Scholarlink

<p>Cross-laminated timber (CLT) is a sustainable construction material that can store up to 50 % of its weight in biogenic carbon. However, U.S. CLT production depends heavily on imports from Europe and Canada, since locally sourced timber often contains large and frequent knots that compromise structural consistency. This reliance increases both costs and transportation-related carbon emissions. To address this, we propose a reinforcement learning (RL)-based advanced manufacturing framework that enables effective use of local timber. In a virtual design environment, AI agents optimize the arrangement of timber planks with random knot patterns using trial-and-error learning guided by structural performance feedback. This approach reduces variability in mechanical properties by strategically distributing defects across the panel layout. Compared to baseline methods using random placement, the proposed system demonstrates improved consistency and structural performance, highlighting its potential to support cost-effective, sustainable CLT production using domestic timber resources.</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

Advanced manufacturing of locally-sourced cross-laminated timber: proof-of-concept

  • 期刊论文

  • 发表日期：

    2025年11月22日

查看全文
  • SungKu Kang,
  • Michael B. Kane

摘要

Cross-laminated timber (CLT) is a sustainable construction material that can store up to 50 % of its weight in biogenic carbon. However, U.S. CLT production depends heavily on imports from Europe and Canada, since locally sourced timber often contains large and frequent knots that compromise structural consistency. This reliance increases both costs and transportation-related carbon emissions. To address this, we propose a reinforcement learning (RL)-based advanced manufacturing framework that enables effective use of local timber. In a virtual design environment, AI agents optimize the arrangement of timber planks with random knot patterns using trial-and-error learning guided by structural performance feedback. This approach reduces variability in mechanical properties by strategically distributing defects across the panel layout. Compared to baseline methods using random placement, the proposed system demonstrates improved consistency and structural performance, highlighting its potential to support cost-effective, sustainable CLT production using domestic timber resources.