Flexible adjustment method and device of distribution line based on intelligent power distribution

Abstract

The invention discloses a flexible adjustment method and device for a distribution line based on intelligent power distribution, and relates to the technical field of intelligent power distribution methods, and the method comprises the steps: determining a management and control effect grade based on the management and control content of a peak management and control mode, a corresponding management and control dimension, and a voltage fluctuation event corresponding to the distribution line; and according to each management and control effect grade, the electric quantity change data of the corresponding power consumption peak area and the corresponding current power consumption condition, determining the flexible adjustment event of the power consumption peak area, thereby improving the accuracy of the flexible adjustment event of the power consumption peak area, and realizing power consumption management and control of each power consumption peak area. Therefore, the intelligent power distribution event of each power consumption peak area is determined according to the power distribution state of each power consumption peak area and the corresponding electric quantity loss condition, and the dynamic adjustment system of each power consumption peak area is determined according to the intelligent power distribution event, the power distribution line and each power consumption peak area. And the accuracy of the dynamic adjustment system of each power consumption peak region is improved.

Description

Technical Field

[0001] This invention relates to the technical field of intelligent power distribution methods, and more particularly to a flexible adjustment method and apparatus for power distribution lines based on intelligent power distribution. Background Technology

[0002] Power distribution lines are a common part of the power system and are widely used in urban areas. They are laid out according to the distribution pattern of the city. The power system directly transmits electrical energy from substations or distribution hubs to various users in the city. In the existing technology, current data of each power consumption area in the city is collected, and the current usage status of each power consumption area is determined based on the current data. Current distribution is triggered by the current usage status of each power consumption area. However, there is no control for peak power consumption areas, which affects the accuracy of flexible adjustment events in peak power consumption areas and results in low accuracy of the dynamic adjustment system in each peak power consumption area. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of the prior art. This invention provides a flexible adjustment method and device for power distribution lines based on intelligent power distribution.

[0004] This invention provides a flexible regulation method for power distribution lines based on intelligent power distribution, comprising: Collect current data for each set of power distribution lines, determine peak electricity consumption areas based on the current data and corresponding current detection locations, and determine multiple peak electricity consumption characteristics based on each peak electricity consumption area and corresponding peak electricity consumption time. Multiple power distribution control paths are determined based on the characteristic patterns of multiple peak electricity consumption patterns and power distribution lines. Based on the control scope, corresponding control content, and power consumption status of the corresponding peak electricity consumption area of ​​each power distribution control path, the corresponding peak control method is determined. Based on the control content, corresponding control dimensions, and voltage fluctuation events corresponding to the distribution lines of this peak control method, the control effect level is determined. Based on each control effect level, the corresponding power consumption change data of the peak power consumption area, and the corresponding current power consumption situation, the flexible adjustment event of the peak power consumption area is determined. The flexible adjustment event includes a unique event ID, a clear target area, specific triggering conditions, quantified adjustment target, and an ordered execution sequence. Based on multiple flexible regulation projects and peak electricity consumption areas based on flexible regulation events, the corresponding power control mode is determined. Based on the power control mode, the current power consumption situation of each peak electricity consumption area and the current time, the power distribution status of each peak electricity consumption area is determined. Based on the power distribution status and corresponding power loss of each peak power consumption area, intelligent power distribution events for each peak power consumption area are determined. Based on these intelligent power distribution events, power distribution lines, and each peak power consumption area, a dynamic adjustment system for each peak power consumption area is determined. The dynamic adjustment system is an adaptive closed-loop control system.

[0005] This invention provides a flexible adjustment device for a power distribution line based on intelligent power distribution, which is applied to the aforementioned flexible adjustment method for a power distribution line based on intelligent power distribution.

[0006] Compared with the prior art, the beneficial effects of the present invention are: In this embodiment of the invention, the method determines the control effect level based on the control content, corresponding control dimensions, and voltage fluctuation events corresponding to the power distribution lines of the peak control mode. It then determines the flexible adjustment event for each peak power consumption area based on the power consumption change data and current power consumption status of each control effect level. This introduces multiple power distribution control paths, incorporating a holistic consideration of each control effect level, the corresponding power consumption change data, and the current power consumption status, thereby improving the accuracy of the flexible adjustment event for the peak power consumption area and achieving power consumption control for each peak power consumption area.

[0007] Therefore, based on multiple flexible adjustment projects and peak electricity consumption areas based on flexible adjustment events, corresponding power control modes are determined. Based on this power control mode, the current power consumption situation in each peak electricity consumption area, and the current time, the power distribution status of each peak electricity consumption area is determined. Based on the power distribution status and corresponding power loss in each peak electricity consumption area, intelligent power distribution events in each peak electricity consumption area are determined. Based on these intelligent power distribution events, power distribution lines, and each peak electricity consumption area, a dynamic adjustment system for each peak electricity consumption area is determined. The power distribution status of each peak electricity consumption area is introduced, further controlling the intelligent power distribution events in each peak electricity consumption area. This achieves a holistic consideration of the intelligent power distribution events, power distribution lines, and each peak electricity consumption area, improving the accuracy of the dynamic adjustment system for each peak electricity consumption area. Attached Figure Description

[0008] Figure 1 This is a flowchart illustrating the flexible adjustment method for power distribution lines based on intelligent power distribution in an embodiment of the present invention. Figure 2 This is a flowchart illustrating step S11 in the flexible adjustment method for power distribution lines based on intelligent power distribution in an embodiment of the present invention. Figure 3 This is a flowchart illustrating step S12 in the flexible adjustment method for power distribution lines based on intelligent power distribution in an embodiment of the present invention. Figure 4 This is a flowchart illustrating step S13 in the flexible adjustment method for power distribution lines based on intelligent power distribution in an embodiment of the present invention. Figure 5 This is a flowchart illustrating step S14 in the flexible adjustment method for power distribution lines based on intelligent power distribution in an embodiment of the present invention. Figure 6 This is a flowchart illustrating step S15 in the flexible adjustment method for power distribution lines based on intelligent power distribution in an embodiment of the present invention. Figure 7 This is a schematic diagram of the structural composition of a flexible adjustment device for power distribution lines based on intelligent power distribution in an embodiment of the present invention. Detailed Implementation

[0009] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0010] Please see Figures 1 to 7 A flexible regulation method for power distribution lines based on intelligent power distribution is proposed and applied to intelligent power distribution scenarios. The flexible regulation method for power distribution lines based on intelligent power distribution includes: Step S11: Collect current data of each group of power distribution lines, determine the peak electricity consumption area based on the current data of each group of power distribution lines and the corresponding current detection location, and determine multiple peak electricity consumption characteristics based on each peak electricity consumption area and the corresponding peak electricity consumption time. Step S12: Determine multiple power distribution control paths based on the characteristic patterns of multiple peak electricity consumption patterns and power distribution lines. Determine the corresponding peak control method based on the control scope, corresponding control content, and power consumption status of the corresponding peak electricity consumption area of ​​each power distribution control path. Step S13: Determine the control effect level based on the control content, corresponding control dimensions, and voltage fluctuation events corresponding to the power distribution line of the peak control method; determine the flexible adjustment event of the peak power consumption area based on each control effect level, the corresponding power change data of the peak power consumption area, and the corresponding current power consumption situation. Step S14: Determine the corresponding power control mode based on multiple flexible control items and peak power consumption areas based on the flexible control event, and determine the power distribution status of each peak power consumption area based on the power control mode, the current power consumption situation of each peak power consumption area and the current time. Step S15: Determine the intelligent power distribution events for each peak power consumption area based on the power distribution status and corresponding power loss of each peak power consumption area, and determine the dynamic adjustment system for each peak power consumption area based on the intelligent power distribution events, power distribution lines and each peak power consumption area.

[0011] refer to Figure 2 In step S11, the specific steps are as follows: S111: Real-time monitoring of power distribution lines and collection of power distribution line database; determination of each group of current data of the power distribution line based on the detection of the power distribution line database; determination of peak electricity consumption areas based on each group of current data, the corresponding current detection location and the power distribution line. S112: Mark the location of each peak electricity consumption area, determine the first electricity consumption combination based on the location of the peak electricity consumption area and the corresponding peak electricity consumption time, and determine the second electricity consumption combination based on the area shape of the peak electricity consumption area and the corresponding peak electricity consumption time; S113: Based on the matching of the first and second power consumption combinations, multiple peak power consumption characteristics are determined. These multiple peak power consumption characteristics are distributed in different spaces of the corresponding peak power consumption areas, and present the peak power consumption characteristics of the peak power consumption areas.

[0012] In the embodiments of this application, high-precision synchronous phasor measurement units (PMUs) or intelligent electronic devices (IEDs) deployed on key nodes of the distribution network continuously acquire data streams at a frequency of tens or even hundreds of times per second, thereby capturing rapid load fluctuations, intermittent output of distributed energy resources, and transient events of the power grid. The acquired data (such as three-phase current, voltage, power factor, timestamps, etc.) are transmitted in real time and reliably to the distribution management system (DMS) of the master station or the cloud data platform through standardized communication protocols such as IEC61850 and high-speed, low-latency communication channels such as optical fiber or 5G power private network. These massive amounts of raw data are ultimately stored in databases (such as InfluxDB) or distributed data lakes optimized for time series data. Their architecture design must support high-concurrency writes and efficient time range queries, laying a solid data foundation for subsequent in-depth analysis and model training.

[0013] The system identifies and handles data anomalies: For missing values ​​caused by communication interruptions, the system uses methods such as linear interpolation, forward imputation, or Kalman filtering based on neighboring sensor data for intelligent filling; for abnormal spikes or abrupt changes in data caused by sensor malfunctions or electromagnetic interference, statistical methods (such as the 3σ principle) or machine learning models (such as isolated forests) are used for identification and removal; in addition, due to network transmission delays, there are micro-hour differences in data from different monitoring points, and the system uses precise timestamp alignment to ensure that all current data are logically a network snapshot at the same moment; after this series of refined processing, the system finally obtains a set of high-quality, time-synchronized, and reliable current data sequences that correspond one-to-one with the physical monitoring points.

[0014] The system integrates the geographic information system (GIS) and topology data of the power distribution network with time-series current data to construct a spatial-temporal load model, assigning geographic coordinates and topological location information to each current data point. Then, the system abandons static thresholds and establishes a dynamic load baseline for each monitoring point. This baseline is usually generated by a moving average of historical data from the same period (such as the same period in the past few weeks) or by a more advanced predictive model (such as a long short-term memory network LSTM).

[0015] When the real-time current value continuously exceeds its dynamic baseline by a certain percentage (e.g., 20%) and remains at that level for a preset duration (e.g., 15 minutes), the monitoring point is marked as being in a peak state. The system uses graph theory algorithms (e.g., connected component analysis) to search the power grid topology and aggregate monitoring points that are geographically adjacent and are all in a peak state to form one or more continuous peak electricity consumption areas. The system will also further calculate key characteristic parameters such as the total load and load density of the area to provide a basis for decision-making in subsequent refined regulation.

[0016] Furthermore, the system precisely correlates power grid operation data with physical spatial locations, laying the foundation for subsequent visualization analysis and spatial decision-making. Its technical implementation primarily relies on a Geographic Information System (GIS). The system combines the latitude and longitude coordinates of monitoring points (such as FTU locations) within the peak electricity consumption areas identified in S111 with the GIS topology information of the distribution network, and uses spatial aggregation algorithms (such as convex hull algorithms or rasterized aggregation) to calculate the geographical boundaries and center point coordinates of the peak area. The system generates a standardized set of geographic location information for each peak area, including but not limited to the area boundary coordinate string, center point latitude and longitude, street names, major building or community names, etc., and stores it as a key attribute.

[0017] Based on the location of the peak electricity consumption area and the corresponding peak electricity consumption time, the first electricity consumption combination is determined. After the location is marked, the system enters the combination analysis stage to form the first electricity consumption combination. This combination focuses on classifying the peak area from the perspective of macro-spatial correlation. Its combination logic is mainly based on two dimensions: geographical location and time information. In terms of geographical dimension, the system will analyze the GIS coordinates of all peak areas and aggregate spatially adjacent and physically connected areas (such as multiple residential communities belonging to the same block, powered by the same substation, or only one street apart) into a combination.

[0018] In terms of time, the system examines the time windows in which peak loads occur in peak areas and groups together areas that enter peak conditions simultaneously during the same time period (such as the evening peak for residents from 18:00 to 20:00 or the commercial lunch peak from 11:30 to 13:30). Through this spatiotemporal correlation analysis, the system can identify regional groups with synergistic effects, providing a basis for implementing regional coordinated regulation strategies.

[0019] Based on the regional morphology and corresponding peak electricity consumption time of the peak electricity consumption area, a second level of electricity consumption combination is determined. Building upon the first level, the system further constructs this second level, which focuses more on the similarity of the internal electricity load characteristics and patterns within the region; its combination logic transcends simple spatiotemporal proximity. In terms of regional morphology, the system analyzes the area, shape, load density, and user type composition of the peak area (such as purely residential areas, mixed-use commercial and residential areas, industrial areas, etc.), grouping areas with similar morphology and composition into one category; for example, multiple newly built residential communities with similar area and building density exhibit highly similar load growth curves and peak characteristics.

[0020] In the time dimension, the time information here focuses more on the duration and periodicity of peak loads; for example, grouping together multiple office areas that experience continuous, flat peaks in the afternoons of summer due to concentrated use of air conditioning; through this deep clustering based on load patterns and characteristics, the system can identify regional groups with homogeneous electricity consumption behaviors, providing data support for developing more targeted and standardized flexible adjustment schemes (such as unified demand response strategies for similar communities).

[0021] Therefore, multiple peak electricity consumption characteristics are determined based on the matching of the first and second electricity consumption combinations. These multiple peak electricity consumption characteristics are distributed in different spaces of the corresponding peak electricity consumption areas and present the peak electricity consumption characteristics of the peak electricity consumption areas. This approach is compatible with the overall consideration of matching the first and second electricity consumption combinations, ensuring the accuracy of the multiple peak electricity consumption characteristics.

[0022] At this point, a high-dimensional feature vector is constructed for each peak electricity consumption area. This vector not only contains the basic statistics of the load curve itself (such as peak value, mean, and variance), but more importantly, it incorporates the two combined information formed in S112 as context, such as [peak load, load growth rate, duration, first combination ID, second combination ID, area type], etc. Subsequently, the system enters the matching stage, which is essentially a multi-dimensional clustering or pattern recognition process. The system will use advanced algorithms such as DBSCAN (density-based clustering) or spectral clustering to intelligently group all peak areas in the feature space.

[0023] In this process, the first layer of information-guided algorithm prioritizes grouping areas that are spatially and temporally close and have a linkage effect into the same cluster; while the second layer of information drives the algorithm to discover homogeneous areas that are geographically distant but have highly similar load patterns and behavior patterns; after clustering, each cluster represents a unique peak electricity consumption pattern, and the system analyzes the commonalities of the areas within the cluster and assigns it a specific and interpretable label.

[0024] After identifying various peak electricity consumption characteristics, the system needs to re-associate these abstract characteristics with the physical world and present them quantitatively. The system will then map these feature labels back onto a Geographic Information System (GIS). Each identified peak electricity consumption area will be assigned one or more of the most representative feature labels, allowing dispatchers to visually see the geographical distribution patterns of different electricity consumption characteristics on a map. For example, the evening air conditioning load characteristics of residential areas are concentrated in several large residential areas of the town, while the mixed load characteristics of commercial areas during the day are scattered in the city center. Secondly, the presentation of the characteristics is accomplished through a series of refined quantitative indicators.

[0025] For each region marked with a specific characteristic, the system calculates and updates the detailed parameter set of that characteristic in real time. This includes: time characteristics (peak start / end time, peak time, duration), amplitude characteristics (peak / average power, power factor range), dynamic characteristics (load ramp rate, fluctuation frequency, harmonic content), and correlation characteristics (correlation coefficient with meteorological factors, sensitivity to electricity price signals, etc.).

[0026] refer to Figure 3 In step S12, the specific steps are as follows: S121: Real-time monitoring of multiple peak electricity consumption characteristics, marking the feature patterns of multiple peak electricity consumption characteristics, and determining multiple power distribution control paths based on the feature patterns of multiple peak electricity consumption characteristics, the corresponding peak electricity consumption areas and power distribution lines; S122: In multiple power distribution control paths, the corresponding power distribution control range is determined based on the detection of each power distribution control path, and the control content of each power distribution control path is marked. At the same time, the power consumption status of the peak power consumption area corresponding to the power distribution control path is collected. The corresponding peak control method is determined according to the control range of each power distribution control path, the corresponding control content, and the power consumption status of the corresponding peak power consumption area.

[0027] In the embodiments of this application, the system establishes a digital state machine for each feature, which includes at least the states of inactive, warning, activated, and deactivated. By continuously analyzing the real-time collected data streams such as current, voltage, and timestamps, the system can automatically determine and update the state of each feature. At the same time, for each feature that enters the warning or activated state, the system will start the real-time calculation and tracking of key performance indicators (KPIs), such as the current total load, load growth rate, load margin from the predicted peak, and voltage deviation rate of key nodes. These KPIs are like the vital signs of the feature and are the direct basis for the system to judge the severity of the peak and trigger subsequent control decisions.

[0028] While monitoring feature status, the system needs to assign precise quantitative labels to each activated feature; this labeling is a process of digitizing feature behavior patterns. The system calculates a set of multi-dimensional morphological parameters for each feature, forming a unique feature vector. These parameters include: temporal morphological parameters, such as the ramp rate during load increases (kW / min), the duration of peak periods, and the fluctuation index of the load curve; spatial morphological parameters, such as the geographical coverage radius of peak areas, load density (kW / km²), and electrical distance from adjacent peak areas; and amplitude morphological parameters, such as peak power and peak-valley difference rate. Based on these parameters, the system can use clustering algorithms (such as K-means) to automatically classify all activated features, categorizing them into more general morphological types such as short-term impact type, long-term plateau type, or high-frequency fluctuation type, thereby providing input for subsequent differentiated management strategies.

[0029] The system is based on a real-time topology model of the distribution network (a graph structure), treating each peak electricity consumption area as a target node requiring intervention. The inputs for path planning are the morphological parameters of the features (e.g., rapid climbing features require rapid path response), the geographical location of the area (determining the physical starting point of the path), and the electrical parameters of the lines (such as line impedance and current carrying capacity). The planning process must meet a series of strict constraints, such as the switchgear on the path must be controllable, the lines must not be overloaded, and the operation must not trigger new safety limits. Its objective function is usually multi-objective, aiming to minimize the total path impedance, total response delay, and operating cost.

[0030] To this end, the system will use heuristic algorithms (such as variants of Dijkstra's algorithm) or intelligent optimization algorithms (such as genetic algorithms) to solve the problem. The final generated power distribution control path is a logical sequence that not only includes a series of physical switches and lines from the control center to the target area, but more importantly, it also integrates all available distributed resources on the path (such as energy storage systems, photovoltaic inverters, flexible loads, HLVR equipment, etc.) to form a complete and collaborative control resource pool.

[0031] Specifically, in the town's power distribution lines, the town system is continuously monitoring characteristic A (residential area evening air conditioning load characteristics) of Xingfuli Community. At 19:30, based on real-time data stream, the system determined that the status of this characteristic had transitioned from warning to activated. Its KPI panel showed that the current load was 7.8MW, the load growth rate was 0.5MW / min, and the load margin to the predicted peak of 8.5MW was only 0.7MW. In addition, the voltage on the low-voltage side of the regional distribution transformer was slowly decreasing at a rate of 0.5V / min, indicating that the peak was rapidly approaching and beginning to affect power quality.

[0032] The urban system is continuously monitoring feature A (residential area evening air conditioning load characteristics) of Xingfuli Community. At 19:30, based on real-time data stream, the system determined that the status of this feature had transitioned from warning to activated. Its KPI panel shows that the current load is 7.8MW, the load growth rate is 0.5MW / min, and the load margin is only 0.7MW away from the predicted peak of 8.5MW. In addition, the voltage on the low-voltage side of the regional distribution transformer is slowly decreasing at a rate of 0.5V / min, which indicates that the peak is rapidly approaching and has begun to affect power quality.

[0033] Based on the high-density, rapid-climbing morphological tags and the geographical location of Xingfuli Community, the system immediately launched a multi-objective path planning algorithm. The algorithm evaluated multiple intervention paths in the urban power distribution network topology model: Path 1: Adjust the entire network through the main transformer of Chengnan Substation; Evaluation conclusion: Slow response speed (minutes), wide impact range, unable to achieve precise control of a single community, therefore not adopted; Path 2: Adjust directly through the intelligent terminal (DTU) of Xingfuli Community's own distribution room; Evaluation conclusion: Fastest response speed (seconds), accurate impact range, ideal candidate path.

[0034] The system ultimately determined an optimal power distribution control path: [Dispatch Center] > [Chengnan Substation 10kV Outgoing Switch CB-205] > [Main Line Section Switch FS-01] > [Xingfuli Community Power Distribution Room DTU-07] > [Community Energy Storage System]. Simultaneously, the system detected a high-voltage line regulator (HLVR) connected in series along this path, and therefore automatically included it in the available resource pool for this control path. The final path logic description is: by controlling the FS-01 switch downstream of CB-205, via DTU-07, the active power of the community energy storage system is prioritized for rapid peak shaving, and the HLVR equipment is coordinated for fine reactive power / voltage regulation, thereby achieving rapid, precise, and multi-dimensional flexible control of the 'Xingfuli Community'.

[0035] Furthermore, among multiple power distribution control paths, the corresponding power distribution control range is determined based on the detection of each power distribution control path, and the control content of each power distribution control path is marked. At the same time, the power consumption status of the peak power consumption area corresponding to the power distribution control path is collected. Based on the control range of each power distribution control path, the corresponding control content, and the power consumption status of the corresponding peak power consumption area, the corresponding peak control method is determined. This method takes into account the overall consideration of the control range of each power distribution control path, the corresponding control content, and the power consumption status of the corresponding peak power consumption area, ensuring the accuracy of the corresponding peak control method.

[0036] At this point, the system will start from the target peak electricity consumption area at the end of the path and use graph theory algorithms such as depth-first search (DFS) or breadth-first search (BFS) to trace upstream and identify all electrical equipment controlled by key switches (such as sectionalizing switches and tie switches) on the path, including transformers, line segments and user loads.

[0037] The system performs online or offline power flow calculations to precisely define the electrical boundaries that the path regulation can affect, because the boundaries depend not only on the physical connections, but also on the direction and distribution of the real-time power flow; the control range is defined as a complete set that includes all significantly affected devices and their key parameters (such as rated capacity and impedance).

[0038] Once the control scope is defined, the system needs to inventory and mark the capabilities of all controllable resources within that scope, i.e., mark the controllable content. The system will create a detailed list of controllable resources for each control path. For each controllable device on the path (such as energy storage, photovoltaic inverters, HLVR devices, smart circuit breakers, and controllable load aggregators), the system will call its digital twin model to parameterize its specific control content into a series of standardized control command sets. For example, the control content of the energy storage system is marked as {active power adjustment range: -500kW~+500kW, response time: <200ms}; the HLVR device is marked as {voltage compensation range: -50V~+50V, response time: <20ms}. This process transforms abstract resources into concrete and quantifiable capabilities.

[0039] Having established the path, scope, and resource capabilities, the system must monitor the power consumption status of the target area in real time, which is the basis for decision-making. The system uses multiple data sources, such as SCADA, PMU, and Advanced Metering Management (AMI), to collect multi-dimensional and high-frequency data on the area covered by the control path. The collected data includes not only electrical quantities such as three-phase voltage, current, and active / reactive power, but also status quantities such as switch opening and closing status and transformer tap positions, as well as forecasts such as ultra-short-term load forecasts and distributed energy output forecasts. The system compares this real-time data with preset safety operation constraints (such as voltage upper and lower limits and line thermal stability limits) to make an accurate assessment of the current power grid health status, safety margin, and development trend.

[0040] The system uses the control scope, control content, and power consumption status as the three core inputs for decision-making, and performs comprehensive analysis through its built-in decision engine. This engine can operate based on an expert system rule base (e.g., if voltage deviation > 5% AND energy storage SOC > 20% THEN, execute energy storage discharge + HLVR voltage compensation), or it can use more advanced optimization algorithms (such as linear programming and model predictive control, MPC) to solve the problem. Its goal is usually to achieve optimization objectives such as minimizing network losses, minimizing regulation costs, or maximizing voltage compliance rate while satisfying all safety constraints. The output of the decision engine is a refined control scheme that includes priorities, specific control objects, regulation targets, and quantitative instructions.

[0041] Specifically, the urban system has planned a power distribution control path for Xingfuli Community; the system immediately performed topology analysis on the path [Dispatch Center>CB-205>FS-01>DTU-07]; through the BFS algorithm, the system identified that the downstream electrical islands include: Xingfuli Community power distribution room, 3 low-voltage outgoing circuits, 800 residential users, 1 community charging pile and public lighting system. This complete set of equipment is formally defined as the power distribution control scope of the path.

[0042] The system then scans all controllable resources within the control area and marks their specific control content: Community energy storage under DTU-07: marked as {P regulation: -500~500kW, Q regulation: -250~250kVar, response time <200ms}; HLVR equipment connected in series on the main line: marked as {V compensation: -50~+50V, response time <20ms}; Air conditioning load aggregator contracted with the community: marked as {load reduction capacity: 300kW, response delay: 5min}.

[0043] At 19:35, the system collected the real-time power consumption status of the controlled area through multi-source data fusion: {Total active power: 8.0MW, reactive power: 1.2MVar, minimum voltage: 212V, predicted peak load: 8.6MW (expected to arrive at 20:10), energy storage SOC: 85%}; the status assessment module immediately issued an early warning: if no intervention is taken, according to the load growth trend prediction, the minimum voltage in this area will drop below 207V by 20:10, exceeding the national standard lower limit of -7%.

[0044] The decision engine integrates all the above information and initiates an optimization algorithm based on model predictive control (MPC) with the goal of stabilizing the voltage within the range of 220V±3% at the lowest cost within the next 30 minutes. This ultimately generates a peak load management method called energy storage-voltage coordinated peak shaving, with the following specific instruction sequence: Immediate Execution: Send an instruction to the HLVR device to calculate and output +18V voltage compensation, quickly raising the voltage at critical nodes to a safe level; Immediate Execution: Send an instruction to the community energy storage to perform constant power discharge at 400kW for 30 minutes to reduce peak load; Conditional Trigger: Continuously monitor voltage and energy storage SOC. If the voltage continues to decline after 30 minutes and the energy storage SOC is below 40%, automatically initiate demand-side response to the air conditioning load aggregator to reduce the load by 200kW.

[0045] refer to Figure 4 In step S13, the specific steps are as follows: S131: Based on the identification of the peak control method, determine the corresponding control content, and determine the control dimension corresponding to the peak control method according to the control content of the peak control method and the corresponding peak electricity consumption area; at the same time, construct a voltage dynamic diagram of the distribution line according to multiple voltage data of the distribution line, and determine the voltage fluctuation event corresponding to the distribution line based on the voltage dynamic diagram of the distribution line and multiple peak electricity consumption areas. S132: Determine the first control effect coefficient based on the control content of the peak control method and the voltage fluctuation event corresponding to the distribution line; determine the second control effect coefficient based on the control dimension of the peak control method and the voltage fluctuation event corresponding to the distribution line; and determine the control effect level based on the mapping relationship between the first control effect coefficient, the second control effect coefficient, and the control effect level. S133: Based on the power consumption detection of each peak power consumption area, determine the corresponding power consumption change data, determine the first flexible adjustment content according to each control effect level and the corresponding power consumption change data of the peak power consumption area, determine the second flexible adjustment content according to each control effect level and the current power consumption of the corresponding peak power consumption area, and determine the flexible adjustment event of the peak power consumption area based on the first flexible adjustment content and the second flexible adjustment content.

[0046] In the embodiments of this application, the corresponding control content is determined based on the identification of the peak control method, and the control dimension corresponding to the peak control method is determined according to the control content of the peak control method and the corresponding peak electricity consumption area. At the same time, a voltage dynamic map of the distribution line is constructed based on multiple voltage data of the distribution line, and the voltage fluctuation event corresponding to the distribution line is determined based on the voltage dynamic map of the distribution line and multiple peak electricity consumption areas. This approach takes into account both the voltage dynamic map of the distribution line and the overall consideration of multiple peak electricity consumption areas, ensuring the accuracy of the voltage fluctuation event corresponding to the distribution line.

[0047] At this point, the system receives control methods such as energy storage-voltage coordinated peak shaving through a built-in semantic parser, and decomposes them into a series of indivisible, device-specific atomic instructions. These atomic instructions constitute the control content at the execution level. To ensure machine readability and executableness, each control content is structured into a standardized data object, which clearly includes key fields such as {target device ID, control type, regulation parameter, expected response time}. For example, an instruction for energy storage is structured as {target device: ESS-07, control type: Discharge, regulation parameter: 400kW, expected response time: <1s}, thereby transforming a policy intent into a precise action command.

[0048] After obtaining the specific control content, the system needs to understand the physical mechanism of the control method at a higher level, that is, to determine its control dimensions. The system will perform cluster analysis on all the parsed atomic control content, and automatically map these content categories to abstract control dimensions through a predefined content-dimension mapping rule base. These dimensions represent the physical level at which the regulation strategy affects the power grid. Common dimensions include: active power dimension (when energy storage charging and discharging, load shedding, etc. are involved), reactive power / voltage dimension (when HLVR, SVG, etc. are involved), network topology dimension (when switching operations are involved), and time dimension (when time-of-use pricing or demand response are involved). Through this step, the system can summarize whether the current control strategy has a single-dimensional effect or a multi-dimensional synergistic effect.

[0049] To intuitively and in real-time grasp the operating status of the power grid, especially power quality, the system constructs a dynamic voltage map using multi-source data. The system collects synchronous voltage amplitude and phase angle data at a high frequency (e.g., 10-50 times per second) from PMUs (phasor measurement units), FTUs (feeder terminal units), and smart meters (AMIs) deployed throughout the distribution network. Subsequently, a powerful visualization engine renders this massive amount of spatiotemporal data into a dynamic image, which can be a two-dimensional heat map, where the horizontal axis represents time, the vertical axis represents different nodes on the line, and the color depth represents the voltage level.

[0050] After constructing the voltage dynamic graph, the system enters the intelligent diagnostic phase, automatically identifying voltage fluctuation events. The system uses statistical algorithms based on sliding windows (such as calculating the standard deviation of voltage within the window) or more complex machine learning models (such as Isolation Forest and Long Short-Term Memory Network LSTM) to analyze the data flow of the voltage dynamic graph in real time. When the system detects that the pattern of the voltage data sequence deviates significantly from the preset normal operating baseline model, it triggers a voltage fluctuation event. At the same time, based on the characteristics of the deviation (such as drop, rise, oscillation), the system automatically classifies the event and records its key attributes, forming a structured event record: {event type, location of occurrence, start time, duration, maximum deviation amplitude}, providing accurate input for subsequent effect evaluation and cause analysis.

[0051] Specifically, the town system has formulated a peak control method for the Xingfuli community, which combines energy storage and voltage shaving. The system immediately parses the instructions of this method and decomposes them into two atomized control contents: Content 1: {Target device: Community Energy Storage-ESS-07, Control type: Constant power discharge, Adjustment parameter: 400kW, Duration: 30min}; Content 2: {Target device: HLVR device-HLVR-01, Control type: Dynamic voltage compensation, Adjustment parameter: Target voltage 225V, Response mode: Continuous}.

[0052] The system analyzes the above content: Content 1 clearly involves the regulation of active power, while Content 2 directly affects voltage; by querying the internal mapping rule base, the system quickly determines that the current peak control method simultaneously affects two physical dimensions and marks them as: {active power dimension, reactive power / voltage dimension}, indicating that this is a multi-dimensional collaborative composite control strategy.

[0053] At 19:38, the system was collecting real-time voltage data from the low-voltage side of the distribution transformer and three key branch boxes in Xingfuli Community at a frequency of 20 times per second. The visualization engine drew a dynamic heat map of these data streams in real time. On the dispatch screen, it could be clearly seen that the color of the area representing Xingfuli Community was rapidly changing from green, which represented normal voltage, to yellow and orange, which represented low voltage.

[0054] When the system's LSTM anomaly detection model continuously analyzed the data stream of the dynamic graph, it found that the current voltage sequence change pattern highly matched the trained summer evening load surge model, and the voltage drop rate had exceeded the preset safety threshold. Therefore, the system automatically triggered and recorded a voltage fluctuation event at 19:38:15: {Event type: voltage sag, Location: downstream of DTU-07 in Xingfuli Community, Start time: 19:38:10, Current duration: 5s, Maximum deviation: -13V (from 225V to 212V)}.

[0055] Furthermore, a first control effect coefficient is determined based on the control content of the peak control method and the voltage fluctuation events corresponding to the distribution lines. A second control effect coefficient is determined based on the control dimension of the peak control method and the voltage fluctuation events corresponding to the distribution lines. The control effect level is determined based on the mapping relationship between the first control effect coefficient, the second control effect coefficient, and the control effect level. This approach takes into account the overall consideration of the mapping relationship between the first control effect coefficient, the second control effect coefficient, and the control effect level, ensuring the accuracy of the control effect level.

[0056] At this point, the first control effect coefficient is determined based on the control content of the peak control method and the voltage fluctuation event corresponding to the distribution line. The first control effect coefficient aims to evaluate the matching degree between the direct physical capability of the control method and the severity of the problem. Its calculation usually adopts the capacity-demand ratio model. The system extracts key capability parameters (such as the maximum voltage compensation capability of HLVR and the maximum discharge power of energy storage) from the atomic control content and compares them with the key demand parameters (such as voltage deviation amplitude and power deficit required to eliminate the event) of the voltage fluctuation event identified in S131.

[0057] For example, for voltage regulation, E1 = (amplitude of available voltage compensation capability) / (amplitude of deviation of voltage fluctuation event); E1 is a dimensionless ratio, and its value intuitively reflects the sufficiency of capability: E1≥1 indicates that the capability fully covers the demand and the expected effect is excellent; 0.5≤E1<1 indicates that the capability partially covers the demand and auxiliary means are required; while E1<0.5 indicates that the capability is seriously insufficient and the control method is ineffective.

[0058] The second control effect coefficient is determined based on the control dimensions of the peak control method and the voltage fluctuation events corresponding to the distribution lines. The second control effect coefficient aims to evaluate the compatibility between the control method's effect dimensions and the physical nature of the problem. The calculation of E2 is usually based on an expert rule base or a simple classification scoring model. The system will match and score according to the type of voltage fluctuation event (such as voltage sag, frequency fluctuation) and the control dimensions of the peak control method (such as reactive power / voltage dimension, active power dimension).

[0059] For example, the rule base defines that the adaptability score for addressing voltage sag using the reactive power / voltage dimension is 0.9, while the adaptability score for addressing the same problem using the active power dimension is 0.6 (because adjusting active power can also affect voltage, but it is not as accurate and fast as directly adjusting voltage); if a control method involves multiple dimensions, E2 can be a weighted average of multiple scores; the value of E2 is usually between [0,1], and the closer it is to 1, the higher the dimension adaptability.

[0060] After obtaining two quantitative indicators, capability matching degree (E1) and dimension fit degree (E2), the system needs to integrate them into an intuitive, qualitative evaluation result, namely the control effectiveness level. The system has a built-in mapping relationship from two-dimensional input (E1, E2) to one-dimensional output. This relationship can be a decision tree, a fuzzy logic inferencer, or a simple two-dimensional lookup table. For example, a decision tree rule definition is: IFE1≥1.0 AND E2≥0.8 THEN = Excellent; IFE1≥0.7 AND E2≥0.8 THEN = Good; IFE1≥0.7 AND 0.5≤E2<0.8 THEN = Average; ELSE = Poor. Through this mapping relationship, the system finally outputs a discrete label, such as {Excellent, Good, Average, Poor}, providing a clear and intuitive basis for subsequent decision-making.

[0061] Specifically, the urban system has identified the voltage dip event in Xingfuli Community and parsed the corresponding control content. In order to calculate E1, the system extracted the available voltage compensation capability of the HLVR device from the control content as ±50V, and extracted the deviation amplitude of -13V from the voltage dip event record. The calculation showed that: E1=50V / 13V≈3.85. This result is much greater than 1, indicating that the physical capability of the HLVR device is fully sufficient to cover this voltage drop, and its direct regulation capability is very strong.

[0062] The system then calculates E2; it knows that this control has {active power dimension, reactive power / voltage dimension}, and the event type is voltage sag; the system queries the rule base and finds that the matching score between voltage sag and reactive power / voltage dimension is 0.9; the matching score between voltage sag and active power dimension (voltage is increased by energy storage discharge) is 0.6; since the current adjustment strategy is mainly based on HLVR fast voltage compensation, with energy storage as a secondary measure, the system gives a higher weight to the voltage dimension (e.g., 0.8) and a lower weight to the active power dimension (e.g., 0.2); the comprehensive calculation yields: E2=(0.9×0.8)+(0.6×0.2)=0.72+0.12=0.84, this high score indicates that the dimension selection for this control is highly suitable for the nature of the problem.

[0063] The system inputs the calculated values ​​(E1=3.85, E2=0.84) into the preset decision tree mapping relationship; the decision process is as follows: Condition 1: E1≥1.0 (3.85≥1.0)>True; Condition 2: E2≥0.8 (0.84≥0.8)>True; According to the decision tree rule IFE1≥1.0 AND E2≥0.8 THEN grade='Excellent', the system finally determines the control effect level of this peak control method as excellent; Through step S132, the system gives an excellent rating to the initially formulated control strategy through a quantitative and logically rigorous evaluation process. This rating greatly enhances the system's confidence in implementing the strategy and provides key input for generating more refined flexible adjustment events in the future.

[0064] Therefore, based on the power consumption detection of each peak power consumption area, the corresponding power consumption change data is determined. The first flexible adjustment content is determined according to each control effectiveness level and the corresponding power consumption change data of the peak power consumption area. The second flexible adjustment content is determined according to each control effectiveness level and the current power consumption situation of the corresponding peak power consumption area. The flexible adjustment event for that peak power consumption area is determined based on the first and second flexible adjustment contents. This approach considers both the first and second flexible adjustment contents as a whole, ensuring the accuracy of the flexible adjustment event for the peak power consumption area. Simultaneously, multiple power distribution control paths are introduced, considering the overall situation of each control effectiveness level, the corresponding power consumption change data of the peak power consumption area, and the corresponding current power consumption situation, thus improving the accuracy of the flexible adjustment event for the peak power consumption area and realizing power consumption control for each peak power consumption area.

[0065] At this point, the system continuously collects electrical quantities such as active power and reactive power at a high frequency of seconds or minutes by aggregating data from smart metering units (such as DTUs and FTUs) deployed at the area entrance or from smart meters (AMIs) within the area. However, the system does not simply record instantaneous values, but calculates a series of power change data that can quantify the dynamic behavior of the load in real time based on a sliding time window (such as the most recent 5 minutes). These key data include: load change rate (dP / dt, in kW / min), which measures the speed of load growth; load fluctuation index (the ratio of the standard deviation to the mean of the load data within the window), which reflects the stability of the load; and prediction deviation rate ((actual load - predicted load) / predicted load), which measures the degree to which the current load exceeds expectations.

[0066] Based on the power consumption change data of each control effect level and the corresponding peak power consumption area, the first flexible adjustment content is determined. After grasping the dynamic trend of the load, the system begins to formulate proactive adjustment strategies, namely the first flexible adjustment content. Its generation follows a capacity-trend matching logic, combining the expected effect (effect level) of the control strategy evaluated in S132 and the current dynamic trend of the load (power consumption change data). The system has a built-in rule engine that makes decisions based on the combination of effect level and power consumption change data. For example, when the effect level is excellent and the load change rate exceeds the threshold, the system will determine it as an excellent effect, rapid change scenario, and generate forward-looking peak shaving content, that is, make full use of efficient adjustment resources to intervene in advance, smooth the load curve, and avoid the problem from worsening.

[0067] Conversely, if the effectiveness level is good or moderate but the load changes rapidly, the system will generate more conservative preventative support content, prioritizing key safety indicators; the generated content is specific, parameterized instructions, such as discharging at power A for B minutes.

[0068] Based on the current electricity consumption situation in each peak electricity consumption area and the level of control effectiveness, the second flexible adjustment content is determined. As a supplement or backup to the first line of defense, the system will further formulate responsive adjustment strategies, namely the second flexible adjustment content. Its generation follows a capacity-state matching logic, combining the expected effect of the control strategy and the real-time status of the power grid (current electricity consumption situation). Similarly, through the rule engine, the system performs scenario matching. For example, when the effectiveness level is excellent and the current voltage, frequency and other status indicators are all within the normal range, the system will generate refined optimization content to utilize surplus capacity for network loss optimization or power quality improvement.

[0069] When the effect level is poor and the current voltage is close to the limit, the system will generate emergency load shedding content as a last resort to avoid safety accidents; the second flexible adjustment content usually has priority labels, such as standby, immediate execution, etc., which clarifies its role in the whole strategy.

[0070] By integrating and encapsulating the aforementioned proactive and reactive strategies, a complete and executable flexible adjustment event is formed. A flexible adjustment event is an instruction package containing all necessary information, defining all the details of the adjustment action. A flexible adjustment event includes a unique event ID, a clearly defined target area, specific triggering conditions, quantified adjustment goals, and an ordered execution sequence containing the first and second contents (and defining their execution logic, such as parallel, serial, or conditional triggering). In addition, a complete event also includes the expected duration and a rollback mechanism, that is, defining how the system can safely return to the initial state after the adjustment ends or fails.

[0071] Specifically, at 19:40, the urban system was continuously monitoring the electricity data of Xingfuli Community; the system detected that the load in the area rapidly increased from 7.5MW to 8.2MW within 5 minutes; based on sliding window calculations, the system obtained key electricity change data: {load change rate: +84kW / min, load fluctuation index: 0.05 (relatively stable), prediction deviation rate: +15% (exceeding expectations)}, indicating that the load is growing steadily at a rate exceeding expectations.

[0072] The system takes {Effect Level: Excellent} and {Electricity Change Data: {Change Rate: +84kW / min>Threshold 60kW / min}} as input; the decision engine matches the scenario with excellent effect and rapid change; therefore, the system determines the first flexible adjustment content as proactive peak shaving and generates a specific instruction: commanding the community energy storage (ESS-07) to discharge at full power of 500kW for 30 minutes to actively suppress the rapid growth of load.

[0073] The system then assesses the current power consumption: {Current voltage: 213V (still low but controllable), energy storage SOC: 85% (sufficient)}; combined with {Effect level: Excellent}, the decision engine matches the scenario with excellent effect and acceptable status; therefore, the system determines the second flexible adjustment content as refined voltage support and generates specific instructions: command the HLVR device (HLVR-01) to enter dynamic compensation mode, calculate and output △U in real time, with the goal of accurately stabilizing the user-end voltage at 225V as a supplement to energy storage peak shaving.

[0074] The system integrates and encapsulates the above two contents to generate a flexible regulation event named "Xingfuli Community - Superior Collaborative Peak Shaving and Voltage Stabilization". Its structure is as follows: Event ID: FLE-20231025-1940-C-XFL; Target area: Xingfuli Community; Execution sequence: Parallel execution: Task 1 (first content): Issue a 500kW discharge command to ESS-07; Task 2 (second content): Issue a 225V target voltage command to HLVR-01; Condition trigger (backup): IF After 15 minutes of execution, if the voltage is <218V, THEN, initiate demand-side response and reduce the load by 100kW; Rollback mechanism: After 30 minutes or after the voltage stabilizes, the energy storage stops discharging, and the HLVR switches back to bypass mode; Through step S133, the system successfully transforms an abstract superior rating and a series of dynamic data into a specific flexible regulation event that includes primary and secondary strategies and backup plans, realizing a closed loop from evaluating the effect to formulating the final action plan.

[0075] refer to Figure 5 In step S14, the specific steps are as follows: S141: Collect flexible adjustment events, identify multiple flexible adjustment items based on the identification of flexible adjustment events, and determine the corresponding flexible control level according to the item content of multiple flexible adjustment items, the location of the corresponding peak electricity consumption area and the current electricity consumption situation; S142: Determine the first power control coefficient based on the project content of multiple flexible adjustment projects and the regional shape of peak power consumption areas; determine the second power control coefficient based on the flexible control level of multiple flexible adjustment projects and the regional shape of peak power consumption areas; and determine the corresponding power control mode based on the mapping relationship between the first power control coefficient, the second power control coefficient, and the power control mode. S143: Collect the current electricity consumption of each peak electricity consumption area, determine the power distribution content of each peak electricity consumption area based on the current electricity consumption of each peak electricity consumption area and the power control mode, and determine the power distribution status of each peak electricity consumption area based on the power distribution content of each peak electricity consumption area, the corresponding flexible adjustment events and time.

[0076] In the embodiments of this application, flexible adjustment events are collected, and multiple flexible adjustment items are determined based on the identification of flexible adjustment events. The corresponding flexible control level is determined according to the item content of multiple flexible adjustment items, the location of the corresponding peak electricity consumption area, and the current electricity consumption situation. This approach takes into account the overall consideration of the item content of multiple flexible adjustment items, the location of the corresponding peak electricity consumption area, and the current electricity consumption situation, thereby ensuring the accuracy of the corresponding flexible control level.

[0077] At this point, the system collects structured flexible adjustment events, which are complete instruction packages containing objectives, strategies, and conditions. Through a built-in event parser, the system performs in-depth parsing of the event according to predefined syntax or patterns (such as XMLSchema or JSONSchema). The core function of the parser is to decompose a composite event into multiple atomic, single-objective flexible adjustment items. This decomposition process strictly follows the execution sequence field in the event, separating the defined serial or parallel tasks one by one. Each identified item is instantiated as a standardized data object, which clearly contains key attributes such as {item ID, associated event ID, target device ID, control action type, adjustment parameters, execution priority}.

[0078] After breaking down the event into specific projects, the system needs to assign a reasonable execution intensity level to each project to ensure the accuracy and safety of regulation. Determining the flexible regulation level is a risk assessment and decision-making process based on multi-factor input. Its core is to quantify the urgency of the current power grid status and match an appropriate regulation intensity accordingly. The system will quantify input from three dimensions: first, the project content, assessing the regulation capacity and potential impact of the project itself (e.g., the impact of energy storage discharge is mild, while load shedding has a greater impact); second, the regional location, querying the user attributes of the area through the GIS system (e.g., whether it is an important user such as a hospital or government agency); and finally, the current electricity consumption situation, calculating an urgency index by collecting real-time electrical data and comparing it with safety thresholds.

[0079] The system uses a fuzzy logic controller or a multi-level decision tree as the decision model, mapping these quantified inputs to an ordered set of labels, such as {Level 1 - Fine-tuning, Level 2 - Regular adjustment, Level 3 - Strong intervention, Level 4 - Emergency control}. This level defines the intensity of the adjustment and its impact on the user, ensuring that subsequent actions can effectively solve the problem while avoiding over-adjustment.

[0080] Furthermore, a first power regulation coefficient is determined based on the project content of multiple flexible regulation projects and the regional shape of peak electricity consumption areas. A second power regulation coefficient is determined based on the flexible regulation level of multiple flexible regulation projects and the regional shape of peak electricity consumption areas. The corresponding power regulation mode is determined based on the mapping relationship between the first power regulation coefficient, the second power regulation coefficient, and the power regulation mode. This approach takes into account the overall consideration of the mapping relationship between the first power regulation coefficient, the second power regulation coefficient, and the power regulation mode, ensuring the accuracy of the corresponding power regulation mode.

[0081] At this point, the first power regulation coefficient is determined based on the project content of multiple flexible regulation projects and the regional shape of peak electricity consumption areas. This coefficient aims to evaluate the matching degree between the physical characteristics of the regulation means and the spatial distribution characteristics of the load. Its calculation model is usually a multi-parameter weighted function. The system converts the project content into quantifiable physical attributes, such as regulation accuracy, response speed, and range of action (point / line / area). At the same time, the regional shape determined in S11 (such as point-like concentrated type) is quantified into indicators such as load density and spatial dispersion. Through a matching degree function, the system performs comprehensive calculations on these parameters. For example, a high-precision regulation project with a point-like range of action (such as energy storage) in a high-density, low-dispersion point-like concentrated type area will have a very high C1 value. The value range of C1 is usually between [0,1]. The closer it is to 1, the higher the physical matching degree between the project content and the regional shape, and the higher the regulation efficiency.

[0082] The second power regulation coefficient is determined based on the flexible regulation levels of multiple flexible regulation projects and the regional morphology of peak electricity consumption areas. This coefficient aims to evaluate the adaptability between the intensity of the regulation strategy and the spatial distribution characteristics of the load. Its calculation is also based on a weighting function. The system maps discrete regulation levels (e.g., Level 1 - fine adjustment, Level 2 - normal regulation) to numerical values ​​(e.g., Level 1 > 0.25, Level 2 > 0.5). At the same time, scale parameters such as area and total number of users are extracted from the regional morphology. Through an adaptability function, the system performs a correlation analysis between regulation intensity and regional scale. For example, a Level 4 - emergency control strong regulation level will have a higher C2 value in a large area with many users because strong regulation needs to cover a sufficiently large area to be effective. Conversely, the C2 value will be lower when strong regulation is used in a small area. The value range of C2 is also between [0,1], and the closer it is to 1, the better the adaptability between the regulation level and the regional scale.

[0083] After obtaining two quantitative indicators, physical matching degree (C1) and intensity adaptability degree (C2), the system needs to integrate them to select the most suitable power control mode. The power control mode is a predefined, higher-level strategy template that specifies how to organize and coordinate different regulatory resources to achieve a specific goal. The system has a built-in mapping relationship from two-dimensional input (C1, C2) to mode output. This relationship is usually a two-dimensional decision matrix or a K-means clustering model. For example, a decision matrix is ​​defined as follows: when C1 is high and C2 is low, select the precise point control mode; when both C1 and C2 are high, select the collaborative optimization mode. The system finds the corresponding cell in the matrix based on the calculated C1 and C2 values ​​to determine the most suitable power control mode.

[0084] Therefore, by collecting the current electricity consumption data of each peak electricity consumption area, determining the power distribution content of each peak electricity consumption area based on the current electricity consumption data and the power control mode, and determining the power distribution status of each peak electricity consumption area based on the power distribution content, corresponding flexible adjustment events, and time, this method takes into account the overall consideration of the power distribution content, corresponding flexible adjustment events, and time of each peak electricity consumption area, ensuring the accuracy of the power distribution status of each peak electricity consumption area.

[0085] At this time, the system collects real-time electrical data snapshots of the target peak electricity consumption area at the decision time through SCADA, AMI and other systems, including total active / reactive power, key node voltage, frequency and ultra-short-term load forecast curves, etc.; then, the system starts a mode-command mapping engine, which uses different algorithms to generate commands according to the power control mode determined in S142.

[0086] For example, in the precise point control mode, the engine uses a feedforward-feedback composite control algorithm, combining load forecasting (feedforward) and current voltage deviation (feedback) to generate precise adjustment commands for a single or a few high-precision devices; in the regional balancing mode, the optimal power flow (OPF) simplification algorithm is used to calculate the coordinated output of multiple distributed resources within the region; and in the powerful peak shaving mode, a preset emergency control sequence list is activated; the final generated power distribution content is a set of structured instructions, each of which clearly defines {target device ID, control type, setpoint / target value, rate of change limit, execution timestamp}.

[0087] After generating specific power distribution content, the system needs to integrate these actions with contextual information to form a complete power distribution state that describes the power grid's operating status. This can be seen as a digital twin snapshot of the power grid after executing regulation commands. This state is a comprehensive information package, whose components include: the core power distribution content, which defines what is being done; the flexible regulation events that trigger this state, which provide the context and cause of why this is being done; and complete time-dimensional information, including the state generation timestamp, the expected start time, and the duration.

[0088] More importantly, the system will quickly simulate the power grid state after the power distribution content is executed based on the power flow calculation model, and generate key indicators such as expected voltage curve, expected load curve, and expected network loss change. This complete power distribution state will be assigned a unique ID and published to the real-time database of the power grid for monitoring system, operation and maintenance personnel and upper-level applications to call. At the same time, it will be recorded in the historical database for post-event analysis and model optimization.

[0089] refer to Figure 6 In step S15, the specific steps are as follows: S151: In multiple peak electricity consumption areas, multiple power loss data are determined based on the detection of each peak electricity consumption area, and the power loss situation of the peak electricity consumption area is determined according to the multiple power loss data, time and the corresponding regional location of the peak electricity consumption area. S152: Collect the overall power supply of the power distribution lines, determine the smart power distribution events for each peak power consumption area based on the power distribution status, corresponding power loss, and overall power supply of the power distribution lines; and determine multiple smart power distribution projects based on the identification of these smart power distribution events. S153: Determine the first level of dynamic adjustment content based on multiple smart power distribution projects and power distribution lines; determine the second level of dynamic adjustment content based on multiple smart power distribution projects and various peak electricity consumption areas; and determine the dynamic adjustment system for each peak electricity consumption area based on the first and second level of dynamic adjustment content.

[0090] In the embodiments of this application, multiple power loss data are determined based on the detection of each power consumption peak area. The power loss situation of the power consumption peak area is determined according to the multiple power loss data, time and the corresponding regional location of the power consumption peak area. This approach takes into account the overall consideration of multiple power loss data, time and the corresponding regional location of the power consumption peak area, ensuring the accuracy of the power loss situation in the power consumption peak area.

[0091] At this point, the system acquires raw data through a multi-level measurement system, including the output power of the main transformer at the substation level, the power of each outgoing line at the feeder level, and the total power of each peak power consumption area at the user / transformer level. After obtaining this data, the system uses a combination of topology analysis and power flow calculation to determine the loss data. On the one hand, the system calculates the theoretical power loss of each line segment based on the topology of the power grid, line parameters (resistance R, reactance X) and real-time current data.

[0092] On the other hand, considering the limited number of measurement points, the system will run state estimation algorithms (such as weighted least squares) to calculate the state of all nodes in the entire network using limited measurement data, thereby more accurately calculating the theoretical loss distribution of the entire network; the system outputs a time series power loss dataset, with each data point recording in detail {timestamp, area ID, line loss, transformer loss, total loss}.

[0093] After obtaining accurate loss data, it is transformed into a power loss assessment report with decision-making value. The system deeply correlates the calculated loss data with time and regional location to construct a multi-dimensional loss profile. The system uses historical loss data to establish a loss baseline model for each peak electricity consumption area at different time scales. This model represents the typical loss level of the area under normal operating conditions.

[0094] The system compares real-time loss data with a baseline model to quantify loss anomalies, thereby identifying uneconomical operating modes or potential equipment failures. Simultaneously, through a GIS system, the system maps loss data onto a geographic map, using prominent colors to mark lines or transformers with high loss values, visually locating loss hotspots. The power loss situation of a region is defined as a comprehensive assessment report, whose structure includes: {Region ID, Current Real-Time Loss Rate, Loss Anomaly Level, List of Loss Hotspots, Comparison with Historical Data, Analysis of Main Causes of Loss}.

[0095] Furthermore, the overall power supply of the distribution lines is collected, and intelligent distribution events for each peak power consumption area are determined based on the distribution status, corresponding power loss, and overall power supply of the distribution lines. Based on the identification of these intelligent distribution events, multiple intelligent distribution projects are identified, which takes into account the overall considerations for the identification of intelligent distribution events and ensures the accuracy of multiple intelligent distribution projects.

[0096] At this point, the system collects the total active power and total power supply of the distribution line in real time from the substation's SCADA system to grasp the overall macroscopic operating status and load-bearing pressure of the entire line. Subsequently, the system constructs a composite trigger rule engine, which continuously integrates and evaluates information from three core dimensions: first, the distribution status of the area, to determine whether it is currently in adjustment mode and whether the effect has been achieved; second, the power loss situation of the area, to assess whether the loss rate is abnormal and whether there are loss hotspots; and third, the overall power supply of the line, to determine whether the line load rate is close to the limit and whether there is sufficient margin to support more complex operations.

[0097] The triggering logic of intelligent power distribution events is combinatorial, aiming to achieve system-level optimization. For example, when the system determines that a region is under adjustment and has a high degree of loss abnormality, while the line load rate still has a margin, an economic operation optimization event will be triggered. This event is a higher-order decision, marking the system's shift from dealing with a single problem to pursuing overall performance improvement. It contains key information such as {event ID, event type, triggering region, optimization target}.

[0098] When a smart distribution event is triggered, the system needs to decompose it into a series of specific, executable tasks, i.e., smart distribution projects. The system predefines a project decomposition template for each type of smart distribution event. When an event is triggered, the system will call the corresponding template according to the event type and decompose it into multiple projects.

[0099] Unlike the flexible regulation projects in S141 that focus on single-device control, smart distribution projects emphasize system-level analysis, strategy adjustment, and long-term optimization. They do not directly control equipment but provide decision-making basis for the generation of subsequent control strategies. Each project is instantiated as a task object, containing {project ID, associated event ID, project type, objective, input data requirements, expected output}. Project types can be diverse, such as analysis-type three-phase imbalance root cause analysis, strategy-type formulation of intelligent commutation regulation strategies, or coordination-type optimization of multi-energy storage system coordinated charging and discharging strategies.

[0100] Therefore, the first level of dynamic adjustment content is determined based on multiple smart power distribution projects and power distribution lines, and the second level of dynamic adjustment content is determined based on multiple smart power distribution projects and various peak electricity consumption areas. The dynamic adjustment system for each peak electricity consumption area is then determined based on the first and second levels of dynamic adjustment content, incorporating the overall consideration of both levels to ensure the accuracy of the dynamic adjustment system for each peak electricity consumption area. Simultaneously, the power distribution status of each peak electricity consumption area is introduced to further control smart power distribution events in each peak electricity consumption area. This achieves an overall consideration of the smart power distribution events, power distribution lines, and each peak electricity consumption area, thereby improving the accuracy of the dynamic adjustment system for each peak electricity consumption area.

[0101] At this point, the first level of dynamic adjustment content is determined based on multiple smart distribution projects and distribution lines. This first level of dynamic adjustment content is a specification at the technical implementation layer, which clarifies which specific hardware resources need to be mobilized, what technical means need to be adopted, and the physical parameters and performance indicators of these technical means in order to complete the smart distribution project. The system queries and matches in a technical resource library according to the type and objectives of each smart distribution project. For example, for analysis-type projects, the first level of content includes calling the data interface of the Advanced Measurement System (AMI) and enabling specific data mining algorithms; while for strategy-type projects, it includes identifying and locking all available smart phase-changing switches in the area, defining their communication protocols and safety interlocking conditions.

[0102] Based on multiple smart power distribution projects and various peak electricity consumption areas, the second layer of dynamic adjustment content is determined. This second layer of dynamic adjustment content is a specification of the strategy logic layer, defining in what scenarios, when, and how the technical tools defined in the first layer of content are used. It includes the adjustment objectives, constraints, priorities, and execution logic. The system will instantiate this in a strategy rule base by combining project objectives, regional characteristics (such as load curves and user types), and system operation constraints. For example, for a commutation strategy project, the second layer of content includes: setting an execution window (such as off-peak hours at night), defining trigger conditions (such as imbalance exceeding a threshold continuously), clarifying adjustment objectives (such as reducing imbalance to below 5%), and setting priority and constraint rules (such as avoiding frequent operations on the same user). The second layer of content is a series of if-then-else rules and strategy parameters, which endow the system with the ability to make intelligent decisions, rather than simply executing blindly.

[0103] By organically combining technical tools (the first layer of content) and strategic logic (the second layer of content), a complete dynamic adjustment system is constructed. The dynamic adjustment system is an adaptive closed-loop control system. Its internal structure typically includes: a sensing module that continuously monitors key electrical quantities, a decision engine with a built-in strategy rule base, an execution interface for calling technical resources, and a feedback and learning module for evaluating the adjustment effect and fine-tuning strategy parameters. The completed dynamic adjustment system is deployed to the corresponding peak electricity consumption area controller, where it will run as a long-term program to continuously and autonomously manage the power quality, economy, and reliability of the area.

[0104] Please see Figure 7 , Figure 7 This is a schematic diagram of the structural composition of a flexible adjustment device for a power distribution line based on intelligent power distribution in an embodiment of the present invention; the flexible adjustment device for the power distribution line based on intelligent power distribution includes: Peak electricity consumption feature module 21 is used to collect current data of each group of power distribution lines, determine peak electricity consumption areas based on the current data of each group of power distribution lines and the corresponding current detection locations, and determine multiple peak electricity consumption features based on each peak electricity consumption area and the corresponding peak electricity consumption time. The peak control mode module 22 is used to determine multiple power distribution control paths based on the characteristic patterns of multiple peak electricity consumption and power distribution lines, and to determine the corresponding peak control mode based on the control scope, corresponding control content and power consumption status of the corresponding peak electricity consumption area of ​​each power distribution control path. The flexible adjustment event module 23 is used to determine the control effect level based on the control content, corresponding control dimension and voltage fluctuation event corresponding to the power distribution line of the peak control method, and to determine the flexible adjustment event of the peak power consumption area based on each control effect level, the power change data of the corresponding peak power consumption area and the corresponding current power consumption situation. The power distribution status module 24 is used to determine the corresponding power control mode based on multiple flexible control items and peak power consumption areas based on flexible control events, and to determine the power distribution status of each peak power consumption area based on the power control mode, the current power consumption situation of each peak power consumption area and the current time. The dynamic adjustment system module 25 is used to determine the intelligent power distribution events of each peak power consumption area based on the power distribution status and corresponding power loss of each peak power consumption area, and to determine the dynamic adjustment system of each peak power consumption area based on the intelligent power distribution events, power distribution lines and each peak power consumption area.

[0105] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

Claims

1. A flexible regulation method for power distribution lines based on intelligent power distribution, characterized in that, include: Collect current data for each set of power distribution lines, determine peak electricity consumption areas based on the current data and corresponding current detection locations, and determine multiple peak electricity consumption characteristics based on each peak electricity consumption area and corresponding peak electricity consumption time. Multiple power distribution control paths are determined based on the characteristic patterns of multiple peak electricity consumption patterns and power distribution lines. Based on the control scope, corresponding control content, and power consumption status of the corresponding peak electricity consumption area of ​​each power distribution control path, the corresponding peak control method is determined. Based on the control content, corresponding control dimensions, and voltage fluctuation events corresponding to the distribution lines of this peak control method, the control effect level is determined. Based on each control effect level, the corresponding power consumption change data of the peak power consumption area, and the corresponding current power consumption situation, the flexible adjustment event of the peak power consumption area is determined. The flexible adjustment event includes a unique event ID, a clear target area, specific triggering conditions, quantified adjustment target, and an ordered execution sequence. Based on multiple flexible regulation projects and peak electricity consumption areas based on flexible regulation events, the corresponding power control mode is determined. Based on the power control mode, the current power consumption situation of each peak electricity consumption area and the current time, the power distribution status of each peak electricity consumption area is determined. Based on the power distribution status and corresponding power loss of each peak power consumption area, intelligent power distribution events for each peak power consumption area are determined. Based on these intelligent power distribution events, power distribution lines, and each peak power consumption area, a dynamic adjustment system for each peak power consumption area is determined. The dynamic adjustment system is an adaptive closed-loop control system.

2. The flexible regulation method for power distribution lines based on intelligent power distribution according to claim 1, characterized in that, The process involves collecting current data from various sets of power distribution lines, determining peak electricity consumption areas based on the current data and corresponding current detection locations, and identifying multiple peak electricity consumption characteristics based on each peak area and its corresponding peak time. These characteristics include: Real-time monitoring of power distribution lines and collection of power distribution line database; determination of each group of current data of power distribution lines based on the detection of the power distribution line database; determination of peak electricity consumption areas based on each group of current data, the corresponding current detection location and power distribution line. Mark the location of each peak electricity consumption area, determine the first electricity consumption combination based on the location of the peak electricity consumption area and the corresponding peak electricity consumption time, and determine the second electricity consumption combination based on the regional shape of the peak electricity consumption area and the corresponding peak electricity consumption time; Multiple peak electricity consumption characteristics are determined based on the matching of the first and second electricity consumption combinations. These characteristics are distributed in different spaces within the corresponding peak electricity consumption regions, thus revealing the peak electricity consumption characteristics of those regions.

3. The flexible regulation method for power distribution lines based on intelligent power distribution according to claim 1, characterized in that, The process involves determining multiple power distribution control paths based on the characteristic patterns of multiple peak electricity consumption patterns and power distribution lines. Based on the control scope, corresponding control content, and power consumption status of the corresponding peak electricity consumption area for each control path, the corresponding peak control method is determined, including: Real-time monitoring of multiple peak electricity consumption characteristics, marking the feature patterns of multiple peak electricity consumption characteristics, and determining multiple power distribution control paths based on the feature patterns of multiple peak electricity consumption characteristics, the corresponding peak electricity consumption areas and power distribution lines; In multiple power distribution control paths, the corresponding power distribution control range is determined based on the detection of each power distribution control path, and the control content of each power distribution control path is marked. At the same time, the power consumption status of the peak power consumption area corresponding to the power distribution control path is collected. Based on the control range of each power distribution control path, the corresponding control content, and the power consumption status of the corresponding peak power consumption area, the corresponding peak control method is determined.

4. The flexible regulation method for power distribution lines based on intelligent power distribution according to claim 1, characterized in that, The control effectiveness level is determined based on the control content, corresponding control dimensions, and voltage fluctuation events corresponding to the distribution lines according to the peak control method. The flexible adjustment events for the peak electricity consumption area are determined based on each control effectiveness level, the corresponding electricity consumption change data for the peak area, and the corresponding current electricity consumption situation. These events include: Based on the identification of the peak control method, the corresponding control content is determined. Based on the control content of the peak control method and the corresponding peak electricity consumption area, the control dimension corresponding to the peak control method is determined. At the same time, a voltage dynamic map of the distribution line is constructed based on multiple voltage data of the distribution line. Based on the voltage dynamic map of the distribution line and multiple peak electricity consumption areas, the voltage fluctuation events corresponding to the distribution line are determined. The first control effect coefficient is determined based on the control content of the peak control method and the voltage fluctuation events corresponding to the power distribution lines. The second control effect coefficient is determined based on the control dimension of the peak control method and the voltage fluctuation events corresponding to the power distribution lines. The control effect level is determined based on the mapping relationship between the first control effect coefficient, the second control effect coefficient, and the control effect level.

5. The flexible adjustment method for power distribution lines based on intelligent power distribution according to claim 4, characterized in that, The method for determining the control effectiveness level based on the control content, corresponding control dimensions, and voltage fluctuation events corresponding to the distribution lines, and determining the flexible adjustment events for the peak electricity consumption area based on each control effectiveness level, the corresponding electricity consumption change data for the peak electricity consumption area, and the corresponding current electricity consumption situation, also includes: Based on the electricity consumption detection of each peak electricity consumption area, the corresponding electricity change data is determined. The first flexible adjustment content is determined according to each control effect level and the corresponding electricity change data of the peak electricity consumption area. The second flexible adjustment content is determined according to each control effect level and the current electricity consumption of the corresponding peak electricity consumption area. The flexible adjustment event of the peak electricity consumption area is determined based on the first and second flexible adjustment content.

6. The flexible regulation method for power distribution lines based on intelligent power distribution according to claim 1, characterized in that, The system determines corresponding power control modes based on multiple flexible control items and peak electricity consumption areas based on flexible control events. Based on these power control modes, the current electricity consumption situation in each peak electricity consumption area, and the current time, the power distribution status of each peak electricity consumption area is determined, including: Collect flexible regulation events, identify multiple flexible regulation items based on the identification of flexible regulation events, and determine the corresponding flexible regulation level according to the item content of multiple flexible regulation items, the location of the corresponding peak electricity consumption area and the current electricity consumption situation; The first power regulation coefficient is determined based on the project content of multiple flexible regulation projects and the regional shape of peak electricity consumption areas. The second power regulation coefficient is determined based on the flexible regulation level of multiple flexible regulation projects and the regional shape of peak electricity consumption areas. The corresponding power regulation mode is determined based on the mapping relationship between the first power regulation coefficient, the second power regulation coefficient, and the power regulation mode.

7. The flexible regulation method for power distribution lines based on intelligent power distribution according to claim 6, characterized in that, The method for determining corresponding power control modes based on multiple flexible control items and peak electricity consumption areas based on flexible control events, and determining the power distribution status of each peak electricity consumption area based on this power control mode, the current electricity consumption situation of each peak electricity consumption area, and the current time, further includes: Collect the current electricity consumption data for each peak electricity consumption area, determine the power distribution content for each peak electricity consumption area based on the current electricity consumption data and the power control mode, and determine the power distribution status for each peak electricity consumption area based on the power distribution content, the corresponding flexible adjustment events, and the time.

8. The flexible regulation method for power distribution lines based on intelligent power distribution according to claim 1, characterized in that, The process involves determining intelligent power distribution events for each peak electricity consumption area based on its power distribution status and corresponding power loss, and then determining a dynamic adjustment system for each peak electricity consumption area based on these intelligent power distribution events, power distribution lines, and the specific peak electricity consumption areas. This includes: In multiple peak electricity consumption areas, multiple power loss data are determined based on the detection of each peak electricity consumption area. The power loss situation in the peak electricity consumption area is determined based on the multiple power loss data, time and the corresponding regional location of the peak electricity consumption area. Collect the overall power supply of the power distribution lines, and determine the smart power distribution events for each peak power consumption area based on the power distribution status, corresponding power loss, and overall power supply of the power distribution lines; identify multiple smart power distribution projects based on the identification of these smart power distribution events.

9. The flexible regulation method for power distribution lines based on intelligent power distribution according to claim 8, characterized in that, The method of determining intelligent power distribution events for each peak electricity consumption area based on the power distribution status and corresponding power loss, and determining the dynamic adjustment system for each peak electricity consumption area based on the intelligent power distribution events, power distribution lines, and each peak electricity consumption area, further includes: The first level of dynamic regulation is determined based on multiple smart power distribution projects and power lines. The second level of dynamic regulation is determined based on multiple smart power distribution projects and various peak electricity consumption areas. The dynamic regulation system for each peak electricity consumption area is determined based on the first and second levels of dynamic regulation.

10. A flexible regulating device for a power distribution line based on intelligent power distribution, characterized in that, The flexible adjustment device for the power distribution line based on intelligent power distribution is applied to the flexible adjustment method for the power distribution line based on intelligent power distribution as described in any one of claims 1-9.

Images