Power distribution network wind-solar-hydrogen multi-energy complementary regulation method and system based on cloud edge end cooperation

Abstract

The application provides a power distribution network wind-solar-hydro multi-energy complementary regulation method and system based on cloud edge-end cooperation, and the regulation method is applied to a dispatching system including a dispatching center layer, a multi-energy complementary cluster layer and a unit control layer, and specifically comprises the following steps: clustering wind turbine generators and photovoltaic generators respectively to obtain wind power clusters and photovoltaic clusters, and complementarily pairing according to a wind-solar complementary degree to form a wind-solar complementary aggregate, combining a hydrogen energy system to form a multi-energy complementary cluster, constructing an optimal dispatching model in combination with conventional units, operating loads and power distribution network operation constraints, solving to obtain power distribution network dispatching strategies in each dispatching period, formulating and issuing corresponding output plans and operating strategies according to the power distribution network dispatching strategies, and controlling the operation of corresponding wind turbine generators, photovoltaic generators and hydrogen energy system equipment. The application can reduce the complexity of dispatching, inhibit and relieve source side power fluctuation and uncertainty through wind-solar complementation and hydrogen energy system consumption, and guarantee the operating stability of the power distribution network.

Description

Technical Field

[0001] This invention relates to the field of power dispatching technology, and in particular to a method and system for multi-energy complementary regulation of distribution networks based on cloud-edge-device collaboration, encompassing wind, solar, and hydrogen energy. Background Technology

[0002] The installed capacity of distributed renewable energy sources such as wind power and photovoltaic power in the distribution network is showing a continuous upward trend. Affected by the inherent intermittency, volatility and spatial dispersion of these energy sources, problems such as voltage fluctuation, power flow reversal and line overload are becoming increasingly prominent, posing a severe challenge to the safe and stable operation of traditional distribution networks.

[0003] The existing distribution network control mostly adopts a two-layer structure of dispatch center and substation. The dispatch center centrally completes the optimization of the whole network and issues instructions, while the substation layer only performs specific control. In scenarios with a high proportion of distributed new energy and multiple energy coupling, this architecture faces problems such as frequent information interaction, high dependence on communication, and computational pressure caused by massive resources that are prone to the curse of dimensionality, resulting in high dispatch complexity.

[0004] While some have proposed using methods like zoning and aggregation to cluster distributed renewable energy sources and reduce the complexity of global dispatching in the dispatch center, these methods often focus on a single energy source or perform simple statistical aggregation, failing to reflect the comprehensive regulation capabilities of the source side. They also ignore the complementary nature of wind and solar power and cannot alleviate the drastic fluctuations and uncertainties in source-side power. Furthermore, a hydrogen energy system, composed of electrolytic hydrogen production devices, hydrogen storage facilities, and fuel cells, can utilize its electricity-hydrogen-electricity link to absorb surplus electricity during periods of high renewable energy output and compensate for power gaps when output is insufficient. This effectively mitigates the output fluctuations of wind and solar power, contributing to the safe and stable operation of the distribution network. However, existing distribution network dispatching largely ignores the ability of hydrogen energy systems to suppress output fluctuations in distributed energy sources, treating them merely as simple loads or independent energy storage units. This still fails to balance the contradiction between renewable energy consumption and grid safety, making it difficult to guarantee the safe and stable operation of the distribution network. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies that use distributed energy cluster management for power grid regulation. These shortcomings include the difficulty in reflecting the comprehensive regulation capacity of the source side through the aggregation of distributed energy resources, the neglect of the renewable energy absorption and fluctuation suppression capabilities of hydrogen energy systems, and the difficulty in balancing the contradiction between renewable energy absorption and grid safety operation. This invention provides a multi-energy complementary regulation method and system for power grids based on cloud-edge-device collaboration, which clusters wind power and photovoltaic power according to their output characteristics and regulation capabilities. Wind power and photovoltaic power clusters are formed by pairing them with the goal of maximizing the total wind-solar complementarity. Hydrogen energy systems are then coupled to construct multi-energy complementary clusters. This reduces the complexity of scheduling while mitigating power fluctuations and uncertainties on the source side through wind-solar complementarity and hydrogen energy system absorption and suppression. Furthermore, scheduling analysis is performed based on the formed multi-energy complementary clusters, and corresponding strategies are issued hierarchically to improve the accuracy and reliability of the generated power grid scheduling strategies, thereby ensuring the safe and stable operation of the power grid.

[0006] The objective of this invention is achieved through the following technical solution:

[0007] A multi-energy complementary regulation method for power distribution networks based on cloud-edge-device collaboration includes:

[0008] Based on power output characteristics and regulation capabilities, wind turbine units and photovoltaic units are clustered separately, and wind power clusters and photovoltaic clusters are obtained by aggregating the clustering results.

[0009] Based on the wind-solar complementarity index, with the goal of maximizing the total wind-solar complementarity, wind power clusters and photovoltaic clusters are paired to obtain the optimal combination scheme.

[0010] The optimal combination scheme is used to form a wind-solar complementary aggregate, and combined with the hydrogen energy system to form a multi-energy complementary cluster;

[0011] Based on multi-energy complementary clusters, an optimal scheduling model is constructed by combining conventional units, operating loads and distribution network operation constraints. The optimal scheduling model is solved to obtain the distribution network scheduling strategy for each scheduling period.

[0012] The power output plans for wind power clusters and photovoltaic clusters, as well as the operation strategies for hydrogen energy systems, are determined based on the power distribution network dispatch strategy.

[0013] According to the corresponding scheduling period, the power generation plan and operation strategy are sent to the corresponding wind power cluster, photovoltaic cluster and hydrogen energy system to control the operation of each equipment in the corresponding wind turbine, photovoltaic unit and hydrogen energy system.

[0014] Furthermore, the clustering of wind turbine units and photovoltaic units based on output characteristics and regulation capabilities includes:

[0015] Historical power generation curves of each wind turbine and photovoltaic unit were collected. Based on the historical power generation curves, corresponding output characteristics were extracted. The original dataset was constructed by combining the corresponding installed capacity and regulation capacity. The original dataset was then preprocessed.

[0016] Several distinct data points are randomly selected from the preprocessed dataset as initial cluster centers. The DTW distance between each data point in the dataset and each initial cluster center is calculated based on the corresponding power output characteristics. The corresponding Euclidean distance is calculated based on the installed capacity and regulation capability.

[0017] The corresponding DTW distance and Euclidean distance are weighted and summed to obtain the corresponding comprehensive distance.

[0018] Based on the corresponding comprehensive distance, each data point in the dataset is assigned to the cluster corresponding to the nearest initial cluster center;

[0019] Calculate the cumulative distance within each cluster for each unit, update the cluster center based on the cumulative distance, recalculate the comprehensive distance between each data point in the dataset and each cluster center, and allocate the data points until the iteration termination condition is met.

[0020] Furthermore, the step of aggregating and obtaining wind power clusters and photovoltaic clusters based on clustering results includes:

[0021] Based on the clustering results, all clusters are obtained. All wind turbine units belonging to the same cluster are merged into an initial wind power cluster, and all photovoltaic units belonging to the same cluster are merged into an initial photovoltaic cluster.

[0022] The initial wind power cluster and the initial photovoltaic cluster are subjected to consistency verification. Based on the preset threshold, the initial wind power cluster and the initial photovoltaic cluster are split or merged respectively to obtain the final wind power cluster and photovoltaic cluster.

[0023] Based on the historical output curves, regulation capacity, and installed capacity of all units within the wind power cluster and photovoltaic cluster, corresponding cluster-standardized feature parameters are extracted, and annotations are added to each wind power cluster and photovoltaic cluster according to the corresponding cluster-standardized feature parameters.

[0024] Furthermore, the calculation process of the wind-solar complementarity index includes:

[0025] Obtain the corresponding power output sequence based on the cluster standardization parameters labeled by wind power clusters and photovoltaic clusters;

[0026] The correlation coefficient between the wind power cluster and the photovoltaic cluster is calculated based on the corresponding power output sequence. The correlation coefficient is then mapped to a predetermined interval to obtain the corresponding correlation complementarity index.

[0027] Based on the corresponding power output sequence and combined with the preset cluster power output state threshold, the state-type complementarity index between wind power clusters and photovoltaic clusters is calculated through the indicator function.

[0028] The correlation-type complementarity index and the state-type complementarity index are weighted and summed according to the preset weights to obtain the wind-solar complementarity index value between the corresponding wind power cluster and photovoltaic cluster.

[0029] Furthermore, based on the wind-solar complementarity index, and with the goal of maximizing the total wind-solar complementarity, the wind power clusters and photovoltaic clusters are paired complementaryly to obtain the optimal combination scheme, including:

[0030] Calculate the wind-solar complementarity index for each pair of wind power clusters and photovoltaic clusters, and construct a wind-solar complementarity matrix.

[0031] Based on the wind-solar complementarity matrix, with the optimization objective of maximizing the total wind-solar complementarity of all paired clusters and the constraint of one-to-one cluster pairing, an integer programming pairing model is constructed between wind power clusters and photovoltaic clusters.

[0032] Solve the integer programming pairing model to obtain the optimal photovoltaic cluster corresponding to each wind power cluster, and form the optimal combination scheme of wind power cluster and photovoltaic cluster.

[0033] Furthermore, the process of forming a wind-solar hybrid aggregate based on the optimal combination scheme and systematically forming a multi-energy complementary cluster by combining it with hydrogen energy includes:

[0034] Based on each wind power cluster and photovoltaic cluster combination in the optimal combination scheme, a corresponding wind-solar hybrid aggregate is formed, and the equivalent parameters and operating constraints of each wind-solar hybrid aggregate are calculated based on all units in the corresponding wind power cluster and photovoltaic cluster.

[0035] Based on the equipment type of the hydrogen energy system, corresponding operating parameters and constraints are set. Combined with the equivalent parameters and operating constraints of each wind-solar hybrid agglomeration, the hydrogen energy system and the wind-solar hybrid agglomeration are integrated and modeled to form a multi-energy complementary cluster.

[0036] Furthermore, the optimization scheduling model, based on multi-energy complementary clusters and combined with conventional units, operating loads, and distribution network operating constraints, is constructed. Solving the optimization scheduling model yields the distribution network scheduling strategy for each scheduling period, including:

[0037] The optimization objective is to minimize the sum of operating costs, renewable energy abandonment penalty costs, and carbon emission costs, while maximizing hydrogen sales revenue. The constraints are the operation constraints of the distribution network, conventional units, and multi-energy complementary clusters. An optimization scheduling model is constructed based on the equivalent parameters of the multi-energy complementary clusters, the operating parameters of conventional units, and the load data of the operating load.

[0038] Solve the optimized scheduling model to obtain the distribution network scheduling strategy for each scheduling period.

[0039] Furthermore, the determination of the output plans for wind power clusters and photovoltaic clusters and the operation strategy for the hydrogen energy system based on the distribution network dispatch strategy includes:

[0040] The control targets of each multi-energy complementary cluster are obtained based on the distribution network dispatch strategy, and the control targets are decomposed according to the cluster standardized characteristic parameters of the corresponding wind power cluster and photovoltaic cluster to obtain the time-period control targets of the corresponding wind power cluster and photovoltaic cluster.

[0041] Predict the available power generation of the corresponding wind power cluster and photovoltaic cluster during the specified time period. Combine the wind-solar complementarity index of the corresponding wind power cluster and photovoltaic cluster with the operating parameters and operating constraints of the hydrogen energy system to allocate the time period control targets, and obtain the output plan of the corresponding wind power cluster and photovoltaic cluster and the operation strategy of the hydrogen energy system for each multi-energy complementary cluster.

[0042] Furthermore, the step of issuing power generation plans and operating strategies to the corresponding wind power clusters, photovoltaic clusters, and hydrogen energy systems according to the corresponding scheduling period, and controlling the operation of each piece of equipment in the corresponding wind turbines, photovoltaic units, and hydrogen energy systems, includes:

[0043] Based on the corresponding scheduling period, the power generation plan and operation strategy are sent to the corresponding wind power cluster, photovoltaic cluster and hydrogen energy system, and the corresponding unit operation instructions are formulated in combination with the real-time operation status of the wind power cluster, photovoltaic cluster and hydrogen energy system;

[0044] Each piece of equipment in the corresponding wind turbine, photovoltaic unit, and hydrogen energy system responds to the corresponding unit's operating command and adjusts its corresponding operating status.

[0045] A cloud-edge-device collaborative power distribution network wind-solar-hydrogen multi-energy complementary regulation system, used to execute any of the above-mentioned multi-energy complementary regulation methods, including:

[0046] The dispatch center layer is used to screen complementary pairings of wind power clusters and photovoltaic clusters, and to form multi-energy complementary clusters by combining hydrogen energy systems. At the same time, it constructs an optimized dispatch model by combining conventional units, operating loads and distribution network operating constraints, and solves the optimized dispatch model to obtain the distribution network dispatch strategy for each dispatch period.

[0047] The multi-energy complementary cluster layer responds to the distribution network dispatch strategy of the dispatch center and formulates corresponding unit operation instructions based on the real-time operating status of wind power clusters, photovoltaic clusters and hydrogen energy systems.

[0048] The unit control layer, including wind turbines, photovoltaic units, and hydrogen energy systems, adjusts its corresponding operating status in response to the unit operation commands from the multi-energy complementary cluster layer.

[0049] The beneficial effects of this invention are:

[0050] By combining DTW distance and Euclidean distance weighted calculation, a large number of dispersed wind turbines and photovoltaic units are aggregated into a small number of standardized single energy clusters. After consistency verification and splitting and merging optimization, the optimal pairing of wind power and photovoltaic clusters is further completed to form aggregates with the goal of maximizing the total wind-solar complementarity. Peak-valley misalignment characteristics are explored to alleviate intraday power fluctuations from the source side. In addition, a multi-energy complementary cluster is constructed by coupling a hydrogen energy system to suppress power gaps under short-term fluctuations and extreme operating conditions. While reducing the number of dispatch objects and reducing the data processing pressure and computational complexity of distribution network dispatch, it also reduces safety risks such as voltage fluctuations, power flow reversal, and line overload, ensuring the safe and stable operation of the distribution network.

[0051] Based on the obtained multi-energy complementary clusters, the distribution network dispatch strategy is obtained by solving the solution. Then, by extracting the cluster standardized characteristic parameters and decomposing the control target, combined with the prediction of wind and solar power generation during the wind and solar seasons, the power output plan can be accurately allocated. Combined with the new energy consumption capacity of the hydrogen energy system, the local consumption rate of wind and solar power can be improved, and the problem of wind and solar curtailment can be reduced.

[0052] Furthermore, relying on the three-layer collaborative architecture of cloud, edge, and terminal, the scheduling center layer is responsible for global optimization and strategy formulation, the multi-energy complementary cluster layer undertakes regional coordination and instruction refinement, and the unit control layer executes equipment-level adjustment and autonomous response, realizing the hierarchical distribution of scheduling strategies, reducing dependence on a single scheduling center and communication link, and improving response speed and operational reliability under disturbances and extreme operating conditions. Attached Figure Description

[0053] Figure 1 This is a schematic diagram of a process of the present invention;

[0054] Figure 2 This is a schematic diagram of the cloud-edge architecture of the multi-energy complementary control system according to an embodiment of the present invention. Detailed Implementation

[0055] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0056] Example:

[0057] A multi-energy complementary regulation method for power distribution networks based on cloud-edge-device collaboration, such as wind, solar, and hydrogen energy, etc. Figure 1 As shown, it includes:

[0058] Based on power output characteristics and regulation capabilities, wind turbine units and photovoltaic units are clustered separately, and wind power clusters and photovoltaic clusters are obtained by aggregating the clustering results.

[0059] Based on the wind-solar complementarity index, with the goal of maximizing the total wind-solar complementarity, wind power clusters and photovoltaic clusters are paired to obtain the optimal combination scheme.

[0060] The optimal combination scheme is used to form a wind-solar complementary aggregate, and combined with the hydrogen energy system to form a multi-energy complementary cluster;

[0061] Based on multi-energy complementary clusters, an optimal scheduling model is constructed by combining conventional units, operating loads and distribution network operation constraints. The optimal scheduling model is solved to obtain the distribution network scheduling strategy for each scheduling period.

[0062] The power output plans for wind power clusters and photovoltaic clusters, as well as the operation strategies for hydrogen energy systems, are determined based on the power distribution network dispatch strategy.

[0063] According to the corresponding scheduling period, the power generation plan and operation strategy are sent to the corresponding wind power cluster, photovoltaic cluster and hydrogen energy system to control the operation of each equipment in the corresponding wind turbine, photovoltaic unit and hydrogen energy system.

[0064] For multi-energy coupling scenarios involving high proportions of wind power, photovoltaics, and hydrogen energy systems including electrolysis hydrogen production and fuel cells, a hierarchical autonomous collaborative control framework based on cloud-edge-device architecture is constructed to fully utilize the complementary characteristics of wind and solar power and the flexibility of hydrogen energy systems, thereby mitigating the volatility and uncertainty issues brought about by large-scale grid connection of new energy sources.

[0065] Distributed renewable energy sources are numerous and widely distributed, with many small-scale access points. Managing them individually would place enormous pressure on the cloud-based dispatch center in terms of data processing and real-time control. Therefore, by pre-aggregating and partitioning renewable energy systems that are geographically close and have similar performance characteristics, renewable energy clusters can be formed. This reduces the number of systems to be managed, facilitates unified dispatch and management, reduces communication latency, and improves the grid's rapid response capabilities.

[0066] This embodiment uses an optimized K-medoids clustering algorithm based on time series clustering and DTW (Dynamic Time Warping) distance to classify and aggregate new energy systems. The DTW distance is used to measure the similarity between time series data of different new energy plants and stations. By leveraging its adaptability to time shifts and nonlinear deformations in time series, the accuracy of clustering analysis is improved.

[0067] The clustering of wind turbine units and photovoltaic units based on output characteristics and regulation capabilities includes:

[0068] Historical power generation curves of each wind turbine and photovoltaic unit were collected. Based on the historical power generation curves, corresponding output characteristics were extracted. The original dataset was constructed by combining the corresponding installed capacity and regulation capacity. The original dataset was then preprocessed.

[0069] Several distinct data points are randomly selected from the preprocessed dataset as initial cluster centers. The DTW distance between each data point in the dataset and each initial cluster center is calculated based on the corresponding power output characteristics. The corresponding Euclidean distance is calculated based on the installed capacity and regulation capability.

[0070] The corresponding DTW distance and Euclidean distance are weighted and summed to obtain the corresponding comprehensive distance.

[0071] Based on the corresponding comprehensive distance, each data point in the dataset is assigned to the cluster corresponding to the nearest initial cluster center;

[0072] Calculate the cumulative distance within each cluster for each unit, update the cluster center based on the cumulative distance, recalculate the comprehensive distance between each data point in the dataset and each cluster center, and allocate the data points until the iteration termination condition is met.

[0073] Using installed capacity, regulation capability, and historical power generation curves as clustering indicators, this study measures the similarity among new energy sources from two dimensions: dynamic time-series characteristics and static parameters, thereby aggregating new energy sources with similar characteristics. Installed capacity directly reflects power generation potential and economies of scale, while regulation capability reflects the flexibility of individual new energy sources in operation and their responsiveness to grid dispatch. Historical power generation curves integrate output characteristics over time, containing information on the seasonality, diurnal variation, and weather dependence of new energy power generation, reflecting the spatiotemporal distribution patterns of distributed new energy output.

[0074] To optimize subsequent clustering efficiency, the historical power generation curves are used to extract output characteristics that reflect the power generation time sequence, including peak and valley periods, fluctuation range, and duration of continuous power generation.

[0075] An original dataset is constructed based on power output characteristics, installed capacity, and regulation capability, and then preprocessed. The preprocessing operations include outlier removal, missing value completion, and data normalization to ensure a uniform data format and provide accurate input for subsequent clustering.

[0076] Then, select several distinct data points from the preprocessed original dataset as initial cluster centers.

[0077] For each remaining data point in the preprocessed original dataset, calculate its distance to all initial cluster centers, and then assign it to the cluster corresponding to the nearest initial cluster center.

[0078] Specifically, DTW distance and Euclidean distance are used to measure the distance between the two dimensions of data.

[0079] For time series data on output characteristics, the DTW distance is used to measure the distance between two data points. When calculating the DTW distance, for any two time series... and initialize a matrix .

[0080] Fill the matrix using dynamic programming Its expression is:

[0081] ;

[0082] .

[0083] in, for The first element and The DTW distance between the first elements, express The former elements and The former DTW distance between elements.

[0084] The final result That is, two time series and The DTW distance between them.

[0085] For static parameters such as installed capacity and regulation capability, the distance between two data points is measured using Euclidean distance.

[0086] Then, based on the preset distance weights, the DTW distance of the output characteristic dimension, the Euclidean distance of the installed capacity dimension, and the Euclidean distance of the regulation capability dimension are weighted and summed to obtain a comprehensive distance that takes into account both time-series characteristics and static parameters, which is used as the basis for clustering.

[0087] Then, following the principle of proximity, each data point is assigned to the cluster corresponding to the initial cluster center with the closest comprehensive distance, thus completing the first round of clustering assignment.

[0088] Then, calculate the cumulative distance between all units in each cluster and other units in the cluster. Select the unit with the smallest cumulative distance as the new cluster center. After updating the cluster centers, recalculate the comprehensive distance between each data point and each new cluster center and redistribute the data points. Repeat the above process of updating the cluster centers and redistributing the data points until the set of cluster centers no longer changes after two consecutive iterations, or until the preset number of iterations is reached. At this point, the clustering converges.

[0089] Units in the same cluster have similar installed capacity, output characteristics and regulation capabilities. Based on the clustering results, wind turbines and photovoltaic units are aggregated to form wind power clusters and photovoltaic clusters, thereby reducing the number of management objects to be scheduled and reducing the computational pressure caused by the dimensionality explosion problem.

[0090] The step of aggregating wind power clusters and photovoltaic clusters based on clustering results includes:

[0091] Based on the clustering results, all clusters are obtained. All wind turbine units belonging to the same cluster are merged into an initial wind power cluster, and all photovoltaic units belonging to the same cluster are merged into an initial photovoltaic cluster.

[0092] The initial wind power cluster and the initial photovoltaic cluster are subjected to consistency verification. Based on the preset threshold, the initial wind power cluster and the initial photovoltaic cluster are split or merged respectively to obtain the final wind power cluster and photovoltaic cluster.

[0093] Based on the historical output curves, regulation capacity, and installed capacity of all units within the wind power cluster and photovoltaic cluster, corresponding cluster-standardized feature parameters are extracted, and annotations are added to each wind power cluster and photovoltaic cluster according to the corresponding cluster-standardized feature parameters.

[0094] By merging wind turbines and photovoltaic units within the same cluster, units with similar characteristics are integrated into a whole to form an initial cluster. However, since the initial cluster is formed solely based on the clustering algorithm results, there may be biases, such as individual abnormal units being mixed into a certain cluster, some clusters being too small in size, or insufficient purity of characteristics. Therefore, further consistency checks are used to correct the cluster biases caused by clustering.

[0095] The consistency check primarily verifies whether the output characteristics and regulation capabilities of all units within the initial cluster meet preset consistency thresholds, such as the upper limit of the standard deviation of output fluctuations and the range of regulation capability differences. If the characteristics of some units within an initial cluster differ from those of other units beyond the threshold, it indicates a deviation in their classification, requiring splitting to ensure internal consistency within each sub-cluster after splitting. If the characteristics of different initial clusters are highly similar, i.e., the differences are below the threshold, merging them can improve the cluster's scale effect and output stability, then merging is performed. Through splitting or merging, the final wind power cluster and photovoltaic cluster are formed.

[0096] Furthermore, for the wind power clusters and photovoltaic clusters formed, corresponding standardized cluster characteristic parameters are generated based on their historical output curves, regulation capabilities, and installed capacity, and corresponding annotations are added to optimize the efficiency of subsequent scheduling calculations.

[0097] The standardized characteristic parameters of the cluster include the typical output ratio curve after standardization, the output fluctuation coefficient, the maximum adjustable power ratio, the installed capacity ratio, etc., which can be set according to actual needs.

[0098] The output characteristics of wind and solar power naturally exhibit a complementary tendency. Solar power generation peaks during periods of abundant sunshine, while wind power generation is more advantageous at night or during periods of weaker sunlight. Previously formed wind and solar power clusters already possess standardized output characteristics and adjustment capabilities. Further optimization through complementary pairing between clusters maximizes the synergistic effect of wind and solar power generation, addressing the inherent volatility and intermittency of wind and solar output. Specifically, pairing is targeted at maximizing the overall wind-solar complementarity, precisely matching wind and solar clusters with opposite peak and valley periods and complementary fluctuation trends. This achieves peak shaving and valley filling, improving the overall stability and predictability of combined wind and solar power generation, and avoiding the impact of concentrated output or deficits of a single energy source on the power distribution network.

[0099] Before performing cluster matching, a wind-solar complementarity index is first constructed to quantify the complementarity between wind power clusters and photovoltaic clusters.

[0100] The calculation process for the wind-solar complementarity index includes:

[0101] Obtain the corresponding power output sequence based on the cluster standardization parameters labeled by wind power clusters and photovoltaic clusters;

[0102] The correlation coefficient between the wind power cluster and the photovoltaic cluster is calculated based on the corresponding power output sequence. The correlation coefficient is then mapped to a predetermined interval to obtain the corresponding correlation complementarity index.

[0103] Based on the corresponding power output sequence and combined with the preset cluster power output state threshold, the state-type complementarity index between wind power clusters and photovoltaic clusters is calculated through the indicator function.

[0104] The correlation-type complementarity index and the state-type complementarity index are weighted and summed according to the preset weights to obtain the wind-solar complementarity index value between the corresponding wind power cluster and photovoltaic cluster.

[0105] By extracting the time-based output characteristics of each cluster from the standardized feature parameters labeled in each cluster, typical output ratio curves reflecting the output pattern of the clusters in different scheduling periods were recorded. By extracting the time-series data of these curves, they were transformed into numerical sequences arranged by scheduling periods to obtain the output sequences of wind power clusters and photovoltaic clusters.

[0106] Then, the correlation coefficient between the wind power cluster and the photovoltaic cluster is calculated based on the corresponding power output sequence, and mapped to a predetermined interval to obtain a correlation-type complementarity index, so as to evaluate the complementarity from the perspective of the overall power output trend.

[0107] This embodiment specifically calculates the correlation coefficient between wind power clusters and photovoltaic clusters using the Pearson correlation coefficient, the expression of which is:

[0108] ;

[0109] in, For the first The first wind power cluster and the first The correlation coefficient between individual photovoltaic clusters The stronger the negative correlation, the greater the misalignment between peak and valley power outputs, and the better the complementarity. For the first A wind power cluster in A typical example of someone who contributes effort at any given moment. For the first A photovoltaic cluster in A typical example of someone who contributes effort at any given moment. and The first The first wind power cluster and the first The average power output of a photovoltaic cluster This represents the number of time periods corresponding to the output sequence.

[0110] To facilitate the unification of dimensions with other indicators, the correlation coefficient is also mapped to a complementarity coefficient within a preset interval of [0, 1], i.e., the correlation-type complementarity index value, the expression of which is:

[0111] ;

[0112] in, For the first The first wind power cluster and the first The correlation complementarity index value between photovoltaic clusters, when hour , indicating complete anticorrelation and the strongest complementarity; when hour This indicates that they are completely in the same direction and have virtually no complementarity.

[0113] Based on this, a state-type complementarity index is calculated using an indicator function based on the output sequence and a preset threshold, which evaluates complementarity from the perspective of time period operation status and makes up for the limitation that the correlation coefficient only reflects a linear relationship.

[0114] The preset cluster output state threshold includes a low output threshold. and high output threshold , and , , .

[0115] Based on the low output threshold and high output threshold The formula for calculating the state-type complementarity index is:

[0116] ;

[0117] in, For the first The first wind power cluster and the first State-type complementarity index values ​​among photovoltaic clusters Its value reflects the proportion of time that the wind-low-light-high or wind-high-light-low conditions occur within the statistical time domain; the larger the value, the stronger the complementarity. For the first The minimum active power output of a wind power cluster in all time periods of the corresponding power output sequence. For the first The maximum active power output of a wind power cluster in all time periods of the corresponding power output sequence. For the first The minimum active power output of a photovoltaic cluster in all time periods of the corresponding power output sequence. For the first The maximum active power output of a photovoltaic cluster in all time periods of the corresponding power output sequence. This is an indicator function; it takes the value 1 when the condition within the parentheses is met, and 0 otherwise.

[0118] Based on preset weights, a weighted fusion of correlation-type complementarity indicators and state-type complementarity indicators is performed to obtain the wind power cluster. and photovoltaic clusters The combined wind-solar complementarity index is used to comprehensively and objectively measure the degree of complementarity between wind power clusters and photovoltaic clusters. The weights can be determined based on the distribution network's emphasis on output trend stability and time-period state matching; for example, in scenarios where long-term dispatch stability is the dispatch objective, the weight of correlation-type complementarity can be increased.

[0119] The formula for calculating the wind-solar complementarity index is as follows:

[0120] ;

[0121] in, and These are the preset weights for correlation-type complementarity indicators and state-type complementarity indicators, respectively.

[0122] Based on the wind-solar complementarity index, we seek the optimal combination of wind power clusters and photovoltaic clusters.

[0123] The step of matching wind power clusters and photovoltaic clusters based on the wind-solar complementarity index, with the goal of maximizing the total wind-solar complementarity, to obtain the optimal combination scheme includes:

[0124] Calculate the wind-solar complementarity index for each pair of wind power clusters and photovoltaic clusters, and construct a wind-solar complementarity matrix.

[0125] Based on the wind-solar complementarity matrix, with the optimization objective of maximizing the total wind-solar complementarity of all paired clusters and the constraint of one-to-one cluster pairing, an integer programming pairing model is constructed between wind power clusters and photovoltaic clusters.

[0126] Solve the integer programming pairing model to obtain the optimal photovoltaic cluster corresponding to each wind power cluster, and form the optimal combination scheme of wind power cluster and photovoltaic cluster.

[0127] For each wind power cluster and each photovoltaic cluster, the corresponding wind-solar complementarity index value is calculated. All wind power clusters are arranged by row, and all photovoltaic clusters are arranged by column, so that the paired wind-solar complementarity index values ​​are filled into the corresponding positions, forming a wind-solar complementarity matrix. The wind-solar complementarity matrix transforms the discrete complementarity data into structured two-dimensional data, intuitively displaying the degree of complementarity of all possible pairings.

[0128] Based on the established wind-solar complementarity matrix, decision variables are defined, and their expressions are as follows:

[0129] ;

[0130] in, As decision variables, It is a combination of wind power clusters and photovoltaic clusters. For wind power clusters, It is a collection of photovoltaic clusters.

[0131] The system sets cluster pairing as a constraint and maximizes the total wind-solar complementarity of all paired clusters as the optimization objective, thus constructing an integer programming pairing model.

[0132] The expression for the constraints of the integer programming pairing model is as follows:

[0133] ;

[0134] .

[0135] The objective function of the integer programming pairing model is expressed as follows:

[0136] ;

[0137] in, The total wind-solar complementarity.

[0138] Based on the constructed integer programming pairing model, the optimal values ​​of the decision variables are quickly obtained by using algorithms such as the Hungarian algorithm and the branch and bound method. Then, based on the solution results, the optimal pairing combination of all wind power clusters and photovoltaic clusters with a value of 1 is selected.

[0139] Then, based on the selected optimal pairing, the corresponding wind power clusters and photovoltaic clusters are aggregated into the same dispatch target, namely, a wind-solar complementary aggregate, to achieve peak shaving and valley filling of power output. However, even if the complementarity reaches the optimal level, the aggregate may still experience a large output gap or surplus under extreme weather or load change scenarios. Relying solely on wind and solar self-regulation cannot meet the distribution network's requirements for power supply stability and reliability.

[0140] Hydrogen energy systems have the characteristic of bidirectional electricity-hydrogen-electricity conversion. When the polymer has surplus power, the excess electricity can be used to electrolyze water to produce hydrogen and store it. When the polymer has insufficient power, hydrogen can be converted into electricity through fuel cells and fed into the grid. This effectively smooths out the output fluctuations of the polymer, solves the problem of wind and solar curtailment, fills the power supply gap, and improves the stability of energy supply.

[0141] Therefore, based on the formation of a wind-solar complementary agglomeration through the optimal combination scheme, a hydrogen energy system is further introduced to integrate it into a multi-energy complementary cluster, thereby improving the flexibility of the cluster in participating in power grid dispatch.

[0142] The step of forming a wind-solar complementary aggregate based on the optimal combination scheme and systematically forming a multi-energy complementary cluster by combining it with hydrogen energy includes:

[0143] Based on each wind power cluster and photovoltaic cluster combination in the optimal combination scheme, a corresponding wind-solar hybrid aggregate is formed, and the equivalent parameters and operating constraints of each wind-solar hybrid aggregate are calculated based on all units in the corresponding wind power cluster and photovoltaic cluster.

[0144] Based on the equipment type of the hydrogen energy system, corresponding operating parameters and constraints are set. Combined with the equivalent parameters and operating constraints of each wind-solar hybrid agglomeration, the hydrogen energy system and the wind-solar hybrid agglomeration are integrated and modeled to form a multi-energy complementary cluster.

[0145] Let the set of wind-solar hybrid aggregates formed by combining wind power clusters and photovoltaic clusters corresponding to the optimal combination scheme be denoted as . ,gather The Middle The set of units contained in a wind-solar hybrid complex is denoted as The units in a wind-solar hybrid complex include all the units within the wind power cluster and the solar power cluster of that complementary pair.

[0146] The equivalent parameters of the wind-solar hybrid power complex include the equivalent power output curve, the equivalent total installed capacity, and the equivalent regulation capacity boundary. Taking the equivalent power output curve as an example, the equivalent active power output for each time period is obtained by calculating the sum of the power output of all units within the wind-solar hybrid power complex, and then the corresponding equivalent power output curve is formed by sorting according to the time period.

[0147] The formula for calculating the equivalent active power output is:

[0148] ;

[0149] in, For the first A wind-solar hybrid complex in The equivalent effort exerted at any given moment For the first Each unit Power generation at any given moment For the first A collection of units in a wind-solar hybrid complex. For the first The first of the wind-solar hybrid complexes Each unit Wind power generation capacity at any given time For the first The first of the wind-solar hybrid complexes Each unit Photovoltaic power generation at any given time.

[0150] The operational constraints include equivalent power constraints and equivalent ramp constraints.

[0151] The expression for the equivalent power constraint is as follows:

[0152] ;

[0153] in, For the first The equivalent active power output of a wind-solar hybrid system For the first The maximum equivalent active power output of a wind-solar hybrid power system For the first The first of the wind-solar hybrid complexes Each unit Maximum power generation at any given time.

[0154] Because the output peaks and valleys of the units within the wind-solar hybrid complex are misaligned, when the output of some units approaches the upper limit, the output of others often still has a certain adjustment margin. Based on this, a cluster ramp enhancement coefficient is introduced to construct the corresponding equivalent ramp constraint.

[0155] The expression for the cluster climbing enhancement coefficient is as follows:

[0156] ;

[0157] in, For the first The cluster slope enhancement coefficient of a wind-solar hybrid complex These are empirical parameters. For the first The wind-solar complementarity of a wind-solar hybrid cluster. When the cluster complementarity is weak. ,have It degenerates into a nominal ramp constraint; when complementarity increases This corresponds to a larger equivalent climbing space.

[0158] Based on the cluster ramp enhancement coefficient, the expression for the equivalent ramp constraint is as follows:

[0159] ;

[0160] in, For the first A wind-solar hybrid complex in Equivalent work output at any given moment For the first A wind-solar hybrid complex in The equivalent effort exerted at any given moment For the first The first of the wind-solar hybrid complexes The ramp-up capability of a unit, i.e., the maximum regulation capacity per unit time.

[0161] The hydrogen energy system includes hydrogen electrolysis equipment, hydrogen storage equipment, and fuel cells. Based on the corresponding equipment, corresponding operating parameters and constraints are set, and then mathematical models of each device in the hydrogen energy system are performed to provide a data foundation for the establishment of a multi-energy complementary cluster.

[0162] The operating parameters of the hydrogen energy system include the rated hydrogen production power and hydrogen production efficiency of the electrolyzer, the maximum hydrogen storage capacity and hydrogen charging / discharging rate limit of the hydrogen storage tank, and the rated power generation and power generation efficiency of the fuel cell. The operating constraints include equipment start-up and shutdown constraints, hydrogen storage constraints, and supply and demand balance constraints for various devices.

[0163] The mathematical model of the integrated multi-energy complementary cluster is as follows:

[0164] ;

[0165] in, For the first Each wind-solar hybrid complex corresponds to a multi-energy complementary cluster. The equivalent effort exerted at any given moment For the first Each wind-solar hybrid polymer complex corresponds to an integrated hydrogen energy system and its corresponding electrolysis hydrogen production equipment. Input power at time , For the first Each wind-solar hybrid power plant corresponds to a fuel cell integrating a hydrogen energy system. Electricity generation at any given moment.

[0166] The equivalent power operating range constraint for the multi-energy complementary cluster is:

[0167] ;

[0168] in, For the first The lower limit of the input power of the electrolytic hydrogen production equipment corresponding to the integrated hydrogen energy system of each wind-solar hybrid polymer. For the first The upper limit of the input power of the electrolytic hydrogen production equipment corresponding to the integrated hydrogen energy system of each wind-solar hybrid polymer.

[0169] The equivalent ramp constraint is expressed as follows:

[0170] ;

[0171] in, For the first Each wind-solar hybrid complex corresponds to a multi-energy complementary cluster. The equivalent effort exerted at any given moment and The first Each wind-solar hybrid unit corresponds to the upper limit of the regulation capacity of the electrolytic hydrogen production equipment and fuel cell of the integrated hydrogen energy system.

[0172] By forming a multi-energy complementary cluster, resources such as wind farms, photovoltaic power plants, electrolytic hydrogen production devices, hydrogen storage, and fuel cells within the distribution network dispatch area are modeled and aggregated in a unified spatiotemporal dimension. This reduces the dimensionality curse caused by modeling individual wind turbines, photovoltaic arrays, and hydrogen energy devices one by one at the model level. It allows wind-solar-hydrogen systems to participate in distribution network or regional power grid dispatch optimization as a small number of equivalent cluster units, enhancing the solvability and engineering feasibility of the dispatch model. Simultaneously, through wind-solar complementary aggregation and hydrogen energy systems such as electrolytic hydrogen production and fuel cells, peak shaving and valley filling are performed in time, mitigating the impact of uncertainties in new energy output on grid operation, reducing wind and solar curtailment, and improving system flexibility and power supply reliability.

[0173] Based on the constructed multi-energy complementary cluster, and combined with conventional units and load models, an optimized scheduling model including power flow and voltage constraints is constructed, and then the initial distribution network scheduling strategy is obtained by solving the model.

[0174] The aforementioned optimization scheduling model, based on multi-energy complementary clusters and combined with conventional units, operating loads, and distribution network operating constraints, is used to solve the optimization scheduling model to obtain the distribution network scheduling strategy for each scheduling period. This includes:

[0175] The optimization objective is to minimize the sum of operating costs, renewable energy abandonment penalty costs, and carbon emission costs, while maximizing hydrogen sales revenue. The constraints are the operation constraints of the distribution network, conventional units, and multi-energy complementary clusters. An optimization scheduling model is constructed based on the equivalent parameters of the multi-energy complementary clusters, the operating parameters of conventional units, and the load data of the operating load.

[0176] Solve the optimized scheduling model to obtain the distribution network scheduling strategy for each scheduling period.

[0177] The objective function of the optimized scheduling model is expressed as follows:

[0178] ;

[0179] ;

[0180] in, The total operating cost of the distribution network. , , and These are thermal power generation, penalties for abandoning renewable energy sources, carbon emission costs, and revenue from hydrogen sales. and The first Thermal power units at each power grid node The cost and amount of electricity generated at any given time. and The first Each thermal power unit Start-up and shutdown costs at any given moment; The cost of penalizing units that abandon new energy sources and The first At the first power grid node The predicted maximum power generation and actual power generation of all wind turbines in a multi-energy complementary cluster. and The first At the first power grid node All photovoltaic units in a multi-energy complementary cluster The predicted maximum power generation and the actual power generation at any given time; Penalty costs for carbon emission units For the first Carbon emission factors of thermal power units at each power grid node For the first Thermal power units at each power grid node Power generation at any given moment For the price of hydrogen, For the first At the first power grid node Hydrogen energy systems in multi-energy complementary clusters Hydrogen production at any given moment.

[0181] The constraints of the optimized scheduling model include the operational constraints of each device in the previously established multi-energy complementary cluster, such as wind turbines, photovoltaic units, and hydrogen energy systems, as well as the operational constraints of thermal power units, power flow constraints of the distribution network, and voltage constraints.

[0182] The constructed optimization scheduling model can be directly solved using MATLAB software and the Gurobi solver to obtain the distribution network scheduling strategy for each scheduling period. The distribution network scheduling strategy is the overall control objective of each multi-energy complementary cluster, which mainly includes the upper and lower limits of the cluster's total output in each scheduling period, the renewable energy absorption rate requirement, and the power collaborative allocation ratio with conventional thermal power units.

[0183] Since each multi-energy complementary cluster is formed by a pair of wind power clusters and photovoltaic clusters, and the two wind and solar clusters have different standardized characteristic parameters, the overall control target cannot be evenly distributed. It is necessary to combine these parameters to perform differentiated decomposition in order to obtain the time-period control target of each wind power cluster and photovoltaic cluster in each scheduling period, determine the output level and adjustment direction that it needs to achieve in the corresponding time period, improve the degree of scheduling refinement, and ensure the scheduling effect.

[0184] The determination of the output plans for wind power clusters and photovoltaic clusters and the operation strategy for the hydrogen energy system based on the distribution network dispatch strategy includes:

[0185] The control targets of each multi-energy complementary cluster are obtained based on the distribution network dispatch strategy, and the control targets are decomposed according to the cluster standardized characteristic parameters of the corresponding wind power cluster and photovoltaic cluster to obtain the time-period control targets of the corresponding wind power cluster and photovoltaic cluster.

[0186] Predict the available power generation of the corresponding wind power cluster and photovoltaic cluster during the specified time period. Combine the wind-solar complementarity index of the corresponding wind power cluster and photovoltaic cluster with the operating parameters and operating constraints of the hydrogen energy system to allocate the time period control targets, and obtain the output plan of the corresponding wind power cluster and photovoltaic cluster and the operation strategy of the hydrogen energy system for each multi-energy complementary cluster.

[0187] When decomposing the control targets, the basic output share of each wind power cluster and photovoltaic cluster is determined based on the standardized installed capacity ratio of each multi-energy complementary cluster. The cluster with a higher installed capacity ratio undertakes more basic output tasks, and the lower limit of the cluster's new energy absorption rate is directly linked to the two clusters to constrain the minimum requirements that the combined absorption rate of the two must meet. Finally, a time-based control target is formed for each cluster, which includes a basic output target, an adjustable power range, and an absorption rate constraint.

[0188] By combining the standardized typical power output curves of each cluster with real-time meteorological forecast data, the maximum power output of wind power clusters and photovoltaic clusters in each time period is calculated and the error range is marked as the boundary condition for target allocation. Then, the wind-solar complementarity index of the corresponding cluster is retrieved. Based on the time period matching characteristics corresponding to the state-type complementarity, the clusters in the current time period with the power output advantage range are identified. Their power output targets are prioritized to maximize the utilization of natural complementary characteristics and reduce the dependence on the regulation of the hydrogen energy system.

[0189] Then, following the principle of prioritizing wind and solar power consumption and using hydrogen energy for auxiliary regulation, dynamic allocation and constraint verification are carried out in stages. If the total renewable power of wind and solar power is not lower than the overall control target, the target is first allocated to the advantageous clusters according to priority until the upper limit of renewable power is reached. The remaining target is allocated to another cluster. If there is still a power surplus, the rated power of the hydrogen energy system's electrolyzer and the capacity of the hydrogen storage tank are verified. If hydrogen can be stored, the hydrogen production mode is activated. If the hydrogen storage tank is full, hydrogen is sold first. The remaining power that cannot be consumed is then treated as curtailment. If the total renewable power of wind and solar power is lower than the overall control target and the gap does not exceed the rated power of the fuel cell, the wind and solar clusters output at their maximum renewable power. The gap is made up by the fuel cell power generation of the hydrogen energy system. At the same time, the lower limit of the hydrogen storage tank capacity is verified. If the hydrogen storage is insufficient, the output target of the wind and solar clusters is lowered to balance the power. If the total renewable power of wind and solar power is 0, all control targets are borne by the hydrogen energy system, and the rated power of the fuel cell and the hydrogen storage constraints are verified simultaneously. Finally, the output plans for each time period of the wind power cluster and photovoltaic cluster are finalized, the corresponding operation mode and related power parameters of the hydrogen energy system are determined, a complete time period execution plan for each multi-energy complementary cluster is formed, and the output plans of the corresponding wind power cluster and photovoltaic cluster and the operation strategy of the hydrogen energy system are determined.

[0190] The established output plans and operational strategies are issued according to the scheduling periods, and the issuance process follows hierarchical transmission, forming a hierarchical execution link of the distribution network scheduling layer, multi-energy complementary cluster layer, and unit layer. This ensures that the goals at each level are accurately received, facilitates real-time monitoring and tracing of equipment operating status, timely detection and correction of operational behaviors that deviate from the plan, and ensures the ultimate achievement of scheduling goals such as increasing the renewable energy consumption rate, reducing operating costs, and reducing carbon emissions, thereby ensuring the safe and stable operation of the distribution network.

[0191] The step of issuing power generation plans and operating strategies to the corresponding wind power clusters, photovoltaic clusters, and hydrogen energy systems according to the corresponding scheduling period, and controlling the operation of each piece of equipment in the corresponding wind turbines, photovoltaic units, and hydrogen energy systems, includes:

[0192] Based on the corresponding scheduling period, the power generation plan and operation strategy are sent to the corresponding wind power cluster, photovoltaic cluster and hydrogen energy system, and the corresponding unit operation instructions are formulated in combination with the real-time operation status of the wind power cluster, photovoltaic cluster and hydrogen energy system;

[0193] Each piece of equipment in the corresponding wind turbine, photovoltaic unit, and hydrogen energy system responds to the corresponding unit's operating command and adjusts its corresponding operating status.

[0194] As source-side devices, wind power clusters, photovoltaic clusters, and hydrogen energy systems, in addition to receiving cloud-side plans and parameters, further perceive the real-time operating status of wind farms, photovoltaic power plants, electrolysis hydrogen production stations, and fuel cell stations within the region. They perform online verification of equivalent ramp-up and regulation capacity and execute multi-energy collaborative rolling coordination. When wind and solar output is too high, excess electricity is absorbed by increasing electrolysis hydrogen production power or reducing fuel cell output. When wind and solar output is too low, compensation is achieved by increasing fuel cell output or reducing electrolysis load, thus achieving dynamic matching between wind and solar output and hydrogen storage and release within the cluster. Simultaneously, the cluster-level scheduling strategy is further refined into active and reactive power setpoints for the stations and equipment adjustment ranges, and the end-side autonomous control parameters are updated based on actual performance. Even in the event of cloud-edge communication anomalies, regional autonomous operation can still be achieved based on the last received valid plan and local measurements.

[0195] As end-side devices, wind turbines, photovoltaic units, and hydrogen energy systems can drive corresponding physical equipment to perform active and reactive power regulation and start-up / stop operations based on the received unit operation commands. They can also report the execution results and status information in real time, enabling rapid autonomous response and coordinated control of various resources such as wind power, photovoltaics, electrolytic hydrogen production, hydrogen storage, and fuel cells.

[0196] Another aspect of this embodiment provides a multi-energy complementary control system for wind, solar, and hydrogen power distribution networks based on cloud-edge-device collaboration, including:

[0197] The dispatch center layer is used to screen complementary pairings of wind power clusters and photovoltaic clusters, and to form multi-energy complementary clusters by combining hydrogen energy systems. At the same time, it constructs an optimized dispatch model by combining conventional units, operating loads and distribution network operating constraints, and solves the optimized dispatch model to obtain the distribution network dispatch strategy for each dispatch period.

[0198] The multi-energy complementary cluster layer responds to the distribution network dispatch strategy of the dispatch center and formulates corresponding unit operation instructions based on the real-time operating status of wind power clusters, photovoltaic clusters and hydrogen energy systems.

[0199] The unit control layer, including wind turbines, photovoltaic units, and hydrogen energy systems, adjusts its corresponding operating status in response to the unit operation commands from the multi-energy complementary cluster layer.

[0200] The dispatch center layer is cloud-based and is mainly used to manage the dispatch and allocation of various energy sources within the distribution network area, forming a system-level dispatch strategy.

[0201] The multi-energy complementary cluster layer is a side layer, including multiple multi-energy complementary clusters formed by wind power, photovoltaic and hydrogen energy. It can respond to the system-level scheduling strategy of the upper level, perceive the operating status of wind farms, photovoltaic power stations, electrolysis hydrogen production stations and fuel cell stations in the region in real time, perform online verification of equivalent ramp-up and regulation capacity, and perform multi-energy collaborative rolling coordination.

[0202] The unit control layer is the end side, which includes multiple wind turbines, photovoltaic units and hydrogen energy equipment. It can respond to instructions issued by the superior unit, complete the power distribution, voltage and power factor control and protection functions within the station, and track the superior unit settings when the instructions are normal. When the frequency or voltage exceeds the limit or communication is interrupted, it automatically adjusts the output according to the preset autonomous strategy, giving priority to the safety of the local area and the grid connection point.

[0203] The dispatch center layer, multi-energy complementary cluster layer, and unit control layer together constitute the cloud-edge-device system for distribution network dispatching. Based on global optimization, it can achieve regional coordination and local autonomy, ensuring the efficiency and effectiveness of distribution network dispatching.

[0204] The cloud-edge architecture of the multi-energy complementary control system is as follows: Figure 2 As shown.

[0205] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any way. Other variations and modifications are possible without departing from the technical solutions described in the claims.

Claims

1. A method for multi-energy complementary regulation of wind, solar, and hydrogen power distribution networks based on cloud-edge-device collaboration, characterized in that, include: Based on power output characteristics and regulation capabilities, wind turbine units and photovoltaic units are clustered separately, and wind power clusters and photovoltaic clusters are obtained by aggregating the clustering results. Based on the wind-solar complementarity index, with the goal of maximizing the total wind-solar complementarity, wind power clusters and photovoltaic clusters are paired to obtain the optimal combination scheme. The optimal combination scheme is used to form a wind-solar complementary aggregate, and combined with the hydrogen energy system to form a multi-energy complementary cluster; Based on multi-energy complementary clusters, an optimal scheduling model is constructed by combining conventional units, operating loads and distribution network operation constraints. The optimal scheduling model is solved to obtain the distribution network scheduling strategy for each scheduling period. The power output plans for wind power clusters and photovoltaic clusters, as well as the operation strategies for hydrogen energy systems, are determined based on the power distribution network dispatch strategy. According to the corresponding scheduling period, the power generation plan and operation strategy are sent to the corresponding wind power cluster, photovoltaic cluster and hydrogen energy system to control the operation of each equipment in the corresponding wind turbine, photovoltaic unit and hydrogen energy system; The calculation process for the wind-solar complementarity index includes: Obtain the corresponding power output sequence based on the cluster standardization parameters labeled by wind power clusters and photovoltaic clusters; The correlation coefficient between the wind power cluster and the photovoltaic cluster is calculated based on the corresponding power output sequence. The correlation coefficient is then mapped to a predetermined interval to obtain the corresponding correlation complementarity index. Based on the corresponding power output sequence and combined with the preset cluster power output state threshold, the state-type complementarity index between wind power clusters and photovoltaic clusters is calculated through the indicator function. The correlation-type complementarity index and the state-type complementarity index are weighted and summed according to the preset weights to obtain the wind-solar complementarity index value between the corresponding wind power cluster and photovoltaic cluster.

2. The method for multi-energy complementary regulation of wind, solar, and hydrogen power distribution networks based on cloud-edge-device collaboration according to claim 1, characterized in that, The clustering of wind turbines and photovoltaic units based on output characteristics and regulation capabilities includes: Historical power generation curves of each wind turbine and photovoltaic unit were collected. Based on the historical power generation curves, corresponding output characteristics were extracted. The original dataset was constructed by combining the corresponding installed capacity and regulation capacity. The original dataset was then preprocessed. Several distinct data points are randomly selected from the preprocessed dataset as initial cluster centers. The DTW distance between each data point in the dataset and each initial cluster center is calculated based on the corresponding power output characteristics. The corresponding Euclidean distance is calculated based on the installed capacity and regulation capability. The corresponding DTW distance and Euclidean distance are weighted and summed to obtain the corresponding comprehensive distance. Based on the corresponding comprehensive distance, each data point in the dataset is assigned to the cluster corresponding to the nearest initial cluster center; Calculate the cumulative distance within each cluster for each unit, update the cluster center based on the cumulative distance, recalculate the comprehensive distance between each data point in the dataset and each cluster center, and allocate the data points until the iteration termination condition is met.

3. The method for multi-energy complementary regulation of wind, solar, and hydrogen power distribution networks based on cloud-edge-device collaboration according to claim 1, characterized in that, The process of aggregating wind power clusters and photovoltaic clusters based on clustering results includes: Based on the clustering results, all clusters are obtained. All wind turbine units belonging to the same cluster are merged into an initial wind power cluster, and all photovoltaic units belonging to the same cluster are merged into an initial photovoltaic cluster. The initial wind power cluster and the initial photovoltaic cluster are subjected to consistency verification. Based on the preset threshold, the initial wind power cluster and the initial photovoltaic cluster are split or merged respectively to obtain the final wind power cluster and photovoltaic cluster. Based on the historical output curves, regulation capacity, and installed capacity of all units within the wind power cluster and photovoltaic cluster, corresponding cluster-standardized feature parameters are extracted, and annotations are added to each wind power cluster and photovoltaic cluster according to the corresponding cluster-standardized feature parameters.

4. The method for multi-energy complementary regulation of wind, solar, and hydrogen power distribution networks based on cloud-edge-device collaboration according to claim 1, characterized in that, The method, based on the wind-solar complementarity index and aiming to maximize the total wind-solar complementarity, involves complementary pairing of wind power clusters and photovoltaic clusters to obtain the optimal combination scheme, including: Calculate the wind-solar complementarity index for each pair of wind power clusters and photovoltaic clusters, and construct a wind-solar complementarity matrix. Based on the wind-solar complementarity matrix, with the optimization objective of maximizing the total wind-solar complementarity of all paired clusters and the constraint of one-to-one cluster pairing, an integer programming pairing model is constructed between wind power clusters and photovoltaic clusters. Solve the integer programming pairing model to obtain the optimal photovoltaic cluster corresponding to each wind power cluster, and form the optimal combination scheme of wind power cluster and photovoltaic cluster.

5. The method for multi-energy complementary regulation of wind, solar, and hydrogen power distribution networks based on cloud-edge-device collaboration according to claim 1, characterized in that, The formation of a wind-solar complementary aggregate based on the optimal combination scheme, and the systematic formation of a multi-energy complementary cluster by combining hydrogen energy, includes: Based on each wind power cluster and photovoltaic cluster combination in the optimal combination scheme, a corresponding wind-solar hybrid aggregate is formed, and the equivalent parameters and operating constraints of each wind-solar hybrid aggregate are calculated based on all units in the corresponding wind power cluster and photovoltaic cluster. Based on the equipment type of the hydrogen energy system, corresponding operating parameters and constraints are set. Combined with the equivalent parameters and operating constraints of each wind-solar hybrid agglomeration, the hydrogen energy system and the wind-solar hybrid agglomeration are integrated and modeled to form a multi-energy complementary cluster.

6. The method for multi-energy complementary regulation of wind, solar, and hydrogen power distribution networks based on cloud-edge-device collaboration according to claim 5, characterized in that, The optimization scheduling model, based on multi-energy complementary clusters and combined with conventional units, operating loads, and distribution network operating constraints, is constructed. Solving the optimization scheduling model yields the distribution network scheduling strategy for each scheduling period, including: The optimization objective is to minimize the sum of operating costs, renewable energy abandonment penalty costs, and carbon emission costs, while maximizing hydrogen sales revenue. The constraints are the operation constraints of the distribution network, conventional units, and multi-energy complementary clusters. An optimization scheduling model is constructed based on the equivalent parameters of the multi-energy complementary clusters, the operating parameters of conventional units, and the load data of the operating load. Solve the optimized scheduling model to obtain the distribution network scheduling strategy for each scheduling period.

7. The method for multi-energy complementary regulation of wind, solar, and hydrogen power distribution networks based on cloud-edge-device collaboration according to claim 1, characterized in that, The method for determining the output plans of wind power clusters and photovoltaic clusters and the operation strategies of hydrogen energy systems based on distribution network dispatching strategies includes: The control targets of each multi-energy complementary cluster are obtained based on the distribution network dispatch strategy, and the control targets are decomposed according to the cluster standardized characteristic parameters of the corresponding wind power cluster and photovoltaic cluster to obtain the time-period control targets of the corresponding wind power cluster and photovoltaic cluster. Predict the available power generation of the corresponding wind power cluster and photovoltaic cluster during the specified time period. Combine the wind-solar complementarity index of the corresponding wind power cluster and photovoltaic cluster with the operating parameters and operating constraints of the hydrogen energy system to allocate the time period control targets, and obtain the output plan of the corresponding wind power cluster and photovoltaic cluster and the operation strategy of the hydrogen energy system for each multi-energy complementary cluster.

8. The method for multi-energy complementary regulation of wind, solar, and hydrogen power distribution networks based on cloud-edge-device collaboration according to claim 1, characterized in that, The process of sending power generation plans and operating strategies to the corresponding wind power clusters, photovoltaic clusters, and hydrogen energy systems according to the corresponding scheduling period, and controlling the operation of each piece of equipment in the corresponding wind turbines, photovoltaic units, and hydrogen energy systems, includes: Based on the corresponding scheduling period, the power generation plan and operation strategy are sent to the corresponding wind power cluster, photovoltaic cluster and hydrogen energy system, and the corresponding unit operation instructions are formulated in combination with the real-time operation status of the wind power cluster, photovoltaic cluster and hydrogen energy system; Each piece of equipment in the corresponding wind turbine, photovoltaic unit, and hydrogen energy system responds to the corresponding unit's operating command and adjusts its corresponding operating status.

9. A multi-energy complementary regulation system for power distribution networks based on cloud-edge-device collaboration, used to execute the multi-energy complementary regulation method according to any one of claims 1 to 8, characterized in that, include: The dispatch center layer is used to screen complementary pairings of wind power clusters and photovoltaic clusters, and to form multi-energy complementary clusters by combining hydrogen energy systems. At the same time, it constructs an optimized dispatch model by combining conventional units, operating loads and distribution network operating constraints, and solves the optimized dispatch model to obtain the distribution network dispatch strategy for each dispatch period. The multi-energy complementary cluster layer responds to the distribution network dispatch strategy of the dispatch center and formulates corresponding unit operation instructions based on the real-time operating status of wind power clusters, photovoltaic clusters and hydrogen energy systems. The unit control layer, including wind turbines, photovoltaic units, and hydrogen energy systems, adjusts its corresponding operating status in response to the unit operation commands from the multi-energy complementary cluster layer.

Images