大规格气阀电镦成形晶粒匀细化控制研究

Study at Controlling Grain Refinement and Distribution Uniformity in Electric Upsetting Process of Large-scale Valve

提出一套可以控制电镦过程晶粒匀细化的方法,首先从Ni80A镍基超合金热加工图识别出晶粒细化稳健变形参数区间并初步规划电镦工艺参数加载路径;构建晶粒尺寸和晶粒分布不均匀度双目标控制函数,建立电镦成形电-热-力多场、变形-晶粒多尺度动态耦合有限元模型;开展模拟夹持长度、砧子后退速度、镦粗缸速度和加热电流等不同电镦决策变量的加载方案,构建双目标与电镦决策变量的响应关系;基于NSGA-Ⅱ遗传算法求解获得四项决策变量对双目标交互作用的最优折中解方案。结果表明,根据最优折衷解模拟得到的晶粒尺寸及其分布不均匀度分别为29.95μm和6.52μm,双目标值相较于优化前分别减小了10.86%和14.90%。对中等规格气阀进行电镦工艺试验和组织表征,电镦“蒜头”的晶粒尺寸在轴向、径向和45°方向的模拟值与试验值之间的平均相对误差分别为4.41%、1.39%和5.68%,验证了电镦成形晶粒匀细化控制方法的有效性。这套方法将为调控大规格气阀电镦成形晶粒匀细化提供一般方法。

A method for controlling grain refinement and uniform distribution in electric upsetting process is proposed.Firstly,the grain refinement domains are identified from the hot processing map of Ni80A superalloy,and then the parameters loading paths in electric upsetting process are preliminarily designed.The double objectives control functions of grain size and inhomogeneous degree of grain distribution are developed,and the electrical-thermal-mechanical multi-filed and deformation-grain multi-scale coupling finite element model in the electric upsetting of Ni80A superalloy is established.Additionally,the decision variables loading schemes with different levels of clamping length,anvil astern speed,upsetting cylinder speed and heating current are simulated,and the relationships between grain size,inhomogeneous degree of grain distribution and decision variables are constructed.The optimal compromise solution for the interaction between four decision variables and double objectives is solved based on NSGA-Ⅱgenetic algorithm.The simulated results of optimal compromise solution show that grain size and inhomogeneous degree of grain distribution are 29.95μm and 6.52μm,compared with before optimization,they are decreased 10.86%and 14.90%,respectively.Finally,the trial-produce experiment for the electric upsetting of medium-size valve and microstructure characterization are carried out.The average relative errors of grain size in axial,radial and 45 degree directions between simulated values and experimental ones are 4.41%,1.39%and 5.68%,respectively.The results also verify the validity of the proposed method.This method will provide a basic approach for controlling grain refinement and uniform distribution in the electric upsetting process of large-scale valves.