Electricity market transaction strategy recommendation method and device, equipment and storage medium

Abstract

The invention discloses an electricity market transaction strategy recommendation method and device, equipment and a storage medium. The method comprises the following steps: acquiring all running day city-level weather forecast features of a running day, and all historical city-level weather features and historical market running boundary information of a historical associated day, respectively aggregating the city-level weather forecast features of each running day and the historical city-level weather features, and combining the historical market running boundary information to obtain a city-level weather forecast feature of each running day and a city-level weather forecast feature of each historical associated day; predicting the day-ahead market supply and demand power of the operation day and the real-time market supply and demand power of the operation day by adopting an electric power prediction model; determining a day-ahead bidding space and a real-time bidding space based on the day-ahead market supply and demand power of the operation day and the real-time market supply and demand power of the operation day; and according to the day-ahead bidding space and the real-time bidding space, in combination with the user risk preference type, recommending a market transaction strategy of the running day. According to the invention, the method achieves the accurate prediction of the transaction trend of the electric power spot market, gives a corresponding transaction strategy, and achieves the maximization of the power generation income of a station.

Description

Technical Field

[0001] The present invention relates to the technical field of electricity market trading, and particularly to a method, device, equipment and storage medium for recommending electricity market trading strategies. Background Art

[0002] In recent years, the reform of the electricity system has been continuously deepened, the proportion of electricity traded in the electricity market has been continuously expanding, and relevant departments have also issued documents requiring the steady and orderly full coverage of the electricity spot market. New energy power generation stations face both opportunities and challenges.

[0003] The power generation of new energy power stations is divided into three parts: base plan electricity, medium- and long-term trading plan electricity, and spot trading plan electricity. With the annual reduction of base electricity, more and more electricity will participate in spot market trading. The clearing prices in the day-ahead market and real-time market fluctuate greatly, and the market price differences alternate between positive and negative. At present, the electricity spot market trading strategy is mainly obtained by electricity market traders collecting reference data and relying on experience and recent trends for reasoning, lacking numerical quantification and risk management, and the prediction effect is unstable.

[0004] How to formulate strategies in the spot market trading of new energy power stations to reasonably allocate the trading electricity in the day-ahead market and real-time market, reduce prediction deviation, and ultimately maximize the overall power generation revenue is an urgent problem to be solved. Summary of the Invention

[0005] The present invention provides a method, device, equipment and storage medium for recommending electricity market trading strategies to accurately predict the trading trend of the electricity spot market, give corresponding trading strategies, and maximize the power generation revenue of the power station.

[0006] According to one aspect of the present invention, there is provided a method for recommending electricity market trading strategies, the method comprising:

[0007] Obtaining all city-level meteorological forecast features of the operating day, as well as all historical city-level meteorological features and historical market operation boundary information of historical associated days, wherein the historical city-level meteorological features include historical city-level meteorological forecast features and historical city-level meteorological measured features, and the historical market operation boundary information includes historical day-ahead market supply and demand power and historical real-time market supply and demand power;

[0008] Aggregating each of the city-level meteorological forecast features of the operating day, each of the historical city-level meteorological forecast features, and each of the historical city-level meteorological measured features to obtain provincial-level meteorological forecast features of the operating day, historical provincial-level meteorological forecast features, and historical provincial-level meteorological measured features;

[0009] Based on the provincial meteorological forecast characteristics of the operating day, the historical provincial meteorological forecast characteristics, the historical provincial meteorological measured characteristics, the historical pre-day market supply and demand power, and the historical real-time market supply and demand power, use an electric power prediction model to predict the pre-day market supply and demand power and the real-time market supply and demand power of the operating day;

[0010] Based on the pre-day market supply and demand power of the operating day and the real-time market supply and demand power of the operating day, determine the pre-day bidding space and the real-time bidding space;

[0011] According to the pre-day bidding space and the real-time bidding space, combined with the user's risk preference type, recommend the market trading strategy for the operating day.

[0012] Optionally, the provincial meteorological forecast characteristics of the operating day include the provincial centralized photovoltaic meteorological forecast characteristics of the operating day, the provincial distributed photovoltaic meteorological forecast characteristics of the operating day, the provincial centralized wind power meteorological forecast characteristics of the operating day, and the provincial whole-network load meteorological forecast characteristics of the operating day;

[0013] The historical provincial meteorological forecast characteristics include the historical provincial centralized photovoltaic meteorological forecast characteristics, the historical provincial distributed photovoltaic meteorological forecast characteristics, the historical provincial centralized wind power meteorological forecast characteristics, and the historical provincial whole-network load meteorological forecast characteristics;

[0014] The historical provincial meteorological measured characteristics include the historical provincial centralized photovoltaic meteorological measured characteristics, the historical provincial distributed photovoltaic meteorological measured characteristics, the historical provincial centralized wind power meteorological measured characteristics, and the historical provincial whole-network load meteorological measured characteristics.

[0015] Optionally, the aggregation of the city-level meteorological forecast characteristics, the historical city-level meteorological forecast characteristics, and the historical city-level meteorological measured characteristics of each operating day to obtain the provincial meteorological forecast characteristics, the historical provincial meteorological forecast characteristics, and the historical provincial meteorological measured characteristics of the operating day includes:

[0016] When the city-level meteorological forecast characteristics of the operating day are used as city-level meteorological characteristics, all city-level meteorological characteristics are weighted and fused after being divided by meteorological characteristic elements to obtain provincial centralized photovoltaic meteorological characteristics, provincial distributed photovoltaic meteorological characteristics, provincial centralized wind power meteorological characteristics, and provincial whole-network load meteorological characteristics, and are respectively used as the provincial centralized photovoltaic meteorological forecast characteristics of the operating day, the provincial distributed photovoltaic meteorological forecast characteristics of the operating day, the provincial centralized wind power meteorological forecast characteristics of the operating day, and the provincial whole-network load meteorological forecast characteristics of the operating day;

[0017] When the historical urban-level meteorological forecast features are used as urban-level meteorological features, all urban-level meteorological features are divided according to meteorological feature elements and then weighted and fused to obtain provincial-level centralized photovoltaic meteorological features, provincial-level distributed photovoltaic meteorological features, provincial-level centralized wind power meteorological features, and provincial-level whole-network load meteorological features, which are respectively used as the historical provincial-level centralized photovoltaic meteorological forecast features, the historical provincial-level distributed photovoltaic meteorological forecast features, the historical provincial-level centralized wind power meteorological forecast features, and the historical provincial-level whole-network load meteorological forecast features;

[0018] When the historical urban-level meteorological measured features are used as urban-level meteorological features, all urban-level meteorological features are divided according to meteorological feature elements and then weighted and fused to obtain provincial-level centralized photovoltaic meteorological features, provincial-level distributed photovoltaic meteorological features, provincial-level centralized wind power meteorological features, and provincial-level whole-network load meteorological features, which are respectively used as the historical provincial-level centralized photovoltaic meteorological measured features, the historical provincial-level distributed photovoltaic meteorological measured features, the historical provincial-level centralized wind power meteorological measured features, and the historical provincial-level whole-network load meteorological measured features.

[0019] Optionally, the step of dividing all urban-level meteorological features according to meteorological feature elements and then weighted and fused to obtain provincial-level centralized photovoltaic meteorological features, provincial-level distributed photovoltaic meteorological features, provincial-level centralized wind power meteorological features, and provincial-level whole-network load meteorological features includes:

[0020] All urban-level meteorological features are divided into basic meteorological features, light-related meteorological features, wind-related meteorological features, and extreme meteorological features;

[0021] Based on the share of centralized photovoltaic installed capacity in the city, each of the basic meteorological features, each of the light-related meteorological features, and each of the extreme meteorological features are weighted and fused to obtain provincial-level centralized photovoltaic meteorological features;

[0022] Based on the share of distributed photovoltaic installed capacity in the city, each of the basic meteorological features, each of the light-related meteorological features, and each of the extreme meteorological features are weighted and fused to obtain provincial-level distributed photovoltaic meteorological features;

[0023] Based on the share of centralized wind power installed capacity in the city, each of the basic meteorological features, each of the wind-related meteorological features, and each of the extreme meteorological features are weighted and fused to obtain provincial-level centralized wind power meteorological features;

[0024] Based on the share of electricity consumption in the city, each of the basic meteorological features and each of the extreme meteorological features are weighted and fused to obtain provincial-level whole-network load meteorological features.

[0025] Optionally, the historical day-ahead market supply and demand power includes historical day-ahead cleared centralized photovoltaic power, historical day-ahead cleared distributed photovoltaic power, historical day-ahead cleared centralized wind power, and historical day-ahead cleared total network load;

[0026] The historical real-time market supply and demand power includes historical real-time centralized photovoltaic power, historical real-time distributed photovoltaic power, historical real-time centralized wind power, and historical real-time total network load;

[0027] Correspondingly, the operating day's day-ahead market supply and demand power includes operating day's day-ahead centralized photovoltaic power, operating day's day-ahead distributed photovoltaic power, operating day's day-ahead centralized wind power, and operating day's day-ahead total network load;

[0028] The operating day's real-time market supply and demand power includes operating day's real-time centralized photovoltaic power, operating day's real-time distributed photovoltaic power, operating day's real-time centralized wind power, and operating day's real-time total network load.

[0029] Optionally, based on the operating day's provincial meteorological forecast characteristics, the historical provincial meteorological forecast characteristics, the historical provincial meteorological measured characteristics, the historical day-ahead market supply and demand power, and the historical real-time market supply and demand power, using an electric power prediction model to predict the operating day's day-ahead market supply and demand power and operating day's real-time market supply and demand power, includes:

[0030] Based on the operating day's provincial centralized photovoltaic meteorological forecast characteristics, the historical provincial centralized photovoltaic meteorological forecast characteristics, and the historical day-ahead cleared centralized photovoltaic power, using an electric power prediction model to obtain the operating day's day-ahead centralized photovoltaic power;

[0031] Based on the operating day's provincial distributed photovoltaic meteorological forecast characteristics, the historical provincial distributed photovoltaic meteorological forecast characteristics, and the historical day-ahead cleared distributed photovoltaic power, using the electric power prediction model to obtain the operating day's day-ahead distributed photovoltaic power;

[0032] Based on the operating day's provincial centralized wind power meteorological forecast characteristics, the historical provincial centralized wind power meteorological forecast characteristics, and the historical day-ahead cleared centralized wind power, using the electric power prediction model to obtain the operating day's day-ahead centralized wind power;

[0033] Based on the operating day's provincial total network load meteorological forecast characteristics, the historical provincial total network load meteorological forecast characteristics, and the historical day-ahead cleared total network load, using the electric power prediction model to obtain the operating day's day-ahead total network load;

[0034] Based on the provincial centralized PV meteorological forecast characteristics on the operation day, the historical provincial centralized PV meteorological measured characteristics, and the historical real-time centralized PV power, the real-time centralized PV power on the operation day is obtained by using the electric power prediction model;

[0035] Based on the provincial distributed PV meteorological forecast characteristics on the operation day, the historical provincial distributed PV meteorological measured characteristics, and the historical real-time distributed PV power, the real-time distributed PV power on the operation day is obtained by using the electric power prediction model;

[0036] Based on the provincial centralized wind power meteorological forecast characteristics on the operation day, the historical provincial centralized wind power meteorological measured characteristics, and the historical real-time centralized wind power, the real-time centralized wind power on the operation day is obtained by using the electric power prediction model;

[0037] Based on the provincial whole-network load meteorological forecast characteristics on the operation day, the historical provincial whole-network load meteorological measured characteristics, and the historical real-time whole-network load, the real-time whole-network load on the operation day is obtained by using the electric power prediction model.

[0038] Optionally, determining the day-ahead bidding space and the real-time bidding space based on the day-ahead market supply-demand power and the real-time market supply-demand power on the operation day includes:

[0039] Obtaining the pre-disclosed power of local power plants, the day-ahead tie-line planned power, and the real-time tie-line planned power;

[0040] Taking the sum of the day-ahead centralized PV power, the day-ahead distributed PV power, the day-ahead centralized wind power, the pre-disclosed power of local power plants, and the day-ahead tie-line planned power on the operation day as the day-ahead supply-side power, and taking the difference between the day-ahead whole-network load and the day-ahead supply-side power on the operation day as the day-ahead bidding space;

[0041] Taking the sum of the real-time centralized PV power, the real-time distributed PV power, the real-time centralized wind power, the pre-disclosed power of local power plants, and the real-time tie-line planned power on the operation day as the real-time supply-side power, and taking the difference between the real-time whole-network load and the real-time supply-side power on the operation day as the real-time bidding space.

[0042] Optionally, recommending the market trading strategy on the operation day according to the day-ahead bidding space and the real-time bidding space, in combination with the user's risk preference type, includes:

[0043] Taking the difference between the day-ahead bidding space and the real-time bidding space as the bidding space deviation;

[0044] According to the user risk preference type and the bidding space deviation, using a preset deviation coefficient mapping function, and combining with the installed capacity of the power station, determine the strategy coefficient;

[0045] Based on the strategy coefficient, determine the market trading strategy for the operating day.

[0046] Optionally, the step of determining the strategy coefficient according to the user risk preference type and the bidding space deviation, using a preset deviation coefficient mapping function, and combining with the installed capacity of the power station, includes:

[0047] Obtain the user risk preference type, and determine the risk preference coefficient corresponding to the user risk preference type;

[0048] Substitute the bidding space deviation and the risk preference coefficient into the preset deviation coefficient mapping function to obtain an initial strategy coefficient;

[0049] Obtain the installed capacity of the power station and the initial power of the power station, and determine the maximum output power of the power station according to the installed capacity of the power station and the initial power of the power station;

[0050] Based on the maximum output power of the power station, adjust the initial strategy coefficient to obtain the strategy coefficient.

[0051] Optionally, the expression of the deviation coefficient mapping function is:

[0052]

[0053] where, I t represents the initial strategy coefficient within the time period t, β represents the risk preference coefficient, and a and b are the function coefficients of the deviation coefficient mapping function, which are obtained by performing function fitting on historical samples by calling the nonlinear least squares method.

[0054] According to another aspect of the present invention, there is provided a power market trading strategy recommendation device, which includes:

[0055] A prediction data acquisition module, configured to acquire all operating day city-level meteorological forecast features of the operating day, as well as all historical city-level meteorological features and historical market operation boundary information of historical associated days, where the historical city-level meteorological features include historical city-level meteorological forecast features and historical city-level meteorological measured features, and the historical market operation boundary information includes historical day-ahead market supply and demand power and historical real-time market supply and demand power;

[0056] A meteorological feature aggregation module, configured to respectively aggregate each of the operating day city-level meteorological forecast features, each of the historical city-level meteorological forecast features, and each of the historical city-level meteorological measured features to obtain operating day provincial-level meteorological forecast features, historical provincial-level meteorological forecast features, and historical provincial-level meteorological measured features;

[0057] A market power prediction module, which is used to predict the power supply and demand of the day-ahead market and the real-time market of the operating day by using an electric power prediction model according to the provincial meteorological forecast characteristics of the operating day, the historical provincial meteorological forecast characteristics, the historical provincial meteorological measured characteristics, the historical day-ahead market power supply and demand, and the historical real-time market power supply and demand;

[0058] A bidding space determination module, which is used to determine the day-ahead bidding space and the real-time bidding space based on the power supply and demand of the day-ahead market and the real-time market of the operating day;

[0059] A trading strategy recommendation module, which is used to recommend the market trading strategy of the operating day according to the day-ahead bidding space and the real-time bidding space, in combination with the user's risk preference type.

[0060] According to another aspect of the present invention, an electronic device is provided, and the electronic device includes:

[0061] At least one processor; and

[0062] A memory communicatively connected to the at least one processor; wherein,

[0063] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the power market trading strategy recommendation method according to any embodiment of the present invention.

[0064] According to another aspect of the present invention, a computer-readable storage medium is provided, and the computer-readable storage medium stores computer instructions, and the computer instructions are used to implement the power market trading strategy recommendation method according to any embodiment of the present invention when executed by a processor.

[0065] The technical solution of the embodiment of the present invention is as follows: By obtaining all the city-level meteorological forecast features of the operating day, as well as all the historical city-level meteorological features and historical market operation boundary information of the historical associated days. Among them, the historical city-level meteorological features include historical city-level meteorological forecast features and historical city-level meteorological measured features, and the historical market operation boundary information includes historical day-ahead market supply and demand power and historical real-time market supply and demand power; Aggregate each operating day's city-level meteorological forecast feature, each historical city-level meteorological forecast feature, and each historical city-level meteorological measured feature respectively to obtain the provincial-level meteorological forecast feature of the operating day, the historical provincial-level meteorological forecast feature, and the historical provincial-level meteorological measured feature; According to the provincial-level meteorological forecast feature of the operating day, the historical provincial-level meteorological forecast feature, the historical provincial-level meteorological measured feature, the historical day-ahead market supply and demand power, and the historical real-time market supply and demand power, use an electric power prediction model to predict the day-ahead market supply and demand power and the real-time market supply and demand power of the operating day; Based on the day-ahead market supply and demand power and the real-time market supply and demand power of the operating day, determine the day-ahead bidding space and the real-time bidding space; According to the day-ahead bidding space and the real-time bidding space, combined with the user's risk preference type, recommend the market trading strategy of the operating day. Through the provincial-level meteorological features fused from the city-level meteorological features in this embodiment, combined with the spot market trading conditions of historical similar days, it can accurately predict the spot market trading trend and give corresponding trading strategies, realizing the maximization of the power generation income of the power station, and solving the problem that the current power spot market trading strategy mainly relies on manual experience and recent trend reasoning, lacking numerical quantification and risk management, and the prediction effect is unstable.

[0066] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present invention, nor is it used to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. Brief Description of the Drawings

[0067] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0068] Figure 1 is a flowchart of a method for recommending an electric power market trading strategy provided in Embodiment 1 of the present invention;

[0069] Figure 2a is a flowchart of a method for recommending an electric power market trading strategy provided in Embodiment 2 of the present invention;

[0070] Figure 2bIt is a schematic diagram of the principle of the electric power prediction model in a method for recommending electric power market trading strategies provided in Embodiment 2 of the present invention;

[0071] Figure 3 It is a schematic structural diagram of a device for recommending electric power market trading strategies provided in Embodiment 3 of the present invention;

[0072] Figure 4 It is a schematic structural diagram of an electronic device for implementing the method for recommending electric power market trading strategies in the embodiments of the present invention. Detailed implementation manners

[0073] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0074] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily need to be used to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units does not necessarily need to be limited to those clearly listed steps or units, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0075] Embodiment 1

[0076] Figure 1 The present invention provides a flowchart of a method for recommending electric power market trading strategies in Embodiment 1. This embodiment is applicable to predicting the trading strategies of the electric power spot market. This method can be executed by a device for recommending electric power market trading strategies. The device for recommending electric power market trading strategies can be implemented in the form of hardware and / or software, and the device for recommending electric power market trading strategies can be configured in a computer device. As Figure 1 shown, the method includes:

[0077] S110. Obtain all the city-level meteorological forecast features of the operating day, as well as all the historical city-level meteorological features and historical market operation boundary information of the historical associated days. Among them, the historical city-level meteorological features include historical city-level meteorological forecast features and historical city-level meteorological measured features, and the historical market operation boundary information includes historical day-ahead market supply-demand power and historical real-time market supply-demand power.

[0078] Among them, the operating day is a certain day for which a market trading strategy recommendation is required. Generally, the next day of the date where the current moment is located can be selected as the operating day. If the operating day is denoted as day D, then today can be denoted as day D - 1. The historical associated days can be understood as the days before the operating day that are similar or relevant in terms of features such as date, holiday type, and occurrence of major social events. There can be multiple historical associated days in this embodiment.

[0079] The city-level meteorological forecast features can be understood as the feature data extracted based on the meteorological forecast information of a certain city within the province. The city-level meteorological forecast features of the operating day are the city-level meteorological forecast features of the operating day, and the historical city-level meteorological forecast features are the city-level meteorological forecast features of the historical associated days. The city-level meteorological measured features can be understood as the feature data extracted based on the actually measured meteorological information of a certain city within the province. The historical city-level meteorological measured features are the city-level meteorological measured features of the historical associated days.

[0080] The historical market operation boundary information can include various supply-demand condition information related to power trading such as the provincial-wide network load, provincial dispatching load, centralized photovoltaic power, wind power, tie-line plan, day-ahead clearing price, real-time clearing price, pumped storage, nuclear power, etc. of the historical associated days disclosed by the power trading center. In this embodiment, the historical day-ahead market supply-demand power and historical real-time market supply-demand power can be mainly obtained. The historical day-ahead market supply-demand power can be understood as the power data related to the power supply side and power demand side in the day-ahead market of the historical associated days, and the historical real-time market supply-demand power can be understood as the power data related to the power supply side and power demand side in the real-time market of the historical associated days.

[0081] S120. Aggregate the city-level meteorological forecast features of each operating day, the historical city-level meteorological forecast features, and the historical city-level meteorological measured features respectively to obtain the provincial-level meteorological forecast features of the operating day, the historical provincial-level meteorological forecast features, and the historical provincial-level meteorological measured features.

[0082] This embodiment mainly analyzes and predicts the provincial power spot trading market. However, the provincial meteorological forecast is obviously not as accurate as the city-level meteorological forecast. Therefore, the city-level meteorological features can be fused into provincial-level meteorological features, and the trading trend of the power spot market can be predicted based on the fused meteorological features, which can improve the prediction accuracy.

[0083] Specifically, the operation-day city-level meteorological forecast features corresponding to all cities within the province can be aggregated to obtain the operation-day provincial-level meteorological forecast features; the historical city-level meteorological forecast features corresponding to all cities within the province can be aggregated to obtain the historical provincial-level meteorological forecast features; and the historical provincial-level meteorological measured features corresponding to all cities within the province can be aggregated to obtain the historical provincial-level meteorological measured features.

[0084] S130. According to the operation-day provincial-level meteorological forecast features, historical provincial-level meteorological forecast features, historical provincial-level meteorological measured features, historical day-ahead market supply-demand power, and historical real-time market supply-demand power, use an electric power prediction model to predict the operation-day day-ahead market supply-demand power and operation-day real-time market supply-demand power of the operation day.

[0085] Among them, the operation-day day-ahead market supply-demand power can be understood as the power data related to the power supply side and power demand side in the operation-day day-ahead market, and the operation-day real-time market supply-demand power can be understood as the power data related to the power supply side and power demand side in the operation-day real-time market.

[0086] The electric power prediction model used in this embodiment can be trained based on historical sample data. Taking the operation-day provincial-level meteorological forecast features, historical provincial-level meteorological forecast features, and historical day-ahead market supply-demand power as input data, input them into the electric power prediction model, and output the predicted operation-day day-ahead market supply-demand power; taking the operation-day provincial-level meteorological forecast features, historical provincial-level meteorological measured features, and historical real-time market supply-demand power as input data, input them into the electric power prediction model, and output the predicted operation-day real-time market supply-demand power.

[0087] S140. Based on the operation-day day-ahead market supply-demand power and operation-day real-time market supply-demand power, determine the day-ahead bidding space and real-time bidding space.

[0088] Among them, the day-ahead bidding space can reflect the comparison of the amount of electricity that the power supply side in the day-ahead market can provide and the amount of electricity that the power demand side needs to consume. The real-time bidding space can reflect the comparison of the amount of electricity that the power supply side in the real-time market can provide and the amount of electricity that the power demand side needs to consume.

[0089] Since the operation-day day-ahead market supply-demand power includes the predicted power of the electricity that the power supply side in the day-ahead market can provide and the power of the electricity that the power demand side needs, the day-ahead bidding space can be obtained through the analysis of the operation-day day-ahead market supply-demand power; similarly, the operation-day real-time market supply-demand power includes the predicted power of the electricity that the power supply side in the real-time market can provide and the power of the electricity that the power demand side needs, and the real-time bidding space can be obtained through the analysis of the operation-day real-time market supply-demand power.

[0090] S150. Recommend the market trading strategy for the operation day based on the day-ahead bidding space and the real-time bidding space, combined with the user's risk preference type.

[0091] Among them, the user's risk preference type can be divided into conservative, balanced, and aggressive types, etc., which reflects the degree of risk that the user can bear. The more aggressive the user's risk preference, the higher the possible return, but the greater the risk borne. Similarly, the more conservative the user's risk preference, the less possible return, but the smaller the risk borne.

[0092] Specifically, by comparing the day-ahead bidding space and the real-time bidding space, it can be analyzed which market has a greater demand for electricity in the day-ahead market and the real-time market. It can be understood that according to the law of market supply and demand, the clearing price of the electricity spot market mainly depends on the balance degree of the electricity market supply and demand. A larger demand corresponds to a higher price. Therefore, a trading strategy can be selected to bias more trading volume towards the spot market with a larger demand to improve trading returns. And for the degree of bias of the trading volume, it can be determined according to the user's risk preference type.

[0093] In the embodiment of the present invention, all city-level meteorological forecast characteristics of the operation day, all historical city-level meteorological characteristics of the historical associated day, and historical market operation boundary information are obtained. Among them, the historical city-level meteorological characteristics include historical city-level meteorological forecast characteristics and historical city-level meteorological measured characteristics, and the historical market operation boundary information includes historical day-ahead market supply and demand power and historical real-time market supply and demand power; the city-level meteorological forecast characteristics of each operation day, the historical city-level meteorological forecast characteristics of each historical day, and the historical city-level meteorological measured characteristics of each historical day are respectively aggregated to obtain provincial-level meteorological forecast characteristics of the operation day, historical provincial-level meteorological forecast characteristics, and historical provincial-level meteorological measured characteristics; according to the provincial-level meteorological forecast characteristics of the operation day, historical provincial-level meteorological forecast characteristics, historical provincial-level meteorological measured characteristics, historical day-ahead market supply and demand power, and historical real-time market supply and demand power, a power prediction model is used to predict the day-ahead market supply and demand power of the operation day and the real-time market supply and demand power of the operation day; based on the day-ahead market supply and demand power of the operation day and the real-time market supply and demand power of the operation day, the day-ahead bidding space and the real-time bidding space are determined; according to the day-ahead bidding space and the real-time bidding space, combined with the user's risk preference type, the market trading strategy for the operation day is recommended. Through the provincial-level meteorological characteristics fused from the city-level meteorological characteristics in this embodiment, combined with the spot market trading conditions of historical similar days, the trading trend of the spot market can be accurately predicted, and the corresponding trading strategy can be given to maximize the power generation income of the power station, solving the problem that the current trading strategy of the electricity spot market mainly relies on manual experience and recent trend reasoning, lacks numerical quantification and risk management, and the prediction effect is unstable.

[0094] Embodiment 2

[0095] Figure 2aFIG. 0 is a flowchart of a method for recommending an electricity market trading strategy provided by the second embodiment of the present invention. Based on the above embodiment, the method for recommending the electricity market trading strategy is further optimized. As Figure 2a shown, the method includes:

[0096] S210. Obtain all city-level meteorological forecast features of the operating day, as well as all historical city-level meteorological forecast features, historical city-level meteorological measured features, historical day-ahead market supply and demand power, and historical real-time market supply and demand power of the historical associated days. Among them, the historical day-ahead market supply and demand power includes historical day-ahead cleared centralized photovoltaic power, historical day-ahead cleared distributed photovoltaic power, historical day-ahead cleared centralized wind power, and historical day-ahead cleared total network load. The historical real-time market supply and demand power includes historical real-time centralized photovoltaic power, historical real-time distributed photovoltaic power, historical real-time centralized wind power, and historical real-time total network load.

[0097] Currently, there is no directly open API interface for new energy power stations to directly retrieve meteorological data and market supply and demand data from each meteorological reporting center and power trading center. It can only rely on manual login to the trading center website to copy or download, and then manually complete data integration, which is time-consuming, laborious and error-prone. In this embodiment, a timed scheduling task can be used to start a multi-threaded high-concurrency crawler to complete network requests, web page parsing, and data integration operations, realizing the full automation of the data collection process disclosed by the meteorological reporting center and the power trading center, directly parsing and generating a structured data set, storing it in an online database, and directly obtaining relevant data from the online database when recommending an electricity market trading strategy.

[0098] After S210, S211, S212, and S213 can be performed respectively.

[0099] S211. When the city-level meteorological forecast features of the operating day are used as city-level meteorological features, all city-level meteorological features are weighted and fused after being divided by meteorological feature elements to obtain provincial centralized photovoltaic meteorological features, provincial distributed photovoltaic meteorological features, provincial centralized wind power meteorological features, and provincial total network load meteorological features, which are respectively used as the operating day provincial centralized photovoltaic meteorological forecast features, the operating day provincial distributed photovoltaic meteorological forecast features, the operating day provincial centralized wind power meteorological forecast features, and the operating day provincial total network load meteorological forecast features.

[0100] S212. When using the historical city-level meteorological forecast features as the city-level meteorological features, all the city-level meteorological features are divided according to meteorological feature elements and then weighted and fused to obtain provincial centralized photovoltaic meteorological features, provincial distributed photovoltaic meteorological features, provincial centralized wind power meteorological features, and provincial overall network load meteorological features, which are respectively used as historical provincial centralized photovoltaic meteorological forecast features, historical provincial distributed photovoltaic meteorological forecast features, historical provincial centralized wind power meteorological forecast features, and historical provincial overall network load meteorological forecast features.

[0101] S213. When using the historical city-level meteorological measured features as the city-level meteorological features, all the city-level meteorological features are divided according to meteorological feature elements and then weighted and fused to obtain provincial centralized photovoltaic meteorological features, provincial distributed photovoltaic meteorological features, provincial centralized wind power meteorological features, and provincial overall network load meteorological features, which are respectively used as historical provincial centralized photovoltaic meteorological measured features, historical provincial distributed photovoltaic meteorological measured features, historical provincial centralized wind power meteorological measured features, and historical provincial overall network load meteorological measured features.

[0102] In this embodiment, by introducing high-precision meteorological features at the city level and aggregating them into provincial-level meteorological features, it is used to predict provincial-level new energy power and load, which is closer to the actual scenario and improves the prediction accuracy.

[0103] Optionally, in S211, S212, and S213, the step of dividing all the city-level meteorological features according to meteorological feature elements and then weighted and fused to obtain provincial centralized photovoltaic meteorological features, provincial distributed photovoltaic meteorological features, provincial centralized wind power meteorological features, and provincial overall network load meteorological features can be achieved through the following steps:

[0104] A1. Divide all the city-level meteorological features into basic meteorological features, light-related meteorological features, wind-related meteorological features, and extreme meteorological features.

[0105] A2. Based on the share of centralized photovoltaic installed capacity in the city, respectively perform weighted fusion on each basic meteorological feature, each light-related meteorological feature, and each extreme meteorological feature to obtain provincial centralized photovoltaic meteorological features.

[0106] A3. Based on the share of distributed photovoltaic installed capacity in the city, respectively perform weighted fusion on each basic meteorological feature, each light-related meteorological feature, and each extreme meteorological feature to obtain provincial distributed photovoltaic meteorological features.

[0107] A4. Based on the share of centralized wind power installed capacity in the city, respectively perform weighted fusion on each basic meteorological feature, each wind-related meteorological feature, and each extreme meteorological feature to obtain provincial centralized wind power meteorological features.

[0108] A5. Based on the share of urban electricity consumption, each basic meteorological feature and each extreme meteorological feature are weighted and fused respectively to obtain the provincial-wide network load meteorological feature.

[0109] In this embodiment, the urban-level meteorological features can be divided into basic meteorological features, light-related meteorological features, wind-related meteorological features, and extreme meteorological features.

[0110] The basic meteorological features may include: meteorological features such as temperature, relative humidity, air pressure, precipitation, snowfall, etc.

[0111] The light-related meteorological features may include: meteorological features such as direct radiation, diffuse radiation, horizontal total radiation, cloud coverages of low, medium, and high altitudes, visibility, etc.

[0112] The wind-related meteorological features may include: meteorological features such as ground 10m wind speed, 100m high-altitude wind speed, ground 10m gust, 100m high-altitude gust, etc.

[0113] The extreme meteorological features may include: meteorological features such as sand and dust intensity, haze index, strong wind storm, cold wave, high temperature, heavy rain, heavy snow, freezing, etc.

[0114] The original meteorological features are at the urban-level granularity. Due to regional differences, the values of meteorological features in each city are not the same, and the new energy installed capacity shares or population and economic distributions in each city are not uniform. The contributions of each city to photovoltaic output or the whole network load cannot be regarded as equal. Therefore, when aggregating from urban-level meteorology to provincial-level meteorology, it cannot be simply averaged. When predicting the provincial-level centralized photovoltaic meteorological features, the corresponding meteorological features are weighted by the centralized photovoltaic installed capacity shares of each city. For example, if the city with the largest centralized photovoltaic installed capacity share is City B, then when predicting the provincial-level centralized photovoltaic power contribution, the contribution of solar radiation in City B to the provincial-level solar radiation is correspondingly the largest; and so on. When predicting the provincial-level distributed photovoltaic meteorological features, the corresponding meteorological features are weighted by the distributed photovoltaic installed capacity shares of each city; when predicting the provincial-level centralized wind power meteorological features, the corresponding meteorological features are weighted by the centralized wind power installed capacity shares of each city; when predicting the provincial-level whole network load meteorological features, the corresponding meteorological features are weighted by the electricity consumption shares of each city.

[0115] On the other hand, since the power of centralized photovoltaic power generation and distributed photovoltaic power generation are mainly affected by basic meteorological conditions, light-related meteorological conditions, and extreme meteorological conditions, when predicting the meteorological characteristics of provincial centralized photovoltaics, the basic meteorological characteristics, light-related meteorological characteristics, and extreme meteorological characteristics corresponding to each city are weighted and fused; the power of centralized wind power generation is mainly affected by basic meteorological conditions, wind-related meteorological conditions, and extreme meteorological conditions, so when predicting the meteorological characteristics of provincial centralized wind power, the basic meteorological characteristics, wind-related meteorological characteristics, and extreme meteorological characteristics corresponding to each city are weighted and fused; the power of the whole network load is mainly affected by basic meteorological conditions and extreme meteorological conditions, so when predicting the meteorological characteristics of the provincial whole network load, the basic meteorological characteristics and extreme meteorological characteristics corresponding to each city are weighted and fused.

[0116] After S211 and S212, S214, S215, S216, and S217 can be carried out respectively.

[0117] After S211 and S212, S218, S219, S220, and S221 can be carried out respectively.

[0118] S214. Based on the meteorological forecast characteristics of provincial centralized photovoltaics on the operating day, the historical meteorological forecast characteristics of provincial centralized photovoltaics, and the historical day-ahead cleared centralized photovoltaic power, using the electric power prediction model, obtain the day-ahead centralized photovoltaic power on the operating day.

[0119] S215. Based on the meteorological forecast characteristics of provincial distributed photovoltaics on the operating day, the historical meteorological forecast characteristics of provincial distributed photovoltaics, and the historical day-ahead cleared distributed photovoltaic power, using the electric power prediction model, obtain the day-ahead distributed photovoltaic power on the operating day.

[0120] S216. Based on the meteorological forecast characteristics of provincial centralized wind power on the operating day, the historical meteorological forecast characteristics of provincial centralized wind power, and the historical day-ahead cleared centralized wind power, using the electric power prediction model, obtain the day-ahead centralized wind power on the operating day.

[0121] S217. Based on the meteorological forecast characteristics of provincial whole network load on the operating day, the historical meteorological forecast characteristics of provincial whole network load, and the historical day-ahead cleared whole network load, using the electric power prediction model, obtain the day-ahead whole network load on the operating day.

[0122] S218. Based on the meteorological forecast characteristics of provincial centralized photovoltaics on the operating day, the historical measured meteorological characteristics of provincial centralized photovoltaics, and the historical real-time centralized photovoltaic power, using the electric power prediction model, obtain the real-time centralized photovoltaic power on the operating day.

[0123] S219. Based on the provincial distributed PV meteorological forecast characteristics on the operating day, the historical provincial distributed PV meteorological measured characteristics, and the historical real-time distributed PV power, use the electric power prediction model to obtain the real-time distributed PV power on the operating day.

[0124] S220. Based on the provincial centralized wind power meteorological forecast characteristics on the operating day, the historical provincial centralized wind power meteorological measured characteristics, and the historical real-time centralized wind power, use the electric power prediction model to obtain the real-time centralized wind power on the operating day.

[0125] S221. Based on the provincial whole-network load meteorological forecast characteristics on the operating day, the historical provincial whole-network load meteorological measured characteristics, and the historical real-time whole-network load, use the electric power prediction model to obtain the real-time whole-network load on the operating day.

[0126] In this embodiment, for the clearing power prediction of the day-ahead market, the meteorological forecast characteristics are mainly used, and for the prediction of the real-time market, the meteorological measured characteristics are mainly used.

[0127] After S214, S215, S216, S217, S218, S219, S220, and S221, S222 can be carried out.

[0128] S222. Obtain the pre-disclosed power of local power plants, the day-ahead tie-line planned power, and the real-time tie-line planned power.

[0129] In practical applications, the pre-disclosed power of local power plants, the day-ahead tie-line planned power, and the real-time tie-line planned power can be obtained from the publicly available market operation boundary conditions of the power trading center.

[0130] After S222, S223 and S224 can be carried out respectively.

[0131] S223. Take the sum of the day-ahead centralized PV power, the day-ahead distributed PV power, the day-ahead centralized wind power, the pre-disclosed power of local power plants, and the day-ahead tie-line planned power on the operating day as the day-ahead supply-side power, and take the difference between the day-ahead whole-network load and the day-ahead supply-side power on the operating day as the day-ahead bidding space.

[0132] Specifically, the day-ahead provincial dispatching load on the operating day = the day-ahead whole-network load on the operating day - the day-ahead distributed PV power on the operating day - the pre-disclosed power of local power plants; the day-ahead bidding space = the day-ahead provincial dispatching load on the operating day - the day-ahead centralized PV power on the operating day - the day-ahead centralized wind power on the operating day - the day-ahead tie-line planned power. Therefore, the above relationships can be integrated to obtain:

[0133] Day-ahead bidding space = Total network load before the operating day - (Centralized PV power before the operating day + Distributed PV power before the operating day + Centralized wind power before the operating day + Pre-disclosed power of local power plants + Day-ahead tie-line planned power).

[0134] S224. Take the sum of the real-time centralized PV power on the operating day, the real-time distributed PV power on the operating day, the real-time centralized wind power on the operating day, the pre-disclosed power of local power plants, and the real-time tie-line planned power as the real-time supply-side power, and take the difference between the real-time total network load on the operating day and the real-time supply-side power as the real-time bidding space.

[0135] Specifically, Real-time provincial dispatching load on the operating day = Real-time total network load on the operating day - Real-time distributed PV power on the operating day - Pre-disclosed power of local power plants; Real-time bidding space = Real-time provincial dispatching load on the operating day - Real-time centralized PV power on the operating day - Real-time centralized wind power on the operating day - Real-time tie-line planned power. Therefore, the above relationships can be integrated to obtain:

[0136] Real-time bidding space = Real-time total network load on the operating day - (Real-time centralized PV power on the operating day + Real-time distributed PV power on the operating day + Real-time centralized wind power on the operating day + Pre-disclosed power of local power plants + Real-time tie-line planned power)

[0137] After S223 and S224, S225 can be carried out.

[0138] S225. Take the difference between the day-ahead bidding space and the real-time bidding space as the bidding space deviation.

[0139] In practical applications, the power dispatching agency determines the real-time market clearing price based on the declared information of power generation enterprises sealed in the day-ahead market. Therefore, the price difference direction in the day-ahead market is equivalent to the deviation direction between the day-ahead bidding space and the real-time bidding space, that is, in which spot market the electricity demand is large, the price in that spot market is high. Therefore, it can be formulated as:

[0140] Bidding space deviation = Day-ahead bidding space - Real-time bidding space.

[0141] S226. According to the user's risk preference type and the bidding space deviation, adopt a preset deviation coefficient mapping function, and combine with the installed capacity of the power station to determine the strategy coefficient.

[0142] Optionally, S226 can be implemented through the following steps:

[0143] S2261. Obtain the user's risk preference type and determine the risk preference coefficient corresponding to the user's risk preference type.

[0144] Specifically, the corresponding relationship between the user risk preference type and the risk preference coefficient can be established in advance. The user risk preference type can be divided into conservative, balanced, and aggressive types. Compared with the aggressive type, the risk preference coefficient determined according to the balanced type is smaller, and thus the recommended market trading strategy is also milder. Similarly, compared with the balanced type, the risk preference coefficient determined according to the conservative type is smaller, and thus the recommended market trading strategy is also milder.

[0145] S2262. Substitute the bidding space deviation and the risk preference coefficient into the preset deviation coefficient mapping function to obtain the initial strategy coefficient.

[0146] Furthermore, the expression of the deviation coefficient mapping function is:

[0147]

[0148] Among them, I t can represent the initial strategy coefficient within the time period t, β can represent the risk preference coefficient, and a and b are the function coefficients of the deviation coefficient mapping function, which can be obtained by fitting the function to the historical samples by calling the nonlinear least squares method.

[0149] When the bidding space deviation > 0, it can represent that the estimated day-ahead market electricity price > the real-time market electricity price. The corresponding strategy coefficient I t > 1, which can represent that more electricity should be allocated to the day-ahead market; conversely, the strategy coefficient I t < 1, which can represent that more electricity should be allocated to the real-time market.

[0150] When the user risk preference type is conservative, the risk preference coefficient β can take 0.2; when the user risk preference type is balanced, the risk preference coefficient β can take 0.6; when the user risk preference type is aggressive, the risk preference coefficient β can take 1.

[0151] S2263. Obtain the station installed capacity and the initial power of the station, and determine the maximum output power of the station according to the station installed capacity and the initial power of the station.

[0152] Among them, the initial power of the station can be understood as the current power generation power of the station.

[0153] The maximum power that the station can output, that is, the maximum output power of the station, can be determined according to the station installed capacity and the initial power of the station. S2264. Adjust the initial strategy coefficient based on the maximum output power of the station to obtain the strategy coefficient.

[0154] In practical applications, due to the limitation of the installed capacity of the power station, after adjusting the output power by the initial strategy coefficient, the output power of the power station cannot exceed the maximum output power of the power station. If the output power adjusted by the strategy coefficient exceeds the maximum output power of the power station, the initial strategy coefficient can be adjusted to obtain the strategy coefficient.

[0155] S227. Determine the market trading strategy for the operating day based on the strategy coefficient.

[0156] Specifically, after obtaining the strategy coefficient, the output power of the power station can be adjusted according to the strategy coefficient to obtain the market trading strategy for the operating day, that is, the day-ahead market power finally declared.

[0157] In this embodiment, the time granularity for obtaining data can be divided according to the price clearing frequency of the power trading market.

[0158] Exemplarily, currently the power trading market usually adopts a 96-period system, that is, each 15 minutes is a trading clearing period. The meteorological information generally cannot reach this update frequency. Therefore, 24-hour-level meteorological information can be used for feature extraction and fusion. The obtained historical market supply and demand power is aggregated from 15-minute level to hourly level. After predicting the hourly market supply and demand power of the operating day and obtaining the hourly strategy coefficient, the hourly strategy coefficient can be decomposed into 96-point strategy coefficients, and a 96-point market trading strategy can be formulated for the operating day. The specific implementation process can be as follows:

[0159] Determine 6 historical associated days associated with the operating day. These 6 historical associated days can be the 6 consecutive days before the operating day. Obtain the hourly operating-day city-level meteorological forecast features of each city in the province on the operating day to form an hourly sequence of the operating-day city-level meteorological forecast features. Similarly, obtain the hourly sequence of the historical city-level meteorological forecast features and the hourly sequence of the historical city-level measured meteorological features.

[0160] After dividing the hourly sequences of urban-level meteorological forecasts for operation days by meteorological characteristic elements, they are respectively fused into hourly sequences of provincial-level centralized photovoltaic meteorological forecasts for operation days, hourly sequences of provincial-level distributed photovoltaic meteorological forecasts for operation days, hourly sequences of provincial-level centralized wind power meteorological forecasts for operation days, and hourly sequences of provincial-level overall network load meteorological forecasts for operation days. After dividing the hourly sequences of historical urban-level meteorological forecasts by meteorological characteristic elements, they are respectively fused into hourly sequences of historical provincial-level centralized photovoltaic meteorological forecasts, hourly sequences of historical provincial-level distributed photovoltaic meteorological forecasts, hourly sequences of historical provincial-level centralized wind power meteorological forecasts, and hourly sequences of historical provincial-level overall network load meteorological forecasts. After dividing the hourly sequences of historical urban-level meteorological measured characteristics by meteorological characteristic elements, they are respectively fused into hourly sequences of historical provincial-level centralized photovoltaic meteorological measured characteristics, hourly sequences of historical provincial-level distributed photovoltaic meteorological measured characteristics, hourly sequences of historical provincial-level centralized wind power meteorological measured characteristics, and hourly sequences of historical provincial-level overall network load meteorological measured characteristics.

[0161] Obtain the 96-point sequences of historical day-ahead cleared centralized photovoltaic power, historical day-ahead cleared distributed photovoltaic power, historical day-ahead cleared centralized wind power, and historical day-ahead cleared overall network load for each historical associated day. Aggregate the 96-point sequences into 24-hourly sequences to obtain the hourly sequences of historical day-ahead cleared centralized photovoltaic power, the hourly sequences of historical day-ahead cleared distributed photovoltaic power, the hourly sequences of historical day-ahead cleared centralized wind power, and the hourly sequences of historical day-ahead cleared overall network load. Similarly, the hourly sequences of historical real-time centralized photovoltaic power, the hourly sequences of historical real-time distributed photovoltaic power, the hourly sequences of historical real-time centralized wind power, and the hourly sequences of historical real-time overall network load can be obtained.

[0162] Since the data used in this embodiment is large in quantity, has numerous input features, and high computational complexity, it belongs to the problem of multi-dimensional multi-step time series prediction, which can also be called the problem of long sequence prediction. For long sequence time series prediction, the model is required to have high prediction ability, that is, to be able to accurately capture the long-term dependence relationship between the output and the input. Compared with RNN-based models (including LSTM), Transformer shows high potential in expressing long-distance dependencies. One very important reason is that Transformer applies the self-attention mechanism to reduce the maximum path length of signal propagation to the shortest O(1) and avoids the recurrent structure. However, the Transformer model still has three problems that limit the direct application of Transformer to long sequence problems:

[0163] 1) The quadratic computational complexity of self-attention is high. The atomic operation of self-attention, the normalized dot product, results in a time complexity and memory usage of O(L 2 ) for each layer, where L is the length of the input sequence.

[0164] 2) Stacking layers is memory-limited when dealing with long input sequences. The J-layer stacking framework of Encoder-Decoder leads to a memory usage of O(J·L 2 ) when inputting long sequences, which limits the scalability of the model when dealing with long input sequences.

[0165] 3) The speed drops significantly when predicting the output of long sequences. The dynamic decoding process of Transformer results in slow inference speed when inferring the output of long sequences, and the actual effect may be as poor as the RNN model mentioned above.

[0166] Therefore, this embodiment adopts an LSTF model based on an improved Transformer as the electric power prediction model, which has three remarkable features:

[0167] 1) The ProbSparse self-attention mechanism can replace the normalized self-attention and achieve a time complexity and memory usage of O(L·logL).

[0168] 2) The self-attention distilling operation can extract the main attention scores in the stacked J layers, significantly reducing the total space complexity.

[0169] 3) A generative style decoder is proposed to obtain the output of long sequences. It only requires one forward step to output the entire decoded sequence, significantly improving the inference speed of long sequence prediction and avoiding the cumulative error propagation during inference.

[0170] The electric power prediction model in this embodiment uses a multi-dimensional feature matrix (multiple time steps × multiple feature dimensions) as input for modeling, extracts information at critical moments in different time steps, such as possible photovoltaic icing after snowfall and reduced air floating dust after rain stops and clears, etc., which can improve the prediction accuracy and efficiency for long time series.

[0171] An overly short time series input may cause the neural network to be unable to utilize past information, while an overly long time series input may result in an excessive amount of historical information, making it impossible for the neural network to focus on key information. Therefore, considering the acquired data, the length of the input time series is set to 7×24 after comprehensive consideration. During prediction, the input data is normalized. Figure 2b is a schematic diagram of the principle of the electric power prediction model in a power market trading strategy recommendation method provided in Embodiment 2 of the present invention, as Figure 2b shown, taking [X1, X2, …, X 7×24 and [Y1, Y2, …, Y 6×24 as input data, and outputting [Y 6×24+1 , Y 6×24+2 , …, Y 7×24 .

[0172] Taking the historical provincial centralized photovoltaic meteorological forecast feature hourly series of 6 historical associated days as [X1, X2, …, X 6×24 , the operating day provincial centralized photovoltaic meteorological forecast feature hourly series as [X 6×24+1 , X 6×24+2 , …, X 7×24 , the historical day-ahead cleared centralized photovoltaic power hourly series of 6 historical associated days as [Y1, Y2, …, Y 6×24 , inputting into the electric power prediction model, and the output [Y 6×24+1 , Y 6×24+2 , …, Y 7×24 is the operating day day-ahead centralized photovoltaic power hourly series.

[0173] Taking the historical provincial distributed photovoltaic meteorological forecast feature hourly series of 6 historical associated days as [X1, X2, …, X 6×24 , the operating day provincial distributed photovoltaic meteorological forecast feature hourly series as [X 6×24+1 , X 6×24+2 , …, X 7×24 , the historical day-ahead cleared distributed photovoltaic power hourly series of 6 historical associated days as [Y1, Y2, …, Y 6×24 , inputting into the electric power prediction model, and the output [Y 6×24+1 , Y 6×24+2 , …, Y 7×24 is the operating day day-ahead distributed photovoltaic power hourly series.

[0174] Taking the historical provincial centralized wind power meteorological forecast feature hourly series of 6 historical associated days as [X1, X2, …, X 6×24 , the operating day provincial centralized wind power meteorological forecast feature hourly series as [X 6×24+1 , X6×24+2 ,…,X 7×24 , 6 hourly sequences of centralized wind power cleared before the historical days of 6 historical associated days are used as [Y1, Y2, …, Y 6×24 , and input into the electric power prediction model, the output [Y 6×24+1 , Y 6×24+2 , …, Y 7×24 is the hourly sequence of centralized wind power cleared before the operation day.

[0175] Use the hourly sequences of historical provincial-wide network load meteorological forecast characteristics of 6 historical associated days as [X1, X2, …, X 6×24 , and the hourly sequence of provincial-wide network load meteorological forecast characteristics of the operation day as [X 6×24+1 , X 6×24+2 , …, X 7×24 , 6 hourly sequences of historical provincial-wide network loads cleared before the historical days of 6 historical associated days are used as [Y1, Y2, …, Y 6×24 , and input into the electric power prediction model, the output [Y 6×24+1 , Y 6×24+2 , …, Y 7×24 is the hourly sequence of provincial-wide network load before the operation day.

[0176] Use the hourly sequences of historical provincial centralized photovoltaic meteorological measured characteristics of 6 historical associated days as [X1, X2, …, X 6×24 , and the hourly sequence of provincial centralized photovoltaic meteorological forecast characteristics of the operation day as [X 6×24+1 , X 6×24+2 , …, X 7×24 , 6 hourly sequences of historical real-time centralized photovoltaic power of 6 historical associated days are used as [Y1, Y2, …, Y 6×24 , and input into the electric power prediction model, the output [Y 6×24+1 , Y 6×24+2 , …, Y 7×24 is the hourly sequence of real-time centralized photovoltaic power on the operation day.

[0177] Use the hourly sequences of historical provincial distributed photovoltaic meteorological measured characteristics of 6 historical associated days as [X1, X2, …, X 6×24 , and the hourly sequence of provincial distributed photovoltaic meteorological forecast characteristics of the operation day as [X 6×24+1 , X 6×24+2 , …, X 7×24 , 6 hourly sequences of historical real-time distributed photovoltaic power of 6 historical associated days are used as [Y1, Y2, …, Y 6×24 , and input into the electric power prediction model, the output [Y 6×24+1 , Y 6×24+2 , …, Y 7×24That is the hourly sequence of the real-time distributed PV power on the operating day.

[0178] Take the hourly sequences of the historical provincial centralized wind power meteorological measured characteristics for 6 historical related days as [X1, X2, …, X 6×24 , and the hourly sequence of the provincial centralized wind power meteorological forecast characteristics on the operating day as [X 6×24+1 , X 6×24+2 , …, X 7×24 , and the hourly sequences of the historical real-time centralized wind power for 6 historical related days as [Y1, Y2, …, Y 6×24 . Input them into the electric power prediction model, and the output [Y 6×24+1 , Y 6×24+2 , …, Y 7×24 is the hourly sequence of the real-time centralized wind power on the operating day.

[0179] Take the hourly sequences of the historical provincial grid-wide load meteorological measured characteristics for 6 historical related days as [X1, X2, …, X 6×24 , and the hourly sequence of the provincial grid-wide load meteorological forecast characteristics on the operating day as [X 6×24+1 , X 6×24+2 , …, X 7×24 , and the hourly sequences of the historical real-time grid-wide load for 6 historical related days as [Y1, Y2, …, Y 6×24 . Input them into the electric power prediction model, and the output [Y 6×24+1 , Y 6×24+2 , …, Y 7×24 is the hourly sequence of the real-time grid-wide load on the operating day.

[0180] The obtained 96-point sequences of the pre-disclosed power of local power plants, the 96-point sequences of the planned power of the interconnection lines on the day-ahead, and the 96-point sequences of the real-time planned power of the interconnection lines can also be aggregated into hourly sequences. Combine with the hourly sequences of the day-ahead centralized PV power, the day-ahead distributed PV power, the day-ahead centralized wind power, and the day-ahead grid-wide load on the operating day to calculate the hourly sequence of the corresponding day-ahead bidding space on the operating day. Combine with the hourly sequences of the real-time centralized PV power, the real-time distributed PV power, the real-time centralized wind power, and the real-time grid-wide load on the operating day to calculate the hourly sequence of the corresponding real-time bidding space on the operating day, so as to obtain the hourly sequence of the bidding space deviation. After selecting the risk preference coefficient corresponding to the user's risk preference type, substitute each bidding space deviation in the hourly sequence of the bidding space deviation into the deviation coefficient mapping function respectively, and the corresponding initial strategy coefficient can be obtained, forming the hourly coefficient of the initial strategy coefficient.

[0181] The hourly sequence of the initial strategy coefficient is decomposed into a 96-point sequence. For the 4 fifteen-minute periods within each hour, the initial strategy coefficient of this hour is uniformly adopted, and a 96-point sequence of the initial strategy coefficient is obtained after decomposition.

[0182] Obtain the installed capacity C of the power station and the 96-point sequence of the initial power of the power station [P1, P2, …, P 96 , since the power generation power of the power station must be ≥0, let the minimum value of the power station output power be i min = 0; let the maximum value of the power station output power be i max = 1 + (1 - Max([P1, P2, …, P96]) / C). The initial strategy coefficient cannot be greater than i max , if the initial strategy coefficient cannot be greater than i max , then adjust this initial strategy coefficient to i max . The adjusted strategy coefficients form a 96-point sequence of strategy coefficients I = [i1, i2, …, i 96 .

[0183] The finally declared day-ahead market power can be expressed as [P1, P2, …, P 96 * [i1, i2, …, i 96 T .

[0184] In the embodiment of the present invention, after all urban-level meteorological characteristics are divided into basic meteorological characteristics, light-related meteorological characteristics, wind-related meteorological characteristics, and extreme meteorological characteristics according to meteorological characteristic elements, weighted fusion is performed according to the new energy installed capacity share or electricity consumption share of each city, and provincial centralized photovoltaic meteorological characteristics, provincial distributed photovoltaic meteorological characteristics, provincial centralized wind power meteorological characteristics, and provincial whole-network load meteorological characteristics are obtained respectively, which are closer to the actual scenario, improve the accuracy of meteorological characteristic extraction, and thus increase the accuracy of the prediction result; based on the intelligent recommendation strategy of user risk preference, the maximization of benefits under risk control is realized. This embodiment solves the problem that the current power spot market trading strategy mainly relies on manual experience and recent trend reasoning, lacks numerical quantification and risk management, and the prediction effect is unstable.

[0185] Embodiment III

[0186] Figure 3 FIG. shows a schematic structural diagram of a power market trading strategy recommendation device provided by Embodiment III of the present invention. As Figure 3 shown, the device includes a prediction data acquisition module 310, a meteorological characteristic aggregation module 320, a market power prediction module 330, a bidding space determination module 340, and a trading strategy recommendation module 350.

[0187] The prediction data acquisition module 310 is used to obtain all the city-level meteorological forecast features of the operation day, as well as all the historical city-level meteorological features and historical market operation boundary information of the historical associated days. Among them, the historical city-level meteorological features include historical city-level meteorological forecast features and historical city-level meteorological measured features, and the historical market operation boundary information includes historical day-ahead market supply and demand power and historical real-time market supply and demand power.

[0188] The meteorological feature aggregation module 320 is used to aggregate each of the city-level meteorological forecast features of the operation day, each of the historical city-level meteorological forecast features, and each of the historical city-level meteorological measured features to obtain provincial-level meteorological forecast features of the operation day, historical provincial-level meteorological forecast features, and historical provincial-level meteorological measured features.

[0189] The market power prediction module 330 is used to predict the day-ahead market supply and demand power and real-time market supply and demand power of the operation day by using an electric power prediction model according to the provincial-level meteorological forecast features of the operation day, the historical provincial-level meteorological forecast features, the historical provincial-level meteorological measured features, the historical day-ahead market supply and demand power, and the historical real-time market supply and demand power.

[0190] The bidding space determination module 340 is used to determine the day-ahead bidding space and real-time bidding space based on the day-ahead market supply and demand power and the real-time market supply and demand power of the operation day.

[0191] The trading strategy recommendation module 350 is used to recommend the market trading strategy of the operation day according to the day-ahead bidding space and the real-time bidding space, in combination with the user risk preference type.

[0192] In an embodiment of the present invention, all city-level meteorological forecast features of the operating day are obtained, as well as all historical city-level meteorological features and historical market operation boundary information of the historical associated days. Among them, the historical city-level meteorological features include historical city-level meteorological forecast features and historical city-level meteorological measured features, and the historical market operation boundary information includes historical day-ahead market supply-demand power and historical real-time market supply-demand power; the city-level meteorological forecast features of each operating day, the historical city-level meteorological forecast features of each historical day, and the historical city-level meteorological measured features of each historical day are respectively aggregated to obtain provincial-level meteorological forecast features of the operating day, historical provincial-level meteorological forecast features, and historical provincial-level meteorological measured features; based on the provincial-level meteorological forecast features of the operating day, historical provincial-level meteorological forecast features, historical provincial-level meteorological measured features, historical day-ahead market supply-demand power, and historical real-time market supply-demand power, a power prediction model is used to predict the day-ahead market supply-demand power and real-time market supply-demand power of the operating day; based on the day-ahead market supply-demand power and real-time market supply-demand power of the operating day, the day-ahead bidding space and real-time bidding space are determined; according to the day-ahead bidding space and real-time bidding space, combined with the user's risk preference type, a market trading strategy for the operating day is recommended. In this embodiment, through the provincial-level meteorological features fused from the city-level meteorological features, combined with the spot market trading conditions of historical similar days, the trading trend of the spot market can be accurately predicted, and the corresponding trading strategy can be given to maximize the power generation revenue of the power station, solving the problem that the current power spot market trading strategy mainly relies on manual experience and recent trend reasoning, lacking numerical quantification and risk management, and the prediction effect is unstable.

[0193] Optionally, the provincial-level meteorological forecast features of the operating day include provincial-level centralized photovoltaic meteorological forecast features of the operating day, provincial-level distributed photovoltaic meteorological forecast features of the operating day, provincial-level centralized wind power meteorological forecast features of the operating day, and provincial-level whole-network load meteorological forecast features of the operating day;

[0194] The historical provincial-level meteorological forecast features include historical provincial-level centralized photovoltaic meteorological forecast features, historical provincial-level distributed photovoltaic meteorological forecast features, historical provincial-level centralized wind power meteorological forecast features, and historical provincial-level whole-network load meteorological forecast features;

[0195] The historical provincial-level meteorological measured features include historical provincial-level centralized photovoltaic meteorological measured features, historical provincial-level distributed photovoltaic meteorological measured features, historical provincial-level centralized wind power meteorological measured features, and historical provincial-level whole-network load meteorological measured features.

[0196] Optionally, the meteorological feature aggregation module 320 includes:

[0197] The operating-day provincial meteorological forecast feature fusion unit is used to take the operating-day city-level meteorological forecast features as city-level meteorological features, divide all city-level meteorological features according to meteorological feature elements, and then perform weighted fusion to obtain provincial centralized photovoltaic meteorological features, provincial distributed photovoltaic meteorological features, provincial centralized wind power meteorological features, and provincial whole-network load meteorological features, and respectively use them as the operating-day provincial centralized photovoltaic meteorological forecast features, the operating-day provincial distributed photovoltaic meteorological forecast features, the operating-day provincial centralized wind power meteorological forecast features, and the operating-day provincial whole-network load meteorological forecast features;

[0198] The historical provincial meteorological forecast feature fusion unit is used to take the historical city-level meteorological forecast features as city-level meteorological features, divide all city-level meteorological features according to meteorological feature elements, and then perform weighted fusion to obtain provincial centralized photovoltaic meteorological features, provincial distributed photovoltaic meteorological features, provincial centralized wind power meteorological features, and provincial whole-network load meteorological features, and respectively use them as the historical provincial centralized photovoltaic meteorological forecast features, the historical provincial distributed photovoltaic meteorological forecast features, the historical provincial centralized wind power meteorological forecast features, and the historical provincial whole-network load meteorological forecast features;

[0199] The historical provincial meteorological measured feature fusion unit is used to take the historical city-level meteorological measured features as city-level meteorological features, divide all city-level meteorological features according to meteorological feature elements, and then perform weighted fusion to obtain provincial centralized photovoltaic meteorological features, provincial distributed photovoltaic meteorological features, provincial centralized wind power meteorological features, and provincial whole-network load meteorological features, and respectively use them as the historical provincial centralized photovoltaic meteorological measured features, the historical provincial distributed photovoltaic meteorological measured features, the historical provincial centralized wind power meteorological measured features, and the historical provincial whole-network load meteorological measured features.

[0200] Optionally, the step of dividing all city-level meteorological features according to meteorological feature elements and then performing weighted fusion to obtain provincial centralized photovoltaic meteorological features, provincial distributed photovoltaic meteorological features, provincial centralized wind power meteorological features, and provincial whole-network load meteorological features includes:

[0201] Dividing all city-level meteorological features into basic meteorological features, light-related meteorological features, wind-related meteorological features, and extreme meteorological features;

[0202] Based on the share of the centralized photovoltaic installed capacity in the city, perform weighted fusion on each of the basic meteorological features, each of the light-related meteorological features, and each of the extreme meteorological features to obtain provincial centralized photovoltaic meteorological features;

[0203] Based on the share of distributed photovoltaic installed capacity in the city, the basic meteorological characteristics, the light-related meteorological characteristics, and the extreme meteorological characteristics are weighted and fused respectively to obtain provincial distributed photovoltaic meteorological characteristics;

[0204] Based on the share of centralized wind power installed capacity in the city, the basic meteorological characteristics, the wind-related meteorological characteristics, and the extreme meteorological characteristics are weighted and fused respectively to obtain provincial centralized wind power meteorological characteristics;

[0205] Based on the share of electricity consumption in the city, the basic meteorological characteristics and the extreme meteorological characteristics are weighted and fused respectively to obtain provincial whole-network load meteorological characteristics.

[0206] Optionally, the historical day-ahead market supply and demand power includes historical day-ahead cleared centralized photovoltaic power, historical day-ahead cleared distributed photovoltaic power, historical day-ahead cleared centralized wind power, and historical day-ahead cleared whole-network load;

[0207] The historical real-time market supply and demand power includes historical real-time centralized photovoltaic power, historical real-time distributed photovoltaic power, historical real-time centralized wind power, and historical real-time whole-network load;

[0208] Correspondingly, the operating day's day-ahead market supply and demand power includes the operating day's day-ahead centralized photovoltaic power, the operating day's day-ahead distributed photovoltaic power, the operating day's day-ahead centralized wind power, and the operating day's day-ahead whole-network load;

[0209] The operating day's real-time market supply and demand power includes the operating day's real-time centralized photovoltaic power, the operating day's real-time distributed photovoltaic power, the operating day's real-time centralized wind power, and the operating day's real-time whole-network load.

[0210] Optionally, the market power prediction module 330 includes:

[0211] An operating day's day-ahead centralized photovoltaic power prediction unit, configured to obtain the operating day's day-ahead centralized photovoltaic power by using an electric power prediction model according to the operating day's provincial centralized photovoltaic meteorological forecast characteristics, the historical provincial centralized photovoltaic meteorological forecast characteristics, and the historical day-ahead cleared centralized photovoltaic power;

[0212] An operating day's day-ahead distributed photovoltaic power prediction unit, configured to obtain the operating day's day-ahead distributed photovoltaic power by using the electric power prediction model according to the operating day's provincial distributed photovoltaic meteorological forecast characteristics, the historical provincial distributed photovoltaic meteorological forecast characteristics, and the historical day-ahead cleared distributed photovoltaic power;

[0213] The centralized wind power prediction unit for the day before the operation date is used to obtain the centralized wind power for the day before the operation date by using the electric power prediction model according to the provincial centralized wind power meteorological forecast characteristics for the day before the operation date, the historical provincial centralized wind power meteorological forecast characteristics and the historical day-ahead cleared centralized wind power;

[0214] The network load forecasting unit before the operation date is used to obtain the network load before the operation date by using the electric power forecasting model according to the meteorological forecast characteristics of the provincial network load on the operation date, the meteorological forecast characteristics of the historical provincial network load and the historical cleared network load before the operation date;

[0215] The real-time distributed photovoltaic power prediction unit on the operation day is used to obtain the real-time centralized photovoltaic power on the operation day by using the electric power prediction model according to the provincial centralized photovoltaic meteorological forecast characteristics on the operation day, the historical provincial centralized photovoltaic meteorological measured characteristics and the historical real-time centralized photovoltaic power;

[0216] The real-time distributed photovoltaic power prediction unit on the operation day is used to obtain the real-time distributed photovoltaic power on the operation day by using the electric power prediction model according to the provincial distributed photovoltaic meteorological forecast characteristics on the operation day, the historical provincial distributed photovoltaic meteorological measured characteristics and the historical real-time distributed photovoltaic power;

[0217] The real-time centralized wind power prediction unit on the operation day is used to obtain the real-time centralized wind power on the operation day by using the electric power prediction model according to the provincial centralized wind power meteorological forecast characteristics on the operation day, the historical provincial centralized wind power meteorological measured characteristics and the historical real-time centralized wind power;

[0218] The real-time network load forecasting unit on the operation day is used to obtain the real-time network load on the operation day by using the electric power forecasting model according to the meteorological forecast characteristics of the provincial network load on the operation day, the meteorological measured characteristics of the historical provincial network load and the historical real-time network load.

[0219] Optionally, the bidding space determination module 340 includes:

[0220] The market operation boundary condition acquisition unit is used to obtain the pre-disclosed power of local power plants, the day-ahead interconnection line planned power, and the real-time interconnection line planned power;

[0221] a day-ahead bidding space determination unit, configured to take the sum of the day-ahead centralized photovoltaic power on the operation date, the day-ahead distributed photovoltaic power on the operation date, the day-ahead centralized wind power on the operation date, the pre-disclosed power of the local power plant and the day-ahead tie line planned power as the day-ahead supply side power, and take the difference between the day-ahead total grid load on the operation date and the day-ahead supply side power as the day-ahead bidding space;

[0222] A real-time bidding space determination unit, which is used to take the sum of the real-time centralized photovoltaic power on the operating day, the real-time distributed photovoltaic power on the operating day, the real-time centralized wind power on the operating day, the pre-disclosed power of the local power plant, and the real-time tie-line plan power as the real-time supply-side power, and take the difference between the real-time network-wide load on the operating day and the real-time supply-side power as the real-time bidding space.

[0223] Optionally, the trading strategy recommendation module 350 includes:

[0224] A bidding space deviation determination unit, which is used to take the difference between the day-ahead bidding space and the real-time bidding space as the bidding space deviation;

[0225] A strategy coefficient determination unit, which is used to determine a strategy coefficient according to the user's risk preference type and the bidding space deviation, by using a preset deviation coefficient mapping function and combining with the installed capacity of the power station;

[0226] A market trading strategy determination unit, which is used to determine the market trading strategy for the operating day based on the strategy coefficient.

[0227] Optionally, the strategy coefficient determination unit is specifically used for:

[0228] Obtain the user's risk preference type, and determine the risk preference coefficient corresponding to the user's risk preference type;

[0229] Substitute the bidding space deviation and the risk preference coefficient into the preset deviation coefficient mapping function to obtain an initial strategy coefficient;

[0230] Obtain the installed capacity of the power station and the initial power of the power station, and determine the maximum output power of the power station according to the installed capacity of the power station and the initial power of the power station;

[0231] Based on the maximum output power of the power station, adjust the initial strategy coefficient to obtain the strategy coefficient.

[0232] Optionally, the expression of the deviation coefficient mapping function is:

[0233]

[0234] where I t represents the initial strategy coefficient in the time period t, β represents the risk preference coefficient, and a and b are the function coefficients of the deviation coefficient mapping function, which are obtained by calling the nonlinear least squares method for function fitting of historical samples.

[0235] The power market trading strategy recommendation device provided by an embodiment of the present invention can execute the power market trading strategy recommendation method provided by any embodiment of the present invention, and has corresponding functional modules and beneficial effects for executing the method.

[0236] Embodiment 4

[0237] Figure 4 FIG. shows a schematic structural diagram of an electronic device 10 that can be used to implement an embodiment of the present invention. The electronic device is intended to represent various forms of digital computers, such as, laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as, personal digital processors, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0238] As Figure 4 shown, the electronic device 10 includes at least one processor 11, and a memory communicatively connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc. The memory stores a computer program executable by the at least one processor. The processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 into the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. The input / output (I / O) interface 15 is also connected to the bus 14.

[0239] Multiple components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a magnetic disk, an optical disc, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0240] The processor 11 may be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the power market trading strategy recommendation method.

[0241] In some embodiments, the power market trading strategy recommendation method may be implemented as a computer program tangibly embodied in a computer-readable storage medium, such as the storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed onto the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the power market trading strategy recommendation method described above may be executed. Alternatively, in other embodiments, the processor 11 may be configured to execute the power market trading strategy recommendation method by any other suitable means (e.g., by means of firmware).

[0242] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuitry, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special or general-purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit the data and instructions to the storage system, the at least one input device, and the at least one output device.

[0243] The computer program for implementing the method of the present invention can be written in any combination of one or more programming languages. These computer programs can be provided to the processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the computer program is executed by the processor, the functions / operations specified in the flowchart and / or block diagram are implemented. The computer program can be executed entirely on the machine, partially on the machine, as a stand-alone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0244] In the context of the present invention, a computer-readable storage medium can be a tangible medium that can contain or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. The computer-readable storage medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. Alternatively, the computer-readable storage medium can be a machine-readable signal medium. More specific examples of the machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0245] For providing interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the electronic device. Other kinds of devices can also be used for providing interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, speech input, or tactile input).

[0246] The systems and techniques described herein can be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer having a graphical user interface or a web browser through which the user can interact with an implementation of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system can be interconnected to each other by any form or medium of digital data communication (e.g., a communication network). Examples of the communication network include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0247] A computing system may include a client and a server. The client and the server are generally far from each other and usually interact via a communication network. The relationship between the client and the server is created by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or a cloud host, which is a host product in the cloud computing service system, solving the defects of difficult management and weak business scalability existing in traditional physical hosts and VPS services.

[0248] It should be understood that various forms of the processes shown above can be used, with steps reordered, added, or deleted. For example, the steps recited in the present invention can be executed in parallel, sequentially, or in a different order, as long as the desired results of the technical solution of the present invention can be achieved, and no limitation is made herein.

[0249] The above specific embodiments do not constitute a limitation on the protection scope of the present invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for recommending electricity market trading strategies, characterized in that, Including: Obtain all city-level meteorological forecast features of the operating day, as well as all historical city-level meteorological features and historical market operation boundary information of the historical associated days. Among them, the historical city-level meteorological features include historical city-level meteorological forecast features and historical city-level meteorological measured features, and the historical market operation boundary information includes historical day-ahead market supply-demand power and historical real-time market supply-demand power; Aggregate each of the operating-day city-level meteorological forecast features, each of the historical city-level meteorological forecast features, and each of the historical city-level meteorological measured features to obtain operating-day provincial meteorological forecast features, historical provincial meteorological forecast features, and historical provincial meteorological measured features; According to the operating-day provincial meteorological forecast features, the historical provincial meteorological forecast features, the historical provincial meteorological measured features, the historical day-ahead market supply-demand power, and the historical real-time market supply-demand power, use an electric power prediction model to predict the operating-day day-ahead market supply-demand power and operating-day real-time market supply-demand power of the operating day; Based on the operating-day day-ahead market supply-demand power and the operating-day real-time market supply-demand power, determine the day-ahead bidding space and the real-time bidding space; According to the day-ahead bidding space and the real-time bidding space, combined with the user's risk preference type, recommend the market trading strategy of the operating day.

2. The method according to claim 1, wherein The operating-day provincial meteorological forecast features include operating-day provincial centralized photovoltaic meteorological forecast features, operating-day provincial distributed photovoltaic meteorological forecast features, operating-day provincial centralized wind power meteorological forecast features, and operating-day provincial overall network load meteorological forecast features; The historical provincial meteorological forecast features include historical provincial centralized photovoltaic meteorological forecast features, historical provincial distributed photovoltaic meteorological forecast features, historical provincial centralized wind power meteorological forecast features, and historical provincial overall network load meteorological forecast features; The historical provincial meteorological measured features include historical provincial centralized photovoltaic meteorological measured features, historical provincial distributed photovoltaic meteorological measured features, historical provincial centralized wind power meteorological measured features, and historical provincial overall network load meteorological measured features.

3. The method according to claim 2, wherein The aggregating each of the operating-day city-level meteorological forecast features, each of the historical city-level meteorological forecast features, and each of the historical city-level meteorological measured features to obtain operating-day provincial meteorological forecast features, historical provincial meteorological forecast features, and historical provincial meteorological measured features includes: When the operating-day city-level meteorological forecast features are used as city-level meteorological features, all city-level meteorological features are divided by meteorological feature elements and then weighted and fused to obtain provincial centralized photovoltaic meteorological features, provincial distributed photovoltaic meteorological features, provincial centralized wind power meteorological features, and provincial overall network load meteorological features, and are respectively used as the operating-day provincial centralized photovoltaic meteorological forecast features, the operating-day provincial distributed photovoltaic meteorological forecast features, the operating-day provincial centralized wind power meteorological forecast features, and the operating-day provincial overall network load meteorological forecast features; When the historical city-level meteorological forecast features are used as city-level meteorological features, all city-level meteorological features are divided according to meteorological feature elements and then weighted and fused to obtain provincial centralized photovoltaic meteorological features, provincial distributed photovoltaic meteorological features, provincial centralized wind power meteorological features, and provincial overall network load meteorological features, which are respectively used as the historical provincial centralized photovoltaic meteorological forecast features, the historical provincial distributed photovoltaic meteorological forecast features, the historical provincial centralized wind power meteorological forecast features, and the historical provincial overall network load meteorological forecast features; When the historical city-level meteorological measured features are used as city-level meteorological features, all city-level meteorological features are divided according to meteorological feature elements and then weighted and fused to obtain provincial centralized photovoltaic meteorological features, provincial distributed photovoltaic meteorological features, provincial centralized wind power meteorological features, and provincial overall network load meteorological features, which are respectively used as the historical provincial centralized photovoltaic meteorological measured features, the historical provincial distributed photovoltaic meteorological measured features, the historical provincial centralized wind power meteorological measured features, and the historical provincial overall network load meteorological measured features.

4. The method according to claim 3, characterized in that, The step of dividing all city-level meteorological features according to meteorological feature elements and then weighted and fused to obtain provincial centralized photovoltaic meteorological features, provincial distributed photovoltaic meteorological features, provincial centralized wind power meteorological features, and provincial overall network load meteorological features includes: Dividing all city-level meteorological features into basic meteorological features, light-related meteorological features, wind-related meteorological features, and extreme meteorological features; Based on the share of centralized photovoltaic installed capacity in the city, weighted and fuse each of the basic meteorological features, each of the light-related meteorological features, and each of the extreme meteorological features to obtain provincial centralized photovoltaic meteorological features; Based on the share of distributed photovoltaic installed capacity in the city, weighted and fuse each of the basic meteorological features, each of the light-related meteorological features, and each of the extreme meteorological features to obtain provincial distributed photovoltaic meteorological features; Based on the share of centralized wind power installed capacity in the city, weighted and fuse each of the basic meteorological features, each of the wind-related meteorological features, and each of the extreme meteorological features to obtain provincial centralized wind power meteorological features; Based on the share of electricity consumption in the city, weighted and fuse each of the basic meteorological features and each of the extreme meteorological features to obtain provincial overall network load meteorological features.

5. The method according to claim 2, wherein The historical day-ahead market supply and demand power includes historical day-ahead cleared centralized photovoltaic power, historical day-ahead cleared distributed photovoltaic power, historical day-ahead cleared centralized wind power, and historical day-ahead overall network load; The historical real-time market supply and demand power includes historical real-time centralized photovoltaic power, historical real-time distributed photovoltaic power, historical real-time centralized wind power, and historical real-time overall network load; Correspondingly, the operating day's day-ahead market supply and demand power includes operating day's day-ahead centralized photovoltaic power, operating day's day-ahead distributed photovoltaic power, operating day's day-ahead centralized wind power, and operating day's day-ahead overall network load; The real-time market supply and demand power on the operation day includes the real-time centralized photovoltaic power on the operation day, the real-time distributed photovoltaic power on the operation day, the real-time centralized wind power on the operation day and the real-time whole network load on the operation day.

6. The method according to claim 5, wherein The method uses an electric power prediction model to predict the market supply and demand power on the day before the operation date and the real-time market supply and demand power on the day of the operation date based on the provincial meteorological forecast characteristics on the operation date, the historical provincial meteorological forecast characteristics, the historical provincial meteorological measured characteristics, the historical day-ahead market supply and demand power, and the historical real-time market supply and demand power, including: According to the provincial centralized photovoltaic meteorological forecast characteristics on the operation day, the historical provincial centralized photovoltaic meteorological forecast characteristics and the historical day-ahead cleared centralized photovoltaic power, an electric power prediction model is used to obtain the day-ahead centralized photovoltaic power on the operation day; According to the provincial distributed photovoltaic meteorological forecast characteristics on the operation day, the historical provincial distributed photovoltaic meteorological forecast characteristics and the historical day-ahead cleared distributed photovoltaic power, the electric power prediction model is used to obtain the day-ahead distributed photovoltaic power on the operation day; According to the provincial centralized wind power meteorological forecast characteristics on the operation day, the historical provincial centralized wind power meteorological forecast characteristics and the historical day-ahead centralized wind power clearing, the electric power prediction model is used to obtain the day-ahead centralized wind power on the operation day; According to the meteorological forecast characteristics of the provincial-level full-grid load on the operation day, the meteorological forecast characteristics of the historical provincial-level full-grid load and the historical full-grid load cleared a day ago, the electric power prediction model is used to obtain the full-grid load a day before the operation day; According to the provincial centralized photovoltaic meteorological forecast characteristics of the operation day, the historical provincial centralized photovoltaic meteorological measured characteristics and the historical real-time centralized photovoltaic power, the electric power prediction model is used to obtain the real-time centralized photovoltaic power of the operation day; According to the provincial distributed photovoltaic meteorological forecast characteristics on the operation day, the historical provincial distributed photovoltaic meteorological measured characteristics and the historical real-time distributed photovoltaic power, the electric power prediction model is used to obtain the real-time distributed photovoltaic power on the operation day; According to the provincial centralized wind power meteorological forecast characteristics on the operation day, the historical provincial centralized wind power meteorological measured characteristics and the historical real-time centralized wind power, the electric power prediction model is used to obtain the real-time centralized wind power on the operation day; The real-time network load on the operation day is obtained by using the electric power prediction model based on the meteorological forecast characteristics of the provincial network load on the operation day, the measured meteorological characteristics of the historical provincial network load and the historical real-time network load.

7. The method according to claim 5, wherein The determining of the day-ahead bidding space and the real-time bidding space based on the day-ahead market supply and demand power and the real-time market supply and demand power on the operation day includes: Obtain the pre-disclosed power of local power plants, the planned power of the day-ahead interconnection lines, and the planned power of the real-time interconnection lines; Taking the sum of the centralized PV power before the operation date, the distributed PV power before the operation date, the centralized wind power before the operation date, the pre-disclosed power of the local power plants, and the planned power of the tie lines before the operation date as the power supply side power before the operation date, and taking the difference between the total network load before the operation date and the power supply side power before the operation date as the bidding space before the operation date; Taking the sum of the real-time centralized PV power on the operation date, the real-time distributed PV power on the operation date, the real-time centralized wind power on the operation date, the pre-disclosed power of the local power plants, and the planned power of the tie lines in real time as the power supply side power in real time, and taking the difference between the total network load in real time on the operation date and the power supply side power in real time as the bidding space in real time.

8. The method according to claim 1, wherein According to the bidding space before the operation date and the bidding space in real time, and combining with the user's risk preference type, recommending the market trading strategy for the operation date, including: Taking the difference between the bidding space before the operation date and the bidding space in real time as the deviation of the bidding space; According to the user's risk preference type and the deviation of the bidding space, using a preset deviation coefficient mapping function, and combining with the installed capacity of the power station, determining the strategy coefficient; Based on the strategy coefficient, determining the market trading strategy for the operation date.

9. The method according to claim 8, wherein According to the user's risk preference type and the deviation of the bidding space, using a preset deviation coefficient mapping function, and combining with the installed capacity of the power station, determining the strategy coefficient, including: Obtaining the user's risk preference type and determining the risk preference coefficient corresponding to the user's risk preference type; Substituting the deviation of the bidding space and the risk preference coefficient into the preset deviation coefficient mapping function to obtain the initial strategy coefficient; Obtaining the installed capacity of the power station and the initial power of the power station, and determining the maximum output power of the power station according to the installed capacity of the power station and the initial power of the power station; Based on the maximum output power of the power station, adjusting the initial strategy coefficient to obtain the strategy coefficient.

10. The method according to claim 9, wherein The expression of the deviation coefficient mapping function is: Among them, I t represents the initial strategy coefficient within the time period t, β represents the risk preference coefficient, and a and b are the function coefficients of the deviation coefficient mapping function, which are obtained by performing function fitting on historical samples by calling the nonlinear least squares method.

11. A power market trading strategy recommendation device, characterized in that, Including: A prediction data acquisition module, configured to acquire all the city-level meteorological forecast features of the operation date, as well as all the historical city-level meteorological features and historical market operation boundary information of the historical associated dates, where the historical city-level meteorological features include historical city-level meteorological forecast features and historical city-level meteorological measured features, and the historical market operation boundary information includes historical power supply and demand power in the day-ahead market and historical power supply and demand power in the real-time market; A meteorological feature aggregation module, configured to aggregate each of the city-level meteorological forecast features of the operation date, each of the historical city-level meteorological forecast features, and each of the historical city-level meteorological measured features, to obtain provincial-level meteorological forecast features of the operation date, historical provincial-level meteorological forecast features, and historical provincial-level meteorological measured features; A market power prediction module, configured to predict the power supply and demand power in the day-ahead market and the power supply and demand power in the real-time market of the operation date by using a power prediction model according to the provincial-level meteorological forecast features of the operation date, the historical provincial-level meteorological forecast features, the historical provincial-level meteorological measured features, the historical power supply and demand power in the day-ahead market, and the historical power supply and demand power in the real-time market; A bidding space determination module, configured to determine a day-ahead bidding space and a real-time bidding space based on the day-ahead market supply and demand power before the operation day and the real-time market supply and demand power on the operation day; A trading strategy recommendation module, configured to recommend a market trading strategy for the operation day according to the day-ahead bidding space and the real-time bidding space, in combination with the user risk preference type.

12. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the electricity market trading strategy recommendation method according to any one of claims 1-10.

13. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions, and the computer instructions are used to implement the electricity market trading strategy recommendation method according to any one of claims 1-10 when executed by a processor.

Images