【作者】 史志远；

【导师】 梅瑞斌；

【作者基本信息】 东北大学 ， 材料工程（专业学位）， 2020， 硕士

【摘要】 热轧过程中温度对轧制过程力能参数及成品组织性能具有重要影响,构建快速准确的温度在线预测模型对实现轧制过程智能制造具有重要意义。本文以热传导有限元理论为基础,在保证计算精度的条件下,尽可能缩短温度预测的响应时间,进而构建在线预测模型。主要研究工作如下:(1)基于传热学原理和实际轧制流程,研究了轧制生产线各工艺参数间的数学关系及换热系数与内热源的计算公式,开发了棒材、板材、H型钢等钢材空冷过程及水冷过程温度场有限元求解程序。以棒材为研究对象,其空冷温度求解时间仅为15.625 ms,程序计算最大绝对温差为38.3℃,最大相对误差为5.5%;水冷程序计算值在高于300℃时其最大相对误差为3.75%,二者都满足在线生产对温度场计算精度及速度要求。(2)以棒材为研究对象研究了网格划分方式对于瞬态温度振荡、温度收敛值和计算速度的影响规律。结果表明:空冷过程网格数量为200时计算速度快,计算精度较高,为最优网格数量;基于单元边长和时间步长提出了消除温度振荡的条件,相比于ANSYS软件中条件参数β缩小了 10倍,即本程序计算稳定性更好;逐层细分网格可以有效抑制温度振荡并提高程序运算速度与计算精度(等差划分方式使计算精度提高了 0.21%);分析了等差、等比划分方式消除温度振荡具体条件,其中等比方式边缘单元径向长度drn为等差方式的8.6倍;缩小单元内相邻节点编号差距可以减少系数矩阵带宽,计算效率提高了约22%。(3)为实现连续轧制过程温度场求解和动态网格划分,在研究分析了反比插值法、多项式插值算法(Lagrange插值、Newton插值)和分段插值法等插值方法后,提出了 Newton二次优化插值法,避免了高次插值多项式的病态性质并且可以在保证精度的情况下节约计算时间,完成插值程序设计及编写。以棒材轧制变形过程为研究对象,综合考虑了塑性变形做功及摩擦引起的热效应,利用插值程序开发了棒材热轧变形过程温度场模拟程序,对棒材热轧过程温度变化情况进行了分析并对反红现象做出了解释。(4)使用Python语言开发了人机交互界面,成功实现了不同断面钢材热轧过程温度场的预测与优化,使本研究的自编有限元程序转变成为一款专用温度场模拟软件。采集了棒材实际轧制过程工艺参数和实测温度场数据,利用自行开发的有限元预测软件对成品规格为Φ22 mm的棒材热轧过程温度进行了预测,发现实测值和计算值吻合良好,最大绝对温度偏差为13℃,最大相对误差约为1.4%,计算精度高。在单元总数为50的条件下,整个轧制过程计算时间约为200 ms,计算效率高。完全满足实际生产在线预测精度和速度的要求。更多还原

【Abstract】 During hot rolling,the temperature has an important influence on the rolling energy parameters and the microstructure and properties of the finished product.The construction of quick and accurate temperature online prediction model is of great significance for the realization of intelligent manufacturing in the rolling process.Based on the heat conduction finite element theory and the condition of calculation accuracy,this paper tries to shorten the response time of temperature prediction as much as possible,and then construct an on-line prediction model.The main research work is as follows:(1)Based on the heat transfer principle and actual rolling process,the research concentrates on the mathematical relationships among the process parameters of the rolling production line,the calculation formulas of heat transfer coefficient and as well as the internal heat source are included,developing the finite element program for solving the temperature field of air cooling process and water cooling process like bar,plate,and H-beam etc.Taking the bar as the research point,the solution time of the air cooling temperature is only 15.625 ms,the maximum difference of absolute temperature calculated is 38.3℃,and the maximum relative error is 5.5%;Besides,the maximum relative error calculated in the water-cooling program is higher than 300℃,which is 3.75%.Both of them meet the requirements of the accuracy and speed of the temperature field calculation in online production.(2)Taking the bar as the research object,the affecting rules of meshing methods on transient temperature oscillation,temperature convergence and calculation speed are studied.Here come to the results that when the number of grids in the air cooling process reaches 200,the calculation speed is faster and the calculation accuracy is higher,which is the optimal number of grids;Based on the element edge length and time step,the conditions for eliminating temperature oscillations are proposed;The parameters are reduced by 10 times compared with the conditions in ANSYS software,which means that the program has better calculation stability;What’s more,the layer-by-layer subdivision mesh can effectively restrain the temperature oscillation and improve the program operation speed and calculation accuracy(the calculation accuracy is improved by 0.21%);Furthermore,the specific conditions for eliminating the temperature oscillation are analyzed.Among them,the radial length of the edge element of the proportional method is 8.6 times than that of the equal difference mode.Narrowing the numbering gap of adjacent nodes in the unit can reduce the bandwidth of the coefficient matrix and improve the computational efficiency by about 22%.(3)In order to realize the solution of temperature field and dynamic meshing in continuous rolling process,newton quadratic optimal interpolation is proposed after studying and analyzing the interpolation methods such as inverse interpolation,polynomial interpolation(Lagrange interpolation,Newton interpolation)and piecewise interpolation method,which avoids the ill-conditioned properties of high-order interpolation polynomials and saves calculation time under the condition of ensuring accuracy,therefore to complete the design and programming of interpolation program.Taking the rolling deformation process of the bar as a research object,the thermal effects caused by plastic deformation work and friction has been comprehensively considered.The interpolation program is utilized to develop a temperature field simulation program for the hot rolling deformation process of the bar,and then the anti-red phenomenon is analyzed and explained.(4)The graphical user interface was developed by using Python language,which successfully realized the prediction and optimization of the temperature field during the hot rolling process of different sections of steel,transforming the self-made finite element program of this research into special temperature field simulation software.The process parameters and measured temperature field data of the actual rolling process of the bar were collected,and the self-developed finite element prediction software was used to predict the hot rolling process temperature of the finished product with a size of Φ22 mm.The absolute temperature deviation is 13℃,the maximum relative error is about 1.4%,and the calculation accuracy is enough.Under the condition that the total number of units is 50,the calculation time of the entire rolling process is about 200 ms,and the calculation efficiency is admirable.Therefore it fully meets the requirements of accuracy and speed of online prediction of actual production.更多还原