激光立体成形Ti-Al-Mo合金成分-组织相关性研究  

Relationship between Composition and Microstructure of Ti-Al-Mo Alloys Prepared by Laser Solid Forming

作　　者：武文璐 李静 赵可馨 邱莹 王猛 张凤英[1] Wu Wenu;Li Jing;Zhao Kexin;Qiu Ying;Wang Meng;Zhang Fengying(School of Materials Science and Engineering,Chang'an Univeristiy,Xi'an 710064,China;Shaanxi Aero-space Power High-Tech Co.,Ltd.,Xi'an 710077,China;State Key Laboratory of Solidification Processing,Northwestern Polytechnical University,Xi'an 710072,China)

机构地区：[1]长安大学材料科学与工程学院,陕西西安710064 [2]陕西航天动力高科技股份有限公司,陕西西安710077 [3]西北工业大学凝固技术国家重点实验室,陕西西安710072

出　　处：《稀有金属》2023年第5期714-724,共11页Chinese Journal of Rare Metals

基　　金：陕西省重点研发计划项目(2023-YBGY-359);西安市科技计划项目(21ZCZZHXJS-QCY6-0001,21CXLHTJSGG-QCY8-0003);空军工程大学等离子体动力学重点实验室开放基金(6142202210203);中央高校基本科研业务费项目(300102312407,X202210710372)资助。

摘　　要：研发高性能专用钛合金材料是激光增材制造钛合金在高端装备领域发展和应用的关键。然而,现有的合金设计方法效率低,难以建立钛合金成分-显微组织的定量关联。利用送粉激光增材制造(激光立体成形(LSF))高效制备系列成分钛合金的技术优势,制备60组不同成分的Ti-Al-Mo合金试样,结合显微组织观察和图像处理方法,获得不同成分Ti-Al-Mo合金的显微组织参数(晶内初生α板条宽度(W_(α))及α相体积分数(F_(α))),进而结合Back Propagation(BP)神经网络模型的建立,获得激光立体成形Ti-xAl-yMo(2.0≤x≤7.5,2.0≤y≤9.0)合金成分-显微组织特征的定量关系。结果表明:W_(α)测量值主要分布在0.1~1.1μm范围内,Al含量一定时,W_(α)随Mo含量的升高而下降。在Al含量相对较低时(2.0%~4.5%,质量分数),W_(α)呈现先缓慢下降、然后快速下降、进而缓慢下降的非线性趋势;而Al含量相对较高时(5.0%~7.5%),W_(α)呈现先缓慢下降、然后加速下降、再近似线性下降的趋势。Al含量一定时,F_(α)随Mo含量的增加并未呈现连续下降的趋势;而是在某些成分范围内,出现了随Mo含量增高,F_(α)反而增高的现象;这表明在激光立体成形条件下,Mo元素的合理添加可实现α板条细化和F_(α)提升的协同作用效果。分析这种现象的产生是由于激光往复热循环作用下,β转变基体上析出大量次生α相而导致。BP神经网络模型的预测结果与典型成分合金的实验观察结果吻合较好。Laser additive manufacturing can create complicated titanium alloy parts in near-net shape and enormous sizes without us-ing molds.It has important application prospects in the field of high-end equipment manufacturing.However,due to the unique charac-teristics of non-equilibrium rapid solidification and complex thermal cycling during laser additive manufacturing,the microstructures of the laser deposited titanium alloy parts are different from those prepared by conventional processing techniques,resulting in their high strength and low plasticity,as well as the anisotropy of mechanical properties.There is an urgent need to develop special titanium alloys according to the characteristics of laser additive manufacturing.However,current alloy design methods are time-consuming and inefficient.It is also difficult to establish quantitative correlations between the composition of titanium alloys and their microstructure.In this work,by combing the technical advantages of powder feeding laser additive manufacturing(laser solid forming,LSF)and the establishment of Back Propagation(BP)neural network model,the quantitative relationship between the composition and microstruc-ture of LSF titanium alloys was studied.Ti,Al and Mo element powders were used as raw materials,and 60 sets of Ti-Al-Mo alloy specimens with different compositions were prepared by LSF.Electron-probe micro-analysis(EPMA)was used for quantitative analy-sis of alloy compositions,and scanning electron microscope(SEM)was used for microstructure observation.The microstructure param-eters of Ti-Al-Mo alloys(width ofαlath(W_(α)),and volume fraction ofαphase(F_(α)))were quantitatively characterized by Image-Pro Plus(IPP)software,so as to obtain the microstructure parameters of Ti-Al-Mo alloys prepared by LSF.Further,BP neural network model with a double hidden layer was established,and the prediction accuracy analysis of the model was performed.The results showed that the relative prediction errors of W_(α)and F_(α)were less than 15%,and the determinatio

关 键 词：激光增材制造 Ti-Al-Mo合金 Back Propagation(BP)神经网络 显微组织 

分 类 号：TF803.21[冶金工程—有色金属冶金]