Scholarlink

<p>Wire arc additive manufacturing (WAAM) technology has become a very promising alternative in many manufacturing industries due to the possibility to build large-size components at low cost. With the rise of industry 4.0, availability of massive data and computational resources particularly the machine learning (ML) algorithms boosted the necessity to integrate the ML and WAAM technology to develop an intelligent WAAM system which has potential to adjust the errors in real time and enhance the product quality, process stability and productivity. Recent WAAM literature shows that nearly 80% of studies now use ML techniques, with deep-learning methods contributing approx. 45% and classical ML ~ 40%. Vision-based sensors dominate (~ 55%), followed by acoustic (~ 20%), electrical (~ 15%), and thermal (~ 10%) systems. Therefore, this paper aims to provide an exhaustive review of current developments in WAAM technique with particular focus on process monitoring, and control with the application of sensor systems and ML techniques. The paper concludes by proposing a framework for multiple sensor-based monitoring and control system for the gas metal arc welding (GMAW) based WAAM process. This proposed framework provides a blueprint for the monitoring and control which aims to control the geometrical aspects of deposited materials and identify/reduce defects with the help of sensor data fusion techniques and ML algorithms.</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

Applications of machine learning in monitoring and controlling the wire arc additive manufacturing: a review

  • 期刊论文

  • 发表日期：

    2026年2月5日

查看全文
  • Pratishtha Sharma,
  • S. Aravindan,
  • Kusum Meena

摘要

Wire arc additive manufacturing (WAAM) technology has become a very promising alternative in many manufacturing industries due to the possibility to build large-size components at low cost. With the rise of industry 4.0, availability of massive data and computational resources particularly the machine learning (ML) algorithms boosted the necessity to integrate the ML and WAAM technology to develop an intelligent WAAM system which has potential to adjust the errors in real time and enhance the product quality, process stability and productivity. Recent WAAM literature shows that nearly 80% of studies now use ML techniques, with deep-learning methods contributing approx. 45% and classical ML ~ 40%. Vision-based sensors dominate (~ 55%), followed by acoustic (~ 20%), electrical (~ 15%), and thermal (~ 10%) systems. Therefore, this paper aims to provide an exhaustive review of current developments in WAAM technique with particular focus on process monitoring, and control with the application of sensor systems and ML techniques. The paper concludes by proposing a framework for multiple sensor-based monitoring and control system for the gas metal arc welding (GMAW) based WAAM process. This proposed framework provides a blueprint for the monitoring and control which aims to control the geometrical aspects of deposited materials and identify/reduce defects with the help of sensor data fusion techniques and ML algorithms.