Power grid load dynamic prediction and optimal scheduling method, device, equipment and medium

Abstract

The invention relates to the technical field of power distribution network dispatching. By providing a power grid load dynamic prediction and optimal scheduling method, device, equipment and medium, the method comprises the following steps: performing multi-source heterogeneous fusion processing on meteorological parameters, historical load curves and new energy output data to generate a dynamic load prediction map; constructing a dynamic network model, and performing power flow distribution simulation processing based on the dynamic network model to obtain a stability margin calculation result and a preset safety threshold boundary; performing stage decomposition processing on the global scheduling target to generate a progressive scheduling stage sequence; performing matching processing on the response characteristics of the power generation equipment to generate a self-adaptive progressive scheduling instruction sequence; and executing an adaptive progressive scheduling instruction sequence, performing feedback processing on the real-time state of the power grid, and generating a dynamic adjustment instruction so as to realize multi-dimensional data association modeling, stability margin quantitative analysis and dynamic instruction optimization, thereby improving the load prediction precision and reducing fault diffusion and transient oscillation.

Description

Technical Field

[0001] This invention relates to the field of power distribution network dispatching technology, and in particular to methods, devices, equipment and media for dynamic prediction and optimized dispatching of power grid load. Background Technology

[0002] With the accelerated construction of new power systems, the high proportion of new energy access and the rapid growth of diversified loads have placed higher demands on power grid dispatching technology.

[0003] However, the relevant power grid dispatching technologies have the following problems in the dynamic coupling modeling of multi-source heterogeneous data: the accuracy of dynamic load forecasting is limited by insufficient multi-dimensional correlation modeling of meteorological, new energy and historical loads; the dispatching model lacks a quantitative mapping relationship between the volatility of new energy and the dynamic stability margin of the power grid, which is prone to inducing the risk of cascading failures; the dispatching command generation mechanism is mismatched with the dynamic response characteristics of equipment, and the absence of closed-loop feedback mechanism leads to weakened transient oscillation suppression capability and dynamic deterioration of trajectory deviation. Summary of the Invention

[0004] Therefore, it is necessary to provide methods, devices, equipment and media for dynamic prediction and optimized scheduling of power grid load to address the above-mentioned technical problems, so as to realize multi-dimensional data correlation modeling, quantitative analysis of stability margin and dynamic command optimization, thereby improving the accuracy of load prediction and reducing fault propagation and transient oscillations.

[0005] Firstly, this application provides a method for dynamic forecasting and optimal scheduling of power grid load, the method comprising:

[0006] Acquire meteorological parameters, historical load curves, and renewable energy output data; perform multi-source heterogeneous fusion processing on the meteorological parameters, historical load curves, and renewable energy output data to generate dynamic load forecast maps;

[0007] The power grid topology is analyzed and processed to construct a dynamic network model. Based on the dynamic network model, power flow distribution simulation is performed to obtain the stability margin calculation results of each node and the preset safety threshold boundary.

[0008] Based on the dynamic load forecast map and stability margin calculation results, the global scheduling target is decomposed into stages to generate a progressive scheduling stage sequence.

[0009] The response characteristics of the power generation equipment are matched, and an adaptive progressive scheduling instruction sequence is generated based on the progressive scheduling stage sequence.

[0010] Execute the adaptive progressive scheduling instruction sequence and process the feedback of the real-time power grid status to generate dynamic adjustment instructions.

[0011] Furthermore, based on the dynamic load forecast map and stability margin calculation results, the global scheduling objective is decomposed into stages to generate a progressive scheduling stage sequence, including:

[0012] Using the following formula, based on the load gradient distribution characteristics and stability margin calculation results of the dynamic load forecast map, bottleneck areas are identified in the entire network transmission channels, generating a multi-dimensional risk heat map:

[0013]

[0014] Among them, L g This represents the load gradient index, where n represents the total number of nodes, and P... i V represents the active power at node i. i The voltage magnitude at node i is represented by ΔV. i S represents the voltage deviation. m λ represents the stability margin index. max λ represents the largest eigenvalue. min Let λ represent the smallest eigenvalue, m represent the total number of eigenvalues, and λ represent the smallest eigenvalue. k This represents the k-th eigenvalue;

[0015] Based on the spatiotemporal evolution trend of the multidimensional risk heatmap, the global scheduling objectives are prioritized for load transfer and segmented into time windows to generate a progressive scheduling phase sequence.

[0016] Furthermore, based on the load gradient distribution characteristics and stability margin calculation results of the dynamic load forecast map, bottleneck areas are identified in the entire network transmission channels, generating a multi-dimensional risk heat map, including:

[0017] Based on the load gradient distribution characteristics of the dynamic load forecast map, the trend analysis of the load change direction and rate of the entire network is performed to generate gradient evolution paths.

[0018] Based on the stability margin calculation results and gradient evolution path, the node carrying capacity attenuation region of the transmission channel is located and an initial bottleneck node set is generated.

[0019] Using the following formula, based on the topological connectivity of the initial bottleneck node set, a correlation analysis is performed on the power coupling strength of adjacent transmission channels to generate risk propagation paths:

[0020]

[0021] Among them, P ij α represents the power coupling strength between node i and node j. ij N represents the coupling coefficient. i Let d represent the set of neighboring nodes of node i. ibR represents the physical distance between node i and node b, σ represents the standard deviation parameter of the Gaussian kernel function, and R i β represents the risk propagation intensity at node i, β represents the risk propagation coefficient, and ω represents the risk propagation intensity at node i. ij t represents the weight coefficient between node i and node j. ij τ represents the signal propagation time, h represents the time decay constant, and h represents the total number of nodes in the network.

[0022] By integrating the gradient evolution path, the initial bottleneck node set, and the risk diffusion path, a multi-dimensional risk heat map is generated, which includes labels for risk level, propagation direction, and urgency.

[0023] Furthermore, the response characteristics of the power generation equipment are matched, and an adaptive progressive scheduling instruction sequence is generated based on the progressive scheduling stage sequence, including:

[0024] Dynamic capability mapping is performed on the response characteristics of the power generation equipment to generate a ramp rate constraint domain;

[0025] Based on the spatiotemporal domain partitioning results of the progressive scheduling phase sequence, the capacity matching degree is calculated for the load adjustment requirements of each sub-phase to generate the upper limit of the phase instruction amplitude.

[0026] Based on the climbing rate constraint domain and the upper limit of the stage instruction amplitude, the timing execution interval of the scheduling instructions is dynamically planned to generate instruction step size optimization parameters.

[0027] By integrating instruction step size optimization parameters with priority weights of the incremental scheduling phase sequence, an adaptive incremental scheduling instruction sequence is generated.

[0028] Furthermore, based on the ramp rate constraint domain and the upper limit of the stage instruction amplitude, dynamic programming is performed on the timing execution interval of the scheduling instructions to generate instruction step size optimization parameters, including:

[0029] Based on the margin decay rate of the ramp rate constraint domain and the load mutation threshold of the upper limit of the stage instruction amplitude, the timing interval of the scheduling instructions is subjected to backpropagation risk analysis to generate a safe execution time window.

[0030] Based on the boundary constraints of the safe execution time window, the instruction superposition effect between adjacent scheduling stages is dynamically divided into time windows to generate instruction execution interval thresholds.

[0031] By combining the safe execution time window and the instruction execution interval threshold, instruction step size optimization parameters are generated.

[0032] Furthermore, an adaptive progressive scheduling instruction sequence is executed, and feedback processing is performed on the real-time status of the power grid to generate dynamic adjustment instructions, including:

[0033] Based on the spatiotemporal constraints of the adaptive progressive scheduling instruction sequence, step-by-step instruction issuance is performed on the generator set and energy storage device to generate the initial scheduling execution trajectory.

[0034] Real-time status monitoring and processing are performed on the node voltage, frequency, and power flow of the initial scheduling execution trajectory to generate a real-time power grid status dataset;

[0035] Dynamic trajectory deviation analysis is performed on the real-time power grid status dataset and the expected scheduling path to generate trajectory deviation parameters and safety risk level labels.

[0036] Based on trajectory deviation parameters and safety risk level labels, the step size and direction of the adaptive progressive scheduling instruction sequence are dynamically compensated to generate dynamically adjusted instructions.

[0037] Furthermore, dynamic trajectory deviation analysis is performed on the real-time power grid status dataset and the expected scheduling path to generate trajectory deviation parameters and safety risk level labels, including:

[0038] Spatiotemporal alignment processing is performed on the real-time power grid status dataset to generate standardized status monitoring sequences that match the expected scheduling path;

[0039] Based on the time-domain waveform characteristics of the standardized state monitoring sequence and the expected scheduling path, dynamic trajectory similarity comparison processing is performed to generate trajectory deviation parameters.

[0040] The trajectory deviation parameter is processed by time window cumulative effect analysis to generate a risk diffusion trend prediction index.

[0041] Based on the risk diffusion trend prediction indicators and the preset safety threshold boundary, the deviation area is classified into risk levels to generate trajectory deviation parameters and safety risk level labels.

[0042] Secondly, this application also provides a device for dynamic forecasting and optimized dispatching of power grid load, the device comprising:

[0043] The multi-source data fusion module is used to acquire meteorological parameters, historical load curves and new energy output data, and to perform multi-source heterogeneous fusion processing on the meteorological parameters, historical load curves and new energy output data to generate dynamic load forecast maps.

[0044] The network modeling and power flow analysis module is used to analyze the power grid topology, construct a dynamic network model, simulate the power flow distribution based on the dynamic network model, and obtain the stability margin calculation results and preset safety threshold boundaries for each node.

[0045] The phase decomposition optimization module is used to decompose the global scheduling target into phases based on the dynamic load forecast map and the stability margin calculation results, and generate a progressive scheduling phase sequence.

[0046] The response characteristic matching module is used to match the response characteristics of the power generation equipment and generate an adaptive progressive scheduling instruction sequence based on the progressive scheduling stage sequence.

[0047] The dynamic feedback execution module is used to execute the adaptive progressive scheduling instruction sequence, process the real-time status of the power grid, and generate dynamic adjustment instructions.

[0048] 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 implement the steps of any of the methods in the first aspect of this application.

[0049] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of any of the methods in the first aspect of this application.

[0050] The technical solution provided in this application includes the following technical effects: By providing a method, device, equipment, and medium for dynamic prediction and optimized scheduling of power grid load, the method includes: acquiring meteorological parameters, historical load curves, and renewable energy output data; performing multi-source heterogeneous fusion processing on the meteorological parameters, historical load curves, and renewable energy output data to generate a dynamic load prediction map; performing analytical processing on the power grid topology to construct a dynamic network model; performing power flow distribution simulation processing based on the dynamic network model to obtain the stability margin calculation results and preset safety threshold boundaries for each node; performing stage decomposition processing on the global scheduling objective based on the dynamic load prediction map and stability margin calculation results to generate an incremental scheduling stage sequence; performing matching processing on the response characteristics of power generation equipment to generate an adaptive incremental scheduling command sequence based on the incremental scheduling stage sequence; executing the adaptive incremental scheduling command sequence and performing feedback processing on the real-time status of the power grid to generate dynamic adjustment commands, thereby realizing multi-dimensional data association modeling, stability margin quantitative analysis, and dynamic command optimization, thereby improving load prediction accuracy and reducing fault propagation and transient oscillations. Attached Figure Description

[0051] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0052] Figure 1 This is a flowchart of a power grid load dynamic prediction and optimal scheduling method in one embodiment of the present invention;

[0053] Figure 2 The flowchart illustrates the matching process for the response characteristics of power generation equipment in one embodiment of the present invention, which generates an adaptive progressive scheduling instruction sequence based on the progressive scheduling stage sequence.

[0054] Figure 3 This is a structural diagram of a power grid load dynamic prediction and optimization scheduling device according to one embodiment of the present invention. Detailed Implementation

[0055] To make the above-mentioned objects, features, and advantages of this application more apparent and understandable, the specific implementation methods of this application will be described in detail below with reference to the accompanying drawings. Many specific details are set forth in the following description to provide a full understanding of this application. However, this application can be implemented in many other ways different from those described herein, and those skilled in the art can make similar modifications without departing from the spirit of the application. Therefore, this application is not limited to the specific embodiments disclosed below.

[0056] like Figure 1 As shown, this application provides a method for dynamic forecasting and optimal scheduling of power grid load, which includes:

[0057] S101: Acquire meteorological parameters, historical load curves, and renewable energy output data; perform multi-source heterogeneous fusion processing on the meteorological parameters, historical load curves, and renewable energy output data to generate dynamic load forecast maps.

[0058] Specifically, real-time meteorological parameters, historical load curves, and renewable energy output data are collected. This data originates from multiple sources, including the power grid operation monitoring system, meteorological departments, the electricity trading market, and renewable energy power plants. It covers key information such as temperature, humidity, wind speed, historical electricity load data for different time periods, and power generation curves for renewable energy sources like photovoltaics and wind power. Next, multi-source heterogeneous fusion processing is performed. This includes data standardization, unifying data from different sources and formats into a standardized framework to ensure comparability and fusion. For example, the temporal resolution of meteorological data is adapted to that of electricity data, and feature extraction is performed to identify key feature dimensions that reflect the trend of power grid load changes. This includes extracting temperature and humidity indicators affecting electricity load from meteorological data, and extracting peak, trough, and fluctuation characteristics from historical load curves. Simultaneously, data correlation and integration are performed to establish a correlation model between meteorological parameters, historical load, and renewable energy output, analyzing their mutual influence relationships. Finally, a dynamic load forecast map is generated.

[0059] S102: Analyze the power grid topology, construct a dynamic network model, simulate power flow distribution based on the dynamic network model, and obtain the stability margin calculation results and preset safety threshold boundaries for each node.

[0060] Specifically, the process involves acquiring grid connection information, including line connection methods, substation distribution, power plant access points, and load center locations, to clarify the node and branch structure of the grid. Then, based on the electrical parameters of grid components, such as line impedance, transformer turns ratio, and generator output limits, a network model reflecting the dynamic characteristics of the grid is constructed, incorporating the grid's steady-state operating parameters and transient response characteristics into the model. On this basis, power flow calculation algorithms, such as the Newton-Raphson method or PQ decomposition method, are used to simulate the power flow distribution of the constructed dynamic network model. Taking into account factors such as power balance, voltage level, and reactive power distribution under different operating conditions, the stability margin calculation results for each node are determined, clarifying the stable operating range of the grid under current and predicted conditions. Simultaneously, combining historical grid operation data and safe operation standards, preset safety threshold boundaries are set for each node to enable real-time monitoring of the grid's operating status during dispatching, providing crucial information for subsequent dispatching decisions.

[0061] S103: Based on the dynamic load forecast map and stability margin calculation results, the global scheduling target is decomposed into stages to generate a progressive scheduling stage sequence.

[0062] Specifically, a thorough analysis of the dynamic load forecast map is conducted to obtain key information such as load change trends, fluctuation amplitudes, and peak and trough periods for various future time periods. Simultaneously, the stability margin calculation results for each node are meticulously reviewed to clarify the stability limits and safety boundaries of the power grid under different operating conditions. Based on a comprehensive understanding of load characteristics and stability margins, the global scheduling objectives are rationally decomposed according to time and spatial dimensions. In the time dimension, scheduling stages are divided based on the natural laws and fluctuation characteristics of load changes, ensuring relatively consistent load characteristics in each stage to facilitate the development of targeted scheduling strategies. In the spatial dimension, factors such as the geographical distribution of the power grid, load density, and power source access points are comprehensively considered to rationally allocate scheduling tasks to different regions. Through this refined decomposition, a sequence containing multiple progressive scheduling stages is generated. Each stage has clear load adjustment objectives, power balance requirements, and safety and stability constraints to gradually achieve stable operation and optimized scheduling of the power grid.

[0063] S104: Match the response characteristics of the power generation equipment and generate an adaptive progressive scheduling instruction sequence based on the progressive scheduling phase sequence.

[0064] Specifically, an in-depth analysis of the response characteristics of power generation equipment is conducted, including key indicators such as power regulation capabilities, ramp rate limits, and sensitivity to frequency and voltage changes. Then, the progressive scheduling phase sequence is broken down, clarifying the specific tasks for each phase, including load adjustment requirements, power balance targets, and node voltage and frequency stability requirements. Based on the response characteristics of the power generation equipment, suitable combinations of power generation equipment are matched to the scheduling tasks of each phase, determining the output adjustment magnitude and timing of each piece of equipment, while also considering the equipment's operating constraints, such as maximum output limits and minimum stable operating power.

[0065] The dispatch instructions are broken down into specific operational steps, including instructions for increasing or decreasing generator output, start-up and shutdown operations, and reactive power regulation. These are then arranged according to chronological order and logical relationships to generate a preliminary dispatch instruction sequence. The generated sequence is then optimized to ensure feasibility and grid stability. Simultaneously, the sequence is verified to meet grid safety and economic operation standards. This process generates an adaptive, progressive dispatch instruction sequence that guides generators to make appropriate adjustments at different stages to meet grid dispatch requirements.

[0066] S105: Executes the adaptive progressive scheduling instruction sequence, processes the feedback of the real-time grid status, and generates dynamic adjustment instructions.

[0067] Specifically, executing the adaptive progressive dispatch command sequence involves issuing dispatch commands to various generator units and energy storage devices, adjusting their power output according to the time and magnitude specified in the commands. Simultaneously, a real-time monitoring system is established to continuously track the grid's operating status, focusing on key indicators such as whether node voltages are stable within the rated range, whether frequencies remain at standard levels, and whether power flow is evenly distributed. Real-time data is collected by sensors and monitoring equipment installed at each node of the grid, clearly reflecting the actual operating status of the grid after the execution of dispatch commands. Subsequently, the data obtained from the real-time monitoring is compared and analyzed with the expected targets of the dispatch commands to obtain the deviations between the actual operating status and the expected targets, such as node voltage deviations, frequency offsets, and power imbalances. These deviations will serve as the basis for subsequent adjustments.

[0068] Based on the results of the deviation analysis, a scheduling optimization algorithm is used to recalculate and generate dynamic adjustment instructions. These instructions will correct and optimize the original scheduling plan to ensure the safe, stable, and efficient operation of the power grid. The dynamic adjustment instructions will then be issued to each execution unit, forming a closed-loop scheduling control process that continuously adjusts according to the actual operating conditions of the power grid to achieve the optimal operating state of the power grid.

[0069] One embodiment of this application also provides a method for dynamic forecasting and optimized scheduling of power grid load, including: acquiring meteorological parameters, historical load curves, and renewable energy output data; performing multi-source heterogeneous fusion processing on the meteorological parameters, historical load curves, and renewable energy output data to generate a dynamic load forecast map; performing analytical processing on the power grid topology to construct a dynamic network model; performing power flow distribution simulation processing based on the dynamic network model to obtain the stability margin calculation results and preset safety threshold boundaries for each node; performing stage decomposition processing on the global scheduling objective based on the dynamic load forecast map and stability margin calculation results to generate an incremental scheduling stage sequence; performing matching processing on the response characteristics of power generation equipment to generate an adaptive incremental scheduling instruction sequence based on the incremental scheduling stage sequence; executing the adaptive incremental scheduling instruction sequence and performing feedback processing on the real-time state of the power grid to generate dynamic adjustment instructions, so as to realize multi-dimensional data association modeling, stability margin quantitative analysis, and dynamic instruction optimization, thereby improving load forecasting accuracy and reducing fault propagation and transient oscillations.

[0070] Furthermore, based on the dynamic load forecast map and stability margin calculation results, the global scheduling objective is decomposed into stages to generate a progressive scheduling stage sequence, including:

[0071] Using the following formula, based on the load gradient distribution characteristics and stability margin calculation results of the dynamic load forecast map, bottleneck areas are identified in the entire network transmission channels, generating a multi-dimensional risk heat map:

[0072]

[0073]

[0074] Among them, L g This represents the load gradient index, where n represents the total number of nodes, and P... i V represents the active power at node i. i The voltage magnitude at node i is represented by ΔV. i S represents the voltage deviation. m λ represents the stability margin index. max λ represents the largest eigenvalue. min Let λ represent the smallest eigenvalue, m represent the total number of eigenvalues, and λ represent the smallest eigenvalue. k This represents the k-th eigenvalue;

[0075] Based on the spatiotemporal evolution trend of the multidimensional risk heatmap, the global scheduling objectives are prioritized for load transfer and segmented into time windows to generate a progressive scheduling phase sequence.

[0076] Specifically, the load gradient distribution characteristics in the dynamic load forecast map are analyzed to clarify the rate and direction of load change in different regions and time periods. Simultaneously, combined with stability margin calculation results, the stability limits and risk levels of each node in the power grid are determined. Based on this information, a specific algorithm is used to comprehensively evaluate the entire network transmission channels, identifying potential bottleneck areas. These bottleneck areas include congestion points during peak load periods and weak links with low stability margins. Subsequently, the identified bottleneck areas and their risk levels are integrated to generate a multi-dimensional risk heat map. This heat map not only reflects the current risk distribution of the power grid but also predicts its spatiotemporal evolution trend.

[0077] Based on the evolution trend of the heat map, the overall scheduling objectives are refined: on the one hand, the load transfer tasks are prioritized according to the urgency and scope of the risks, and the load congestion problem in high-risk areas is addressed first; on the other hand, the scheduling time periods are reasonably divided, and corresponding scheduling strategies are formulated for time windows with different risk levels and load characteristics, thereby generating a phased and prioritized progressive scheduling phase sequence, providing clear guidance and planning for subsequent scheduling operations.

[0078] Furthermore, based on the load gradient distribution characteristics and stability margin calculation results of the dynamic load forecast map, bottleneck areas are identified in the entire network transmission channels, generating a multi-dimensional risk heat map, including:

[0079] Based on the load gradient distribution characteristics of the dynamic load forecast map, the trend analysis of the load change direction and rate of the entire network is performed to generate gradient evolution paths.

[0080] Based on the stability margin calculation results and gradient evolution path, the node carrying capacity attenuation region of the transmission channel is located and an initial bottleneck node set is generated.

[0081] Using the following formula, based on the topological connectivity of the initial bottleneck node set, a correlation analysis is performed on the power coupling strength of adjacent transmission channels to generate risk propagation paths:

[0082]

[0083] Among them, P ij α represents the power coupling strength between node i and node j. ij N represents the coupling coefficient. i Let d represent the set of neighboring nodes of node i. ib R represents the physical distance between node i and node b, σ represents the standard deviation parameter of the Gaussian kernel function, and R i β represents the risk propagation intensity at node i, β represents the risk propagation coefficient, and ω represents the risk propagation intensity at node i. ij t represents the weight coefficient between node i and node j. ijτ represents the signal propagation time, h represents the time decay constant, and h represents the total number of nodes in the network.

[0084] By integrating the gradient evolution path, the initial bottleneck node set, and the risk diffusion path, a multi-dimensional risk heat map is generated, which includes labels for risk level, propagation direction, and urgency.

[0085] Specifically, based on dynamic load forecast maps, the direction and rate of load changes in each region are extracted to construct gradient evolution paths and clarify the future development trend of load within the spatiotemporal range. Then, combined with the stability margin calculation results, the carrying capacity of nodes within the transmission channels of the network is analyzed, identifying nodes that may experience capacity attenuation under high load conditions, and generating an initial bottleneck node set. Next, topological connectivity is introduced, and by analyzing parameters such as the physical distance and connection strength between the initial bottleneck nodes and their adjacent nodes, the coupling strength of power between adjacent transmission channels is calculated, thereby determining the risk propagation path. The gradient evolution path, the initial bottleneck node set, and the risk propagation path are comprehensively analyzed, generating a multi-dimensional risk heat map based on risk level, propagation direction, and urgency, providing a more comprehensive and intuitive risk assessment basis for power grid dispatching decisions.

[0086] like Figure 2 The process of matching the response characteristics of the power generation equipment and generating an adaptive progressive scheduling instruction sequence based on the progressive scheduling stage sequence includes:

[0087] S201: Perform dynamic capability mapping on the response characteristics of the power generation equipment to generate a ramp rate constraint domain;

[0088] S202: Based on the spatiotemporal domain partitioning results of the progressive scheduling phase sequence, the capacity matching degree is calculated for the load adjustment requirements of each sub-phase, and the upper limit of the phase instruction amplitude is generated.

[0089] S203: Based on the ramp rate constraint domain and the upper limit of the stage instruction amplitude, perform dynamic planning on the timing execution interval of the scheduling instructions to generate instruction step size optimization parameters;

[0090] S204: Integrate instruction step size optimization parameters with priority weights of the incremental scheduling phase sequence to generate an adaptive incremental scheduling instruction sequence.

[0091] Specifically, a comprehensive dynamic capability mapping is performed on the response characteristics of power generation equipment to obtain its power output ramp rate limit, generating a clear ramp rate constraint domain to ensure that the equipment does not exceed its regulation capacity during scheduling. Then, based on the spatiotemporal domain division results of the progressive scheduling phase sequence, the capacity matching degree of the power generation equipment is calculated for the load adjustment requirements of each sub-stage, thereby determining the upper limit of the command amplitude for each stage, ensuring that load demand is met while avoiding equipment overload. Taking into account both the ramp rate constraint domain and the stage command amplitude upper limit, dynamic programming techniques are used to rationally arrange the timing execution interval of scheduling commands, generating scientific command step size optimization parameters to ensure the orderliness and executability of scheduling commands. Finally, the command step size optimization parameters are fully integrated with the priority weights of the progressive scheduling phase sequence, comprehensively considering the urgency and importance of each stage, to generate an adaptive progressive scheduling command sequence, achieving more precise, efficient, and stable grid scheduling, ensuring the safe and reliable operation of the grid at different stages.

[0092] Furthermore, based on the ramp rate constraint domain and the upper limit of the stage instruction amplitude, dynamic programming is performed on the timing execution interval of the scheduling instructions to generate instruction step size optimization parameters, including:

[0093] Based on the margin decay rate of the ramp rate constraint domain and the load mutation threshold of the upper limit of the stage instruction amplitude, the timing interval of the scheduling instructions is subjected to backpropagation risk analysis to generate a safe execution time window.

[0094] Based on the boundary constraints of the safe execution time window, the instruction superposition effect between adjacent scheduling stages is dynamically divided into time windows to generate instruction execution interval thresholds.

[0095] By combining the safe execution time window and the instruction execution interval threshold, instruction step size optimization parameters are generated.

[0096] Specifically, by combining the margin decay rate of the ramp rate constraint domain with the load mutation threshold of the stage command amplitude upper limit, backpropagation risk analysis technology is used to evaluate the timing interval of dispatch commands, identify potential risk points, and then determine the safe execution time window to ensure the safe and stable operation of the power grid. Based on the boundary constraints of the safe execution time window, the problem of power grid load fluctuations caused by the superposition of commands between adjacent dispatch stages is analyzed. The time interval is reasonably divided through dynamic time window segmentation technology, and command execution interval thresholds are set to ensure that commands in each stage are executed in an orderly manner without interference. By organically integrating the safe execution time window and the command execution interval threshold, and comprehensively considering multiple factors such as the safety, stability, and economy of power grid operation, optimization algorithms are used to calculate and generate command step size optimization parameters, providing a key basis for generating a scientific and reasonable adaptive progressive dispatch command sequence, and realizing more precise and efficient dispatch control of the power grid.

[0097] Furthermore, an adaptive progressive scheduling instruction sequence is executed, and feedback processing is performed on the real-time status of the power grid to generate dynamic adjustment instructions, including:

[0098] Based on the spatiotemporal constraints of the adaptive progressive scheduling instruction sequence, step-by-step instruction issuance is performed on the generator set and energy storage device to generate the initial scheduling execution trajectory.

[0099] Real-time status monitoring and processing are performed on the node voltage, frequency, and power flow of the initial scheduling execution trajectory to generate a real-time power grid status dataset;

[0100] Dynamic trajectory deviation analysis is performed on the real-time power grid status dataset and the expected scheduling path to generate trajectory deviation parameters and safety risk level labels.

[0101] Based on trajectory deviation parameters and safety risk level labels, the step size and direction of the adaptive progressive scheduling instruction sequence are dynamically compensated to generate dynamically adjusted instructions.

[0102] Specifically, based on the spatiotemporal constraints of the instruction sequence, instructions are distributed step-by-step to generator sets and energy storage devices, thereby generating an initial scheduling execution trajectory to ensure that each device adjusts gradually according to the predetermined plan. Simultaneously, a real-time monitoring system is established to continuously track the grid's operating status, focusing on key indicators such as node voltage stability, frequency normality, and power flow balance. This real-time data is collected and integrated to generate a real-time grid status dataset. Then, the real-time status dataset is compared and analyzed with the expected scheduling path. Using dynamic trajectory deviation analysis technology, the deviation parameters between the actual operating trajectory and the expected trajectory are calculated. Based on the magnitude and trend of the deviation, combined with a safety risk assessment model, corresponding safety risk level labels are generated. Based on the trajectory deviation parameters and safety risk level labels, a scheduling optimization algorithm is used to dynamically compensate and adjust the step size and direction of the adaptive progressive scheduling instruction sequence, generating dynamic adjustment instructions to correct deviations, reduce risks, and ensure the accuracy and stability of grid scheduling.

[0103] Furthermore, dynamic trajectory deviation analysis is performed on the real-time power grid status dataset and the expected scheduling path to generate trajectory deviation parameters and safety risk level labels, including:

[0104] Spatiotemporal alignment processing is performed on the real-time power grid status dataset to generate standardized status monitoring sequences that match the expected scheduling path;

[0105] Based on the time-domain waveform characteristics of the standardized state monitoring sequence and the expected scheduling path, dynamic trajectory similarity comparison processing is performed to generate trajectory deviation parameters.

[0106] The trajectory deviation parameter is processed by time window cumulative effect analysis to generate a risk diffusion trend prediction index.

[0107] Based on the risk diffusion trend prediction indicators and the preset safety threshold boundary, the deviation area is classified into risk levels to generate trajectory deviation parameters and safety risk level labels.

[0108] Specifically, the real-time status dataset undergoes spatiotemporal alignment to ensure that the data's temporal sequence and spatial distribution match the expected scheduling path, generating standardized status monitoring sequences. Based on the standardized sequences and the temporal waveform characteristics of the expected path, a dynamic trajectory similarity comparison algorithm is used to calculate the trajectory deviation parameter, quantifying the difference between the actual running trajectory and the expected trajectory. Through time window cumulative effect analysis, the cumulative impact of the trajectory deviation parameter in the time dimension is evaluated, risk diffusion trends are predicted, and risk diffusion trend prediction indicators are generated. Combining the risk diffusion trend prediction indicators with preset safety threshold boundaries, areas with deviations are classified into risk levels, their risk levels are determined, and trajectory deviation parameters and corresponding safety risk level labels are generated, providing a basis for subsequent scheduling adjustments.

[0109] 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 of other steps.

[0110] In one embodiment, such as Figure 3 As shown, this application also provides a power grid load dynamic forecasting and optimized dispatching device 300, which includes:

[0111] The multi-source data fusion module 301 is used to acquire meteorological parameters, historical load curves and new energy output data, and to perform multi-source heterogeneous fusion processing on the meteorological parameters, historical load curves and new energy output data to generate a dynamic load forecast map.

[0112] The network modeling and power flow analysis module 302 is used to analyze the power grid topology, construct a dynamic network model, perform power flow distribution simulation based on the dynamic network model, and obtain the stability margin calculation results and preset safety threshold boundaries for each node.

[0113] The stage decomposition optimization module 303 is used to perform stage decomposition processing on the global scheduling target based on the dynamic load prediction map and the stability margin calculation results, and generate a progressive scheduling stage sequence.

[0114] The response characteristic matching module 304 is used to match the response characteristics of the power generation equipment and generate an adaptive progressive scheduling instruction sequence based on the progressive scheduling stage sequence.

[0115] The dynamic feedback execution module 305 is used to execute the adaptive progressive scheduling instruction sequence and perform feedback processing on the real-time status of the power grid to generate dynamic adjustment instructions.

[0116] Specifically, the multi-source data fusion module 301 acquires meteorological parameters, historical load curves, and renewable energy output data. These data sources are diverse and vary in format. This module integrates the heterogeneous data into a unified dynamic load forecast map through data cleaning, standardization, and feature extraction, providing a foundation for subsequent analysis. The network modeling and power flow analysis module 302 is responsible for analyzing the power grid topology, constructing a dynamic network model including nodes, branches, and electrical parameters, and simulating the power flow distribution of the power grid using power flow calculation algorithms to obtain the stability margin and safety threshold boundaries of each node. The stage decomposition and optimization module 303 combines the dynamic load forecast map and stability margin results, using methods such as time-series segmentation and priority ranking to decompose the global scheduling objective into multiple progressive scheduling stage sequences, making the scheduling task more operable. The response characteristic matching module 304 analyzes the response characteristics of generating equipment, such as power regulation capability and ramp rate, and generates an adaptive progressive scheduling command sequence that matches the equipment characteristics based on the requirements of the progressive scheduling stage sequence. The dynamic feedback execution module 305 executes the sequence of dispatching instructions, monitors the power grid status in real time, and generates dynamic adjustment instructions through the feedback mechanism to ensure that the power grid operation meets the dispatching objectives.

[0117] The stage decomposition and optimization module 303 is also used for:

[0118] Using the following formula, based on the load gradient distribution characteristics and stability margin calculation results of the dynamic load forecast map, bottleneck areas are identified in the entire network transmission channels, generating a multi-dimensional risk heat map:

[0119]

[0120] Among them, L g This represents the load gradient index, where n represents the total number of nodes, and P... i V represents the active power at node i. i The voltage magnitude at node i is represented by ΔV. i S represents the voltage deviation. m λ represents the stability margin index. maxλ represents the largest eigenvalue. min Let λ represent the smallest eigenvalue, m represent the total number of eigenvalues, and λ represent the smallest eigenvalue. k This represents the k-th eigenvalue;

[0121] Based on the spatiotemporal evolution trend of the multidimensional risk heatmap, the global scheduling objectives are prioritized for load transfer and segmented into time windows to generate a progressive scheduling phase sequence.

[0122] The stage decomposition and optimization module 303 is also used for:

[0123] Based on the load gradient distribution characteristics of the dynamic load forecast map, the trend analysis of the load change direction and rate of the entire network is performed to generate gradient evolution paths.

[0124] Based on the stability margin calculation results and gradient evolution path, the node carrying capacity attenuation region of the transmission channel is located and an initial bottleneck node set is generated.

[0125] Using the following formula, based on the topological connectivity of the initial bottleneck node set, a correlation analysis is performed on the power coupling strength of adjacent transmission channels to generate risk propagation paths:

[0126]

[0127] Among them, P ij α represents the power coupling strength between node i and node j. ij N represents the coupling coefficient. i Let d represent the set of neighboring nodes of node i. ib R represents the physical distance between node i and node b, σ represents the standard deviation parameter of the Gaussian kernel function, and R i β represents the risk propagation intensity at node i, β represents the risk propagation coefficient, and ω represents the risk propagation intensity at node i. ij t represents the weight coefficient between node i and node j. ij τ represents the signal propagation time, h represents the time decay constant, and h represents the total number of nodes in the network.

[0128] By integrating the gradient evolution path, the initial bottleneck node set, and the risk diffusion path, a multi-dimensional risk heat map is generated, which includes labels for risk level, propagation direction, and urgency.

[0129] The response characteristic matching module 304 is also used for:

[0130] Dynamic capability mapping is performed on the response characteristics of the power generation equipment to generate a ramp rate constraint domain;

[0131] Based on the spatiotemporal domain partitioning results of the progressive scheduling phase sequence, the capacity matching degree is calculated for the load adjustment requirements of each sub-phase to generate the upper limit of the phase instruction amplitude.

[0132] Based on the climbing rate constraint domain and the upper limit of the stage instruction amplitude, the timing execution interval of the scheduling instructions is dynamically planned to generate instruction step size optimization parameters.

[0133] By integrating instruction step size optimization parameters with priority weights of the incremental scheduling phase sequence, an adaptive incremental scheduling instruction sequence is generated.

[0134] The response characteristic matching module 304 is also used for:

[0135] Based on the margin decay rate of the ramp rate constraint domain and the load mutation threshold of the upper limit of the stage instruction amplitude, the timing interval of the scheduling instructions is subjected to backpropagation risk analysis to generate a safe execution time window.

[0136] Based on the boundary constraints of the safe execution time window, the instruction superposition effect between adjacent scheduling stages is dynamically divided into time windows to generate instruction execution interval thresholds.

[0137] By combining the safe execution time window and the instruction execution interval threshold, instruction step size optimization parameters are generated.

[0138] The dynamic feedback execution module 305 is also used for:

[0139] Based on the spatiotemporal constraints of the adaptive progressive scheduling instruction sequence, step-by-step instruction issuance is performed on the generator set and energy storage device to generate the initial scheduling execution trajectory.

[0140] Real-time status monitoring and processing are performed on the node voltage, frequency, and power flow of the initial scheduling execution trajectory to generate a real-time power grid status dataset;

[0141] Dynamic trajectory deviation analysis is performed on the real-time power grid status dataset and the expected scheduling path to generate trajectory deviation parameters and safety risk level labels.

[0142] Based on trajectory deviation parameters and safety risk level labels, the step size and direction of the adaptive progressive scheduling instruction sequence are dynamically compensated to generate dynamically adjusted instructions.

[0143] The dynamic feedback execution module 305 is also used for:

[0144] Spatiotemporal alignment processing is performed on the real-time power grid status dataset to generate standardized status monitoring sequences that match the expected scheduling path;

[0145] Based on the time-domain waveform characteristics of the standardized state monitoring sequence and the expected scheduling path, dynamic trajectory similarity comparison processing is performed to generate trajectory deviation parameters.

[0146] The trajectory deviation parameter is processed by time window cumulative effect analysis to generate a risk diffusion trend prediction index.

[0147] Based on the risk diffusion trend prediction indicators and the preset safety threshold boundary, the deviation area is classified into risk levels to generate trajectory deviation parameters and safety risk level labels.

[0148] In one embodiment, 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 implement the steps in the above-described method embodiments.

[0149] In one embodiment, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the above-described method embodiments.

[0150] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The components described as separate parts may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this disclosure according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0151] The above-described embodiments are merely illustrative of several implementation methods of the embodiments of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the embodiments of this application, and these modifications and improvements all fall within the protection scope of the embodiments of this application.

Claims

1. A method for dynamic forecasting and optimal scheduling of power grid load, characterized in that, The method includes: Meteorological parameters, historical load curves, and renewable energy output data are acquired. Multi-source heterogeneous fusion processing is performed on the meteorological parameters, historical load curves, and renewable energy output data to generate a dynamic load prediction map. The power grid topology is analyzed and processed to construct a dynamic network model. Based on the dynamic network model, power flow distribution simulation is performed to obtain the stability margin calculation results and preset safety threshold boundaries for each node. Based on the dynamic load prediction map and the stability margin calculation results, the global scheduling target is decomposed into stages to generate a progressive scheduling stage sequence. The response characteristics of the power generation equipment are matched, and an adaptive progressive scheduling instruction sequence is generated based on the progressive scheduling stage sequence. The adaptive progressive scheduling instruction sequence is executed, and feedback processing is performed on the real-time status of the power grid to generate dynamic adjustment instructions.

2. The method for dynamic forecasting and optimized scheduling of power grid load according to claim 1, characterized in that, The step of performing a phased decomposition of the global scheduling objective based on the dynamic load forecast map and the stability margin calculation results to generate a progressive scheduling phase sequence includes: Using the following formula, based on the load gradient distribution characteristics of the dynamic load prediction map and the stability margin calculation results, bottleneck areas are identified in the entire network transmission channels, generating a multi-dimensional risk heat map: Among them, L g This represents the load gradient index, where n represents the total number of nodes, and P... i V represents the active power at node i. i The voltage magnitude at node i is represented by ΔV. i S represents the voltage deviation. m λ represents the stability margin index. max λ represents the largest eigenvalue. min Let λ represent the smallest eigenvalue, m represent the total number of eigenvalues, and λ represent the smallest eigenvalue. k This represents the k-th eigenvalue; Based on the spatiotemporal evolution trend of the multidimensional risk heatmap, the global scheduling target is sorted by load transfer priority and segmented by time window to generate the progressive scheduling phase sequence.

3. The method for dynamic forecasting and optimized scheduling of power grid load according to claim 2, characterized in that, The step of identifying bottleneck areas in the entire network transmission channels based on the load gradient distribution characteristics of the dynamic load prediction map and the stability margin calculation results, and generating a multi-dimensional risk heat map, includes: Based on the load gradient distribution characteristics of the dynamic load prediction map, the trend analysis of the load change direction and rate of the entire network is performed to generate a gradient evolution path. Based on the stability margin calculation results and the gradient evolution path, the node carrying capacity attenuation region of the transmission channel is located and an initial bottleneck node set is generated. Using the following formula, based on the topological connectivity of the initial bottleneck node set, a correlation analysis is performed on the power coupling strength of adjacent transmission channels to generate risk propagation paths: Among them, P ij α represents the power coupling strength between node i and node j. ij N represents the coupling coefficient. i Let d represent the set of neighboring nodes of node i. ib R represents the physical distance between node i and node b, σ represents the standard deviation parameter of the Gaussian kernel function, and R i β represents the risk propagation intensity at node i, β represents the risk propagation coefficient, and ω represents the risk propagation intensity at node i. ij t represents the weight coefficient between node i and node j. ij τ represents the signal propagation time, h represents the time decay constant, and h represents the total number of nodes in the network. By integrating the gradient evolution path, the initial bottleneck node set, and the risk diffusion path, a multidimensional risk heat map is generated, which includes labels for risk level, propagation direction, and urgency.

4. The method for dynamic forecasting and optimized scheduling of power grid load according to claim 1, characterized in that, The process of matching the response characteristics of the power generation equipment and generating an adaptive progressive scheduling instruction sequence based on the progressive scheduling phase sequence includes: The response characteristics of the power generation equipment are subjected to dynamic capability mapping processing to generate a ramp rate constraint domain. Based on the spatiotemporal domain division results of the progressive scheduling phase sequence, the capacity matching degree is calculated for the load adjustment requirements of each sub-phase to generate the upper limit of the phase instruction amplitude. Based on the climbing rate constraint domain and the upper limit of the stage instruction amplitude, the timing execution interval of the scheduling instructions is dynamically planned to generate instruction step size optimization parameters. The adaptive progressive scheduling instruction sequence is generated by combining the instruction step size optimization parameters with the priority weights of the progressive scheduling stage sequence.

5. The method for dynamic prediction and optimal scheduling of power grid load according to claim 4, characterized in that, The step of dynamically planning the timing execution interval of the scheduling instructions based on the climbing rate constraint domain and the upper limit of the stage instruction amplitude to generate instruction step size optimization parameters includes: Based on the margin decay rate of the ramp rate constraint domain and the load mutation threshold of the upper limit of the stage command amplitude, the timing interval of the scheduling command is subjected to backpropagation risk analysis to generate a safe execution time window. Based on the boundary constraints of the safe execution time window, the instruction superposition effect between adjacent scheduling stages is dynamically divided into time windows to generate an instruction execution interval threshold. The instruction step size optimization parameters are generated by combining the safe execution time window and the instruction execution interval threshold.

6. The method for dynamic forecasting and optimized scheduling of power grid load according to claim 1, characterized in that, The execution of the adaptive progressive scheduling instruction sequence and the feedback processing of the real-time power grid status to generate dynamic adjustment instructions include: Based on the spatiotemporal constraints of the adaptive progressive scheduling instruction sequence, step-by-step instruction issuance is performed on the generator set and energy storage device to generate the initial scheduling execution trajectory. Real-time status monitoring and processing are performed on the node voltage, frequency, and power flow of the initial scheduling execution trajectory to generate a real-time power grid status dataset; Dynamic trajectory deviation analysis is performed on the real-time power grid status dataset and the expected scheduling path to generate trajectory deviation parameters and safety risk level labels. Based on the trajectory deviation parameters and safety risk level labels, the step size and direction of the adaptive progressive scheduling instruction sequence are dynamically compensated to generate the dynamic adjustment instructions.

7. The method for dynamic forecasting and optimized scheduling of power grid load according to claim 6, characterized in that, The step of performing dynamic trajectory deviation analysis on the real-time power grid status dataset and the expected scheduling path to generate trajectory deviation parameters and safety risk level labels includes: The real-time power grid status dataset is spatiotemporally aligned to generate a standardized status monitoring sequence that matches the expected scheduling path. Based on the time-domain waveform characteristics of the standardized state monitoring sequence and the expected scheduling path, dynamic trajectory similarity comparison processing is performed to generate trajectory deviation parameters. The trajectory deviation parameter is subjected to time window cumulative effect analysis to generate a risk diffusion trend prediction index; Based on the risk diffusion trend prediction index and the preset safety threshold boundary, the deviation area is classified into risk levels to generate the trajectory deviation parameters and safety risk level labels.

8. A device for dynamic forecasting and optimized dispatching of power grid load, characterized in that, The device includes: The multi-source data fusion module is used to acquire meteorological parameters, historical load curves and new energy output data, and to perform multi-source heterogeneous fusion processing on the meteorological parameters, historical load curves and new energy output data to generate a dynamic load forecast map. The network modeling and power flow analysis module is used to analyze the power grid topology, construct a dynamic network model, perform power flow distribution simulation based on the dynamic network model, and obtain the stability margin calculation results and preset safety threshold boundaries for each node. The phase decomposition optimization module is used to perform phase decomposition processing on the global scheduling target based on the dynamic load prediction map and the stability margin calculation results, and generate a progressive scheduling phase sequence. The response characteristic matching module is used to match the response characteristics of the power generation equipment and generate an adaptive progressive scheduling instruction sequence based on the progressive scheduling stage sequence. The dynamic feedback execution module is used to execute the adaptive progressive scheduling instruction sequence, and to process the feedback of the real-time status of the power grid to generate dynamic adjustment instructions.

9. 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 power grid load dynamic prediction and optimal scheduling method as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the power grid load dynamic prediction and optimal scheduling method as described in any one of claims 1 to 7.

Images