Procurement Strategy Optimization Of Thermal Power Enterprises Based On Role Analysis For Short-term Coal Price Forecasting

Thermal coal is a key raw material in the production process of power plants,and the production and operation profit and cost of power plants are all centered on thermal coal.With the increasing competitive pressure brought by power market and clean energy utilization to thermal power enterprises,thermal power enterprises urgently need to adjust their business strategies,reduce costs and improve their competitiveness.In the operation activities of thermal power enterprises,the cost related to thermal coal accounts for more than 50% of the total operating cost of enterprises.Therefore,reducing the cost of thermal coal has an important impact on reducing the cost of power plants.However,the purchase plan of thermal coal involves multiple links such as power generation plan,inventory planning,procurement and transportation of the power plant,and the price of thermal coal fluctuates with market demand under the market economy,which affects the cost of each control link of the power plant.Therefore,it has become an important research direction of power plant cost optimization to summarize and analyze the change law of thermal coal price in the market and formulate appropriate purchase plan in combination with each link of power plant thermal coal control.In this paper,the procurement strategy optimization of power plant is taken as the research object,and each link of the procurement process of power plant and its cost generation process are analyzed.Combined with the current situation that the thermal coal price fluctuates with the market demand under the background of thermal coal marketization,the procurement strategy optimization based on the role analysis of GAN short-term thermal coal price prediction is proposed.Firstly,in order to establish an appropriate power market model,a three-role model under the complex system of thermal coal is established through system dynamics with the system and feedback as the basic elements,and the mathematical relations among variables are described to provide a theoretical basis for selecting appropriate indicators.Secondly,the VAR model is used to test the variable indexes,and the appropriate relevant variable indexes are screened to provide an accurate feature set for the prediction model.Then the historical feature mapping data set is established for the feature space to solve the problems of missing features and the inconspicuous learning of change rules in the process of prediction.In addition,seasonal difference was used to extract the trend sequence from the thermal coal price change sequence,so as to ensure the accuracy of the sequence for the study of price change law and eliminate the influence of cycle change law on the accuracy of the model.Then,a GAN adversation network was established,with LSTM as generator and CNN as discriminator,to integrate the prediction results of the three-role model in complex markets and obtain more accurate prediction results.Finally,the complete purchasing process of thermal coal in power plant is analyzed,and an optimization model is established for the purchasing plan of power plant,and the optimal purchasing strategy of thermal coal in power plant is solved by using chaotic improved particle swarm optimization algorithm.The validity of the proposed method for optimizing the purchasing strategy of power plant is verified by setting the purchasing case of power plant.