Data-driven Electricity Market Analysis

Since further deepening the reform of the electricity system in 2015,the construction process of China’s electricity market has accelerated.In order to further realize the sharing,mutual assistance,and optimal allocation of electric power resources on a larger scale,and improve the stability and flexible adjustment ability of the electric power system,China has accelerated the construction of a unified national electric power market system.The proportion of wind power and energy storage in the electricity spot market is continuously increasing,but the randomness of wind power output and the uncertainty of electricity prices are not conducive to the formulation of bidding strategies on the power generation side,while thermal power units and energy storage can effectively address uncertainty.Therefore,this article first predicts the day-ahead clearing electricity price,and on this basis,studies the"wind fire storage" system participating in the day-ahead market bidding strategy to reduce carbon emissions,absorb clean energy,and improve bidding revenue.The main research results of this article are as follows:Firstly,it briefly introduces the current situation and development process of domestic and foreign power markets,as well as the application of data driven research methods in the field of power market.Secondly,a graph convolution residual long and short term memory neural network model for predicting day-ahead clearing price considering topological structure information and power flow is proposed.Aiming at the existence of power transmission between different price zones,a graph convolution neural network is used to extract the topological structure information between price zones;Using residual network to extract the connection of electricity prices at different time scales;The short-term and short-term memory neural network is used to extract the characteristics of other factors that affect the electricity price.Finally,the short-term and short-term memory neural network with attention mechanism is used to fuse the characteristics of the three branches and predict the day-ahead clearing electricity price of the next day.The validity of the prediction model is verified by the actual operation data of four price zones in Northern Europe.Then,based on different electricity price scenarios in electricity price zones,the bidding strategy of "wind storage" system participating in the electricity market is studied.Considering the uncertainty of daily clearing electricity prices,the daily electricity prices in the price zone are clustered into five typical scenarios.Combining the electricity price scenario and system operation constraints,with the maximum system revenue as the objective function,a day-ahead bidding strategy model for the "wind storage" system is established,and the bidding strategy is evaluated from two aspects:the operation of energy storage equipment and the allocation of energy storage capacity.Finally,the"wind storage" system is extended to a "wind fire storage" system,and a bidding strategy model considering conditional risk value and carbon market trading is proposed.Considering the uncertainty of wind power output,conditional risk value is used to characterize the bidding strategy model.Considering the carbon emissions during the operation of traditional thermal power units,the relevant constraints of the carbon market are introduced into the system day-ahead bidding strategy model.Using MATLAB to solve the bidding strategy model,the system bidding volume and energy storage charging and discharging capacity are obtained.On this basis,the forward bidding strategy is analyzed and discussed from both risk weighting coefficients and carbon trading.