Scholarlink

<p>In this study, DN400 large-diameter PE electrofusion fittings were selected as the research object, and a theoretical model of electrofusion welding was established and experimentally validated. Based on this, the influence of welding parameters and wire distribution methods on the morphology of the fusion zone and welding performance were analyzed. A welding database was constructed using finite element simulation combined with Latin hypercube sampling (LHS). A random forest (RF) was used to classify the fusion state under different welding parameters and wire distribution methods, and A BP neural network (BPNN) was employed to predict welding performance. The particle swarm (PSO) algorithm was applied to optimize the the fusion zone morphology and welding performance. The results show that, compared with uniform wire distribution, non-uniform wire distribution has a better effect on controlling and improving the morphology and performance indicators of the fusion zone. Both the RF classification model and BPNN achieved accuracies exceeding 96% in classifying the fusion state and predicting welding performance. The PSO optimization algorithm significantly improve the morphology of the fusion zone, with all welding performance indicators improved by more than 8%.</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

Morphological analysis and intelligent optimization of the fusion zone in electrofusion welding of PE pipes

  • 期刊论文

  • 发表日期：

    2026年6月18日

查看全文
  • Tao Shen,
  • Yuan Chen,
  • Weiyu Liu,
  • Jingtai Liu,
  • Xiaogang Wang,
  • Bingqing Wang,
  • Yuntang Li

摘要

In this study, DN400 large-diameter PE electrofusion fittings were selected as the research object, and a theoretical model of electrofusion welding was established and experimentally validated. Based on this, the influence of welding parameters and wire distribution methods on the morphology of the fusion zone and welding performance were analyzed. A welding database was constructed using finite element simulation combined with Latin hypercube sampling (LHS). A random forest (RF) was used to classify the fusion state under different welding parameters and wire distribution methods, and A BP neural network (BPNN) was employed to predict welding performance. The particle swarm (PSO) algorithm was applied to optimize the the fusion zone morphology and welding performance. The results show that, compared with uniform wire distribution, non-uniform wire distribution has a better effect on controlling and improving the morphology and performance indicators of the fusion zone. Both the RF classification model and BPNN achieved accuracies exceeding 96% in classifying the fusion state and predicting welding performance. The PSO optimization algorithm significantly improve the morphology of the fusion zone, with all welding performance indicators improved by more than 8%.