Virtual power plant power transaction optimization method and system based on data driving

Abstract

The invention relates to the technical field of electric energy management, and discloses a virtual power plant electric power transaction optimization method and system based on data driving, and the method comprises the steps: obtaining an electric power price change curve, and carrying out the extension operation, and obtaining a primary electric power price prediction curve; calculating a real-time change rate corresponding to each direct factor according to a direct factor historical data set related to the electric power price, calculating to obtain a comprehensive electric power change rate, and calculating the real-time electric power price according to the comprehensive electric power change rate to obtain an advanced electric power price prediction curve; carrying out weighted average on the power price prediction curve and the primary power price prediction curve to obtain a power price prediction curve; obtaining a residual electric quantity prediction curve; and generating a power transaction storage scheme by using a pre-trained power transaction scheme generation model according to the power price prediction curve and the remaining power prediction curve. The power transmission and storage distribution proportion of the virtual power plant can be accurately adjusted, and the overall operation cost of the power plant is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of electric energy management, and in particular to a data-driven virtual power plant power trading optimization method and system. Background Art

[0002] As the concept of green development has gained popularity, clean energy such as solar energy and wind energy has gradually occupied an increasing proportion of energy. Unlike traditional thermal power plants, clean energy such as wind power occupies a large area and is difficult to form a more effective centralized management. Therefore, virtual power plants have emerged. The virtual power plant refers to a method of modeling equipment such as generators and energy storage equipment scattered in various places on a single platform, which is conducive to the centralized control of power equipment.

[0003] In recent years, the electricity futures market has developed, and electricity resources can be bought and sold more flexibly. Today, virtual power plants can identify the total power generation of power plants. However, there is no clear prediction for how to handle the total power generation. The more electricity is stored, the greater the cost of activating and maintaining the energy storage equipment. Unplanned storage of electricity may greatly increase the operating costs of virtual power plants. Summary of the invention

[0004] The present invention provides a data-driven virtual power plant power trading optimization method, the main purpose of which is to adjust the power transmission and storage allocation ratio of the virtual power plant and reduce the overall operating cost of the power plant.

[0005] To achieve the above-mentioned purpose, the present invention provides a data-driven virtual power plant power trading optimization method, comprising: obtaining real-time power prices, and obtaining a power price change curve within a preset time period; performing an extension operation on the power price change curve based on the curve law to obtain a primary power price prediction curve; querying a set of direct factors of power prices according to a pre-constructed power-related literature set; obtaining historical data corresponding to each direct factor in the direct factor set to obtain a direct factor historical data set, and performing a multivariate regression analysis on the power price change curve according to the direct factor historical data set to obtain an influence coefficient of each direct factor on the power price; calculating the real-time change rate corresponding to each direct factor in the direct factor set according to the direct factor historical data set, performing weighted calculation on each real-time change rate and each influence coefficient to obtain a comprehensive power change rate, and calculating the real-time power price according to the comprehensive power change rate to obtain an advanced power price prediction curve. power price forecast curve; use the weighted average of the advanced power price forecast curve and the primary power price forecast curve to obtain the power price forecast curve; obtain the total power generation forecast curve, and obtain the sold power curve, calculate the difference between the total power generation forecast curve and the sold power curve, and obtain the remaining power forecast curve; use the pre-trained power trading scheme generation model to perform a curve feature extraction operation on the power price forecast curve to obtain a predicted power price trend feature set, and perform a curve feature extraction operation on the remaining power forecast curve to obtain remaining power information; according to the pre-constructed trading incentive strategy and the predicted power price trend feature set, perform a binary classification judgment based on storage and sale on the remaining power information to obtain the power trading storage scheme with the highest incentive score, wherein the power trading storage scheme includes the highest incentive score, the distribution of power sold at each time point, and the number of pre-constructed power storage devices occupied when all power is sold.

[0006] Optionally, before using the pre-trained power trading scheme to generate the model, the method further includes: constructing an incentive score calculation formula according to a pre-constructed trading incentive strategy, wherein the incentive score calculation formula is expressed as: , where represents the incentive score, express The amount of electricity sold at a given time, express The predicted electricity price at the time, represents the initial electricity price, represents the curve before the highest price point in the electricity price forecast curve, , In the electricity price forecast curve A collection of high points, In the electricity price forecast curve The low point collection, represents the remaining power prediction curve, represents the number of the power storage devices in use, represents a single storage capacity of the power storage device, represents the remainder symbol, Indicates the amount of electricity stored in the power storage device that is not full. represents the utilization rate of the underfilled power storage device, , , and represents the reward and punishment intensity coefficient; according to the pre-constructed constraint condition, the maximum value of the incentive score calculation formula is solved to obtain the maximum incentive score formula, wherein the constraint condition is that when the time is at the highest point of the electricity price forecast curve, all the electricity in the pre-constructed power storage device is sold; the incentive score maximum formula is configured as the loss function of the pre-constructed power trading scheme generation network to obtain the power trading scheme generation model.

[0007] Optionally, the trading incentive strategy includes: obtaining an initialized cumulative score; within a preset trading time period, when executing an electricity selling action, obtaining the electricity quantity difference and the sold electricity quantity corresponding to the electricity selling action, multiplying the electricity quantity difference and the sold electricity quantity to obtain a trading score, and updating the cumulative score according to the trading score; when the number of power storage devices in use changes, obtaining newly added or disabled power change devices, and obtaining the device type corresponding to the power change device, obtaining a cost score based on the device type, and updating the cumulative score based on the cost score; obtaining the utilization rate of the power storage device, obtaining a waste score based on the utilization rate, and updating the cumulative score based on the waste score; when the trading time period ends, using the final cumulative score as the incentive score.

[0008] Optionally, after obtaining the electricity trading storage scheme with the highest incentive score, the method further includes: obtaining a real electricity change curve, and using the electricity trading storage scheme to perform electricity trading simulation calculations based on the trading incentive strategy on the real electricity change curve to obtain a real incentive score; calculating the ratio of the real incentive score to the highest incentive score to obtain a verification score; determining whether the verification score is less than a preset qualification threshold; when the verification score is less than the qualification threshold, determining that the electricity trading scheme generation model needs to be retrained.

[0009] Optionally, the electricity price change curve is extended based on the law of the curve to obtain a primary electricity price prediction curve, including: performing a curve feature extraction operation on the electricity price change curve to obtain historical electricity price features; using a pre-constructed long short-term memory network to perform a curve feature recognition operation based on the long and short term on the historical electricity price features to obtain short-term fluctuation features and long-term fluctuation features; calculating a curve change rate sequence based on the short-term fluctuation features and the long-term fluctuation features; using the curve change rate sequence to perform an electricity price prediction operation on the real-time electricity price to obtain a primary electricity price prediction curve.

[0010] Optionally, the method of querying the set of direct factors of electricity price based on a pre-constructed set of electricity-related literature includes: performing a big data query operation based on preset electricity keywords on the pre-constructed set of electricity-related literature to obtain electricity-related texts; performing text semantic analysis on the electricity-related texts to obtain a set of electricity influencing factors; using a pre-constructed clustering algorithm to perform weight calculation on each electricity influencing factor in the set of electricity influencing factors to obtain a weight sequence of electricity influencing factors, and performing a Top-N-based weight sorting operation on the weight sequence of electricity influencing factors to obtain a set of potential factors; using a pre-constructed data interface to obtain a historical data sequence of each potential factor in the potential factor set to obtain a potential factor historical data set, and using a pre-constructed Granger causality test algorithm to calculate the causal relationship score of each potential factor to the electricity price based on the potential factor historical data set and the electricity price change curve; and screening potential factors whose causal relationship scores are greater than a preset strong causal threshold to obtain a set of direct factors.

[0011] Optionally, the real-time change rate corresponding to each direct factor in the direct factor set is calculated based on the direct factor historical data set, including: identifying the time influence of each direct factor in the direct factor set to obtain the influence time of each direct factor, wherein the influence time includes units of day, week, month, season and year; configuring each influence time as a unit time corresponding to each direct factor; according to the unit time corresponding to each direct factor, obtaining the most recent historical data sequence of each direct factor from the direct factor historical data set; and calculating the real-time change rate corresponding to each direct factor based on the most recent historical data sequence of each direct factor.

[0012] Optionally, the real-time change rates and the influence coefficients are weightedly calculated to obtain a comprehensive power change rate, including: constructing a sub-predicted power change rate of each direct factor for the power price according to the real-time change rate and the influence coefficient of each direct factor; intercepting each sub-predicted power change rate by time length according to the influence time of each direct factor to obtain each effective sub-power change rate; and superimposing the curves of each effective sub-power change rate to obtain a comprehensive power change rate.

[0013] Optionally, the electricity price prediction curve is subjected to a curve feature extraction operation to obtain a predicted electricity price trend feature set, comprising: performing a convolution operation on the electricity price prediction curve to obtain a convolution matrix set, and performing feature dimension reduction on the convolution matrix set to obtain a curve feature sequence set; performing feature recognition on the curve feature sequence set to obtain a predicted electricity price trend feature set, wherein the predicted electricity price trend feature set includes the highest and lowest price points in the electricity price prediction curve, the time period in which the electricity price is greater than the real-time electricity price in the electricity price prediction curve, and the time period in which the electricity price is less than the real-time electricity price in the electricity price prediction curve.

[0014] To achieve the above-mentioned purpose, the present invention also provides a data-driven virtual power plant power trading optimization system, including: a primary power price prediction module, used to obtain real-time power prices, and obtain power price change curves within a preset time period, and extend the power price change curve based on the curve law to obtain a primary power price prediction curve; a causal relationship price prediction module, used to query a set of direct factors of power prices based on a pre-constructed power-related literature set, and obtain historical data corresponding to each direct factor in the direct factor set to obtain a direct factor historical data set, and perform a multivariate regression analysis on the power price change curve based on the direct factor historical data set to obtain the influence coefficient of each direct factor on the power price, and calculate the real-time change rate corresponding to each direct factor in the direct factor set based on the direct factor historical data set, weightedly calculate each real-time change rate and each influence coefficient to obtain a comprehensive power change rate, and calculate the real-time change rate based on the comprehensive power change rate. The power change rate calculates the real-time power price to obtain an advanced power price forecast curve; a power generation surplus identification module is used to obtain a total power generation forecast curve, and obtain a sold power curve, calculate the difference between the total power generation forecast curve and the sold power curve, and obtain a remaining power forecast curve; a transaction plan generation module is used to use a pre-trained power transaction plan generation model to perform a curve feature extraction operation on the power price forecast curve to obtain a predicted power price trend feature set, perform a curve feature extraction operation on the remaining power forecast curve to obtain remaining power information, and perform a binary classification judgment based on storage and selling on the remaining power information according to a pre-constructed transaction incentive strategy and the predicted power price trend feature set to obtain a power transaction storage plan with the highest incentive score, wherein the power transaction storage plan includes the highest incentive score, the distribution of power sold at each time point, and the number of pre-constructed power storage devices occupied when all power is sold.

[0015] In order to solve the above problems, the present invention also provides an electronic device, which includes: a memory storing at least one instruction; and a processor executing the instructions stored in the memory to implement the above-mentioned data-driven virtual power plant electricity trading optimization method.

[0016] In order to solve the above problems, the present invention also provides a computer-readable storage medium, in which at least one instruction is stored, and the at least one instruction is executed by a processor in an electronic device to implement the above-mentioned data-driven virtual power plant electricity trading optimization method.

[0017] The present invention is to solve the problem described in the background technology. First, the present invention performs a prediction operation based on the curve law according to the electricity price change curve to obtain a primary electricity price prediction curve, and the primary electricity price prediction curve can preliminarily predict the general trend of electricity prices; further, the direct factors of electricity prices are queried through big data identification, and the influence coefficients of each direct factor on the electricity price are obtained through causal analysis, so as to obtain a high-level electricity price prediction curve by weighting, wherein the electricity price prediction curve is composed of polynomials based on each direct factor, and when each direct factor mutates, the electricity price will show changes, wherein the high-level electricity price prediction curve is a detail change, and the primary electricity price prediction curve is a general trend. The present invention performs weighted averaging on the high-level electricity price prediction curve and the primary electricity price prediction curve to obtain the electricity price prediction curve, which can better balance the law of electricity changes; then the remaining power prediction curve is obtained, and the pre-trained power transaction plan generation model is used to generate a power transaction storage plan, wherein the loss function of the power transaction plan generation model is a pre-constructed transaction incentive strategy, which can calculate the most reasonable power transaction storage plan and improve the plan generation efficiency. Therefore, the present invention can accurately adjust the power transmission and storage allocation ratio of the virtual power plant and reduce the overall operating cost of the power plant. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A schematic diagram of a flow chart of a data-driven virtual power plant power trading optimization method provided in one embodiment of the present invention; Figure 2 A functional module diagram of a data-driven virtual power plant power trading optimization system provided by an embodiment of the present invention; Figure 3 A schematic diagram of the structure of an electronic device for implementing the data-driven virtual power plant power trading optimization method provided in one embodiment of the present invention.

[0019] Explanation of the reference numerals: 1. electronic device; 10. processor; 11. memory; 12. bus.

[0020] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0021] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0022] The embodiment of the present application provides a data-driven virtual power plant power trading optimization method. The execution subject of the data-driven virtual power plant power trading optimization method includes but is not limited to at least one of the electronic devices such as a server and a terminal that can be configured to execute the method provided by the embodiment of the present application. In other words, the data-driven virtual power plant power trading optimization method can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc.

[0023] Reference Figure 1 As shown, it is a flow chart of a data-driven virtual power plant power trading optimization method provided by an embodiment of the present invention. In this embodiment, the data-driven virtual power plant power trading optimization method includes: S1, obtaining real-time power prices, and obtaining a power price change curve within a preset time period.

[0024] In an embodiment of the present invention, in order to predict future electricity prices, it is necessary to first record the real-time electricity prices of a target area, such as an industrial area, city or province, based on a preset time period, such as a week or a month, to obtain an electricity price change curve, wherein the last set of electricity data of the electricity price change curve is the real-time electricity price.

[0025] In addition, the real-time electricity price can also be used as the initial electricity price in the subsequent electricity prediction process.

[0026] S2. Perform an extension operation on the electricity price change curve based on the curve law to obtain a primary electricity price prediction curve.

[0027] In detail, in an embodiment of the present invention, the electricity price change curve is extended based on the curve law to obtain a primary electricity price prediction curve, including: performing a curve feature extraction operation on the electricity price change curve to obtain historical electricity price features; using a pre-constructed long short-term memory network to perform a curve feature recognition operation based on the long and short term on the historical electricity price features to obtain short-term fluctuation features and long-term fluctuation features; calculating a curve change rate sequence based on the short-term fluctuation features and the long-term fluctuation features; and performing an electricity price prediction operation on the real-time electricity price using the curve change rate sequence to obtain a primary electricity price prediction curve.

[0028] Among them, the long short-term memory network (LSTM) is a special type of recurrent neural network (RNN) that can learn and memorize long-term information and is suitable for capturing long-term dependencies in time series data.

[0029] Specifically, in the embodiment of the present invention, the curve feature extraction is performed through a pre-constructed neural network to obtain the historical electricity price feature, wherein the historical electricity price feature is obtained by convolution calculation of different regions of the electricity price change curve. Then, the long short-term memory network is used to perform feature recognition operations based on different extraction fields and calculation scales to obtain short-term fluctuation features and long-term fluctuation features, which include the initial electricity price of the curve, the rising time period, the falling time period, the turning points of the highest point and the lowest point, and other mathematical features of the curve.

[0030] According to the short-term fluctuation characteristics and the long-term fluctuation characteristics, the embodiment of the present invention can calculate the curve change amount in each time period, thereby obtaining a curve change rate sequence, for example, the price decreases every 12-hour time period, and then increases steadily to form a long cycle, and in each long cycle, the price fluctuates 30 times. According to the curve change rate sequence, the initial electricity price can be calculated to obtain a primary electricity price prediction curve.

[0031] S3. According to the pre-built electricity-related literature set, query the set of direct factors of electricity price.

[0032] In the embodiment of the present invention, the power-related document collection may be collected and constructed by the enterprise itself, or a plurality of enterprises may construct an alliance database, or the documents may be retrieved from various websites and data interfaces.

[0033] Specifically, in an embodiment of the present invention, a virtual power plant enterprise needs to collect electricity-related news information, electricity-related research literature, government electricity-related policies or notifications, natural disasters and environmental change information, and obtain a collection of electricity-related literature.

[0034] In detail, in an embodiment of the present invention, the method of querying a set of direct factors of electricity price based on a pre-constructed set of electricity-related literature includes: performing a big data query operation based on preset electricity keywords on the pre-constructed set of electricity-related literature to obtain electricity-related texts; performing text semantic analysis on the electricity-related texts to obtain a set of electricity influencing factors; using a pre-constructed clustering algorithm to perform weight calculation on each electricity influencing factor in the set of electricity influencing factors to obtain a weight sequence of electricity influencing factors, and performing a Top-N-based weight sorting operation on the weight sequence of electricity influencing factors to obtain a set of potential factors; using a pre-constructed data interface to obtain a historical data sequence of each potential factor in the set of potential factors to obtain a potential factor historical data set, and using a pre-constructed Granger causality test algorithm to calculate the causal relationship score of each potential factor to the electricity price based on the potential factor historical data set and the electricity price change curve; screening potential factors whose causal relationship scores are greater than a preset strong causal threshold to obtain a set of direct factors.

[0035] Among them, the preset electricity keywords include: electricity, power generation technology, electricity cost, electricity price, electricity policy planning, electricity price increase, electricity storage, clean energy, peak electricity consumption, electricity consumption benefits, factors affecting electricity prices, etc.

[0036] The clustering algorithm refers to an algorithm for keyword classification based on similarity calculation, which is used to summarize similar words, such as power plant operating costs and surplus power storage costs, and calculate the magnitude of the data.

[0037] Furthermore, the Granger Causality Test algorithm is a statistical method used to determine whether a time series can be used to predict another time series. The basic idea of ​​the test is that if the past value of one variable has a significant impact on the current value of another variable, it can be considered that the former has a "causal relationship" with the latter.

[0038] Specifically, in an embodiment of the present invention, firstly, a big data power keyword query is performed through a pre-built data engine to obtain power-related texts, such as articles, news paragraphs, etc., to achieve dimensionality reduction of a large amount of data. Then, a text semantic analysis is performed on the power-related text through a pre-built text semantic recognition network to obtain a set of power influencing factors, wherein the power influencing factor set is [power demand, power supply, market price changes, temperature, humidity, precipitation, wind speed, GDP, industrial output value, unemployment rate, power industry policy, subsidies, carbon tax, proportion of renewable energy, progress in power generation technology...]. Then, each similar word is aggregated together according to a clustering algorithm, and the proportion of word occurrence is statistically divided by weight to obtain a weight sequence of power influencing factors, such as [economic factors 10%, market factors 40%, policy factors 10%, meteorological factors 20%...]. Then, through a sorting operation, a potential factor set [market factors 40%, meteorological factors 20%, economic factors 10%, policy factors 10%...] is obtained. Among them, each proportion in the potential factor set is affected by the number of text occurrences, so it cannot be used as a basis for the degree of influence of each potential factor on the power price for the time being.

[0039] In an embodiment of the present invention, after obtaining the potential factor set, the historical data of each potential factor in the potential factor set is obtained through a pre-constructed data interface to obtain the historical data sequence of each potential factor, and the various historical data sequences are integrated together as the potential factor historical data set.

[0040] Furthermore, the present invention identifies causal relationships on a set of potential factors according to the Granger causality test algorithm, obtains a causal relationship score for each potential factor, extracts potential factors with obvious causal relationships, i.e., potential factors with causal relationship scores greater than a preset strong causal threshold, and obtains a direct factor set [market factors, meteorological factors, economic factors, policy factors].

[0041] S4. Obtain historical data corresponding to each direct factor in the direct factor set to obtain a direct factor historical data set, and perform a multivariate regression analysis on the electricity price change curve based on the direct factor historical data set to obtain the influence coefficient of each direct factor on the electricity price.

[0042] Specifically, in an embodiment of the present invention, historical data corresponding to each direct factor is obtained through a pre-built data interface, such as statistical data released by government and industry organizations, meteorological data providers (such as NOAA, weather service agencies, etc.), economic indicators and databases and APIs released by economic statistics agencies (such as electricity market trading platforms, financial market data services, etc.).

[0043] Furthermore, the present invention performs a multiple regression analysis on the electricity price change curve according to the direct factor historical data set to obtain the influence coefficient of each direct factor on the electricity price. The multiple regression analysis is a statistical analysis method used to study the relationship between a dependent variable (explained variable) and two or more independent variables (explanatory variables). Through this analysis, the influence of the independent variable on the dependent variable can be evaluated and the value of the dependent variable can be predicted.

[0044] Specifically, in an embodiment of the present invention, market factors, meteorological factors, economic factors, and policy factors are used as independent variables, and electricity prices are used as dependent variables for analysis to obtain the influence coefficient of each direct factor on the electricity price, for example [market factors: 20%; meteorological factors: 20%; economic factors: 20%; policy factors: 30%].

[0045] S5. According to the direct factor historical data set, calculate the real-time change rate corresponding to each direct factor in the direct factor set, perform weighted calculation on each real-time change rate and each influence coefficient to obtain a comprehensive power change rate, and calculate the real-time power price according to the comprehensive power change rate to obtain an advanced power price prediction curve.

[0046] It should be known that the time influence of each direct factor is different. For example, meteorological information can only affect the electricity price in the next day to a week, while policy factors may have a greater impact on a region in different quarters or years. Therefore, it is necessary to calculate the real-time change rate for each direct factor separately.

[0047] In detail, in an embodiment of the present invention, the real-time change rate corresponding to each direct factor in the direct factor set is calculated based on the direct factor historical data set, including: identifying the time influence of each direct factor in the direct factor set to obtain the influence time of each direct factor, wherein the influence time includes units of day, week, month, season and year; configuring each influence time as a unit time corresponding to each direct factor; according to the unit time corresponding to each direct factor, obtaining the most recent historical data sequence of each direct factor from the direct factor historical data set; and calculating the real-time change rate corresponding to each direct factor based on the most recent historical data sequence of each direct factor.

[0048] Specifically, in an embodiment of the present invention, the update time period of each direct factor in life is used as the impact time of each direct factor. For example, meteorological factors are updated daily, so that the impact time is day, then the unit time is day. The change amount of today's and yesterday's meteorological information is predicted to obtain the real-time change rate of the meteorological factors. The market factors are updated every quarter through the market transaction summary work report. The impact time is quarter, then the unit time is also quarter. The market transaction historical data series of this quarter and the previous quarter are obtained, and the real-time change rate of the market factors is calculated.

[0049] In detail, in an embodiment of the present invention, the weighted calculation of each real-time change rate and each influence coefficient is performed to obtain a comprehensive power change rate, including: constructing a sub-predicted power change rate of each direct factor for the power price according to the real-time change rate and the influence coefficient of each direct factor; according to the influence time of each direct factor, intercepting each sub-predicted power change rate by time length to obtain each effective sub-power change rate; and superimposing the curves of each effective sub-power change rate to obtain a comprehensive power change rate.

[0050] In the embodiment of the present invention, since the influence coefficients of various direct factors on the electricity price are different, the sub-predicted electricity change rates of various direct factors on the electricity price are constructed through the real-time change rates and influence coefficients of various direct factors. Since the influence times of the real-time change rates corresponding to various direct factors are different, the time lengths of various sub-predicted electricity change rates are intercepted through the influence time to obtain various effective sub-power change rates, and finally the comprehensive electricity change rate is obtained by superimposing the corresponding numerical values ​​of the curves.

[0051] Furthermore, the present invention calculates the real-time electricity price according to the comprehensive electricity change to obtain an advanced electricity price prediction curve.

[0052] S6. Perform weighted averaging of the advanced electricity price prediction curve and the primary electricity price prediction curve to obtain an electricity price prediction curve.

[0053] In an embodiment of the present invention, the advanced electricity price prediction curve is the change inferred from sub-factors such as market, weather, and policy, and the primary electricity price prediction curve is the change inferred based on the mathematical law of the curve. The present invention performs weighted averaging of the advanced electricity price prediction curve and the primary electricity price prediction curve to obtain an electricity price prediction curve, which can better balance the law of electricity changes.

[0054] S7. Obtain a total power generation prediction curve and a sold power generation curve, calculate the difference between the total power generation prediction curve and the sold power generation curve, and obtain a remaining power generation prediction curve.

[0055] It should be known that the virtual power plant can record the values ​​of each wind power and photovoltaic power generation unit, and by monitoring the wind speed and light intensity of the environment, it can predict the total power generation forecast curve.

[0056] Furthermore, the sold electricity curve refers to the electricity that has been pre-sold through orders, and this part of electricity cannot be used elsewhere.

[0057] The present invention can obtain a remaining power prediction curve by calculating the difference between the total power generation prediction curve and the sold power curve.

[0058] S8. Using the pre-trained electricity trading scheme generation model, perform a curve feature extraction operation on the electricity price prediction curve to obtain a predicted electricity price trend feature set, and perform a curve feature extraction operation on the remaining power prediction curve to obtain remaining power information.

[0059] In an embodiment of the present invention, the electricity trading plan generation model is a neural network model based on image (curve image) recognition, which is used to arrange trading plans according to the remaining power prediction curve and the electricity price prediction curve, wherein the trading plan is a time-ordered sequence, for example, selling 10,000 kWh of electricity at 10 o'clock and selling 5,000 kWh of electricity at 11 o'clock.

[0060] In detail, in an embodiment of the present invention, before using the pre-trained power trading scheme to generate the model, the method further includes: constructing an incentive score calculation formula according to a pre-constructed trading incentive strategy, wherein the incentive score calculation formula is expressed as: , where represents the incentive score, express The amount of electricity sold at a given time, express The predicted electricity price at the time, represents the initial electricity price, represents the curve before the highest price point in the electricity price forecast curve, , In the electricity price forecast curve A collection of high points, In the electricity price forecast curve The low point collection, represents the remaining power prediction curve, represents the number of the power storage devices in use, represents a single storage capacity of the power storage device, represents the remainder symbol, Indicates the amount of electricity stored in the power storage device that is not full. represents the utilization rate of the underfilled power storage device, , , and represents the reward and punishment intensity coefficient; according to the pre-constructed constraint condition, the maximum value of the incentive score calculation formula is solved to obtain the maximum incentive score formula, wherein the constraint condition is that when the time is at the highest point of the electricity price forecast curve, all the electricity in the pre-constructed power storage device is sold; the incentive score maximum formula is configured as the loss function of the pre-constructed power trading scheme generation network to obtain the power trading scheme generation model.

[0061] Specifically, in an embodiment of the present invention, the power trading scheme generation model includes two parts, a network for extracting and identifying curve features, and a network for calculating the maximum incentive score, wherein the network for calculating the maximum incentive score can be used as the loss function of the network for extracting and identifying curve features.

[0062] Furthermore, the trading incentive strategy includes scores for common actions in power trading.

[0063] In detail, in an embodiment of the present invention, the trading incentive strategy includes: obtaining an initialized cumulative score; within a preset trading time period, when executing an electricity selling action, obtaining the electricity quantity difference and the sold electricity quantity corresponding to the electricity selling action, multiplying the electricity quantity difference and the sold electricity quantity to obtain a trading score, and updating the cumulative score according to the trading score; when the number of used power storage devices changes, obtaining newly added or disabled power change devices, and obtaining the device type corresponding to the power change device, obtaining a cost score based on the device type, and updating the cumulative score based on the cost score; obtaining the utilization rate of the power storage device, obtaining a waste score based on the utilization rate, and updating the cumulative score based on the waste score; when the trading time period ends, using the final cumulative score as the incentive score.

[0064] Specifically, in the embodiment of the present invention, by The difference between the electricity price when selling electricity and the initial electricity price is obtained by calculating the electricity sales volume. Multiply the product to get the revenue from selling electricity, then Reward and punishment intensity coefficient ,get : In the embodiment of the present invention, according to the constraint condition that all the electricity in the pre-built power storage device is sold when the time is at the highest point of the power price prediction curve, it can be obtained ,in, The total remaining power in the power storage device.

[0065] In addition, considering the maintenance and activation costs of power storage equipment, In the present embodiment, only the selling of electricity is considered, but the buying at a low price is not considered. The item may indicate that new power storage equipment should not be started unless necessary.

[0066] Specifically, when executing an electricity selling action, the present invention will obtain the electricity quantity difference and the sold electricity quantity corresponding to the electricity selling action. Since the electricity price when selling electricity may be higher or lower, the electricity quantity difference may be positive or negative, resulting in the transaction score may be positive or negative.

[0067] In the embodiment of the present invention, the cumulative score is updated according to the cost score: when the power change equipment is increased, indicating that the operating cost is increased, the cumulative score is subtracted; conversely, when the power change equipment is disabled, the cumulative score is added.

[0068] In the embodiment of the present invention, the waste score is used to update the accumulated score to control the change of the power equipment, so that when the equipment utilization rate is low, for example, below a fixed threshold of 20%, the power storage equipment will not be added. The specific value of the fixed threshold can be based on Modify the relationship between them.

[0069] In the embodiment of the present invention, the cumulative score is initialized, and then each change of the cumulative score at the end of the trading time period is recorded and integrated, so that the incentive score calculation formula can be obtained. When the incentive score calculation formula takes the maximum value, it indicates that the power selling action, the change in the number of power storage devices in use, and the utilization rate of the power storage devices all have better solutions.

[0070] Furthermore, after the power trading scheme generation model is constructed, although the loss function of the present invention is relatively unique, considering that the model structure is relatively simple, it can be put into use by obtaining real data samples for simple training.

[0071] In detail, in an embodiment of the present invention, the curve feature extraction operation is performed on the electricity price prediction curve to obtain a predicted electricity price trend feature set, including: performing a convolution operation on the electricity price prediction curve to obtain a convolution matrix set, and performing feature dimensionality reduction on the convolution matrix set to obtain a curve feature sequence set; performing feature recognition on the curve feature sequence set to obtain a predicted electricity price trend feature set, wherein the predicted electricity price trend feature set includes the highest and lowest price points in the electricity price prediction curve, the time period when the electricity price in the electricity price prediction curve is greater than the real-time electricity price, and the time period when the electricity price in the electricity price prediction curve is less than the real-time electricity price.

[0072] The power trading scheme generation model completed by training in the present invention can perform feature extraction and dimensionality reduction operations through the convolution layer, pooling layer and flattening layer in the neural network respectively to obtain a curve feature sequence set, and then perform feature recognition on the remaining power prediction curve to obtain remaining power information.

[0073] S9. According to the pre-constructed trading incentive strategy and the predicted electricity price trend feature set, the remaining electricity information is subjected to a binary classification judgment based on storage and selling to obtain an electricity trading storage plan with the highest incentive score, wherein the electricity trading storage plan includes the highest incentive score, the distribution of electricity sold at each time point, and the number of pre-constructed electricity storage devices occupied when all electricity is sold.

[0074] In an embodiment of the present invention, a binary classification judgment based on storage and sale is performed on the remaining power information according to the curve feature set, and finally a power trading sequence is obtained. Then, an incentive score is calculated for the power trading sequence according to the trading incentive strategy, and the incentive score is maximized through the loss function to obtain a power trading storage scheme with the highest incentive score.

[0075] The power trading storage scheme includes the highest incentive score, the distribution of power sold at each time point, and the number of pre-built power storage devices used when all power is sold. The distribution of the electricity sold at each time point is: sequence, the number of pre-built power storage devices used when all electricity is sold is .

[0076] Furthermore, in an embodiment of the present invention, after obtaining the electricity trading storage scheme with the highest incentive score, the method also includes: obtaining a real electricity change curve, using the electricity trading storage scheme to perform electricity trading simulation calculations based on the trading incentive strategy on the real electricity change curve to obtain a real incentive score; calculating the ratio of the real incentive score to the highest incentive score to obtain a verification score; determining whether the verification score is less than a preset qualified threshold; when the verification score is less than the qualified threshold, determining that the electricity trading scheme generation model needs to be retrained.

[0077] In the embodiment of the present invention, by predicting the electricity price forecast curve through the above process, the trend of domestic electricity prices in the future can be understood, and an electricity trading storage plan is formulated through the electricity price forecast curve. However, whether the electricity trading storage plan is suitable still needs to be verified.

[0078] Specifically, the present invention obtains the real power change curve, and then calculates the incentive score according to the power trading storage plan to obtain the real incentive score. If the real incentive score can reach the preset qualified threshold, such as 60%, it indicates that the power trading storage plan is qualified. Otherwise, the model needs to be retrained. By updating at any time, it is beneficial to maintain the reliability of the power trading plan generation model for a long time.

[0079] The present invention is to solve the problem described in the background technology. First, the present invention performs a prediction operation based on the curve law according to the electricity price change curve to obtain a primary electricity price prediction curve, and the primary electricity price prediction curve can preliminarily predict the general trend of electricity prices; further, the direct factors of electricity prices are queried through big data identification, and the influence coefficients of each direct factor on the electricity price are obtained through causal analysis, so as to obtain a high-level electricity price prediction curve by weighting, wherein the electricity price prediction curve is composed of polynomials based on each direct factor, and when each direct factor mutates, the electricity price will show changes, wherein the high-level electricity price prediction curve is a detail change, and the primary electricity price prediction curve is a general trend. The present invention performs weighted averaging on the high-level electricity price prediction curve and the primary electricity price prediction curve to obtain the electricity price prediction curve, which can better balance the law of electricity changes; then the remaining power prediction curve is obtained, and the pre-trained power transaction plan generation model is used to generate a power transaction storage plan, wherein the loss function of the power transaction plan generation model is a pre-constructed transaction incentive strategy, which can calculate the most reasonable power transaction storage plan and improve the plan generation efficiency. Therefore, the present invention can accurately adjust the power transmission and storage allocation ratio of the virtual power plant and reduce the overall operating cost of the power plant.

[0080] like Figure 2 As shown, it is a functional module diagram of a data-driven virtual power plant power trading optimization system provided by one embodiment of the present invention.

[0081] The data-driven virtual power plant power trading optimization system 100 of the present invention can be installed in an electronic device. According to the functions implemented, the data-driven virtual power plant power trading optimization system 100 can include a primary power price prediction module 101, a causal price prediction module 102, a power generation surplus identification module 103 and a trading scheme generation module 104. The module of the present invention can also be called a unit, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.

[0082] The primary electricity price prediction module 101 is used to obtain the real-time electricity price, obtain the electricity price change curve within a preset time period, and extend the electricity price change curve based on the curve law to obtain the primary electricity price prediction curve; the causal relationship price prediction module 102 is used to query the direct factor set of electricity price according to the pre-constructed electricity-related literature set, and obtain the historical data corresponding to each direct factor in the direct factor set to obtain the direct factor historical data set, and perform a multivariate regression analysis on the electricity price change curve according to the direct factor historical data set to obtain the influence coefficient of each direct factor on the electricity price, and calculate the real-time change rate corresponding to each direct factor in the direct factor set according to the direct factor historical data set, perform weighted calculation on each real-time change rate and each influence coefficient to obtain a comprehensive electricity change rate, and calculate the real-time electricity price according to the comprehensive electricity change rate to obtain to the advanced electricity price forecast curve; the power generation surplus identification module 103 is used to obtain the total power generation forecast curve, and obtain the sold power curve, calculate the difference between the total power generation forecast curve and the sold power curve, and obtain the surplus power forecast curve; the transaction plan generation module 104 is used to use the pre-trained power transaction plan generation model to perform a curve feature extraction operation on the power price forecast curve to obtain a predicted power price trend feature set, perform a curve feature extraction operation on the surplus power forecast curve to obtain surplus power information, and perform a binary classification judgment based on storage and selling on the surplus power information according to the pre-constructed transaction incentive strategy and the predicted power price trend feature set to obtain the power transaction storage plan with the highest incentive score, wherein the power transaction storage plan includes the highest incentive score, the distribution of power sold at each time point, and the number of pre-constructed power storage devices occupied when all power is sold.

[0083] In detail, each module in the data-driven virtual power plant power trading optimization system 100 in the embodiment of the present invention is used in the same manner as described above. Figure 1 The technical means are the same as the data-driven virtual power plant power trading optimization method described in, and can produce the same technical effects, so I will not go into details here.

[0084] like Figure 3 ] The figure is a schematic diagram of the structure of an electronic device for implementing a data-driven virtual power plant power trading optimization method provided by an embodiment of the present invention.

[0085] The electronic device 1 may include a processor 10, a memory 11 and a bus 12, and may also include a computer program stored in the memory 11 and executable on the processor 10, such as a data-driven virtual power plant electricity trading optimization method program.

[0086] The memory 11 includes at least one type of readable storage medium, and the readable storage medium includes a flash memory, a mobile hard disk, a multimedia card, a card-type memory (e.g., SD or DX memory, etc.), a magnetic memory, a disk, an optical disk, etc. In some embodiments, the memory 11 may be an internal storage unit of the electronic device 1, such as a mobile hard disk of the electronic device 1. In other embodiments, the memory 11 may also be an external storage device of the electronic device 1, such as a plug-in mobile hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device 1. Further, the memory 11 also includes an internal storage unit of the electronic device 1 and an external storage device. The memory 11 can not only be used to store application software and various types of data installed in the electronic device 1, such as the code of the data-driven virtual power plant power trading optimization method program, but also can be used to temporarily store data that has been output or is to be output.

[0087] In some embodiments, the processor 10 may be composed of an integrated circuit, for example, a single packaged integrated circuit, or a plurality of packaged integrated circuits with the same or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and combinations of various control chips, etc. The processor 10 is the control core (Control Unit) of the electronic device, and uses various interfaces and lines to connect various components of the entire electronic device, and executes or executes programs or modules (such as data-driven virtual power plant power trading optimization method programs, etc.) stored in the memory 11, and calls data stored in the memory 11 to execute various functions of the electronic device 1 and process data.

[0088] The bus 12 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus 12 may be divided into an address bus, a data bus, a control bus, etc. The bus 12 is configured to realize connection and communication between the memory 11 and at least one processor 10, etc.

[0089] Figure 3 Only an electronic device with components is shown, and those skilled in the art will understand that Figure 3The structure shown does not constitute a limitation on the electronic device 1, and may include fewer or more components than shown in the figure, or combine certain components, or arrange the components differently.

[0090] For example, although not shown, the electronic device 1 may also include a power source (such as a battery) for supplying power to various components. Preferably, the power source may be logically connected to the at least one processor 10 through a power management device, so that the power management device can realize functions such as charging management, discharging management, and power consumption management. The power source may also include one or more DC or AC power sources, recharging devices, power failure detection circuits, power converters or inverters, power status indicators, and other arbitrary components. The electronic device 1 may also include a variety of sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be repeated here.

[0091] Furthermore, the electronic device 1 may also include a network interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), which is generally used to establish a communication connection between the electronic device 1 and other electronic devices.

[0092] Optionally, the electronic device 1 may further include a user interface, which may be a display, an input unit (such as a keyboard), or a standard wired interface or a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, and an OLED (Organic Light-Emitting Diode) touch device. The display may also be appropriately referred to as a display screen or a display unit, which is used to display information processed in the electronic device 1 and to display a visual user interface.

[0093] The data-driven virtual power plant power trading optimization method program stored in the memory 11 in the electronic device 1 is a combination of multiple instructions. When running in the processor 10, it can achieve: obtaining real-time electricity prices, and obtaining the electricity price change curve within a preset time period; extending the electricity price change curve based on the curve law to obtain a primary electricity price prediction curve; querying the direct factor set of electricity prices according to a pre-constructed electricity-related literature set; obtaining the historical data corresponding to each direct factor in the direct factor set to obtain a direct factor historical data set, and performing a multivariate regression analysis on the electricity price change curve based on the direct factor historical data set to obtain the influence coefficient of each direct factor on the electricity price; calculating the real-time change rate corresponding to each direct factor in the direct factor set based on the direct factor historical data set, weighting each real-time change rate and each influence coefficient to obtain a comprehensive electricity change rate, and The power price is calculated to obtain an advanced power price forecast curve; the advanced power price forecast curve and the primary power price forecast curve are weighted averaged to obtain the power price forecast curve; the total power generation forecast curve and the sold power curve are obtained, and the difference between the total power generation forecast curve and the sold power curve is calculated to obtain the remaining power forecast curve; the pre-trained power trading scheme generation model is used to perform a curve feature extraction operation on the power price forecast curve to obtain a predicted power price trend feature set, and the curve feature extraction operation is performed on the remaining power forecast curve to obtain the remaining power information; according to the pre-constructed trading incentive strategy and the predicted power price trend feature set, the remaining power information is subjected to a binary classification judgment based on storage and sale to obtain the power trading storage scheme with the highest incentive score, wherein the power trading storage scheme includes the highest incentive score, the distribution of power sold at each time point, and the number of pre-constructed power storage devices occupied when all power is sold.

[0094] Specifically, the specific implementation method of the processor 10 for the above instructions can refer to Figures 1 to 3 The description of the relevant steps in the corresponding embodiments will not be repeated here.

[0095] Furthermore, if the module / unit integrated in the electronic device 1 is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a mobile hard disk, a magnetic disk, an optical disk, a computer memory, and a read-only memory (ROM).

[0096] The present invention also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor of an electronic device, it can achieve the following: obtaining real-time electricity prices and obtaining an electricity price change curve within a preset time period; performing an extension operation on the electricity price change curve based on the curve law to obtain a primary electricity price prediction curve; querying a set of direct factors of electricity prices based on a pre-constructed electricity-related literature set; obtaining historical data corresponding to each direct factor in the direct factor set to obtain a direct factor historical data set, and performing a multivariate regression analysis on the electricity price change curve based on the direct factor historical data set to obtain an influence coefficient of each direct factor on the electricity price; calculating the real-time change rate corresponding to each direct factor in the direct factor set based on the direct factor historical data set, performing weighted calculation on each real-time change rate and each influence coefficient to obtain a comprehensive electricity change rate, and predicting the real-time electricity price based on the comprehensive electricity change rate. The method comprises the following steps: calculating the grid to obtain an advanced electricity price forecast curve; performing weighted average of the advanced electricity price forecast curve and the primary electricity price forecast curve to obtain an electricity price forecast curve; obtaining a total power generation forecast curve and a sold power curve, calculating the difference between the total power generation forecast curve and the sold power curve to obtain a remaining power forecast curve; using a pre-trained power trading scheme generation model, performing a curve feature extraction operation on the power price forecast curve to obtain a predicted power price trend feature set, performing a curve feature extraction operation on the remaining power forecast curve to obtain remaining power information; performing a binary classification judgment based on storage and sale on the remaining power information according to a pre-constructed trading incentive strategy and the predicted power price trend feature set to obtain an electricity trading storage scheme with the highest incentive score, wherein the electricity trading storage scheme includes the highest incentive score, the distribution of the power sold at each time point, and the number of pre-constructed power storage devices occupied when all the power is sold.

[0097] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, systems and methods can be implemented in other ways. For example, the system embodiments described above are only illustrative, and actual implementation may have other division methods.

[0098] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0099] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0100] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0101] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A data-driven virtual power plant power trading optimization method, characterized in that: The method comprises: obtaining real-time electricity prices, and obtaining an electricity price variation curve within a preset time period; performing an extension operation on the electricity price variation curve based on the curve law to obtain a primary electricity price prediction curve; querying a set of direct factors of electricity prices according to a pre-constructed electricity-related literature set; obtaining historical data corresponding to each direct factor in the direct factor set to obtain a direct factor historical data set, and performing a multivariate regression analysis on the electricity price variation curve according to the direct factor historical data set to obtain an influence coefficient of each direct factor on the electricity price; calculating a real-time change rate corresponding to each direct factor in the direct factor set according to the direct factor historical data set, performing a weighted calculation on each real-time change rate and each influence coefficient to obtain a comprehensive electricity change rate, and calculating the real-time electricity price according to the comprehensive electricity change rate to obtain an advanced electricity price prediction curve; using the advanced electricity The price forecast curve is weighted averaged with the primary electricity price forecast curve to obtain the electricity price forecast curve; the total power generation forecast curve and the sold power curve are obtained, and the difference between the total power generation forecast curve and the sold power curve is calculated to obtain the remaining power forecast curve; the pre-trained power trading scheme generation model is used to perform a curve feature extraction operation on the power price forecast curve to obtain a predicted power price trend feature set, and the curve feature extraction operation is performed on the remaining power forecast curve to obtain the remaining power information; according to the pre-constructed trading incentive strategy and the predicted power price trend feature set, the remaining power information is subjected to a binary classification judgment based on storage and sale to obtain the power trading storage scheme with the highest incentive score, wherein the power trading storage scheme includes the highest incentive score, the distribution of power sold at each time point, and the number of pre-constructed power storage devices occupied when all power is sold.

2. The data-driven virtual power plant power trading optimization method according to claim 1, characterized in that: Before using the pre-trained power trading scheme to generate the model, the method further includes: constructing an incentive score calculation formula according to a pre-constructed trading incentive strategy, wherein the incentive score calculation formula is expressed as: , where represents the incentive score, express The amount of electricity sold at a given time, express The predicted electricity price at the time, represents the initial electricity price, represents the curve before the highest price point in the electricity price forecast curve, , In the electricity price forecast curve A collection of high points, In the electricity price forecast curve The low point collection, represents the remaining power prediction curve, represents the number of the power storage devices in use, represents a single storage capacity of the power storage device, represents the remainder symbol, Indicates the amount of electricity stored in the power storage device that is not full. represents the utilization rate of the underfilled power storage device, , , and represents the reward and punishment intensity coefficient; according to the pre-constructed constraint condition, the maximum value of the incentive score calculation formula is solved to obtain the maximum incentive score formula, wherein the constraint condition is that when the time is at the highest point of the electricity price forecast curve, all the electricity in the pre-constructed power storage device is sold; the incentive score maximum formula is configured as the loss function of the pre-constructed power trading scheme generation network to obtain the power trading scheme generation model.

3. The data-driven virtual power plant power trading optimization method according to claim 2, characterized in that: The trading incentive strategy includes: obtaining an initialized cumulative score; within a preset trading time period, when executing an electricity selling action, obtaining the power difference and the sold power corresponding to the electricity selling action, multiplying the power difference and the sold power to obtain a trading score, and updating the cumulative score according to the trading score; when the number of power storage devices in use changes, obtaining newly added or disabled power change devices and the device type corresponding to the power change device, obtaining a cost score based on the device type, and updating the cumulative score based on the cost score; obtaining the utilization rate of the power storage device, obtaining a waste score based on the utilization rate, and updating the cumulative score based on the waste score; when the trading time period ends, using the final cumulative score as the incentive score.

4. The data-driven virtual power plant power trading optimization method according to claim 1, characterized in that: After obtaining the electricity trading storage plan with the highest incentive score, the method also includes: obtaining a real electricity change curve, using the electricity trading storage plan to perform electricity trading simulation calculations based on the trading incentive strategy on the real electricity change curve to obtain a real incentive score; calculating the ratio of the real incentive score to the highest incentive score to obtain a verification score; judging whether the verification score is less than a preset qualified threshold; when the verification score is less than the qualified threshold, determining that the electricity trading plan generation model needs to be retrained.

5. The data-driven virtual power plant power trading optimization method according to claim 1, characterized in that: The method of performing an extension operation on the electricity price change curve based on the curve law to obtain a primary electricity price prediction curve includes: performing a curve feature extraction operation on the electricity price change curve to obtain historical electricity price features; using a pre-constructed long short-term memory network to perform a curve feature recognition operation based on the long and short term on the historical electricity price features to obtain short-term fluctuation features and long-term fluctuation features; calculating a curve change rate sequence based on the short-term fluctuation features and the long-term fluctuation features; and performing an electricity price prediction operation on the real-time electricity price using the curve change rate sequence to obtain a primary electricity price prediction curve.

6. The data-driven virtual power plant power trading optimization method according to claim 1, characterized in that: The method of querying a set of direct factors of electricity price based on a pre-constructed set of electricity-related literature comprises: performing a big data query operation based on preset electricity keywords on the pre-constructed set of electricity-related literature to obtain electricity-related texts; performing text semantic analysis on the electricity-related texts to obtain a set of electricity influencing factors; using a pre-constructed clustering algorithm to perform weight calculation on each electricity influencing factor in the set of electricity influencing factors to obtain a weight sequence of electricity influencing factors, and performing a Top-N-based weight sorting operation on the weight sequence of electricity influencing factors to obtain a set of potential factors; using a pre-constructed data interface to obtain a historical data sequence of each potential factor in the set of potential factors to obtain a potential factor historical data set, and using a pre-constructed Granger causality test algorithm to calculate a causal relationship score of each potential factor to the electricity price based on the potential factor historical data set and the electricity price change curve; and screening potential factors whose causal relationship scores are greater than a preset strong causal threshold to obtain a set of direct factors.

7. The data-driven virtual power plant power trading optimization method according to claim 1, characterized in that: The method of calculating the real-time change rate corresponding to each direct factor in the direct factor set based on the direct factor historical data set includes: identifying the time influence of each direct factor in the direct factor set to obtain the influence time of each direct factor, wherein the influence time includes units of day, week, month, season and year; configuring each influence time as a unit time corresponding to each direct factor; obtaining the most recent historical data sequence of each direct factor from the direct factor historical data set based on the unit time corresponding to each direct factor; and calculating the real-time change rate corresponding to each direct factor based on the most recent historical data sequence of each direct factor.

8. The data-driven virtual power plant power trading optimization method according to claim 1, characterized in that: The weighted calculation of each real-time change rate and each influence coefficient is performed to obtain a comprehensive power change rate, including: constructing a sub-predicted power change rate of each direct factor for the power price according to the real-time change rate and the influence coefficient of each direct factor; intercepting each sub-predicted power change rate by time length according to the influence time of each direct factor to obtain each effective sub-power change rate; and superimposing curves of each effective sub-power change rate to obtain a comprehensive power change rate.

9. The data-driven virtual power plant power trading optimization method according to claim 1, characterized in that: The method of performing a curve feature extraction operation on the electricity price prediction curve to obtain a predicted electricity price trend feature set includes: performing a convolution operation on the electricity price prediction curve to obtain a convolution matrix set, and performing feature dimension reduction on the convolution matrix set to obtain a curve feature sequence set; performing feature recognition on the curve feature sequence set to obtain a predicted electricity price trend feature set, wherein the predicted electricity price trend feature set includes the highest and lowest price points in the electricity price prediction curve, the time period in which the electricity price is greater than the real-time electricity price in the electricity price prediction curve, and the time period in which the electricity price is less than the real-time electricity price in the electricity price prediction curve.

10. A data-driven virtual power plant power trading optimization system, characterized in that: The system includes: a primary electricity price prediction module, which is used to obtain real-time electricity prices, obtain electricity price change curves within a preset time period, and extend the electricity price change curves based on curve rules to obtain primary electricity price prediction curves; a causal relationship price prediction module, which is used to query a set of direct factors of electricity prices based on a pre-constructed electricity-related literature set, obtain historical data corresponding to each direct factor in the direct factor set, obtain a direct factor historical data set, and perform multiple regression analysis on the electricity price change curve based on the direct factor historical data set to obtain the influence coefficient of each direct factor on the electricity price, and calculate the real-time change rate corresponding to each direct factor in the direct factor set based on the direct factor historical data set, perform weighted calculation on each real-time change rate and each influence coefficient to obtain a comprehensive electricity change rate, and perform a causal relationship price prediction on the real-time electricity price based on the comprehensive electricity change rate. Calculate to obtain an advanced electricity price forecast curve; a power generation surplus identification module is used to obtain a total power generation forecast curve, and obtain a sold power curve, calculate the difference between the total power generation forecast curve and the sold power curve, and obtain a surplus power forecast curve; a transaction plan generation module is used to use a pre-trained power transaction plan generation model to perform a curve feature extraction operation on the power price forecast curve to obtain a predicted power price trend feature set, perform a curve feature extraction operation on the surplus power forecast curve to obtain surplus power information, and perform a binary classification judgment based on storage and selling on the surplus power information according to a pre-constructed transaction incentive strategy and the predicted power price trend feature set to obtain a power transaction storage plan with the highest incentive score, wherein the power transaction storage plan includes the highest incentive score, the distribution of power sold at each time point, and the number of pre-constructed power storage devices occupied when all power is sold.

Images