基于多目标遗传算法的工作辊温度场计算与分析  

Work roll temperature field calculation and analysis based on multi-objective genetic algorithm

作　　者：居龙[1] 李洪波[1] 张杰[1] 胡超[2] 张超[2] 陈良[2] 

机构地区：[1]北京科技大学机械工程学院,北京100083 [2]武汉钢铁(集团)公司,武汉430083

出　　处：《北京科技大学学报》2014年第9期1255-1259,共5页Journal of University of Science and Technology Beijing

基　　金：北京科技大学冶金研究基金资助项目(YJ2010-010)

摘　　要：等效换热系数是热连轧机工作辊温度场仿真模型的核心输入参数，多采用遗传算法优化得到，某1800 mm 热连轧机存在品种、规格交替轧制，等效换热系数的准确计算比较困难.选取多组典型工艺条件下的工作辊下机后表面温度作为优化目标，采用多目标遗传算法进行优化，并通过改变遗传算子有效避免了算法早熟及局部收敛等问题，获取了具有较强适应性的等效换热系数.仿真和实测数据的对比结果证明了优化模型的可靠性.利用仿真模型分析了主要工艺参数对工作辊热凸度的影响，并提出同宽交替时，工作辊热凸度随轧制进程呈指数变化，而在品种、规格交替编排轧制工艺下相邻带钢轧制时工作辊热凸度存在6-21.8μm 的波动，且随轧制进程趋于稳定.Equivalent heat transfer coefficients are key input factors for work roll temperature field analysis models in hot strip roll-ing, and the coefficients are usually calculated with genetic algorithms, but the rolling process of alternating material and width in a work roll service period for a 1800 mm hot strip rolling mill makes the coefficients difficult to be calculated accurately. The surface tem-perature distributions of a serviced work roll under multiple rolling schedules were taken as the optimization goals, and the equivalent heat transfer coefficients with a higher adaptability were obtained by using a multi-objective genetic algorithm. The optimizing process was improved by changing the genetic operators, which avoided the disadvantages of premature convergence and local convergence. The optimization model was proved effectively by comparing the simulation results with measured data. The effects of rolling parameters on the thermal crown were analyzed with the optimization model. It is predicted that under alternating width rolling, the thermal crown ex-ponentially increases; but under alternating material and width rolling schedules, the thermal crown has a fluctuation of 6 to 21. 8 μm between the continuous two strips, and towards stability gradually.

关 键 词：热轧机 轧辊 温度场 换热系数 遗传算法 多目标优化 

分 类 号：TG333.17[金属学及工艺—金属压力加工]