Study on high temperature mechanical behavior and microstructure evolution of Ni3Al-based superalloy JG4246A

Journal of Materials Research and Technology

Changbo Sun¹  Tongtong Li¹  Jiantao Wu²  Qingyan Xu³  Jiangwei Zhong³ 

[1] AECC Shenyang Liming Aero-Engine Group Corporation Ltd, Shenyang 110043, China;Central Iron and Steel Research Institute, Beijing 100081, China;Key laboratory for Advanced Materials Processing Technology(MOE), School of Materials Science and Engineering, Tsinghua University, Beijing 100084, China;

关键词: Ni3Al-based superalloy;    True stress-strain;    Constitutive model;    Microstructure;    Mechanical properties;   

DOI  :  

来源: DOAJ

【 摘 要 】

Ni3Al-based polycrystalline intermetallic compound is a kind of engineering superalloy, which have been used in the adjustment flaps of aeroengine tail nozzle. In order to explore its mechanical properties and microstructure evolution in service, the high temperature flow characteristics and microstructural changes of alloys were investigated by uniaxial isothermal compression tests on Gleeble-1500. The true stress-true strain curves were obtained and the microstructures were analyzed by OM, SEM, and EBSD. Based on the experimental data, three different constitutive models, belonging to phenomenological model, physical-based model, and radial basis function (RBF) neural network model, separately, were established to predict the deformation stress. The results show that the true stress is very sensitive to temperature and stain rate. Flow instability occurred during the deformation process, which caused some decline in the stress. Microstructure evolution under different deformation conditions indicated that the γ' to γn' phase transformation was also related to the degree of deformation and the four stage evolutionary process was identified. The SEM and EBSD results showed that the DRX appeared at the later stage of the deformation. The prediction capabilities of the three models were assessed by correlation coefficient (R), the absolute value of relative error (ARE), and the absolute average error (AARE). The result demonstrated that the RBF neural network had the best ability to predict the flow stress and the coefficient.

【 授权许可】

Unknown