Coal Inventory Optimization Of H Thermal Power Plant Based On Thermal Coal Price Trend Analysis

As a cogeneration power plant put into operation by G Group in B province,H power plant undertakes the power supply and winter heat supply in the surrounding areas.At the same time,as a peak shaving power plant,it has made outstanding contributions to the deep peak shaving task of the power grid.At present,the power plant mainly includes three kinds of coal,which are SH coal signed with the group ’s internal coal plant,high calorific value coal and economic coal signed with other state-owned coal mines.The power plant lacks the early grasp of the power coal market,and the inventory strategy formulation does not consider the impact of coal blending and deep peak shaving on the power coal inventory management strategy.The single-stage monthly inventory plan cannot meet the power plant management needs,which is not conducive to the cost reduction and efficiency increase of thermal power enterprises.In view of the problems existing in H thermal power plant under the above background,this paper first analyzes the influencing factors of coal price and determines the lead time of related factors on the basis of consulting domestic and foreign references,collecting historical economic environment data and power plant data collection.A variety of machine learning algorithms are used to predict the trend of monthly,weekly and daily coal price changes,and the prediction accuracy and stability of each model method are analyzed and compared.Then,taking the week as the cycle,the different types of coal inventory of H thermal power plant are optimized on the basis of fully considering the problems of complex coal source blending,power plant production requirements and coal source supply.Aiming at maximizing the inventory value of thermal coal and minimizing the inventory cost,a dynamic multi-objective optimization problems(DMOPs)model is constructed,and the genetic algorithm(GA)is used to quickly solve the model.Finally,the real data of the power plant is used for simulation comparison to verify the feasibility and optimization effect of the inventory optimization model.The research of this paper helps H thermal power enterprises to mine the influencing factors of power coal price and predict the price of power coal.At the same time,the dynamic optimization model of power coal inventory constructed by using the prediction results improves the scientificity of inventory strategy formulation of thermal power enterprises.The results show that the Adaboost Random Forest(Ada-RF)model has better prediction effect on monthly coal price index,and the Informer model has better prediction effect on weekly and daily long-term coal price trend.At the same time,the Ada-RF model retains the interpretability of the model while accurately predicting the coal price,and the Informer model solves the long-term dependence of the time series prediction model.The weekly coal price trend prediction results obtained by using the daily coal price trend prediction results can better help thermal power enterprises to achieve scientific and refined coal inventory management.Based on the prediction results of coal price trend,the coal inventory optimization model constructed in this study reduces the production cost of power plants and ensures the safety production of thermal power enterprises on the basis of ensuring the safety production,blending requirements of thermal power enterprises and enhance the competitiveness of enterprises.At the same time,the research has practical significance for thermal power enterprises to participate in deep peak shaving and ensure regional energy supply.