基于Logistic回归模型的钛合金件激光直接成形工艺参数分析

Analysis of Process Parameters about Direct Laser Metal Deposition Shaping Process of Titanium Alloys Based on Logistic Regression Model

如果使用激光金属直接成形方法制备金属零件时,能够通过最优的工艺参数来有效地控制金属成形件内部的不同区域及成形件与基板间的温度梯度和能量分布,使成形件内部产生缺陷的几率减小,会极大地提升成形件的力学性能。针对激光金属直接成形方法制备金属零件内部是否存在缺陷的二分类问题,通过使用正交旋转实验设计的方法进行工艺实验,得到不同参数下的钛合金激光金属直接成形试样是否存在内部缺陷情况。对实验结果进行Logistic回归建模,使用Logistic回归模型计算的工艺参数进行验证实验,样件内部未发现内部缺陷,说明Logistic回归模型结果与实际结果有较高的一致性。Logistic回归模型对激光金属直接成形工艺参数的选择具有积极的指导意义。

When machining metal parts with laser metal deposition shaping （LMDS） method, if the temperature gradients and energy distributions of different areas inside cladding layer and that between cladding layer and substrate can be effectively controlled with optimal process parameters, then the probability of internal defects of cladding layer can be reduced, and the mechanical properties will be greatly improved. To the problem whether the metal parts made by LMDS contain internal defects, rotatable orthogonal experiments are designed for acquiring the distributions of internal defects of the metal parts under different process parameters. A logistic regression model is built based on these experimental data. Finally an authentication test is taken with the process parameters calculated by the logistic regression model, and no internal defects are found. The results show that the logistic regression model is in good agreement with the actual situation. The logistic regression model proposed in this paper is proved to have the guiding significance for selecting LMDS process parameters.