【作者】 陈飞；

【导师】 崔振山；

【作者基本信息】 上海交通大学 ， 材料加工工程， 2012， 博士

【摘要】 金属的微观组织是决定其宏观力学性能的主要因素，在热锻过程中，金属的微观组织会发生动态回复、动态再结晶、静态回复和静态再结晶等一系列变化，形成新的晶粒。在材料成分一定的条件下，影响晶粒演化的外部因素是温度、应变和应变速率。因此，如何通过控制热锻成形中的温度、变形量和变形速度，来达到控制微观组织及产品力学性能的目的，近年来已成为热锻成形中的热点研究问题，在大锻件成形中尤其如此，乃至于成性的重要性远大于成形。通过一定条件下的物理实验实测结果，建立微观组织演变的唯象数学模型，进而用来预报热成形过程的微观组织演变，是目前热锻成形中的常用研究方法，但这种研究思路只适用于变形比较简单、变形次数少的热成形过程。大锻件的成形一般是多个道次，每个道次包含多次压下，属于增量成形过程。每次压下中和相邻两次压下之间锻件可能发生完全或不完全动态再结晶、亚动态再结晶和静态再结晶，部分晶粒得到细化，应变能降低或消失，而未发生再结晶的晶粒则仍然保留高应变能，这种变形能的差异导致在随后的变形中再结晶驱动力不同，微观组织演变非常复杂，依靠实验方法建立的唯象组织演变模型已无法描述这样复杂的晶粒演变过程。本论文以典型能源大锻件材料为研究对象，探讨更能反映物理本质的微观组织演变建模方法。实际上，材料的位错密度是联系宏观变形过程与微观组织演变之间的桥梁，变形导致材料的位错密度增加，当位错能增加到一定程度时将导致再结晶发生，再结晶后的晶粒位错密度急剧降低，并在新的变形中再逐步累积和循环。根据“应变-位错密度-再结晶-流变应力”之间的宏微观相互影响规律，本文旨在建立包含“变形-回复-动态再结晶-亚动态再结晶和静态再结晶”各过程在内的微观组织演变元胞自动机（Cellular Automaton, CA）模拟方法，并与热成形过程宏观有限元模拟相结合，实现大锻件成形过程微观组织演变的预报。为此开展了以下主要研究工作：在Gleeble-3500热模拟实验机上，对典型大锻件材料进行了等温压缩实验。研究了不同应变速率和变形温度条件下材料流变规律与动态再结晶行为；根据经典应力-位错关系和动态再结晶动力学方程，分别对加工硬化-动态回复和动态再结晶两阶段建立了流变应力模型；同时从实验获得的高温流变应力曲线中提取了建立CA模型所需的材料特性参数。基于晶粒长大的热力学机制、曲率驱动机制和能量耗散机制，制定了能够反映晶粒长大物理机制的元胞转变规则，建立了模拟晶粒正常长大的CA模型。采用该模型模拟了高温奥氏体化晶粒正常长大过程，分析了晶粒正常长大过程CA模拟结果的形态学、动力学和拓扑学等典型特征。基于位错驱动的形核条件和晶粒长大动力学理论，建立了模拟动态再结晶过程的CA模型（DRX-CA Model）。DRX-CA模型综合考虑了位错密度的演变、再结晶形核及晶粒长大等一系列过程，同时考虑了合金元素及第二相粒子对再结晶的影响。应用所建立的DRX-CA模型预测了典型大锻件材料30Cr2Ni4MoV低压转子钢在不同应变、变形温度、应变速率和初始晶粒尺寸条件下动态再结晶行为，检验了DRX-CA模型对典型实验特征的预测能力。为了反映晶粒形状变化对再结晶形核与晶粒长大的影响，建立了包含晶粒拓扑变形技术在内的CA模型。该模型采用既相互关联又各自独立的两套坐标系，即物质坐标系和元胞坐标系，来描述材料动态变形中的再结晶形核与晶粒长大过程。其中物质坐标系用于描述晶粒的变形，晶粒形状及尺寸随着应变的发展而发生改变；元胞坐标系用于描述晶粒的形核和长大，元胞尺寸不发生改变，同时其转变规则也与不变形时相同，从而保证了新晶粒的等轴长大。通过对比采用晶粒拓扑变形技术的CA模拟结果、传统CA方法的模拟结果和金相实验结果，检验了包含晶粒拓扑变形技术的CA模型描述晶粒演化的准确性。建立了包含动态再结晶、亚动态再结晶和静态再结晶的多次非连续热变形过程的微观组织演变CA模拟模型，该模型以位错密度为基本变量，通过追踪非连续热变形过程中元胞内位错密度的变化，模拟复杂再结晶过程中微观组织的演变。采用该方法对典型大锻件材料30Cr2Ni4MoV低压转子钢四个道次热压缩过程微观组织演变进行了模拟。在此基础上，将有限元计算与CA模拟方法相集成，定义了CA计算时宏观热变形参数的更新规则，通过软件用户子程序开发，建立了预报热锻非连续热变形微观组织演变的多尺度数字化仿真系统，对典型大锻件成形工艺-多道次平砧拔长宏观物理场量和微观组织演变进行了多尺度数字化仿真，模拟结果对锻造工艺的优化提供了指导。研究结果表明，CA模拟结果比较准确地再现了温度、应变和初始晶粒度对晶粒演化过程的影响规律，CA模拟给出的晶粒尺寸及形貌与金相实验结果接近一致，根据CA模拟得到的位错密度而计算的流变应力曲线和实验条件下的流变应力曲线吻合良好，证明了以位错密度为基本变量追踪变形历史的CA模拟方法的预测结果符合材料热变形过程中的宏微观演变特征，可以用于模拟非连续热变形过程复杂再结晶微观组织的演变。将宏观有限元计算与微观组织CA模拟相集成，可以对热锻非连续变形过程宏观物理量场和晶粒演变进行多尺度数字化仿真，研究结果给出了多道次热变形的晶粒演化状态，根据模拟结果可为优化大锻件非连续热变形的锻造工艺提供指导。更多还原

【Abstract】 The prediction and control of microstructure evolution in heavy forgings during discontinuoushot deformation play a key role in improving the mechanical properties of heavy forgings. Heavyforging process usually consists of multi-pass and each pass consists of multiple reductions, so it is atypical process of incremental bulk forming. During multi-pass deformation, different parts of theforging at different time will stay in two states: deformation and non-deforamtion. Different statesinclude different mechanisms. Both incomplete and complete dynamic recrystallization (DRX) occurin the deformation state. In non-deforamtion state, incomplete DRX will result in meta-dynamicrecrystallization (MDRX) and static recrystallization (SRX). Complete DRX will result in graingrowth. The complicated recrystallization behavior leads to the non-uniform distribution of storedenergy among individual grains and within grain interiors. However, it is difficult to conceive atraditional phenomenological model to describe the complicated recrystallization processes.Therefore, it is highly necessary to develop a new algorithm that better reflects the physical behaviorobserved in heavy forging production. According to the micro-and macroscopic relationship among“strain, dislocation density, DRX and flow stress”, a mesoscale Celullar Automon (CA) model isdeveloped to predict the microstructure evolution for typical heavy forged materials during heavyforging production, during which DRX, SRX and MDRX may occur. The main research content is asfollows:The compression tests are conducted with Gleeble-3500thermo-mechaincal simulator to studythe DRX behavior and rheological regularity of LP and HP-IP rotor steels at various temperaturesand strain rates. Based on trandional stress-dislocation relation and kinematics of DRX, the flowstress constitutive equations of the work hardening-dynamical recovery period and dynamicalrecrystallization period are established to accurately describe the relationships of the flow stress,strain rate and temperature of LP and HP-IP rotor steels. More importantly, the necessaryexperimental data for the estabilishment of the CA model is abtained by conducting these physicsexperiments.The CA model based on the curvature-driven mechanism, thermodynamic driving mechanismand lowest energy principle has been developed to simulate normal grain growth during high-temperature austenitizing for LP rotor steel. In this model, the morphology, grain size distribution,topological aspects and local kinetics of austenite grain growth are annlyzed under differenttemperatures and calculation time, i.e.CAS.The DRX-CA model coupling foundamental metallurgical principles has been developed topredict microstructure evolution during DRX in heavy forged materials. The evolution of dislocationdensity, DRX nucleation and grain growth are under consideration. At the same time, the effects ofalloying elements and the second-phase particles on the recyrstallization nucleation and growth arealso taken into account in the developed DRX-CA model. In order to examine the prediction performance, DRX behavior of LP rotor steel is simulated under different strains, deformationtemperatures, strain rates and initial grain sizes.The effect of deformation on grain surface area per unit volume and edge length per unit volumeis another issure in DRX. During deformation, the dislocation density increases gradually and whenit reaches a critical density, new nucleus will appear on the grain boundaries and grow with themodel of equiaxed growth. This growth relies on the difference of dislocation density between thedynamic recrystallized grains and deformed grains. When the dislocation density in the new dynamicrecrystallized grains reaches the critical density again, the next round of DRX occurs. Thecompression only occurs on the grains that have not completed the DRX in each round. To describethe compression effect more accurately, an update CA model is propsed in which a cellularcoordinate system and a material coordinate system are establishe separately. The cellular coordinatesystem remains unchangeable, but the material coordinate system and the corresponding grainboundary shape will change with deformation. The simulation results show that the topologydeformation has an accelerated affect on the rate of DRX. The results also indicate that the averagegrain size exhibits a slight decreas because of the effect of topology deformation on therecrystallized grain growth. The average grain size is closer to the experimental results using thetopological CA model with a comparison of the conventional CA simulation.On the basis of the topological CA model, a mesoscale CA model is established to simulatemicrostructure evolution during discontinuous hot deformation for LP rotor steel. The model has thecapability of tracking the deformation history of each cell with dislocation density being an internalvariable, and accurately describes even complex recrystallization process. To examine the validity ofthe developed CA model, we attempted to apply the mesoscale simulation method to four-pass hotdefroamtion for LP rotor steel. The multi-scale simulation platform for prediction the microstructureevolution during discontinuous hot deformation is established by coupling FEM method and CAmethod. Bsed on the established simulation platform for microstructure evolution, numericalsimulations for stretching process with multi-stroke and multi-pass are performed. It proveides anovey way for investigating the microstructure evolution during the complicated recrystallizationprocesses observed in heavy forging production on a multi-scale.Research results show that the impact of the law of temperature, strain rate and intial grain sizeon grain evolution is revealed by CA simulation. Using the developed CA model, the relationshipbetween the flow stress, volume fraction recrystallized and recrystallized grain size and thethermomechanical parameters is established. By comparison of the simulated results by the CAmodel with the flow stress and metallographs from the experiment, it is obviously seen that thesimulation results closely resemble the experimental results. It is proved that the developed CAmodel accurately reflects the interaction mechanism between micro and macro evolution. Therefore,this model can then be used to simulate the microstructure evolution during the discontinuousforming processes. Multi-scale digital simulation can be achieved to predicte the evolution of themacro-physical fields and mesoscopic grain during discontinuous hot forging by coupling mesoscopic CA model with macro-scale finite element analysis. It shows that, based on the multi-scale simulation platform, the optimum forging method can be designed.更多还原