IWOA-KELM prediction model of silicon content in molten iron based on big data
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
万方数据
精准预测高炉铁水硅含量对炉温调控和节能减排具有重要价值.针对高炉炼铁过程的非线性、时变、高维等特点,提出一种基于大数据的IWOA-KELM铁水硅含量预测模型.采用Tent混沌映射种群初始化并设计捕猎速度控制因子迭代更新策略对鲸鱼算法进行改进,用改进鲸鱼算法(IWOA)对核极限学习机(KELM)中的核函数进行优化,使用斯皮尔曼秩相关系数对模型输入变量进行约简,利用高炉现场采集的大量数据进行学习和验证.结果表明,选取与铁水硅含量相关性最高的7个参数作为模型的输入变量,该预测模型均方根误差为0.036,允许误差在0.10内时,命中率为92.23％,预测时间4.23 s.
Accurately predicting the silicon content of molten iron in a blast furnace is crucial for temperature control,energy conservation and emission reduction.A big-data-based IWOA-KELM model for predicting silicon content in molten iron was proposed to address the nonlinear,time-varying,and high-dimensional characteristics of the blast furnace ironmaking process.The whale algorithm was improved by using Tent chaotic mapping population initialization and a novel iterative updating strategy for hunting speed control factors.The kernel function in the kernel extreme learning machine(KELM)was optimized via an improved whale algorithm(IWOA),while a Spearman rank correlation coefficient was employed to reduce the model's input variables.Robust learning and verification were carried out on large quantities of data collected from blast furnace sites.The results show that selecting the seven parameters with the highest correlation with silicon content in molten iron as the input variables of the model,the root mean square error of the prediction model is 0.036,and when the allowable error is within 0.10,the hit rate is 92.23％,and the prediction time is 4.23 seconds.
NETL
NSTL
万方数据
