Decarbonizing precast concrete building components: Cradle-to-site carbon modeling and optimization, explainable machine learning, and a transportation efficiency index
プレキャストコンクリート建築部材の脱炭素化：クレードルからサイトへの炭素モデリングと最適化、説明可能な機械学習、および輸送効率指数 (AI 翻訳)
Peyman Naghipour, Afshin Naghipour, Tarana Bakirova, Hussein Ghiyasi, Faraneh Soltani Gerd Faramarzi, Farazin Soltani Gerd Faramarzi
🤖 gxceed AI 要約
日本語
プレキャストコンクリート部材の生産・輸送・設置における温室効果ガス排出を定量化・予測する枠組みを提案。機械学習と説明可能AIを用い、部品の標準化と物流計画の統合により、工場から現場までの排出を20～30%削減可能と示す。テヘランの大規模プレキャスト工場の実データに基づく実証研究。
English
This paper presents a cradle-to-site carbon modeling framework for precast concrete components, integrating machine learning and explainable AI to predict emissions from production, transportation, and installation. Using data from a large precast plant in Tehran, the study shows that joint optimization of design standardization and logistics can reduce combined emissions by 20-30% and improve a newly proposed Transportation Efficiency Index by ~25%.
Unofficial AI-generated summary based on the public title and abstract. Not an official translation.
📝 gxceed 編集解説 — Why this matters
日本のGX文脈において
日本の建設業界でもプレキャストコンクリートの利用が進んでおり、本論文の枠組みはサプライチェーン全体の排出削減に貢献し得る。設計標準化と物流最適化の統合は、日本の工場生産型住宅などにも応用可能だが、日本への適用には国内データによる検証が求められる。
In the global GX context
This paper provides a replicable methodology for embodied carbon reduction in prefabricated construction, relevant to global efforts to decarbonize the built environment. The use of explainable AI to identify key drivers offers transparent guidance for supply chain optimization, and the Transportation Efficiency Index could be adopted in emission reporting frameworks like the GHG Protocol.
👥 読者別の含意
🔬研究者:Novel ML-XAI framework for supply chain emission modeling and optimization, with empirical validation.
🏢実務担当者:Actionable insights for precast manufacturers to reduce emissions through design standardization and logistics coordination.
🏛政策担当者:Offers a metric (TEI) and evidence-based targets for low-carbon prefabrication supply chains.
📄 Abstract（原文）
Reducing carbon in prefabricated buildings demands component-scale evidence, yet most assessments remain confined to factory production and provide limited, non-transparent guidance on how transportation and on-site installation decisions reshape emissions. This study delivers a consistent framework for quantifying and predicting emissions from the production, transportation, and installation of precast concrete components. It explores the concept that integrating coordinated design standards with logistical planning leads to considerable reductions in cradle-to-site emissions. The framework contributes: (i) a tri-stage system boundary; (ii) a machine-learning plus explainable-AI (XAI) model for transport coupled with a new Transportation Efficiency Index (TEI), defined as delivered component volume-distance per unit CO2e; and (iii) joint optimization of design standardization and logistics parameters. Empirical data were obtained from a prefabrication plant in Tehran, Iran (156,000 m2 footprint; 300,000 m3·yr−1 capacity), including 411 daily energy/resource records, bills of materials and mold-use logs, 408 manufactured components, and matched delivery/installation activities. Gradient-boosted trees yield high predictive accuracy (coefficient of determination R2 = 0.99 for production and R2 = 0.97 for transportation; mean absolute percentage error MAPE < 6%), while XAI identifies component volume, design standardization, route distance, and truck utilization as dominant drivers; materials account for ~91–98% of production emissions and mold amortization falls from ~9% to <3% when standardization exceeds 0.90 and reuse surpasses ~60 cycles. Scenario optimization improves TEI by ~25% and reduces combined production-to-installation emissions by ~20–30%, providing actionable guidance for manufacturers, contractors, and policymakers seeking low-carbon prefabrication supply chains.
🔗 Provenance — このレコードを発見したソース
- semanticscholar https://doi.org/10.59400/be4056first seen 2026-05-05 22:30:14
gxceed は公開メタデータに基づく研究支援データセットです。要約・翻訳・解説は AI 支援で生成されています。 最終的な解釈・検証は利用者が原典資料に基づいて行うことを前提とします。