Multivariable Dynamic Modeling for Molten Iron Quality Using Incremental Random Vector Functional-link Networks

Molten iron temperature as well as Si,P,and S contents is the most essential molten iron quality(MIQ)indices in the blast furnace(BF)ironmaking,which requires strict monitoring during the whole ironmaking production.However,these MIQ parameters are difficult to be directly measured online,and large-time delay exists in offline analysis through laboratory sampling.Focusing on the practical challenge,a data-driven modeling method was presented for the prediction of MIQ using the improved multivariable incremental random vector functional-link networks(M-I-RVFLNs).Compared with the conventional random vector functional-link networks(RVFLNs)and the online sequential RVFLNs,the M-I-RVFLNs have solved the problem of deciding the optimal number of hidden nodes and overcome the overfitting problems.Moreover,the proposed M-I-RVFLNs model has exhibited the potential for multivariable prediction of the MIQ and improved the terminal condition for the multiple-input multiple-output(MIMO)dynamic system,which is suitable for the BF ironmaking process in practice.Ultimately,industrial experiments and contrastive researches have been conducted on the BF No.2in Liuzhou Iron and Steel Group Co.Ltd.of China using the proposed method,and the results demonstrate that the established model produces better estimating accuracy than other MIQ modeling methods.     

主成分分析和最小二乘支持向量机模型在铁水硫和硅含量预测中的应用

良好的铁水质量是铸铁性能可靠性和稳定性的保证,而铁水中硫(S)含量和硅(Si)含量是衡量铁水质量的主要指标,因此在出铁前精准获取铁水S含量和Si含量具有非常重要的意义。实验提出一种结合主成分分析(PCA)和最小二乘支持向量机(LS-SVM)模型的铁水S含量和Si含量的预测方法。将某钢厂大型高炉的在线采集数据作为研究对象,首先对影响铁水中S含量和Si含量变化因素的数据做主成分分析,求取主成分作为模型的输入变量,其次建立最小二乘支持向量机预测模型对铁水S含量和Si含量进行预测。在S含量预测过程中,正则化参数gam和核函数参数sig分别取20、700时,预测误差最小,其均方根误差为0.0012,仿真时间为0.423105s;Si含量预测过程中正则化参数gam和核函数参数sig分别取40、500时预测误差最小,均方根误差为0.0238,仿真时间为0.079522s。最后将实验结果与传统最小二乘支持向量机(LS-SVM)和结合PCA的BP神经网络预测模型(PCA+BP神经网络)的结果对比,后两组对比实验关于S含量预测的均方根误差分别为0.0015和0.0014,仿真时间分别为1.320842s和2.245967s;后两种对比实验关于Si含量预测的均方根误差分别为0.0316和0.0325,仿真时间分别为0.459671s和2.061576s。实验结果表明,实验方法更加全面地考虑了所有因素对铁水中S含量和Si含量变化的影响,具有训练时间短、预测精度高等优点。     

基于分布式神经网络模型的高炉炉温预测建模

高炉炼铁通常采用铁水Si含量间接反映炉温的变化,模型预测精度低。以影响炉温的6个变量为输入变量,采用基于自组织的分布式RBF神经网络模型分别对铁水温度和铁水Si含量建立预测模型,先用自组织神经网络划分输入输出样本空间,然后对每个子空间建立RBF神经网络子网模型,再使用子网模型对测试样本集的同一个样本点进行预测,并以测试样本点对每一子空间的隶属度为权值,对子网预测值进行加权求和,得到最终预测值。对比使用同一输入变量数据的铁水温度和铁水Si含量的预测模型命中率,研究表明,高炉铁水温度的命中率更高,具有更好的炉温预测效果。     

石钢高炉铁水含硅量神经网络预报模型

按照现代控制理论,利用人工神经网络方法,把高炉视为多输入—单输出系统,结合高炉生产实际建立了石钢高炉铁水含硅量神经网络预报模型。通过引入动态步长和惯性项系数提高了网络收敛速度。采用不断更新学习样本集的方法提高了铁水含硅量预报的命中率。结果表明:在允许误差为0.1%时,命中率达到了86.67%,可以为高炉操作提供指导。     

包钢高炉炼铁优化配矿平台的开发与应用

为了降低高炉炼铁系统原料成本,实现自铁矿石采购到高炉炼铁全过程协同优化,开发了一个高炉炼铁全系统、全流程优化配矿平台.全流程是指从铁矿石采购到高炉产出铁水的整个工艺流程,全系统是指钢铁企业所有高炉炼铁整体系统.优化配矿平台包括数据库系统、单座高炉优化配矿平台、全系统高炉优化配矿平台、生产数据采集与分析平台4部分.平台以...     

Artificial neural nets for prediction of silicon content of blast furnace hot metal

Conventional models for prediction of silicon content of blast furnace hot metal are briefly reviewed. Four different artificial neural net (ANN) models, namely, back propagation algorithm (BPA), dynamic learning rate algorithm, functional link network (FLN) and fuzzy neural network (FNN), are trained and tested on operational data from blast furnace (BF1) at Visakhapatnam Steel Plant. FNN can predict silicon mass content of hot metal with a standard error (actual versus predicted) of 0.09% and correlation coefficient of 0.86; standard back propagation predicts with a standard error of 0.08 % and correlation coefficient of 0.79.     

基于改进人工神经网络的LF钢水终点温度预报

采用改进的人工神经网络算法,开发了40t钢包炉精炼时钢水终点温度预报模型。与传统BP网络算法相比较,改进算法可提高预测速度和精度。生产现场实验表明,传统BP神经网络算法,钢水温度预测误差±5℃的炉次仅为77%,用改进的BP神经网络算法,其误差±5℃的炉次为90%。     

Temperature measurement of molten iron in taphole of blast furnace combined temperature drop model with heat transfer model

The molten iron temperature in the taphole of a blast furnace (BF) is an important variable which reflects the internal heat of the BF. However, there is no method to detect the temperature directly. To solve this difficulty, an approach to measure the temperature of molten iron in taphole is proposed in this paper. First, during the stable tapping period, a heat transfer model is established according to the principle of equal heat flux of the trench walls, which is used to calculate the molten iron temperature at different positions in the main pipeline according to the temperature measured by thermocouples buried in trench walls. Next, a main pipeline temperature drop model is established based on the pipe temperature drop theory. Finally, the output of heat transfer model is used as the data source for parameter identification in the temperature drop model. The least-squares method is adopted to identify the molten iron temperature in the taphole. The proposed approach was validated in a practical industrial experiment at the #2 BF in an iron making plant. The results illustrate the effectiveness and rationality of the proposed approach and provides credible temperature data for the control of the BF.     

End Temperature Prediction of Molten Steel in LF Based on CBR

In order to improve the temperature control level of molten steel in ladle furnace (LF), a case‐based reasoning (CBR) method has been proposed for predicting end temperature of molten steel in LF. To predict the temperature accurately and efficiently, this paper develops two‐step retrieval approach and the correlation based feature weighting (CFW) method for CBR. And, the study evaluates the prediction effect of CBR method by the experiment of comparison with back propagation neural network (BPNN) model and CBR model. Experimental results show that CBR model achieves better accuracy than BPNN model and the CBR method is effective to predict end temperature of molten steel in LF.     

温度—静水压力下非线性位移时序模型及应用

岩土结构的位移一般是在多种内外因素的共同作用下产生的,而目前模型大多仅考虑时序对岩土结构位移的影响,忽略了外界因素的变化对岩土结构位移的影响。考虑温度、静水压力外界因素的变化对岩土结构位移的影响,提出一种复杂环境影响下的非线性位移时间序列建模方法。该方法用人工神经网络建模取代传统的分析方法,与遗传算法结合,自动确定输入时步长度和神经网络模型结构,建立温度、静水压力等外界因素影响下的非线性位移时间序列模型。例证表明,该模型具有较好的预测与外推预测功能。     

电弧炉炼钢全过程钢水碳含量动态预报模型

根据电弧炉物料平衡理论与利用BP神经网络的方法,建立了理论模型结合神经网络的电弧炉炼钢全程钢水碳质量分数实时预测模型。通过模型得出冶炼过程中碳质量分数变化曲线,实现对全程钢水碳质量分数的实时监控。在接近冶炼终点时,由于脱碳反应中碳氧积的存在,因此模型对影响终点碳质量分数的因素进行分析,采用BP神经网络方法进行预测,满足了对电弧炉冶炼终点碳质量分数预报准确度的要求。     

炼铁工序铁素流的系统动力学分析

运用系统动力学原理,采用因果关系和存量流量分析方法,在相对宏观的层面构建了钢铁生产流程炼铁工序的铁素流动态模型。仿真结果准确,验证了所建炼铁工序系统动力学模型的正确性并分析了不同返回情况下高炉铁素、铁水铁素流和损失铁素流的动态特性。计算结果表明:从物料投入高炉到铁水稳定输出时段,铁水铁素流随时间的增加而不断增大,然后逐渐趋于稳定,且铁水铁素流在炼铁初始时段增加速度最快;高炉铁素和损失铁素流均与本单元铁素流返回率呈正相关,且不随上游铁素流返回率的变化而改变;铁水铁素流与上游铁素流返回率、本单元铁素流返回率均呈负相关;上游铁素流返回率的增大会使铁水铁素流减小至新的稳定输出状态;本单元铁素流返回率的增大会使铁水铁素流先减小,然后逐渐增大至新的稳定输出状态。     

基于改进遗传算法优化Elman网络的板形识别方法

针对目前的板形缺陷识别方法精度不高、识别速度慢的问题,根据Elman神经网络模型可以反映系统动态特性,而且可以逼近任意非线性函数的特点,提出了一种利用改进的遗传算法优化Elman神经网络,使其泛化能力强、学习速度快、识别精度高,并建立板形缺陷模式识别模型的方法。为了验证该方法的识别能力,在隐层节点数与学习次数相同的条件下,分别与遗传算法优化的Elman网络和BP网络模型进行板形识别仿真对比分析。试验结果表明,改进遗传算法优化的Elman神经网络模型对板形缺陷识别精度高于BP网络等模型,并且具有收敛速度快的优点。     

基于BP神经网络的钢水温度预定模型

针对目前钢水温度预定方法存在不足,在分析钢水温度预定原理的基础上,在邯钢邯宝炼钢厂建立了基于BP神经网络的精炼终点目标温度和转炉终点目标温度的动态预定模型。利用邯宝炼钢厂的历史生产数据对模型进行了训练和测试,并进行了现场应用试验。结果表明,预定模型对转炉和精炼终点目标温度进行了优化,应用预定模型后,LF开始温度命中率提高到75%,中间包温度命中率提高到96.7%。     

Neural networks as models of psychopathology

Neural network modeling is situated between neurobiology, cognitive science, and neuropsychology. The structural and functional resemblance with biological computation has made artificial neural networks (ANN) useful for exploring the relationship between neurobiology and computational performance, i.e., cognition and behavior. This review provides an introduction to the theory of ANN and how they have linked theories from neurobiology and psychopathology in schizophrenia, affective disorders, and dementia.     

Prediction of Endpoint Phosphorus Content of Molten Steel in BOF Using Weighted K-Means and GMDH Neural Network

 The hybrid method composed of clustering and predicting stages is proposed to predict the endpoint phosphorus content of molten steel in BOF (Basic Oxygen Furnace). At the clustering stage, the weighted K-means is performed to generate some clusters with homogeneous data. The weights of factors influencing the target are calculated using EWM (Entropy Weight Method). At the predicting stage, one GMDH (Group Method of Data Handling) polynomial neural network is built for each cluster. And the predictive results from all the GMDH polynomial neural networks are integrated into a whole to be the result for the hybrid method. The hybrid method, GMDH polynomial neural network and BP neural network are employed for a comparison. The results show that the proposed hybrid method is effective in predicting the endpoint phosphorus content of molten steel in BOF. Furthermore, the hybrid method outperforms BP neural network and GMDH polynomial neural network.     

小波模糊神经网络在高炉炉温预测中的应用

在运用模糊神经网络进行预测的基础上,建立了一种应用小波理论对时间信号进行去噪。根据去噪处理对模糊神经网络作相应处理的预测模型,并将所建模型应用于高炉炉温预测。仿真结果证明小波模糊神经网络比模糊神经网络更具优越性,预测准确率明显提高。     

加热炉炉温神经网络预测模型的开发

针对加热炉系统非线性、大滞后、大惯性,炉温难以有效预测的问题,以山东钢铁莱芜分公司宽厚板加热炉为研究对象,通过神经网络训练获得充分逼近仿真对象的系统参数,最后使用该方法对莱钢宽厚板加热炉炉温进行预测,结果说明该方法预测准确,具有较强的实践意义,为炉温控制提供了可靠依据,提高了生产效率,降低了能耗。     

一种基于模糊神经网络FNN在加热炉温度控制中的应用

从实际出发,以昆明钢铁集团公司中板厂加热炉为研究对象,对具有时变性、非线性、模糊性的随机过程进行了研究。着重研究了神经网络与模糊系统融合的可行性及融合方式,采用了一种新型的智能控制方案——模糊神经网络控制。对提出的模糊神经网络控制算法进行了仿真试验,仿真结果表明,对比PID控制和自整定PID控制,采用本文所提出的模糊神经网络控制算法对加热炉进行控制,具有推理速度快,跟踪性能好,抗干扰能力强的优点,它完全能够满足工业生产需要,具有较强的可行性和实用性。