选择性激光熔化致密度预测模型及工艺参数优化

Selective Laser Melting Density Prediction Model and Process Parameter Optimization

针对选择性激光熔化(SLM)成形件质量低和工艺参数难以控制的问题,选取激光功率、扫描速度、铺粉层厚、扫描间距为优化变量,成形件的致密度为优化目标,设计正交实验获取训练样本,应用BP神经网络建立了针对316l不锈钢材料的致密度预测模型,然后通过遗传算法对网络模型进行优化和工艺参数寻优。结果表明:优化后的致密度模型预测相对误差在0.73%左右,预测能力较好且波动较小,并基于模型寻优到激光功率197.28W,扫描速度623.85mm/s,铺粉层厚0.1379mm,扫描间距0.1139mm的最佳工艺方案。模型能准确地反映出工艺参数与致密度之间的映射关系,为SLM成形参数优化提供了新的思路。

For the problem of low quality of selective laser melting(SLM)forming parts and difficult control of process param-eters,Select the laser power,scanning speed,coating thickness,scanning interval as the optimization variables,and the density of the shaped parts as the optimization target,Design orthogonal experiments to obtain training samples,using BP neural net-work to establish the density prediction model for 316L stainless steel,then use the genetic model to optimize the network model and optimize the process parameters.The results show that the relative error of the optimized density model prediction is about 0.73%,with good forecasting ability and small fluctuations.Based on the model,the optimal process solution with laser power of 197.28W,scanning speed of 623.85mm/s,coating thickness of 0.1379mm and scanning distance of 0.1139mm was found.The model can accurately reflect the mapping relationship between process parameters and density,which provides a new idea for SLM forming parameter optimization.