A method and apparatus for energy control of a power grid based on an energy monitoring platform

Abstract

This invention discloses a power grid energy control method and apparatus based on a power monitoring platform. The invention relates to the technical field of energy control methods. It determines the energy control coefficient of an energy region based on multiple sub-energy control characteristics and corresponding energy load curves. It then determines power grid energy control events based on the location of each energy region, its corresponding energy control coefficient, and the current operating state of the power grid, thus improving the accuracy of power grid energy control events. Therefore, multiple sub-energy control contents are determined based on the identification of power grid energy control events. Abnormal maintenance events for that energy region are determined based on these sub-energy control contents, the corresponding energy regions, and the corresponding energy anomaly events. Finally, the power grid energy control level is determined based on the abnormal maintenance events for each energy region, the current operating state of the power grid, and the power grid's energy fluctuation range, thereby improving the accuracy of the power grid energy control level.

Description

Technical Field

[0001] This invention relates to the technical field of energy control methods, and more particularly to an energy control method and apparatus for a power grid based on an energy monitoring platform. Background Technology

[0002] In power grid technology, the power grid is gradually being applied to people's lives, supplying power to residential or industrial areas. The power monitoring platform comprehensively, precisely, and intelligently senses and manages the operating status of the power system (from the power grid to the user side). In existing technologies, the power monitoring platform collects multiple working data from the power grid and determines the corresponding energy area based on these data. However, it ignores the energy dynamics of the energy area and the corresponding energy control coefficient, which affects the accuracy of the power grid's energy control events and results in a low accuracy of the power grid's energy control level. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of the prior art. This invention provides a power grid energy control method and device based on an energy monitoring platform.

[0004] This invention provides an energy control method for a power grid based on an energy monitoring platform, comprising:

[0005] During the process of monitoring the power grid on the power monitoring platform, multiple energy regions of the power grid are determined based on multiple working data of the power grid and the power distribution map of the power grid.

[0006] Based on the regional location of each energy region, the corresponding working data combination, and the surrounding environmental data combination, the energy dynamic map of the energy region is determined, and multiple sub-energy controlled features are determined based on the identification of the energy dynamic map of the energy region.

[0007] Based on the load detection of the energy region, the corresponding energy load curve is determined. Based on the multiple sub-energy control characteristics of the energy region and the corresponding energy load curve, the energy control coefficient of the energy region is determined. Based on the regional location of each energy region, the corresponding energy control coefficient, and the current operating state of the power grid, the energy control event of the power grid is determined.

[0008] Multiple abnormal energy data are identified by combining working data from various energy regions. The corresponding energy anomaly event is determined based on the location of the multiple abnormal energy data, the corresponding energy region, and the current working mode of the power grid.

[0009] Multiple sub-energy control contents are determined based on the identification of energy control events in the power grid. Abnormal maintenance events in the energy area are determined based on the multiple sub-energy control contents, the corresponding energy areas, and the corresponding abnormal energy events. The energy control level of the power grid is determined based on the abnormal maintenance events in each energy area, the current operating status of the power grid, and the energy fluctuation range of the power grid.

[0010] This invention provides an energy control device for a power grid based on a power monitoring platform. The energy control device is applied to the aforementioned energy control method for a power grid based on a power monitoring platform. The energy control device includes:

[0011] The energy zone module is used to determine multiple energy zones of the power grid based on multiple working data of the power grid and the power distribution map of the power grid during the monitoring of the power grid by the power monitoring platform.

[0012] The sub-energy controlled feature module is used to determine the energy dynamic map of each energy region based on its regional location, the corresponding working data combination, and the surrounding environmental data combination, and to determine multiple sub-energy controlled features based on the identification of the energy dynamic map of the energy region.

[0013] The energy control event module is used to determine the corresponding energy load curve based on the load detection of the energy region, determine the energy control coefficient of the energy region based on the multiple sub-energy control characteristics of the energy region and the corresponding energy load curve, and determine the energy control event of the power grid based on the regional location of each energy region, the corresponding energy control coefficient and the current operating state of the power grid.

[0014] The energy anomaly event module is used to identify multiple abnormal energy data based on the detection of working data combinations from various energy regions, and to determine the corresponding energy anomaly event based on the location of the multiple abnormal energy data, the corresponding energy region, and the current working mode of the power grid.

[0015] The energy control level module is used to determine multiple sub-energy control contents based on the identification of energy control events in the power grid, determine the abnormal maintenance events of the energy area based on the multiple sub-energy control contents, the corresponding energy area and the corresponding energy abnormal events, and determine the energy control level of the power grid based on the abnormal maintenance events of each energy area, the current operating status of the power grid and the energy fluctuation range of the power grid.

[0016] Compared with the prior art, the beneficial effects of the present invention are:

[0017] In this embodiment of the invention, the method described herein, during the monitoring of the power grid by the power monitoring platform, determines multiple energy regions of the power grid based on multiple operating data and the power distribution map of the power grid; determines the energy dynamic map of each energy region based on its regional location, corresponding operating data combination, and surrounding environmental data combination; identifies multiple sub-energy controlled features based on the identification of the energy dynamic map of the energy region; determines the corresponding energy load curve based on the load detection of the energy region; determines the energy control coefficient of the energy region based on the multiple sub-energy controlled features and the corresponding energy load curve; and determines the power grid energy control events based on the regional location of each energy region, the corresponding energy control coefficient, and the current operating state of the power grid. The introduction of the energy dynamic map of the energy region further controls multiple sub-energy controlled features, incorporating a holistic consideration of the regional location of each energy region, the corresponding energy control coefficient, and the current operating state of the power grid, thereby improving the accuracy of the power grid energy control events.

[0018] Therefore, multiple abnormal energy data are identified based on the detection of combined working data from various energy regions. Corresponding energy anomaly events are determined based on the location of these abnormal energy data, the corresponding energy region, and the current operating mode of the power grid. Multiple sub-energy control contents are identified based on the recognition of power grid energy control events. Abnormal maintenance events for that energy region are determined based on these sub-energy control contents, the corresponding energy region, and the corresponding energy anomaly events. The power grid's energy control level is determined based on the abnormal maintenance events in each energy region, the current operating status of the power grid, and the energy fluctuation range of the power grid. By introducing energy anomaly events, the abnormal maintenance events in that energy region are further controlled. This achieves a holistic consideration of abnormal maintenance events in each energy region, the current operating status of the power grid, and the energy fluctuation range of the power grid, thereby improving the accuracy of the power grid's energy control level. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating the power grid energy control method based on a power monitoring platform in an embodiment of the present invention.

[0020] Figure 2 This is a flowchart illustrating step S11 in the power grid energy control method based on a power monitoring platform according to an embodiment of the present invention.

[0021] Figure 3 This is a flowchart illustrating step S12 in the power grid energy control method based on a power monitoring platform according to an embodiment of the present invention.

[0022] Figure 4 This is a flowchart illustrating step S13 in the power grid energy control method based on a power monitoring platform according to an embodiment of the present invention.

[0023] Figure 5 This is a flowchart illustrating step S14 in the power grid energy control method based on a power monitoring platform according to an embodiment of the present invention.

[0024] Figure 6 This is a flowchart illustrating step S15 in the power grid energy control method based on a power monitoring platform according to an embodiment of the present invention.

[0025] Figure 7 This is a schematic diagram of the structural composition of the power grid energy control device based on the power monitoring platform in an embodiment of the present invention. Detailed Implementation

[0026] 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.

[0027] Please see Figures 1 to 7 An energy control method for a power grid based on a power monitoring platform, applied to energy control scenarios; the energy control method for a power grid based on a power monitoring platform includes:

[0028] Step S11: During the process of monitoring the power grid on the power monitoring platform, multiple energy regions of the power grid are determined based on multiple working data of the power grid and the power distribution map of the power grid;

[0029] Step S12: Determine the energy dynamic map of each energy region based on its regional location, corresponding working data combination, and surrounding environmental data combination; and determine multiple sub-energy controlled features based on the identification of the energy dynamic map of the energy region.

[0030] Step S13: Determine the corresponding energy load curve based on the load detection of the energy region, determine the energy control coefficient of the energy region according to the multiple sub-energy control characteristics of the energy region and the corresponding energy load curve, and determine the energy control event of the power grid according to the regional location of each energy region, the corresponding energy control coefficient and the current operating state of the power grid;

[0031] Step S14: Based on the detection of the combination of working data from each energy region, determine multiple abnormal energy data, and determine the corresponding energy anomaly event according to the location of the multiple abnormal energy data, the corresponding energy region, and the current working mode of the power grid;

[0032] Step S15: Based on the identification of energy control events in the power grid, determine multiple sub-energy control contents; based on the multiple sub-energy control contents, the corresponding energy regions, and the corresponding energy anomaly events, determine the abnormal maintenance events of the energy regions; based on the abnormal maintenance events of each energy region, the current operating status of the power grid, and the energy fluctuation range of the power grid, determine the energy control level of the power grid.

[0033] refer to Figure 2 In step S11, the specific steps are as follows:

[0034] S111: The power monitoring platform communicates with the power grid and performs dynamic monitoring of the power grid; it monitors the power grid monitoring process of the power monitoring platform in real time; it collects multiple working data of the power grid, determines multiple data combinations based on the multiple working data of the power grid and the overall shape of the power grid, and determines the corresponding power distribution area based on the identification of each data combination;

[0035] S112: Collect data on multiple power distribution areas and mark the power distribution in each area. Based on the location of each power distribution area, the corresponding power distribution, and the power distribution map of the power grid, determine multiple energy regions of the power grid.

[0036] In the embodiments of this application, the power monitoring platform communicates with the power grid and performs dynamic monitoring of the power grid; the monitoring process of the power monitoring platform on the power grid is monitored in real time; multiple working data of the power grid are collected, and multiple data combinations are determined based on the multiple working data of the power grid and the overall shape of the power grid. Based on the identification of each data combination, the corresponding power distribution area is determined, which is compatible with the overall consideration of the identification of each data combination and ensures the accuracy of the corresponding power distribution area.

[0037] At this point, at the communication level, the platform needs to establish a reliable connection with smart devices at each node of the power grid through industrial-grade protocols (such as IEC61850, Modbus TCP / IP). Whether it is wired fiber optic or wireless 5G, low latency and high reliability of data transmission must be ensured. At the dynamic monitoring level, the platform works continuously. It continuously collects the operating parameters of the power grid at a set frequency. The purpose is to capture dynamic behaviors such as load changes, voltage fluctuations, frequency deviations, and the randomness of new energy output. At the same time, the platform itself will also perform monitoring.

[0038] The platform's built-in diagnostic module continuously monitors data integrity, checks for lost or out-of-order data packets, assesses data timeliness, calculates whether transmission latency exceeds limits, monitors device status, confirms whether remote devices are online, and evaluates link health to predict potential communication interruption risks.

[0039] It collects multiple operational data from the power grid, including not only steady-state data such as three-phase voltage / current, active / reactive power, and frequency, but also power quality data such as voltage flicker, harmonics, and three-phase imbalance, as well as status data such as switch opening and closing status and protection action signals. Each data point has a precise timestamp and geographic location tag, providing a foundation for subsequent data aggregation.

[0040] The process of data aggregation and feature extraction transforms isolated data points into meaningful information. The platform combines data points based on the power grid topology model using clustering algorithms or rule engines. For example, it aggregates data from the same feeder into feeder data sets, or aggregates user data with similar load curves, or combines monitoring point data with highly correlated power quality indicators, thereby discovering the inherent correlation between data.

[0041] The platform will match each data combination with a preset feature library based on its characteristics and assign it a functional label, thereby forming a power distribution area with clear physical meaning. For example, a data combination with high harmonics and three-phase imbalance will be identified as an industrial load power distribution area; a data combination with bi-peak load characteristics will be identified as a residential load power distribution area; and a data combination with power fluctuating with light intensity will be identified as a new energy access power distribution area.

[0042] Furthermore, multiple power distribution areas are collected and the power distribution of each power distribution area is marked. Based on the location of each power distribution area, the corresponding power distribution, and the power distribution map of the power grid, multiple energy regions of the power grid are determined. This approach takes into account the overall consideration of the location of each power distribution area, the corresponding power distribution, and the power distribution map of the power grid, ensuring the accuracy of multiple energy regions of the power grid.

[0043] At this point, the platform will load multiple logical entities of power distribution areas that have been generated by S111. Each entity contains a set of associated monitoring point IDs, real-time data streams, and preliminary functional tags. The platform will create a list of areas and establish an index for each entry pointing to its underlying data source.

[0044] The platform calculates and labels a series of key operational characteristics for each region, forming a multi-dimensional feature vector. These labels are very rich in content, including load characteristics (such as peak-to-valley difference and load density), power quality characteristics (such as harmonic distortion rate and three-phase imbalance), power source characteristics (such as distributed power source type and output fluctuation), and network characteristics (such as line type and short-circuit capacity). The abstract region is transformed into a quantifiable and comparable digital twin model.

[0045] Based on the regional location of each power distribution area, the corresponding power distribution situation, and the power distribution map of the power grid, multiple energy regions of the power grid are determined. The key steps of spatially integrating the logical region with the physical power grid and refining the boundaries form energy regions for fine-grained control. The platform integrates regional location, detailed feature vectors, and accurate GIS topology maps, and uses complex algorithms to determine the energy regions. The algorithm refines the spatial boundaries, delineates the physical power supply boundaries along the actual feeder branches, performs homogeneity verification to ensure consistent features within the boundaries, and conducts control accessibility analysis to assess whether each region has independent control means. The output energy region is an execution unit with clear boundaries, well-defined features, and control potential.

[0046] refer to Figure 3 In step S12, the specific steps are as follows:

[0047] S121: Collect data from each energy region, determine the corresponding region location based on the detection of each energy region, collect multiple working data in the region location of each energy region, and construct a corresponding working data combination based on the multiple working data.

[0048] S122: Based on the regional location of each energy region, multiple surrounding spaces are determined. In the multiple surrounding spaces, the combination of environmental data is determined according to the spatial location of each surrounding space, the corresponding multiple environmental data, and the spatial morphology of each surrounding space.

[0049] S123: Determine the first energy change map based on the regional location of each energy region and the corresponding working data combination; determine the second energy change map based on the regional location of each energy region and the surrounding environmental data combination; and determine the energy dynamic map of the energy region based on the first and second energy change maps.

[0050] S124: Based on the detection of the energy dynamic map of the energy region, multiple energy dynamic regions are determined, and the corresponding sub-energy controlled features are determined according to the identification of each energy dynamic region, so as to collect multiple sub-energy controlled features.

[0051] In the embodiments of this application, each energy region is collected, and the corresponding region location is determined based on the detection of each energy region. In the region location of each energy region, multiple working data are collected, and a corresponding working data combination is constructed based on the multiple working data. This approach takes into account the overall consideration of the detection of each energy region and ensures the accuracy of the corresponding region location.

[0052] At this point, the platform acquires the metadata of each energy zone defined in step S11; the platform collects the unique ID of each zone and the list of devices that define its boundaries; the platform determines the composite location of the zone by parsing and verifying this metadata. This location includes two levels: one is the electrical connection relationship in the power grid topology model, i.e., the topological location; the other is the set of geographic coordinates in the GIS system, i.e., the geographic spatial location. This process ensures that each energy zone is a unit that is clearly defined both logically and physically.

[0053] After defining the area's location and boundaries, the platform precisely locates all data collection points within that area, including outgoing switches of the upstream substation, feeder terminal units (FTUs), transformer terminal units (TTUs) within the area, and smart meters of key users. The platform initiates data polling or subscribes to real-time data streams from these devices through pre-set communication links to collect multi-dimensional operational data, such as three-phase voltage / current, active / reactive power, power quality parameters, and switch status quantities. To ensure data consistency, all data is given a unified timestamp and aligned.

[0054] The platform will organize the collected, time-synchronized working data into a structured form according to energy regions, forming a high-dimensional vector or time series data frame. This working data combination is a complete digital snapshot of the energy region at a certain moment. It not only contains all the key measurement values ​​in the region, but also encapsulates metadata such as region ID, timestamp, and topological location. This structured data combination provides high-quality input for subsequent dynamic analysis and feature extraction.

[0055] Specifically, the power monitoring platform has identified the photovoltaic energy storage zone of the power grid; the power monitoring platform retrieved the photovoltaic energy storage zone definition with ID ER_PV_Storage from the database; the platform determined its topological location as the F5 feeder on the 10kV section bus of the 110kV substation of the power grid by detecting its boundary device list; at the same time, by querying the GIS system, the platform drew the geographic spatial location of the area, which is a polygonal area covering approximately 2 square kilometers.

[0056] Based on the equipment list of the F5 feeder, the power monitoring platform initiated data collection from multiple devices in the area. These devices included the FTU at the F5 feeder outlet, the EMS system of the 500kW / 1000kWh energy storage power station, the grid-connected inverter of the 2MWp photovoltaic power station, the TTUs on the three public distribution transformers, and the smart meters of 10 representative industrial and commercial users. At 10:00:00, the power monitoring platform collected a set of synchronous data, such as photovoltaic output of 1.8MW, energy storage SOC of 60%, total regional load of 1.0MW, and feeder outlet voltage of 10.4kV.

[0057] The power monitoring platform encapsulates all the above data points into a structured working data set with 10:00:00 as the unified timestamp. This data set clearly reveals that at exactly 10:00, a net power surplus of 0.8MW (1.8MW photovoltaic - 1.0MW load) was generated within the photovoltaic energy storage zone. This surplus is being fed back to the upstream grid, leading to an increase in the regional voltage. This accurate and comprehensive digital snapshot lays a solid foundation for the power monitoring platform to further analyze the energy dynamics of the region and extract control characteristics.

[0058] Furthermore, multiple surrounding spaces are determined based on the surrounding detection of the regional location of each energy region. In these multiple surrounding spaces, environmental data combinations are determined according to the spatial location of each surrounding space, the corresponding multiple environmental data, and the spatial morphology of each surrounding space. This approach takes into account the overall consideration of the spatial location of each surrounding space, the corresponding multiple environmental data, and the spatial morphology of each surrounding space, ensuring the accuracy of the environmental data combinations.

[0059] At this point, the platform will use spatial analysis algorithms to expand one or more dynamic buffer zones outward based on the geographical boundaries of the energy region. These buffer zones are called the surrounding space, and their size and shape are not fixed. Instead, they are dynamically determined according to the type of energy region (such as a new energy region that requires a larger area) and the propagation path of key environmental factors (such as extending along the wind direction). The platform can even define multiple surrounding spaces of different scales to conduct multi-scale environmental analysis.

[0060] After determining the surrounding space, the platform initiates query and collection requests to multiple heterogeneous data sources based on its geographical coordinate range. These data sources include meteorological service APIs, GIS databases, satellite remote sensing data, and socio-economic information systems. Since the collected environmental data have different spatiotemporal resolutions, the platform needs to perform strict spatiotemporal alignment processing to ensure that all data are on a unified timestamp and spatial grid. The platform then structures and encapsulates the aligned multi-source environmental data according to its surrounding space to form a comprehensive environmental data combination that describes the overall state of the external environment.

[0061] Specifically, the power monitoring platform is analyzing the photovoltaic and energy storage area of ​​the power grid; the power monitoring platform conducts surrounding detection with the geographical boundary of the photovoltaic and energy storage area as the center; considering that the core of this area is photovoltaic power generation, the platform defines two surrounding spaces: one is a circular near-field meteorological space with a radius of 2 kilometers, used to capture local micro-meteorological changes; the other is a 5-kilometer × 3-kilometer rectangular cloud trajectory space delineated according to the historical cloud movement path, used to predict the impact of upcoming cloud clusters.

[0062] The power monitoring platform constructed environmental data combinations for these two surrounding spaces respectively. For the near-field meteorological space, the platform queried the internal micro-weather stations, collected data such as temperature, humidity, wind speed and surface type, and packaged them into environmental data combination A. For the cloud trajectory space, the platform called the meteorological radar and satellite cloud image API to collect key information such as cloud coverage, movement speed, direction and expected time to reach the energy zone boundary, and packaged it into environmental data combination B.

[0063] The power monitoring platform not only grasps the internal operating status of the photovoltaic energy storage area, but also accurately captures the current status and future trends of its external environment. In particular, the key disturbance information revealed by environmental data combination B, which indicates that cloud cover will occur in 15 minutes, provides a crucial data foundation for the platform to predict power fluctuations and formulate energy storage charging and discharging strategies in advance. It is a key link in realizing the transformation from passive response to active predictive control.

[0064] Furthermore, a first-level energy change map is determined based on the regional location of each energy region and the corresponding combination of working data. A second-level energy change map is determined based on the regional location of each energy region and the combination of surrounding environmental data. An energy dynamic map for the energy region is determined based on the first-level and second-level energy change maps, which takes into account the overall consideration of the first-level and second-level energy change maps and ensures the accuracy of the energy dynamic map of the energy region.

[0065] At this point, the first energy change map is determined based on the regional location of each energy region and the corresponding working data combination. The aim is to construct a dynamic model that reflects the electrical behavior inside the energy region, which can be understood as a self-portrait of the region. The platform mainly uses the working data combination generated by S121 and uses power system analysis techniques such as state estimation, time series analysis and power flow calculation to generate this map. The output first energy change map is a multi-dimensional dynamic model that accurately describes the energy flow, voltage distribution, power quality level and its time-varying laws within the region.

[0066] The second energy change map is determined based on the regional location of each energy region and the combination of surrounding environmental data. The aim is to construct a dynamic model that reflects how external environmental factors affect the energy region, i.e., a regional environmental response profile. The platform mainly uses the combination of environmental data generated by S122 to establish the mapping relationship between the environment and energy state through correlation analysis, machine learning or physical modeling. The output second energy change map is a causal relationship model that reveals how changes in the external environment will drive or affect changes in the regional energy state.

[0067] By combining internal behavior with external influences, a comprehensive dynamic model with predictive and explanatory capabilities is formed. The platform adopts multimodal data fusion technologies, such as Kalman filtering, model fusion, or deep learning fusion, to integrate the first two graphs. The energy dynamic graph is a high-fidelity, predictable dynamic system model that not only describes the current and past energy states, but more importantly, it proactively predicts future energy states (such as voltage over-limit risks, power fluctuation amplitudes, etc.) based on environmental changes, providing a basis for decision-making in active control.

[0068] Specifically, the power monitoring platform is analyzing the solar-storage energy zone of the power grid. Using the working data combination of S121 (1.8MW photovoltaic output, 1.0MW load, 10.4kV voltage), the platform, through state estimation and power flow calculation, found that the voltage of all nodes in this zone is slightly high, and there is a net power of 0.8MW being fed upstream. Therefore, the generated first energy change map is described as follows: the current zone is in a power backflow state, the voltage is generally high, the system is stable but there is a risk of exceeding limits.

[0069] The power monitoring platform uses environmental data combination B from S122 (the cloud will arrive in 15 minutes) to call the pre-trained photovoltaic power output prediction model and predicts that at T+15 minutes, the photovoltaic power output will drop sharply from 1.8MW to 0.3MW. Therefore, the generated second energy change map is described as: a strong external disturbance (cloud shading) is about to occur, which will cause the power output in the region to decrease by 1.5MW in a short period of time.

[0070] The power monitoring platform uses a model fusion method to input the prediction results (-1.5MW output) of the second power flow model into the power flow model of the first power flow model, and performs a forward-looking power flow calculation. The calculation results show that at T+15 minutes, the region will experience a power deficit of 0.5MW, and the voltage will drop sharply from 10.4kV to 9.8kV, triggering low voltage protection. The generated energy dynamic diagram gives a dynamic prediction: the current voltage is high, but a severe power deficit is expected to occur in 15 minutes, resulting in a sharp voltage drop and a risk of low voltage.

[0071] Therefore, multiple energy dynamic regions are determined based on the detection of the energy dynamic map of the energy region, and corresponding sub-energy controlled features are determined according to the identification of each energy dynamic region, so as to collect multiple sub-energy controlled features, which takes into account the overall consideration of the identification of each energy dynamic region and ensures the accuracy of the corresponding sub-energy controlled features.

[0072] At this point, the complex energy dynamic map generated by S123 is decomposed into patterns to identify sub-regions with significantly different dynamic behaviors. This division is no longer based on physical location, but purely on dynamic behavior patterns. The platform treats the dynamic map as multi-dimensional time-series data and applies methods such as clustering algorithms, change point detection, or frequency domain analysis to cluster data points or time periods with similar change patterns into one category. Each clustering result is an energy dynamic region, such as a voltage-sensitive dynamic region or a power-fluctuation dynamic region.

[0073] The platform will conduct in-depth analysis of the data contained in each energy dynamic region, extracting its most core, quantifiable, and control-related features. These features include time-domain features (such as voltage fluctuation amplitude and rate of change), frequency-domain features (such as dominant frequency and harmonic amplitude), and model parameters (such as damping ratio). The output sub-energy controlled features are a structured feature vector, providing a precise basis for the formulation of control strategies.

[0074] The platform aggregates and manages the sub-energy controlled features of all energy dynamic regions to form a complete control feature profile of that energy region. The platform then creates a feature set, encapsulates all the features extracted in the previous step, and attaches metadata such as the dynamic region ID and timestamp. It outputs a comprehensive, multi-dimensional feature list as direct input for the next step, providing accurate and quantitative basis for the formulation of control strategies.

[0075] Specifically, the power monitoring platform has generated a dynamic energy map for the photovoltaic energy storage area, predicting that the voltage will rise first and then fall. The power monitoring platform performs pattern detection on this dynamic map, which includes the prediction of the voltage rise and fall. Through change point detection and cluster analysis, the platform identifies two energy dynamic regions that are sequential in time but have different behaviors: Energy dynamic region 1, whose behavior is a slow unidirectional rise in voltage caused by power surplus; and Energy dynamic region 2, whose behavior is a rapid drop and oscillation in voltage caused by power deficit.

[0076] The power monitoring platform extracted specific control features for these two dynamic regions respectively. For energy dynamic region 1, the extracted features are: {Feature type: voltage limit exceeded, current voltage deviation: +4.0%, control resource: energy storage rechargeable power 500kW}. For energy dynamic region 2, three features were extracted: {Feature type: power deficit, predicted deficit size: 1.5MW}, {Feature type: voltage drop, predicted minimum voltage: 9.8kV}, and {Feature type: voltage oscillation, predicted oscillation frequency: 1.2Hz}.

[0077] The power monitoring platform summarizes the above four characteristics to form a complete set of control characteristics for the photovoltaic energy storage area at the current moment. This set of characteristics clearly tells the control decision module that the area faces a two-stage dynamic problem. The first stage requires voltage suppression, and the second stage requires power and voltage support, and there is also a risk of voltage oscillation.

[0078] refer to Figure 4 In step S13, the specific steps are as follows:

[0079] S131: Collect data on the energy region and perform load detection on the energy region. Determine multiple energy load data based on the load detection of the energy region. Determine the corresponding energy load curve based on the multiple energy load data and the energy fluctuation range corresponding to the energy region.

[0080] S132: Based on the identification of energy load curves, multiple energy load nodes are identified, and the energy control coefficient of the energy region is determined according to the node positions, corresponding node shapes, and multiple sub-energy control characteristics of the energy region.

[0081] S133: Collect the current operating status of the power grid, determine the first level of energy control content based on the current operating status of the power grid and the regional location of each energy region, determine the second level of energy control content based on the current operating status of the power grid and the energy control coefficient of each energy region, and determine the energy control event of the power grid based on the first level of energy control content and the second level of energy control content.

[0082] In the embodiments of this application, the energy region is collected and the load of the energy region is detected. Multiple energy load data are determined based on the load detection of the energy region. The corresponding energy load curve is determined based on the multiple energy load data and the energy fluctuation range corresponding to the energy region. This approach takes into account the overall consideration of multiple energy load data and the energy fluctuation range corresponding to the energy region, ensuring the accuracy of the corresponding energy load curve.

[0083] At this point, the platform locks onto the target area from its database, such as the photovoltaic energy storage zone with the ID ER_PV_Storage. Load detection is not a one-time measurement, but a continuous data polling and subscription process targeting all load-related data sources within the area. The platform initiates data requests to all relevant field devices within the area, including: smart meters (AMIs), installed on the industrial, commercial, and residential user sides, providing accurate load data; transformer terminal units (TTUs), monitoring the total load on the low-voltage side of public distribution transformers; feeder terminal units (FTUs), monitoring the output power of the upstream feeders in the area, which can be used to verify the total load of the area; and distributed energy (DG) controllers, such as photovoltaic inverters and energy storage management systems (EMS), whose power data can be considered as negative loads. This detection process ensures the comprehensiveness of the data sources, enabling the capture of regional load dynamics from different levels and dimensions.

[0084] Through load detection, the platform acquires a series of raw, timestamped energy load data, primarily time series of active power (P) and reactive power (Q). The platform performs a series of preprocessing steps on this raw data to ensure its usability. First, time alignment is performed; due to differences in sampling periods and communication delays among different devices, the platform uses interpolation or resampling techniques to align all data points to a unified time base (e.g., one data point per minute). Next, data cleaning identifies and handles outliers (e.g., sudden changes due to communication interruptions) and missing values ​​(e.g., filling with mean values ​​or predictions based on historical patterns). Finally, data aggregation algebraically sums the load data of all users, transformer data, and distributed energy data within the region (total load = Σ user load + Σ line loss - Σ DG output) to calculate the net load of the energy region. The output is a set of high-quality, time-series energy load data, representing the true changes in the net load of the energy region.

[0085] The platform connects the discrete time series data points obtained in the previous step using mathematical methods (such as spline interpolation and piecewise linear fitting) to form a visualized, continuous energy load curve. This curve intuitively shows the trajectory of load changes over time. A key step is the fusion of energy fluctuation ranges, which adds credibility and risk dimensions to the curve. The platform retrieves the energy fluctuation range statistically derived from historical operating data for the region. For example, it calculates the 95% confidence interval of the load based on historical data from the same period and plots it above and below the load curve as a shaded area, or draws the envelope of the historical maximum and minimum loads. Therefore, the energy load curve is not just a line, but a composite chart that includes a main trend line, confidence intervals, or envelopes.

[0086] Specifically, the power monitoring platform is analyzing the photovoltaic and energy storage area of ​​the power grid. The power monitoring platform locks the area with ID ER_PV_Storage and initiates real-time data polling to the EMS systems of 10 industrial and commercial smart meters, 3 public distribution transformers (TTUs), 1 photovoltaic power station, and 1 energy storage power station within it, comprehensively covering the load and generation units in the area.

[0087] At 10:00, the platform collected raw data: the total load of industrial and commercial users was 1.2MW, the total load of distribution transformers was 1.25MW (including line losses), the photovoltaic output was 1.8MW, and the energy storage was currently in standby mode (output was 0). The platform then performed data aggregation calculation: net load = 1.25MW - 1.8MW = -0.55MW. This negative value clearly indicates that at the current moment, the region is feeding power back to the upper-level grid. The platform aligned the net load data points processed in the past hour to form a time series.

[0088] The platform connects these discrete data points to generate a curve showing that the current net load is negative. At the same time, the platform retrieves the historical database of the region and calculates that the historical fluctuation range of the net load at 10:00 is [-0.7MW, -0.3MW]. The energy load chart shows that the current net load of -0.55MW is within the historical normal range, but it is close to the upper limit of power backflow. This composite curve intuitively reveals the key information that there is a significant power surplus in the current region, providing direct and strong data support for subsequent decisions on whether to start energy storage charging to absorb excess energy.

[0089] Furthermore, multiple energy load nodes are identified based on the energy load curve. The energy control coefficient of the energy region is determined according to the node location, corresponding node shape, and multiple sub-energy control characteristics of the energy region. This approach takes into account the overall consideration of the node location, corresponding node shape, and multiple sub-energy control characteristics of the energy region, ensuring the accuracy of the energy control coefficient of the energy region.

[0090] At this point, key points of particular significance are identified from the continuous energy load curve, marking shifts in load behavior patterns. To achieve this, the platform employs signal processing and pattern recognition algorithms to analyze the load curve. Specific methods include: first / second derivative analysis, which calculates the first derivative (rate of change) and second derivative (acceleration) of the load curve with respect to time to accurately identify the curve's extreme points (peaks and troughs) and inflection points (points of maximum / minimum rate of change); and piecewise linear fitting, which uses algorithms (such as top-down segmentation methods) to approximate the complex curve as a series of beginnings and endings. Connected straight line segments, with the start and end points of each segment representing an energy load node, signify a relatively stable load state; along with moving averages and standard deviations, the crossover points of short-term and long-term moving averages (similar to golden crosses / death crosses in finance) can identify turning points in load trends; the output energy load nodes are one or more key time points on the time series, each node accompanied by its corresponding load value and timestamp, for example: {Node 1: 09:00, Load: 0.2MW (climbing start)}, {Node 2: 12:00, Load: 1.5MW (peak point)}.

[0091] From continuous energy load curves, key points of special significance are identified, marking shifts in load behavior patterns, similar to identifying P waves, QRS complexes, and T waves on an electrocardiogram. To achieve this, the platform employs signal processing and pattern recognition algorithms to analyze the load curves.

[0092] Specific methods include: first / second derivative analysis, which calculates the first derivative (rate of change) and second derivative (acceleration) of the load curve with respect to time to accurately find the extreme points (peaks and troughs) and inflection points (points with the maximum / minimum rate of change); piecewise linear fitting, which uses algorithms (such as top-down segmentation) to approximate complex curves as a series of connected straight line segments, with the start and end points of each segment being an energy load node, representing a relatively stable load state; and moving averages and standard deviations, which identify turning points in load trends by calculating the intersections of short-term and long-term moving averages (similar to golden crosses / death crosses in finance); the output energy load nodes are one or more key time points on the time series, each node accompanied by its corresponding load value and timestamp, for example: {Node 1: 09:00, Load: 0.2MW (climbing start)}, {Node 2: 12:00, Load: 1.5MW (peak point)}.

[0093] Specifically, the power monitoring platform has generated a load curve for the photovoltaic-storage energy zone that includes the current power surplus. The power monitoring platform analyzes the net load curve (including photovoltaic output prediction) of the photovoltaic-storage energy zone. Through second derivative analysis, the platform identifies two key energy load nodes: Energy load node A (current), time 10:00, net load -0.55MW, which is a peak point of power backflow, and the node shape is a gentle peak; Energy load node B (future), time 10:15, net load +0.95MW, which is the end point of a power reverse ramp caused by cloud cover, and the node shape is a steep valley (changing sharply from -0.55MW to +0.95MW).

[0094] The power monitoring platform's evaluation model began analyzing the controlled characteristics of the two nodes and the sub-energy provided by S12. For node A, its location is in the daytime flat period, with average economic efficiency, but its shape is a power backflow peak. Combined with the characteristics of S12 {voltage deviation: +4.5%}, the model determined that it has moderate control value. For node B, its location is 15 minutes later, with extremely high urgency, and its shape is a steep valley, which means huge power fluctuations. Most importantly, it is directly related to the characteristics of S12 {predicted power deficit: 1.5MW, predicted minimum voltage: 9.8kV}, which indicates that without control, a serious voltage instability event will occur.

[0095] During the model calculation phase, the evaluation model assigns the highest weight to risk. Since node B is associated with a serious risk that could lead to voltage collapse, the model calculates a very high energy control coefficient for the entire photovoltaic energy storage area, such as 0.92 (out of 1.0). Through step S132, the power monitoring platform no longer just sees power surplus or impending fluctuations, but through quantitative calculation, it clearly marks this area as the highest priority control object in the current power grid. This coefficient of 0.92 will directly drive step S133, allocating the most urgent and powerful control resources to it.

[0096] Therefore, by collecting the current operating status of the power grid, determining the first level of energy control content based on the current operating status of the power grid and the regional location of each energy region, determining the second level of energy control content based on the current operating status of the power grid and the energy control coefficient of each energy region, and determining the power grid's energy control events based on the first and second level of energy control content, the overall consideration of the first and second level of energy control content is taken into account, ensuring the accuracy of the power grid's energy control events. At the same time, the energy dynamic map of the energy region is introduced to further control the controlled characteristics of multiple sub-energys, taking into account the regional location of each energy region, the corresponding energy control coefficient, and the overall operating status of the power grid, thereby improving the accuracy of the power grid's energy control events.

[0097] At this point, the platform collects the current operating status of the power grid and obtains macro-level global data from the Power Grid Dispatch Automation System (EMS) and Wide Area Measurement System (WAMS), including total system power generation, total load, system frequency, power flow of critical lines in the backbone network, voltage levels of major hubs, and spinning reserve capacity. This constitutes a global perspective for control decisions. Based on this, the platform treats the power grid as a whole and runs optimization algorithms (such as Optimal Power Flow (OPF) and Security Constrained Economic Dispatch (SCED)) to determine the first level of energy control.

[0098] The goals of algorithms are usually global, such as minimizing the overall power generation cost, minimizing overall network losses, or removing the limits of a critical line. The current operating state of the power grid is the main input, while the regional location of each energy region (i.e., their electrical connection relationship in the power grid topology) determines their way and ability to participate in global optimization.

[0099] The first level of energy control output is a macro-level, guiding strategy, not specific equipment instructions. Instead, it specifies the expected power adjustment amount or voltage adjustment range for each energy region. For example, it may require the F5 feeder region at the end of the grid to provide 500kW of reactive power support to support the voltage in that region.

[0100] Besides the current operating status of the power grid (used to determine the control direction, such as whether charging or discharging is required), the most important input is the energy control coefficient calculated by S132. The platform will sort all energy regions from high to low according to their energy control coefficients. The higher the coefficient, the more urgent the control demand and the more significant the control benefits.

[0101] The platform will prioritize allocating control resources to these areas and formulate more detailed and aggressive local control strategies. The second layer of energy control output is a specific, executable set of instructions for a specific controlled object. For example, for the photovoltaic energy storage zone with the highest coefficient, the content is to command the 500kW / 1000kWh energy storage system in the zone to immediately charge at 400kW for 15 minutes.

[0102] The platform coordinates and integrates the macro-level first-level content (global objective) and the micro-level second-level content (local optimal solution) to form conflict-free, executable energy control events; the platform typically uses event-driven or rule engine methods for fusion.

[0103] The platform will check whether the second layer of content is consistent with the goal of the first layer of content; if there is a conflict, for example, the first layer of content requires region A to reduce the voltage, while the second layer of content is preparing to discharge the energy storage in region A (which will increase the voltage), then the conflict resolution mechanism needs to be activated.

[0104] When the objectives are aligned, the platform will synthesize the instructions. For example, if the first requirement is to provide 500kVar reactive power support and the second requirement specifies the parameters for the energy storage SVG mode, the platform will combine the two into a complete energy storage control instruction.

[0105] The platform encapsulates all synthesized and verified control instructions into one or more energy control events; each energy control event contains a clear and unique ID, triggering condition, target area, controlled object, control parameters, execution time and duration, to form a standardized control work order.

[0106] Specifically, the power monitoring platform has calculated an energy control coefficient as high as 0.92 for the photovoltaic energy storage area. The power monitoring platform collects the current operating status of the power grid: the system frequency is 50.02Hz (slightly high), the main grid voltage is generally high (10.5kV-10.7kV), but there is no line congestion; the platform runs a global optimization algorithm with the goal of reducing the overall grid voltage level and reducing network losses; based on the location of the photovoltaic energy storage area at the end of the grid and in a region with high voltage, the algorithm generates the first level of energy control content: it is recommended that the photovoltaic energy storage area absorb about 500kVar of reactive power to assist in regulating the end voltage.

[0107] The power monitoring platform found that the energy control coefficient of the photovoltaic energy storage area was 0.92, the highest among all areas, indicating that there was a severe voltage collapse risk predicted by S12 in this area. The platform immediately formulated the highest priority local control strategy for this area. Based on the characteristics of S124 ({predicted power deficit: 1.5MW, T+15min}), the platform generated a second layer of energy control content: immediately execute two-stage control: 1. Command the energy storage system to charge at 400kW power for 15 minutes to suppress the current voltage; 2. At T+15 minutes, command the energy storage system to discharge at 1.5MW power for 10 minutes to support the voltage.

[0108] In the consistency verification, the first step of the second layer (charging 400kW) absorbs both active and reactive power, which is completely consistent with the goal of the first layer to absorb reactive power, and the effect is stronger. The second step of the second layer (discharging) is to solve the serious safety problem that is about to occur, and its priority is higher than the conventional voltage optimization. Therefore, the platform decided to adopt the more refined and urgent second layer as the final solution, because it solves the global voltage problem while preventing major future safety risks.

[0109] The power monitoring platform generates an energy control event named EVT_Voltage_Stab_ER_PV_Storage_001 and sends it to the energy storage management system (EMS) of the photovoltaic energy storage area. This event contains all the specific parameters of the two-stage control mentioned above, ensuring the accuracy and timeliness of the control.

[0110] refer to Figure 5 In step S14, the specific steps are as follows:

[0111] S141: Collect working data combinations from each energy region, determine corresponding data association parameters based on the identification of working data combinations from each energy region, determine abnormal data regions based on multiple working data combinations, corresponding data association parameters, and multiple sub-energy controlled characteristics of the corresponding energy region; determine multiple abnormal energy data based on the traversal of abnormal data regions, and mark the positions of multiple abnormal energy data.

[0112] S142: Collect the current working mode of the power grid, determine multiple current working contents based on the identification of the current working mode of the power grid, and determine the first energy anomaly content based on the multiple current working contents and the location of the corresponding abnormal energy data.

[0113] S143: Determine the second energy anomaly content based on multiple current working contents and corresponding energy regions, and determine the corresponding energy anomaly event based on the first energy anomaly content, the second energy anomaly content, and the regional morphology of multiple energy regions.

[0114] In the embodiments of this application, working data combinations of each energy region are collected, and corresponding data association parameters are determined based on the identification of working data combinations of each energy region. Abnormal data regions are determined based on multiple working data combinations, corresponding data association parameters, and multiple sub-energy controlled features of the corresponding energy region. Multiple abnormal energy data are determined based on the traversal of abnormal data regions, and the positions of multiple abnormal energy data are marked. This approach is compatible with the overall consideration of traversing abnormal data regions and ensures the accuracy of multiple abnormal energy data.

[0115] At this point, the structured working data combination generated in real time for each energy region in step S121 is obtained, which includes the synchronous electrical quantity data of all monitoring points in the region; identifying data correlation parameters is the core of establishing a normal behavior baseline. The platform does not view each data point in isolation, but rather mines the inherent correlation between data through advanced algorithms.

[0116] Specific methods include: statistical correlation analysis, calculating Pearson correlation coefficients, Spearman rank correlation coefficients, etc., between different variables to quantify linear or monotonic relationships between variables; causal relationship inference, using Granger causality tests or transfer entropy to determine whether a change in one time series is the cause of a change in another time series; and physical model constraints, establishing equality or inequality constraints between data based on physical principles such as Kirchhoff's laws and energy conservation; the output data correlation parameters are a complex set of parameters, which is a correlation coefficient matrix, a causal network diagram, or a set of constraint equations of a physical model, accurately quantifying the mathematical relationships that should exist between data points in the region during normal operation.

[0117] Deviation detection identifies energy regions with abnormal operating conditions by comparing real-time data with a baseline of normal behavior. The platform continuously inputs real-time collected working data into a baseline model defined by data correlation parameters to calculate residuals or inconsistencies. This can be achieved through residual analysis, which involves substituting real-time data into the physical model equations and calculating the difference between the two sides of the equation. A significant non-zero residual indicates that the model has been violated. Alternatively, correlation violation detection can be used to check whether the correlation coefficient of the real-time data deviates significantly from the baseline correlation coefficient matrix.

[0118] The key fusion step involves combining the sub-energy controlled characteristics, and the platform will pay special attention to the correlation parameters related to these characteristics. For example, if a region is characterized by voltage sensitivity to load changes, the platform will focus on monitoring the load-voltage correlation parameter. When a slight load change is detected but the voltage fluctuates drastically, the anomaly score of that region will be increased even if other parameters are normal. When the overall anomaly score of a region exceeds a preset threshold, the region is marked as an abnormal data region, indicating that its overall operating status has deviated from the normal track.

[0119] Within the identified abnormal data area, the platform will locate the specific data points that caused the abnormality in the area and perform a detailed point-by-point scan and diagnosis of all working data combinations within the abnormal data area.

[0120] The methods include: traversal and single-point detection, applying univariate anomaly detection algorithms, such as Z-score (standard score) and IQR (interquartile range), to each data point within the region to determine whether it is statistically an outlier; contribution analysis, calculating the contribution of each variable to the overall residual or inconsistency in a multivariate model, with the variable with the largest contribution being the source of the anomaly; and pattern matching, matching the change patterns of outlier data points with a predefined fault mode library to help determine the anomaly type.

[0121] The output anomalous energy data consists of one or more specific, tagged data points; each tag contains rich metadata: {data point ID, data type, outlier, normal range, timestamp, physical location description, confidence level}. This precise location is the basis for subsequent fault isolation and root cause analysis.

[0122] Furthermore, the current operating mode of the power grid is collected, and multiple current operating contents are determined based on the identification of the current operating mode of the power grid. The first energy anomaly content is determined based on the multiple current operating contents and the location of the corresponding abnormal energy data. This overall consideration of multiple current operating contents and the location of the corresponding abnormal energy data ensures the accuracy of the first energy anomaly content.

[0123] At this point, the platform collects the current operating mode of the power grid. It obtains the current macro-operation strategy and status of the power grid from the power grid dispatch management system (EMS) or operation logs. This is not electrical quantity data, but descriptive labels or enumerated values. Common modes include: normal operation mode, where the power grid is in a standard dispatch state with sufficient safety margin; planned maintenance mode, where some equipment is out of service for maintenance, the power grid operation mode has changed, and there are weak links; peak power supply mode, where the power grid load is close to or has reached its peak, and spinning reserve capacity is tight; fault recovery mode, where the power grid has just experienced a fault and the system is relatively vulnerable; and high penetration of new energy mode, where the system inertia is low and volatility is high.

[0124] After determining the mode, the platform will parse these abstract mode labels into a series of specific, quantifiable operational constraints and objectives, i.e., the current work content. The platform will map each mode to a set of work content by querying the preset rule base or configuration file. For example, the peak power supply mode is parsed as: {prohibit unplanned power outage operations, narrow the voltage fluctuation range to ±3%, adjust the frequency protection setting, and prioritize the use of rotating standby}. The output current work content is a structured set of rules or a list of constraints.

[0125] Therefore, the second energy anomaly content is determined based on multiple current working contents and corresponding energy regions, and the corresponding energy anomaly event is determined based on the first energy anomaly content, the second energy anomaly content, and the regional morphology of multiple energy regions. This approach takes into account the overall consideration of the first energy anomaly content, the second energy anomaly content, and the regional morphology of multiple energy regions, ensuring the accuracy of the corresponding energy anomaly event.

[0126] At this point, based on multiple current work contents and corresponding energy regions, the second energy anomaly is identified, and a systemic risk assessment based on regional function is conducted. This goes beyond the physical impact and deeply analyzes the potential impact of the anomaly on the social, economic, or specific functions of the entire energy region. The platform will perform a functional impact analysis, which requires pre-storing a functional profile for each energy region, defining its main load type, importance level, sensitivity to power quality, and role in the socio-economic sphere. For example, Region A: a financial data center with extremely high requirements for power supply continuity and power quality; Region B: a residential area, sensitive to power outages but not sensitive to short-term voltage fluctuations. The platform cross-compares the current work contents with the functional profiles of the energy regions.

[0127] In peak power supply mode, a load anomaly in a residential area only affects comfort; however, in fault recovery mode, the same anomaly hinders the entire black start process; the output of the second energy anomaly is a descriptive conclusion.

[0128] The platform will use a multi-dimensional fusion diagnostic model to fuse information. Taking the first energy anomaly content and the second energy anomaly content as input, the platform will retrieve the regional morphological information of the area where the anomaly is located, including the topology (whether it is radial, ring or mesh), network characteristics (whether it is an overhead line or a cable) and power structure (whether there is a distributed power source in the area).

[0129] The platform incorporates a diagnostic engine based on expert rules or deep learning. This engine integrates all information to attribute and characterize anomalies, including event classification (such as equipment failure or power quality pollution) and root cause inference (inferring the root cause by combining regional morphology and anomaly characteristics). The output energy anomaly event is a structured, highly condensed diagnostic report that not only describes the phenomenon and its impact but, more importantly, provides the event type, root cause inference, and preliminary handling recommendations. For example: Event type: Power quality pollution event; Root cause inference: Failure of large rectifier equipment or filter failure on the company's dedicated line within the industrial zone; Handling recommendation: Immediately notify the user to inspect the equipment and consider temporarily restricting the user's harmonic emissions.

[0130] refer to Figure 6 In step S15, the specific steps are as follows:

[0131] S151: Collect energy control events of the power grid, determine multiple sub-energy control items of different dimensions based on the detection of energy control events of the power grid, identify the corresponding sub-energy control content according to the identification of multiple sub-energy control items, and mark the priority of each sub-energy control content;

[0132] S152: Determine the first abnormal maintenance content based on multiple sub-energy control contents and corresponding energy abnormal events; determine the second abnormal maintenance content based on multiple sub-energy control contents, corresponding content priorities and corresponding energy regions; and determine the abnormal maintenance events of the energy region based on the first abnormal maintenance content, the second abnormal maintenance content and multiple sub-energy controlled characteristics.

[0133] S153: Collect the current operating status of the power grid, determine the first energy control coefficient based on the current operating status of the power grid and abnormal maintenance events in each energy region, and determine the energy fluctuation range of the power grid based on the identification of the energy dynamic diagram of each energy region; determine the second energy control coefficient based on the energy fluctuation range of the power grid and the current operating status of the power grid, and determine the energy control level of the power grid based on the mapping relationship between the first energy control coefficient, the second energy control coefficient and the energy control level.

[0134] In the embodiments of this application, energy control events of the power grid are collected, and multiple sub-energy control items of different dimensions are determined based on the detection of energy control events of the power grid. Sub-energy control content corresponding to the identification of multiple sub-energy control items is then determined, and the priority of each sub-energy control content is marked. This approach is compatible with the overall consideration of energy control event detection of the power grid and ensures the accuracy of multiple sub-energy control items of different dimensions.

[0135] At this point, energy control events that have already undergone preliminary decision-making are acquired. These events are typically macroscopic and goal-oriented instructions, such as smoothing voltage fluctuations in the photovoltaic energy storage area or eliminating power overruns in the main line L1. Since a macroscopic control event often cannot be completed by a single device or a single action, the platform must decompose it into a set of smaller, more manageable tasks, namely sub-energy control projects. This decomposition process is carried out along multiple dimensions to ensure the comprehensiveness and systematic nature of the control.

[0136] It can be decomposed into scheduling tasks for different assets such as energy storage, SVG, and OLTC according to the control resource dimension; into tasks with different response speeds such as second-level rapid response and minute-level dynamic adjustment according to the time scale dimension; or into tasks with different objectives such as voltage quality improvement and frequency support according to the control objective dimension; a complex control objective is broken down into a list of sub-tasks in multiple different domains and time frames.

[0137] The platform transforms logical task items into specific, executable content. Based on the specific requirements of each sub-energy control item, it queries the real-time status of the corresponding control resources and applies optimization algorithms or preset rule bases to generate control instructions. These instructions must include all necessary parameters, such as the control object ID, action type, target value, and rate of change. The platform must prioritize these control contents. This is a multi-criteria decision-making process. Typically, instructions related to preventing system instability are marked as highest priority (P0), instructions aimed at saving costs are marked as medium priority (P1), and instructions aimed at improving power quality are marked as ordinary priority (P2). The output is a structured instruction queue with priority labels, which can be directly issued for execution.

[0138] Furthermore, the first abnormal maintenance content is determined based on multiple sub-energy control contents and corresponding energy anomaly events. The second abnormal maintenance content is determined based on multiple sub-energy control contents, corresponding content priorities, and corresponding energy regions. The abnormal maintenance events of the energy region are determined based on the first abnormal maintenance content, the second abnormal maintenance content, and multiple sub-energy controlled characteristics. This approach takes into account the overall consideration of the first abnormal maintenance content, the second abnormal maintenance content, and multiple sub-energy controlled characteristics, ensuring the accuracy of the abnormal maintenance events of the energy region.

[0139] At this point, the platform performs correlation mapping analysis to cross-compare the list of sub-energy control content generated in S151 with the energy anomaly events determined in S14. This comparison is based on spatiotemporal correlation and functional correlation, that is, checking whether the execution area of ​​the control content is consistent with or electrically adjacent to the occurrence area of ​​the anomaly event, and analyzing whether the control function directly offsets the impact of the anomaly. The first anomaly maintenance content output is not a new action, but a reclassification of the existing control content, marking those control instructions with emergency anomaly mitigation functions, indicating that they are not only optimized operations, but also critical emergency responses.

[0140] The platform generates maintenance recommendations through a decision engine based on priority and regional characteristics. It prioritizes sub-energy control content marked as high priority (such as P0) because scenarios requiring the highest priority control often indicate that the underlying physical problems are very serious. At the same time, the platform combines the functional profile of the energy area and the type of energy anomaly event to retrieve the most matching maintenance strategy from the knowledge base. The output of the second anomaly maintenance content is one or more specific, operation-oriented action instructions, usually in the form of a work order, which includes information such as maintenance objectives, area, recommended measures, and urgency level.

[0141] Emergency control (first abnormal maintenance content) and root cause maintenance (second abnormal maintenance content) are integrated into a unified, structured abnormal maintenance event. The platform adopts an event fusion engine to integrate multi-dimensional information. The engine takes the first and second abnormal maintenance contents as core inputs and refers to the sub-energy controlled characteristics (such as the predicted minimum voltage value) in step S12 to verify and reinforce the urgency of the maintenance event.

[0142] The engine packages all information into a standardized abnormal maintenance event object, which includes all key information such as event ID, root cause inference, impact assessment, emergency control measures taken, and recommended maintenance measures. This comprehensive information package can be distributed to different roles, realizing the extension from automated control to intelligent operation and maintenance.

[0143] Therefore, the current operating status of the power grid is collected, and a first energy control coefficient is determined based on the current operating status of the power grid and abnormal maintenance events in each energy region. Simultaneously, the energy fluctuation range of the power grid is determined based on the identification of the energy dynamic diagrams of each energy region. A second energy control coefficient is determined based on the energy fluctuation range of the power grid and the current operating status of the power grid. The energy control level of the power grid is determined based on the mapping relationship between the first energy control coefficient, the second energy control coefficient, and the energy control level. This comprehensive approach considers the mapping relationship between the first energy control coefficient, the second energy control coefficient, and the energy control level, ensuring the accuracy of the power grid's energy control level. Furthermore, abnormal energy events are introduced to further control abnormal maintenance events in the energy region. This comprehensive approach considers abnormal maintenance events in each energy region, the current operating status of the power grid, and the energy fluctuation range of the power grid, thereby improving the accuracy of the power grid's energy control level.

[0144] At this time, the platform quantifies the current vulnerability and health status of the power grid, and assesses the grid's ability to withstand stress and anomalies at this moment. The platform collects indicators that reflect the macro-health level of the power grid, such as system frequency deviation and total spinning reserve percentage, as the basic physical condition of the power grid.

[0145] At the same time, the platform counts the number, type, and severity of abnormal maintenance events in all energy areas. As a symptom that the power grid is currently suffering from, the platform uses a weighted scoring algorithm or a trained model to fuse these two types of inputs and calculate the first energy control coefficient. This coefficient is a normalized value. The higher the coefficient, the more strained the current operating state of the power grid, the more serious the anomaly, and the more fragile the system as a whole.

[0146] The platform quantifies the uncertainties and potential risks of the power grid in the future, assessing the extent of the impact the power grid will face in the near future. It aggregates energy dynamic maps of all energy regions and uses Monte Carlo simulation or interval analysis to deduce the probability distribution of key state variables of the entire power grid within a future time window, thereby determining the range of energy fluctuations. The wider this range, the greater the uncertainty.

[0147] The platform compares this fluctuation range with the current operating status (such as the level of reserve capacity), assesses the distance between the fluctuation range and the current safety margin, and calculates the second energy control coefficient. This coefficient is also a normalized value. The higher the value, the greater the uncertainty the power grid will face in the future and the higher the potential safety risks.

[0148] The platform transforms complex quantitative assessment results into an intuitive macro-level situation that dispatchers can quickly understand. It integrates a first energy control coefficient representing current vulnerability and a second energy control coefficient representing future risks. The logic is that healthy systems need to focus on future risks, while vulnerable systems must prioritize addressing current issues. The platform uses a predefined mapping table to look up the integrated index to determine the energy control level. The output level is a discrete label with clear action guidance, such as normal, caution, warning, severe, or emergency. It serves as the overall guideline for power grid dispatch operations.

[0149] Please see Figure 7 , Figure 7 This is a schematic diagram of the structural composition of the power grid energy control device based on the power monitoring platform in an embodiment of the present invention; the power grid energy control device based on the power monitoring platform includes:

[0150] Energy region module 21 is used to determine multiple energy regions of the power grid based on multiple working data of the power grid and the power distribution map of the power grid during the process of monitoring the power grid on the power monitoring platform.

[0151] The sub-energy controlled feature module 22 is used to determine the energy dynamic map of the energy region based on the regional location of each energy region, the corresponding working data combination and the surrounding environmental data combination, and to determine multiple sub-energy controlled features based on the identification of the energy dynamic map of the energy region.

[0152] The energy control event module 23 is used to determine the corresponding energy load curve based on the load detection of the energy area, determine the energy control coefficient of the energy area according to the multiple sub-energy control characteristics of the energy area and the corresponding energy load curve, and determine the energy control event of the power grid according to the regional location of each energy area, the corresponding energy control coefficient and the current working state of the power grid.

[0153] The energy anomaly event module 24 is used to determine multiple abnormal energy data based on the detection of working data combinations of various energy regions, and to determine the corresponding energy anomaly event according to the location of multiple abnormal energy data, the corresponding energy region and the current working mode of the power grid.

[0154] The energy control level module 25 is used to determine multiple sub-energy control contents based on the identification of energy control events in the power grid, determine the abnormal maintenance events of the energy area based on the multiple sub-energy control contents, the corresponding energy area and the corresponding energy abnormal events, and determine the energy control level of the power grid based on the abnormal maintenance events of each energy area, the current operating status of the power grid and the energy fluctuation range of the power grid.

[0155] 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 power grid energy control method based on a power monitoring platform, characterized in that, include: During the process of monitoring the power grid on the power monitoring platform, multiple energy regions of the power grid are determined based on multiple working data of the power grid and the power distribution map of the power grid. Based on the regional location of each energy region, the corresponding working data combination, and the surrounding environmental data combination, an energy dynamic map of the energy region is determined. Multiple sub-energy controlled features are then identified based on the identification of the energy dynamic map. This includes: collecting data on each energy region; determining the corresponding regional location based on the detection of each energy region; collecting multiple corresponding working data within each energy region's regional location; constructing corresponding working data combinations based on these multiple working data; determining multiple surrounding spaces based on the detection of the surrounding locations of each energy region; determining environmental data combinations within these surrounding spaces based on their spatial locations, corresponding multiple environmental data, and spatial morphology; determining a first-level energy change map based on the regional location of each energy region and the corresponding working data combination; determining a second-level energy change map based on the regional location of each energy region and the surrounding environmental data combination; and determining the energy dynamic map of the energy region based on the first and second-level energy change maps. Multiple energy dynamic regions are then identified based on the detection of the energy dynamic map of the energy region, and corresponding sub-energy controlled features are determined based on the identification of each energy dynamic region, with the collection of multiple sub-energy controlled features. The energy dynamic map not only describes the current and past energy states but, more importantly, proactively predicts future energy states based on environmental changes. Based on load detection of the energy region, the corresponding energy load curve is determined. The energy control coefficient of the energy region is determined based on the controlled characteristics of multiple sub-energys within the energy region and the corresponding energy load curve. Energy control events for the power grid are determined based on the location of each energy region, its corresponding energy control coefficient, and the current operating state of the power grid. This includes: collecting data from the energy region and performing load detection on it; determining multiple energy load data points based on the load detection; determining the corresponding energy load curve based on the multiple energy load data points and the energy fluctuation range corresponding to the energy region; identifying multiple energy load nodes based on the identification of the energy load curve; and determining the energy control coefficient of the energy region based on the node location, corresponding node shape, and controlled characteristics of multiple sub-energys within the energy region. Each energy control event includes a unique ID, trigger condition, target region, controlled object, control parameters, execution time, and duration. Multiple abnormal energy data are identified by combining working data from various energy regions. Based on the location of the multiple abnormal energy data, the corresponding energy region, and the current working mode of the power grid, the corresponding energy anomaly event is determined. The energy anomaly event not only describes the phenomenon and its impact, but more importantly, it provides the event type, root cause inference, and preliminary handling suggestions. Multiple sub-energy control contents are determined based on the identification of energy control events in the power grid. Abnormal maintenance events in the energy area are determined based on the multiple sub-energy control contents, the corresponding energy areas, and the corresponding abnormal energy events. The energy control level of the power grid is determined based on the abnormal maintenance events in each energy area, the current operating status of the power grid, and the energy fluctuation range of the power grid.

2. The power grid energy control method based on a power monitoring platform according to claim 1, characterized in that, During the process of monitoring the power grid on the power monitoring platform, multiple energy regions of the power grid are determined based on multiple operating data of the power grid and the power distribution map of the power grid, including: The power monitoring platform communicates with the power grid and performs dynamic monitoring of the power grid; it monitors the power grid monitoring process of the power monitoring platform in real time; it collects multiple working data of the power grid, determines multiple data combinations based on the multiple working data of the power grid and the overall shape of the power grid, and determines the corresponding power distribution area based on the identification of each data combination; Multiple power distribution areas are collected and the power distribution of each area is marked. Based on the location of each power distribution area, the corresponding power distribution, and the power distribution map of the power grid, multiple energy regions of the power grid are determined.

3. The power grid energy control method based on a power monitoring platform according to claim 1, characterized in that, The method of determining the corresponding energy load curve based on load detection of the energy region, determining the energy control coefficient of the energy region based on the multiple sub-energy control characteristics of the energy region and the corresponding energy load curve, and determining the energy control event of the power grid based on the regional location of each energy region, the corresponding energy control coefficient, and the current operating state of the power grid, further includes: The system collects the current operating status of the power grid, determines the first level of energy control content based on the current operating status of the power grid and the regional location of each energy region, determines the second level of energy control content based on the current operating status of the power grid and the energy control coefficient of each energy region, and determines the energy control events of the power grid based on the first and second level of energy control content.

4. The power grid energy control method based on a power monitoring platform according to claim 1, characterized in that, The detection of multiple abnormal energy data based on the combination of working data from various energy regions, and the determination of corresponding energy anomaly events based on the location of the multiple abnormal energy data, the corresponding energy region, and the current operating mode of the power grid, including: The system collects working data combinations from various energy regions, determines corresponding data association parameters based on the identification of these combinations, identifies anomalous data regions based on multiple working data combinations, corresponding data association parameters, and multiple sub-energy controlled features of the corresponding energy regions, and identifies multiple anomalous energy data based on the traversal of these anomalous data regions, and marks the locations of these anomalous energy data.

5. The power grid energy control method based on a power monitoring platform according to claim 4, characterized in that, The method of detecting multiple abnormal energy data based on the combination of working data from various energy regions, and determining the corresponding energy anomaly event based on the location of the multiple abnormal energy data, the corresponding energy region, and the current operating mode of the power grid, further includes: The current working mode of the power grid is collected, and multiple current working contents are determined based on the identification of the current working mode of the power grid. The first energy anomaly content is determined based on the multiple current working contents and the location of the corresponding abnormal energy data. The second energy anomaly content is determined based on multiple current working contents and corresponding energy regions, and the corresponding energy anomaly event is determined based on the first energy anomaly content, the second energy anomaly content, and the regional morphology of multiple energy regions.

6. The power grid energy control method based on a power monitoring platform according to claim 1, characterized in that, The process involves identifying multiple sub-energy control contents based on the recognition of power grid energy control events, determining abnormal maintenance events for each energy region based on these sub-energy control contents, their corresponding energy regions, and corresponding energy anomaly events, and determining the power grid's energy control level based on the abnormal maintenance events for each energy region, the current operating status of the power grid, and the power grid's energy fluctuation range. This includes: Collect energy control events from the power grid, determine multiple sub-energy control items of different dimensions based on the detection of energy control events from the power grid, identify the corresponding sub-energy control content based on the identification of multiple sub-energy control items, and mark the priority of each sub-energy control content; The first abnormal maintenance content is determined based on multiple sub-energy control contents and corresponding energy abnormal events. The second abnormal maintenance content is determined based on multiple sub-energy control contents, corresponding content priorities, and corresponding energy regions. The abnormal maintenance events of the energy region are determined based on the first abnormal maintenance content, the second abnormal maintenance content, and multiple sub-energy controlled characteristics.

7. The power grid energy control method based on a power monitoring platform according to claim 6, characterized in that, The process of determining multiple sub-energy control contents based on the identification of power grid energy control events, determining abnormal maintenance events for the energy region based on the multiple sub-energy control contents, the corresponding energy region, and the corresponding energy anomaly events, and determining the power grid energy control level based on the abnormal maintenance events of each energy region, the current operating status of the power grid, and the energy fluctuation range of the power grid, further includes: The system collects the current operating status of the power grid, determines the first energy control coefficient based on the current operating status of the power grid and abnormal maintenance events in each energy region, and determines the energy fluctuation range of the power grid based on the identification of the energy dynamic map of each energy region. The system then determines the second energy control coefficient based on the energy fluctuation range of the power grid and the current operating status of the power grid, and determines the energy control level of the power grid based on the mapping relationship between the first energy control coefficient, the second energy control coefficient and the energy control level.

8. An energy control device for a power grid based on an energy monitoring platform, characterized in that, The power grid energy control device based on the power monitoring platform is applied to the power grid energy control method based on the power monitoring platform as described in any one of claims 1-7, wherein the power grid energy control device based on the power monitoring platform comprises: The energy zone module is used to determine multiple energy zones of the power grid based on multiple working data of the power grid and the power distribution map of the power grid during the monitoring of the power grid by the power monitoring platform. The sub-energy controlled feature module is used to determine the energy dynamic map of each energy region based on its regional location, the corresponding working data combination, and the surrounding environmental data combination, and to determine multiple sub-energy controlled features based on the identification of the energy dynamic map of the energy region. The energy control event module is used to determine the corresponding energy load curve based on the load detection of the energy region, determine the energy control coefficient of the energy region based on the multiple sub-energy control characteristics of the energy region and the corresponding energy load curve, and determine the energy control event of the power grid based on the regional location of each energy region, the corresponding energy control coefficient and the current operating state of the power grid. The energy anomaly event module is used to identify multiple abnormal energy data based on the detection of working data combinations from various energy regions, and to determine the corresponding energy anomaly event based on the location of the multiple abnormal energy data, the corresponding energy region, and the current working mode of the power grid. The energy control level module is used to determine multiple sub-energy control contents based on the identification of energy control events in the power grid, determine the abnormal maintenance events of the energy area based on the multiple sub-energy control contents, the corresponding energy area and the corresponding energy abnormal events, and determine the energy control level of the power grid based on the abnormal maintenance events of each energy area, the current operating status of the power grid and the energy fluctuation range of the power grid.

Images