Optimal Scheduling Of Steam In Iron And Steel Industry

Energy is power,which to promote the development and production of Iron & Steel enterprises.Iron & Steel Industry is not only the pillar industry of national economy,but also an energy-intensive industry,in promoting the economic development at the same time,also increasing the energy consumption.The energy consumption in Iron & Steel enterprises accounted for 15% of the national energy consumption,saving energy and reducing consumption has become the iron and steel enterprises urgently needs to solve the problem.The steam is a very important energy resource medium in steel industry in the process of production,its proportion in total energy consumption in iron and steel enterprises consume more than 10%.Therefore,the optimal scheduling of steam plays an important role on saving energy and reducing consumption in iron and steel enterprises.This paper based on iron and steel enterprise steam system as the research object,the optimal schedule model were studied,in order to achieve the purpose of saving energy and reducing consumption in iron and steel enterprises.The main research results are as follows:(1)Due to the production of steam production consumption of irregular change,which make the steam pipe network pressure to present the irregular change,existing the certain errors between the detection value and the actual value,affecting the precision of the steam pipe network pressure prediction.Therefore,a wavelet transform-Bayesian Neural Network method is selected.Through the instance and model comparisons,which prove the prediction accuracy is higher.(2)Due to fluctuation of coking process steam production is rules,and less volatile,less influence factors,choose regression analysis method to predict;And sintering,steelmaking and rolling process steam generation is influenced by many factors,the steam output has the characteristics of randomness and volatility,three kinds of process steam production with similar characteristics.Therefore,these three processes can use the same method to forecast,which namely using stepwise regression and wavelet neural network prediction method.Through examples of validation,the prediction method of steam in various processes yield prediction fitting degree is higher,which has a good practicability.(3)Base on the research of the steam consumption users,they have different features.According to their different consuming characters,the consuming users of steam are divided into three kinds.So the steam consumption users by linear regression analysis method,exponential smooth method and BP neural network to predict.All these prediction methods are selected pertinently,the higher prediction precision.(4)The steam system production cost are selected for objective function,it includes equipment maintenance,boiler fuel cost,cost of purchased electricity,outsourcing cost of water,outsourcing cost of steam,equipment start-stop cost and sell electricity cost;to the constraint conditions on equipment capacity,energy demand,energy conservation,etc;mixed integer linear programming model is adopted to optimize the steam system scheduling.Finally,the research on the content and the corresponding research achievements are summarized,and the content of the improvement in the study as well as the future direction of the field are prospected.