Scholarlink

<p>Ultra-high performance concrete (UHPC) exhibits exceptional mechanical properties and durability, but its adoption is limited by high material costs and significant carbon emissions. This study presents a novel, integrated machine learning (ML) and life-cycle assessment (LCA) framework for the multi-objective optimization of UHPC mix designs that balance strength, workability, cost, and environmental impact. A comprehensive dataset of 514 UHPC mixes, incorporating 13 key mix variables, was used to train and evaluate multiple ML models. The XGBoost Regressor demonstrated superior performance (R² = 0.951) for predicting compressive strength and slump flow. A detailed life-cycle cost and carbon emission model was constructed using real-world pricing, transport, and embodied carbon data. Optimization was performed using NSGA-III and AGE-MOEA algorithms, and optimal trade-off solutions were ranked using the TOPSIS method. To enhance transparency, SHAP analysis was applied to interpret model predictions and quantify the influence of each variable. Selected Pareto-optimal mixes were validated experimentally, showing an average prediction error of less than 2%. The proposed framework offers a scalable, data-driven alternative to traditional empirical methods, enabling faster, more sustainable UHPC design with built-in decision-support tools. This work demonstrates the potential of ML–LCA integration for intelligent, performance-based, and environmentally responsible concrete mix design.</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

Integrated machine learning and life-cycle optimization framework for sustainable ultra-high performance concrete mix design

  • 期刊论文

  • 发表日期：

    2025年9月25日

查看全文
  • Amit Gautam,
  • Viswanadham Sangeeta,
  • Jagannath Padhy,
  • Pooja,
  • A Ganpathi Rao

摘要

Ultra-high performance concrete (UHPC) exhibits exceptional mechanical properties and durability, but its adoption is limited by high material costs and significant carbon emissions. This study presents a novel, integrated machine learning (ML) and life-cycle assessment (LCA) framework for the multi-objective optimization of UHPC mix designs that balance strength, workability, cost, and environmental impact. A comprehensive dataset of 514 UHPC mixes, incorporating 13 key mix variables, was used to train and evaluate multiple ML models. The XGBoost Regressor demonstrated superior performance (R² = 0.951) for predicting compressive strength and slump flow. A detailed life-cycle cost and carbon emission model was constructed using real-world pricing, transport, and embodied carbon data. Optimization was performed using NSGA-III and AGE-MOEA algorithms, and optimal trade-off solutions were ranked using the TOPSIS method. To enhance transparency, SHAP analysis was applied to interpret model predictions and quantify the influence of each variable. Selected Pareto-optimal mixes were validated experimentally, showing an average prediction error of less than 2%. The proposed framework offers a scalable, data-driven alternative to traditional empirical methods, enabling faster, more sustainable UHPC design with built-in decision-support tools. This work demonstrates the potential of ML–LCA integration for intelligent, performance-based, and environmentally responsible concrete mix design.