Scholarlink

<p>The energy consumption dynamic characteristics of laser welding process are unclear restricts its development towards sustainable manufacturing. Considering the multi-source carbon emission feature of laser welding, achieving intelligent decision-making of the incidence relation between laser welding process parameters and carbon emissions and welding quality faces great challenges. To tackle the gap, a novel algorithm based on dung beetle optimizer (DBO)- back propagation neural network (BP) and NSGA-II is proposed for multi-objective optimization of process parameters. First, laser welding experiments are carried out through carbon emission modeling and optimal Latin hypercube sampling method. Subsequently, based on the experimental data, the hybrid DBO-BP algorithm is employed to establish nonlinear mapping relationships between laser power, welding speed, defocusing amount, carbon emissions, tensile strength, and weld integrity. Finally, the optimal process parameters are determined through the NSGA-II multi-objective optimization algorithm and entropy weight TOPSIS decision method. This optimization not only reduces carbon emissions but also enhances the quality of fiber laser welding, aligning with the objective of achieving ‘low carbon and high quality’.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Optimization of laser welding process parameters considering carbon emissions and weld quality based on DBO-BP and NSGA-II

  • 期刊论文

  • 发表日期：

    2025年2月26日

查看全文
  • Chunjin Li,
  • Shuai Li,
  • Yang Xie,
  • Jianzhao Wu,
  • Chaoyong Zhang

摘要

The energy consumption dynamic characteristics of laser welding process are unclear restricts its development towards sustainable manufacturing. Considering the multi-source carbon emission feature of laser welding, achieving intelligent decision-making of the incidence relation between laser welding process parameters and carbon emissions and welding quality faces great challenges. To tackle the gap, a novel algorithm based on dung beetle optimizer (DBO)- back propagation neural network (BP) and NSGA-II is proposed for multi-objective optimization of process parameters. First, laser welding experiments are carried out through carbon emission modeling and optimal Latin hypercube sampling method. Subsequently, based on the experimental data, the hybrid DBO-BP algorithm is employed to establish nonlinear mapping relationships between laser power, welding speed, defocusing amount, carbon emissions, tensile strength, and weld integrity. Finally, the optimal process parameters are determined through the NSGA-II multi-objective optimization algorithm and entropy weight TOPSIS decision method. This optimization not only reduces carbon emissions but also enhances the quality of fiber laser welding, aligning with the objective of achieving ‘low carbon and high quality’.