基于神经网络和模拟退火算法的配煤智能优化方法

炼焦配煤过程是存在诸多不确定性的复杂工业过程,传统控制方法难以实施控制。在炼焦配煤中,配煤比是影响焦炭质量的主要因素。本文通过建立智能优化系统,提出一种确定配煤比的优化策略。首先根据焦化反应机理,分别建立配合煤和焦炭质量预测模型,然后通过配比智能优化模型,运用模拟退火算法对配煤比进行优化计算。实际效果证明了这种控制策略的有效性。     

Dissolution behaviour of a beryl ore for optimal industrial beryllium compound production

This study involves the leaching of the beryl ore with sulphuric acid (H₂SO₄) solution for predicting optimal beryllium extraction conditions with the aim of assessing the importance of leachant concentration, reaction temperature and particle size on the extent of dissolution. A kinetic model to represent the effects of these variables on the leaching rate was developed. It was observed that the dissolution of beryl ore increases with increasing H₂SO₄ concentration, temperature, decreasing particle size and solid to liquid ratio. At optimal leaching conditions, 89.3% of the ore was reacted by 1.25?mol/L at 75°C temperature and 120 minutes with moderate stirring, where 1612.0?mg/L Be²⁺, 786.7?mg/L Al³⁺, 98.1?mg/L Fe³⁺ and 63.4?mg/L Ag⁺ were found as major species in the leach liquor. The unleached products constituting about 10.7% were examined by X-ray diffraction (XRD) and found to contain primarily, siliceous compounds such as Xonotlite, Antigorite, Chrysolite and Kaolinite.     

Effect of manganese ore blends on performance of submerged arc furnace for ferromanganese production

Abstract

Manganese ore blends are used in ferromanganese production. The blend composition controls the operational performance of submerged arc furnace. A case study has been carried out at FAP, Joda, Tata Steel to better understand the process. The results of experiments revealed that the phase decomposition and decrepitation of Mn ore at low temperatures (500–900°C) in the upper part of the furnace increased the furnace top temperature and thereby promoted agglomeration of the charge, which caused the violent eruptions in the furnace. The root cause of the problems in reaction zone (bottom part) of the furnace was changes in the composition of slag, i.e. low silica and high alumina, which was also due to selection of Mn ores in the blends. Various options for silica addition were examined and compared. The pretreatment of the Mn ores and use of synthetic slag for silica adjustment options were identified to overcome the operating problems and to utilise the captive Mn ores.     

铜锍吹炼过程操作参数优化决策模型研究

在分析了铜转炉的生产工艺特点的基础上 ,提出了基于经验公式、热量和物料衡算以及计算智能的铜转炉优化操作决策模型。在该模型中 ,本文推导了最佳入炉铜锍量的公式 ;以铜转炉的热量衡算模型来确定冷料的加入制度 ;提出一个以神经网络和遗传算法为基础的参数寻优模型来确定熔剂加入制度和鼓风制度。最后 ,以该模型为基础 ,开发了一套铜转炉优化操作决策支持系统。工业现场运行结果表明 ,转炉产量提高了 6 % ,冷料处理量提高了 7.8%。     

热镀锌机组圆盘剪工艺参数优化技术的开发

热镀锌机组圆盘剪工作时,由于剪切工艺参数设定不当等原因造成圆盘剪设备的故障以及带钢边部出现有扣边、毛刺和切断比失衡等主要切边质量问题,以某钢厂热镀锌机组圆盘剪设备为研究对象,对其设备本身和工艺参数展开研究.通过对影响带钢边部质量缺陷因素的分析,以及BP神经网络输入变量相关性要求选择前张力、间隙量、重叠量、切边量4个因素...     

真空感应炉熔化期工艺优化控制

采用神经网络进行真空感应炉熔化期工艺过程的建模,利用实际工业生产数据进行仿真,并对某一钢种的熔炼工艺操作,采用遗传算法对熔化期电能的输入与合理分配进行了寻优,仿真结果表明该方法的有效性和优越性。     

遗传神经网络-X射线荧光光谱法测定铁矿石中铅砷

为测定铁矿石中微量的铅和砷,采用行星球磨机研磨铁矿石样品,经800℃预熔融后,于1 050℃熔融5min,冷却后再次熔融8min的方法制备了稳定的低稀释比熔片(样品与熔剂的质量比为1∶2)。采用无标样分析软件对合成样品熔片中铅和砷元素荧光强度进行扫描,不同扫描角度下的X射线荧光光谱强度数据作为神经网络的输入,铅、砷元素含量作为输出,用遗传算法对网络的权值和阈值进行优化,对测试集中样片的铅、砷谱线重叠进行校正,克服了低稀释比导致的背景强度高的缺点。对预测集中铅、砷元素含量模型预测的均方根误差(RMSEC)分别为0.39和0.42,相关系数均为0.98。可见实验方法与理论α系数回归方程法没有明显区别。     

高铁锰矿粉熔融还原生产富锰渣新技术

富锰渣是电炉冶炼硅锰合金的原料,而高炉法是冶炼富锰渣的主要方法,但该方法存在工艺流程长,焦炭消耗量大,铅、锌等易挥发金属影响冶炼等问题。提出了高铁锰矿熔融还原冶炼富锰渣新技术,该技术具有工艺流程短,不需要消耗焦炭,铅、锌等金属可以综合回收等优点。试验研究结果表明,高铁锰矿熔融还原生产富锰渣工艺是可行的,在还原温度为1 400 ℃、还原时间为20 min的条件下,渣铁可以完全分离,铁、锰回收率高。     

Benefits of oxygen enrichment in sponge iron rotary kiln emerging from data-driven modelling

Optimal oxygen enrichment conditions for sponge iron rotary kiln have been successfully explored on an industrial scale using a data-driven model. A multi-objective optimisation by genetic algorithm (MOGA) is employed to find the favourable conditions. The objective function for MOGA is derived from neural networks using pre-processed operational data. From industrial experimentations guided by the optimum conditions predicted by the present model, it emerged that when the coal fines injection is maintained at 1.75?tph and the oxygen enrichment is 8 Nm³?t^?1 of sponge iron, a reduction in the specific air requirement from 2609 to 2150?Nm³?t^?1 was obtained, while the end-zone bed temperature remained under control at 1132°C. These conditions resulted in a reduction of specific coal consumption by 6%, an enhancement in the sponge iron production by 6% and an increase in the rotary kiln campaign life from 50 to 100 days.     

Artificial neural network based modeling of liquid membranes for separation of dysprosium

In recent years,the liquid membrane process has been widely investigated to remove rare earth metals.However,transport modeling of this process requires the accurate values of several parameters,which are difficult to measure.Thus,the accurate simulation of this process is a challenging task.In this study,the artificial neural network(ANN) based approach is used to model the liquid membrane process for removing dysprosium.Experimental results from a previous study were used to train the ANN.Init...     

BP神经网络在连铸板坯质量在线诊断中的应用

为了更好地应用BP神经网络对连铸板坯质量进行在线诊断,基于连铸生产特点,利用采集的过程数据建立了符合生产实际的均一化函数.通过分析BP神经网络中各参数对网络性能及诊断准确率的影响,对BP神经网络的结构及学习算法进行修正,使该网络有选择和有区分地学习铸坯质量知识.结合某钢厂连铸现场数据,以黏结为例,建立了6种网络模型,对各模型算法进行了比较测试.结果表明:采用自定义函数均一化样本或采用提出的差异性算法训练神经网络,均可明显提高诊断准确率;采用选择性算法可确保诊断准确率不变的同时,提高学习速度;修正的算法更能很好地符合连铸生产实际.     

基于LM算法BP神经网络的高炉-转炉界面铁水温度预报模型

通过研究高炉-转炉界面铁水运输过程温度的主要影响因素,确定了影响高炉-转炉界面铁水运输过程温度的参数,建立了基于Levenberg-Marquardt (LM)算法BP神经网络的高炉-转炉界面铁水温度及铁水过程温降的预报模型.用沙钢100包铁水数据进行模型训练,50包铁水数据进行现场预报,结果表明:在高炉-转炉界面“一包到底”模式下,当绝对误差| X |≤20℃时,铁水温度命中率为94％,铁水温降命中率为78％;当绝对误差|X|≤40℃时,铁水温度命中率为100％,铁水温降命中率为92％,该预报模型能够满足现场实际生产需求,对炼钢生产有很好的指导意义.     

Application of Artificial Neural Networks for Evaluating Pressure Filtration of Coal Refuse Slurries

Bench-scale pressure filtration testing was performed to evaluate the dewatering characteristics of two coal refuse slurries, which were collected from thickener underflow streams of two coal preparation plants. Pressure filtration provides an opportunity to produce drier solids (filter cake) that can be stacked or mixed with the coarse refuse and improved water (filtrate) recovery that can be reused in the plant. This paper examines the effects of some of the major influencing variables such as pressure, slurry pH, feed solids concentration, fines fraction of solids in the slurry, filtration time, and temperature on dewatering thickener underflow slurry. Experimental results indicated that the overall filtrate flux increased with increase in pressure and temperature while it decreased with increase in fines fraction, pH, filtration time, and solids concentration. A total of 82 experimental results were used to develop a feed forward back-propagating artificial neural network (ANN) model. The model had R² values over 0.9 for both the training and the testing datasets, indicating the goodness of fit. Sensitivity analysis performed using the ANN model indicated that filtration time and pH were the most significant variables influencing filtrate flux.     

A New Correction Method of Electronic-weighing System for Blast Furnace

Ｉｎｔｈｅｐｕｆｉｎｇｓｙｓｔｅｍｏｆｂｌａｓｔｆｕｒｎａｃｅｐｌａｎｔ，ｔｈｅｐｒｅｓｓｕｒｅｈｅａｄｏｆｅｌｅｃｔｒｏｎｉｃ－ｗｅｉｇｈｉｎｇｓｙｓｔｅｍｉｓｉｎｓｔａｌｅｄｏｎｐｕｆｉｎｇｔａｎｋ，ｗｈｉｌｅｔｈｅｃｏｎｎｅｃｔｉｏｎｂｅ－ｔｗ...     

基于神经网络的真空感应炉终点钢水温度预报

针对真空感应炉生产过程中温度测量成本较高及精度较差等不足,建立了基于RBF神经网络的真空感应炉终点钢水温度预报模型。对输入参数作了详细的分析、筛选,并运用聚类算法对该模型进行了训练。结合现场数据进行了学习和预报,预报命中率较高,表明采用该方法可很好地预报钢水温度。     

热轧轧制力小波神经网络预报模型的研究

为了进一步提高热连轧精轧机组轧制力的设定精度,采用小波神经网络建立轧制力预报模型。并采用改进的快速BP算法来训练网络。仿真结果表明：建立的轧制力预报模型的预报值与实际值之间的相对误差在±6％以内,且学习算法收敛速度快。     

基于遗传神经网络的不锈钢带冷轧轧制力模型

 为了提高工厂从国外引进的以Bland Ford公式为基础的冷轧不锈钢带轧制力模型的计算精度,将基于遗传算法的BP神经网络与现有变形阻力和轧制压力解析数学模型相结合,建立了变形阻力和轧制压力修正模型。将在生产现场采集的部分过程记录数据,进行分类和预处理后作为训练样本用于训练遗传神经网络模型。将其他现场实测数据用于验证所建的轧制力模型,计算结果表明所建的轧制力模型具有较高的计算精度。     

炼钢连铸流程在线钢水温度控制

为了实现炼钢过程钢水温度的精确控制,在分析了实际炼钢厂钢水温度控制现状和钢水温度影响因素的基础上,建立了关键工序节点钢水温度的正向预测模型和逆向预定模型.同时,为了克服现有钢水温度预报方法的不足,提出一种基于钢包热状态和BP神经网络的混合模型方法.该方法以钢包热状态跟踪模型为基础,充分考虑了钢包热状态对钢水温度的影响,并与BP神经网络结合,可有效提高预测精度.     

智能控制综述及展望

本文综述和分析了主要智能控制方法和特点,指出了有待解决的问题;并针对两类非线性系统,对智能控制发展趋势进行了展望,提出了研究方向.     

Fuzzy Neural Model for Flatness Pattern Recognition

For the problems occurring in a least square method model, a fuzzy model, and a neural network model for flatness pattern recognition, a fuzzy neural network model for flatness pattern recognition with only three-input and three output signals was proposed with Legendre orthodoxy polynomial as basic pattern, based on fuzzy logic expert experiential knowledge and genetic-BP hybrid optimization algorithm. The model not only had definite physical meanings in its inner nodes, but also had strong self-adaptability, anti interference ability, high recognition precision, and high velocity, thereby meeting the demand of high-precision flatness control for cold strip mill and providing a convenient, practical, and novel method for flatness pattern recognition.