A Modeling Method for Thermal State of Blast Furnace
The paper proposes a kernel based modeling method for thermal state of blast furnace, and an improved Support Vector Regression(SVR) to enhance the accuracy of modeling. Firstly, we extract features from many factors influencing thermal state of blast furnace via Kernel Principal Component Analysis (KPCA) in order to remove correlations to each other and avoid their informations overlapping or counteracting. Then we set up model of thermal state of blast furnace adopting SVR, and propose an improved SVR against the problem of lower-accuracy at peak points. The system realization shows that the proposed modeling method, which combines KPCA with the improved SVR, has higher accuracy compared to normal SVR method, explaining that the method is effective to thermal state modeling of blast furnace.
Shukuan Lin Jianzhong Qiao
College of Information Science and Engineering North-eastern University Shenyang 110004, China
国际会议
青岛
英文
2006-07-21（万方平台首次上网日期，不代表论文的发表时间）