首页|基于GWO-RFR的激光熔覆多道成形层形貌的预测方法

基于GWO-RFR的激光熔覆多道成形层形貌的预测方法

扫码查看
原文链接
  • NETL
  • NSTL
  • 万方数据
激光熔覆多道成形层形貌受激光熔覆过程中多个工艺参数的综合影响,为获得良好的熔覆层形貌,提出了一种基于灰狼优化(GWO)算法优化随机森林回归(RFR)算法(GWO-RFR)的激光熔覆多道成形层形貌预测方法.以 12Cr13 不锈钢为基体,Fe60 为熔覆粉末,设计试错法结合中心复合实验,测量成形层宽高比和稀释率.基于多道激光熔覆实验数据,建立激光熔覆工艺参数与成形层形貌间的GWO-RFR回归预测模型,并与RFR模型、响应面模型(RSM)的预测结果进行比较.结果表明:与RFR模型和RSM模型相比,GWO-RFR模型的预测结果和评价指标均优于RFR模型和RSM模型,GWO-RFR预测模型能够更准确地预测熔覆层形貌,更接近实际值,可为获得优异的激光熔覆多道成形层形貌提供理论依据.
Prediction method for multi-track laser cladding layer morphology based on GWO-RFR
The morphology of laser cladding multi-track forming layer is influenced by multiple process parameters in the process of laser cladding.In order to obtain a good morphology of the cladding layer,a method for predicting the morphology of laser cladding multi-track forming layer based on grey wolf optimization(GWO)algorithm optimized random forest regression(RFR)algorithm(GWO-RFR)was proposed.Using 12Cr13 stainless steel as the substrate and Fe60 as the cladding powder,a trial and error method combined with a central composite experiment was designed to measure the aspect ratio and dilution rate of the formed layer.Based on multi-track laser cladding experimental data,establish a GWO-RFR regression prediction model between laser cladding process parameters and formed layer morphology,and compared with the prediction results of RFR model and response surface model(RSM).The results show that compared with the RFR model and RSM model,the GWO-RFR model has better prediction results and evaluation indicators than the RFR model and RSM model.The GWO-RFR prediction model can more accurately predict the morphology of the cladding layer,which is closer to the actual value,and can provide a theoretical basis for obtaining excellent laser cladding multi-track forming layer morphology.

laser claddingmorphologygray wolf optimization algorithmrandom forest regression algorithm

毛恺奕、杜彦斌、何国华、彭云川、李志强

展开 >

重庆工商大学制造装备机构设计与控制重庆市重点实验室,重庆 400067

激光熔覆 形貌 灰狼优化算法 随机森林回归算法

重庆市自然科学基金联合基金重点项目重庆市高校创新研究群体资助项目重庆英才计划"包干制项目"重庆英才计划

CSTB2022NSCQ-LZX0011CXQT21024cstc2022ycjhbgzxm0056CQYC20210302226

2024

10.13289/j.issn.1009-6264.2023-0289

材料热处理学报
中国机械工程学会

材料热处理学报

CSTPCD北大核心
影响因子：0.958
ISSN：1009-6264
年,卷(期)：2024.45(2)
  • 26