点击次数：

论文类型：会议论文

发表时间：2018-01-01

收录刊物：CPCI-S

卷号：51

期号：21

页面范围：7-12

关键字：Coke oven gas system; Causal model; Scheduling

摘要：Reasonable scheduling of coke oven gas (COG) in steel industry will save fossil energy resources and increase gas consumption efficiency. In this study, a causal model-based scheduling approach is proposed to provide guidance for the coke oven gas scheduling. The causal relationship of the variables related to the gas tank level is discovered and the causal diagram is established, according to which the training sample is constructed and the predicting model of gas tank level is trained. Then, an objective function that considers the scheduling solution and its result is designed. To calculate the most reasonable solution, a modified particle swarm optimization algorithm is employed. The validation experiments are carried out by using practical data coming from a steel plant in China, where the gas tank level approaching the higher and the lower level zones are both considered. The human experience -based method and a data based scheduling one are conducted as comparative studies. The results indicate that the proposed method is efficient for COG scheduling. (C) 2018, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.