金属和合金高温变形过程本构模型的研究进展

Advances in Constitutive Models of Metals and Alloys During Hot Deformation

本构模型是预测金属和合金高温变形行为的重要途径,在不同金属和合金选择合适的变形工艺参数、预防缺陷等方面起着至关重要的作用。在近年对金属和合金高温变形过程的研究中,常通过不同工艺参数下的各类高温变形试验来获取建立本构模型的原始数据,并将所获本构模型导入Deform、Ansys等模拟软件相应模块,以预测材料在锻造等过程中应力、应变速率、温度的分布规律,进而优化实际加工参数、避免缺陷的产生,同时减少耗材及资源浪费。鉴于本构模型在优化加工参数、预防缺陷等方面的重要作用,对金属和合金本构模型的建立、选择等方面的研究较多,选择何种试验方法来获取建立材料本构模型的试验数据、运用何种数学或物理方法来建立材料的本构模型、选择何种本构模型进行预测、各类本构模型的优缺点及修正方法等都是金属和合金高温变形过程本构模型的研究重点。在近些年的研究中,常运用热压缩、热拉伸、热扭转、分离式霍布金森压杆等高温变形试验方法来获取材料不同高温变形工艺参数下的原始数据,进而建立其本构模型。常用的本构模型大致可分为唯象型、物理基型及基于人工神经网络型。各类模型分别具有不同的适用性及优缺点,而缺点最终主要体现在部分工艺参数下的拟合偏差较大,针对该现象,各国学者不断对模型进行完善、修正,其中,除了模型本身的原因外,引起偏差的原因还包括没有考虑摩擦及变形热等宏观问题的影响。目前,常用的典型唯象型本构模型包括Arrhenius型本构模型、Johnson-Cook模型等,物理基模型如Zerilli-Armstrong型等,而基于人工神经网络模型则主要是利用输入层、隐含层及输出层进行预测,各类模型在数据处理的复杂性、物理意义等方面各有优缺点。文章从金属和合金高温变形过程获取本构模型原始数据的试验方法、本构模型的种类及修正、模型的应用等方面综述了本构模型的研究及发展,分析并总结了不同本构模型的优缺点,指出了模型预测过程中个别参数下预测值与试验值偏差较大的现象及其修正方法,并展望了金属和合金高温变形过程本构模型未来的研究方向。

Constitutive model is an important way to predict the hot deformation behavior of metals and alloys,which plays a significant role in selecting suitable deformation parameters and preventing defects for various metals and alloys.In the recent investigations of hot deformation behavior of metals and alloys,the original data for constitutive model is generally obtained by various kinds of hot deformation tests under different parameters,and then obtained constitutive model is imported to the corresponding part of simulation software like Deform and Ansys to predict the distribution law of strain,strain rate and temperature of materials during forging process,and further optimize the actual processing parameters,avoid defects,and reduce the waste of materials and resource.Considering the importance of constitutive models in optimizing the working parameters and preventing the defects,tremendous strudies on construction and selection of constitutive models were conducted.The research focuses of constitutive model of metals and alloys in hot deformation can be concluded in the following aspects：the test method to obtain the original data for constitutive model construction,the mathematical or physical method for establishing the constitutive model,the selection of constitutive models for specified object,the advantage and disadvantage of various models and their revision.In past years,the original data for constitutive model are usually acquired by hot compression,hot extension,hot torsion and Split Hopkinson Pressure Bar tests under different hot deformation parameters.The common constitutive models can be generallyclassified as phenomenological method,physical-based method and the ones using artificial neural network.Each kind of model has its unique applicability,advantages and disadvantages.The disadvantages finally lead to the apparent fitting deviation of several deformation parameters.Aiming at reducing the deviation,scholars over the world have been devoted to revise and improve the models.Apart to the intrinsic reasons of the models,the lack of consideration of the grand factors like friction and deformation heat would also lead to the deviation.Presently,the common phenomenological constitutive models include Arrhenius model,Johnson-Cook model and so forth,the physical-based model includes Zerilli-Armstrong model,while the artificial neural network model is always conducted by three layers.These models have various advantages and disadvantages in dealing with data or physical meaning.This article summarizes the researches and developing direction of constitutive models from the aspects of the experimental methods to achieve the original data,the kinds and revision,and applications of the constitutive models.Also,the advantages and disadvantages of the typical constitutive models are discussed.The apparent deviation phenomenon of predicted and experimental data under several deformation parameters and the modified methods are pointed out.Finally,the future investigation direction of constitutive models of metals and alloys during hot deformation is predicted.