登录 EN

添加临时用户

镍基单晶高温合金定向凝固微观组织相场模拟

Phase-field Simulation of Microstructure Evolution in Nickel-based Single Crystal Superalloy Directional Solidification

作者:杨聪
  • 学号
    2015******
  • 学位
    博士
  • 电子邮箱
    yan******.cn
  • 答辩日期
    2020.05.21
  • 导师
    许庆彦
  • 学科名
    材料科学与工程
  • 页码
    143
  • 保密级别
    公开
  • 培养单位
    035 材料学院
  • 中文关键词
    镍基高温合金,微观组织,多相场模拟,定向凝固
  • 英文关键词
    Nickel-based superalloy, Microstructure, Multiphase-field simulation, Directional solidification

摘要

镍基单晶高温合金叶片是航空发动机中最重要的热端部件之一,它通常由定向凝固技术制造。单晶叶片的微观组织特征决定了其高温力学性能,而定向凝固工艺对微观组织形成具有重要影响。使用相场模拟与实验相结合的方法研究了镍基单晶高温合金定向凝固中微观组织的形成过程,并探究了不同凝固条件下微观组织的演化规律,研究结果对优化单晶叶片生产工艺、提高合格率具有重要意义。建立了耦合高温合金热力学数据库的多元合金多相场模型。提出了求解多相场模型的高精度插值算法,该算法能够将相变热力学驱动力计算的相对误差降低至1%以下。开发了基于GPU的多相场模拟并行计算方法,并提出了异步并行数据传输与计算算法解决GPU显存不足的问题。实现了凝固过程三维枝晶生长的相场模拟,结果表明GPU能提速100倍以上,效果明显。系统开展了不同定向凝固条件下枝晶组织生长的相场模拟,建立了更精确的单晶高温合金的一次臂间距解析表达式。开发了耦合格子-玻尔兹曼方法的多相场模型,实现了对流作用下枝晶生长模拟,结果表明在低冷速下熔体自然对流强度大,能造成枝晶偏斜、重熔与分枝,并减小一次枝晶臂间距。基于不同凝固条件下的糊状区熔体流速预测了雀斑缺陷形成敏感区。对高温合金枝晶间组织生长过程进行了实验和多相场模拟研究。模拟得到了与实验一致的组织形貌,证实了共晶反应的发生,并揭示了共晶组织粗化原因和粗大γ'相附近γ枝晶重熔机制。模拟和实验中得到了相吻合的元素偏析规律,其中Al、Nb、Ta富集在γ'相中,Co、Re、W富集在γ相中,Cr、Mo元素富集在最后凝固的γ相中。对单晶试棒定向凝固温度场进行了模拟,并实现了不同冷却区域枝晶-共晶组织耦合生长模拟,模拟得到的枝晶-共晶组织形貌结果与实验结果吻合良好。对高温合金固溶热处理过程进行了模拟与实验验证,基于热力学计算和多相场模拟分别得到了合金初熔温度与γ'相分数随保温时间的变化规律,模拟结果能进行固溶热处理工艺设计。

Nickel-based single-crystal (SX) superalloy blades, which are mainly produced by directional solidification (DS) technique, are among the most important high-temperature parts in aero engines. The high-temperature performance of SX blades is determined by its microstructure, and DS process can essentially affect the microstructure formation. In this work, the phase-field simulation coupled with experimental study was used to study the microstructure formation during nickel-based superalloy DS. The results are helpful for optimizing the SX blade DS process and improving the product qualification rate.The multiphase-field model coupled with superalloy thermodynamic database was established to study the microstructure evolution during multicomponent superalloy DS. To solve the multiphase-field equation, a high precision extrapolation algorithm was developed, which significantly decreases the relative computing error of phase transition thermodynamic driving force to below 1%. Besides, a GPU-based parallel computing scheme was developed to accelerate the phase-field simulation, and an asynchronous concurrent data transfer and computation algorithm was proposed to solve the problem of insufficient GPU memory. Then, 3D dendrite growth simulation was performed to demonstrate the ability and efficiency of the developed GPU-based computing algorithm.Systematic simulation study on dendrite growth under various of solidification conditions was conducted, and a more accurate analytical expression for predicting superalloy primary dendritic arm spacing (PDAS) was built based on the simulation results. To consider fluid flow during dendrite growth, the multiphase-field-lattice Boltzmann model was developed. The simulation results have demonstrated that the segregation induced natural convection is much stronger when under lower cooling rates, and the interdendritic fluid flow can cause dendrite deflection, remelting, branching and even reduce the PDAS. Based on the simulated interdendritic fluid velocity, the freckle prone region was predicted, which agreed well with the experimental results. The interdendritic microstructure growth during superalloy solidification was investigated by both experiments and multiphase-field simulations. The simulated interdendritic microstructure was in consistent with experimental finding. The simulation results have confirmed the eutectic reaction at the final stage of solidification, revealed the eutectic microstructure coarsening and γ dendrite remelting mechanism. The simulated microsegration patterns of the alloy elements are in quantitative agreement with the experimental results, which indicates Al, Nb and Ta segregate to the γ' phase, Co, Re and W segregate to the γ phase, and Cr, Mo segregate to the last solidified γ phase. To achieve macro and micro coupled simulation during superalloy DS, the macro temperature field of a SX bar casting was first simulated, and the cooling curves in four typical solidification positions were extracted. The cooling curves were used as the input for the microstructure simulation, and the simulated dendrite and interdendritic microstructure agreed well with the experimental results. Then the superalloy solution heat treatment (SHT) process was simulated and the results were verified by experiments. Based on the thermodynamic calculation and multiphase-field simulation, the incipient melting temperature and the microstructure evolution were obtained respectively, which can be used to design the SHT process for SX superalloy.