基于M-SVR与RVFLNs的高炉十字测温中心温度估计

发布时间：2018-03-16 21:25

  本文选题：高炉炼铁　切入点：十字测温　出处：《东北大学学报(自然科学版)》2017年05期 　论文类型：期刊论文

【摘要】：由于高炉中心温度较高,十字测温中心位置传感器极易损坏,并且更换周期长,因而导致无法及时判断炉顶煤气流分布.采用多输出支持向量回归(M-SVR)和随机权神经网络(RVFLNs)两种数据驱动智能建模方法建立高炉十字测温中心带温度估计模型,并基于实际工业数据对建立的模型进行验证和比较分析.结果表明,在样本数量较小时,M-SVR模型和RVFLNs模型都具有较好的温度估计效果,但当样本数量充足时,M-SVR模型的泛化性能和估计精度更优于RVFLNs模型.
[Abstract]:Because of the high temperature in the center of blast furnace, the cross temperature measuring center position sensor is easily damaged, and the replacement period is long. As a result, it is impossible to judge the gas flow distribution on the top of the furnace in time. Two data driven intelligent modeling methods, multi-output support vector regression (M-SVR) and random weight neural network (RVFLNs), are used to establish the temperature estimation model of the cross temperature measuring center belt of blast furnace. The results show that both M-SVR model and RVFLNs model have good temperature estimation effect when the number of samples is small. However, when the number of samples is sufficient, the generalization performance and estimation accuracy of M-SVR model are better than that of RVFLNs model.
【作者单位】： 东北大学流程工业综合自动化国家重点实验室;
【基金】：国家自然科学基金资助项目(61473064,61290323,61333007,61290321,61621004) 中央高校基本科研业务费专项资金资助项目(N160805001,N160801001) 辽宁省教育厅科技项目(L20150186)
【分类号】：TF321;TP18
，

本文编号：1621708