CN-121998363-A - Virtual deduction method and system for promoting electric power market transaction

CN121998363ACN 121998363 ACN121998363 ACN 121998363ACN-121998363-A

Abstract

The embodiment of the invention provides a virtual deduction method and a virtual deduction system for promoting electric power market trading, and belongs to the technical field of electric power trading simulation. The virtual deduction method comprises the steps of obtaining multi-source heterogeneous distributed data related to electric power operation, processing the multi-source heterogeneous distributed data to obtain processed data, predicting energy data and requirements in a future period of time based on the processed data after the processed data are obtained, constructing a transaction rule under the common influence of the energy data and the requirements on the basis of an engine rule of current electric power market transaction, and virtually deducting according to the transaction rule and the predicted energy data and requirements in the future period of time to obtain a deduction result of the market transaction. The virtual deduction method can simulate the transaction process through a virtual deduction technology so as to predict and evaluate the effects of different transaction strategies.

Inventors

SUN HONGYAN
LIU JUN
WANG LILI
QI YURONG
WANG PENGCHENG
ZHANG RUI
WANG HUAN
ZHANG SHANRUI
ZHANG LANXI

Assignees

四川中电启明星信息技术有限公司
国网信息通信产业集团有限公司

Dates

Publication Date: 20260508
Application Date: 20260129

Claims (10)

1. A virtual deduction method for facilitating electric power market transactions, the virtual deduction method comprising: Acquiring multi-source heterogeneous distributed data about power operation, and processing the multi-source heterogeneous distributed data to obtain processed data; After the processing data is acquired, predicting energy data and requirements in a future period of time based on the processing data; Based on the engine rule of the current electric power market transaction, constructing a transaction rule under the common influence of the energy data and the demand forecast; And virtually deducting according to the transaction rules and the predicted energy data and requirements in a future period of time to obtain a deduction result of the market transaction.
2. The virtual deduction method according to claim 1, wherein the multi-source heterogeneous distributed data comprises distributed energy data, namely historical output of photovoltaic power, energy storage SOC, charge and discharge power, equipment state, inverter/fan operation parameters, load data, namely historical data of user side electricity, heat and cold loads, wherein the resolution is 15 minutes or 1 hour, meteorological data, namely historical and predicted wind speed, wind direction, illumination intensity, temperature, humidity and cloud cover, market data, namely historical electricity price, transaction rules and network blocking information, and power grid data, namely network topology, line capacity, node voltage constraint and protection fixed value.
3. The virtual deduction method according to claim 2, wherein the multi-source heterogeneous distributed data on the power operation is acquired and processed to obtain the processed data, comprising: Acquiring the multi-source heterogeneous distributed data, and processing the missing value and the abnormal value in the multi-source heterogeneous distributed data; After the processing is finished, unifying all the data to the same time stamp, and constructing time sequence characteristics of the processed multi-source heterogeneous distributed data according to the characteristics of hours, weeks, holidays and seasons based on the time stamp; weather data in the multi-source heterogeneous distributed data are built into weather derivative features according to effective illumination time length, temperature difference and wind speed cubes; and carrying out normalization processing on all the characteristics to obtain final processing data.
4. The virtual deduction method according to claim 1, wherein after acquiring the processing data, predicting energy data and demand for a future period of time based on the processing data, comprises: Dividing multisource heterogeneous distributed data about power operations into a training set and a validation set; The data of the training set is sent into a distributed energy prediction model taking LSTM as a core so as to train; After the data training of the training set is completed, the data of the verification set is sent into a distributed energy prediction model, and root mean square error and continuous ranking probability score verification are carried out according to the output result of the distributed energy prediction model; after verification and training are completed, the processing data are sent to the trained distributed energy prediction model, so that energy data and requirements in a future period of time are obtained.
5. The virtual deduction method according to claim 4, wherein constructing a transaction rule under the combined influence of the energy data and the demand based on the engine rule of the current electric power market transaction comprises: Encoding engine rules simulating a target power market, wherein the engine rules comprise transaction rules, timelines, quotation formats and clearing algorithms; constructing an intelligent agent for executing the strategy and an intelligent agent for clearing the market according to the engine rules, wherein the intelligent agent for executing the strategy integrates energy data and requirements inside the intelligent agent for executing the strategy, and the intelligent agent for clearing the market can simulate the clearing of the market according to the quotations of all participants to generate clearing electricity price and power; The method comprises the steps that an intelligent agent is executed through a strategy execution mode, and an optimal decision is solved through an optimization algorithm based on energy data and requirements, current physical states, market information and a preset transaction strategy; The method comprises the steps that an intelligent agent for executing the strategy submits the solved optimal decision to a market clearing agent, and the market clearing agent simulates clearing according to each submitted decision and returns a winning result and market price; And executing actual charge and discharge/power generation instructions by the intelligent agent according to the winning result, and updating the states of all the devices.
6. A virtual deduction system for facilitating electric market transactions, the virtual deduction system comprising: The data acquisition module acquires multi-source heterogeneous distributed data about power operation and processes the multi-source heterogeneous distributed data to obtain processed data; The prediction module predicts energy data and requirements in a future period of time based on the processing data after the processing data are acquired; And the virtual deduction module constructs a transaction rule under the joint influence of the energy data and the demand forecast based on the engine rule of the current electric power market transaction, and performs virtual deduction according to the transaction rule and the energy data and the demand in the forecast future period of time so as to obtain a deduction result of the market transaction.
7. The virtual deduction system according to claim 6, wherein the multi-source heterogeneous distributed data comprises distributed energy data, wherein the distributed energy data comprises historical output of photovoltaic/wind power, energy storage SOC, charge and discharge power, equipment states and inverter/fan operation parameters, load data comprises historical data of user side electricity, heat and cold loads, resolution is 15 minutes or 1 hour, meteorological data comprises historical and forecast wind speed, wind direction, illumination intensity, temperature, humidity and cloud amount, market data comprises historical electricity price, transaction rules and network blocking information, and power grid data comprises network topology, line capacity, node voltage constraint and protection fixed values.
8. The virtual deduction system of claim 7, wherein the multi-source heterogeneous distributed data about the power operation is obtained and processed to obtain processed data, comprising: Acquiring the multi-source heterogeneous distributed data, and processing the missing value and the abnormal value in the multi-source heterogeneous distributed data; After the processing is finished, unifying all the data to the same time stamp, and constructing time sequence characteristics of the processed multi-source heterogeneous distributed data according to the characteristics of hours, weeks, holidays and seasons based on the time stamp; weather data in the multi-source heterogeneous distributed data are built into weather derivative features according to effective illumination time length, temperature difference and wind speed cubes; and carrying out normalization processing on all the characteristics to obtain final processing data.
9. The virtual deduction system according to claim 6, wherein after acquiring the processing data, predicting energy data and demand for a future period of time based on the processing data comprises: Dividing multisource heterogeneous distributed data about power operations into a training set and a validation set; The data of the training set is sent into a distributed energy prediction model taking LSTM as a core so as to train; After the data training of the training set is completed, the data of the verification set is sent into a distributed energy prediction model, and root mean square error and continuous ranking probability score verification are carried out according to the output result of the distributed energy prediction model; after verification and training are completed, the processing data are sent to the trained distributed energy prediction model, so that energy data and requirements in a future period of time are obtained.
10. The virtual deduction system of claim 9, wherein constructing a transaction rule under the combined influence of the energy data and the demand based on the engine rule of the current electric market transaction comprises: Encoding engine rules simulating a target power market, wherein the engine rules comprise transaction rules, timelines, quotation formats and clearing algorithms; constructing an intelligent agent for executing the strategy and an intelligent agent for clearing the market according to the engine rules, wherein the intelligent agent for executing the strategy integrates energy data and requirements inside the intelligent agent for executing the strategy, and the intelligent agent for clearing the market can simulate the clearing of the market according to the quotations of all participants to generate clearing electricity price and power; The method comprises the steps that an intelligent agent is executed through a strategy execution mode, and an optimal decision is solved through an optimization algorithm based on energy data and requirements, current physical states, market information and a preset transaction strategy; The method comprises the steps that an intelligent agent for executing the strategy submits the solved optimal decision to a market clearing agent, and the market clearing agent simulates clearing according to each submitted decision and returns a winning result and market price; And executing actual charge and discharge/power generation instructions by the intelligent agent according to the winning result, and updating the states of all the devices.

Description

Virtual deduction method and system for promoting electric power market transaction Technical Field The invention relates to the technical field of electric power trade simulation, in particular to a virtual deduction method and a virtual deduction system for promoting electric power market trade. Background With the improvement of the renewable energy duty ratio, the volatility and the uncertainty of the renewable energy duty ratio bring higher requirements to market balance, deduction can simulate a high-proportion new energy scene, verify whether the market can effectively coordinate flexible resources, and promote the formation of a novel power system with stronger new energy consumption capability. Accordingly, a virtual deduction method and system for facilitating electric market trading is constructed to predict and evaluate the effects of different trading strategies by simulating the trading process using virtual deduction techniques in the trading environment space. Disclosure of Invention An object of an embodiment of the present invention is to provide a virtual deduction method and system for promoting electric power market trading, which can simulate a trading process through a virtual deduction technology to predict and evaluate effects of different trading strategies. In order to achieve the above object, an embodiment of the present invention provides a virtual deduction method for promoting an electric power market transaction, the virtual deduction method including: Acquiring multi-source heterogeneous distributed data about power operation, and processing the multi-source heterogeneous distributed data to obtain processed data; After the processing data is acquired, predicting energy data and requirements in a future period of time based on the processing data; Based on the engine rule of the current electric power market transaction, constructing a transaction rule under the common influence of the energy data and the demand forecast; And virtually deducting according to the transaction rules and the predicted energy data and requirements in a future period of time to obtain a deduction result of the market transaction. The multi-source heterogeneous distributed data comprises distributed energy data, load data, meteorological data, historical and forecast wind speed, wind direction, illumination intensity, temperature, humidity and cloud cover, market data, historical electricity price, transaction rules and network blocking information, power grid data, network topology, line capacity, node voltage constraint and protection fixed values, wherein the historical output of photovoltaic/wind power, energy storage SOC, charging and discharging power, equipment state and inverter/fan operation parameters, the load data comprises historical data of user side electricity, heat and cold loads, the resolution is 15 minutes or 1 hour. Optionally, the multi-source heterogeneous distributed data about the power operation is acquired and processed to obtain processed data, including: Acquiring the multi-source heterogeneous distributed data, and processing the missing value and the abnormal value in the multi-source heterogeneous distributed data; After the processing is finished, unifying all the data to the same time stamp, and constructing time sequence characteristics of the processed multi-source heterogeneous distributed data according to the characteristics of hours, weeks, holidays and seasons based on the time stamp; weather data in the multi-source heterogeneous distributed data are built into weather derivative features according to effective illumination time length, temperature difference and wind speed cubes; and carrying out normalization processing on all the characteristics to obtain final processing data. Optionally, after acquiring the processing data, predicting energy data and requirements over a future period of time based on the processing data, including: Dividing multisource heterogeneous distributed data about power operations into a training set and a validation set; The data of the training set is sent into a distributed energy prediction model taking LSTM as a core so as to train; After the data training of the training set is completed, the data of the verification set is sent into a distributed energy prediction model, and root mean square error and continuous ranking probability score verification are carried out according to the output result of the distributed energy prediction model; after verification and training are completed, the processing data are sent to the trained distributed energy prediction model, so that energy data and requirements in a future period of time are obtained. Optionally, constructing a transaction rule under the combined influence of the energy data and the demand based on the engine rule of the current electric power market transaction includes: Encoding engine rules simulating a target power market, wherein the engine rules comprise tr