基于响应面法和支持向量机模型的选区激光熔化参数优化
Optimization of Selective Laser Melt Parameters Based on Response Surface Method and Support Vector Machine Model
刘玉德 1高钰淳 1石文天 1林宇翔 1贾世龙1
作者信息
- 1. 北京工商大学人工智能学院,北京 100048
- 折叠
摘要
合理的激光预烧结参数可降低选区激光熔化样件的表面粗糙度.为获得优质的加工参数,研究预烧结过程中激光功率、曝光时间、线间距和点间距对表面粗糙度的影响,并建立响应面法(RSM)和经蛇算法(SO)优化支持向量机(SVM)模型,使其能够预测并优化最小表面粗糙度对应的输入参数,结果表明,两模型都具有优秀的预测能力,但在优化能力和泛化能力上SO-SVM模型更为优秀,经SO-SVM模型优化后的工艺参数得到的最小表面粗糙度为17.7μm,小于响应面法优化得到的19.3 μm.研究成果可为降低表面粗糙度,大大减少加工过程的试错成本并获得更优质的加工产品提供一种参考.
Abstract
Optimal laser pre-sintering parameters are essential for reducing the surface roughness of laser melting samples.This paper investigates the impact of process parameters,including laser power,exposure time,line spacing,and point spacing,on surface roughness during the pre-sintering process.To achieve high-quality processing parameters,a response surface method(RSM)and a support vector machine(SVM)model optimized by the warp snake algorithm(SO)were established.The results indicate that both models possess strong predictive capabilities,with the SO-SVM model demonstrating superior optimization and generalization performance.The minimum surface roughness achieved through the SVM model optimized by the snake algo-rithm for the refined process parameters is 17.7 μm,which is lower than the 19.3 μm obtained using the response surface method.This study provides a reference for minimizing surface roughness,significantly reducing the trial-and-error costs in the machining process,and facilitating the production of higher quality machined parts.
关键词
选区激光熔化/优化支持向量机/表面粗糙度/参数优化Key words
selective laser melting/optimized support vector machine/surface roughness/parameter optimization引用本文复制引用
出版年
2024
NETL
NSTL
万方数据