Scholarlink

<p>The rapid diffusion of Additive Manufacturing (AM) within the Industry 4.0 paradigm has introduced new challenges related to process reliability, energy efficiency, and sustainable production planning. While Artificial Intelligence (AI) has increasingly been adopted to enhance AM process optimization, existing approaches often address quality prediction and energy analysis separately, limiting the ability of operators to perform informed trade-off decisions between product integrity and sustainability. To address this gap, this study proposes a <i>Dual-module Machine Learning Framework</i> that integrates explainable supervised Machine Learning (ML) techniques for simultaneous quality risk assessment and energy-aware process optimization in AM. To support the proposed framework, this study introduces the AM Quality-Energy Dataset (AM-QED), a novel experimental dataset designed to systematically investigate the relationship between key printing parameters, structural quality risk, and energy consumption in Material Extrusion AM. Decision-tree-based classification models are first employed to evaluate combinations of printing parameters and classify them into two quality categories (<i>SAFE or RISKY</i>), providing interpretable predictions and associated confidence levels to support transparent decision-making. In parallel, supervised regression models estimate the energy consumption associated with each parameter configuration, enabling the identification of settings that balance manufacturing reliability and energy demand. Furthermore, a comprehensive statistical analysis is performed to quantify the influence of key process parameters on both quality risk and energy consumption. The proposed framework enables operators to proactively evaluate parameter configurations in terms of both structural reliability and sustainability objectives. Overall, the results demonstrate the potential of integrating explainable ML within AM workflows to improve process robustness, reduce energy inefficiencies and support sustainable production strategies. This work contributes to the growing body of research at the intersection of AI and AM by providing an interpretable and energy-aware approach for intelligent and sustainable manufacturing.</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

A dual-module machine learning framework for joint quality risk assessment and energy-aware process optimization in additive manufacturing

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月14日

查看全文
  • Fabrizia Devito,
  • Vincenzo Gattulli,
  • Donato Impedovo,
  • Luca Musti

摘要

The rapid diffusion of Additive Manufacturing (AM) within the Industry 4.0 paradigm has introduced new challenges related to process reliability, energy efficiency, and sustainable production planning. While Artificial Intelligence (AI) has increasingly been adopted to enhance AM process optimization, existing approaches often address quality prediction and energy analysis separately, limiting the ability of operators to perform informed trade-off decisions between product integrity and sustainability. To address this gap, this study proposes a Dual-module Machine Learning Framework that integrates explainable supervised Machine Learning (ML) techniques for simultaneous quality risk assessment and energy-aware process optimization in AM. To support the proposed framework, this study introduces the AM Quality-Energy Dataset (AM-QED), a novel experimental dataset designed to systematically investigate the relationship between key printing parameters, structural quality risk, and energy consumption in Material Extrusion AM. Decision-tree-based classification models are first employed to evaluate combinations of printing parameters and classify them into two quality categories (SAFE or RISKY), providing interpretable predictions and associated confidence levels to support transparent decision-making. In parallel, supervised regression models estimate the energy consumption associated with each parameter configuration, enabling the identification of settings that balance manufacturing reliability and energy demand. Furthermore, a comprehensive statistical analysis is performed to quantify the influence of key process parameters on both quality risk and energy consumption. The proposed framework enables operators to proactively evaluate parameter configurations in terms of both structural reliability and sustainability objectives. Overall, the results demonstrate the potential of integrating explainable ML within AM workflows to improve process robustness, reduce energy inefficiencies and support sustainable production strategies. This work contributes to the growing body of research at the intersection of AI and AM by providing an interpretable and energy-aware approach for intelligent and sustainable manufacturing.