基于热源参数反向识别的定向能量沉积热力仿真

定向能量沉积(directed energy deposition,DED)热力仿真是有效预测沉积件残余应力和变形、优化工艺参数的重要方法,其仿真精度取决于输入参数的准确性.针对传统方法难以直接、准确获取输入参数的问题,文中以热源参数为例,提出基于支持向量机和遗传算法的参数反向识别方法,并用于精确构建实际DED工件热力...

Thermo-mechanical modelling of direct energy deposition using inversely identified heat source parameter

Thermal-mechanical simulation of direct energy deposition (DED) is an effective method to predict the residual stress and deformation and optimize deposition process parameters. The simulation accuracy depends on input parameters which are difficult to be directly and accurately measured in most cases. This paper, taking heat source parameters as examples, proposes an inverse identification method based on support vector machine and genetic algorithm, and the inversely-identified heat parameters are further used to improve the thermal-mechanical model accuracy of real DED applications. Firstly, simulation errors of simple single-track deposition models under different heat source parameters are obtained. Secondly, the relationship between the heat source parameters and simulation errors is established using support vector regression, and the optimized initial heat sources are identified using genetic algorithm. Thirdly, a forward-inverse closed loop is applied to narrow the ranges of heat source parameters for more precise parameter identification. Finally, a thermal-mechanical model for a turbine blade using the optimal parameters is constructed to further verify the proposed method. The results show that the optimal heat source parameters extracted from single-track deposition models can be extended to accurately predict the thermal and mechanical results of a turbine blade DED case, which provide a practical method for DED process optimization (e.g. distortion compensation) in industry applications.