Virtual power plant participation electric power spot market clearing optimization method

Abstract

The invention relates to a virtual power plant participation electric power spot market clearing optimization method comprising the following steps: obtaining artificial intelligence-based source network load storage real-time data, and processing the data; constructing a source network load storage virtual power plant prediction and evaluation model, and performing test optimization to obtain an optimal source network load storage virtual power plant prediction and evaluation model; and inputting the source network load storage characteristic data into the optimal source network load storage virtual power plant prediction and evaluation model, and performing prediction and evaluation to obtain a source network load storage virtual power plant prediction and evaluation result. The method can effectively process the source network load storage real-time data, guarantees the quality of the source network load storage real-time data, improves the prediction precision and decision-making effect of the model, can optimize the participation of the source network load storage virtual power plant in the spot market transaction, can achieve the reasonable allocation and optimal utilization of power resources, and improves the efficiency of the model. And the output of the source network load storage virtual power plant can be intelligently adjusted, the market demand can be met, and the stable operation of the power grid can be ensured.

Description

Technical Field

[0001] The present invention relates to the field of source-grid-load-storage technology, and in particular to a method for optimizing the clearing of a virtual power plant participating in a spot electricity market. Background Art

[0002] A virtual power plant is not a power plant in the traditional sense, but a regional source, grid, load and storage concentration. It can serve as a unified agent for distributed energy sources to participate in the electricity market. It can effectively utilize cutting-edge communication and aggregation technologies combined with multi-energy complementary effects to achieve multi-source integrated scheduling. With the rapid development of distributed power sources, electric vehicles, energy storage and microgrids, deepening the construction of the electricity spot market, further promoting the coordinated interaction of sources, grids, loads and storage, and establishing a market-oriented operation mechanism for virtual power plants under the spot background will help to fully tap the system's flexible regulation capabilities and demand-side resources, coordinate the development and scientific allocation of various resources, improve the system's operating efficiency and the comprehensive benefits of power development, and improve the power system's real-time balance and safe supply capabilities.

[0003] The current virtual power plant system includes an information center terminal, distributed energy, a dispatch center terminal, and a virtual power plant; the information center terminal is used to publish the virtual power plant's power purchase and sale information; the distributed energy uses the virtual power plant's power purchase and sale information published by the information center terminal, combined with its own power supply and demand characteristics and historical information for dynamic aggregation, and selects its preferred virtual power plant; the dispatch center terminal is used to publish dispatch requirements; the virtual power plant is used to optimize the dispatch of the distributed energy aggregated in itself in combination with the dispatch requirements published by the dispatch center terminal, and obtain the operation plan of each distributed energy. However, the existing virtual power plant system cannot process the real-time data of the source, grid, load and storage, and cannot guarantee the quality of the real-time data of the source, grid, load and storage, resulting in the subsequent inability to effectively optimize the source, grid, load and storage virtual power plant's participation in spot market transactions, and cannot achieve the reasonable allocation and optimal utilization of power resources, and cannot intelligently adjust the output of the virtual power plant, and cannot meet market demand and ensure the stable operation of the power grid. Summary of the invention

[0004] In order to solve the problems in the prior art that the participation of source, grid, load and storage virtual power plants in spot market transactions cannot be effectively optimized, the reasonable allocation and optimal utilization of power resources cannot be achieved, and the output of virtual power plants cannot be intelligently adjusted to meet market demand and ensure stable operation of the power grid, the purpose of the present invention is to provide a method for optimizing the participation of source, grid, load and storage virtual power plants in spot market transactions, which can achieve reasonable allocation and optimal utilization of power resources, and intelligently adjust the output of virtual power plants to meet market demand and ensure stable operation of the power grid.

[0005] To achieve the above object, the present invention adopts the following technical solution: a virtual power plant participating in the electricity spot market clearing optimization method, the method comprising the following steps in order:

[0006] (1) Obtaining real-time data of source, grid, load and storage based on artificial intelligence, processing the real-time data of source, grid, load and storage, and obtaining characteristic data of source, grid, load and storage;

[0007] (2) According to the optimization needs of the source-grid-load-storage virtual power plant participating in spot market transactions, a source-grid-load-storage virtual power plant prediction and evaluation model is constructed, and the source-grid-load-storage virtual power plant prediction and evaluation model is tested and optimized to obtain the optimal source-grid-load-storage virtual power plant prediction and evaluation model;

[0008] (3) Inputting the source-grid-load-storage characteristic data into the optimal source-grid-load-storage virtual power plant prediction and evaluation model, predicting and evaluating the source-grid-load-storage real-time data, obtaining the source-grid-load-storage virtual power plant prediction and evaluation results, formulating the source-grid-load-storage virtual power plant intelligent regulation and control plan, and intelligently regulating and controlling the energy storage output of the virtual power plant;

[0009] (4) Obtain the status, rights and responsibilities of virtual power plants with power sources, grids, loads and storage in the spot market, as well as the corresponding market rules and policies, adapt to the implementation path of the provincial spot market operation and settlement mechanism in which virtual power plants with power sources, grids, loads and storage participate, and build a declaration, clearing, pricing and settlement sharing mechanism for virtual power plants with power sources, grids, loads and storage in the spot market, determine their degree of connection with the existing electricity market system and the corresponding risk prevention and arbitrage prevention mechanisms and measures.

[0010] In step (1), the acquisition of artificial intelligence-based source-grid-load-storage real-time data specifically refers to: real-time monitoring of the charging and discharging power and storage capacity of the energy storage equipment in the source-grid-load-storage virtual power plant to obtain the virtual power plant energy storage data; real-time monitoring of the power consumption of the power-consuming equipment in the power grid system to obtain the power consumption data of the power grid system; and obtaining artificial intelligence-based source-grid-load-storage real-time data based on the virtual power plant energy storage data and the power consumption data of the power grid system.

[0011] In step (1), the processing of the source-grid-load-storage real-time data to obtain source-grid-load-storage characteristic data specifically includes the following steps in sequence:

[0012] (1a) Cleaning and processing the source-grid-load-storage real-time data based on artificial intelligence to obtain the cleaned source-grid-load-storage real-time data that is valuable for the source-grid-load-storage virtual power plant to participate in spot market transaction optimization, i.e., the cleaned source-grid-load-storage real-time data;

[0013] (1b) converting and processing the cleaned source-grid-load-storage real-time data, i.e., standardizing it, eliminating the dimensional differences between the source-grid-load-storage real-time data that are valuable for the source-grid-load-storage virtual power plant to participate in spot market transaction optimization, and obtaining standardized source-grid-load-storage real-time data based on artificial intelligence;

[0014] (1c) Feature extraction is performed on the standardized AI-based real-time data of source, grid, load and storage, and features that can reflect the optimization of source, grid, load and storage virtual power plants participating in spot market transactions are extracted to obtain AI-based source, grid, load and storage feature data.

[0015] The step (2) specifically comprises the following steps in order:

[0016] (2a) According to the optimization requirements of the source-grid-load-storage virtual power plant participating in spot market transactions, historical data of the source-grid-load-storage are collected and divided to obtain a training data set and a test data set;

[0017] (2b) Select a neural network model framework suitable for the prediction and evaluation of source-grid-load-storage virtual power plants;

[0018] (2c) inputting the training data set into the neural network model framework for training to obtain a source-grid-load-storage virtual power plant prediction and evaluation model based on artificial intelligence;

[0019] (2d) Inputting the test data set into the source-grid-load-storage virtual power plant prediction and evaluation model to perform performance testing and obtain performance test results; according to the performance test results, adjusting and optimizing the parameters of the source-grid-load-storage virtual power plant prediction and evaluation model to obtain the optimal source-grid-load-storage virtual power plant prediction and evaluation model.

[0020] In step (4), the risk prevention and arbitrage prevention mechanism measures include strengthening data monitoring and risk warning, strict entry and exit mechanisms, formulating detailed operating procedures, establishing a sound information disclosure system, introducing third-party institutions for supervision, establishing a sound emergency plan and continuously optimizing the market supervision system;

[0021] Strengthening data monitoring and risk warning means: establishing a real-time and comprehensive power market data monitoring system, timely discovering potential risk points and abnormal situations by comparing abnormal transactions and price fluctuation information, forming a risk warning model through the analysis of historical data, and predicting possible risk events in advance;

[0022] The strict entry and exit mechanism means: a strict entry and exit mechanism is implemented for virtual power plants participating in the power market to ensure fair competition in the market. For virtual power plants that do not comply with regulations or have potential risks, transactions with them will be stopped immediately to prevent the spread of risks.

[0023] The formulation of detailed operating procedures refers to: formulating detailed and clear operating procedures for the declaration, clearance, pricing and settlement of virtual power plants in the spot market, regulating the behavior of market players and avoiding risks caused by illegal operations;

[0024] The establishment of a sound information disclosure system refers to requiring virtual power plants to fully disclose relevant information during the transaction process and accept public supervision. At the same time, government departments should also increase the disclosure of market information to ensure market transparency and prevent arbitrage caused by information asymmetry;

[0025] The introduction of a third-party agency for supervision refers to: introducing an independent third-party agency to supervise the electricity market;

[0026] The establishment of a comprehensive emergency plan means: when a risk event occurs, the emergency plan is quickly activated and effective measures are taken to reduce the impact of risk spread;

[0027] The continuous optimization of the market supervision system refers to: constantly adjusting and improving the market supervision system according to market developments and risk changes, so that it can better adapt to the needs of power market operation under the new situation.

[0028] The step (1a) specifically comprises the following steps in order:

[0029] (1a1) Perform consistency check on the source, grid, load and storage real-time data based on artificial intelligence: Check whether the source, grid, load and storage real-time data based on artificial intelligence meets the requirements according to the reasonable value range and mutual relationship of each parameter in the source, grid, load and storage real-time data based on artificial intelligence;

[0030] (1a2) Remove inconsistent data that is beyond the normal range, logically unreasonable or contradictory in the real-time data of source, grid, load and storage based on artificial intelligence;

[0031] (1a3) Processing invalid values ​​and missing values ​​of the source, grid, load and storage real-time data based on artificial intelligence, and removing invalid data and missing data that are of no value to the source, grid, load and storage virtual power plant participating in the spot market transaction optimization in the source, grid, load and storage real-time data based on artificial intelligence: extracting the amount of invalid data and the time when the missing data is generated corresponding to the missing position of the missing data; obtaining a bad data evaluation coefficient according to the amount of invalid data and the time when the missing data is generated corresponding to the missing position of the missing data;

[0032] The bad data evaluation coefficient is obtained by the following formula:

[0033]

[0034] Among them, g represents the bad data evaluation coefficient; n represents the number of invalid data; C w Indicates the total amount of invalid data; C y The amount of valid data that is valuable for optimizing the source-grid-load-storage virtual power plant's participation in spot market transactions; T 01i represents the collection time corresponding to the i-th invalid data; T 01i-1represents the collection time corresponding to the i-1th invalid data; T c01 represents the preset reference value of the collection time interval corresponding to the invalid data; m represents the number of missing data; T 02i represents the time when the missing data corresponding to the i-th missing data is generated; T 02i-1 represents the time when the missing data corresponding to the i-1th missing data is generated; T c02 represents the reference value of the generation time interval corresponding to the preset missing data; k represents the number of valid data that are valuable for the source-grid-load-storage virtual power plant to participate in the spot market transaction optimization;

[0035] The bad data evaluation coefficient is compared with a preset bad data evaluation coefficient threshold. When the bad data evaluation coefficient exceeds the preset bad data evaluation coefficient threshold, data filling is performed on invalid data and missing data.

[0036] The data filling of invalid data specifically refers to: extracting the data value of valid data that is closest to the time when the invalid data is collected and is valuable for the source grid load storage virtual power plant to participate in the spot market transaction optimization; using the data value of valid data that is closest to the time when the invalid data is collected and is valuable for the source grid load storage virtual power plant to participate in the spot market transaction optimization to obtain the data filling adjustment coefficient;

[0037] The data filling adjustment coefficient is obtained by the following formula:

[0038]

[0039] Where S represents the data filling adjustment coefficient; X 01 Indicates the data value of the valid data that is closest to the collection time corresponding to the invalid data and is valuable for the source-grid-load-storage virtual power plant to participate in the spot market transaction optimization; X 02 The data value representing the closest valid data after the collection time corresponding to the invalid data, which is valuable for the source-grid-load-storage virtual power plant to participate in the spot market transaction optimization; X up Indicates the upper limit of the normal range corresponding to the real-time data of source, grid, load and storage based on artificial intelligence; X down It indicates the lower limit of the normal range corresponding to the real-time data of the source, grid, load and storage based on artificial intelligence; sqrt indicates the square root operation function symbol; exp indicates the exponential function symbol; ε indicates the preset minimum constant, which is used to prevent the numerator from being 0;

[0040] Utilize the data values ​​of valid data other than the valid data that is closest to the collection time of each invalid data and is valuable for the source-grid-load-storage virtual power plant to participate in the optimization of spot market transactions, and combine the data filling adjustment coefficient to obtain the filling data corresponding to each invalid data; and replace the corresponding invalid data with the filling data corresponding to the invalid data;

[0041] The filling data corresponding to each invalid data is obtained by the following formula:

[0042]

[0043] Among them, X w Indicates the padding data corresponding to each invalid data.

[0044] The data filling of missing data specifically refers to: extracting the data value of the valid data closest to the missing data before and after the missing data generation time as the first data value; extracting the data value of the filling data corresponding to the invalid data closest to the missing data before and after the missing data generation time as the second data value; calling the data filling adjustment coefficient; using the first data value and the second data value, combined with the data filling adjustment coefficient, to obtain the filling data corresponding to the missing data;

[0045] The filling data corresponding to the missing data is obtained by the following formula:

[0046]

[0047] Among them, X q represents the filling data corresponding to the missing data; S represents the data filling adjustment coefficient; X y01 The data value representing the closest valid data before the time when the missing data is generated, which is valuable for the source-grid-load-storage virtual power plant to participate in the spot market transaction optimization; X y02 The data value representing the closest valid data after the missing data is generated, which is valuable for the source-grid-load-storage virtual power plant to participate in the spot market transaction optimization; X w01 Indicates the data value of the filling data corresponding to the invalid data closest to the time when the missing data is generated; X w02 Indicates the data value of the filling data corresponding to the invalid data closest to the time when the missing data is generated;

[0048] The missing data corresponding to the missing data is replaced with the filling data corresponding to the missing data.

[0049] It can be seen from the above technical scheme that the beneficial effects of the present invention are: first, the present invention obtains the real-time data of source, grid, load and storage based on artificial intelligence, processes the real-time data of source, grid, load and storage based on artificial intelligence, determines the characteristic data of source, grid, load and storage based on artificial intelligence, and constructs a prediction and evaluation model of a source, grid, load and storage virtual power plant according to the optimization needs of source, grid, load and storage virtual power plant participating in spot market transactions. The prediction and evaluation model of the source, grid, load and storage virtual power plant is tested and optimized to determine the optimal prediction and evaluation model of the source, grid, load and storage virtual power plant. Based on the optimal prediction and evaluation model of the source, grid, load and storage virtual power plant, the real-time data of source, grid, load and storage based on artificial intelligence is predicted and evaluated to determine the prediction and evaluation results of the source, grid, load and storage virtual power plant; second, the present invention formulates an intelligent regulation and control system for the source, grid, load and storage virtual power plant A plan is developed to intelligently regulate and control the energy storage output of the source-grid-load-storage virtual power plant, and to adapt to the implementation path of the provincial spot market operation and settlement mechanism in which the source-grid-load-storage virtual power plant participates. A declaration, clearing, pricing, and settlement sharing mechanism for the source-grid-load-storage virtual power plant in the spot market is established, and its degree of connection with the existing power market system and the corresponding risk prevention and arbitrage prevention mechanisms and measures are judged. The real-time data of the source-grid-load-storage can be effectively processed to ensure the quality of the real-time data, thereby improving the prediction accuracy and decision-making effect of the model, optimizing the participation of the source-grid-load-storage virtual power plant in spot market transactions, realizing the rational allocation and optimal utilization of power resources, and intelligently regulating the output of the source-grid-load-storage virtual power plant to meet market demand and ensure the stable operation of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 is a system structure block diagram of the present invention;

[0051] Figure 2 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0052] like Figure 1 , Figure 2 As shown, a virtual power plant participates in the electricity spot market clearing optimization method, the method comprises the following steps in sequence:

[0053] (1) Obtaining real-time data of source, grid, load and storage based on artificial intelligence, processing the real-time data of source, grid, load and storage, and obtaining characteristic data of source, grid, load and storage;

[0054] (2) According to the optimization needs of the source-grid-load-storage virtual power plant participating in spot market transactions, a source-grid-load-storage virtual power plant prediction and evaluation model is constructed, and the source-grid-load-storage virtual power plant prediction and evaluation model is tested and optimized to obtain the optimal source-grid-load-storage virtual power plant prediction and evaluation model;

[0055] (3) Inputting the source-grid-load-storage characteristic data into the optimal source-grid-load-storage virtual power plant prediction and evaluation model, predicting and evaluating the source-grid-load-storage real-time data, obtaining the source-grid-load-storage virtual power plant prediction and evaluation results, formulating the source-grid-load-storage virtual power plant intelligent regulation and control plan, and intelligently regulating and controlling the energy storage output of the virtual power plant;

[0056] (4) Obtain the status, rights and responsibilities of virtual power plants with power sources, grids, loads and storage in the spot market, as well as the corresponding market rules and policies, adapt to the implementation path of the provincial spot market operation and settlement mechanism in which virtual power plants with power sources, grids, loads and storage participate, and build a declaration, clearing, pricing and settlement sharing mechanism for virtual power plants with power sources, grids, loads and storage in the spot market, determine their degree of connection with the existing electricity market system and the corresponding risk prevention and arbitrage prevention mechanisms and measures.

[0057] In step (1), the acquisition of artificial intelligence-based source-grid-load-storage real-time data specifically refers to: real-time monitoring of the charging and discharging power and storage capacity of the energy storage equipment in the source-grid-load-storage virtual power plant to obtain the virtual power plant energy storage data; real-time monitoring of the power consumption of the power-consuming equipment in the power grid system to obtain the power consumption data of the power grid system; and obtaining artificial intelligence-based source-grid-load-storage real-time data based on the virtual power plant energy storage data and the power consumption data of the power grid system.

[0058] In step (1), the processing of the source-grid-load-storage real-time data to obtain source-grid-load-storage characteristic data specifically includes the following steps in sequence:

[0059] (1a) Cleaning and processing the source-grid-load-storage real-time data based on artificial intelligence to obtain the cleaned source-grid-load-storage real-time data that is valuable for the source-grid-load-storage virtual power plant to participate in spot market transaction optimization, i.e., the cleaned source-grid-load-storage real-time data;

[0060] (1b) converting and processing the cleaned source-grid-load-storage real-time data, i.e., standardizing it, eliminating the dimensional differences between the source-grid-load-storage real-time data that are valuable for the source-grid-load-storage virtual power plant to participate in spot market transaction optimization, and obtaining standardized source-grid-load-storage real-time data based on artificial intelligence;

[0061] (1c) Feature extraction is performed on the standardized AI-based real-time data of source, grid, load and storage, and features that can reflect the optimization of source, grid, load and storage virtual power plants participating in spot market transactions are extracted to obtain AI-based source, grid, load and storage feature data.

[0062] The step (2) specifically comprises the following steps in order:

[0063] (2a) According to the optimization requirements of the source-grid-load-storage virtual power plant participating in spot market transactions, historical data of the source-grid-load-storage are collected and divided to obtain a training data set and a test data set;

[0064] (2b) Select a neural network model framework suitable for the prediction and evaluation of source-grid-load-storage virtual power plants;

[0065] (2c) inputting the training data set into the neural network model framework for training to obtain a source-grid-load-storage virtual power plant prediction and evaluation model based on artificial intelligence;

[0066] (2d) Inputting the test data set into the source-grid-load-storage virtual power plant prediction and evaluation model to perform performance testing and obtain performance test results; according to the performance test results, adjusting and optimizing the parameters of the source-grid-load-storage virtual power plant prediction and evaluation model to obtain the optimal source-grid-load-storage virtual power plant prediction and evaluation model.

[0067] In step (4), the risk prevention and arbitrage prevention mechanism measures include strengthening data monitoring and risk warning, strict entry and exit mechanisms, formulating detailed operating procedures, establishing a sound information disclosure system, introducing third-party institutions for supervision, establishing a sound emergency plan and continuously optimizing the market supervision system;

[0068] Strengthening data monitoring and risk warning means: establishing a real-time and comprehensive power market data monitoring system, timely discovering potential risk points and abnormal situations by comparing abnormal transactions and price fluctuation information, forming a risk warning model through the analysis of historical data, and predicting possible risk events in advance;

[0069] The strict entry and exit mechanism means: a strict entry and exit mechanism is implemented for virtual power plants participating in the power market to ensure fair competition in the market. For virtual power plants that do not comply with regulations or have potential risks, transactions with them will be stopped immediately to prevent the spread of risks.

[0070] The formulation of detailed operating procedures refers to: formulating detailed and clear operating procedures for the declaration, clearance, pricing and settlement of virtual power plants in the spot market, regulating the behavior of market players and avoiding risks caused by illegal operations;

[0071] The establishment of a sound information disclosure system refers to requiring virtual power plants to fully disclose relevant information during the transaction process and accept public supervision. At the same time, government departments should also increase the disclosure of market information to ensure market transparency and prevent arbitrage caused by information asymmetry;

[0072] The introduction of a third-party agency for supervision refers to: introducing an independent third-party agency to supervise the electricity market;

[0073] The establishment of a comprehensive emergency plan means: when a risk event occurs, the emergency plan is quickly activated and effective measures are taken to reduce the impact of risk spread;

[0074] The continuous optimization of the market supervision system refers to: constantly adjusting and improving the market supervision system according to market developments and risk changes, so that it can better adapt to the needs of power market operation under the new situation.

[0075] The step (1a) specifically comprises the following steps in order:

[0076] (1a1) Perform consistency check on the source, grid, load and storage real-time data based on artificial intelligence: Check whether the source, grid, load and storage real-time data based on artificial intelligence meets the requirements according to the reasonable value range and mutual relationship of each parameter in the source, grid, load and storage real-time data based on artificial intelligence;

[0077] (1a2) Remove inconsistent data that is beyond the normal range, logically unreasonable or contradictory in the real-time data of source, grid, load and storage based on artificial intelligence;

[0078] (1a3) Processing invalid values ​​and missing values ​​of the source, grid, load and storage real-time data based on artificial intelligence, and removing invalid data and missing data that are of no value to the source, grid, load and storage virtual power plant participating in the spot market transaction optimization in the source, grid, load and storage real-time data based on artificial intelligence: extracting the amount of invalid data and the time when the missing data is generated corresponding to the missing position of the missing data; obtaining a bad data evaluation coefficient according to the amount of invalid data and the time when the missing data is generated corresponding to the missing position of the missing data;

[0079] The bad data evaluation coefficient is obtained by the following formula:

[0080]

[0081] Among them, g represents the bad data evaluation coefficient; n represents the number of invalid data; C w Indicates the total amount of invalid data; C y The amount of valid data that is valuable for optimizing the source-grid-load-storage virtual power plant's participation in spot market transactions; T 01i represents the collection time corresponding to the i-th invalid data; T 01i-1 represents the collection time corresponding to the i-1th invalid data; T c01 represents the preset reference value of the collection time interval corresponding to the invalid data; m represents the number of missing data; T 02i represents the time when the missing data corresponding to the i-th missing data is generated; T 02i-1 represents the time when the missing data corresponding to the i-1th missing data is generated; T c02 represents the reference value of the generation time interval corresponding to the preset missing data; k represents the number of valid data that are valuable for the source-grid-load-storage virtual power plant to participate in the spot market transaction optimization;

[0082] The bad data evaluation coefficient is compared with a preset bad data evaluation coefficient threshold. When the bad data evaluation coefficient exceeds the preset bad data evaluation coefficient threshold, data filling is performed on invalid data and missing data.

[0083] By calculating the bad data evaluation coefficient, the number of invalid data, the number of valid data, the time interval between invalid data, and the time interval between missing data are comprehensively considered, providing a comprehensive quantitative indicator for data quality. This multi-dimensional evaluation method is more comprehensive than a single data integrity or accuracy check, and can more accurately reflect the overall quality of the data source. By setting the bad data evaluation coefficient threshold, it can automatically determine whether the data needs further processing (such as data filling). When the data quality does not meet the requirements (that is, g exceeds the preset threshold), the system can automatically trigger the data filling mechanism to reduce manual intervention and improve processing efficiency.

[0084] At the same time, the accuracy and completeness of valid data are crucial for the source-grid-load-storage virtual power plant to participate in spot market transaction optimization. Therefore, through the data processing process of the present invention, the quality of data input into the transaction optimization model can be ensured, thereby improving the prediction accuracy and decision-making effect of the model, and ultimately enhancing the market competitiveness of the virtual power plant. In addition, if invalid values ​​and missing values ​​in the data are not processed, they may have a negative impact on the normal operation of the system. Through the data filling mechanism, system anomalies or erroneous decisions caused by data problems can be reduced, and the reliability and stability of the system can be enhanced.

[0085] The present invention processes the real-time data of source, grid, load and storage based on artificial intelligence technology, which reflects the advantages of artificial intelligence technology in the field of data processing and analysis. By continuously optimizing the data processing process and method, the application effect of artificial intelligence technology in the power system can be further improved, and the intelligent upgrade of the power system can be promoted. By accurately evaluating data quality, optimizing data processing processes, improving transaction optimization capabilities, enhancing system reliability and stability, and promoting the effective application of artificial intelligence technology, it provides strong technical support for the optimization of spot market transactions of source, grid, load and storage virtual power plants.

[0086] The data filling of invalid data specifically refers to: extracting the data value of valid data that is closest to the time when the invalid data is collected and is valuable for the source grid load storage virtual power plant to participate in the spot market transaction optimization; using the data value of valid data that is closest to the time when the invalid data is collected and is valuable for the source grid load storage virtual power plant to participate in the spot market transaction optimization to obtain the data filling adjustment coefficient;

[0087] The data filling adjustment coefficient is obtained by the following formula:

[0088]

[0089] Where S represents the data filling adjustment coefficient; X 01 Indicates the data value of the valid data that is closest to the collection time corresponding to the invalid data and is valuable for the source-grid-load-storage virtual power plant to participate in the spot market transaction optimization; X 02 The data value representing the closest valid data after the collection time corresponding to the invalid data, which is valuable for the source-grid-load-storage virtual power plant to participate in the spot market transaction optimization; X up Indicates the upper limit of the normal range corresponding to the real-time data of source, grid, load and storage based on artificial intelligence; X down It indicates the lower limit of the normal range corresponding to the real-time data of the source, grid, load and storage based on artificial intelligence; sqrt indicates the square root operation function symbol; exp indicates the exponential function symbol; ε indicates the preset minimum constant, which is used to prevent the numerator from being 0;

[0090] Utilize the data values ​​of valid data other than the valid data that is closest to the collection time of each invalid data and is valuable for the source-grid-load-storage virtual power plant to participate in the optimization of spot market transactions, and combine the data filling adjustment coefficient to obtain the filling data corresponding to each invalid data; and replace the corresponding invalid data with the filling data corresponding to the invalid data;

[0091] The filling data corresponding to each invalid data is obtained by the following formula:

[0092]

[0093] Among them, X w Indicates the padding data corresponding to each invalid data.

[0094] By extracting the valid data values ​​closest to the invalid data collection time, and using these data values ​​to calculate the data filling adjustment coefficient, a reasonable filling value can be provided for the invalid data. This data filling method not only considers the time series of the data, but also combines the normal range of the data, thereby improving the accuracy and rationality of the filled data.

[0095] The introduction of the data filling adjustment coefficient S enables the filling data to be dynamically adjusted according to the distribution of valid data around the invalid data. By using exponential functions and square root operations, the difference between the filling data and the surrounding valid data can be more finely controlled to avoid the filling data from deviating too much from the normal range. Effective data filling can reduce decision errors caused by missing or invalid data, and can more accurately reflect the real-time status of the source-grid-load-storage virtual power plant, thereby providing more reliable data support for its participation in spot market transaction optimization. The data filling mechanism can reduce system anomalies or wrong decisions caused by data problems, can better cope with situations where data is missing or invalid, and improve the robustness and stability of the system. Combining the advantages of artificial intelligence technology in data processing and analysis, through a sophisticated data filling algorithm, the availability and quality of data are improved, which not only solves practical problems, but also promotes the in-depth application and development of artificial intelligence technology in power systems.

[0096] Through the introduction of sophisticated data filling algorithms and adjustment coefficients, the integrity and accuracy of the data are improved, the data filling process is optimized, the decision-making ability of the source-grid-load-storage virtual power plant is enhanced, the robustness and stability of the system are improved, and the in-depth application of artificial intelligence technology is promoted. These technical effects together provide strong support for the optimization of spot market transactions of the source-grid-load-storage virtual power plant.

[0097] The data filling of missing data specifically refers to: extracting the data value of the valid data closest to the missing data before and after the missing data generation time as the first data value; extracting the data value of the filling data corresponding to the invalid data closest to the missing data before and after the missing data generation time as the second data value; calling the data filling adjustment coefficient; using the first data value and the second data value, combined with the data filling adjustment coefficient, to obtain the filling data corresponding to the missing data;

[0098] The filling data corresponding to the missing data is obtained by the following formula:

[0099]

[0100] Among them, X q represents the filling data corresponding to the missing data; S represents the data filling adjustment coefficient; X y01 The data value representing the closest valid data before the time when the missing data is generated, which is valuable for the source-grid-load-storage virtual power plant to participate in the spot market transaction optimization; X y02 The data value representing the closest valid data after the missing data is generated, which is valuable for the source-grid-load-storage virtual power plant to participate in the spot market transaction optimization; X w01 Indicates the data value of the filling data corresponding to the invalid data closest to the time when the missing data is generated; X w02Indicates the data value of the filling data corresponding to the invalid data closest to the time when the missing data is generated;

[0101] The missing data corresponding to the missing data is replaced with the filling data corresponding to the missing data.

[0102] By extracting the valid data value (first data value) closest to the time when the missing data is generated and the filling data value (second data value) corresponding to the invalid data, and combining the data filling adjustment coefficient S, a reasonable filling value can be provided for the missing data. This data filling method not only takes into account the time series and continuity of the data, but also combines the information of the valid data and the filled invalid data, thereby improving the accuracy and rationality of the filled data and making the overall data set more complete and continuous.

[0103] The introduction of the data filling adjustment coefficient S enables the filling data to be dynamically adjusted according to the distribution of valid data around the missing data and the filled invalid data. By comprehensively considering the first data value, the second data value and the data filling adjustment coefficient, the difference between the filling data and the surrounding data can be more finely controlled to avoid the filling data from deviating too much from the normal range or producing unreasonable mutations. Effective data filling can reduce the analysis errors and decision-making errors caused by missing data, and can more accurately reflect the real-time status and historical trends of the source-grid-load-storage virtual power plant, thereby providing more reliable data support for its participation in spot market transaction optimization. The data filling mechanism can reduce system anomalies or wrong decisions caused by missing data. It can better cope with the situation of missing data, improve the robustness and reliability of the system, and ensure the stable operation of the system. Combining data filling technology with the concept of data-driven decision-making, through fine data processing and analysis, it provides strong data support for the decision-making of the source-grid-load-storage virtual power plant, which not only solves the problem of missing data, but also promotes the in-depth application and development of data-driven decision-making in the power system.

[0104] By comprehensively considering valid data, filled invalid data, and data filling adjustment coefficients, a reasonable filling value is provided for missing data, which improves the integrity and continuity of data, optimizes data filling strategies, enhances the accuracy of data analysis and decision-making, improves the robustness and reliability of the system, and promotes the development of data-driven decision-making. These technical effects together provide strong support for the optimization of spot market transactions of source-grid-load-storage virtual power plants.

[0105] To sum up, the present invention formulates an intelligent regulation and control scheme for a source-grid-load-storage virtual power plant, performs intelligent regulation and control on the energy storage output of the source-grid-load-storage virtual power plant, and adapts to the implementation path of the provincial spot market operation and settlement mechanism in which the source-grid-load-storage virtual power plant participates, constructs a declaration, clearing, pricing, and settlement sharing mechanism for the source-grid-load-storage virtual power plant in the spot market, judges its degree of connection with the existing power market system and the corresponding risk prevention and arbitrage prevention mechanisms and measures, can effectively process the real-time data of the source-grid-load-storage virtual power plant, ensure the quality of the real-time data of the source-grid-load-storage virtual power plant, thereby improving the prediction accuracy and decision-making effect of the model, can optimize the source-grid-load-storage virtual power plant's participation in spot market transactions, can realize the reasonable allocation and optimal utilization of power resources, and can intelligently regulate the output of the source-grid-load-storage virtual power plant, can meet market demand and ensure the stable operation of the power grid.

[0106] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A virtual power plant participating in the electricity spot market clearing optimization method, characterized by: The method comprises the following steps in order: (1) Obtaining real-time data of source, grid, load and storage based on artificial intelligence, processing the real-time data of source, grid, load and storage, and obtaining characteristic data of source, grid, load and storage; (2) According to the optimization needs of the source-grid-load-storage virtual power plant participating in spot market transactions, a source-grid-load-storage virtual power plant prediction and evaluation model is constructed, and the source-grid-load-storage virtual power plant prediction and evaluation model is tested and optimized to obtain the optimal source-grid-load-storage virtual power plant prediction and evaluation model; (3) Inputting the source-grid-load-storage characteristic data into the optimal source-grid-load-storage virtual power plant prediction and evaluation model, predicting and evaluating the source-grid-load-storage real-time data, obtaining the source-grid-load-storage virtual power plant prediction and evaluation results, formulating the source-grid-load-storage virtual power plant intelligent regulation and control plan, and intelligently regulating and controlling the energy storage output of the virtual power plant; (4) Obtain the status, rights and responsibilities of virtual power plants with power sources, grids, loads and storage in the spot market, as well as the corresponding market rules and policies, adapt to the implementation path of the provincial spot market operation and settlement mechanism in which virtual power plants with power sources, grids, loads and storage participate, and build a declaration, clearing, pricing and settlement sharing mechanism for virtual power plants with power sources, grids, loads and storage in the spot market, determine their degree of connection with the existing electricity market system and the corresponding risk prevention and arbitrage prevention mechanisms and measures.

2. The method for optimizing the clearing of the electricity spot market by virtual power plants according to claim 1, characterized in that: In step (1), the acquisition of artificial intelligence-based source-grid-load-storage real-time data specifically refers to: real-time monitoring of the charging and discharging power and storage capacity of the energy storage equipment in the source-grid-load-storage virtual power plant to obtain the virtual power plant energy storage data; real-time monitoring of the power consumption of the power-consuming equipment in the power grid system to obtain the power consumption data of the power grid system; and obtaining artificial intelligence-based source-grid-load-storage real-time data based on the virtual power plant energy storage data and the power consumption data of the power grid system.

3. The method for optimizing the clearing of the electricity spot market by a virtual power plant according to claim 1, characterized in that: In step (1), the processing of the source-grid-load-storage real-time data to obtain source-grid-load-storage characteristic data specifically includes the following steps in sequence: (1a) Cleaning and processing the source-grid-load-storage real-time data based on artificial intelligence to obtain the cleaned source-grid-load-storage real-time data that is valuable for the source-grid-load-storage virtual power plant to participate in spot market transaction optimization, i.e., the cleaned source-grid-load-storage real-time data; (1b) converting and processing the cleaned source-grid-load-storage real-time data, i.e., standardizing it, eliminating the dimensional differences between the source-grid-load-storage real-time data that are valuable for the source-grid-load-storage virtual power plant to participate in spot market transaction optimization, and obtaining standardized source-grid-load-storage real-time data based on artificial intelligence; (1c) Feature extraction is performed on the standardized AI-based real-time data of source, grid, load and storage, and features that can reflect the optimization of source, grid, load and storage virtual power plants participating in spot market transactions are extracted to obtain AI-based source, grid, load and storage feature data.

4. The method for optimizing the clearing of the electricity spot market by a virtual power plant according to claim 1, characterized in that: The step (2) specifically comprises the following steps in order: (2a) According to the optimization requirements of the source-grid-load-storage virtual power plant participating in spot market transactions, historical data of the source-grid-load-storage are collected and divided to obtain a training data set and a test data set; (2b) Select a neural network model framework suitable for the prediction and evaluation of source-grid-load-storage virtual power plants; (2c) inputting the training data set into the neural network model framework for training to obtain a source-grid-load-storage virtual power plant prediction and evaluation model based on artificial intelligence; (2d) Inputting the test data set into the source-grid-load-storage virtual power plant prediction and evaluation model to perform performance testing and obtain performance test results; according to the performance test results, adjusting and optimizing the parameters of the source-grid-load-storage virtual power plant prediction and evaluation model to obtain the optimal source-grid-load-storage virtual power plant prediction and evaluation model.

5. The method for optimizing the clearing of the electricity spot market by virtual power plants according to claim 1, characterized in that: In step (4), the risk prevention and arbitrage prevention mechanism measures include strengthening data monitoring and risk warning, strict entry and exit mechanisms, formulating detailed operating procedures, establishing a sound information disclosure system, introducing third-party institutions for supervision, establishing a sound emergency plan and continuously optimizing the market supervision system; Strengthening data monitoring and risk warning means: establishing a real-time and comprehensive power market data monitoring system, timely discovering potential risk points and abnormal situations by comparing abnormal transactions and price fluctuation information, forming a risk warning model through the analysis of historical data, and predicting possible risk events in advance; The strict entry and exit mechanism means: a strict entry and exit mechanism is implemented for virtual power plants participating in the power market to ensure fair competition in the market. For virtual power plants that do not comply with regulations or have potential risks, transactions with them will be stopped immediately to prevent the spread of risks. The formulation of detailed operating procedures refers to: formulating detailed and clear operating procedures for the declaration, clearance, pricing and settlement of virtual power plants in the spot market, regulating the behavior of market players and avoiding risks caused by illegal operations; The establishment of a sound information disclosure system refers to requiring virtual power plants to fully disclose relevant information during the transaction process and accept public supervision. At the same time, government departments should also increase the disclosure of market information to ensure market transparency and prevent arbitrage caused by information asymmetry; The introduction of a third-party agency for supervision refers to: introducing an independent third-party agency to supervise the electricity market; The establishment of a comprehensive emergency plan means: when a risk event occurs, the emergency plan is quickly activated and effective measures are taken to reduce the impact of risk spread; The continuous optimization of the market supervision system refers to: constantly adjusting and improving the market supervision system according to market developments and risk changes, so that it can better adapt to the needs of power market operation under the new situation.

6. The method for optimizing the clearing of the electricity spot market by virtual power plants according to claim 3, characterized in that: The step (1a) specifically comprises the following steps in order: (1a1) Perform consistency check on the source, grid, load and storage real-time data based on artificial intelligence: Check whether the source, grid, load and storage real-time data based on artificial intelligence meets the requirements according to the reasonable value range and mutual relationship of each parameter in the source, grid, load and storage real-time data based on artificial intelligence; (1a2) Remove inconsistent data that is beyond the normal range, logically unreasonable or contradictory in the real-time data of source, grid, load and storage based on artificial intelligence; (1a3) Processing invalid values ​​and missing values ​​of the source, grid, load and storage real-time data based on artificial intelligence, and removing invalid data and missing data that are of no value to the source, grid, load and storage virtual power plant participating in the spot market transaction optimization in the source, grid, load and storage real-time data based on artificial intelligence: extracting the amount of invalid data and the time when the missing data is generated corresponding to the missing position of the missing data; obtaining a bad data evaluation coefficient according to the amount of invalid data and the time when the missing data is generated corresponding to the missing position of the missing data; The bad data evaluation coefficient is obtained by the following formula: Among them, g represents the bad data evaluation coefficient; n represents the number of invalid data; C w Indicates the total amount of invalid data; C y The amount of valid data that is valuable for optimizing the source-grid-load-storage virtual power plant's participation in spot market transactions; T 01i represents the collection time corresponding to the i-th invalid data; T 01i-1 represents the collection time corresponding to the i-1th invalid data; T c01 represents the preset reference value of the collection time interval corresponding to the invalid data; m represents the number of missing data; T 02i represents the time when the missing data corresponding to the i-th missing data is generated; T 02i-1 represents the time when the missing data corresponding to the i-1th missing data is generated; T c02 represents the reference value of the generation time interval corresponding to the preset missing data; k represents the number of valid data that are valuable for the source-grid-load-storage virtual power plant to participate in the spot market transaction optimization; The bad data evaluation coefficient is compared with a preset bad data evaluation coefficient threshold. When the bad data evaluation coefficient exceeds the preset bad data evaluation coefficient threshold, data filling is performed on invalid data and missing data.

7. The method for optimizing the clearing of the electricity spot market by virtual power plants according to claim 6, characterized in that: The data filling of invalid data specifically refers to: extracting the data value of valid data that is closest to the time when the invalid data is collected and is valuable for the source grid load storage virtual power plant to participate in the spot market transaction optimization; using the data value of valid data that is closest to the time when the invalid data is collected and is valuable for the source grid load storage virtual power plant to participate in the spot market transaction optimization to obtain the data filling adjustment coefficient; The data filling adjustment coefficient is obtained by the following formula: Where S represents the data filling adjustment coefficient; X 01 Indicates the data value of the valid data that is closest to the collection time corresponding to the invalid data and is valuable for the source-grid-load-storage virtual power plant to participate in the spot market transaction optimization; X 02 The data value representing the closest valid data after the collection time corresponding to the invalid data, which is valuable for the source-grid-load-storage virtual power plant to participate in the spot market transaction optimization; X up Indicates the upper limit of the normal range corresponding to the real-time data of source, grid, load and storage based on artificial intelligence; X down It indicates the lower limit of the normal range corresponding to the real-time data of the source, grid, load and storage based on artificial intelligence; sqrt indicates the square root operation function symbol; exp indicates the exponential function symbol; ε indicates the preset minimum constant, which is used to prevent the numerator from being 0; Utilize the data values ​​of valid data other than the valid data that is closest to the collection time of each invalid data and is valuable for the source-grid-load-storage virtual power plant to participate in the optimization of spot market transactions, and combine the data filling adjustment coefficient to obtain the filling data corresponding to each invalid data; and replace the corresponding invalid data with the filling data corresponding to the invalid data; The filling data corresponding to each invalid data is obtained by the following formula: Among them, X w Indicates the padding data corresponding to each invalid data.

8. The method for optimizing the clearing of the electricity spot market by virtual power plants according to claim 6, characterized in that: The data filling of missing data specifically refers to: extracting the data value of the valid data closest to the missing data before and after the missing data generation time as the first data value; extracting the data value of the filling data corresponding to the invalid data closest to the missing data before and after the missing data generation time as the second data value; calling the data filling adjustment coefficient; using the first data value and the second data value, combined with the data filling adjustment coefficient, to obtain the filling data corresponding to the missing data; The filling data corresponding to the missing data is obtained by the following formula: Among them, X q represents the filling data corresponding to the missing data; S represents the data filling adjustment coefficient; X y01 The data value representing the closest valid data before the time when the missing data is generated, which is valuable for the source-grid-load-storage virtual power plant to participate in the spot market transaction optimization; X y02 The data value representing the closest valid data after the missing data is generated, which is valuable for the source-grid-load-storage virtual power plant to participate in the spot market transaction optimization; X w01 Indicates the data value of the filling data corresponding to the invalid data closest to the time when the missing data is generated; X w02 Indicates the data value of the filling data corresponding to the invalid data closest to the time when the missing data is generated; The missing data corresponding to the missing data is replaced with the filling data corresponding to the missing data.

Images