ベイズ最適化に基づく低炭素建材配合・プロセスパラメータの新規高効率探索アルゴリズム
A novel high-efficiency search algorithm for low-carbon building material formulations and process parameters based on Bayesian optimization (原題)
Li Bin
🤖 gxceed AI 要約
日本語
ベイズ最適化とガウス過程サロゲートモデルを組み合わせ、低炭素建材の配合とプロセスパラメータを統合的に最適化するアルゴリズムを提案。アルカリ活性化断熱材のケーススタディで、従来のポルトランドセメント系に対し内包炭素を36%削減しつつ基本物性を維持した。活性化剤濃度と養生条件が性能に支配的であることを感度分析で示し、実験負担の大幅削減と計算効率の高さを実証した。
English
This paper integrates Bayesian optimization with Gaussian Process surrogate modeling to jointly optimize mixture formulations and process parameters for low-carbon building materials, embedding experimental and regulatory constraints. A case study on alkali-activated insulation materials achieved a 36% reduction in embodied carbon versus traditional Portland cement while maintaining mechanical and thermal properties. Sensitivity analysis identifies activator concentration and curing conditions as dominant factors, and complexity assessments confirm computational efficiency and robustness.
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
建材の内包炭素削減はScope 3上流（購入資材）排出量の算定・削減に直結し、建設業のSSBJ開示やサプライチェーン脱炭素目標の達成手段として日本企業にも実務的示唆がある。
In the global GX context
Embodied-carbon reduction in construction materials feeds directly into Scope 3 Category 1 purchased-goods accounting under GHG Protocol and ISSB/CSRD disclosure, offering a data-driven pathway for construction-sector transition planning.
👥 読者別の含意
🔬研究者:ベイズ最適化を材料設計の高次元非線形空間に適用する手法と制約埋め込みの実装が参考になる。
🏢実務担当者:建材調達における内包炭素削減とScope 3上流排出量の低減策として活用可能。
🏛政策担当者:建設分野の低炭素資材普及に向けた技術的裏付けとして参照できる。
📄 Abstract（原文）
Bayesian optimization, combined with Gaussian Process surrogate modeling, offers a powerful solution to the challenges of low-carbon construction material design. This paper presents an integrated algorithmic approach that leverages these advanced machine learning techniques to systematically optimize both mixture formulations and process parameters for environmentally sustainable building materials. By embedding experimental and regulatory constraints into the optimization loop, the method ensures that the selected solutions are computationally valid and practically feasible. Thru a comprehensive case study on alkali-activated insulation materials, the algorithm was found to be able to navigate in a high-dimensional nonlinear design space while significantly reducing the experimental burden. The proposed formulation reduces the embodied carbon by 36% compared to a traditional Portland cement system, while maintaining the basic mechanical and thermal properties. Sensitivity analysis shows the dominant influence of activator concentration and curing conditions on performance indicators. This provides guidance for targeted material development. Comparative time complexity assessment validates the computational efficiency and robustness under parameter perturbation and simulated noise. The results show that this machine learning-driven approach can accelerate the development of sustainable materials and achieve repeatable and scalable optimization. These findings lay the technical groundwork for the wider adoption of the technology in the construction industry.
🔗 Provenance — このレコードを発見したソース
- openalex https://doi.org/10.1117/12.3124330first seen 2026-11-07 04:54:43
🔔 こうした論文の新着を逃したくない方は キーワードアラート に登録（無料・3キーワードまで）。
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。