基于联合概率分布的近β钛合金片层α相形核与各向异性长大动力学

To expand the availability of classical methods in predicting the behaviors of anisotropic growth and correlate the distribution of microstructure features to the yield strength of alloys, the project investigates the kinetics of nucleation and anisotropic growth of lamellar α phase in multicomponent near β titanium alloys based on the concept of joint probability distribution. It is a comprehensive research by combining microstructure characterizations, theoretical analyses and numerical simulations. We first modify the classical models of nucleation, growth and coarsening by introducing a joint probability distribution of nucleation density for tips and broad faces in lamellar α phase and then calculate the evolution of sizes and chemical compositions under the constraints of mass conservation and geometry of spherical crown. The model is validated against the experimental findings in Ti-1023 and Ti-5553 alloys. With the modeling, the changes of phase stability and the auxiliary diffusion flux due to the Gibbs-Thomson effect are to be evaluated. The role of multicomponent chemical coupling on interface migration is to be figured out. The captured mechanisms can help to precisely control the β↔ lamellar α phase transformation in industry. Subsequently, the distribution information of microstructure features will be passed to the internal-variable method to predict the yield strength of alloys. Not only mean values of microstructure features but also extrema and covariance of them are considered. We evaluate the weights of these statistical parameters in the determination of yield strength. The model and the simulation tool will be packaged in a software. The software together with the methodology opens a new window for application of physical metallurgical modeling in processing. It helps to optimize the processing parameters of heat treatment and is capable of predicting mechanical properties, thus does contributions to Chinese innovations in manufacture.

为弥补经典理论不能准确预测钛合金片层α相各向异性长大的缺陷，并建立片层组织特征参数的分布与屈服强度的关系，本项目结合实验表征、理论建模和计算，开展基于联合概率分布的片层α相形核与各向异性长大动力学研究，并以近β型Ti-1023和Ti-5553合金进行例证。项目从α片层尖端和宽面形核率的联合分布出发，建立新模型；在物质守恒和可变长宽比的球冠几何条件下，计算不同界面尺寸和成分的演化，揭示多组元耦合作用对β/α界面移动的影响机制，探明Gibbs-Thomson效应引起的近场扩散对α尖端和宽面形核长大的影响，为β↔片层α相变的定量调控提供理论依据；基于组织特征参数的分布演化，系统地评估屈服强度与片层组织特征平均值、极值、协方差等的关联强度，更全面地建立组织-屈服强度关系；开发具有工业应用价值的组织与性能调控软件，拓展经典方法在钛合金热处理工艺制定和力学性能预测中的应用，为我国制造业的创新做出贡献。

近β钛合金具有高强度和良好的塑性-断裂韧性匹配，已成为航空承力构件的主要材料。本项目以Ti-55531近β钛合金为研究对象，1)开展了β↔片层α相变的实验表征。针对β固溶+时效、双级时效、STA三种近β钛合金的典型热处理工艺，通过热膨胀测试和定量组织法建立了等温β→α相变动力学的TTT图。实验表征了各种热处理微观组织，系统评估了组织特征变化的普适趋势，即在有ω协助析出时，片层α的析出遵循非经典形核长大模式，无ω协助析出时片层α的析出遵循经典形核长大模式，析出动力学与Al的扩散同步。2）开展了片层α尖端与宽面独立生长动力学建模与预测研究。采用CALPHAD技术，校核了Ti-Al-Mo-Cr-V-Zr体系的β和α相的Gibbs模型和参数、β相的原子移动性参数，开发了计算多元化学驱动力和扩散势与独立演化程序实时通讯的脚本，预测了近β钛合金中片层α相尖端和宽面的独立生长和回溶，并与实验进行了对比，结果表明Gibbs-Thomson效应不仅降低了α尖端的稳定性，还附加扩散流，引起的尖端和宽面的成分差异，导致尖端和宽面生长之间的竞争。3）开展了片层α尖端与宽面联合演化生长建模与预测研究。在α相尖端和宽面独立生长模型的基础上，创新性地引入了尖端和宽面之间由Gibbs-Thomson效应引起的附加扩散流的作用，进一步修正了α相尖端和宽面生长。构建了可变长宽比条件下的几何形状因子，突破了传统KWN分布模型只适用于几何形状因子不变的析出模拟，并关联等效球形析出作为中介，实现了片层α相宽面尺寸空间分布和时间演化的计算。与实验结果进行比较，揭示了附加扩散流是相变过程中细长的针状α相向短棒状α相演化的物理本源，并对片层α相的几何形状因子、数量、尺寸和分布产生影响。4）采用机器学习方法，构建了热处理工艺、β+片层α组织特征和拉伸性能之间的对应关系，筛选了片层α的宽度体积比、β+α片层间距等基于组织特征的强韧化因子，并量化了其权重。采用“决策树分类”和“梯度下降”算法，准确地预测了合金的屈服强度和抗拉强度。