CN-122026369-A - Price difference transaction auxiliary decision-making method based on day-ahead price and related equipment

CN122026369ACN 122026369 ACN122026369 ACN 122026369ACN-122026369-A

Abstract

The invention relates to the technical field of electric power transaction management, in particular to a price difference transaction auxiliary decision-making method based on a day-ahead price and related equipment; the method comprises the steps of carrying out short-term prediction by using a pre-trained price difference prediction model to generate a price difference value and a confidence interval of electricity price prediction before time sharing, establishing a benefit objective function by taking the price difference as a core variable, optimizing internal and external arbitrage combinations of cross-date and time intervals by combining a basic rolling direction and electric quantity to generate a decision strategy, carrying out online self-adaptive correction and risk control on the strategy based on real-time disc opening data and transaction feedback to form a closed-loop decision system, and improving transaction efficiency and benefit certainty.

Inventors

Ma Zhien
DONG XIWEI
Qiu Zhuocheng
LI JUNQING
CHEN JIHUI
GAO WENQING
XU WENTAO
LIU SHIXUAN
ZHANG JINGMIN
LI YU
LU ZHIWEN
LI JIAN
XU MINGLIANG

Assignees

华能国际电力股份有限公司湖南清洁能源分公司

Dates

Publication Date: 20260512
Application Date: 20260123

Claims (10)

1. A day-ahead price-based price difference transaction aid decision-making method, comprising: The method comprises the steps of obtaining a ten-day degree rule boundary parameter of an electric power market, wherein the rule boundary parameter comprises a ten-day degree contract electric quantity lower limit proportion K1, a contract electric quantity upper limit proportion K2, a first recovery coefficient h1 and a second recovery coefficient h2; calculating a target load rate interval of the current ten-day residual period based on the rule boundary parameters, and determining a basic rolling direction and a basic rolling quantity according to the target load rate interval; Short-term prediction is carried out on the predicted price difference of the day-ahead price-market rolling price based on a pre-trained price difference prediction model, a time-sharing day-ahead electricity price prediction price difference value is obtained, and a corresponding confidence interval is generated; establishing a benefit objective function according to the day-ahead electricity price prediction valence difference value, taking a basic rolling direction and a basic rolling quantity as input, and optimizing an internal arbitrage and an external arbitrage combination crossing date and time intervals based on the benefit objective function to obtain an optimized decision strategy; and on-line self-adaptive correction and risk control are carried out on the optimized decision strategy based on real-time disk data and transaction feedback, so as to form a closed-loop decision.
2. The price difference transaction aiding decision-making method based on the day-ahead price according to claim 1, wherein the calculating the current ten-day remaining days should load rate comprises: Accumulating the current ten-day-degree clear internet surfing electric quantity and the middle-and-long-term contract electric quantity, and reversely calculating to obtain a target load rate interval of the remaining days by combining the ten-day-degree contract electric quantity lower limit proportion K1 and the contract electric quantity upper limit proportion K2.
3. The method for assisting decision-making of price difference transaction based on day-ahead price according to claim 1, wherein the basic rolling direction determination comprises comparing a current ten-day total contract electric quantity, a product of a contract electric quantity upper limit proportion K2 and a ten-day accumulated spot-charge settlement electric quantity, and a product of a contract electric quantity lower limit proportion K1 and a ten-day accumulated spot-charge settlement electric quantity, and determining a buyer warehouse-reduction or seller warehouse-increase direction by combining preset safety margin parameters.
4. The day-ahead price-based price difference transaction aid decision-making method according to claim 1, wherein the price difference prediction model adopts a time series analysis algorithm model.
5. The day-ahead price-based price difference transaction aid decision-making method according to claim 4, wherein the time series analysis algorithm model adopts an ARIMA model.
6. The price difference transaction assistant decision-making method based on the day-ahead price according to claim 4, wherein the time series analysis algorithm model is a combination model of an ARIMA model and a machine learning model, and the machine learning model adopts an LSTM model, a GRU model or a XGBoost model.
7. The day-ahead price-based price difference transaction aid decision-making method according to claim 1, wherein the step of establishing a revenue objective function comprises: setting a benefit objective function taking price difference as a core variable, and covering an inner benefit and an outer benefit; Embedding a ten-day contract electric quantity lower limit proportion K1 and a contract electric quantity upper limit proportion K2 in the income objective function, and restraining a first recovery coefficient h1 and a second recovery coefficient h2, and optimizing the listing price/amount of different time intervals and across days.
8. The price difference transaction auxiliary decision-making system based on the daily price is characterized by comprising a rule boundary calculation module, a price difference prediction module, an objective function optimization module and a closed loop execution module; The rule boundary calculation module is used for obtaining the rule boundary parameters of the power market, calculating the target load rate interval of the current ten-day residual period based on the rule boundary parameters, determining the basic rolling direction and the basic rolling quantity according to the target load rate interval, The price difference prediction module carries out short-term prediction on the expected price difference of the day-ahead price-market rolling price based on a pre-trained price difference prediction model to obtain a time-sharing day-ahead electricity price prediction price difference value, and generates a corresponding confidence interval; the objective function optimization module establishes a benefit objective function according to the current price prediction valence difference value, takes a basic rolling direction and a basic rolling quantity as input, optimizes internal arbitrage and external arbitrage combinations crossing dates and time intervals based on the benefit objective function, and obtains an optimized decision strategy; The closed-loop execution module is used for carrying out on-line self-adaptive correction and risk control on the optimized decision strategy based on real-time disk data and transaction feedback to form a closed-loop decision.
9. A computer readable storage medium storing one or more programs, the one or more programs comprising instructions, which when executed by a computing device, cause the computing device to perform the daily price based price differential transaction assistance decision making method of any one of claims 1 to 7.
10. A computing device, comprising: one or more processors, memory, and one or more programs, wherein one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs comprising steps for performing the daily price based price differential transaction assistance decision making method of any one of claims 1 to 7.

Description

Price difference transaction auxiliary decision-making method based on day-ahead price and related equipment Technical Field The invention relates to the technical field of power transaction management, in particular to a price difference transaction auxiliary decision-making method based on a day-ahead price and related equipment. Background The electric power market in China forms a double-track pattern of 'medium-long term and spot-commodity cooperation', medium-long term trade provides reference guarantee, spot-commodity market reflects real-time supply and demand, rolling matching trade in day/ten-day/month is a key link for connecting the medium-long term contract and the spot-commodity, fine-grained dynamic optimization is carried out on the medium-long term contract under the scenes of clear result updating, unit adjustment, load fluctuation and the like, and related enterprises form a standard flow covering electric quantity prediction, arbitrage decision-making, trade execution and duplication. The main technical path of the current industry around the price rolling-day front price is mainly three, namely firstly, a benefit association model is built based on a psychological expected price, an optimal reporting strategy is obtained through an intelligent algorithm, secondly, a fixed threshold is set according to a historical price difference, buying and selling operations are carried out in combination with target electric quantity, a price of a coil is hung according to the price of the coil, thirdly, high-frequency matching is carried out through the fixed price difference based on the price of the coil and the depth of the coil, simple loss prevention wind control is assisted, and part of schemes can adopt models such as ARIMA to predict electricity price or price difference to support decision. However, the prior art still has many inherent technical limitations in practical application, the mainstream scheme relies on experiences or heuristic methods to judge transaction directions and price in sections, a unified benefit objective function taking the price difference of 'price before day-market rolling and pinching price expectation' as a core is lacked, internal and external, date and time division arbitrage decisions are difficult to cooperate under the same optimization framework, K1/K2 performance proportion and h1/h2 recovery coefficient in the medium and long term transaction rules are often corrected afterwards, explicit embedding is not performed in a strategy generation stage, excessive/deficiency recovery loss or benefit leakage is easily caused, price prediction is performed by adopting models such as ARIMA, but linkage on-line correction and loss stopping mechanisms of prediction uncertainty and disk opening change are insufficient, strategy execution self-adaptability is weak, single point prediction matching with a disk-following mode does not introduce confidence interval and cost constraint and is easy to amplify prediction errors, the market signals which are difficult to meet medium and long term-spot full cooperation are difficult to effectively accepted at the same time constraint of the medium and long term rule, and the market signals are not easy to be accepted in addition, and the method of simple rolling and pinching is not easy to realize the single-state error under the conditions of low price node or low-price error, and high-speed market stability. Disclosure of Invention The invention aims to solve the technical problems of non-uniform target driving and insufficient rule and boundary coupling in the conventional power market day-rolling matching transaction decision. The invention aims at realizing the following technical scheme: in a first aspect, the present invention provides a price difference transaction auxiliary decision method based on a day-ahead price, comprising: The method comprises the steps of obtaining a ten-day degree rule boundary parameter of an electric power market, wherein the rule boundary parameter comprises a ten-day degree contract electric quantity lower limit proportion K1, a contract electric quantity upper limit proportion K2, a first recovery coefficient h1 and a second recovery coefficient h2; calculating a target load rate interval of the current ten-day residual period based on the rule boundary parameters, and determining a basic rolling direction and a basic rolling quantity according to the target load rate interval; Short-term prediction is carried out on the predicted price difference of the day-ahead price-market rolling price based on a pre-trained price difference prediction model, a time-sharing day-ahead electricity price prediction price difference value is obtained, and a corresponding confidence interval is generated; establishing a benefit objective function according to the day-ahead electricity price prediction valence difference value, taking a basic rolling direction and a basic rolling quantity as input, and o