Indexes prediction of blast furnace gas flow using the T-S fuzzy model of weighting rule firing level based on two kinds of cluster prototypes

The distribution state of the blast furnace (BF) gas flow is an important reference for the analysis of BF operation states. Bosh gas volume and wind pressure are two main indexes which reflect the distribution state of the BF gas flow. Therefor, predictions of these two indexes are important guidance for the field operation of BF. In recent years, the T-S fuzzy model plays an important role in nonlinear systems modeling and forecasting. As a method of information extraction, fuzzy clustering is crucial to the modeling quality of the T-S fuzzy model. However, the information extraction of datasets is incomplete based on one cluster prototype. Furthermore, hyper-shape of high-dimensional datasets is unknown. It is also unsuitable to cluster high-dimensional datasets based on one cluster prototype in previous researches. To improve the modeling accuracy, we consider the hyper-spherical-shaped cluster prototype and the hyper-plane-shaped cluster prototype in this paper. Then the T-S fuzzy model of weighting rule firing level is proposed to predict indexes of the BF gas flow. Simulation results verify that our model process superiority of accuracy, which makes the method serve better for practical production.