当前位置：首页 > 过刊浏览->2025年46卷第1期
基于机器学习与多目标优化算法的 电弧增材制造过程优化研究
Parameter Optimization of Wire Arc Additive Manufacturing via Machine Learning and a Multiobjective Optimization Algorithm
浏览(87) 下载(3)
- DOI:
- 作者:
- 刘少杰1，2，彭逸琦1，2，赵宇凡1，2，杨海欧 1，2，林 鑫1，2
LIU Shaojie1,2, PENG Yiqi1,2, ZHAO Yufan1,2, YANG Haiou1,2, LIN Xin1,2
- 作者单位:
- 1. 西北工业大学凝固技术国家重点实验室，陕西西安710072；2.西北工业大学金属高性能增材制造与创新设计工业 和信息化部重点实验室，陕西西安710072
1. State Key Laboratory of Solidification Processing, Northwestern Polytechnical University, Xi'an 710072, China; 2. MIIT Key Laboratory of Metal High Performance Additive Manufacturing and Innovative Design, Northwestern Polytechnical University, Xi'an 710072, China
- 关键词:
- 电弧增材制造；机器学习；多目标优化；2219铝合金
wire arc additive manufacturing; machine learning; multiobjective optimization; 2219 aluminium alloy
- 摘要:
- 在电弧增材制造中，工艺参数之间存在复杂相互作用，难以寻找最优参数组合以获得最佳的成形质量与 预期的几何结构。 为了加速工艺参数优化过程，在3因素3水平全因素实验的基础上，明确了电流、送丝速度和扫描速 度对熔道熔宽、熔高和稀释率的影响规律，建立了神经网络、支持向量机和高斯回归分析模型来预测熔宽、熔高和稀释 率。 对比分析表明，高斯过程回归模型对熔宽的预测性能最好，神经网络模型对熔高的预测性能最好，支持向量回归对 稀释率的预测性能最好。 基于这3种机器学习模型，采用多目标遗传算法(NSGA-II)，实现了以熔宽和熔高最大、稀释率 最小为目标的电弧增材制造多目标优化，最后对优化结果进行了实验验证。In wire arc additive manufacturing (WAAM), the intricate interactions among process parameters complicate the task of finding the best settings for manufacturing metal components with superior molding quality and desired geometries. To expedite parameter optimization, this study investigates the effects of the welding current, wire-feed speed (WFS), and travel speed (TS) on the melt width (W), melt height (H), and dilution ratio (D) via 3-factor and 3-level full-factor tests. Artificial neural network (ANN), support vector regression (SVR), and Gaussian process regression (GPR) models are developed to predict these metrics. Comparative analysis indicates that GPR is most effective for predicting the melt width, ANN excels in predicting the melt height, and SVR is superior for assessing the dilution ratio. Multiobjective optimization, which uses the nondominated sorting genetic algorithm-II (NSGA-II), maximizes the melt width and height while minimizing the dilution ratio. The optimized parameters were validated experimentally, confirming the accuracy and effectiveness of the approach.