Prediction of deformation resistance during hot rolling process based on generalized additive model

A deformation resistance model with generalized additive form for hot rolled strip was established based on the measured data of hot strip rolling process. Firstly, a deformation resistance modeling method for hot rolled strip was proposed under the framework of generalized additive models, including pre- analysis of variables, model setting, model estimation and result analysis; the back- fitting algorithm for the estimation of the univariate function of each independent variable was given. Then, the modeling experiment was carried out for Baosteel 1880 finishing mill, the generalized additive model for deformation resistance was established based on large number of strip sample data which covering many kinds of steel grades, the influence of strain temperature, strain and strain rate on deformation resistance was obtained. Practical calculation results show that the accuracy of new model is higher than that of the Baosteel online model. The new model has the advantages of high calculation precision and applicable to wide range of steel grades, and can be used in online process control of hot continuous rolling production.     

轧制变形抗力数学模型的发展与研究动态

建立变形抗力数学模型对变形抗力预测和电子计算机在线控制生产的普及具有重要意义。从传统解析法和人工神经网络法两个角度,系统介绍了国内外建立变形抗力模型的发展应用与研究动态,结合建模中存在的问题,针对性地提出了一些建议。     

基于RF-LSTM的热连轧板坯粗轧出口温度预报

热连轧板坯温度的连续、实时预报是提高带钢产品性能的关键,针对热连轧过程数据维度过高不利于快速、准确预测板坯温度的问题,提出一种基于随机森林-长短期记忆神经网络(RF-LSTM)的板坯粗轧出口温度预测模型.首先,采用改进随机森林算法对特征变量进行选择,通过分析板坯粗轧出口温度的预测结果变化衡量特征的贡献度,进而构造反映过...     

热连轧变形抗力模型的优化改进

改进后的志田茂变形抗力模型是比较经典的变形抗力模型,但该模型没有考虑残余应变对于变形抗力的影响,不能完全适应现场生产的需要。作者在分析研究济南钢铁集团有限公司热连轧厂大量实际生产数据的基础上,在原变形抗力模型中考虑了碳质量分数的影响,增加了对应的影响系数,用于修正残余应变的影响,从而使该模型更贴近于现场实际。模型实际应用结果表明,改进后的模型预报精度明显提高,效果良好。     

基于改进遗传算法的热轧批量计划优化模型

针对钢铁企业的热轧批量计划编制问题,考虑了热轧生产工艺规程、同宽板坯连轧长度限制等约束,建立基于不确定旅行商数的多旅行商问题(MTSP)的热轧批量计划数学模型,把轧制单元计划数最小化和紧邻轧制板坯间轧制宽度、厚度、硬度差异而造成的总惩罚值最小化作为优化目标。结合模型的实际情况,采用基于两交换启发式的改进遗传算法对问题进行求解。最后,利用某钢厂的实际生产数据进行计算试验,同时对结合传统遗传算法和改进遗传算法的优化情况进行对比和结果分析,模型的正确性和算法的有效性得到了验证。     

Thermal camber model for hot and cold rolling

The MDS profile and flatness control system is based on many physical part models. The thermal camber model is the only one of these part models which, apart from the spatial dimensions, inherently depends on the time dimension as opposed to part models like roll bending or even roll wear. This means that work roll temperature and shape keep changing even if boundary conditions like water distribution or strip contact stay constant. The exact mathematical solution of the resulting dynamic equations Is a very time consuming task usually not suitable for on-line purposes. Here a new mathematical approach is presented, which solves the 2-dimensional Fourier heat conduction equation with 3-dimensional boundary conditions at a speed suitable for on-line applications. Furthermore an approximate solution of the Hooke stress strain relations is derived, which translates the temperature distribution of the roll into an expansion distribution. This thermal camber model has been implemented in several hot and cold rolling mills. Data collected there show good agreement with the model.     

热轧带钢组织性能预报模型及应用

 基于物理冶金理论,研究了热轧带钢过程中的奥氏体晶粒长大模型、奥氏体再结晶模型、奥氏体相变模型以及力学性能模型。奥氏体再结晶模型中,通过研究位错密度的变化来描述由于再结晶不完全造成的变形抗力的变化。奥氏体相变模型中,通过碳扩散理论描述了奥氏体-铁素体相界面随冷却过程的变化规律。基于热轧带钢过程中的冶金物理模型,开发热轧带钢组织性能预报系统。系统包括4个模块,分别用于计算板坯在加热炉、粗轧精轧、层流冷却和卷取完成各阶段的组织和力学性能参数,生产工艺是该系统的重要输入参数。利用该系统对某钢厂实际生产过程的组织性能进行预报,预报的力学性能和现场实测值有较好的一致性。     

基于增量分析和PSO-LSSVM的热轧辊缝调平预测模型

辊缝调平在保证热轧带钢板形质量和轧制稳定性中起着关键作用,目前以操作人员目测后经验调整为主,无法满足未来少人化、智能化轧制技术需求.为此,基于增量分析方法实现工艺参数增量因子提取,有效解决传统离散数据预测中部分信息丢失问题;同时以粒子群算法(PSO)优化最小二乘支持向量机(LSSVM)模型参数,使得参数选取更加科学.采...     

基于行处理算法的带钢热连轧负荷分配

负荷分配是带钢热连轧精轧模型设定的基本问题,传统 Newton-Raphson算法存在初始值要求严格、易发散等计算性能较差的缺点。针对其在求解非线性方程组方面的不足,基于行处理法提出了一套适用于热连轧机组快速厚度分配的计算模型。该方法以中间机架出口厚度为变量,在采用高精度、非简化轧制模型的基础上,计算速度快,弥补了传统方法的不足。同时为了加快收敛速度,根据经验统计提出了一种新的初始值的确定方法。并利用C＋＋根据得出的算法编制了计算程序进行分析计算,在采用实测数据基础上的验证结果表明该算法具有运算速度快、收敛性好等优点。     

一种新型实用热轧工艺参数模型的开发

介绍了一套自主开发的热轧工艺参数模型．该模型内耦合了不同钢种的变形抗力曲线,这些变形抗力方程中耦合了钢的化学成分、温度、应变、应变速率及奥氏体晶粒尺寸等因素．根据输入的工艺参数用西姆斯方程计算每道次的应变速率及应变量,并得到相应道次的变形抗力、热轧轧制力、力矩及功率等参数．模型可根据实测的结果自学习,并修正相应的结果．与攀钢热轧厂的实测结果相比,模型的输出结果吻合较好,预测误差在10%以内．     

Three-dimensional hot rolling model for prediction of camber generation

Abstract

A three-dimensional explicit finite element model of the hot rolling process has been developed and used to investigate the generation of camber during the roughing stages of this process. A number of important contributing factors have been studied in detail in an effort to predict their combined effects on camber generation. Predictive equations have been derived which can be used to estimate the level of camber, given suitable values of the dominant process variables. A parallel experimental study has also generated similar empirical relationships. There is reasonable comparison between the predictive and empirical equations, and causes of minor discrepancies are discussed.     

大型筒节轧制力预报模型

 根据筒节和轧辊的几何关系,得到了筒节上下表面接触弧长的几何方程;根据现场数据和有限元方法,得到了筒节上下表面接触弧长的变形方程;结合几何方程和变形方程,并基于赫希柯克公式计算了考虑弹性压扁的筒节接触弧长。由于筒节外端对轧制力的影响远远大于接触摩擦的影响,结合接触弧长模型和材料变形抗力模型,基于现场数据和优化算法,优化得到了外端应力状态影响系数,从而建立了大型筒节轧制力预报模型。结果表明：上下辊的接触弧长不等,上辊接触弧长稍大于下辊接触弧长,上辊压下量大于下辊压下量,上辊和下辊的接触弧长之比约为1.3左右;将模型应用到筒节轧制中,计算轧制力与实测轧制力平均误差为9.2%,模型计算精度较高,能够满足工业应用要求。     

基于热轧全流程氧化铁皮控制的耐蚀性工艺

针对热轧生产流程实际工况,系统研究了热轧、卷取阶段的三次氧化铁皮演变规律,旨在不增加生产成本的前提下,通过调整热轧生产工艺控制氧化铁皮结构,利用热轧后生成的氧化铁皮作为防护屏障,提高钢材耐蚀性能.结果表明,在不同轧制温度下,三次氧化铁皮结构从外到里分别为Fe2O3、Fe3O4和FeO,由于FeO中的阳离子空位密度大,导...     

Finite element based model of structure development in the hot rolling process

An existing thermomechanical finite-element model of the hot rolling process is used to simulate the structure evolution in the strip. The equations which describe the grain-size variations during the hot deformation are implemented in the model. The local values of temperatures, strain rates and strains are used in calculations, allowing the determination of the grain-size distribution across the strip. The results of calculations carried out for the seven-stand hot-strip mill are presented in this paper.     

热轧HVC工作辊热变形

以HVC工作辊为研究对象,采用二维交替方向差分法建立了工作辊温度场模型,可对任意时刻工作辊温度场分布和热辊形进行仿真,计算时间能够满足在线应用且具有绝对稳定性.通过现场实测数据与模型计算值比较,吻合较好,验证了模型的准确性.     

热轧形变量对共析钢珠光体组织球化的影响

利用Gleeble 1500热模拟机进行热压缩试验,研究了不同热轧形变量下奥氏体区形变对共析钢后续珠光体相变组织球化的影响。结果表明,高温奥氏体区形变增加了奥氏体的形变储存能,导致C曲线左移,并且形变可以明显减小珠光体团直径,随应变量增大,珠光体片层间距减小,渗碳体厚度变薄,片层取向多样化。最终的等温球化试验表明,增大奥氏体形变量有利于珠光体的球化效果。     

热轧前滑模型的研究

前滑模型对热轧生产参数的设定有重要意义。基于Orowan微分方程和Sims假设,建立了热轧前滑理论模型,并把该理论模型中相关的3个参数(轧辊半径、入口厚度、出口厚度)优化为2个无单位的参数,采用统计分析法得到简易的经验模型,并分析了各因素的影响。现场实测数据表明:前滑模型计算的最大绝对误差是0.001,最大相对误差是6.93%,能满足在线生产要求。     

热连轧精轧机组温度控制数学模型研究

从现场实际出发研究了精轧各个环节的温度变化过程,分析了带钢热连轧终轧温度与各影响因素之间的关系.同时将温度模型程序化,采用计算机系统进行分析和计算,提出了行之有效的温度模型方案.该温度模型对其它热轧生产线同样适用.     

基于数字信号处理器DSP的热连轧实时仿真

在热连轧精轧机组的改造或新建过程中 ,如何降低生产风险 ,减少调试时间和降低成本是当前的研究热点。目前主要采用常规的仿真方法来研究 ,但热连轧系统的高度复杂性使得仿真的实时性和准确性受到制约 ,仿真结果对实际的指导作用不大。本文提出的用DSP处理器、工控机、内存映像网以及实物控制器组成的实时仿真系统能对复杂的热连轧过程实现满足精度要求的实时仿真 ,为继续深入这方面研究与应用提供了良好的基础     

Microstructural model for hot strip rolling of high-strength low-alloy steels

The microstructural evolution during hot-strip rolling has been investigated in four commercial high-strength low-alloy (HSLA) steels and compared to that of a plain, low-carbon steel. The recrystallization rates decrease as the Nb microalloying content increases, leading to an increased potential to accumulate retained strain during the final rolling passes. The final microstructure and properties of the hot band primarily depend on the austenite decomposition and precipitation during run-out table cooling and coiling. A combined transformation-ferrite-grain-size model, which was developed for plain, low-carbon steels, can be applied to HSLA steels with some minor modifications. The effect of rolling under no-recrystallization conditions (controlled rolling) on the transformation kinetics and ferrite grain refinement has been evaluated for the Nb-containing steels. Precipitation of carbides, nitrides, and/or carbonitrides takes place primarily during coiling, and particle coarsening controls the associated strengthening effect. The microstructural model has been verified by comparison to structures produced in industrial coil samples.