【作者】 张青；

【导师】 薛克敏； 李萍；

【作者基本信息】 合肥工业大学 ， 材料加工工程， 2008， 博士

【摘要】 随着塑性成形技术的发展以及对产品质量要求的进一步提高,需要更为全面、细致、深入地了解和掌握生产中工件内部组织的变化过程,因而分析变形过程中的微观组织形貌并采用计算机对其演变进行模拟近年来已成为科学工作者研究的热点,具有重要的理论意义和实际应用前景。晶粒形貌作为表征材料微观组织的重要指标,迄今为止还无法采用传统的数学方法对其进行定量描述,分形几何的应用则为解决这一难题提供了强有力的工具。本文对新型近β型钛合金—TB8合金的高温变形行为及变形过程中动态再结晶规律进行了系统研究,首次采用分形理论对其热变形显微组织进行深入分析,建立了组织演变的计算机分形模拟模型,利用该模型对动态再结晶过程进行模拟。本文主要工作如下:1、在热模拟试验机上对TB8合金进行了恒温恒应变速率热模拟压缩实验,研究其高温变形行为,并建立了该合金的高温流变应力模型,该模型能够精确描述高温变形条件下合金的流变应力,是进行有限元数值模拟的先决条件。2、采用刚粘塑性有限元法,引入前面得到的TB8合金流变应力本构模型对热压缩过程进行了热力耦合有限元数值模拟,结果与实验吻合较好,在此基础上得出应变、位移等场变量,为热变形组织模拟提供必要前提。3、精确的显微组织预测模型是提高组织模拟精度的关键环节。本文依据热模拟实验结果,对TB8合金动态再结晶规律进行研究,建立了其动态再结晶晶粒尺寸模型和动态再结晶动力学模型,研究表明所建模型能够比较精确地预测显微组织参数随变形工艺参数的变化而变化的情况。这不仅为提高组织模拟精度创造前提,也为制定合理的热加工工艺提供了理论依据。4、要进行组织的分形模拟,首先必须根据分形理论对热塑性变形显微组织进行分析。本文在图像处理的基础上,分别采用小岛法和盒维数法对再结晶和形变显微组织的分形维数进行计算,研究表明变形过程中的显微组织具有典型的分形结构,可以采用分形维数对其进行定量描述。文章系统分析了显微组织分形维数与变形条件之间的关系,首次采用人工神经网络的方法建立了形变显微组织的分形维数与变形温度、应变速率关系的预测模型,为变形金属显微组织的定量研究及精确的微观组织演变模拟提供了更为科学的理论基础。5、以分形模拟方法—元胞自动机方法为基础,实现正常晶粒生长的二维计算机模拟,获得了与实际金属较为相似的显微组织形貌。分析表明,模拟所得晶粒形貌具有分形特征,且晶粒的演化符合动力学生长规律,采用该方法模拟得出的显微组织是合理有效的。6、为了既能反映动态再结晶组织演变的动力学机制,又能真实再现组织演变的过程,本文在晶粒长大模拟的基础上,将热力耦合刚粘塑性有限元模拟、微观组织的动力学模型和基于分形的随机性模型—元胞自动机模型相结合,建立了动态再结晶显微组织演变的分形模拟模型。结果表明,模拟结果很好地描述了动态再结晶动力学规律,获得的组织与实际组织非常接近,其分形维数数值及变化趋势也均与实际结果符合较好,因而可以精确再现动态再结晶显微组织的整体形貌。以上研究表明,本文所采用的热力耦合有限元模拟、再结晶显微组织的动力学模型、晶粒生长的元胞自动机模型以及将三者进行结合所得到的再结晶显微组织演变模型具有较高的可靠性和稳定性,对于优化工艺参数、提高产品的组织性能具有实际的指导意义。更多还原

【Abstract】 With the development of hot deformation technique and the higher requiring to product quality, it is necessary to know microstructure evolution of products during hot deformation comprehensive and thoroughly. Therefore, analysis and simulation on the evolution of microstructure has been a researching hotspot in recent years. It has important theory significance and utility. Grain shape is one of the important indexes to characterize microstructure. However, classical stereology has some difficulties in describing irregular shapes of grain quantificationally. The use of fractal geometry offers a powerful instrument for solving the problem. In this study, the hot deformation behavior and dynamic recrystallization rule of TB8 alloy were systemic studied. The fractal theory was applied to analyse the microstructure during hot deformation for the first time. On the basis of kinetic theory, a fractal simulation model was established to simulate the microstructure evolution during dynamic recrystallization process. The major content can be summarized as follows:Hot compression test of TB8 alloy were performed on Gleele-1500. Hot deformation behavior was studied and the model of flow stress was obtained by hyperbolic-sine-type Arrhenuis equation. This model can describe the flow stress during hot deformation accurately and become a predeterminate condition of FE numerical simulation.The thermal-mechanical coupled rigid-viscoplastic finite element simulation was carried adopting the model of flow stress of TB8. Simulation result show good agreement with experiment. Strain and displacement during hot deformation were calculated which provided an essential precondition for microstructure simulation and proved the correctness of flow stress model.In order to improve precision of the microstructure simulation, accurate microstructure predicting model is needed. In this paper, dynamic recrystallization rules of TB8 alloy were studied, size model and kinetic model of dynamic recrystallization were established by experience formula. The results show that this method is able to predict the variation of microstructural parameters with technique parameters successfully. It not only creates necessary prerequisite for improving the accuracy of microstructure simulation, but also lays a scientific foundation for determining reasonable hot forming process.To perform fractal simulation for microstructure, the fractal theory was applied to analyze the recrystallized microstructure during hot deformation. Based on the computer image treatment, Slit Island Method and Box Counting Method were applied to calculate the fractal dimension of recrystallized microstructure and deformed microstructure. It is shown that microstructure presents typical fractal character, and fractal dimension can be used to describe its degree of irregularity. With decreasing strain rate and increasing deformation temperature, the fractal dimension of recrystallization grain decreases while grain size increases. At the same time, the variety of deformation temperature and strain rate results in the changes of grain amount in unit volume, recrystallization grain size, the size and shape of original grains to a certain extent, which influences the fractal dimension of deformed microstructure consequently. Owning to the nonlinear and complexity of relationship between fractal dimension and deformed parameter, the predicting model was established with a three-layer feed-forward artificial neural network with a back-propagation learning rule for the first time. It provides more scientific theory foundation for the quantitative study of deformed microstructure and accurate simulation of microstructural evolution.Based on the fractal simulation method—Cellular Automaton method, efficient simulation on normal 2D grain growth was accomplished, the simulated microstructure was similar to actual alloy. Analysis shows that the grain shape by simulation presents fractal character and simulation results accords with the kinetic rule of grain growth. So the initial microstructure by this method is rational and effective.In order to reflect kinetic mechanism of DRX microstructural evolution and show the evolutional process truly, dynamic fractal simulation of recrystallized microstructure was realized by incorporating the coupled thermomechanical rigid-viscoplastic FE model and the kinetic model of microstructure into Cellular Automaton model. Compared with the experimental results, it shows that the simulated microstructures are similar to the actual ones. The fractal dimensions and their variety tendency agree well with the actual results too. So the whole shapes of dynamic recrystallized microstructure can be displayed accurately. The simulated results describe the kinetic rule of DRX well. The study shows that: the thermomechanical rigid-viscoplastic finite element simulation, kinetic model of dynamic recrystallization, Cellular Automaton model of grain growth and microstructural evolution model by combining them are reliable and stable. It has important practical meanings for optimizing the deformation process and improving the microstructure and capability of product.更多还原