基于RBF网络的钢坯温度预报软测量模型研究
Study on Soft Sensor Prediction Model for Slab Temperature Based on RBF Neural Network
姜 磊1， 王德慧2， 朱里红
点击：3709次 下载：0次
作者单位：1. 浙江工业职业技术学院，浙江 绍兴 312000； 2. 长治钢铁(集团)有限公司，山西 长治 046031
中文关键字：RBF网络；钢坯温度；软测量
英文关键字：RBF neural network; temperature prediction model of slab; soft sensor
中文摘要：针对目前的测温技术难以用仪器直接测量出加热炉内被加热钢坯温度的问题，提出了通过RBF神经网络建立钢坯温度软测量预报模型，实现了钢坯温度准确及时预报，达到了减少燃料和钢坯表面氧化的目的。工业试验仿真研究表明，该钢坯温度预报模型精度高、自适应性好、鲁棒性强。
英文摘要： It's difficult to measure temperature of slab by some measuring instruments directly in the heating furnace now. A soft sensor prediction model for slab temperature was put forward based on RBF neural network. The targets which reduce materials and the probability of oxidation on the slab were attained by forecasting temperature accurately and timely. Industry experiments and simulations indicate that the temperature prediction model of slab is high precision, good adaptive ability and strong robustness.