Scholarlink

<p>Additive manufacturing (AM) systems require high precision and stable thermal–mechanical control, yet remain vulnerable to machine degradation, process instability, and defect formation that can cause costly downtime and rejected builds. Conventional reactive and preventive maintenance strategies are often insufficient for AM because they do not respond to real-time process variability or machine-condition changes. This narrative-conceptual review synthesizes literature on predictive maintenance (PdM), artificial intelligence (AI), multisensor monitoring, and Lean Six Sigma (LSS) to address a specific gap: the absence of a structured decision architecture that links AI-derived maintenance insights with LSS-guided process improvement in AM. The principal contribution of this study is an AM-specific AI-LSS-PdM framework that connects multimodal sensor data acquisition, data preprocessing, AI-based analytics, and DMAIC-based governance within a closed-loop workflow for proactive failure anticipation and process stabilization. The review examines condition-based and statistical-based PdM, AI methods including machine learning, deep learning, reinforcement learning, and hybrid models, AM data sources such as thermal, acoustic, optical, vibration, and machine-log signals, and selected literature examples illustrating current implementation patterns. Emerging directions including digital twins, federated learning, explainable AI, and Industry 5.0 are also discussed. Rather than claiming empirical validation, the proposed framework provides a literature-grounded structure for organizing future research and implementation of AI-enabled PdM within AM systems.</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

Predictive maintenance for additive manufacturing through integrated lean six sigma and artificial intelligence approaches

  • 期刊论文

  • 发表日期：

    2026年6月19日

查看全文
  • Hitesh Kishnani,
  • Xiwei Audrey Wang,
  • Shalahudin Muhammad,
  • Samuel Hee Yi,
  • Minahil Khurram,
  • Gema Jimenez Garnica,
  • Kalana Abeywardena,
  • Utkarsh Chadha,
  • Divyansh Srivastava,
  • Swapnil Kumar

摘要

Additive manufacturing (AM) systems require high precision and stable thermal–mechanical control, yet remain vulnerable to machine degradation, process instability, and defect formation that can cause costly downtime and rejected builds. Conventional reactive and preventive maintenance strategies are often insufficient for AM because they do not respond to real-time process variability or machine-condition changes. This narrative-conceptual review synthesizes literature on predictive maintenance (PdM), artificial intelligence (AI), multisensor monitoring, and Lean Six Sigma (LSS) to address a specific gap: the absence of a structured decision architecture that links AI-derived maintenance insights with LSS-guided process improvement in AM. The principal contribution of this study is an AM-specific AI-LSS-PdM framework that connects multimodal sensor data acquisition, data preprocessing, AI-based analytics, and DMAIC-based governance within a closed-loop workflow for proactive failure anticipation and process stabilization. The review examines condition-based and statistical-based PdM, AI methods including machine learning, deep learning, reinforcement learning, and hybrid models, AM data sources such as thermal, acoustic, optical, vibration, and machine-log signals, and selected literature examples illustrating current implementation patterns. Emerging directions including digital twins, federated learning, explainable AI, and Industry 5.0 are also discussed. Rather than claiming empirical validation, the proposed framework provides a literature-grounded structure for organizing future research and implementation of AI-enabled PdM within AM systems.