变结构神经网络在板形信号模式识别方面的应用

提出了基于模糊距离的变结构神经优化算法,并将其用于板形信号的模式识别过程,有效地解决了板宽变化时神经网络拓扑结构不变的问题,提高了识别速度和精度,从而成为一种新的智能板形信号识别方法。     

板形模式识别的GA-BP模型和改进的最小二乘法

针对板宽变化时需要不同拓扑结构的神经网络才能完成板形模式识别任务，网络学习工作量大，网络存在收敛速度慢，易陷入局部极小等结构性能不佳的问题，首次建立了以勒让德正交多项式为基模式的只用6个输入信号、3个输出信号的板形模式识别GA-BP网络模型。该模型不仅结构简单，而且物理意义明确，识别精度较高，解决了板宽变化时神经网络结构形式不变的问题，从而实现了板形模式识别的智能化。又提出了基于勒让德正交多项式的板形模式识别最小二乘法，该方法简单、实用，识别精度较高，克服了传统的最小二乘模型板形模式识别的缺点和不足。为板形模式识别提供了两种简便实用的新方法，发展了板形模式识别理论和方法。     

基于自适应神经模糊推理系统的板形模式识别研究

 针对传统的最小二乘板形模式识别方法的抗干扰能力差、精度低和神经网络板形模式识别方法存在的网络学习时间长、易陷入局部最小值、应用效果不佳等问题,提出了一种基于ANFIS的板形模式识别方法。该方法融合了模糊理论和神经网络的优点,弥补了彼此的不足,有效的解决了上述问题。板形模式识别是一个多输出系统,而MATLAB中ANFIS指令仅有一个输出,针对这个问题,本文提出利用系统拟合的方法,有效的克服了这个缺陷。研究结果表明,该方法能有效识别出常见的板形缺陷,识别速度和精度有所提高,识别结果跟板形仪的实测板形非常接近。     

含有三次板形的新型板形模式识别方法

为了提高板形模式识别的精度,在传统一次、二次和四次板形基模式的基础上增加了三次板形基模式,提出了以一次、二次、三次和四次勒让德正交多项式为基模式的板形模式识别方法。通过该方法对机理模型计算的板形数据和实测的板形数据进行了分析,表明该方法提高了板形模式识别的精度,使识别结果更具有完备性,细化了板形控制的目标,有利于提高板形控制的精度。     

基于循环神经网络的板形模式识别模型

板形模式识别是板形控制关键一环。传统的板形模式识别模型存在识别精度差、抗干扰能力差等缺点。随着数据回归任务复杂程度的增加,以深度学习为基础的深度神经网络已经大量用于数据分类、图像处理、模式识别、特征提取等任务。深度学习可通过学习一种深层非线性网络结构,实现复杂函数逼近。基于这一背景提出了基于循环神经网络RNN的板形模式识别模型。结果表明,基于RNN的板形模式识别模型可以完成板形的大数据训练,模型的识别精度和泛化能力都很高,为进一步提高板形控制精度提供了新方法。     

基于自适应神经模糊推理系统的板形模式识别

针对传统的最小二乘板形模式识别方法的抗干扰能力差、精度低和神经网络方法存在网络学习时间长、易陷入局部最小值等问题,把模糊理论和神经网络的优点融合在一起,通过三个自适应神经模糊推理系统的有效拟合,提出了一种基于自适应神经模糊推理系统的板形模式识别方法.研究结果表明,该方法能够很好地克服以上缺点,而且能够有效识别出常见的板形缺陷,识别速度和精度有所提高,识别结果跟板形仪的实测板形也非常接近.     

智能识别方法在板形识别中的应用及发展趋势

对国内外关于板形模式识别技术的研究现状和发展趋势进行了综述。通过比较,提出了传统的基于最小二乘法的多项式拟合法存在的不足,并对模糊分类、神经网络、遗传算法、混沌优化等智能识别方法在板形模式识别中所具有的优势进行了归纳和总结。最后,对智能方法在板形识别问题中的应用以及板形识别技术的发展趋势进行了展望,为板形检测环节得到理想板形信号提供理论研究方法,并将逐步应用于钢铁板形控制的工业过程中。     

基于Legendre多项式的板形模式识别优化仿真及应用

板形缺陷模式识别是铜带轧机板形控制中最重要的部分.本文在对多种识别方法对比分析的基础上,提出了一种基于Legendre多项式的板形缺陷识别优化算法.利用其正交特性,结合轧制工艺要求,采用Matlab仿真得到板形基函数多项式的参数设定,代替原Legendre多项式的固定参数,同时根据来料板形特征设置迭代方程.结果表明:该优化算法能够有效地得到板形特征值,既弥补了最小二乘法板形识别的缺陷,又避免了Legendre正交多项式对调节机构的特殊要求;经现场验证,板形平均误差从±8 I减小到±6 I,精度提高了10％左右.     

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。     

整辊镶块智能型冷轧带钢板形仪的研制与工业应用

 为提高冷轧带钢板形检测精度,研制出一种新型的冷轧带钢板形仪——整辊镶块智能型板形仪。该板形仪提出与国际上流行的分段式检测辊不同的结构形式,研制了整辊镶块式检测辊,以避免传统检测辊辊片之间由于热膨胀不同对带钢表面造成的划伤;提出与国际上流行的干式集流环不同的信号集流传输方式,研制了喷淋湿式集流环,以提高检测精度和使用寿命;提出与国际上流行的采集卡形式不同的信号处理方式,研制了嵌入式DSP板形信号处理硬件系统,以提高检测精度和抗干扰能力;研发了具有误差补偿、模式识别、目标板形制定和闭环控制计算功能的板形信号处理软件系统,以实现板形仪的智能化。该板形仪经在鞍钢1250mm冷轧机上实际工业考验证明,测量信号稳定可靠,而且实现了板形闭环控制。     

Quantitative Evaluation Method of Shape Indicator for Cold-rolled Strip

Quantitative level evaluation of cold-rolled strip was significant for the shape control and the product structure optimization. Based on the probability statistics method and the shape pattern recognition method, the shape evaluation model of cold-rolled strip was established to evaluate real-time shape of the online strip. The model was embedded in the shape control system, and the online strip shape state was real-time observed whether to meet the technical indicators, so the shape level could be identified and evaluated. Based on the shape evaluation indicators, the shape component could be well controlled to improve the shape hit rate and the control stability, so that the rolled strip shape could meet the technical requirements. At a 1050 cold strip mill, the shape data of the scene typical specifications strip were online collected and real-time evaluated, and the hit of macro shape in less than 8 I was 68.44%, indicating that the mill shape control had considerable room for improvement, as well as the evaluation method was feasible and practical for quantitative evaluation of the strip shape. So the new evaluation model has important guiding significance to increase value-added and reduce product quality objection.     

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.     

Detection and evaluation mechanism of cold rolling strip edge shape

The edge shape problem of strip has great influence on the shape control precision, so it is extremely necessary to establish the high-precision strip edge shape model. Based on the theoretical analysis and the practical engineering, a new high-precision shape detection model, which consists of wide channels in the middle of detection roll and narrow channels on both sides, is established to analyse the problem of strip edge shape. Instance data showed that the strip lateral thickness difference was directly related with the shape state of strip, and the strip edge coverage rate and the strip deviation amount had important influences on the shape detection precision, the pattern recognition accuracy and the shape control effect, so the strip edge shape model has become one of the key factors to improve industrial application effect of the shape closed-loop control system in the present situation.     

Improvement of Shape Recognition Performance of Sendzimir Mill Control Systems Using Echo State Neural Networks

High rigidity twenty-high Sendzimir mills （ZRMs） are widely used for rolling stainless steels, silicon sheets, etc. A ZRM uses a small diameter work roll to produce massive rolling forces. Since a work roll with a small diameter can be bent easily, strips often have complex shapes with mixed quarter and deep edge waves in the shape of plates. In order to solve this problem, fuzzy neural network controls are generally used for shape： recognition in ZRM control systems. Among various neural network types, the multi-layer perceptron （MLP） is typically used in current ZRMs. However, an MLP causes the loss of a large amount of shape recognition data. To improve the shape recognition per- formance of ZRM control systems, echo state networks （ESNs） are proposed to be used. Through simulation re- sults, it is found that shape recognition performance could be improved using the proposed ESN method.     

2800四辊轧机板形控制功能完善及雪橇板控制参数优化

针对武钢中板厂产品的板形问题,立足轧机原有的板形控制装备和技术,研究以先进的理论和算法制定先进的控制策略、挖掘原有手段的板形控制潜力,开发出具有特色的系列板形控制技术--支持辊初始辊形、工作辊初始辊形、轧制序列编组、多道次压下负荷最优分配动态调整板形,形成完整的板形控制系统,使轧机的板形控制性能得到优化和显著提高,创造了较大的经济效益。同时,对轧机原有引进的"雪橇板"控制功能系统和软件进行了消化和完善,通过优化控制模型和参数,提高了控制精度,消除了钢板头尾过度弯曲。     

A Neural Network Recognition Method of Shape Pattern

The final objective of shape control is to makethe output shape data achieve the objective shapedata,then the adjusting magnitude of shape controlactuating mechanism can be defined in accordancewith the deflection value of the output shape dataand the objective shape data.In shape controlsystem,the first thing is to make shape faultpattern recognition based on the speciality andrequirement of shape control actuating mechanism ofrolling mill.The main assignment of shape patternrecognition is to…     

蒙古铁矿块矿直接入炉基础冶金性能

摘要：高炉装料制度是复杂高炉炼铁中调节炉况运行状态的重要上部调剂手段,炉料在布料矩阵操作参数下在炉喉处所形成的空间分布是影响炉内煤气流分布和高炉炉况波动的重要因素之一。合理的调控与优化高炉装料所产生的料面形状,给出布料矩阵优化计算的理论依据,是保证高炉稳定顺行,提高资源利用率和减少污染物排放的有效途径。结合无钟炉顶的设备结构与布料工艺,针对期望料层厚度分布研究布料矩阵的优化计算方法,进行完善与改进;同时以期望料面输出形状为设定目标,建立了期望料面输出形状优化模型并通过遗传算法实现对布料矩阵的优化计算。最后,通过工业过程的实测数据对PSO优化方法和遗传算法优化方法进行对比验证,结果表明,使用基于遗传算法的优化模型能够有效地制定布料矩阵,符合期望目标。