The Formant Criteria Of Degree Paper At USTL

According to the energy management regulations from Energy Center for iron and steel enterprise, this paper presents some research on gas production and consumption forecast and optimization distribution for iron and steel enterprise. The main original research achievements are as follows:(1) Gas production and consumption forecast methods in the process of iron and steel manufacturing are proposed. The Box-Jenkins time series forecasting method is used to predict the amount of gas produced. Based on the actual production data of Bayuquan branch’s 1# blast furnace, the corresponding forecast model is built. Aiming at many factors affecting gas consumption and gas forecasting problem with complex relationship, Levenberg-Marquart algorithms is used to adjust BP neural network’s weights and thresholds in order to improve the efficiency of BP training this algorithm; Bayesian regularization is adopted to improve the generalization ability of BP network. Based on actual production data from Bayuquan branch’s hot-rolling heating furnace, main factors affecting gas consumption are determined by grey correlation degree analysis and then the BP network forecasting model for gas consumption is built.(2) On the basis of predicting gas production and consumption, gas pipe network model and gas unit classification model are established. Balancing gas production-and-consumption and minimizing release amount in gas pipeline network are the objective functions. Technological conditions for gas pipelines and gas production and consumption unit during the process of gas production-and-consumption are constraints. A distribution model of dynamic gas optimization is built.