Power grid operation control method and device, computer equipment, medium and program product

Abstract

The invention relates to a power grid operation control method and device, computer equipment, a medium and a program product, and relates to the field of power grids. Comprising the steps of obtaining state time sequence data of a target power grid and operation time sequence data of at least one power device in the target power grid; dividing the state time sequence data into a plurality of state data subsets according to different time scales, and determining a time scale interval corresponding to each state data subset; aiming at each time scale interval, extracting an operation data subset corresponding to the time scale interval from each operation time sequence data; based on the matching degree between each operation data subset and the state data subset corresponding to the time scale interval, determining a scheduling scheme for each power device; and performing operation control on each power device according to the scheduling scheme so as to enable the target power grid to reach a stable operation state. The method can realize stable operation of the power grid.

Description

Technical Field

[0001] This application relates to the field of power grids, and in particular to a power grid operation control method, apparatus, computer equipment, computer-readable storage medium, and computer program product. Background Technology

[0002] In the current power system, the widespread application of renewable energy has placed higher demands on the stable operation of the power grid, which needs more flexible and efficient control measures to cope with the increasingly complex operating environment.

[0003] Traditional power grid operation and control are based on fixed time divisions and resource allocation, which makes it difficult to adapt to the diverse characteristics of different energy sources. As a result, when faced with sudden fluctuations, some energy equipment cannot respond in time, while other energy equipment may be over-allocated. This imbalance directly affects the operational stability of the power grid. Summary of the Invention

[0004] Therefore, it is necessary to provide a power grid operation control method, device, computer equipment, computer-readable storage medium, and computer program product that can achieve stable power grid operation in response to the above-mentioned technical problems.

[0005] In a first aspect, this application provides a power grid operation control method, including:

[0006] Acquire the state timing data of the target power grid, as well as the operating timing data of at least one power device in the target power grid;

[0007] The state time series data is divided into multiple state data subsets according to different time scales, and the time scale interval corresponding to each state data subset is determined.

[0008] For each time scale interval, extract the runtime data subset corresponding to the time scale interval from each runtime sequence data;

[0009] Based on the degree of matching between each subset of operational data and the corresponding subset of state data for the time scale interval, a scheduling scheme for each power equipment is determined.

[0010] The operation and control of each power equipment are carried out according to the dispatching plan in order to make the target power grid reach a stable operating state.

[0011] In one embodiment, the state time series data is divided into multiple state data subsets according to different time scales, including:

[0012] Extract the state fluctuation characteristics of the target power grid from the state time series data;

[0013] Based on the characteristics of state fluctuations, the state time series data is divided into multiple state data subsets according to different time scales.

[0014] In one embodiment, based on state fluctuation characteristics, the state time series data is divided into multiple state data subsets according to different time scales, including:

[0015] Based on the characteristics of state fluctuations, the state time series data is divided into multiple initial data subsets according to different time scales;

[0016] For each initial data subset, extract the fluctuation characteristics corresponding to the initial data subset;

[0017] If the fluctuation characteristics do not meet the expected fluctuation conditions, the time scale interval to which the initial data subset belongs is divided into multiple sub-time intervals;

[0018] The initial data subset that meets the expected fluctuation conditions, and the target data subsets corresponding to each of the multiple sub-time intervals, are collectively determined as multiple state data subsets.

[0019] In one embodiment, a scheduling scheme for each power device is determined based on the degree of matching between each subset of operational data and the corresponding subset of state data for a time scale interval, including:

[0020] Based on the degree of matching between each subset of operational data and the corresponding subset of state data for the time scale interval, each power device is scheduled to obtain a preliminary scheduling scheme.

[0021] Based on the preliminary scheduling plan, simulated operation was conducted on each power equipment to obtain the simulation results;

[0022] Based on the simulation results, an operational risk analysis was conducted to obtain the operational risks corresponding to the preliminary scheduling scheme.

[0023] Based on operational risks, the preliminary scheduling plan is optimized to obtain the final scheduling plan.

[0024] In one embodiment, the preliminary scheduling scheme is optimized based on operational risks to obtain a scheduling scheme, including:

[0025] Identify the types of operational risks;

[0026] Obtain historical scheduling configuration information corresponding to risk types, as well as the time characteristics of operational risks.

[0027] Based on historical scheduling configuration information and time characteristics, the preliminary scheduling scheme is optimized to obtain the final scheduling scheme.

[0028] In one embodiment, the operation control of each power device is performed according to the scheduling scheme, including:

[0029] The dispatching plan is converted into preliminary control instructions and issued to each power equipment so that each power equipment operates in accordance with the preliminary control instructions;

[0030] If the target power grid does not reach a stable operating state during the operation of various power equipment, the parameters of the initial control command are calibrated to obtain the adjusted control command.

[0031] Continue to control the operation of each power device according to the adjusted control instructions until the target power grid reaches a stable operating state.

[0032] Secondly, this application also provides a power grid operation control device, comprising:

[0033] The data acquisition module is used to acquire the state time-series data of the target power grid, as well as the operation time-series data of at least one power device in the target power grid;

[0034] The data partitioning module is used to divide the state time series data into multiple state data subsets according to different time scales, and determine the time scale interval corresponding to each state data subset.

[0035] The data extraction module is used to extract a subset of runtime data corresponding to each time scale interval from each runtime sequence data.

[0036] The scheduling scheme generation module is used to determine the scheduling scheme for each power equipment based on the degree of matching between each subset of operating data and the corresponding subset of state data in the time scale interval.

[0037] The operation control module is used to control the operation of each power device according to the dispatching plan, so as to make the target power grid reach a stable operating state.

[0038] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0039] Acquire the state timing data of the target power grid, as well as the operating timing data of at least one power device in the target power grid;

[0040] The state time series data is divided into multiple state data subsets according to different time scales, and the time scale interval corresponding to each state data subset is determined.

[0041] For each time scale interval, extract the runtime data subset corresponding to the time scale interval from each runtime sequence data;

[0042] Based on the degree of matching between each subset of operational data and the corresponding subset of state data for the time scale interval, a scheduling scheme for each power equipment is determined.

[0043] The operation and control of each power equipment are carried out according to the dispatching plan in order to make the target power grid reach a stable operating state.

[0044] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0045] Acquire the state timing data of the target power grid, as well as the operating timing data of at least one power device in the target power grid;

[0046] The state time series data is divided into multiple state data subsets according to different time scales, and the time scale interval corresponding to each state data subset is determined.

[0047] For each time scale interval, extract the runtime data subset corresponding to the time scale interval from each runtime sequence data;

[0048] Based on the degree of matching between each subset of operational data and the corresponding subset of state data for the time scale interval, a scheduling scheme for each power equipment is determined.

[0049] The operation and control of each power equipment are carried out according to the dispatching plan in order to make the target power grid reach a stable operating state.

[0050] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0051] Acquire the state timing data of the target power grid, as well as the operating timing data of at least one power device in the target power grid;

[0052] The state time series data is divided into multiple state data subsets according to different time scales, and the time scale interval corresponding to each state data subset is determined.

[0053] For each time scale interval, extract the runtime data subset corresponding to the time scale interval from each runtime sequence data;

[0054] Based on the degree of matching between each subset of operational data and the corresponding subset of state data for the time scale interval, a scheduling scheme for each power equipment is determined.

[0055] The operation and control of each power equipment are carried out according to the dispatching plan in order to make the target power grid reach a stable operating state.

[0056] The aforementioned power grid operation control method, apparatus, computer equipment, computer-readable storage medium, and computer program product first acquire the state-time series data of the target power grid and the operation-time series data of at least one power device in the target power grid. Next, the state-time series data is divided into multiple state data subsets according to different time scales, and the time scale interval corresponding to each state data subset is determined. For each time scale interval, the operation data subset corresponding to the time scale interval is extracted from each operation-time series data. Based on the degree of matching between each operation data subset and the state data subset corresponding to the time scale interval, a scheduling scheme for each power device is determined. Finally, the operation control of each power device is performed according to the scheduling scheme to achieve a stable operating state for the target power grid. Thus, by introducing a multi-time scale division and refined matching mechanism, this scheme can automatically match the power device with the most suitable characteristics for regulation in response to power grid disturbances at different time scales, thereby significantly improving the accuracy, efficiency, and reliability of power grid operation control. It is particularly suitable for modern power systems containing a high proportion of fluctuating renewable energy, effectively enhancing the stability and intelligence level of power grid operation. Attached Figure Description

[0057] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0058] Figure 1 This is an application environment diagram of the power grid operation control method in one embodiment;

[0059] Figure 2 This is a flowchart illustrating a power grid operation control method in one embodiment;

[0060] Figure 3 This is a schematic diagram of the data partitioning process in one embodiment;

[0061] Figure 4 This is a schematic diagram of the data partitioning process in another embodiment;

[0062] Figure 5 This is a flowchart illustrating the optimization of a scheduling scheme in one embodiment;

[0063] Figure 6 This is a flowchart illustrating the scheduling scheme optimization in another embodiment;

[0064] Figure 7 This is a schematic diagram of the operation control process in one embodiment;

[0065] Figure 8 This is a structural block diagram of a power grid operation control device in one embodiment;

[0066] Figure 9 This is an internal structural diagram of a computer device in one embodiment;

[0067] Figure 10 This is a diagram of the internal structure of a computer device in another embodiment. Detailed Implementation

[0068] In modern power systems, the large-scale integration of renewable energy sources such as wind and solar power places higher demands on the stable operation of the power grid, particularly in terms of precise control over the grid's operating frequency. Because renewable energy output is intermittent and fluctuating, the power grid requires more flexible and efficient control measures to cope with the increasingly complex operating environment.

[0069] However, current widely adopted power grid operation and control methods largely rely on fixed time scale divisions and static resource allocation. When facing complex scenarios involving the coupling of multiple energy types such as wind power, photovoltaics, and energy storage, these methods struggle to flexibly adapt to the varying characteristics of different energy devices in terms of response speed, regulation accuracy, and duration. This leads to situations where, when faced with sudden power fluctuations, some slow-responding devices cannot function promptly, while others may experience over-allocation of resources. This imbalance directly impacts the stability of the power grid. Furthermore, traditional methods handle time scales too coarsely; a simple, uniform time window cannot effectively capture fluctuation characteristics across different time scales, from milliseconds to hours. For example, at certain critical moments, millisecond-level frequency deviations may escalate into greater system risks due to a lack of ultra-fast-responding energy devices. This mismatch in time scales further wastes resources, preventing the full exploitation and synergistic utilization of the potential of various energy devices. Therefore, breaking through the constraints of fixed time frames, achieving refined division of operation and control time scales, and accurately matching energy devices with different characteristics to fluctuation demands at corresponding time scales has become a core key to improving the stability and operational efficiency of power grids with high proportions of renewable energy.

[0070] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0071] It should be noted that the terms "first," "second," etc., used in this application can be used to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish the first element from the second element. The terms "comprising" and "having," and any variations thereof, used in this application, are intended to cover non-exclusive inclusion. The term "multiple" used in this application refers to two or more. The term "and / or" used in this application refers to one of the embodiments, or any combination of multiple embodiments.

[0072] The power grid operation control method provided in this application embodiment can be applied to, for example, Figure 1 In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104, or it can be located on the cloud or other network servers. Terminal 102 can be, but is not limited to, at least one of various personal computers, laptops, smartphones, tablets, drones, low-altitude aircraft, IoT devices, etc. Server 104 serves as the control center of the power system and can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services.

[0073] This application embodiment can be executed independently by terminal 102 or server 104, or interactively by terminal 102 and server 104. The following description uses the scenario of interactive execution by terminal 102 and server 104 as an example. Specifically, after receiving a power grid operation control command initiated by the user through terminal 102, server 104 first obtains the state timing data of the target power grid and the operation timing data of at least one power device in the target power grid. Then, it divides the state timing data into multiple state data subsets according to different time scales and determines the time scale interval corresponding to each state data subset. For each time scale interval, server 104 extracts the operation data subset corresponding to the time scale interval from each operation timing data set, thereby determining a scheduling scheme for each power device based on the degree of matching between each operation data subset and the state data subset corresponding to the time scale interval. Finally, server 104 performs operation control on each power device according to the scheduling scheme to ensure the target power grid reaches a stable operating state.

[0074] In one exemplary embodiment, such as Figure 2 As shown, a power grid operation control method is provided, which is applied to... Figure 1 Taking server 104 as an example, the explanation includes the following steps:

[0075] Step S202: Obtain the state timing data of the target power grid and the operation timing data of at least one power device in the target power grid.

[0076] The target power grid is a power network encompassing power generation, transmission, distribution, and consumption units. State-time series data refers to the continuous time-varying sequence of core parameters reflecting the overall operational status of the power grid. These core parameters include, but are not limited to, at least one of the following: grid frequency, voltage, and power. For example, the nominal value of grid frequency is 50 Hz (Hertz), and any instantaneous deviation directly reflects the real-time imbalance between power generation and consumption within the grid. State-time series data can be collected through synchronous phasor measurement units (PMUs) or high-precision sensors deployed at key nodes of the power grid, such as substations. Sampling frequencies can reach 1000 times per second or even higher to capture millisecond-level instantaneous changes. Each data point in the state-time series data contains a precise timestamp and a corresponding measurement value. Power equipment refers to power generation or energy storage equipment connected to the target power grid and whose operating status is controllable, including, but not limited to, at least one of the following: renewable energy power generation equipment (such as wind power generation equipment and photovoltaic power generation equipment), energy storage equipment (such as battery energy storage equipment), and gas-fired power generation equipment. Runtime sequence data refers to the sequence of changes in the key operating parameters of each of the above-mentioned power devices over time. Key operating parameters include, but are not limited to, at least one of the following: power output, voltage, current, etc.

[0077] For example, the server uses synchronized phasor measurement units or high-precision sensors deployed at key nodes of the target power grid to perform high-frequency sampling of core parameters reflecting the power grid's operating status, such as the power grid operating frequency. Through this high-frequency sampling, the server can acquire dynamic data on the continuous changes of core parameters from the real-time operation of the power grid, i.e., state-time series data. Simultaneously, the server synchronously or periodically acquires key operating parameters from the local monitoring units of each power device in the target power grid. These parameters form runtime series data describing the time-varying output of each device. To ensure the accuracy of subsequent analysis, the server can further synchronize and align the aforementioned state-time series data with the runtime series data, ensuring that data from different sources remain consistent in timestamps.

[0078] Optionally, after acquiring the state time-series data, the server can identify the regularity and anomalies in the changes of the power grid state from the data. Specifically, the server can use time series analysis methods, such as autoregressive integral moving average models, to identify hidden periodic patterns in the data, such as daily or hourly fluctuations. Simultaneously, it can use anomaly detection algorithms, such as the Isolation Forest algorithm, to locate and mark anomalous data points that significantly deviate from the normal range, such as sudden frequency drops caused by faults. Through comprehensive analysis of these periodic patterns and anomaly location, the server can determine the characteristic distribution pattern of power grid state fluctuations.

[0079] Optionally, the server can further perform time alignment and correlation analysis on the state timing data and runtime timing data. By comparing the changes of the two at the same time, the server can determine the response characteristics of specific power equipment output changes to grid state fluctuations. For example, whether there are response delays on the order of seconds or mismatches in the direction of power changes. If such unsatisfactory responses are identified, the server will quantitatively calculate the theoretical power adjustment required to eliminate the mismatch and define a quantitative synchronization index to assess the degree of fit between the power equipment response and grid fluctuations.

[0080] Optionally, the server can combine key feature parameters extracted from state time-series data (such as fluctuation amplitude, fluctuation occurrence time, fluctuation duration, etc.) with corresponding key feature parameters extracted from runtime time-series data (such as power adjustment amount, response delay, etc.), and add a synchronization index characterizing the degree of fit between the two to form a structured feature record. Each feature record fully describes the correlation event between a power grid fluctuation and equipment response, and the collection of these records constitutes a comprehensive feature dataset. Further, the server transforms this dataset into a feature vector matrix, where each feature vector represents a complete event, and its dimensions cover multiple attributes such as power grid state fluctuation, power equipment response, and synchronization. Statistical methods such as Pearson correlation coefficient are used to analyze the correlation strength between different feature dimensions. Finally, the server uses this correlation strength data to establish a predictive relationship from power grid state fluctuation characteristics to ideal equipment response parameters and synchronization targets through a fitting model such as a linear regression model, i.e., dynamic mapping. This mapping relationship, together with the core knowledge extracted from the two time series data, such as periodicity patterns, anomaly location, response characteristics, and synchronization patterns, constitutes a systematic set of fluctuation characteristics. This set not only records historical fluctuations and response patterns, but also embeds quantitative criteria for assessing the compatibility between equipment and fluctuations in future events, which can be directly used as the basis for supporting subsequent calculations of matching degree and time scale division.

[0081] For example, a server continuously monitors the power grid frequency using high-precision sensors, sampling up to 1000 times per second, thereby obtaining state-time series data that accurately reflects instantaneous frequency changes. In an exemplary 24-hour monitoring dataset, the normal frequency fluctuation range is between 49.9 Hz and 50.1 Hz. To gain a deeper understanding of the fluctuation patterns, the server performs feature extraction on this data. First, in the periodicity identification stage, an autoregressive integral moving average model is used to calculate the autocorrelation function of the sequence. The analysis results show that the frequency fluctuations exhibit significant hourly periodic peaks, a pattern typically closely related to the periodic changes in power grid load. Second, in the anomaly detection stage, the isolated forest algorithm is applied to analyze the data, successfully identifying anomalies that significantly deviate from the normal range, such as a sudden drop in frequency to 49.5 Hz at a specific moment. Such anomalies are often associated with sudden failures in power generation equipment or transient network disturbances.

[0082] Analyzing the correlation between grid state fluctuations and equipment response reveals a significant relationship between power output changes in renewable energy equipment (such as wind farms) and grid frequency fluctuations. Specifically, assuming a sudden drop in wind farm output power is detected over a certain period, the grid frequency drops by approximately 0.2 Hz almost simultaneously. Further analysis of the response characteristics shows a delay of approximately 2 seconds between the start of the wind power decrease and the subsequent frequency drop, indicating a significant lag in the energy equipment's response to grid fluctuations. To improve this situation, the system attempts to adjust the wind power output signal by rapidly increasing the power output by approximately 10% as compensation after the sudden power drop. Calculations show that after implementing this compensation measure, the synchronicity index, reflecting the degree of following between the two, increased from 0.6 to 0.85, indicating a significant enhancement in the matching relationship between wind power output and grid frequency changes after the adjustment.

[0083] In the example of constructing a fluctuation feature set, the server extracts key parameters from a typical frequency fluctuation event, such as the grid fluctuation amplitude (0.3Hz) and equipment response delay (2 seconds), and integrates them with the corresponding wind power generation equipment's control parameters (such as a 10% power compensation adjustment) to form a structured feature record. Multiple such records constitute a feature vector matrix. By applying the Pearson correlation coefficient to this matrix for statistical analysis, the intrinsic connections between different feature dimensions can be discovered. For example, the correlation coefficient between fluctuation amplitude and power adjustment can reach 0.9, indicating a very strong linear correlation between the two. Based on this high correlation, the system can establish a quantitative predictive relationship from frequency fluctuations to expected power adjustments, i.e., a dynamic mapping, through model fitting such as linear regression. This mapping relationship, along with core knowledge such as periodicity, anomaly location, response characteristics, and synchronicity patterns extracted from two time-series data, together constitute the fluctuation feature set.

[0084] Step S204: Divide the state time series data into multiple state data subsets according to different time scales, and determine the time scale interval corresponding to each state data subset.

[0085] Here, different time scales refer to different time window lengths. It can be understood that power grid fluctuations often encompass multiple states, such as rapid second-level disturbances, medium-speed minute-level fluctuations, and slow hourly trends. By distinguishing different time scales, these diverse fluctuation patterns can be separated and identified. A state data subset refers to a continuous data segment from complete state time-series data, categorized into a specific time scale based on its fluctuation characteristics. For example, the original data corresponding to all periods identified as experiencing severe second-level fluctuations collectively constitute a second-level state data subset. A time scale interval refers to the specific start and end time range corresponding to each of the above state data subsets on the original time axis. It describes the time position of the data subset, for example: [14:30; 14:31]. One time scale interval corresponds to one state data subset.

[0086] For example, after obtaining the state time series data, the server first performs data cleaning and preprocessing to improve data quality. For instance, filtering algorithms can be used to remove high-frequency interference introduced by communication interruptions or sensor noise, or statistical threshold rules can be applied to identify and eliminate outliers that significantly exceed the physical operating range. Of course, other preprocessing methods can be selected depending on the specific circumstances in practical applications, and this embodiment does not impose any limitations on this. After preprocessing, the server can perform signal analysis and feature extraction on the current state time series data. Specifically, it converts the data to a frequency domain representation using frequency domain transformations such as Fourier transforms to identify the periodic patterns of power grid fluctuations. Simultaneously, statistical characteristics, such as fluctuation variance and average rate of change, can be calculated in the time domain. These time-frequency domain features will be integrated and incorporated into a more structured set of fluctuation features. Subsequently, an unsupervised clustering algorithm, such as the K-means algorithm, is used to automatically group the continuous state time series data based on the aforementioned core features. The algorithm calculates the similarity of fluctuation characteristics across different time periods and aggregates data points with similar characteristics and temporal continuity into the same cluster. Each cluster initially corresponds to a specific fluctuation pattern. Next, the server evaluates each initially formed cluster (initial time segment), such as calculating the difference between the feature variance of data points within the cluster and a preset threshold. If this difference indicates that the fluctuations within the cluster are still mixed or the distribution is too discrete, algorithms capable of nested partitioning, such as hierarchical clustering, are invoked to further segment the time period corresponding to that cluster, resulting in more consistent sub-clusters. Finally, the server integrates all the initial clusters that pass the evaluation and the sub-clusters generated from the subdivision, extracting corresponding data segments from the original state time-series data based on their start and end timestamps. These data segments constitute the multiple subsets of state data obtained from the final partitioning, and the time range covered by each subset is explicitly recorded as the corresponding time scale interval.

[0087] Step S206: For each time scale interval, extract the runtime data subset corresponding to the time scale interval from each runtime sequence data.

[0088] In this context, the operational data subset refers to a data segment precisely extracted from the complete operational sequence data of a specific power grid device, whose time range is completely consistent with a given time scale interval. For example, for energy storage power station B, the power data within the interval [14:30; 14:31] constitutes the operational data subset corresponding to that interval.

[0089] For example, for each defined time scale interval, the server first extracts the corresponding subset of operational data from the complete operational sequence data of each power grid device using data slicing technology, based on the start and end timestamps of that interval.

[0090] For example, suppose the runtime sequence data records the power changes of electrical equipment over 24 hours. The time scale interval can be divided into four data segments, each containing the specific numerical value of the power change. When extracting the response speed parameter within each segment, the average slope can be determined by observing the trend of power change within each 6-hour segment.

[0091] In some embodiments, to gain a deeper understanding of the dynamic response characteristics of each device within a given time period, the server further performs fine-grained analysis on each subset of operational data. Specifically, the server first divides the power curve within each subset of operational data into shorter data segments of fixed duration, such as every 30 seconds, forming a first set of data segments. For each segment, the server calculates the average slope of its power change curve using numerical analysis methods such as least-squares fitting, and quantifies this slope as the response speed parameter of the device within that segment. Thus, for each device within a specific time scale interval, a set of response speed parameters reflecting its response speed at different times can be obtained.

[0092] The server then analyzes the distribution of these response speed parameters, identifying and marking anomalous parameters that significantly deviate from the normal range, such as segments with excessively slow response times, by comparing each parameter with a set reasonable threshold range based on historical data. If the distribution of anomalous parameters exhibits a specific pattern, such as being concentrated in the latter half of the time period, it may indicate a systematic deviation between the original timescale interval division and the actual response pattern of the device. In this case, the server triggers an adaptive adjustment mechanism, that is, it re-divides the relevant device's operational data subset based on the distribution characteristics of the anomalous parameters using a dynamic time segmentation algorithm, generating a more reasonable second set of data segments, and recalculates its response speed parameters.

[0093] For example, assuming a segment's response rate parameter is 5 units per hour, and the preset threshold range is 3 to 4 units per hour, this segment can be marked as an anomalous parameter. Further analysis of the distribution of anomalous parameters reveals that these outliers are concentrated in a specific time period, such as during the night when load is low. This may indicate deviations in data acquisition or system operation. If the deviation value of a nighttime segment reaches 20%, exceeding the preset 10% range, the relevant segment needs to be rescaled, for example, by adjusting 6 hours to 3 hours, and a new set of data segments needs to be generated. This adjustment helps to capture changes in response characteristics more precisely.

[0094] Based on the optimized response speed parameters, the server can further quantify the matching degree between the subset of operational data and the corresponding subset of state data for each device. Specifically, the server first extracts fluctuation characteristics, such as fluctuation rate and amplitude, from the subset of state data corresponding to the given time scale. These characteristics define the ideal response speed required to smooth fluctuations at that time scale, i.e., the expected fluctuation demand. Then, statistical correlation analysis methods, such as the Pearson correlation coefficient, are used to evaluate the matching degree between the actual response speed of each device within that interval and the aforementioned expected fluctuation demand, thereby calculating a quantified matching degree value. All devices are sorted according to their matching degree values, forming a matching candidate set for that time scale interval. Each record in this matching candidate set clearly represents the quantified matching relationship between the subset of operational data and the corresponding time scale interval, which can also be understood as the quantified matching relationship between the subset of operational data and the corresponding subset of state data. This provides a direct and clear decision-making basis for subsequently determining the scheduling scheme; that is, the scheduling scheme will prioritize calling devices with high matching degrees to handle grid fluctuations at the corresponding time scale.

[0095] For example, suppose that when re-extracting the response speed parameter and using the correlation coefficient to evaluate the matching degree, the matching degree of the response speed parameter of a certain segment with the current time scale interval is 0.8, indicating a strong correlation between the two. A decision tree classifier is used to classify the matching degree of each segment, eliminating candidates with a matching degree below 0.5 to ensure the reliability of the final candidate set. Finally, the response speed parameter and corresponding matching degree value for each time scale interval can be integrated into a data table for storage, facilitating subsequent querying and analysis.

[0096] Step S208: Based on the degree of matching between each subset of operating data and the corresponding subset of state data for the time scale interval, determine the scheduling scheme for each power equipment.

[0097] Among them, the matching degree comprehensively reflects the level of adaptation between the response characteristics of the equipment and the grid fluctuation demand within a specific time scale interval. The scheduling scheme is an executable instruction plan that specifies the active power setpoint or other key control parameters that each grid device should achieve at various time points or time segments within a future scheduling cycle. It aims to coordinate the devices to optimally mitigate predicted or impending grid fluctuations.

[0098] For example, the server first extracts the response speed parameters of each device from the matching candidate set formed based on the matching degree, and compares them with the minimum standard required to ensure the stable operation of the power grid, i.e., the preset lower threshold, to identify whether there are devices with substandard response capabilities. If such devices are found, the server can initiate a rapid correction. Specifically, the server obtains the scheduling ratio of these devices in the current scheduling scheme, i.e., the proportion of their power regulation tasks in the total regulation demand, and determines whether this ratio is lower than the preset ratio threshold. If not, it means that even if enough tasks are allocated, their performance is still substandard, and the problem may stem from the device's own capacity limit or temporary state. In this case, the server should consider reducing its scheduling ratio, or even temporarily removing it from the current scheduling tasks. If yes, it means that the current task allocation is insufficient, failing to fully stimulate or utilize the potential response capability of the device. For the latter case, the server aims to minimize the adjustment range of the overall scheduling ratio, i.e., the total deviation between the new scheme and the original scheme in the scheduling ratio of each device, while using the requirement that the response speed of all devices must meet the preset lower threshold as a hard constraint, and constructs a linear programming model. A linear programming solver is used to solve the model, and a preliminary scheduling scheme is output by adjusting the scheduling ratio of each device.

[0099] To further improve the robustness and cost-effectiveness of the solution, the server put the initial scheduling scheme into short-term simulation. If the simulation shows that the response speed parameters of some devices are still below the threshold, it indicates that the first round of static proportional adjustment based on linear programming may not have fully considered the dynamic relationship between device response performance and task load. Therefore, the server initiates a second round of fine optimization, that is, for this type of device, a predictive model reflecting the statistical relationship between its response speed parameters and the ideal scheduling ratio is trained using historical operating data. The suggested scheduling ratio output by this model is used as the new decision-making basis to update the initial scheduling scheme and generate the final scheduling scheme.

[0100] Optionally, the prediction model can be a linear regression model, expressed as: y = ax + b, where y is the suggested scheduling ratio, x is the measured or predicted response speed parameter, a represents the slope, i.e., the amount of change in the scheduling ratio required to maintain stable compliance for each unit increase in the response speed parameter, and b represents the intercept, i.e., the theoretical basic scheduling ratio.

[0101] For example, in a specific scenario, the matching candidate set contains 10 elements, namely power devices, and the response speed parameters of each device are obtained through real-time monitoring. Assuming a preset lower threshold of 5 units per second, the response speeds of three of these elements are 3.2, 4.1, and 4.8 units per second, respectively, all below the threshold. This result indicates that, under the current operating conditions, the rapid adjustment capability of these devices is insufficient to handle the fluctuation smoothing task at the corresponding time scale. Therefore, the server will further analyze the scheduling ratio of these devices in the current scheduling scheme, that is, the proportion of the total demand to be allocated to these devices. Assuming a preset ratio threshold of 20%, the scheduling ratio of one device is only 15%, which is clearly insufficient. At this point, the scheduling ratio can be optimized and adjusted using a linear programming solver. The goal of linear programming is to ensure that the total deviation of the scheduling ratio of each device is minimized while satisfying the set constraints. After adjustment, the scheduling ratio of this element may increase to 22%, forming a preliminary allocation scheme. After generating the preliminary allocation scheme, the server can further determine the scheduling priority of each device. Assuming the device with the lowest response speed is marked as high priority, it will be prioritized in the scheduling ratio adjustment. Subsequently, the response speed parameters of each device are obtained from the simulated operation of the initial allocation plan. If the response speed of any device is still below the threshold, the relationship between the speed parameter and the scheduling ratio needs further analysis. For example, historical data analysis might show that for every 1 unit per second increase in the speed parameter, the scheduling ratio needs to be increased by approximately 2 percentage points. Based on this statistical relationship, the server can recalibrate the initial allocation plan to generate the final scheduling plan. Furthermore, adjustments to the scheduling ratio can be dynamically optimized in conjunction with real-time load changes. For instance, during peak periods, priority can be given to allocating core regulation tasks to critical power equipment with the fastest response speed and highest matching degree, thereby maximizing the grid's resilience during critical periods.

[0102] Step S210: Perform operation control on each power equipment according to the dispatching plan to ensure that the target power grid reaches a stable operating state.

[0103] Among them, stable operation state refers to the stable operation state of the power grid, which is typically manifested as a dynamic equilibrium state in which the power grid operating frequency is maintained near the rated value and the fluctuation amplitude and duration are within the safe standard range.

[0104] For example, after obtaining the scheduling scheme, the server can generate the corresponding control instructions and send the instructions to the corresponding power equipment to control the operation of each power equipment, thereby enabling the target power grid to reach a stable operating state.

[0105] In this embodiment, the server first acquires the state time-series data of the target power grid and the runtime time-series data of at least one power device in the target power grid. Then, the state time-series data is divided into multiple state data subsets according to different time scales, and the time scale interval corresponding to each state data subset is determined. For each time scale interval, the runtime data subset corresponding to the time scale interval is extracted from each runtime time-series data. Based on the degree of matching between each runtime data subset and the state data subset corresponding to the time scale interval, a scheduling scheme for each power device is determined. Finally, the operation control of each power device is performed according to the scheduling scheme to ensure the target power grid reaches a stable operating state. Thus, by introducing a multi-time-scale division and refined matching mechanism, this embodiment can automatically match the power device with the most suitable characteristics for regulation in response to power grid disturbances at different time scales, thereby significantly improving the accuracy, efficiency, and reliability of power grid operation control. It is particularly suitable for modern power systems containing a high proportion of fluctuating renewable energy, effectively enhancing the stability and intelligence level of power grid operation.

[0106] In one exemplary embodiment, such as Figure 3 As shown, the state time series data is divided into multiple state data subsets according to different time scales, including:

[0107] Step S302: Extract the state fluctuation characteristics of the target power grid from the state time series data.

[0108] Step S304: Based on the state fluctuation characteristics, the state time series data is divided into multiple state data subsets according to different time scales.

[0109] State fluctuation characteristics are data features used to characterize the fluctuation characteristics of the target power grid, and may include, but are not limited to, at least one of frequency domain characteristics, time domain statistical characteristics, and structural characteristics. Frequency domain characteristics may be one or more dominant fluctuation periods identified through Fourier transform, such as a 24-hour period and a 1-hour period. Time domain statistical characteristics may include, but are not limited to, at least one of the following: fluctuation variance characterizing fluctuation intensity, average rate of change characterizing fluctuation average velocity. Structural characteristics may include, but are not limited to, the distribution locations of stationary and disturbed segments identified through change point detection or cluster preprocessing.

[0110] For example, the server extracts state fluctuation features—specifically describing the fluctuation patterns of the state time series data itself—from the fluctuation feature set, and uses these features as key parameters to drive the time series segmentation and clustering process: First, the server creates an initial macroscopic segmentation framework on the time axis based on the identified main cycle, such as 24 hours. Then, within each macroscopic segment, using the state fluctuation features as clustering features, a clustering algorithm is used to group continuous time windows, grouping time periods with similar features into the same cluster. For significant disturbance segments marked by interval structure features, more granular segmentation can be performed. Finally, the continuous interval corresponding to each cluster or specific segment output by the algorithm on the original time axis is determined as the corresponding time scale interval. Based on the start and end timestamps of these intervals, the server extracts corresponding data segments from the original state time series data; these segments constitute the multiple state data subsets generated in the final segmentation.

[0111] For example, analyzing state-time series data using Fourier transform essentially involves converting the data into an energy distribution signal in the frequency domain. This transformation can clearly reveal hidden periodic patterns of fluctuation. In a typical analysis, the server might identify a significant 12-hour periodic fluctuation component in the frequency sequence. This component is usually closely related to the daily load cycle of the power grid, which is typically higher during the day and lower at night. By extracting this dominant cycle, the server can quickly pinpoint the key patterns driving long-term changes in the power grid's state, thus providing a fundamental basis for subsequent refined analysis and control strategy development based on multiple time scales.

[0112] When using clustering algorithms for grouping, daily frequency data can be divided into several main categories based on fluctuation amplitude and time period, such as morning peak, daily stable period, and evening peak. In a typical analysis, the frequency fluctuation amplitude of the morning peak is relatively large, concentrated between 7 am and 9 am. Clustering results can clearly distinguish the fluctuation characteristics of different time periods, helping to understand the performance of frequency fluctuations at different time scales.

[0113] In this embodiment, feature extraction is performed on the power grid status data to accurately identify its inherent fluctuation patterns. Based on these features, adaptive multi-timescale division is achieved, effectively separating the mixed fluctuations according to different modes, laying the foundation for subsequent matching, thereby significantly improving the precision of power grid operation control.

[0114] In one exemplary embodiment, such as Figure 4 As shown, based on the characteristics of state fluctuations, the state time series data is divided into multiple state data subsets according to different time scales, including:

[0115] Step S402: Based on the state fluctuation characteristics, the state time series data is divided into multiple initial data subsets according to different time scales.

[0116] The initial data subset refers to the set of data segments obtained after preliminary division of the state time series data. At this point, the division may still be relatively coarse.

[0117] For example, the server performs preliminary data segmentation based on the extracted state fluctuation characteristics, such as the identified dominant cycle. For instance, if a significant 12-hour cycle is identified, the system may first use 12 hours as a benchmark to divide the long-term state time series data into several macroscopic daily cycle segments as a coarse-grained initial segmentation, aiming to quickly build a preliminary data view based on global fluctuation patterns.

[0118] Step S404: For each initial data subset, extract the fluctuation characteristics corresponding to the initial data subset.

[0119] Among them, the fluctuation characteristics corresponding to the initial data subset refer to the more refined local characteristics within each initial data subset, such as at least one of the frequency fluctuation amplitude, frequency peak, and data distribution skewness within the segment.

[0120] For example, the server performs independent feature analysis on each initial data subset, that is, for each data subset, it calculates its local statistical characteristics. These local fluctuation characteristics are used to more accurately evaluate the fluctuation characteristics within the initial data subset, thereby determining whether it can be used as a valid time scale interval.

[0121] Step S406: If the fluctuation characteristics do not meet the expected fluctuation conditions, the time scale interval to which the initial data subset belongs is divided into multiple sub-time intervals.

[0122] The expected volatility condition is a preset quantitative standard used to assess the stability of volatility within an initial subset of data. For example, a local volatility amplitude threshold can be set; if the volatility amplitude within a subset exceeds this threshold, it is considered to contain multiple volatility patterns and requires further subdivision. A sub-time interval refers to a shorter time segment generated after subdividing the initial timescale interval that does not meet the expected volatility condition.

[0123] For example, the server compares the local fluctuation characteristics of each initial data subset with the expected fluctuation conditions. If the local characteristics of a subset indicate that its internal fluctuations are still complex or inconsistent, the initial partitioning of that subset is deemed insufficiently refined. In this case, the server initiates a second round of partitioning for the initial timescale interval corresponding to this specific initial data subset. For instance, it may use sliding window analysis or a finer-grained clustering method to further divide the original time-series data within this interval into multiple sub-time intervals with more uniform fluctuation patterns and shorter durations.

[0124] For example, during the time-scale segmentation process, if the frequency fluctuation amplitude is detected to be more than twice the normal fluctuation level during the evening peak period, it is determined that the fluctuation pattern within that interval is still too complex and requires more granular analysis. For instance, a 2-hour evening peak period may be subdivided into multiple 15-minute sub-intervals. Through this subdivision, the system can identify particularly severe fluctuations within a certain 30-minute sub-interval. Combining this with power grid event logs, it can be preliminarily inferred that this abnormal fluctuation may be related to the concentrated startup of a large industrial device. Through refined analysis, the ability to accurately identify and locate complex fluctuation periods is significantly improved.

[0125] When integrating more granular grouping results, the categorized data from all time intervals can be aggregated to form a panoramic view from macro to micro. For example, daily frequency fluctuations can be divided into 3 major categories and 12 subcategories, ensuring that each subcategory reflects a specific fluctuation pattern. This integration method facilitates a comprehensive understanding of the dynamic changes in the frequency sequence. Furthermore, detailed descriptions of fluctuation patterns can be generated for the final time scale intervals. For instance, the morning peak fluctuation pattern can be described as high-frequency, small-amplitude oscillations, while the evening peak fluctuation pattern is described as low-frequency, large-amplitude oscillations. This provides clear data support for subsequent power grid dispatching and optimization decisions.

[0126] Step S408: The initial data subset that meets the expected fluctuation conditions, and the target data subsets corresponding to each of the multiple sub-time intervals, are jointly determined as multiple state data subsets.

[0127] The target data subset refers to the corresponding data fragments that are re-extracted from the original state time series data based on the start and end timestamps of the sub-time interval.

[0128] For example, initial data subsets that meet the expected fluctuation conditions will be directly adopted as part of the final output. For initial data segments that do not meet the expected fluctuation conditions and have been subdivided into multiple sub-time intervals, the server will extract corresponding data fragments from the original state time series data based on the precise start and end times of each sub-time interval; these fragments are the target data subsets. Finally, the server merges all initial data subsets that meet the conditions with all target data subsets generated by refinement to form a complete set of data fragments, which are the multiple state data subsets obtained from the final partitioning. This process ensures high consistency within each final data subset, thereby achieving truly fine-grained and adaptive time-scale partitioning.

[0129] In this embodiment, through a multi-level process of preliminary division, local evaluation, conditional subdivision, and result integration, the initial time period with complex internal fluctuations is automatically identified and broken down. This achieves dynamic optimization and fine adjustment of the time scale division, ensuring that each subset of state data obtained has a high degree of consistency in fluctuation characteristics. This significantly improves the matching accuracy between the time scale and the actual fluctuation pattern of the power grid, providing a more reliable time dimension basis for subsequent matching.

[0130] In one exemplary embodiment, such as Figure 5 As shown, based on the degree of matching between each subset of operational data and the corresponding subset of state data for each time scale interval, a scheduling scheme for each power device is determined, including:

[0131] Step S502: Based on the degree of matching between each subset of operating data and the corresponding subset of state data for the time scale interval, schedule each power device to obtain a preliminary scheduling scheme.

[0132] Step S504: Based on the preliminary scheduling plan, simulate the operation of each power equipment to obtain the simulation results.

[0133] Simulation operation refers to the process of projecting the dynamic operation of the target power grid over a future period of time in a digital twin environment or offline simulation platform, based on the equipment output commands set in the preliminary scheduling plan and combined with data such as power grid models, load forecasts, and renewable energy power forecasts. Simulation operation results refer to a series of time-series data reflecting the simulated operating status of the power grid, output after the simulation, including but not limited to at least one of the following: simulated frequency curves, simulated output curves of each device, line power, and voltage. These data can be used to evaluate the actual effectiveness of the preliminary plan.

[0134] For example, after obtaining a preliminary scheduling plan, the server can load the plan into a power grid dynamic simulation program. This program includes a detailed power grid topology model and power equipment model. The simulation program uses the current power grid state as the initial condition and strictly follows the instructions issued by the preliminary plan to drive the operation of each virtual power device in the model. The simulation program will predict the continuous changes in the power grid state over the next few minutes to tens of minutes and record key simulation results, especially the simulation trajectory of the power grid frequency and the load changes of each virtual power device.

[0135] For example, during simulation, the server can first extract and categorize the equipment information of each power device from the preliminary scheduling plan. Assuming a power system manages multiple distributed power sources, including wind, solar, and energy storage devices, categorizing these resources into a structured resource list based on type, geographical location, and available capacity can provide clear data support for subsequent simulation and analysis. Next, the server inputs this list along with the detailed control commands from the preliminary scheduling plan into the power grid dynamic simulation program. The program then builds corresponding virtual models based on the equipment parameters in the list and drives these models according to the time-series commands in the plan, thereby simulating the overall operating state of the power grid over a future period. The simulation program will record and output key operational trajectory data, such as the simulation curve of the power grid frequency and the load changes of each virtual power device. By analyzing this simulation data, the server can intuitively identify potential operational risks. For example, during a simulated evening peak electricity consumption period, the simulation results showed that the load of a wind farm cluster in a certain area suddenly increased to over 80% in a short period of time, while the load rate of energy storage equipment was only 30% at the same time. This comparison clearly shows that under this initial scheduling scheme, wind power equipment may face the risk of overload. This simulation process is beneficial for identifying potential operational problems and provides a basis for subsequent optimization.

[0136] Step S506: Based on the simulation results, perform an operational risk analysis to obtain the operational risks corresponding to the preliminary scheduling scheme.

[0137] Step S508: Optimize the preliminary scheduling plan based on operational risks to obtain the final scheduling plan.

[0138] Operational risk analysis refers to the process of identifying potential risk points from simulation results that may affect the safe and stable operation of the power grid or indicate that the dispatching effect is not up to standard. Operational risk refers to the potential risk points identified through operational risk analysis.

[0139] For example, the server analyzes the load changes of each device during the simulation process, comparing key indicators such as real-time load rate and adjustment rate with preset safety thresholds. Through this process, the server can extract abnormal data points that exceed the thresholds, i.e., potential risk points, such as identifying a device whose load rate consistently exceeds 80% during a specific period. These potential risk points directly represent potential operational bottlenecks (such as equipment overload) or unstable factors (such as insufficient responsiveness). The server associates each potential risk point with a specific device and time scale, thereby generating a structured operational risk report. For example, assuming a load threshold of 75% is set, if a wind power device consistently exceeds 80% load at a certain moment, or even reaches 90% during certain periods, it is marked as an abnormal point.

[0140] In some embodiments, for detected potential risk points, optimal scheduling adjustment can be predicted. This involves using historical load data corresponding to the potential risk point, and at least one environmental factor such as wind speed and solar intensity as a feature vector, to predict the recommended adjustment range for the scheduling ratio of the equipment. For example, for wind power equipment identified as having overload risk, the model might suggest reducing its planned output by 10% during the relevant period to avoid overload risk, while simultaneously increasing the output of other power equipment, such as energy storage equipment, to 40% to balance the overall load. The resulting scheduling scheme can effectively improve the operational stability of the power system.

[0141] In one exemplary embodiment, such as Figure 6 As shown, based on operational risks, the initial scheduling scheme is optimized to obtain a new scheduling scheme, including:

[0142] Step S602: Identify the type of operational risk.

[0143] Among them, risk type refers to the standardized classification of operational risks.

[0144] For example, the server pre-builds a risk classification knowledge base based on historical data to achieve standardized risk identification. Specifically, the server calls the historical risk case library to classify at least one characteristic of the current potential risk point, such as the equipment involved, the type of constraint violation, and environmental relevance. Through feature matching, it is assigned to a category in a predefined risk category set, such as at least one of the following: equipment overload, response delay, or external disturbance.

[0145] For example, by analyzing the scheduling records of the past year, potential risks are divided into three categories: equipment failure, overload, and external interference. Among them, equipment failure accounts for 40%, overload accounts for 35%, and external interference accounts for 25%.

[0146] Step S604: Obtain the historical scheduling configuration information corresponding to the risk type, as well as the time characteristics of the occurrence of the operational risk.

[0147] Historical dispatch configuration information refers to dispatch strategy templates or parameter configuration sets that have proven effective in similar risk scenarios for a specific risk type. Temporal characteristics refer to the specific manifestation of currently identified potential risk points in the time dimension, including but not limited to at least one of the following: the specific time of risk occurrence, the time scale interval in which the risk occurs, the duration of the risk, and its correlation with the periodic patterns of power grid operation.

[0148] For example, based on the determined risk category, the server retrieves the corresponding response device table from historical scheduling records. This table records valid scheduling records for various types of devices under the same type of risk. Simultaneously, the server employs a time-based analysis method to perform in-depth risk analysis, accurately extracting at least one temporal characteristic, such as the specific time of occurrence, duration, and time scale interval of the current risk. If a temporal mismatch is found between the historical scheduling configuration and the risk distribution—for example, if the historical scheduling configuration concentrates backup power equipment in the afternoon, while the risk is most prevalent during the morning peak—it means that directly applying the historical scheduling configuration may not be accurate enough and calibration based on temporal characteristics is necessary.

[0149] Step S606: Optimize the preliminary scheduling scheme based on historical scheduling configuration information and time characteristics to obtain the final scheduling scheme.

[0150] For example, if a temporal mismatch is found between the historical scheduling configuration and the risk distribution, the server calculates a dynamic adjustment based on the historical scheduling configuration and updates the configuration accordingly. The calculation expression is: ,in The suggested adjustment amount for the dispatch ratio of power equipment i, such as load adjustment. This represents the probability assessment value for risk category j in the current time period. The weighting coefficients vary with the time characteristic k. The server iteratively updates the historical scheduling configuration based on the calculated adjustment amount. The updated configuration serves as the benchmark for the next analysis and comparison. This iterative process continues until the temporal matching degree between the configuration and the risk distribution meets preset requirements. Based on this, the initial scheduling scheme is updated, resulting in the updated scheduling scheme. Furthermore, the server uses the updated scheduling scheme as a new knowledge sample, updates the response device table in real time, and, based on the updated table and related information such as time scale intervals, uses machine learning algorithms such as decision trees to construct or optimize a refined matching model. This model can more accurately predict the scheduling matching effect of various power devices in responding to specific risk types under different time characteristics. To quantify the model's predictive accuracy, the server further introduces a matching accuracy value, A, as an evaluation metric. The formula is: A = (TP + TN) / (TP + TN + FP + FN), where TP represents the number of cases correctly predicted as high matches with good actual scheduling performance, TN represents the number of cases correctly predicted as low matches with poor actual scheduling performance, FP represents the number of cases incorrectly predicted as high matches with poor actual scheduling performance, and FN represents the number of cases incorrectly predicted as low matches with good actual scheduling performance. Subsequently, the server uses this refined matching model to re-evaluate and simulate the updated scheduling scheme. If the model's prediction score for the scheme's performance is not lower than a preset accuracy threshold, such as 0.9, the model is considered to have high reliability, and the updated scheduling scheme can be used as the final scheduling scheme and persistently stored.

[0151] In practical applications, when updating the response device table, the device status can be quickly updated through query statements. For example, if the status of a generator is changed from available to under maintenance, the server will immediately adjust the scheduling scheme and transfer the scheduling task of that generator to other devices.

[0152] In one exemplary embodiment, such as Figure 7 As shown, the operation control of each power equipment is carried out according to the dispatching plan, including:

[0153] Step S702: The dispatching plan is converted into preliminary control instructions and issued to each power equipment so that each power equipment operates in accordance with the preliminary control instructions.

[0154] Among them, the preliminary control instruction is a control command that is transformed from the dispatching scheme and can be directly issued to the power equipment for execution. It may include, but is not limited to, at least one of the following: adjusting power value, voltage value, frequency adjustment mode, execution timestamp, etc.

[0155] For example, after generating the final scheduling plan, the server can encapsulate the settings parameters in the plan into instructions that conform to the protocol, based on the device type and communication protocol. Subsequently, the instructions are sent to the corresponding power equipment via multicast or point-to-point. Upon receiving the instructions, the local controller parses and converts them into control signals, such as adjusting the gas turbine intake valve opening or changing the power reference value of the energy storage converter, driving the equipment to operate and thus causing the equipment's output power to change according to the instructions.

[0156] Optionally, the server can also use clustering algorithms such as k-nearest neighbors to perform similarity matching between historical frequency data and the current frequency, finding the reference point closest to the current frequency state, thereby generating preliminary control commands. For example, in a power grid system, if the current frequency is 49.8Hz, the server will deduce the most suitable command framework based on similar frequency values ​​in historical data, such as records of 49.7Hz and 49.9Hz, combined with the control commands at that time. This method can quickly respond to frequency fluctuations and ensure the rationality of the preliminary commands.

[0157] In some embodiments, the server can determine the priority of generating initial control commands based on the degree to which the power grid operating frequency deviates from the normal frequency; the greater the deviation, the higher the priority of command generation.

[0158] In step S704, if the target power grid does not reach a stable operating state during the operation of each power equipment, the parameters of the preliminary control command are calibrated to obtain the adjusted control command.

[0159] Parameter calibration refers to the process of online correction of at least one key control parameter in the initial control command, such as the power setpoint or adjustment range, based on the deviation between the actual and expected response of the power grid after the initial control command is executed. The adjusted control command refers to the new control command generated after calibration, which is more adapted to the dynamic characteristics of the actual power grid.

[0160] For example, after the command is issued, the server monitors dynamic signals such as the power grid operating frequency in real time. If the frequency continuously exceeds the safe range, it is determined that a stable state has not been reached. At this time, the frequency deviation curve and the actual output curve of each device can be analyzed to identify devices with lagging response or insufficient output, and then control parameters can be adjusted, such as increasing the command execution frequency or increasing power. Finally, a new set of command sequences is generated.

[0161] For example, if after a command is issued, the frequency does not recover to the target value of 50Hz but remains at 49.6Hz, it indicates that the system has not entered a stable operating state. In this case, a parameter calibration operation is initiated. Assuming the original command was to increase the power input by 10MW, after calibration, it is increased to 15MW. The server will monitor in real time whether the frequency recovers to 50Hz.

[0162] Step S706: Continue to control the operation of each power device according to the adjusted control instructions until the target power grid reaches a stable operating state.

[0163] For example, the server reissues the calibrated control commands to the corresponding devices, and the devices adjust their output according to the new commands. Simultaneously, the server continuously collects dynamic data such as grid frequency and voltage at a high sampling rate, and calculates their deviations from rated values ​​and their trends. After each round of data collection, the system evaluates in real time whether the current state has reached a stable state. If not, the parameter calibration steps are repeated. If the state is stable, the grid is determined to have entered a stable operating state, command iteration stops, and the final command parameters and stabilization time are recorded.

[0164] In this embodiment, the power grid can be quickly and accurately restored and maintained in a stable operating state through command calibration operations.

[0165] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. It is understood that the steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.

[0166] Based on the same inventive concept, this application also provides a power grid operation control device for implementing the power grid operation control method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more power grid operation control device embodiments provided below can be found in the limitations of the power grid operation control method described above, and will not be repeated here.

[0167] In one exemplary embodiment, such as Figure 8 As shown, a power grid operation control device is provided, comprising:

[0168] The data acquisition module 802 is used to acquire the state time-series data of the target power grid and the operation time-series data of at least one power device in the target power grid;

[0169] The data partitioning module 804 is used to divide the state time series data into multiple state data subsets according to different time scales, and determine the time scale interval corresponding to each state data subset.

[0170] The data extraction module 806 is used to extract a subset of runtime data corresponding to each time scale interval from each runtime sequence data for each time scale interval.

[0171] The scheduling scheme generation module 808 is used to determine the scheduling scheme for each power equipment based on the degree of matching between each subset of operating data and the subset of state data corresponding to the time scale interval.

[0172] The operation control module 810 is used to control the operation of each power equipment according to the dispatching plan so that the target power grid can reach a stable operating state.

[0173] In one embodiment, the data partitioning module 804 is further configured to:

[0174] Extract the state fluctuation characteristics of the target power grid from the state time series data;

[0175] Based on the characteristics of state fluctuations, the state time series data is divided into multiple state data subsets according to different time scales.

[0176] In one embodiment, the data partitioning module 804 is further configured to:

[0177] Based on the characteristics of state fluctuations, the state time series data is divided into multiple initial data subsets according to different time scales;

[0178] For each initial data subset, extract the fluctuation characteristics corresponding to the initial data subset;

[0179] If the fluctuation characteristics do not meet the expected fluctuation conditions, the time scale interval to which the initial data subset belongs is divided into multiple sub-time intervals;

[0180] The initial data subset that meets the expected fluctuation conditions, and the target data subsets corresponding to each of the multiple sub-time intervals, are collectively determined as multiple state data subsets.

[0181] In one embodiment, the scheduling scheme generation module 808 is further configured to:

[0182] Based on the degree of matching between each subset of operational data and the corresponding subset of state data for the time scale interval, each power device is scheduled to obtain a preliminary scheduling scheme.

[0183] Based on the preliminary scheduling plan, simulated operation was conducted on each power equipment to obtain the simulation results;

[0184] Based on the simulation results, an operational risk analysis was conducted to obtain the operational risks corresponding to the preliminary scheduling scheme.

[0185] Based on operational risks, the preliminary scheduling plan is optimized to obtain the final scheduling plan.

[0186] In one embodiment, the scheduling scheme generation module 808 is further configured to:

[0187] Identify the types of operational risks;

[0188] Obtain historical scheduling configuration information corresponding to risk types, as well as the time characteristics of operational risks.

[0189] Based on historical scheduling configuration information and time characteristics, the preliminary scheduling scheme is optimized to obtain the final scheduling scheme.

[0190] In one embodiment, the operation control module 810 is further configured to:

[0191] The dispatching plan is converted into preliminary control instructions and issued to each power equipment so that each power equipment operates in accordance with the preliminary control instructions;

[0192] If the target power grid does not reach a stable operating state during the operation of various power equipment, the parameters of the initial control command are calibrated to obtain the adjusted control command.

[0193] Continue to control the operation of each power device according to the adjusted control instructions until the target power grid reaches a stable operating state.

[0194] Each module in the aforementioned power grid operation control device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0195] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 9As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores power grid operation control data. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a power grid operation control method.

[0196] In one exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 10 As shown, the computer device includes a processor, memory, input / output interfaces, a communication interface, a display unit, and an input device. The processor, memory, and input / output interfaces are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The input / output interfaces are used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, Near Field Communication (NFC), or other technologies. When the computer program is executed by the processor, it implements a power grid operation control method.

[0197] Those skilled in the art will understand that Figure 9 or Figure 10 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0198] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0199] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0200] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0201] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0202] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0203] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible 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 application.

[0204] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A power grid operation control method, characterized in that, The method includes: Acquire the state timing data of the target power grid, and the operation timing data of at least one power device in the target power grid; The state time series data is divided into multiple state data subsets according to different time scales, and the time scale interval corresponding to each state data subset is determined. For each time scale interval, extract the runtime data subset corresponding to the time scale interval from each runtime sequence data; Based on the degree of matching between each of the operational data subsets and the state data subsets corresponding to the time scale interval, a scheduling scheme for each of the power devices is determined. The operation of each power device is controlled according to the scheduling scheme so that the target power grid can reach a stable operating state.

2. The method according to claim 1, characterized in that, The step of dividing the state time series data into multiple state data subsets according to different time scales includes: Extract the state fluctuation characteristics of the target power grid from the state time series data; Based on the aforementioned state fluctuation characteristics, the state time series data is divided into multiple state data subsets according to different time scales.

3. The method according to claim 2, characterized in that, Based on the state fluctuation characteristics, the state time series data is divided into multiple state data subsets according to different time scales, including: Based on the aforementioned state fluctuation characteristics, the state time series data is divided into multiple initial data subsets according to different time scales; For each of the initial data subsets, extract the fluctuation characteristics from the initial data subsets; If the fluctuation characteristics do not meet the expected fluctuation conditions, the time scale interval to which the initial data subset belongs is divided into multiple sub-time intervals; The initial data subset that meets the expected fluctuation conditions, and the target data subsets corresponding to each of the multiple sub-time intervals, are jointly determined as the multiple state data subsets.

4. The method according to claim 1, characterized in that, The step of determining a scheduling scheme for each power device based on the degree of matching between each subset of operational data and the subset of state data corresponding to the time scale interval includes: Based on the degree of matching between each of the operational data subsets and the state data subsets corresponding to the time scale intervals, the power equipment is scheduled to obtain a preliminary scheduling scheme. Based on the preliminary scheduling plan, the power equipment is simulated to obtain the simulation results. Based on the simulation results, an operational risk analysis is performed to obtain the operational risks corresponding to the preliminary scheduling scheme. Based on the operational risks, the preliminary scheduling scheme is optimized to obtain a new scheduling scheme.

5. The method according to claim 4, characterized in that, The optimization of the preliminary scheduling scheme based on the operational risks to obtain a new scheduling scheme includes: Identify the risk type of the operational risk; Obtain the historical scheduling configuration information corresponding to the risk type, as well as the time characteristics of the occurrence of the operational risk; Based on the historical scheduling configuration information and the time characteristics, the preliminary scheduling scheme is optimized to obtain a new scheduling scheme.

6. The method according to claim 1, characterized in that, The step of controlling the operation of each of the power devices according to the scheduling scheme includes: The scheduling scheme is converted into preliminary control instructions and issued to each of the power devices so that each of the power devices operates in accordance with the preliminary control instructions. If the target power grid fails to reach a stable operating state during the operation of each of the power devices, the parameters of the preliminary control command are calibrated to obtain the adjusted control command. The operation control of each power device continues according to the adjusted control instructions until the target power grid reaches a stable operating state.

7. A power grid operation control device, characterized in that, The device includes: The data acquisition module is used to acquire the state time-series data of the target power grid and the operation time-series data of at least one power device in the target power grid; The data partitioning module is used to divide the state time series data into multiple state data subsets according to different time scales, and determine the time scale interval corresponding to each state data subset. The data extraction module is used to extract a subset of runtime data corresponding to each time scale interval from each runtime sequence data for each time scale interval. The scheduling scheme generation module is used to determine a scheduling scheme for each of the power devices based on the degree of matching between each of the operating data subsets and the state data subsets corresponding to the time scale intervals. The operation control module is used to control the operation of each of the power devices according to the scheduling scheme, so as to make the target power grid reach a stable operating state.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images