuation rate. It is the standard deviation of the total active power sequence within the time window $T$. This is the average value of the sequence. By dividing by the average value, the result becomes a dimensionless relative fluctuation index, which facilitates horizontal comparisons between regions with different load bases.

[0045] The third step is the weighted fusion of multidimensional features. Since the risk distribution variance V and the load power volatility F have different physical meanings and numerical scales, the system first normalizes these two indicators, mapping them to a unified range of 0-1 to eliminate dimensional differences. The min-max normalization method is used, as shown in the formula:

[0046]

[0047] In this formula, These are normalized values. These are the original values. and These are the maximum and minimum values ​​of the indicator obtained by the system over a relatively long historical period, such as the past 24 hours.

[0048] After normalization, the system normalized variance for all county-level power grid areas. and normalized volatility By performing a weighted summation, we obtain the final global risk quantification index Q, as shown in the formula:

[0049]

[0050] In this formula, This refers to risk quantification indicators. These represent different county-level power grid regions. and These are the weighting coefficients for spatial risk distribution and temporal load fluctuation, respectively, and their sum is 1. For example, based on scheduling experience, they can be set as follows: =0.6, A value of 0.4 indicates that the current strategy focuses more on the spatial concentration of risk. The higher the value of this indicator Q, the greater the overall risk faced by the power grid, and the higher the uncertainty and instability of the system.

[0051] Optionally, the adaptive selection of a monitoring display mode includes: comparing the risk quantification index with a preset first system threshold; if the risk quantification index is better than the first system threshold, then selecting a standard panoramic monitoring mode to display the risk distribution globally in a regular manner; if the risk quantification index is worse than the first system threshold, then selecting a high-precision intelligent penetration monitoring mode to highlight and analyze the risk distribution locally.

[0052] Specifically, the engineering objective of this step is to establish an adaptive monitoring and decision-making mechanism based on the overall risk level of the power grid. This mechanism enables the system's visualization interface and analytical depth to dynamically match the real-time operating status of the power grid, thereby achieving a rational allocation of computing resources and dispatcher attention. The core action of this mechanism is to perform threshold judgments on the risk quantification indicators generated in the previous step and switch the system's operating mode based on the judgment results.

[0053] First, obtain the global risk quantification index calculated by the upstream module. This indicator is a dimensionless value that comprehensively reflects the spatial distribution balance of power grid risks and the temporal load stability. The system calculates this in real time. The value is compared with a key parameter of an internal configuration, namely the first system threshold Th1.

[0054] The first system threshold, Th1, is a pre-defined critical value used to distinguish between a "stable" and "alarm" power grid state. This threshold is set based on historical power grid operational data statistical analysis, dispatching procedures, and expert experience, and is typically a value between 0 and 1, for example, 0.75. It represents the upper limit of power grid risk acceptable to operations and maintenance managers. This threshold is configurable, allowing operations and maintenance personnel to adjust it according to seasonal load characteristics or special power grid operating modes.

[0055] The comparison logic executed by the system is as follows, where This represents the monitoring display mode selected by the system:

[0056]

[0057]

[0058] When the judgment result is a risk quantification indicator When it is better than the first system threshold Th1, that is A value less than Th1 indicates that the overall power grid operation is stable and the risks are controllable. The system will automatically select and enter the standard panoramic monitoring mode. In this mode, the system interface will provide a global, routine display of risk distribution, such as on a GIS map or electrical main wiring diagram, using a standardized color spectrum, such as green for low risk, yellow for concern, and orange for warning, to render the risk status of each sub-county power grid area and key equipment. Its core objective is to provide macro-level situational awareness, maintaining the data refresh rate and analysis depth at baseline levels to avoid information overload.

[0059] Conversely, when the risk quantification index Q is worse than the first system threshold Th1 (i.e., Q is greater than or equal to Th1), it indicates that the overall risk level of the power grid exceeds the warning line, potentially indicating severe risks in local areas or a global instability trend. The system will automatically switch to a high-precision intelligent penetration monitoring mode. This mode is an enhanced analysis state of the system, and its actions involve two levels. First, it executes local highlighting alerts, visually enhancing key county-level power grid areas or voltage level sub-bands that cause the Q index to exceed the limit on the risk panoramic distribution map. This can be achieved through methods such as using high-saturation red, dynamic flashing, or automatic zooming of local views to forcibly attract the attention of dispatchers. Second, the system automatically initiates a key penetration analysis process in the background, preloading more detailed topological data and multi-dimensional feature data for these high-risk areas. This prepares the data for subsequent penetrating source tracing analysis, thereby ensuring a rapid and accurate response to potential risks.

[0060] Optionally, generating a dynamic guidance monitoring queue includes: extracting the risk assessment level, associated user priority weight, and real-time device load rate of each risk point in the risk panorama distribution map set as the multidimensional features; calculating the dynamic priority score of each risk point based on the multidimensional features; and sorting all risk points in descending order according to the dynamic priority scores to generate the dynamic guidance monitoring queue.

[0061] Specifically, the engineering objective of this step is to transform the multiple, scattered risk points presented in the risk panorama map into an ordered, executable analysis sequence, namely, a dynamic guidance monitoring queue. The generation of this queue aims to ensure that the analytical resources of system and maintenance personnel can prioritize focusing on the risk points that currently pose the greatest threat to power grid security and have the widest impact, thus achieving a shift from "passive response" to "proactive guidance."

[0062] The first step is multi-dimensional feature extraction and topological association. The system traverses every identified risk point in the risk panorama distribution map set, such as a heavily loaded transformer or a suspected faulty line. For each risk point... Parallel data extraction and correlation analysis are performed to obtain core features across three dimensions:

[0063] Risk assessment level This data comes directly from the output of the preliminary risk assessment module and is usually quantified into a discrete level of 1 to 5, with higher values ​​indicating more severe risks.

[0064] Associated User Priority Weight The system performs a rapid topology analysis, tracing all affected power users along the downstream power supply network from the risk point, and retrieving the priority of these users from the user profile database. For example, it identifies the weights of important users such as hospitals and government agencies. The weight of ordinary residents will be set to a high value close to 1.0, while the weight of ordinary residents will be lower, between 0.1 and 0.3. The system will take the highest priority weight among all affected users as the risk point. value.

[0065] 3. Real-time device load rate The system directly retrieves the real-time active power of the equipment corresponding to the risk point from the SCADA database and divides it by its rated capacity to obtain a percentage value.

[0066] The second step is the calculation of dynamic priority scores. To integrate the three features with different physical meanings and dimensions into a unified ranking criterion, the system first normalizes each feature. Risk assessment level L is mapped to the 0-1 range, and real-time device load rate... It was also normalized, and when When it exceeds 100%, its normalized value increases rapidly and non-linearly to amplify the impact of overload risk. Related user priority weights. The weights themselves are in the 0-1 range and require no additional processing. Subsequently, the system calculates the dynamic priority score for each risk point i using a weighted summation method. The calculation formula is as follows:

[0067]

[0068] In this formula, It is the dynamic priority score of the i-th risk point; It is the normalized risk assessment level; It relates to user priority weights; It is the normalized real-time device load rate; , , These are the weighting coefficients of three characteristics, preset by the system operation and maintenance strategy, and the sum of the three is 1. For example, during peak summer seasons, the load rate weight can be... The priority value can be increased to 0.5, and when it is necessary to ensure power supply to important users, the user priority weight can be increased. This demonstrates the flexibility of the monitoring strategy.

[0069] The third step is to generate the monitoring queue. This involves dynamically prioritizing all risk points. The data is collected and sorted globally in descending order. The resulting list of risk points constitutes the final dynamic guidance and monitoring queue. The first point in this queue represents the highest overall risk, while the last point represents the lowest risk. This queue, as an ordered set of instructions, is directly input into the subsequent active guidance and penetration analysis modules, determining the order in which the system performs penetration-based source tracing analysis.

[0070] Optionally, performing a penetrating source tracing analysis based on an active guidance feedback mechanism includes: actively guiding and prompting risk points according to the queue order to obtain preliminary source tracing conclusions and their corresponding identification confidence levels; comparing the identification confidence levels with preset second system thresholds and third system thresholds, wherein the second system threshold is higher than the third system threshold; if the identification confidence level is higher than the second system threshold, then adopting the preliminary source tracing conclusion as the final risk monitoring result; if the identification confidence level is between the third system threshold and the second system threshold, then generating a deep drilling command, wherein the deep drilling command is used to trigger an increase in the initial drilling depth of the local topology where the risk point is located, and trigger the input of additional multi-source heterogeneous data for re-association analysis, wherein the multi-source heterogeneous data includes the second-level real-time power grid operation data and preset risk knowledge data.

[0071] Specifically, the engineering objective of this step is to construct a closed-loop, intelligent risk tracing and analysis process. This process, through proactive guidance and confidence assessment, dynamically determines the depth and breadth of the analysis, aiming to quickly and accurately identify high-priority risks while initiating deeper analysis of ambiguous risk signals, thereby balancing analytical efficiency and accuracy.

[0072] The first step involves sequentially executing proactive guidance and preliminary source tracing. The system strictly follows the order of the dynamically guided monitoring queue generated upstream, processing risk points one by one. For the first risk point in the queue, the system executes a proactive guidance prompt, meaning that the system automatically invokes a built-in risk diagnostic model, such as a classifier based on a decision tree or lightweight neural network, to quickly analyze the relevant real-time data for that risk point. The output of this process includes two parts: first, a preliminary source tracing conclusion, which is a text description, such as "A phase of the main transformer in a 110kV substation is overloaded"; second, the identification confidence level C, a value between 0 and 1, representing the degree of certainty of the diagnostic model regarding this conclusion.

[0073] The second step is to establish a three-segment confidence decision-making mechanism. The system compares the acquired identification confidence level C with two preset key system parameters: the second system threshold Th2 and the third system threshold Th3. These two thresholds define three decision intervals: a high confidence interval, a medium confidence interval, and a low confidence interval. Based on operational experience, Th2 is usually set to a relatively high value, such as 0.9, representing a highly reliable conclusion; Th3 is set to a medium value, such as 0.6, representing a conclusion with some reference value but not yet clear. These two thresholds can be configured by authorized engineers in the system, and the constraint that Th2 must be higher than Th3 must be met.

[0074] The third step is the direct adoption of high-confidence conclusions. If the identification confidence level C is higher than the second system threshold Th2 (i.e., C>Th2), this indicates that the preliminary source tracing conclusion has extremely high reliability. The system determines that the analysis has converged, directly adopts this preliminary source tracing conclusion as the final risk monitoring result, and pushes it to the main monitoring interface for alarm display. Subsequently, the system automatically marks the risk point as "processed" and immediately begins processing the next risk point in the dynamic guidance monitoring queue to ensure efficient flow of the analysis process.

[0075] The fourth step is to trigger deep drilling for medium-confidence risks. If the identification confidence level C is between the third system threshold and the second system threshold, i.e., Th3≤C≤Th2, this indicates that the preliminary analysis has found some clues, but the chain of evidence is not yet complete, and a definitive conclusion cannot be formed. At this time, the system determines that a more in-depth analysis is needed and automatically generates a deep drilling instruction. This instruction is an internal control signal, and its core function is to trigger two linked operations. First, it increases the drilling depth. The system will call the preset drilling operator to increase the initial drilling depth of the local power grid topology where the risk point is located, for example, from the substation level to the feeder switch level. Second, it expands the analysis data source. The system triggers the input of additional multi-source heterogeneous data. This includes retaining the original second-level real-time power grid operation data and additionally loading preset risk knowledge data, such as the equipment's historical fault reports, maintenance records, and fault mode libraries of similar equipment, etc., which are semi-structured or unstructured data. The system will utilize these enhanced datasets to conduct a more resource-intensive but comprehensive re-association analysis on the risk point, aiming to obtain a conclusion with higher confidence. For cases below Th3, the system will mark them as "low confidence" and temporarily suspend them, awaiting more data or manual intervention.

[0076] Optionally, the high-precision intelligent penetration monitoring mode includes: constructing a multi-level association path based on data drilling technology for the real-time operation data of the power grid to generate a dynamic risk association curve; generating an adaptive risk perception threshold based on the dynamic risk association curve; and comparing the real-time operation data of the power grid with the adaptive risk perception threshold to dynamically update the risk panoramic distribution map.

[0077] Specifically, the engineering objective of this step is to establish a dynamic, condition-adaptive safety operation envelope—an adaptive risk perception threshold—for key power grid sections or paths under high-risk early warning conditions, and to use this as a benchmark for refined risk identification. This model abandons the traditional fixed threshold alarm method and instead adopts a dynamic margin monitoring method that is more in line with the physical laws of the power grid, thereby enabling earlier and more accurate detection of potential stability problems.

[0078] The first step is the construction of multi-level interconnected paths. When the system switches to the high-precision intelligent penetration monitoring mode, its analysis engine activates data drill-down technology. The core of this technology is to perform data correlation and model deduction from top to bottom along the electrical topology of the power grid. The system automatically identifies key power grid areas that cause global risk indicators to exceed limits, and selects one or more high-voltage power sources within these areas as the starting point for path construction, such as the outgoing line of a 220kV substation. Then, the system drills down the topology path level by level, traversing the transmission network, substations, and distribution network, until finally reaching the power user, forming a complete power supply path.

[0079] The second step is the generation of dynamic risk correlation curves. For each multi-level correlation path constructed in the previous step, a dynamic risk correlation curve is calculated in real time. This curve is essentially a predictive model. Based on the real-time injected power and voltage at the path source, combined with the impedance parameter models of all series-connected devices along the path, such as lines and transformers, it calculates the theoretical electrical state values ​​of each key node along the path in real time. The most commonly used indicator is voltage amplitude. This curve dynamically reflects the normal attenuation law of voltage along the power supply path under the current load level. For example, the system will calculate that at the current moment, the theoretical voltage at the end of a certain 10kV feeder should be 9.85kV. The calculation refresh rate of this curve is synchronized with the second-level real-time operation data of the power grid.

[0080] The third step is the generation of the adaptive risk perception threshold. The system uses the theoretical electrical state values ​​of each node on the dynamic risk correlation curve as a benchmark, and superimposes a preset tolerance percentage related to the voltage level and equipment type to generate a dynamic safe operating envelope, i.e., the adaptive risk perception threshold. This threshold includes an upper and lower limit. For example, for a 110kV line, the tolerance can be set to ±2%, while for the 10kV distribution network at the end, it may be relaxed to ±5%, because the voltage fluctuation of the distribution network is inherently larger. Its generation logic is shown in the following formula:

[0081]

[0082]

[0083] In this formula, and These are the upper and lower limits of the adaptive risk perception threshold, respectively. It is the theoretical voltage value at that point calculated from the dynamic risk correlation curve. It is the tolerance coefficient set for that point based on factors such as its voltage level.

[0084] The fourth step is real-time comparison and dynamic updating of risk status. The system will collect the actual measured voltage of each node from the second-level real-time power grid operation data. The adaptive risk perception threshold corresponding to this node and Perform continuous comparisons. Once a node's... This dynamic envelope has been broken, that is > or < The system immediately identifies an abnormal risk at this point. Subsequently, the system dynamically updates the risk panorama distribution map of the node and its upstream associated devices, such as upgrading its color from yellow warning to red alarm and triggering an alarm event, thereby achieving precise detection of minor risks that deviate from normal physical laws.

[0085] Optionally, constructing a multi-level associated path includes: using a first drill operator in the drill operator (which serves as the execution tool for the data drill-up technology) to perform hierarchical processing on the power grid topology corresponding to the voltage level sub-band, generating a main grid topology envelope and a distribution network topology envelope; performing a weighted average calculation on the main grid topology envelope and the distribution network topology envelope to obtain a hierarchical average envelope; and using a second drill operator in the drill operator to perform a topology opening operation on the hierarchical average envelope to generate a dynamic risk association curve from the 110kV voltage level to the power user end.

[0086] Specifically, the core algorithm of the high-precision intelligent penetration monitoring mode, namely the construction of multi-level correlation paths, is described in detail. This process aims to abstract and condense the complex, multi-layered power grid topology into an equivalent calculation path that can accurately characterize the voltage drop characteristics from the high-voltage transmission network to the low-voltage user end, namely the dynamic risk correlation curve, through a series of mathematical and topological operations.

[0087] The first step involves performing topology hierarchical equivalence using the first drill-down operator. The system first activates the first drill-down operator from the drill-down operator set, which serves as the execution tool for data drill-down technology. This operator is a dedicated algorithm module for reducing the order of network models. It receives grid topology data divided by voltage level sub-bands as input and processes the main grid and distribution network separately. For grid topologies composed of voltage levels from 500kV to 220kV, the first drill-down operator simplifies them into a main grid topology envelope through network equivalence calculations, such as Thevenin equivalence or node elimination methods. This envelope is represented in engineering as an equivalent power source and series impedance model, macroscopically characterizing the voltage support characteristics and power transmission capacity of the entire main grid hierarchy to the downstream grid. Similarly, for distribution network topologies of 35kV and below, the operator performs a similar operation, generating a distribution network topology envelope that reflects the overall loss and voltage distribution characteristics of the distribution network.

[0088] The second step involves weighted fusion of the hierarchical models. The system fuses the main network topology envelope and the distribution network topology envelope generated in the previous step to construct a unified model spanning all levels. This fusion is achieved through weighted averaging to obtain the hierarchical average envelope. The calculation focuses on key electrical parameters of the two envelope models, such as equivalent impedance. The calculation formula is as follows:

[0089]

[0090] In this formula, The equivalent impedance representing the average envelope of the hierarchy; and The equivalent impedances of the main network topology envelope and the distribution network topology envelope, respectively; and These are their respective weighting coefficients, which sum to 1. These weights are set based on the typical proportion of different voltage level networks in the total voltage drop. For example, based on historical power flow data analysis, if the voltage drop impact on the main grid side accounts for 30% while that on the distribution network side accounts for 70%, then weights can be set accordingly. =0.3, =0.7.

[0091] The third step involves path refinement using the second drill-up operator. While the hierarchical average envelope obtained by the system unifies the model, it may still contain some side branches or loops with minimal impact on the main path. To obtain a clear computational path from the power source to the load, the system activates the second drill-up operator in the drill-up operator set. This operator performs a topology opening operation on the hierarchical average envelope. Topology opening, a concept derived from mathematical morphology, is used in power grid topology analysis to smooth network structure and eliminate small, isolated topological features. In engineering, this operation can be understood as first performing an "erosion" operation on the network, pruning short, terminal branches, and then performing an "expansion" operation to restore the connectivity of the main path. After this operation, a simplified path representing the path from the power source to the load is obtained. The topology model of the main power supply path from the voltage level power source point to the final power user end is extracted, and the voltage drop calculation curve corresponding to this model is the final output dynamic risk correlation curve.

[0092] Optionally, the process of generating preliminary classification results of causes involves: calculating the power grid impact range and abnormal amplitude change rate of each risk point in the risk panoramic distribution map set to obtain the spatiotemporal characteristics of the risk; comparing the spatiotemporal characteristics of the risk with the risk type feature template preset by the system to generate preliminary classification results of causes including faults, heavy overloads, load mutations, or mode adjustments.

[0093] Specifically, the process enables rapid and automated causal characterization of identified risks, providing maintenance personnel with preliminary diagnostic information beyond risk alarms, thereby significantly shortening the time required to determine faults or anomalies. This process extracts the spatiotemporal propagation characteristics of risk events within the power grid and matches them with pre-defined typical event templates to achieve a preliminary classification of risk causes.

[0094] The first step is the quantitative calculation of the spatiotemporal characteristics of risk. The system traverses every risk point in the risk panorama distribution map set and calculates two key features accordingly:

[0095] Grid impact range (spatial characteristics): Centered on the risk point, the system performs a rapid electrical topology search to count the number of downstream nodes (such as substation busbars and feeder switches) where electrical quantities (such as voltage and current) change significantly due to the risk event. This number is quantified as the grid impact range of the risk. For example, a main network failure could affect dozens of downstream nodes. The value is very large; while for an overloaded terminal transformer, its S value may only be 1.

[0096] Abnormal amplitude change rate (time characteristic): Retrieve key electrical quantities at this risk point, such as current. or power Sample data at the second level within a very short time window before and after the anomaly occurs, such as 2-5 seconds, and calculate its rate of change. As shown in the formula:

[0097]

[0098] In this formula, For abnormal amplitude change rate, This represents the electrical quantity value at the current moment. This is the value from the previous sampling period. This is the data sampling time interval. A line short-circuit fault... The value can reach the order of thousands of amperes per second, and a heavy overload process The values ​​are relatively flat. These two characteristics... and Together, they constitute the spatiotemporal feature vector of the risk point.

[0099] The second step is feature template-based matching and classification. The system has a pre-built risk type feature template library. This library is constructed based on historical power grid event data and simulation analysis, storing standard spatiotemporal feature ranges corresponding to various typical risk causes. For example, the template library might contain the following entries:

[0100] Fault template: Its characteristics are Value greater than 20 and The value is greater than 1000 A / s.

[0101] Heavy overload template: characterized by Values ​​between 2 and 10 and The value is less than 5A / s.

[0102] Burden mutation template: characterized by Value less than 3 and The value is between 100 A / s and 500 A / s.

[0103] Method adjustment template (e.g., capacitor switching): Its characteristics are The value is extremely small, but it can cause a step change in local voltage or reactive power.

[0104] The real-time calculated spatiotemporal feature vector of the risk is compared one by one with all templates in the template library. The comparison logic usually adopts a multi-condition judgment rule. The system classifies the risk point into the cause type corresponding to the template whose feature range can completely contain the spatiotemporal feature vector of the current risk point.

[0105] Finally, the system outputs preliminary classification results of the causes, such as "fault", "heavy overload", "load mutation" or "mode adjustment", and attaches this classification label to the information of the risk point. These results are then presented together on the dynamic guidance monitoring queue and the visualization interface, providing key prior knowledge and directional guidance for subsequent penetrating source tracing analysis.

[0106] Optionally, adjusting the global monitoring sensitivity threshold includes: counting the trigger frequency of the deep drilling command within a preset monitoring window; comparing the trigger frequency with a preset frequency threshold; if the trigger frequency is consistently higher than the frequency threshold, reducing the value of the first system threshold to increase the system's tendency to actively enter the high-precision intelligent penetration monitoring mode.

[0107] Specifically, an adaptive optimization closed loop for global monitoring parameters is established. By monitoring the triggering behavior of deep analysis commands within the system, the sensitivity of the system when switching from conventional monitoring to high-precision monitoring is adjusted in reverse. This enables the system to automatically adapt to the long-term changing trends of power grid risk status and avoids the "passivation" or "hypersensitivity" caused by parameter solidification.

[0108] The first step is to count the triggering frequency of deep drilling commands. A built-in counter and timer are used to monitor and accumulate the number of deep drilling commands generated in real time within a preset, periodically scrolling monitoring window. The duration of this monitoring window is... This is a configurable key parameter, and its setting needs to balance response speed and statistical stability, typically set between 1 and 6 hours. Deep drilling commands are triggered by medium-confidence risk events; their frequent occurrence suggests the presence of numerous ambiguous and difficult-to-determine risk signals in the system. The system calculates the trigger frequency using the following formula. :

[0109]

[0110] In this formula, It is the trigger frequency of deep drilling commands, measured in "times / hour"; In the time window The total number of instruction triggers accumulated within the period.

[0111] The second step is frequency threshold comparison and decision-making. The system will calculate the real-time trigger frequency. With a preset frequency threshold The comparison is then made. This frequency threshold represents the upper limit of the acceptable frequency of "fuzzy events" in the system, as determined by the system designer or operations expert; for example, it could be set to 5 times per hour. This threshold reflects the desired balance between the system's autonomous analysis capabilities and the need for human intervention.

[0112] The third step involves iteratively adjusting the global monitoring sensitivity threshold based on the comparison results. When the system determines that the trigger frequency is consistently higher than the frequency threshold, i.e. > If the state remains stable within one or more consecutive monitoring windows, the system will trigger an automatic adjustment procedure for the global monitoring sensitivity threshold. The core action of this procedure is to perform a slight reduction on the first system threshold Th1, used to distinguish monitoring modes. Its adjustment logic is shown in the following formula:

[0113]

[0114] In this formula, This is the adjusted new threshold. This is the current threshold. It is a preset adjustment step size, usually a small value, such as 0.01, to ensure a smooth adjustment process and avoid system oscillation.

[0115] By lowering the value of the first system threshold Th1, the threshold for the system to enter the high-precision intelligent penetration monitoring mode is effectively lowered. This means that even if the subsequent risk quantification index Q of the power grid increases slightly, the system can be triggered to enter a deep analysis state more quickly. This adjustment increases the system's tendency to proactively enter the high-precision intelligent penetration monitoring mode. The underlying engineering logic is that since the system frequently encounters uncertain risks that require in-depth analysis to determine, powerful analysis tools should be activated earlier and more proactively, rather than waiting until the risk accumulates to a high level before intervention. This allows for early detection and mitigation of potential complex risks, and iterative optimization of the system's visual perception parameters. If the triggering frequency is far below the threshold, the system can also perform a reverse boosting operation to save computational resources.

[0116] Based on the same inventive concept, such as Figure 2 As shown, the present invention also provides an intelligent power grid real-time risk panoramic visualization monitoring system, the system comprising:

[0117] The real-time sensing and assessment module is used to access the second-level real-time operation data of the power grid and perform a preliminary risk assessment using the preset drilling operator to generate risk quantification indicators.

[0118] The intelligent linkage analysis module is used to adaptively select a monitoring and display mode to process the power grid risk distribution based on the risk quantification indicators and generate a panoramic risk distribution map.

[0119] The risk feature and ranking module is used to perform multi-dimensional feature extraction and topological association on each risk point in the risk panorama distribution map set to generate a dynamic guidance monitoring queue.

[0120] The active guidance and penetration analysis module is used to perform penetration-based source tracing analysis according to the order of the dynamic guidance monitoring queue based on the active guidance feedback mechanism, output real-time risk monitoring results and generate deep drilling instructions;

[0121] The global perception optimization module is used to iteratively adjust the global monitoring sensitivity threshold based on the trigger frequency of the deep drilling command, and complete the iterative optimization of the system visualization perception parameters, including the risk quantification indicators.

[0122] To verify the core technical processes of the system, a typical high-load operating section at 14:30 on a certain afternoon was selected for testing. The key parameters preset by the system are as follows: spatial risk weight. =0.6, time fluctuation weight =0.4; First system threshold Th1=0.75; Second system threshold Th2=0.9; Third system threshold Th3=0.6; Dynamic priority score weight (Risk level) = 0.5 (User priority) = 0.3 (Load rate) = 0.2.

[0123] Step 1: Preliminary risk assessment and generation of risk quantification indicators.

[0124] The system accesses the city's power grid's second-level SCADA and EMS real-time operation data and divides it into three sub-county power grid areas: A, B, and C. The system performs parallel calculations on the spatial dimension of risk distribution variance (V) and the temporal dimension (past 15 minutes) of load power fluctuation rate (F) for each area, based on historical 24-hour data.

[0125] Normalization is performed. Finally, a weighted fusion method is used to calculate the global risk quantification index. The detailed calculation process and data are shown in Table 1.

[0126] Table 1. Data Table for Calculation of Risk Characteristics and Global Quantitative Indicators of Power Grid Zones

[0127] County power grid area Risk distribution variance (V) Load power fluctuation rate (F) Normalized variance (NVO) ) Normalized volatility ( ) Area A 0.25 0.040 0.490 0.368 Area B 0.41 0.085 0.816 0.842 Area C 0.18 0.025 0.347 0.211

[0128] Based on the normalized data in Table 1, the system follows the formula... Calculate global risk quantification indicators .

[0129] The contribution to the calculation of region B is: 0.6*0.816+0.4*0.842=0.4896+0.3368=0.8264.

[0130] Assuming the combined impact of areas A and C is relatively small, the global risk quantification index is obtained after weighted summation of all areas. It is 0.81.

[0131] Figure 3 The polar coordinate system was used to display the quantitative indicators of overall risk for each county-level power grid area. The cumulative contribution. Different sectors in the figure represent regions. The radius and depth of the sectors reflect the weighted fusion result of spatial risk variance and temporal load fluctuation. The contribution of different dimensions is distinguished by the shadow texture, which intuitively presents the uneven distribution characteristics of risk in the entire network space.

[0132] Step 2: Adaptive selection of monitoring display mode.

[0133] The calculated risk quantification indicators It is compared with the preset first system threshold Th1=0.75.

[0134] because (0.81)>Th1(0.75), indicating that the overall risk level of the power grid is high, and the system automatically switches from the standard panoramic monitoring mode to the high-precision intelligent penetration monitoring mode. On the visualization interface, area B, which has the greatest risk contribution, is highlighted in red, and its geographical wiring diagram is automatically enlarged. At the same time, the background begins to preload detailed topology data for the risk points in area B.

[0135] Step 3: Dynamically guide the generation of the monitoring queue.

[0136] In high-precision mode, three main risk points were identified in area B: transformer T1, line L2, and switch station S3. Multidimensional features of each risk point were extracted, and their dynamic priority scores were calculated. This is used to generate the analysis queue. Detailed data and calculation process are shown in Table 2.

[0137] Table 2. Data Table of Multidimensional Characteristics of Risk Points and Dynamic Guidance Monitoring Queue Generation

[0138] Risk Point ID Risk Level (L) (1-5) Associated user priority weight ( ) Real-time device load rate ( ) Dynamic priority score ( ) Queue sorting T1 Transformer 5 (normalized to 1.0) 0.9 (Hospital) 98% (normalized to 0.96) 0.5*1.0+0.3*0.9+0.2*0.96=0.962 1 L2 line 4 (normalized to 0.75) 0.5 (Industrial Park) 85% (normalized to 0.82) 0.5*0.75+0.3*0.5+0.2*0.82=0.689 2 S3 switch station 3 (normalized to 0.5) 0.2 (Commercial Area) 70% (normalized to 0.65) 0.5*0.5+0.3*0.2+0.2*0.65=0.440 3

[0139] according to The scores are sorted in descending order, and the system generates a dynamically guided monitoring queue: [T1 transformer, L2 line, S3 switch station]. The dispatcher's attention and system analysis resources will first focus on the T1 transformer.

[0140] Figure 4 A system based on risk level L and load rate was constructed. A dynamic priority isosurface landscape. The background curve represents the priority score. The numerical gradient changes and solid markers show the coordinates of the three risk points T1, L2, and S3 in the assessed landscape. The rigor and scientific nature of the system's sorting logic are verified by the contour intervals in which they are located.

[0141] Step 4: Penetrating source tracing analysis based on proactive guidance and feedback mechanisms.

[0142] The system, following the queue order, first performs proactive guidance prompts on transformer T1. The built-in diagnostic model outputs a preliminary tracing conclusion: "Abnormal A-phase current in transformer T1, suspected heavy overload," with a confidence level of [insert confidence level here]. It is 0.85.

[0143] The system will give this confidence level Compared with the preset threshold: Th3(0.6)≤C(0.85)≤Th2(0.9).

[0144] Since the confidence level was in the medium range, the system's conclusion was still unclear, and a deep drilling command was automatically generated. This command triggered the system to increase the analysis depth from the T1 transformer itself to the level of the 10kV feeder it was connected to, and retrieved the transformer's operating logs and historical maintenance records for the past week (as risk knowledge data) for reanalysis. After reanalysis, the confidence level increased to 0.95, and the final conclusion was confirmed as "the transient overcurrent caused by the short-term grounding fault of the downstream feeder F3 has now been restored, but the insulation status of the equipment needs to be monitored."

[0145] Step 5: Iterative adjustment of global monitoring sensitivity threshold.

[0146] Within a 4-hour monitoring window (14:00-18:00), the system recorded a total of 25 triggers for deep drilling commands.

[0147] Preset monitoring window =4 hours, frequency threshold =5 times / hour, adjust step size Δ=0.01.

[0148] Calculate trigger frequency =25 times / 4 hours = 6.25 times / hour.

[0149] because (6.25)> (5) indicates that there are many ambiguous risks in the system that are difficult to quickly determine, and the current monitoring sensitivity may be too low.

[0150] The global awareness optimization module is automatically triggered to adjust the first system threshold Th1: .

[0151] After the adjustment, the threshold for entering the high-precision intelligent penetration monitoring mode has been lowered, enabling earlier in-depth analysis of potential risks and achieving closed-loop optimization of system parameters.

[0152] Table 3. Data Table of Penetration-Based Source Tracing Analysis and System Sensitivity Adaptive Adjustment

[0153] time Monitoring the first risk point in the queue Preliminary identification confidence level (C) Comparison of decision thresholds (Th3=0.6, Th2=0.9) Source tracing analysis and decision making The cumulative number of deep drilling commands within the 4-hour monitoring window Global sensitivity adjustment 14:30 T1 Transformer 0.85 0.6≤0.85≤0.9 Triggering deep drilling 1 - 14:45 K5 switch 0.92 0.92>0.9 Directly adopt the conclusion 1 - ... .. ... ... ... ... ... 18:00 - - - - 25 Th1 was adjusted from 0.75 to 0.74.

[0154] The detailed process and data of the above embodiments demonstrate the complete closed loop of the present invention, from macro-risk quantification, adaptive mode switching, and micro-risk ranking to in-depth drilling analysis and system self-optimization.

[0155] Table 1 clearly demonstrates how scattered power grid zone operation data can be normalized and weighted to condense into a global risk quantification indicator Q that guides macroeconomic decision-making.

[0156] Table 2 demonstrates, through specific numerical calculations, how the system integrates the severity of risks, the importance of impacts, and the urgency of states, transforming disordered risk points into a logically clear and prioritized analysis queue, embodying the core idea of ​​"intelligent guidance".

[0157] Table 3 uses a specific case and statistics over a period of time to verify how the confidence-based three-segment decision mechanism intelligently determines the depth of analysis, and how the system optimizes key thresholds in reverse by monitoring its own behavior (deep drilling frequency) so that it can maintain its optimal working state in the long term.

[0158] Through the verification of this embodiment, the present invention demonstrates significant advantages over traditional methods in terms of the comprehensiveness of risk identification, the depth of analysis, and the intelligence of response, effectively improving the overall efficiency and security of power grid operation monitoring.

[0159] It should be noted that the electrical connections between the various units described above do not necessarily represent direct or indirect connections. Any indirect connection method can be applied to the embodiments of the present invention as long as it achieves the purpose of the present invention. The above descriptions are merely exemplary embodiments of the present invention and should not be construed as limiting the scope of the present invention.

[0160] All equivalent changes and modifications made in accordance with the teachings of this invention are still within the scope of this invention. Those skilled in the art will readily conceive of other embodiments of this invention upon considering the specification and the disclosure of practical truth. This application is intended to cover any variations, uses, or adaptations of this invention that follow the general principles of this invention and include common knowledge or conventional techniques in the art not described herein.

Claims

1. A method for intelligent power grid real-time risk panoramic visualization monitoring, characterized in that, The method includes: Access the real-time operation data of the power grid at the second level, and use the preset drilling operator to perform a preliminary risk assessment on the real-time operation data of the power grid at the initial drilling depth to generate risk quantification indicators; Based on the aforementioned risk quantification indicators, an adaptive monitoring and display mode is selected to process the power grid risk distribution and generate a panoramic risk distribution map. Perform multi-dimensional feature extraction and topological association on each risk point in the aforementioned risk panorama distribution map set to generate a dynamic guidance monitoring queue; Based on the active guidance feedback mechanism, the penetrating source analysis is performed in the order of the dynamic guidance monitoring queue, outputting real-time risk monitoring results and generating deep drilling instructions. Based on the triggering frequency of the deep drilling command, the global monitoring sensitivity threshold is iteratively adjusted to complete the iterative optimization of the system's visual perception parameters.

2. The intelligent power grid real-time risk panoramic visualization monitoring method as described in claim 1, characterized in that, The generated risk quantification indicators include: The power grid panoramic data, which serves as the real-time operation data set of the power grid, is divided into county-level power grid regions in the spatial dimension and voltage level sub-bands in the hierarchical dimension. For each of the aforementioned county-level power grid areas, calculate the spatial dimension of risk distribution variance and the temporal dimension of load power volatility; A risk quantification index is generated by weighting and fusing the variance of the risk distribution and the load power volatility across all regions.

3. The intelligent power grid real-time risk panoramic visualization monitoring method as described in claim 1, characterized in that, The adaptive selection of a monitoring display mode includes: The risk quantification index is compared with a preset first system threshold. If the risk quantification index is better than the threshold of the first system, then the standard panoramic monitoring mode is selected to display the risk distribution globally in a regular manner. If the risk quantification index is worse than the threshold of the first system, then the high-precision intelligent penetration monitoring mode is selected to perform local highlighting and key penetration analysis on the risk distribution.

4. The intelligent power grid real-time risk panoramic visualization monitoring method as described in claim 1, characterized in that, The generation of the dynamic boot monitoring queue includes: Extract the risk assessment level, associated user priority weight, and real-time device load rate of each risk point in the risk panorama distribution map set as the multidimensional features; Based on the aforementioned multidimensional features, a dynamic priority score is calculated for each risk point; All risk points are sorted in descending order based on the dynamic priority scores to generate a dynamic guidance monitoring queue.

5. The intelligent power grid real-time risk panoramic visualization monitoring method as described in claim 1, characterized in that, The method of performing penetrating source tracing analysis based on the active guidance feedback mechanism includes: Active guidance prompts are executed on the risk points according to the queue order to obtain preliminary source tracing conclusions and their corresponding identification confidence levels; The identification confidence level is compared with a preset second system threshold and a third system threshold, wherein the second system threshold is higher than the third system threshold; If the identification confidence level is higher than the threshold of the second system, the preliminary tracing conclusion will be adopted as the final risk monitoring result; If the identification confidence level is between the third system threshold and the second system threshold, a deep drilling instruction is generated. The deep drilling instruction is used to trigger an increase in the initial drilling depth of the local topology where the risk point is located, and to trigger the input of additional multi-source heterogeneous data for re-association analysis. The multi-source heterogeneous data includes the second-level real-time power grid operation data and preset risk knowledge data.

6. The intelligent power grid real-time risk panoramic visualization monitoring method as described in claim 3, characterized in that, The high-precision intelligent penetration monitoring mode includes: A multi-level correlation path is constructed based on data drilling technology for the real-time operation data of the power grid to generate a dynamic risk correlation curve. Based on the dynamic risk correlation curve, an adaptive risk perception threshold is generated; The real-time operation data of the power grid is compared with the adaptive risk perception threshold, and the risk panorama distribution map is dynamically updated.

7. The intelligent power grid real-time risk panoramic visualization monitoring method as described in claim 6, characterized in that, The construction of multi-level association paths includes: The first drilling operator in the drilling operator, which serves as the execution tool for the data drilling technology, is used to perform hierarchical processing on the power grid topology corresponding to the voltage level sub-band, generating the main network topology envelope and the distribution network topology envelope; The weighted average envelope of the main network topology and the distribution network topology is calculated to obtain the hierarchical average envelope; The second drilling operator in the drilling operator is used to perform a topology opening operation on the hierarchical average envelope to generate a dynamic risk correlation curve from the 110kV voltage level to the power user end.

8. The intelligent power grid real-time risk panoramic visualization monitoring method as described in claim 4, characterized in that, The method also includes a process for generating preliminary causal classification results, including: For each risk point in the aforementioned panoramic risk distribution map set, calculate its power grid impact range and abnormal amplitude change rate to obtain the spatiotemporal characteristics of the risk. The spatiotemporal characteristics of the risk are compared with the risk type characteristic template preset by the system to generate preliminary classification results of causes, including faults, heavy overloads, sudden load changes, or mode adjustments.

9. The intelligent power grid real-time risk panoramic visualization monitoring method as described in claim 1, characterized in that, The iterative adjustment of the global monitoring sensitivity threshold includes: The trigger frequency of the deep drilling command is counted within a preset monitoring window; The trigger frequency is compared with a preset frequency threshold. If the triggering frequency continues to be higher than the frequency threshold, the value of the first system threshold is reduced to increase the system's tendency to actively enter the high-precision intelligent penetration monitoring mode.

10. A smart grid real-time risk panoramic visualization monitoring system, characterized in that... The system is used to implement the intelligent power grid real-time risk panoramic visualization monitoring method according to any one of claims 1-9, the system comprising: The real-time sensing and assessment module is used to access the second-level real-time operation data of the power grid and perform a preliminary risk assessment using the preset drilling operator to generate risk quantification indicators. The intelligent linkage analysis module is used to adaptively select a monitoring and display mode to process the power grid risk distribution based on the risk quantification indicators and generate a panoramic risk distribution map. The risk feature and ranking module is used to perform multi-dimensional feature extraction and topological association on each risk point in the risk panorama distribution map set to generate a dynamic guidance monitoring queue. The active guidance and penetration analysis module is used to perform penetration-based source tracing analysis according to the order of the dynamic guidance monitoring queue based on the active guidance feedback mechanism, output real-time risk monitoring results and generate deep drilling instructions; The global perception optimization module is used to iteratively adjust the global monitoring sensitivity threshold based on the trigger frequency of the deep drilling command, and complete the iterative optimization of the system visualization perception parameters, including the risk quantification indicators.

Images