Wind power and photovoltaic field group resource scheduling optimization method and system

Abstract

The invention provides a wind power and photovoltaic field group resource scheduling optimization method and system, and relates to the technical field of energy scheduling, and the method comprises the steps: constructing a three-dimensional field group model according to a collected equipment set which is a set of all equipment in a wind power and photovoltaic field group; based on the three-dimensional field group model, calling a predetermined clustering strategy to perform clustering analysis on the equipment set to obtain a clustering result; a first node index of first equipment is analyzed, a first initial cluster center is determined, and the first equipment refers to any equipment in a first cluster in the clustering result; performing collaborative analysis on the first resource scheduling record of the first initial cluster center to obtain first prediction information; and carrying out resource scheduling optimization on the wind power and photovoltaic field group based on the first prediction information. According to the invention, the technical problem that the scheduling strategy cannot be adaptively adjusted according to the actual scene in the prior art can be solved, and the technical effect of improving the adaptability of new energy scheduling is achieved.

Description

Technical Field

[0001] This application relates to the field of energy dispatching technology, and in particular to methods and systems for optimizing resource dispatching of wind power and photovoltaic power plants. Background Technology

[0002] With the large-scale integration of renewable energy sources such as wind and solar power, intelligent dispatch and management of new energy power plants has become a crucial link in ensuring the stability and sustainable development of the power system. Currently, traditional wind and solar power plant dispatch methods largely rely on static equipment zoning, empirical rule formulation, and offline optimization strategies. These methods typically depend on control strategies preset by the power grid dispatch center, neglecting the dynamic coupling relationships and real-time operating status between equipment. This leads to untimely responses, low dispatch accuracy, and even energy waste or localized power shortages when facing complex operating scenarios such as resource volatility, temporal uncertainty, and drastic load changes. Especially given the rapid fluctuations and wide distribution of wind and solar power output, dispatch methods based solely on historical averages or rule models cannot accurately reflect the real-time status characteristics of equipment, thus affecting the overall system operating efficiency.

[0003] In summary, existing technologies suffer from a lack of real-time modeling capabilities for the dynamic relationship between wind and photovoltaic equipment, a rapid identification mechanism for changes in equipment operating status, and collaborative prediction methods for multi-source time-series data. This results in scheduling strategies failing to adapt to actual scenarios, inaccurate resource matching, and delayed responses, further impacting the scheduling stability, energy utilization efficiency, and supply-demand balance capabilities of new energy power plants in highly volatile and uncertain environments. Summary of the Invention

[0004] The purpose of this application is to provide a method and system for optimizing resource scheduling in wind power and photovoltaic power plant clusters. This is to address the technical problems in existing technologies, such as the lack of real-time modeling capabilities for the dynamic correlation between wind power and photovoltaic equipment, rapid identification mechanisms for changes in equipment operating status, and collaborative prediction methods for multi-source time-series data. These problems lead to scheduling strategies that cannot be adaptively adjusted according to actual scenarios, inaccurate resource matching, and delayed response, which further affect the scheduling stability, energy utilization efficiency, and supply-demand balance capabilities of new energy power plant clusters in highly volatile and uncertain environments.

[0005] In view of the above problems, this application provides a method and system for optimizing the scheduling of wind power and photovoltaic power plant clusters.

[0006] Firstly, this application provides a method for optimizing resource scheduling in wind and solar power clusters, implemented through a wind and solar power cluster resource scheduling optimization system. The method includes: constructing a three-dimensional cluster model based on a collected set of equipment, wherein the set of equipment is the collection of all equipment in the wind and solar power cluster; performing cluster analysis on the set of equipment based on the three-dimensional cluster model using a predetermined clustering strategy to obtain clustering results; analyzing the first node index of a first device and determining a first initial cluster center, wherein the first device refers to any device in the first cluster of the clustering results; performing collaborative analysis on the first resource scheduling record of the first initial cluster center to obtain first prediction information; and optimizing resource scheduling in the wind and solar power cluster based on the first prediction information.

[0007] Preferably, the wind power and photovoltaic cluster resource scheduling optimization method further includes: extracting any device from the equipment set and collecting arbitrary feature data of the arbitrary device; extracting any power source from the arbitrary feature data according to the predetermined clustering strategy, and performing first-level clustering on the equipment set to obtain a first-level clustering result; extracting any load from the arbitrary feature data according to the predetermined clustering strategy, and performing second-level clustering on the first-level clustering result to obtain the clustering result.

[0008] Preferably, the wind power and photovoltaic cluster resource scheduling optimization method further includes: taking the first device as the first node; obtaining the first neighborhood of the first node and counting the number of first nodes in the first neighborhood; taking the ratio of the number of the first nodes to the total number of first nodes in the first cluster as the first node index.

[0009] Preferably, the wind power and photovoltaic cluster resource scheduling optimization method further includes: sorting the first equipment in descending order using the first node index to obtain a first descending order equipment list, and taking the first and first equipment in the first descending order equipment list as the first initial cluster center.

[0010] Preferably, the wind power and photovoltaic cluster resource scheduling optimization method further includes: dynamically collecting arbitrary real-time feature data of any device; dynamically adjusting the clustering results based on the arbitrary real-time feature data to obtain real-time clustering results; and calling the cluster center update mechanism to analyze the real-time clustering results to obtain the first real-time cluster center of the first cluster.

[0011] Preferably, the wind power and photovoltaic cluster resource scheduling optimization method further includes: extracting a second cluster from the real-time clustering results, wherein the second cluster includes a second device; calculating the sum of distances from the second device to a third device, denoted as a second centrality index, wherein the third device refers to any device in the second cluster that is different from the second device; according to the cluster center update mechanism, sorting the second devices in ascending order using the second centrality index to obtain a first ascending order device list; and taking the first device in the first ascending order device list as the first real-time cluster center.

[0012] Preferably, the wind power and photovoltaic cluster resource scheduling optimization method further includes: performing predictive analysis on the time series of the first power generation in the first resource scheduling record to obtain the first predicted power generation; performing predictive analysis on the time series of the first power consumption in the first resource scheduling record to obtain the first predicted power consumption; performing predictive analysis on the time series of the first resource scheduling quantity in the first resource scheduling record to obtain the first predicted resource scheduling quantity; the first predicted power generation, the first predicted power consumption, and the first predicted resource scheduling quantity constitute the first prediction information.

[0013] Preferably, the wind power and photovoltaic cluster resource scheduling optimization method further includes: constructing a first power generation sample set based on the first power generation time series; performing spline fitting on the first power generation scatter plot drawn based on the first power generation sample set to obtain a first spline curve; obtaining a target time and using the target time as input data for the first polynomial corresponding to the first spline curve to obtain the first predicted power generation.

[0014] Preferably, the wind power and photovoltaic cluster resource scheduling optimization method further includes: comparing the first predicted power generation with the first predicted power consumption to obtain a first predicted power difference; determining whether the first predicted power difference reaches the first predicted resource scheduling amount; if it does, performing scheduling processing of the first cluster with the first predicted resource scheduling amount on the first cluster at the target time; wherein, it includes: obtaining any cluster and predicting any predicted resource scheduling amount of the arbitrary cluster at the target time; when the arbitrary predicted resource scheduling amount and the first predicted resource scheduling amount meet predetermined conditions, adding the arbitrary cluster to a candidate target scheduling cluster set; performing multi-dimensional scheduling feature analysis on the candidate target scheduling cluster set to determine the target scheduling cluster; and scheduling the first predicted resource scheduling amount of the first cluster to the target scheduling cluster at the target time.

[0015] Secondly, this application also provides a wind power and photovoltaic (PV) cluster resource scheduling optimization system for executing the wind power and PV cluster resource scheduling optimization method as described in the first aspect, comprising: a model building module for constructing a three-dimensional cluster model based on a collected set of equipment, wherein the set of equipment is a collection of all equipment in the wind power and PV cluster; a cluster analysis module for performing cluster analysis on the set of equipment based on the three-dimensional cluster model using a predetermined clustering strategy to obtain clustering results; a cluster center determination module for analyzing the first node index of a first device and determining a first initial cluster center, wherein the first device refers to any one device in the first cluster in the clustering results; a collaborative analysis module for performing collaborative analysis on the first resource scheduling record of the first initial cluster center to obtain first prediction information; and a resource scheduling module for optimizing resource scheduling of the wind power and PV cluster based on the first prediction information.

[0016] The technical solution provided in this application has at least the following technical effects or advantages: by achieving the technical goal of constructing a multi-dimensional clustering modeling system for wind power and photovoltaic power plant clusters, core equipment identification and dynamic resource scheduling optimization system driven by predictive information, it can improve the adaptability, real-time performance and overall operating efficiency of new energy system scheduling, and enhance the grid regulation capability and power supply stability in the scenario of large-scale access of renewable energy.

[0017] The above description is merely an overview of the technical solution of this application. To better understand the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the wind power and photovoltaic cluster resource scheduling optimization method of this application.

[0020] Figure 2 This is a schematic diagram of the wind power and photovoltaic power plant cluster resource scheduling optimization system of this application.

[0021] Figure labeling: Model building module 11, cluster analysis module 12, cluster center determination module 13, collaborative analysis module 14, resource scheduling module 15. Detailed Implementation

[0022] This application provides a method and system for optimizing resource scheduling in wind and solar power clusters. It addresses the technical problems in existing technologies where the lack of real-time modeling capabilities for the dynamic relationships between wind and solar equipment, rapid identification mechanisms for changes in equipment operating status, and collaborative prediction methods for multi-source time-series data leads to scheduling strategies that cannot adaptively adjust to actual scenarios, resulting in inaccurate resource matching, delayed response, and further impacting the scheduling stability, energy utilization efficiency, and supply-demand balance of renewable energy clusters in highly volatile and uncertain environments. The application achieves the technical goal of constructing a multi-dimensional clustering modeling system for wind and solar power clusters, core equipment identification, and a dynamic resource scheduling optimization system driven by predictive information. This improves the adaptability, real-time performance, and overall operational efficiency of renewable energy system scheduling, and enhances grid control capabilities and power supply stability in scenarios involving large-scale renewable energy integration.

[0023] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. It should be understood that this application is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. It should also be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all of them.

[0024] Example 1, please refer to the appendix. Figure 1 This application provides a method for optimizing resource scheduling in wind and solar power clusters, which is applied to a wind and solar power cluster resource scheduling optimization system. The method specifically includes the following steps: A three-dimensional cluster model is constructed based on the collected equipment set, which is the collection of all equipment in the wind power and photovoltaic cluster.

[0025] Specifically, all physical equipment in wind farms and photovoltaic farms, such as wind turbines, photovoltaic panels, converters, and energy storage devices, are incorporated into a single dataset. Based on the spatial location information, operating status parameters, and functional categories of the equipment, a three-dimensional model containing spatial, temporal, and energy dimensions is constructed to create a three-dimensional farm cluster model, which represents the structure and operational characteristics of the entire new energy farm cluster.

[0026] A three-dimensional power plant cluster model refers to a multi-dimensional data structure. The first dimension is the geospatial dimension, reflecting the actual geographical distribution of the equipment; the second dimension is the temporal dimension, reflecting the dynamic process of equipment operating parameters changing over time; and the third dimension is the energy dimension, representing the functional role and numerical characteristics of each device in energy generation, consumption, or transmission. By integrating these three dimensions, the correlation between equipment, local congestion within the power plant cluster, or the overall power generation and load distribution can be analyzed more intuitively.

[0027] Based on the three-dimensional field group model, a predetermined clustering strategy is invoked to perform clustering analysis on the device set to obtain clustering results.

[0028] Specifically, based on a three-dimensional field cluster model, and using a pre-constructed new energy field cluster model encompassing spatial, temporal, and energy dimensions as the foundation for analysis, the three-dimensional field cluster model records the characteristics of wind and photovoltaic equipment in terms of geographical distribution, operational sequence, and energy flow. For example, the geographical coordinates, power generation time curve, and power transmission path of a wind turbine are all used as input information. Retrieving a predetermined clustering strategy refers to calling a pre-set grouping algorithm, which may include density-based clustering methods, K-means algorithms, or graph structure partitioning rules, and then dividing the equipment into several categories according to specific rules. Cluster analysis is performed on the equipment set. During algorithm execution, the characteristic data of the equipment is used as input. By comparing the similarity between different devices in dimensions such as spatial location, power attributes, or load response, devices with similar functional characteristics are grouped together. For example, photovoltaic modules that are close to each other and have similar power generation time patterns are clustered into the same category. The final clustering result is that multiple device clusters with related characteristics are formed. Each cluster represents a subsystem with collaborative operation capabilities. For example, a cluster containing 10 photovoltaic arrays and 2 sets of energy storage devices can independently adjust its energy supply and demand balance, while another cluster mainly composed of wind turbines focuses on rapid response to changes in wind speed.

[0029] Analyze the first node index of the first device and determine the first initial cluster center, wherein the first device refers to any device in the first cluster in the clustering results.

[0030] Specifically, the connectivity of the first device within its first cluster is calculated to obtain the first node index, and the first initial cluster center is determined. It is then determined whether to use the first initial cluster center as the representative device of the first cluster. The first device is any device selected from the first cluster; it could be a wind turbine, a photovoltaic module, or an energy storage unit, representing a representative or evaluated individual within the cluster. The first node index measures the connectivity of the first device within the first cluster, and is expressed as the ratio between the number of neighboring nodes of the first device and the total number of devices in the first cluster. A larger first node index indicates more connections between the first device and other devices, stronger centrality, and suitability as the first initial cluster center. By comparing the node indices of multiple devices, the device with the largest first node index can be selected as the first initial cluster center for subsequent prediction and scheduling modeling.

[0031] First prediction information is obtained by collaboratively analyzing the first resource scheduling record of the first initial cluster center.

[0032] Specifically, after determining the first initial cluster center as the representative, the overall behavioral trend is jointly modeled and analyzed based on the past resource scheduling records of the equipment corresponding to the first initial cluster center, that is, by combining the first resource scheduling records with the operational relationships between the first resource scheduling records and other equipment within the first cluster. The first resource scheduling records refer to the historical operational data of the equipment corresponding to the first initial cluster center in energy allocation, power regulation, etc., over the past period, including various time-series information such as power generation, power load, and scheduling response. Collaborative analysis means that it does not only analyze a single variable, but comprehensively correlates multiple scheduling behaviors in the time, equipment, and spatial dimensions to identify the dependencies and evolution patterns between equipment. The first predictive information refers to a set of predictive data output after multivariate collaborative analysis, used to characterize the power generation, power consumption, and resource scheduling trends within the first cluster over a future period.

[0033] Based on the first prediction information, resource scheduling optimization is performed on the wind power and photovoltaic power plant clusters.

[0034] Specifically, the system utilizes a first forecast information comprised of power generation forecasts, power consumption forecasts, and resource allocation forecasts as input to coordinate, manage, and optimize the allocation of resources across multiple wind and solar power systems. This first forecast information comprehensively reflects the system's supply and demand status at a future timeframe, identifying potential power shortages and necessitating cross-cluster energy allocation. Wind and solar power clusters refer to energy systems composed of multiple geographically dispersed but network-connected renewable energy generation units and energy storage devices. Resource allocation optimization, under multi-source coordination and local constraints, dynamically determines the optimal power allocation, load transfer, or energy storage release scheme to maximize power generation utilization, minimize energy consumption, or optimize system operating costs.

[0035] Furthermore, this application also includes: extracting any device from the device set and collecting arbitrary feature data of the arbitrary device; extracting any power source from the arbitrary feature data according to the predetermined clustering strategy, and performing first-level clustering on the device set to obtain a first-level clustering result; extracting any load from the arbitrary feature data according to the predetermined clustering strategy, and performing second-level clustering on the first-level clustering result to obtain the clustering result.

[0036] Specifically, any device in the device set is extracted, and an individual device, such as a wind turbine or a group of photovoltaic arrays, is randomly selected. Any feature data of any device is collected, and information such as device type, rated power, actual power generation, current load, voltage fluctuation range, and access node of any device is extracted and organized as arbitrary feature data.

[0037] Following a pre-defined clustering method, any power source is extracted from arbitrary feature data. Features belonging to the power source category, such as power generation capacity, energy output curve, and power generation type, are selected. Based on these power source features, all devices are first-level clustered, resulting in a primary clustering result. The primary clustering divides the entire device set into several subsets, where devices in each subset exhibit high similarity in power source attributes.

[0038] Based on a predetermined clustering strategy, arbitrary loads are extracted from arbitrary feature data, focusing on load-related features such as equipment energy consumption trends, dispatch response capabilities, and grid load coupling relationships. Secondary clustering is then performed on the primary clustering results, further subdividing the already clustered equipment subsets to form a more accurate and functionally consistent final clustering structure. Table 1 shows the feature extraction and hierarchical clustering data record for wind power and photovoltaic equipment sets.

[0039] Table 1: Data Record Table of Feature Extraction and Hierarchical Clustering for Wind Power and Photovoltaic Equipment Sets

[0040] Furthermore, this application also includes: taking the first device as the first node; obtaining the first neighborhood of the first node and counting the number of first nodes in the first neighborhood; taking the ratio of the number of the first nodes to the total number of first nodes in the first cluster as the first node index.

[0041] Specifically, one device is randomly selected from the clustering results as the reference center in the graph model. This could be a wind turbine, a photovoltaic array, or other energy dispatching device. In the graph structure, each device can be abstracted as a node, representing its connectivity in physical space or energy network. The first device is designated as the first node. The first neighborhood of the first node is obtained, identifying the set of devices directly associated with or closely connected to it. Neighborhood nodes may be devices directly connected by power lines, share communication links, or have similar dispatch response characteristics. The number of nodes in the first neighborhood is then counted as a metric. The number of neighborhood nodes of the first device is divided by the total number of nodes in the first cluster to obtain the first node index, which measures the connectivity or centrality of the first device within the first cluster. A larger ratio indicates that the first node is closely coupled with more devices and has a stronger influence on overall dispatching.

[0042] Furthermore, this application also includes: sorting the first device in descending order using the first node index to obtain a first descending device list, and using the first device in the first descending device list as the first initial cluster center.

[0043] Specifically, the first node index is used to rank all first devices within the first cluster to determine which device has the strongest connectivity or representativeness within the cluster. The first node index is a ratio that measures the connection strength between a device and other devices in its neighborhood, expressed as the ratio of the number of neighborhood nodes to the total number of nodes in the cluster. The first devices are then sorted in descending order of their first node index, forming a first descending order device list. This creates a sequence of devices arranged by influence or centrality, facilitating the rapid identification of devices with stronger structural scheduling value. The first device in the first descending order device list is designated as the initial cluster center, and the device with the highest node index is selected as the most representative core device in the current cluster, serving as the primary reference point in subsequent scheduling analysis and predictive modeling.

[0044] Furthermore, this application also includes: dynamically collecting arbitrary real-time feature data from any device; dynamically adjusting the clustering results based on the arbitrary real-time feature data to obtain real-time clustering results; and calling the cluster center update mechanism to analyze the real-time clustering results to obtain the first real-time cluster center of the first cluster.

[0045] Specifically, it dynamically collects any real-time characteristic data from any device, continuously acquiring current operating status information of various devices in wind and solar power clusters, such as real-time power generation, voltage, current, temperature, wind speed, irradiance, and other environmental parameters, reflecting the dynamic behavior of the equipment at different time periods. "Any device" refers to any wind turbine, solar module, inverter, or energy storage unit, while "any real-time characteristic data" refers to any one or more indicators related to the current operating status of the device.

[0046] The clustering results are dynamically adjusted based on arbitrary real-time feature data to obtain real-time clustering results. The similarity relationships between devices in the static clustering are re-evaluated, thereby updating the device classification results. For example, two devices originally assigned to different clusters may be grouped into the same new cluster after dynamic adjustment due to highly consistent operational characteristics. The real-time clustering result is an updated version of the device classification based on the latest state, and it is updated as the device state continuously changes.

[0047] The cluster center update mechanism is invoked to analyze the real-time clustering results and obtain the first real-time cluster center of the first cluster. This involves recalculating the representativeness or connectivity of each device in the current real-time clustering results to select a new representative device as the cluster center. The cluster center update mechanism can determine the most suitable device to be the new core node based on indicators such as distance between devices, state differences, and energy flow direction.

[0048] Furthermore, this application also includes: extracting a second cluster from the real-time clustering results, wherein the second cluster includes a second device; calculating the sum of distances from the second device to a third device, denoted as a second centrality index, wherein the third device refers to any device in the second cluster that is different from the second device; according to the cluster center update mechanism, sorting the second devices in ascending order using the second centrality index to obtain a first ascending order device list; and taking the first device in the first ascending order device list as the first real-time cluster center.

[0049] Specifically, a second cluster is extracted from the real-time clustering results, and this second cluster includes a second device. A cluster is selected from the latest clustering structure obtained through real-time data updates as the analysis object, designated as the second cluster. Within this second cluster, multiple devices with associated characteristics are included. The second device is any one of the devices to be analyzed within the second cluster, used to participate in the evaluation process of cluster center updates.

[0050] The sum of the distances from the second device to the third device is denoted as the second centrality index. The third device refers to any device in the second cluster that is different from the second device. The sum of the distances between the second device and all other devices in the second cluster is then calculated. Distance can be measured by physical spatial distance, differences in operating states, or differences in power fluctuation curves. The second centrality index reflects the overall proximity of the second device to the other devices in the second cluster. The smaller the second centrality index, the more centrally located the second device is in the second cluster, and the more suitable it is to be a new cluster center.

[0051] According to the cluster center update mechanism, the second devices are sorted in ascending order by the second center index to obtain the first ascending order device list. That is, each device is sorted according to its center index from smallest to largest, forming a device list sorted by comprehensive proximity. Devices with smaller center indices in the first ascending order list are ranked higher. As a result, devices that are more centrally located and more representative in terms of device distribution will appear at the front of the list, thus enabling the rapid identification of the most likely target to become the new cluster center from multiple candidate devices.

[0052] The first device in the first ascending list of devices is designated as the first real-time cluster center, i.e., the device with the smallest center index, and is determined as the new center device for the current cluster. The first real-time cluster center has the closest relationship with other devices in the second cluster, thus it can more accurately represent the behavioral characteristics of the second cluster and guide subsequent scheduling strategies.

[0053] Furthermore, this application also includes: performing predictive analysis on the time series of the first power generation in the first resource scheduling record to obtain the first predicted power generation; performing predictive analysis on the time series of the first power consumption in the first resource scheduling record to obtain the first predicted power consumption; performing predictive analysis on the time series of the first resource scheduling quantity in the first resource scheduling record to obtain the first predicted resource scheduling quantity; the first predicted power generation, the first predicted power consumption, and the first predicted resource scheduling quantity constitute the first prediction information.

[0054] Specifically, predictive analysis is performed on the first power generation time series in the first resource scheduling record to obtain the first predicted power generation. This involves extracting the power generation sequence over time from historical equipment operating data and using methods such as statistical modeling, spline fitting, and neural networks to predict the power generation capacity at a future point in time or over a period of time. For example, modeling the hourly power generation of photovoltaic modules over the past seven days can predict the solar power generation trend over the next 24 hours, thus obtaining the first predicted power generation. Predictive analysis is also performed on the first electricity consumption time series in the first resource scheduling record to obtain the first predicted electricity consumption. This involves analyzing the patterns of electricity consumption changes over time based on historical load or energy consumption data and estimating future electricity demand levels.

[0055] Predictive analysis is performed on the time series of the first resource scheduling quantity in the first resource scheduling record to obtain the first predicted resource scheduling quantity. This involves extracting energy allocation behavior records from historical equipment operating data, establishing the correspondence between scheduling response modes and input conditions, and predicting the required or possible scheduling power or electricity at future points in time. The first predicted resource scheduling quantity reflects the transfer actions between equipment or regions to maintain energy balance, such as releasing electricity from energy storage units to replenish loads or allocating redundant electricity to another region.

[0056] The first predicted power generation, the first predicted power consumption, and the first predicted resource dispatch quantity constitute the first predicted information, forming a set of predictive input data to guide the next step of dispatch optimization. Among them, the first predicted power generation represents the supply-side capacity, the first predicted power consumption represents the demand-side scale, and the first predicted dispatch quantity connects the two to achieve balance control.

[0057] Furthermore, this application also includes: constructing a first power generation sample set based on the first power generation time series; performing spline fitting on the first power generation scatter plot drawn based on the first power generation sample set to obtain a first spline curve; obtaining a target time and using the target time as input data for the first polynomial corresponding to the first spline curve to obtain the first predicted power generation.

[0058] Specifically, the power generation data of the device corresponding to the first initial cluster center in a continuous time period is extracted from the historical records, namely the first power generation time series, and combined into a set of discrete sample points in chronological order. Each sample point includes a specific time and the corresponding power generation value, thereby obtaining the first power generation sample point set.

[0059] The first power generation scatter plot, drawn based on the first power generation sample set (i.e., a scatter plot formed on a plane with time as the horizontal axis and power generation as the vertical axis), is fitted with spline functions to obtain the first spline curve. Spline fitting is a smooth interpolation method that makes the curve continuous at each sample point and its derivative continuous, thereby approximating the original power generation trend and avoiding excessive fluctuations.

[0060] The target time is obtained and used as the input data of the first polynomial corresponding to the first parallel curve to obtain the first predicted power generation. That is, after a certain future time point is given, the future time is used as a variable to input into the first polynomial corresponding to the first parallel curve obtained by fitting, so as to calculate the predicted power generation corresponding to the future time.

[0061] Furthermore, this application also includes: comparing the first predicted power generation with the first predicted power consumption to obtain a first predicted power difference; determining whether the first predicted power difference reaches the first predicted resource scheduling amount; if it does, performing scheduling processing of the first cluster with the first predicted resource scheduling amount on the first cluster at the target time; wherein, it includes: obtaining any cluster and predicting any predicted resource scheduling amount of the arbitrary cluster at the target time; when the arbitrary predicted resource scheduling amount and the first predicted resource scheduling amount meet predetermined conditions, adding the arbitrary cluster to a candidate target scheduling cluster set; performing multi-dimensional scheduling feature analysis on the candidate target scheduling cluster set to determine a target scheduling cluster; and scheduling the first predicted resource scheduling amount of the first cluster to the target scheduling cluster at the target time.

[0062] Specifically, the predicted supply-side electricity (i.e., the first predicted power generation) and the predicted demand-side electricity (i.e., the first predicted electricity consumption) are calculated as the first predicted electricity difference over the same time dimension. The electricity gap or surplus in this first predicted electricity difference is compared with the adjustable electricity estimated based on historical trends and dispatch mechanisms (i.e., the first predicted resource dispatch amount) to determine whether the current supply-demand difference is within the adjustable capacity range. If the first predicted electricity difference reaches the first predicted resource dispatch amount, it is determined that the first predicted electricity difference is within the dispatch capacity range. At the target time, the first cluster is then processed using the first predicted resource dispatch amount to achieve local supply-demand balance.

[0063] This process involves selecting any device cluster from all clusters and predicting the arbitrary resource scheduling amount for that cluster at the target time. The resource scheduling capability of each cluster is independently predicted to obtain the amount of resources that need to be brought in, which is then used as the arbitrary predicted resource scheduling amount. If the arbitrary predicted resource scheduling amount (i.e., the amount of resources that need to be brought in) can numerically compensate for the first predicted resource scheduling amount (i.e., the amount of resources that can be scrambled out), then the arbitrary cluster will be marked as a potential participant and added to the candidate target scheduling cluster set. Multi-dimensional scheduling characteristic analysis is performed on the candidate target scheduling cluster set, including evaluating multiple characteristic dimensions such as grid stability, transmission loss rate, and scheduling efficiency during scheduling. This comprehensive evaluation assesses the merits of the candidate clusters as scheduling objects, ultimately determining one or more optimal target clusters as target scheduling clusters. At the target time, the resources that can be scrambled out are transferred from the first cluster to the determined target cluster, and scheduling commands are executed to achieve cross-cluster energy coordination and meet instantaneous energy demands.

[0064] In summary, the wind power and photovoltaic power plant cluster resource scheduling optimization method provided in this application has the following technical effects: by achieving the technical goals of constructing a multi-dimensional clustering model for wind power and photovoltaic power plant clusters, identifying core equipment, and establishing a dynamic resource scheduling optimization system driven by predictive information, it can improve the adaptability, real-time performance, and overall operating efficiency of the new energy system scheduling, and enhance the grid regulation capability and power supply stability in scenarios of large-scale renewable energy access.

[0065] Example 2: Based on the same inventive concept as the wind power and photovoltaic cluster resource scheduling optimization method in the foregoing examples, this application also provides a wind power and photovoltaic cluster resource scheduling optimization system. Please refer to the appendix. Figure 2 The system includes: a model building module 11, used to build a three-dimensional cluster model based on a collected set of equipment, wherein the set of equipment is a collection of all equipment in the wind power and photovoltaic cluster; a cluster analysis module 12, used to perform cluster analysis on the set of equipment based on the three-dimensional cluster model and a predetermined clustering strategy to obtain clustering results; a cluster center determination module 13, used to analyze the first node index of the first equipment and determine the first initial cluster center, wherein the first equipment refers to any one of the equipment in the first cluster in the clustering results; a collaborative analysis module 14, used to perform collaborative analysis on the first resource scheduling record of the first initial cluster center to obtain first prediction information; and a resource scheduling module 15, used to optimize resource scheduling for the wind power and photovoltaic cluster based on the first prediction information.

[0066] Furthermore, the wind power and photovoltaic cluster resource scheduling optimization system is also used to: extract any device from the equipment set and collect any feature data of the arbitrary device; extract any power source from the arbitrary feature data according to the predetermined clustering strategy, and perform first-level clustering on the equipment set to obtain the first-level clustering result; extract any load from the arbitrary feature data according to the predetermined clustering strategy, and perform second-level clustering on the first-level clustering result to obtain the clustering result.

[0067] Furthermore, the wind power and photovoltaic cluster resource scheduling optimization system is also used to: take the first device as the first node; obtain the first neighborhood of the first node and count the number of first nodes in the first neighborhood; take the ratio of the number of the first nodes to the total number of first nodes in the first cluster as the first node index.

[0068] Furthermore, the wind power and photovoltaic cluster resource scheduling optimization system is also used to: sort the first equipment in descending order using the first node index to obtain a first descending order equipment list, and take the first and first equipment in the first descending order equipment list as the first initial cluster center.

[0069] Furthermore, the wind power and photovoltaic cluster resource scheduling optimization system is also used for: dynamically collecting arbitrary real-time feature data from any device; dynamically adjusting the clustering results based on the arbitrary real-time feature data to obtain real-time clustering results; and calling the cluster center update mechanism to analyze the real-time clustering results to obtain the first real-time cluster center of the first cluster.

[0070] Furthermore, the wind power and photovoltaic power plant cluster resource scheduling optimization system is also used to: extract a second cluster from the real-time clustering results, wherein the second cluster includes a second device; calculate the sum of distances from the second device to a third device, denoted as a second centrality index, wherein the third device refers to any device in the second cluster that is different from the second device; according to the cluster center update mechanism, sort the second devices in ascending order using the second centrality index to obtain a first ascending order device list; and take the first device in the first ascending order device list as the first real-time cluster center.

[0071] Furthermore, the wind power and photovoltaic cluster resource scheduling optimization system is also used for: predicting and analyzing the time series of the first power generation in the first resource scheduling record to obtain the first predicted power generation; predicting and analyzing the time series of the first power consumption in the first resource scheduling record to obtain the first predicted power consumption; and predicting and analyzing the time series of the first resource scheduling quantity in the first resource scheduling record to obtain the first predicted resource scheduling quantity; the first predicted power generation, the first predicted power consumption, and the first predicted resource scheduling quantity constitute the first prediction information.

[0072] Furthermore, the wind power and photovoltaic power plant cluster resource scheduling optimization system is also used for: constructing a first power generation sample set based on the first power generation time series; performing spline fitting on the first power generation scatter plot drawn based on the first power generation sample set to obtain a first spline curve; obtaining a target time and using the target time as input data for the first polynomial corresponding to the first spline curve to obtain the first predicted power generation.

[0073] Furthermore, the wind power and photovoltaic cluster resource scheduling optimization system is also used for: comparing the first predicted power generation with the first predicted power consumption to obtain a first predicted power difference; determining whether the first predicted power difference reaches the first predicted resource scheduling amount; if it does, performing scheduling processing of the first cluster with the first predicted resource scheduling amount at the target time; wherein, it includes: obtaining any cluster and predicting any predicted resource scheduling amount of the arbitrary cluster at the target time; when the arbitrary predicted resource scheduling amount and the first predicted resource scheduling amount meet predetermined conditions, adding the arbitrary cluster to the candidate target scheduling cluster set; performing multi-dimensional scheduling feature analysis on the candidate target scheduling cluster set to determine the target scheduling cluster; and scheduling the first predicted resource scheduling amount of the first cluster to the target scheduling cluster at the target time.

[0074] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The wind power and photovoltaic cluster resource scheduling optimization method and specific examples in the aforementioned embodiment one are also applicable to the wind power and photovoltaic cluster resource scheduling optimization system of this embodiment. Through the foregoing detailed description of the wind power and photovoltaic cluster resource scheduling optimization method, those skilled in the art can clearly understand the wind power and photovoltaic cluster resource scheduling optimization system of this embodiment. Therefore, for the sake of brevity, it will not be described in detail here.

[0075] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0076] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for optimizing resource scheduling in wind and solar power clusters, characterized in that, include: A three-dimensional cluster model is constructed based on the collected equipment set, wherein the equipment set is the collection of all equipment in the wind power and photovoltaic cluster; Based on the three-dimensional field group model, a predetermined clustering strategy is invoked to perform clustering analysis on the device set to obtain clustering results; Analyze the first node index of the first device and determine the first initial cluster center, wherein the first device refers to any device in the first cluster in the clustering results; First prediction information is obtained by collaboratively analyzing the first resource scheduling record of the first initial cluster center. Based on the first prediction information, resource scheduling optimization is performed on the wind power and photovoltaic power plant clusters.

2. The wind power and photovoltaic power plant cluster resource scheduling optimization method according to claim 1, characterized in that, Based on the three-dimensional field group model, a predetermined clustering strategy is invoked to perform clustering analysis on the device set to obtain clustering results, including: Extract any device from the device set and collect any feature data of the arbitrary device; According to the predetermined clustering strategy, any power source is extracted from the arbitrary feature data, and the device set is subjected to first-level clustering to obtain the first-level clustering result; Based on the predetermined clustering strategy, arbitrary loadings are extracted from the arbitrary feature data, and secondary clustering is performed on the first-level clustering results to obtain the clustering results.

3. The wind power and photovoltaic power plant cluster resource scheduling optimization method according to claim 1, characterized in that, Before analyzing the first node index of the first device and determining the first initial cluster center, the process includes: The first device is designated as the first node; Obtain the first neighborhood of the first node and count the number of first nodes in the first neighborhood; The ratio of the number of the first nodes to the total number of the first nodes in the first cluster is taken as the first node index.

4. The wind power and photovoltaic power plant cluster resource scheduling optimization method according to claim 3, characterized in that, Analyzing the first node index of the first device and determining the first initial cluster center includes: sorting the first device in descending order using the first node index to obtain a first descending list of devices, and taking the first device in the first descending list as the first initial cluster center.

5. The wind power and photovoltaic power plant cluster resource scheduling optimization method according to claim 1, characterized in that, After analyzing the first node index of the first device and determining the first initial cluster center, the process also includes: Dynamically collect arbitrary real-time feature data from any device; The clustering results are dynamically adjusted based on the arbitrary real-time feature data to obtain real-time clustering results; The cluster center update mechanism is invoked to analyze the real-time clustering results and obtain the first real-time cluster center of the first cluster.

6. The wind power and photovoltaic power plant cluster resource scheduling optimization method according to claim 5, characterized in that, The cluster center update mechanism is invoked to analyze the real-time clustering results, and the first real-time cluster center of the first cluster is obtained, including: Extract the second cluster from the real-time clustering results, and the second cluster includes the second device; Calculate the sum of distances from the second device to the third device, denoted as the second centrality index, where the third device refers to any device in the second cluster that is different from the second device; According to the cluster center update mechanism, the second devices are sorted in ascending order by the second center index to obtain a first ascending list of devices. The first device in the first ascending device list is designated as the first real-time cluster center.

7. The wind power and photovoltaic power plant cluster resource scheduling optimization method according to claim 1, characterized in that, The first prediction information is obtained by collaboratively analyzing the first resource scheduling record of the first initial cluster center, including: The first predicted power generation is obtained by performing a predictive analysis on the time sequence of the first power generation in the first resource scheduling record. The first predicted electricity consumption is obtained by performing a predictive analysis on the first electricity consumption time sequence in the first resource scheduling record. Perform predictive analysis on the time series of the first resource scheduling quantity in the first resource scheduling record to obtain the first predicted resource scheduling quantity; The first predicted power generation, the first predicted power consumption, and the first predicted resource scheduling amount constitute the first predicted information.

8. The wind power and photovoltaic power plant cluster resource scheduling optimization method according to claim 7, characterized in that, A predictive analysis is performed on the first power generation time series in the first resource scheduling record to obtain the first predicted power generation, including: A first power generation sample set is constructed based on the first power generation time series; Spline fitting is performed on the first power generation scatter plot drawn based on the first power generation sample set to obtain the first spline curve; The target time is obtained and used as the input data of the first polynomial corresponding to the first spline curve to obtain the first predicted power generation.

9. The wind power and photovoltaic power plant cluster resource scheduling optimization method according to claim 8, characterized in that, Based on the first prediction information, resource scheduling optimization is performed on the wind power and photovoltaic power plant clusters, including: The first predicted power generation is compared with the first predicted power consumption to obtain the first predicted power difference; Determine whether the first predicted power difference reaches the first predicted resource scheduling amount; If the target is achieved, the first cluster will be processed by scheduling the first predicted resource allocation at the target time. This includes: Obtain any cluster and predict the arbitrary resource scheduling amount of the arbitrary cluster at the target time; When the arbitrary predicted resource scheduling amount and the first predicted resource scheduling amount meet the predetermined conditions, the arbitrary cluster is added to the candidate target scheduling cluster set. Perform multi-dimensional scheduling feature analysis on the candidate target scheduling clusters to determine the target scheduling clusters; At the target time, the first predicted resource scheduling amount of the first cluster is scheduled to the target scheduling cluster.

10. A wind power and photovoltaic power plant cluster resource scheduling optimization system, characterized in that, The steps for implementing the wind power and photovoltaic cluster resource scheduling optimization method according to any one of claims 1 to 9 include: The model building module is used to build a three-dimensional field cluster model based on the collected equipment set, wherein the equipment set is the collection of all equipment in the wind power and photovoltaic field cluster; The clustering analysis module is used to perform clustering analysis on the device set based on the three-dimensional field group model and a predetermined clustering strategy to obtain clustering results; The cluster center determination module is used to analyze the first node index of the first device and determine the first initial cluster center, wherein the first device refers to any device in the first cluster in the clustering results; The collaborative analysis module is used to perform collaborative analysis on the first resource scheduling record of the first initial cluster center to obtain the first prediction information; The resource scheduling module is used to optimize resource scheduling for the wind power and photovoltaic power plant clusters based on the first prediction information.

Images