Development and application of intelligent coal blending system based on BP neural network and genetic algorithm
In view of the problems of many varieties of coking coal,high fluctuation of coal quality from the same mine,and phenomenon of serious coal mixing,Ningbo Iron and Steel Co.,Ltd.realized the whole-process monitoring of the change of coal coke data by building a coal coke database and developing an intelligent coal blending system.Combined with the historical production data,the intelligent coal blending system analyzed and extracted the key indicators affecting the quality of coke.And by adopting the modeling method of multiple linear regression and BP neural network,a prediction model of key indicators of coke quality was established.Meanwhile,this intelligent system combined with the coke quality prediction model,coal blending expert knowledge and single coking coal inventory information,and adopted optimized genetic algorithm to build the coal blending model,thus achieving the purpose of rapid real-time adjustment in coal proportions,rational use of coking coal resources,stable coke quality,and effective reduction of coking cost.The stable operation of this intelligent coal blending system enabled the rational and efficient utilization of coking coal resources and the purpose of reducing cost and increasing efficiency.
NETL
NSTL
万方数据
