A joint scheduling method for multiple byproduct gases in steel industry

Reasonable scheduling of the byproduct gases produced during steel making procedure is of significance to steel industry because it is helpful to save energy resources, raise economic profit, alleviate the environment pollution and ensure the safety of the production process. In this study, a causal-interval-reasoning (CIR)-based joint scheduling method is proposed for scheduling the multiple categories of byproduct gases. A causal reasoning model is constructed here to predict the gas tank level, in which not only a number of influence factors (i.e., the gas users) on the gas tanks but also the mutual influence of the online tanks are considered. The upper and lower limits of the prediction intervals (PIs) of each gas tank level are constructed based on the granularity partition of the samples and a particle swarm optimization is designed to determine the parameters of granularity. In order for the balance of the whole byproduct gas system, a four-layer causal network is established, in which the operational statuses of boilers, the heat quantity, the steam demand and the gas mixture proportion are all well involved. To further optimize the scheduling solutions, an evaluation index is accordingly proposed. The practical data coming from a steel plant are employed for the validation data experiments, and the human experience based approaches considering only one single category of gas are also conducted as comparable studies so as to indicate the superior performance of the proposed method. For the practical application, a scheduling software system is consequently developed and implemented based on the proposed method, which has been applied in this steel plant for more than 1 year.