摘要
近年来,在带肋钢筋产业供给侧改革背景下,基于DIFT机制的高速棒材产线提质升级是该领域的主要技术进步方向。对于24架轧机布置,3组水箱控冷的高棒产线,调控水箱水量可确保轧件温度达到实现DIFT的条件,是确保成品组织性能的关键。在传热与温度场基本模型基础上,引入遗传算法、BP神经网络智能化算法,实现了高速带肋钢筋棒材轧制中水箱冷却参数的优化与组织性能预报。结果表明:未优化前基础温度模型能较好地反应出轧件由表面至心部温度的变化,但轧后冷速过快会形成不利组织;采用遗传算法成功实现了水冷参数优化,结合基于BP神经网络的组织性能预报,为高棒产线成品组织性能调控提供了基础。
In recent years,under the background of supply-side reform ofribbed bar industry,the upgrading of high-speed bar production line based on DIFT mechanism is the main technological progress direction in this field.For the arrangement of 24 stands rolling mills,3 units water tanks of controlled cooling,the regulation of water volume of the water tank can ensure that the temperature of the rolled pieces meet the conditions for achieving DIFT,which is the key to ensure the microstrusture and performance of finish products.Based on the basic model of heat transfer and temperature field,the genetic algorithm and BP neural network intelligent algorithm were used to realize the optimization of cooling parameters and microstructure performance prediction of products in high-speed ribbed bar rolling.The results show that the basal temperature model before optimization can better reflect the change of temperature from the surface to the core of the rolled pieces,but the cooling speed after rolling will form an unfavorable structure.The optimization of water-cooling parameters was successfully realized by genetic algorithm,combined with the microstructure performance prediction based on BP neural network,which provided a basis for the regulation of production line microstructure performance.
作者
蒲春雷
姜嫄
方实年
沈存
高心宇
王永强
PU Chunlei;JIANG Yuan;FANG Shinian;SHEN Cun;GAO Xinyu;WANG Yongqiang(MCC Huatian Engineering and Technology Corporation,Nanjing 200240,China;School of Metallurgical Engineering,Anhui University of Technology,Maanshan 243002,China;Beijing Aerospace Petrochemical Technology Equipment Enqineering Corporation,Beijing 100036,China)
出处
《轧钢》
2023年第6期44-50,共7页
Steel Rolling
基金
安徽省重点研发计划项目(201904a05020008)
关键词
高速棒材
温度演变
组织性能预报
智能化
优化计算
high-speed ribbed bar
temperature evolution
microstructure and performance prediction
intelligent
optimization cal-culation
