High-proportion renewable energy power generation day-ahead regional transaction clearing model construction method

Abstract

The invention relates to a day-ahead regional transaction clearing model construction method for high-proportion renewable energy power generation. The method comprises the following steps: predicting an electric energy conversion condition according to multi-dimensional power grid operation data of an electric power resource interaction region to obtain an electric energy conversion prediction feature data set; performing clustering processing on each electric energy conversion prediction feature data point in the electric energy conversion prediction feature data set to obtain prediction feature uncertainty information of each electric energy conversion prediction feature data point; calculating clearing scene probability information in different clearing typical distribution scenes according to each electric energy conversion prediction feature data point; and according to the prediction feature uncertainty information, adding each clearing typical distribution scene and each clearing scene probability information to a target clearing function of the power resource interaction region to obtain a day-ahead region transaction clearing model. By adopting the method, the characteristics of high-proportion renewable energy power generation and load can be accurately described, and the economical efficiency of a regional power grid is improved.

Description

Technical Field

[0001] The present application relates to the technical field of electricity spot market, and in particular to a method for constructing a day-ahead regional transaction clearing model for power generation with a high proportion of renewable energy. Background Art

[0002] With the development of electricity spot market technology, a unified multi-level electricity market system has been initially formed, and the idea of ​​regional joint clearing has further expanded the scope of optimal allocation of electricity resources. Under the "dual carbon" goal, the gradual increase in the proportion of renewable energy has increased the uncertainty of the power supply and load ends in the new power system, and also increased the difficulty of regional division of the power market, bringing challenges to the security and stability of the power market. At present, the market-oriented mechanism of electricity prices has been formed, and the decisive role of the market in the optimal allocation of resources has gradually emerged.

[0003] In traditional technology, although the prediction accuracy of wind power and photovoltaic power generation has been improved to a certain extent, in regional electricity market transactions, it is still unable to solve the problem that small prediction deviations will also lead to large actual power deviations, resulting in the inability to accurately describe the characteristics of high-proportion renewable energy generation and load, and improve the economic efficiency of regional power grids. Summary of the invention

[0004] Based on this, it is necessary to provide a method, device, computer equipment, computer-readable storage medium and computer program product for constructing a day-ahead regional transaction clearing model for high-proportion renewable energy power generation that can accurately describe the characteristics of high-proportion renewable energy power generation and load and improve the economic efficiency of the regional power grid in response to the above-mentioned technical problems.

[0005] In a first aspect, the present application provides a method for constructing a day-ahead regional transaction clearing model for high-proportion renewable energy generation. The method comprises:

[0006] According to the multi-dimensional power grid operation data of the power resource interaction area, the power conversion situation of the power resource interaction area is predicted to obtain a power conversion prediction feature data set;

[0007] Performing clustering processing on each electric energy conversion prediction feature data point in the electric energy conversion prediction feature data set to obtain prediction feature uncertainty information of each electric energy conversion prediction feature data point;

[0008] Calculating the clearing scenario probability information of the power resource interaction area under different clearing typical distribution scenarios according to each of the electric energy conversion prediction characteristic data points;

[0009] According to the prediction characteristic uncertainty information, each of the typical distribution scenarios of clearing and the probability information of each of the clearing scenarios are added to the target clearing function of the power resource interaction area to obtain the day-ahead regional transaction clearing model of the power resource interaction area.

[0010] In a second aspect, the present application also provides a device for constructing a day-ahead regional transaction clearing model for high-proportion renewable energy generation. The device comprises:

[0011] A prediction data acquisition module is used to predict the power conversion situation of the power resource interaction area according to the multi-dimensional power grid operation data of the power resource interaction area, and obtain a power conversion prediction feature data set;

[0012] A data point clustering module, used for performing clustering processing on each electric energy conversion prediction feature data point in the electric energy conversion prediction feature data set to obtain prediction feature uncertainty information of each electric energy conversion prediction feature data point;

[0013] A probability information calculation module, used to calculate the clearing scenario probability information of the power resource interaction area under different clearing typical distribution scenarios according to each of the power conversion prediction characteristic data points;

[0014] A clearing model obtaining module is used to add each of the typical distribution scenarios of clearing and the probability information of each of the clearing scenarios to the target clearing function of the power resource interaction area according to the prediction feature uncertainty information, so as to obtain the day-ahead regional transaction clearing model of the power resource interaction area.

[0015] In a third aspect, the present application further provides a computer device. The computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the following steps when executing the computer program:

[0016] According to the multi-dimensional power grid operation data of the power resource interaction area, the power conversion situation of the power resource interaction area is predicted to obtain a power conversion prediction feature data set;

[0017] Performing clustering processing on each electric energy conversion prediction feature data point in the electric energy conversion prediction feature data set to obtain prediction feature uncertainty information of each electric energy conversion prediction feature data point;

[0018] Calculating the clearing scenario probability information of the power resource interaction area under different clearing typical distribution scenarios according to each of the electric energy conversion prediction characteristic data points;

[0019] According to the prediction characteristic uncertainty information, each of the typical distribution scenarios of clearing and the probability information of each of the clearing scenarios are added to the target clearing function of the power resource interaction area to obtain the day-ahead regional transaction clearing model of the power resource interaction area.

[0020] In a fourth aspect, the present application further provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0021] According to the multi-dimensional power grid operation data of the power resource interaction area, the power conversion situation of the power resource interaction area is predicted to obtain a power conversion prediction feature data set;

[0022] Performing clustering processing on each electric energy conversion prediction feature data point in the electric energy conversion prediction feature data set to obtain prediction feature uncertainty information of each electric energy conversion prediction feature data point;

[0023] Calculating the clearing scenario probability information of the power resource interaction area under different clearing typical distribution scenarios according to each of the electric energy conversion prediction characteristic data points;

[0024] According to the prediction characteristic uncertainty information, each of the typical distribution scenarios of clearing and the probability information of each of the clearing scenarios are added to the target clearing function of the power resource interaction area to obtain the day-ahead regional transaction clearing model of the power resource interaction area.

[0025] In a fifth aspect, the present application further provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0026] According to the multi-dimensional power grid operation data of the power resource interaction area, the power conversion situation of the power resource interaction area is predicted to obtain a power conversion prediction feature data set;

[0027] Performing clustering processing on each electric energy conversion prediction feature data point in the electric energy conversion prediction feature data set to obtain prediction feature uncertainty information of each electric energy conversion prediction feature data point;

[0028] Calculating the clearing scenario probability information of the power resource interaction area under different clearing typical distribution scenarios according to each of the electric energy conversion prediction characteristic data points;

[0029] According to the prediction characteristic uncertainty information, each of the typical distribution scenarios of clearing and the probability information of each of the clearing scenarios are added to the target clearing function of the power resource interaction area to obtain the day-ahead regional transaction clearing model of the power resource interaction area.

[0030] The above-mentioned method, device, computer equipment, storage medium and computer program product for constructing a day-ahead regional transaction clearing model for high-proportion renewable energy power generation predicts the electric energy conversion situation in the electric energy resource interaction area based on the multi-dimensional power grid operation data of the electric energy resource interaction area to obtain an electric energy conversion prediction feature data set; clusters each electric energy conversion prediction feature data point in the electric energy conversion prediction feature data set to obtain prediction feature uncertainty information of each electric energy conversion prediction feature data point; calculates the clearing scenario probability information of the electric energy resource interaction area under different clearing typical distribution scenarios based on each electric energy conversion prediction feature data point; and adds each clearing typical distribution scenario and each clearing scenario probability information to the target clearing function of the electric energy resource interaction area based on the prediction feature uncertainty information to obtain a day-ahead regional transaction clearing model for the electric energy resource interaction area.

[0031] Through in-depth analysis of the multi-dimensional operating data of the power resource interaction area, we first make an accurate prediction of the power conversion situation, and then use clustering processing to explore the intrinsic connection between the prediction feature data points, revealing the uncertainty of the prediction characteristics of each data point, thereby improving the depth and breadth of data processing. Further, by calculating the probability information of the clearing scenario and combining it with the typical distribution scenario, we organically integrate uncertainty and probability information, optimize the target clearing function, and enable the model to adaptively respond to different power grid operation scenarios. Ultimately, the model significantly improves the accuracy and stability of the day-ahead regional transaction clearing, can accurately describe the characteristics of high-proportion renewable energy generation and load, and improve the economy of the regional power grid; at the same time, it enhances the system's ability to respond to abnormal fluctuations, ensures the efficiency and safety of power grid operation, reduces the potential risks of market transactions, and has broad application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] FIG1 is an application environment diagram of a method for constructing a day-ahead regional transaction clearing model for high-proportion renewable energy generation in one embodiment;

[0033] FIG2 is a flow chart of a method for constructing a day-ahead regional transaction clearing model for high-proportion renewable energy generation in one embodiment;

[0034] FIG3 is a schematic flow chart of a method for obtaining a power conversion prediction feature data set in one embodiment;

[0035] FIG4 is a flow chart of a method for obtaining a day-ahead regional transaction clearing model in one embodiment;

[0036] FIG5 is a schematic flow chart of a method for determining power clearing constraint conditions in one embodiment;

[0037] FIG6 is a schematic diagram of a flow chart of a method for obtaining prediction feature uncertainty information in one embodiment;

[0038] FIG7 is a schematic flow chart of a method for obtaining prediction feature uncertainty information in another embodiment;

[0039] FIG8 is a structural block diagram of a device for constructing a day-ahead regional transaction clearing model for high-proportion renewable energy generation in one embodiment;

[0040] FIG. 9 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0042] A method for constructing a day-ahead regional transaction clearing model for high-proportion renewable energy generation provided in an embodiment of the present application can be applied in an application environment as shown in FIG1. ​​Among them, the terminal 102 communicates with the server 104 through a network. The data storage system can store data that the server 104 needs to process. The data storage system can be integrated on the server 104, or it can be placed on a cloud or other network server. The server 104 obtains multi-dimensional power grid operation data from the terminal 102, and further predicts the power conversion situation of the power resource interaction area based on the multi-dimensional power grid operation data of the power resource interaction area to obtain a power conversion prediction feature data set; clustering processing is performed on each power conversion prediction feature data point in the power conversion prediction feature data set to obtain prediction feature uncertainty information of each power conversion prediction feature data point; according to each power conversion prediction feature data point, the probability information of the clearing scenario of the power resource interaction area under different typical distribution scenarios of clearing is calculated; according to the prediction feature uncertainty information, each typical distribution scenario of clearing and each clearing scenario probability information are added to the target clearing function of the power resource interaction area to obtain a day-ahead regional transaction clearing model for the power resource interaction area. The terminal 102 may be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, IoT devices, and portable wearable devices. The IoT devices may be smart speakers, smart TVs, smart air conditioners, smart car-mounted devices, etc. The portable wearable devices may be smart watches, smart bracelets, head-mounted devices, etc. The server 104 may be implemented as an independent server or a server cluster consisting of multiple servers.

[0043] In one embodiment, as shown in FIG2 , a method for constructing a day-ahead regional transaction clearing model for high-proportion renewable energy generation is provided, and the method is applied to the server in FIG1 as an example for explanation, including the following steps:

[0044] Step 202 , predicting the power conversion situation in the power resource interaction area based on the multi-dimensional power grid operation data in the power resource interaction area, and obtaining a power conversion prediction feature data set.

[0045] The power resource interaction area can be a specific area in the power grid where power resources can be transmitted and distributed to each other (clearing), usually including the power generation end, the load end and the transmission and distribution network. This area achieves effective supply and demand balance of electric energy through power grid operation, dispatching and market transactions.

[0046] Among them, the multi-dimensional power grid operation data can be a collection of various data related to the power grid operation, including load data, power generation data, meteorological data, market prices, equipment status, etc.

[0047] Among them, the power conversion prediction feature data set can be a data set obtained by analyzing the power grid operation data and combining the power conversion model to predict the future power conversion situation. This data set usually contains multiple characteristic variables, such as load changes, power generation, energy loss, etc., which characterize the different characteristics of energy conversion in the future power system.

[0048] Specifically, when realizing the prediction of power conversion, the multi-dimensional power grid operation data of the power resource interaction area is comprehensively collected and preprocessed, including historical load, power generation plan, renewable energy power generation, meteorological data (such as wind speed, temperature, sunshine), etc. Time series analysis and deep learning models (such as LSTM, GRU, Transformer) are used to predict the power conversion of these data, capture the time dependency and nonlinear characteristics in the data, predict the load and generation of various types of power in the future, and generate a power conversion prediction feature data set. For example, a long short-term memory neural network is used to make multiple predictions of the power grid load and renewable energy power generation (wind power generation and photovoltaic power generation), generating a data set P = [P wind , P PV , P load ].

[0049] Step 204 , clustering is performed on each electric energy conversion prediction feature data point in the electric energy conversion prediction feature data set to obtain prediction feature uncertainty information of each electric energy conversion prediction feature data point.

[0050] Among them, the power conversion prediction feature data point can be an individual data item in the power conversion prediction feature data set, representing a specific power feature value at a certain moment or under a certain condition. These data points are usually generated based on the power grid operation status and prediction model, reflecting the specific situation of power conversion, such as the power supply and demand at a certain moment.

[0051] Among them, the prediction feature uncertainty information can be data describing the uncertainty factors in the power conversion prediction, reflecting the errors and fluctuations that may occur in the prediction process.

[0052] Specifically, clustering algorithms (such as K-means, hierarchical clustering, or density-based clustering algorithm DBSCAN) are applied to perform cluster analysis on each power conversion prediction feature data point in the generated power conversion prediction feature data set. The clustering process discovers the intrinsic pattern of the power conversion feature points by evaluating the similarity between each power conversion prediction feature data point. In order to ensure that the clustering results have good interpretability and application value, the clustering quality is evaluated using indicators such as the silhouette coefficient. The clustering results can identify different power conversion modes and their fluctuation ranges, and further quantify the prediction feature uncertainty information of each power conversion prediction feature data point, such as measuring the deviation or uncertainty of power conversion prediction in different scenarios through statistics such as standard deviation and variance.

[0053] For clustering algorithms, the ant colony sorting clustering algorithm can be used to obtain the uncertainty of the day-ahead forecast. The ant colony algorithm is an algorithm that simulates the path of ants to find the optimal path. When ants look for food sources, they release pheromones along their own paths and sense the pheromones of other ants. The initial path selection of ants is random, and they are more likely to choose a path with a high pheromone concentration while walking and leave the pheromone behind. Therefore, the higher the pheromone concentration on the path, the shorter the path distance. Faced with a large number of disordered prediction data points, the ant colony algorithm with a positive feedback mechanism and heuristic search can effectively characterize the uncertainty of the data point set and achieve the purpose of characterizing the bilateral power characteristics of the power source and load.

[0054] However, the number of clusters in the ant colony algorithm is directly given, but in the face of many data points with unknown rules, the optimal number of clusters cannot be directly determined. Therefore, the iteration of cluster number selection is increased, and the optimal number of clusters can be obtained by calculating and comparing the deviation error from each data point to the cluster center. Therefore, an ant colony sorting clustering algorithm based on the optimal number of clusters is proposed.

[0055] Assume that the number of ants is M, the number of cities is K, the pheromone matrix is ​​Ψ, and the distance matrix is ​​D. Taking the t=1 period as an example, the ants are randomly placed in K cities with W points, where W represents the number of samples. Equation (1) calculates the probability p of ant m moving from city i to j ij m (t), the probability of the path τ ij At the initial time, the pheromone concentration on all paths is the same, assuming τ ij =ε. Equation (2) calculates the heuristic function η ij, evaluates the expected degree of ant migration from city i to city j. Equation (3) calculates the path distance between cities i and j as d ij , (x, y, z) are the coordinates of each city, that is, the coordinates of the cluster center.

[0056] (1)

[0057] (2)

[0058] (3)

[0059] Among them, A m (m=1,2,…,K) is the set of cities that ant m wants to visit. m There are K-1 elements in A, except for the city that ant m starts visiting. m The number of elements in will be reduced by one. When the ants have visited all cities, they are considered to have completed a journey. α and β are the importance factors of pheromones and also the importance factors of the heuristic function. The larger the value, the greater the role in the transfer process.

[0060] After all ants complete a journey, the pheromones on each path are updated according to equation (4). In order to shorten the search time and improve the search efficiency, the weight of the pheromones released by ants with good performance is increased. All ants are ranked according to the length of the journey. The shorter the journey length, the higher the ranking. The top r ants are considered to have better performance. Therefore, the pheromones released by the top r ants will be multiplied by the weight (Rr). In addition, the best travel path found so far is called τ bs ij , with additional weights, as shown in equation (5).

[0061] (4)

[0062] (5)

[0063] (6)

[0064] In the formula, Represents the pheromone volatility coefficient on the path.

[0065] The default number of clusters in the ant colony clustering algorithm is given in advance, but in the face of many data points with unknown rules, it is impossible to directly determine the optimal number of clusters. Therefore, the iteration of cluster number selection is added, and the optimal number of clusters Kopt can be obtained by calculating and comparing the deviation error F from each data point to the cluster center. Finally, the clustering of massive data points can be obtained. The following steps are included:

[0066] (1) Set the initial number of clusters and store the matrix F; (2) Set the initial parameters and initialize the pheromone matrix; (3) Determine and mark the ant path based on the pheromone matrix; (4) Obtain each cluster center based on the path identification; (5) Calculate the deviation from each point to the cluster center and store it in the matrix F; (6) Obtain the optimal path at this time based on the matrix F; (7) Continuously iterate the matrix F and update the pheromone matrix to obtain the optimal number of clusters and find the optimal path; (8) Output the optimal number of clusters and clustering results.

[0067] Step 206, based on each electric energy conversion prediction characteristic data point, calculate the clearing scenario probability information of the electric power resource interaction area under different clearing typical distribution scenarios.

[0068] Among them, typical clearing distribution scenarios can be several typical market supply and demand balance scenarios extracted based on historical data and predictive analysis, such as high load, low load, and large-scale access to renewable energy.

[0069] Among them, the clearing scenario probability information can be the probability of each scenario occurring calculated based on the historical data and prediction models of each typical distribution scenario. These probability information are used to quantify the possibility of clearing in the power market under different scenarios.

[0070] Specifically, several typical distribution scenarios are determined, such as high load, low load, and fluctuating renewable energy generation scenarios, and the power conversion characteristics under each scenario are simulated and calculated. Then, based on the historical data of these typical scenarios, combined with the historical power grid clearing situation and market performance, the Monte Carlo simulation or Bayesian statistical method in the probability model of the clearing scenario is constructed to estimate the clearing scenario probability information of each power resource interaction area under different typical distribution scenarios of clearing, and further determine the market clearing probability distribution under different scenarios. For example, the optimal number of clusters obtains the typical distribution scenarios of data points and calculates the scenario probabilities, and the corresponding probabilities of these scenarios can be derived through equation (7) in the probability model of the clearing scenario.

[0071] (7)

[0072] Step 208, based on the forecast characteristic uncertainty information, each typical distribution scenario of clearing and the probability information of each clearing scenario are added to the target clearing function of the power resource interaction area to obtain the day-ahead regional transaction clearing model of the power resource interaction area.

[0073] Among them, the target clearing function can be a mathematical model used to determine the transaction clearing (supply and demand balance) state in the electricity market, combining market supply and demand, power generation constraints, load demand and other factors to find an optimal way to allocate electricity resources.

[0074] Among them, the day-ahead regional transaction clearing model can be an optimized and generated power market clearing decision model based on the power market data and forecast information of the previous day. The model combines the power conversion situation of different power grid areas, the uncertainty of forecast characteristics and the probability of clearing scenarios to determine the power trading plan of the next day's market, ensuring the stable operation of the power grid and the optimal allocation of resources.

[0075] Specifically, in the optimization process of the target clearing function, the typical distribution scenarios and the probability information of the clearing scenarios are used as input variables and integrated into the target clearing function of the power resource interaction area. In the optimization of the model that integrates the target clearing function, the target clearing function will comprehensively consider constraints such as market supply and demand balance, electricity price fluctuations, generation side constraints, and load side responses. The model that integrates the target clearing function is continuously adjusted through optimization algorithms (such as linear programming, nonlinear programming, genetic algorithms, etc.) so that the final day-ahead regional transaction clearing model can better adapt to market fluctuations under different typical scenarios. This process is solved iteratively, so that the day-ahead regional transaction clearing model can not only adapt to the existing market conditions, but also cope with potential grid operation anomalies and market changes in the future.

[0076] In the above-mentioned method for constructing a day-ahead regional transaction clearing model for high-proportion renewable energy generation, the electric energy conversion situation in the electric energy resource interaction area is predicted based on the multi-dimensional power grid operation data of the electric energy resource interaction area to obtain an electric energy conversion prediction feature data set; clustering processing is performed on each electric energy conversion prediction feature data point in the electric energy conversion prediction feature data set to obtain prediction feature uncertainty information of each electric energy conversion prediction feature data point; based on each electric energy conversion prediction feature data point, the clearing scenario probability information of the electric energy resource interaction area under different clearing typical distribution scenarios is calculated; based on the prediction feature uncertainty information, each clearing typical distribution scenario and each clearing scenario probability information are added to the target clearing function of the electric energy resource interaction area to obtain a day-ahead regional transaction clearing model for the electric energy resource interaction area.

[0077] Through in-depth analysis of the multi-dimensional operating data of the power resource interaction area, we first make an accurate prediction of the power conversion situation, and then use clustering processing to explore the intrinsic connection between the prediction feature data points, revealing the uncertainty of the prediction characteristics of each data point, thereby improving the depth and breadth of data processing. Further, by calculating the probability information of the clearing scenario and combining it with the typical distribution scenario, we organically integrate uncertainty and probability information, optimize the target clearing function, and enable the model to adaptively respond to different power grid operation scenarios. Ultimately, the model significantly improves the accuracy and stability of the day-ahead regional transaction clearing, can accurately describe the characteristics of high-proportion renewable energy generation and load, and improve the economy of the regional power grid; at the same time, it enhances the system's ability to respond to abnormal fluctuations, ensures the efficiency and safety of power grid operation, reduces the potential risks of market transactions, and has broad application prospects.

[0078] Through in-depth analysis of the multi-dimensional operating data of the power resource interaction area, we first make an accurate prediction of the power conversion situation, and then use clustering processing to explore the intrinsic connection between the prediction feature data points, revealing the uncertainty of the prediction characteristics of each data point, thereby improving the depth and breadth of data processing. Further, by calculating the probability information of the clearing scenario and combining it with the typical distribution scenario, we organically integrate uncertainty and probability information, optimize the target clearing function, and enable the model to adaptively respond to different power grid operation scenarios. Ultimately, the model significantly improves the accuracy and stability of the day-ahead regional transaction clearing, can accurately describe the characteristics of high-proportion renewable energy generation and load, and improve the economy of the regional power grid; at the same time, it enhances the system's ability to respond to abnormal fluctuations, ensures the efficiency and safety of power grid operation, reduces the potential risks of market transactions, and has broad application prospects.

[0079] In one embodiment, as shown in FIG3 , based on the multi-dimensional grid operation data of the power resource interaction area, the power conversion situation of the power resource interaction area is predicted to obtain a power conversion prediction feature data set, including:

[0080] Step 302, input the multi-dimensional power grid operation data into the power conversion situation prediction model of the power resource interaction area to obtain a wind energy prediction data set, a solar energy prediction data set and a load prediction data set.

[0081] The wind energy prediction data set may be a set of predicted output data of wind power generation in a future period calculated by a prediction model based on meteorological data (such as wind speed, wind direction, air pressure, etc.).

[0082] The light energy prediction data set may be an analysis of meteorological data such as solar irradiance, temperature, and cloud cover, and a set of prediction models used to estimate solar power generation in future time periods.

[0083] The load forecast data set may be a set of electricity load forecast data formed by estimating future electricity demand using a forecasting model based on historical electricity consumption data, seasonal fluctuations, economic activity levels and other factors.

[0084] Specifically, preprocessed data including meteorological data such as wind speed, light intensity, temperature, the operating status of power plants, and electricity demand on the load side are input into the power conversion prediction model. The power conversion prediction model is based on deep learning or other prediction methods, such as LSTM (long short-term memory network) or a hybrid model based on physical modeling. It uses the time dependence and seasonal trends in historical data to predict the available power generation of wind and solar energy and changes in grid load demand in a certain period of time in the future, and generates wind energy prediction data sets, solar energy prediction data sets, and load prediction data sets respectively, to ensure that the prediction data has sufficient accuracy and real-time performance, which is convenient for subsequent deviation analysis and power dispatching decisions.

[0085] Step 304, calculating the day-ahead forecast deviation according to the disorder characteristics of the wind energy forecast data set, the solar energy forecast data set and the load forecast data set.

[0086] The day-ahead forecast deviation may be an error value calculated based on the difference between the forecast data and the actual observed data of wind energy, solar energy and load.

[0087] Specifically, after the wind energy forecast data set, the solar energy forecast data set and the load forecast data set are generated, the actual observation data set (such as real-time monitored wind energy, solar energy power generation and load demand) is used to compare with the above wind energy forecast data set, solar energy forecast data set and load forecast data set, and statistical analysis methods (such as mean square error, mean absolute percentage error, etc.) are used to quantify the difference between the predicted value and the actual value, simulate the random combination of wind power, photovoltaic and load forecast data in the data set, generate disordered forecast deviations, and obtain the day-ahead forecast deviations of each data set. In this process, the focus is on the randomness and uncontrollability of natural resources such as wind energy and solar energy, which are often the main source of deviations.

[0088] Step 306 , extracting a power conversion prediction feature data set of the power resource interaction area according to the day-ahead prediction deviation.

[0089] Specifically, after calculating the day-ahead forecast deviation, the day-ahead forecast deviation is converted into characteristic variables, such as the direction, size, trend, and time change of the deviation, which are used as input data for the forecast model; these characteristic variables reflect the volatility of renewable energy such as wind energy and solar energy, as well as the dynamic changes in load demand. By deeply analyzing the characteristics and distribution of these deviations, the key characteristic data set for power conversion forecasting is extracted.

[0090] In this embodiment, by inputting multi-dimensional power grid operation data into the power conversion prediction model, the future change trends of wind energy, solar energy and load are accurately predicted to form a prediction data set. Subsequently, the day-ahead prediction deviation is calculated based on the disordered characteristics of these data sets, so that the potential fluctuations and uncertainties in the actual operation of the power system can be identified and quantified. By analyzing the prediction deviation, the power conversion prediction feature data set is extracted to help power grid managers better understand the laws and potential risks of power resource conversion, thereby improving the accuracy of power grid dispatch and the reliability of power supply, and reducing the cost and uncertainty of system operation.

[0091] In one embodiment, as shown in FIG4 , according to the forecast feature uncertainty information, each typical distribution scenario of clearing and the probability information of each clearing scenario are added to the target clearing function of the power resource interaction area, and the day-ahead regional transaction clearing model of the power resource interaction area is obtained, including:

[0092] Step 402, based on the forecast feature uncertainty information and the day-ahead forecast deviation, each typical distribution scenario of clearing and the probability information of each clearing scenario are added to the target clearing function of the power resource interaction area to obtain a generalized regional transaction clearing model.

[0093] Among them, the generalized regional transaction clearing model can be an optimization model for the power market, which comprehensively considers the power supply and demand balance and the operating characteristics of the power resource interaction area based on different typical distribution scenarios and their clearing scenario probability information. The model ensures that the optimal power trading decision can be made under various market conditions by adding the prediction results and uncertainty information of multiple scenarios.

[0094] Specifically, the analysis of historical power trading data and power grid operation data constructs multiple typical distribution scenarios for clearing. These typical distribution scenarios for clearing represent the supply and demand balance of the power market under different circumstances, such as peak load periods, large-scale renewable energy access, or sudden failure scenarios. In view of the current situation, the clearing scenario probability information of each typical distribution scenario for clearing is calculated based on probabilistic statistical methods or machine learning algorithms, combined with the uncertainty information of the prediction characteristics and the day-ahead prediction deviation, to evaluate the possibility of each typical distribution scenario for clearing. Then, these typical distribution scenarios for clearing and their probabilities are added as input variables to the target clearing function of the power resource interaction area to establish a generalized regional transaction clearing model.

[0095] Step 404, determining each power clearing constraint condition based on the forecast characteristic uncertainty information, the day-ahead forecast deviation, and each typical distribution scenario of clearing.

[0096] Among them, power clearing constraints can be technical and economic constraints that must be met by the power system operation during power market transactions and resource allocation. These conditions may include the maximum and minimum output of generators, capacity limitations of transmission lines, system security constraints (such as voltage and frequency stability), and load response.

[0097] Specifically, based on the uncertainty information of forecast characteristics and the day-ahead forecast deviation, the key constraints that may be encountered during the operation of the power system are identified. These constraints may include: the start and stop and power generation capacity of the generator set, the power transmission limit of the transmission line, the response capability of the load side, the system safety constraints (such as voltage stability and frequency control), and the volatility of renewable energy. At the same time, for each typical clearing scenario, the specific constraints are refined. For example, in the scenario of large-scale wind power access, the power balance problem caused by the fluctuation of wind power generation needs to be considered to obtain the constraints for each power clearing.

[0098] Step 406, adding each power clearing constraint condition of the power resource interaction area to the generalized regional transaction clearing model to obtain the day-ahead regional transaction clearing model.

[0099] Specifically, the previously determined various power clearing constraints (such as power generation capacity restrictions, transmission network constraints, system security restrictions, etc.) are integrated into the generalized regional transaction clearing model, and solved in the model by using optimization algorithms (such as linear programming, nonlinear programming or mixed integer programming, etc.) to optimize the allocation of power resources and market clearing results under the premise of satisfying all constraints. In the case of continuous iteration and optimization of the model, it is ensured that the model can find the optimal transaction clearing solution when facing different typical scenarios, and obtain the final day-ahead regional transaction clearing model.

[0100] In this embodiment, by adding the forecast feature uncertainty information, the day-ahead forecast deviation, and the clearing scenario probability information to the target clearing function, a generalized regional transaction clearing model is constructed to ensure that the model can cope with a variety of market scenarios and uncertain factors. At the same time, by further analyzing these uncertainties and deviations, the power clearing constraints, such as power generation capacity restrictions and transmission capacity restrictions, are determined to ensure the safety and reliability of power grid operation. Finally, these constraints are integrated into the generalized regional transaction clearing model to form the final day-ahead regional transaction clearing model, making the allocation of power resources more accurate and flexible, improving market clearing efficiency, reducing the risks and costs of power system operation, and optimizing the overall operation of the power market.

[0101] In one embodiment, as shown in FIG5 , each power clearing constraint condition is determined according to the forecast characteristic uncertainty information, the day-ahead forecast deviation, and each typical distribution scenario of clearing, including:

[0102] Step 502, based on the forecast characteristic uncertainty information, the day-ahead forecast deviation and each typical distribution scenario of clearing, calculate the inter-provincial power trading network loss, inter-provincial power trading performance information, and inter-provincial power trading capacity information.

[0103] Among them, inter-provincial power trading network loss can be the power loss caused by factors such as line impedance, transmission distance and grid load during the inter-provincial power transmission process. It represents the amount of electricity that cannot be effectively utilized during the transmission of electricity from the power generation end to the power consumption end, affecting the economy and power supply efficiency of the power system.

[0104] Among them, inter-provincial power trading performance information can be the performance of each inter-provincial transmission line in power trading, including its ability to respond to load changes, recovery time to cope with faults, and operational stability under different power demand scenarios. This information is used to evaluate the operating efficiency of the power network and its performance under uncertain conditions.

[0105] Among them, inter-provincial power trading capacity information can be the maximum power transmission capacity of inter-provincial transmission lines. It indicates the maximum amount of electricity that can be safely transmitted on each line without overloading. This information is a key parameter in formulating power trading strategies to ensure the effective transmission of power between different regions.

[0106] Specifically, based on the historical data of power transmission between provinces, line parameters (such as impedance, voltage level), load distribution and geographical distance, the power flow calculation model is used to accurately estimate the power loss in different inter-provincial transmission lines, and the inter-provincial power trading network loss is obtained; this calculation takes into account the changes in power loss in different typical scenarios (such as high load or concentrated renewable energy generation scenarios) to ensure that the loss calculation is scenario-adaptive. Then, the performance information of each inter-provincial power trading line is calculated using the power grid dynamic performance evaluation method (such as sensitivity analysis or real-time monitoring data) to evaluate its response ability in dealing with load fluctuations and fault recovery. Finally, the capacity information of inter-provincial power trading is calculated by combining the physical characteristics of the line and historical trading data.

[0107] Step 504, determining power balance constraints based on inter-provincial power trading network losses.

[0108] Among them, the power balance constraint condition can be that in the power system, the power generation must be equal to the load demand plus the loss during transmission. This condition ensures that the supply and demand of the system are balanced at any time to avoid system instability caused by frequency fluctuations, power shortages or surpluses.

[0109] Specifically, based on the predicted demand at the power generation and load ends, combined with the loss information of the inter-provincial power trading network, the output of the generator set is adjusted to ensure that the power generation not only meets the actual load demand, but also compensates for the loss of electricity during transmission. At the same time, by dynamically monitoring the balance state, the supply and demand balance of the power grid in different typical distribution scenarios is ensured. Even in high-loss scenarios, the power generation side can provide sufficient power to prevent system imbalance from causing frequency fluctuations or power outages, and determine the power balance constraints.

[0110] A day-ahead regional transaction clearing model is established that takes into account the transmission costs of inter-provincial power transactions. The optimal clearing model aims to optimize provincial operating costs, including inter-provincial power purchase costs. and the cost of transmitting power Taking into account the strong uncertainty of both sources and loads and the day-ahead forecast deviation, multiple typical equilibrium scenarios are added to the objective function, making the obtained results more adaptable in most scenarios, as shown in Equation (8).

[0111] (8)

[0112] Where N is the number of generating units in the province. and is the power and reserve clearing price of generator n in the province. and is the clearing power capacity and reserve capacity of generator n in the province. J is the number of sending provinces. j is the number of generators in the sending province j. represents the clearing price of generator m in province j in interprovincial electricity trading, Indicates the corresponding clearing power capacity. represents the transmission price in province j.

[0113] The power balance constraint is considered in the constraint conditions, as shown in formula (9).

[0114] (9)

[0115] The transmission loss is taken into account in the formula. Usually, the inter-provincial power trading network loss is calculated based on the proportional function of the power transmitted by the inter-provincial power exchange.

[0116] Step 506, determining the reserve capacity constraint conditions, the upper and lower limit constraints of power generation, and the power generation climbing rate constraint conditions according to the inter-provincial power transaction performance information.

[0117] The upper and lower limits of power generation may refer to the minimum and maximum power ranges that the generator set can output during operation. This constraint ensures that the generator set operates within a safe and economical range, and prevents the unit from being damaged or having reduced efficiency due to excessive overload or underload operation.

[0118] Among them, the power generation rate constraint condition can be a regulation that limits the power adjustment speed of the generator set to ensure that the power generation will not change too fast or too slow, thereby avoiding adverse effects on the stability of the power grid due to overly drastic adjustments. This constraint ensures that the power grid can smoothly transition when the load fluctuates and maintain stable power supply.

[0119] Specifically, based on the performance information of inter-provincial power transactions, combined with the load fluctuations and historical accidents of each province, the proportion of spare power generation capacity is determined to ensure that the system has sufficient spare capacity to respond quickly in the event of an emergency or a surge in load, and to determine the spare capacity constraints. Secondly, through the performance information of inter-provincial power transactions, combined with the analysis of the physical characteristics of each generator set (such as minimum output, maximum output), fuel supply restrictions and technical specifications, to avoid the unit from operating beyond its safe operating range, and to determine the upper and lower limits of power generation constraints. Finally, based on the performance information of inter-provincial power transactions, combined with the technical characteristics and dynamic response capabilities of power generation equipment, the constraints on the power generation climbing rate are determined to ensure that the power generation adjustment speed matches the changes in grid load.

[0120] Formulas (10)-(12) are the required reserve capacity constraints, power generation upper and lower limit constraints, and power generation climbing rate constraints for the province, respectively.

[0121] (10)

[0122] (11)

[0123] (12)

[0124] Where P L,t,k represents the load at time t under scenario k, χ represents the reserve coefficient, and represents the upper and lower limits of generator n, and Limit the ramp rate of generator n.

[0125] Inter-provincial power trading should comply with inter-provincial power trading performance information. If it is considered according to the time period power curve, it will be processed by the power constraints of inter-provincial power trading. If the power of the entire optimization period is considered, it will be processed by the power constraints of inter-provincial power trading.

[0126] Step 508, determining the transmission capacity constraint conditions according to the inter-provincial power trading capacity information.

[0127] Among them, the transmission capacity constraint can be a limit on the maximum safe transmission power of the transmission line to ensure that the line will not cause overheating or other safety hazards due to overload. This constraint takes into account the physical characteristics of the line and the operation requirements of the power grid to ensure that the transmission capacity of the transmission line is within the safe range under any operating state.

[0128] Specifically, the inter-provincial power trading capacity information is analyzed, the physical characteristics of the transmission line (such as the heat capacity of the conductor, the line length, the geographical environment, etc.) are considered, and the power flow calculation results and safety standards of the power system are combined to ensure that the line will not be overheated or damaged due to power overload. In addition, the redundancy capacity of the power system and the maintenance requirements of the line must be considered to determine the transmission capacity constraints of each transmission line under different load conditions.

[0129] The present invention takes into account the time period power curve, so the transmission capacity of the tie line is constrained, as shown in formula (13). In order to ensure the adequacy of the provincial market power purchase capacity information of the inter-provincial power transaction, the sending terminal unit can only have a certain proportion of power participating in the inter-provincial transaction, as shown in formula (14).

[0130] (13)

[0131] (14)

[0132] In this embodiment, by comprehensively analyzing the uncertainty information of forecast characteristics, the day-ahead forecast deviation, and the typical distribution scenarios of clearing, the network loss, transaction performance, and transaction capacity information in the inter-provincial power transaction are accurately calculated, so as to comprehensively evaluate the operation status of the power grid. Based on network loss, the power balance constraint conditions are determined to ensure the balance of supply and demand; according to the transaction performance information, key constraints such as spare capacity, upper and lower limits of power generation, and climbing rate are set to ensure the stability of the system and the ability to respond to emergencies; according to the transaction capacity information, the transmission capacity constraint is determined to avoid line overload and improve the safety of power grid operation. Overall, this method optimizes the dispatch and resource allocation of the power system, effectively reduces the operation risk, and improves the efficiency and reliability of the power market.

[0133] In one embodiment, as shown in FIG6 , clustering is performed on each electric energy conversion prediction feature data point in the electric energy conversion prediction feature data set to obtain prediction feature uncertainty information of each electric energy conversion prediction feature data point, including:

[0134] Step 602: Use the ant colony algorithm to search for each electric energy conversion prediction feature data to obtain the pheromone concentration marking path.

[0135] The pheromone concentration marking path can be a pheromone concentration track left by ants in the ant colony algorithm when they are looking for food, which is used to indicate the correlation between feature data points. As the ants search for paths in the data set, the pheromone concentration gradually accumulates and updates, reflecting the similarity between the data points on the path.

[0136] Specifically, the ant colony algorithm is initialized to simulate multiple virtual ants searching for a suitable path in the power conversion prediction feature data set; each ant starts from a random power conversion feature data point in the algorithm, and gradually builds a path based on the similarity and distance between each feature data point. Pheromone concentration is a key parameter on the path, which indicates the quality of the path. During each iteration, the ant leaves pheromones on the path when traversing the path, and the pheromone concentration between similar data points in the path increases, prompting other ants to choose a better path. As the number of iterations increases, the pheromone concentration is gradually updated, gradually guiding the ants to find the pheromone concentration marked path, reflecting the correlation and pattern between the power conversion prediction feature data points.

[0137] Step 604, determining each cluster center according to the pheromone concentration marked path.

[0138] Specifically, in each pheromone concentration-marked path obtained, the path point with the highest pheromone concentration is selected as the candidate point of the cluster center. At this time, the higher the pheromone concentration on the path, the stronger the similarity between the characteristic data points in the path, and the more clustering potential. In order to determine the final cluster center, the distribution of each path is further analyzed, and the core data points of each cluster, i.e., the cluster center, are found by aggregating the points with the highest pheromone concentration in each path. These center points represent the characteristics of each cluster, have the greatest representativeness and similarity, and can well summarize the patterns and trends of each cluster.

[0139] Step 606, based on the differences from each path point of the pheromone concentration marked path to each cluster center, the pheromone concentration marked path is optimized to obtain prediction feature uncertainty information.

[0140] Specifically, the clustering effect of the path is evaluated by calculating the distance difference between each path point in the pheromone concentration marked path and each cluster center. If the distance difference between the path point and the cluster center is large, it means that the clustering effect of the path is not good, reflecting that the correlation between the data points is not strong and the prediction uncertainty is high. In order to improve the clustering effect, the ant colony algorithm is applied to optimize these paths, adjust the path selection and pheromone update rules, re-optimize the matching degree between the path points and the cluster center, and reduce uncertainty. Finally, the optimized path difference and the change of pheromone concentration are used to extract the prediction feature uncertainty information, which reflects the fluctuation range and error of the prediction in different clusters and helps to quantify the reliability of the model.

[0141] In this embodiment, by using the ant colony algorithm to search the power conversion prediction feature data, the pheromone concentration marking path is obtained, which can effectively explore the similarity between the feature data points. Then, the cluster center is determined by analyzing the pheromone concentration path to ensure that the cluster can accurately represent the characteristics of the data set. Furthermore, by optimizing the difference between the path point and the cluster center, the clustering error is reduced and the prediction feature uncertainty information is extracted. The beneficial effect of this method is that it improves the accuracy and efficiency of clustering, fully captures the potential patterns in the power conversion characteristics, reduces the uncertainty in the model prediction, and provides support for more accurate power system analysis and scheduling.

[0142] In one embodiment, as shown in FIG. 7 , the pheromone concentration marked path is optimized according to the differences from each path point of the pheromone concentration marked path to each cluster center to obtain the prediction feature uncertainty information, including:

[0143] Step 702, constructing a deviation matrix based on the differences from each path point of the pheromone concentration marked path to each cluster center.

[0144] The deviation matrix can be a matrix used to quantify the distance difference between each data point and its cluster center. Each element in the matrix represents the degree of deviation of a data point from a cluster center, which is usually calculated by a distance metric.

[0145] Specifically, the distance between each path point of the pheromone concentration marked path and its corresponding cluster center is calculated, and the Euclidean distance or other appropriate distance measurement method is used to evaluate the similarity between each path point and the cluster center. By summarizing these distance differences, a deviation matrix is ​​constructed, in which each element of the matrix represents the degree of deviation between a specific path point and a certain cluster center. The deviation matrix can fully reveal the degree of deviation of the path points in the cluster structure, especially identifying those path points that are far away from the cluster center, which usually represent data points with greater uncertainty.

[0146] Step 704, iteratively optimize the pheromone concentration marking path according to the deviation matrix to obtain the optimal number of clusters and the optimal marking path.

[0147] The optimal number of clusters can be the best number of clusters determined by iterative optimization in cluster analysis. This value reflects the number of clusters that can balance the similarity within clusters and the differences between clusters when clustering data. The optimal number of clusters ensures that data points can be reasonably classified, without too many subdivisions that blur the differences between clusters, or too few clusters that cause information loss.

[0148] The optimal marking path can be the path with the optimal pheromone concentration selected after iterative optimization by the ant colony algorithm. This path represents the best path that the ants searched between the data points, that is, the path that can most accurately reflect the similarity and clustering relationship between the data points. The optimal marking path helps to build a more accurate clustering structure, thereby improving the overall effect of clustering analysis.

[0149] Specifically, the high-deviation path points in the deviation matrix are analyzed, and the deviation is reduced by adjusting the classification of these path points or optimizing the location of the cluster center. The ant colony algorithm plays a role in this process. Through repeated iterations of path selection, pheromone concentration update, and cluster center adjustment, the path points gradually approach the most suitable cluster center. At the same time, according to different cluster numbers and path distributions, the number of clusters is dynamically adjusted to ensure that the number of clusters can best adapt to the complexity of data distribution. After multiple iterations, the optimization process finally converges and finds the optimal number of clusters and pheromone-marked paths.

[0150] Step 706: Generate prediction feature uncertainty information based on the optimal number of clusters and the optimal labeling path.

[0151] Specifically, after obtaining the optimal number of clusters and the optimal labeling path, the deviation between each data point of the optimal labeling path and the optimal cluster center is evaluated, with special attention paid to data points with large deviations, which usually represent areas with higher uncertainty in the prediction. By analyzing the distribution of these data points, the uncertainty information of the model's prediction features under different characteristics can be quantified, revealing the accuracy and potential risks of the model. In addition, uncertainty information can also be used to quantify the fluctuation range of the entire prediction data set through statistical methods (such as standard deviation or variance), helping decision makers understand the reliability and potential errors of the model output.

[0152] In this embodiment, a deviation matrix is ​​constructed to quantify the difference between each path point and the cluster center in the pheromone concentration marking path, and the distribution of data points and the accuracy of clustering are deeply analyzed. Then, the path is iteratively optimized according to the deviation matrix to determine the optimal number of clusters and the optimal marking path to ensure that the data points are reasonably classified and improve the clustering effect. Finally, the prediction feature uncertainty information is generated by the optimal number of clusters and the marking path, which can more accurately evaluate the uncertainty in the model. The beneficial effect of this process is that it optimizes data clustering, reduces the error of the prediction model, enhances the accuracy and reliability of power conversion prediction, and contributes to the stable operation of the power system under complex conditions.

[0153] It should be understood that, although the steps in the flowcharts involved in the above embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0154] Based on the same inventive concept, the embodiment of the present application also provides a device for constructing a regional transaction clearing model for a day-ahead of a high proportion of renewable energy power generation for realizing the method for constructing a regional transaction clearing model for a day-ahead of a high proportion of renewable energy power generation involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in the embodiments of the device for constructing a regional transaction clearing model for a day-ahead of a high proportion of renewable energy power generation provided below can be referred to the limitations of the method for constructing a regional transaction clearing model for a day-ahead of a high proportion of renewable energy power generation above, and will not be repeated here.

[0155] In one embodiment, as shown in FIG8 , a device for constructing a clearing model for day-ahead regional transactions with a high proportion of renewable energy generation is provided, comprising: a forecast data obtaining module 802, a data point clustering module 804, a probability information calculation module 806 and a clearing model obtaining module 808, wherein:

[0156] The prediction data obtaining module 802 is used to predict the power conversion situation of the power resource interaction area according to the multi-dimensional power grid operation data of the power resource interaction area, and obtain the power conversion prediction feature data set;

[0157] The data point clustering module 804 is used to perform clustering processing on each electric energy conversion prediction feature data point in the electric energy conversion prediction feature data set to obtain prediction feature uncertainty information of each electric energy conversion prediction feature data point;

[0158] The probability information calculation module 806 is used to calculate the clearing scenario probability information of the power resource interaction area under different clearing typical distribution scenarios according to each power conversion prediction characteristic data point;

[0159] The clearing model obtaining module 808 is used to add each clearing typical distribution scenario and each clearing scenario probability information to the target clearing function of the power resource interaction area according to the prediction feature uncertainty information, so as to obtain the day-ahead regional transaction clearing model of the power resource interaction area.

[0160] In one embodiment, the prediction data acquisition module 802 is also used to input the multi-dimensional power grid operation data into the power conversion situation prediction model of the power resource interaction area to obtain a wind energy prediction data set, a solar energy prediction data set and a load prediction data set; according to the disordered characteristics of the wind energy prediction data set, the solar energy prediction data set and the load prediction data set, the day-ahead prediction deviation is calculated; according to the day-ahead prediction deviation, the power conversion prediction feature data set of the power resource interaction area is extracted.

[0161] In one embodiment, the clearing model obtaining module 808 is also used to add each typical distribution scenario of clearing and each probability information of clearing scenarios to the target clearing function of the power resource interaction area according to the prediction feature uncertainty information and the day-ahead prediction deviation, so as to obtain a generalized regional transaction clearing model; determine each power clearing constraint condition according to the prediction feature uncertainty information, the day-ahead prediction deviation and each typical distribution scenario of clearing; add each power clearing constraint condition of the power resource interaction area to the generalized regional transaction clearing model, so as to obtain a day-ahead regional transaction clearing model.

[0162] In one embodiment, the clearing model obtains module 808, which is also used to calculate the inter-provincial power trading network loss, inter-provincial power trading performance information, and inter-provincial power trading capacity information based on the prediction characteristic uncertainty information, the day-ahead prediction deviation, and each typical distribution scenario of clearing; determine the power balance constraints based on the inter-provincial power trading network loss; determine the spare capacity constraints, the upper and lower limits of power generation constraints, and the power generation climbing rate constraints based on the inter-provincial power trading performance information; and determine the transmission capacity constraints based on the inter-provincial power trading capacity information.

[0163] In one embodiment, the data point clustering module 804 is also used to use an ant colony algorithm to search for each electric energy conversion prediction feature data to obtain a pheromone concentration marking path; determine each cluster center based on the pheromone concentration marking path; optimize the pheromone concentration marking path based on the differences from each path point of the pheromone concentration marking path to each cluster center to obtain prediction feature uncertainty information.

[0164] In one embodiment, the data point clustering module 804 is also used to construct a deviation matrix based on the differences from each path point of the pheromone concentration marking path to each cluster center; iteratively optimize the pheromone concentration marking path based on the deviation matrix to obtain the optimal number of clusters and the optimal marking path; and generate prediction feature uncertainty information based on the optimal number of clusters and the optimal marking path.

[0165] Each module in the above-mentioned device for constructing a day-ahead regional transaction clearing model for high-proportion renewable energy generation can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each of the above modules.

[0166] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as shown in FIG9 . The computer device includes a processor, a memory, and a network interface connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store server data. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a method for constructing a day-ahead regional transaction clearing model for high-proportion renewable energy generation is implemented.

[0167] Those skilled in the art will understand that the structure shown in FIG. 9 is merely a block diagram of a partial structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different arrangement of components.

[0168] In one embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above method embodiments when executing the computer program.

[0169] In one embodiment, a computer-readable storage medium is provided, storing a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0170] In one embodiment, a computer program product or computer program is provided, the computer program product or computer program includes computer instructions, the computer instructions are stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the steps in the above-mentioned method embodiments.

[0171] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0172] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0173] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0174] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be construed as limiting the scope of the present application. It should be noted that, for a person of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for constructing a day-ahead regional transaction clearing model for high-proportion renewable energy generation, characterized in that: The method comprises: According to the multi-dimensional power grid operation data of the power resource interaction area, the power conversion situation of the power resource interaction area is predicted to obtain a power conversion prediction feature data set; Performing clustering processing on each electric energy conversion prediction feature data point in the electric energy conversion prediction feature data set to obtain prediction feature uncertainty information of each electric energy conversion prediction feature data point; Calculating the clearing scenario probability information of the power resource interaction area under different clearing typical distribution scenarios according to each of the electric energy conversion prediction characteristic data points; According to the prediction characteristic uncertainty information, each of the typical distribution scenarios of clearing and the probability information of each of the clearing scenarios are added to the target clearing function of the power resource interaction area to obtain the day-ahead regional transaction clearing model of the power resource interaction area.

2. The method according to claim 1, characterized in that The method predicts the power conversion situation of the power resource interaction area according to the multi-dimensional power grid operation data of the power resource interaction area to obtain a power conversion prediction feature data set, including: Inputting the multi-dimensional power grid operation data into the power conversion situation prediction model of the power resource interaction area to obtain a wind energy prediction data set, a light energy prediction data set and a load prediction data set; Calculating a day-ahead forecast deviation according to the disorder characteristics of the wind energy forecast data set, the solar energy forecast data set, and the load forecast data set; According to the day-ahead prediction deviation, an electric energy conversion prediction feature data set of the electric power resource interaction area is extracted.

3. The method according to claim 2, characterized in that According to the forecast feature uncertainty information, each of the typical distribution scenarios of clearing and the probability information of each of the clearing scenarios are added to the target clearing function of the power resource interaction area to obtain the day-ahead regional transaction clearing model of the power resource interaction area, including: According to the forecast characteristic uncertainty information and the day-ahead forecast deviation, each of the typical distribution scenarios of clearing and the probability information of each of the clearing scenarios are added to the target clearing function of the power resource interaction area to obtain a generalized regional transaction clearing model; Determining each power clearing constraint condition according to the forecast characteristic uncertainty information, the day-ahead forecast deviation, and each of the typical clearing distribution scenarios; The power clearing constraints of the power resource interaction area are added to the generalized regional transaction clearing model to obtain the day-ahead regional transaction clearing model.

4. The method according to claim 3, characterized in that The power clearing constraints include power balance constraints, spare capacity constraints, upper and lower limits of power generation constraints, power generation ramp-up rate constraints, and transmission capacity constraints; the power clearing constraints are determined based on the forecast feature uncertainty information, the day-ahead forecast deviation, and each of the typical distribution scenarios of clearing, including: Calculate inter-provincial power trading network loss, inter-provincial power trading performance information, and inter-provincial power trading capacity information based on the forecast characteristic uncertainty information, the day-ahead forecast deviation, and each of the typical distribution scenarios of clearing; Determining the power balance constraint condition according to the inter-provincial power trading network loss; Determining the reserve capacity constraint, the upper and lower limit constraints of power generation, and the power generation climbing rate constraint according to the inter-provincial power transaction performance information; The transmission capacity constraint condition is determined according to the inter-provincial power trading capacity information.

5. The method according to claim 1, characterized in that The clustering process is performed on each electric energy conversion prediction feature data point in the electric energy conversion prediction feature data set to obtain prediction feature uncertainty information of each electric energy conversion prediction feature data point, including: Using an ant colony algorithm to search the electric energy conversion prediction characteristic data to obtain a pheromone concentration marking path; Determining each cluster center according to the pheromone concentration marked path; According to the differences from each path point of the pheromone concentration marked path to each cluster center, the pheromone concentration marked path is optimized to obtain the prediction feature uncertainty information.

6. The method according to claim 5, characterized in that The step of optimizing the pheromone concentration marked path according to the differences from each path point of the pheromone concentration marked path to each cluster center to obtain the prediction feature uncertainty information includes: Constructing a deviation matrix according to the differences from each path point of the pheromone concentration marked path to each cluster center; Iteratively optimizing the pheromone concentration marking path according to the deviation matrix to obtain an optimal cluster number and an optimal marking path; The prediction feature uncertainty information is generated according to the optimal number of clusters and the optimal labeling path.

7. A device for constructing a day-ahead regional transaction clearing model for high-proportion renewable energy generation, characterized in that: The device comprises: A prediction data acquisition module is used to predict the power conversion situation of the power resource interaction area according to the multi-dimensional power grid operation data of the power resource interaction area, and obtain a power conversion prediction feature data set; A data point clustering module, used for performing clustering processing on each electric energy conversion prediction feature data point in the electric energy conversion prediction feature data set to obtain prediction feature uncertainty information of each electric energy conversion prediction feature data point; A probability information calculation module, used to calculate the clearing scenario probability information of the power resource interaction area under different clearing typical distribution scenarios according to each of the power conversion prediction characteristic data points; A clearing model obtaining module is used to add each of the typical distribution scenarios of clearing and the probability information of each of the clearing scenarios to the target clearing function of the power resource interaction area according to the prediction feature uncertainty information, so as to obtain the day-ahead regional transaction clearing model of the power resource interaction area.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

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

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

Images