Scholarlink

<p>Incorporating phase change materials into concrete may increase thermal conductivity but also decrease mechanical properties and potentially cause leakage problems. This research proposes a framework for predicting the macroscopic properties of sustainable concrete using phase change material aggregates and additional cementitious materials, such as fly ash and rice husk ash, through a data-driven approach. PCM aggregate are fabricated using two-stage encapsulation with epoxy-resin and cementitious coating for thermal stability and compatibility. Concrete mixes with different PCM replacement levels and partial cement substitution using RHA and FA are prepared to balance thermal and mechanical performance. The Deep Adaptive Perceptual Generative Adversarial Network (DAPGAN) is used to predict the mechanical and thermal properties based on mix design parameters, demonstrating high predictive accuracy within RHA-FA-PCM blend domain. The results show that incorporating RHA-FA with PCM aggregate improves thermal performance while reducing strength loss caused by PCM. The proposed model achieves higher prediction accuracy and robustness than traditional models, demonstrating the effectiveness of deep learning based PCM encapsulation in designing energy-efficient and sustainable concrete for building envelope applications. With a low RMSE of 1.14&#xa0;MPa and compressive strength prediction R2 of 0.982, the results show that DAPGAN performs better than conventional models. According to experimental results, synergistic use of 15% RHA and 20% FA mitigated strength loss by 12.5% while maintaining structural capacity of 32.5&#xa0;MPa, despite the fact that high PCM replacement levels generally reduce strength. Additionally, the addition of cocktail PCMs led to a notable increase in volumetric heat capacity and 42% decrease in thermal conductivity (to 0.88W/mK).</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

Sustainable Concrete Incorporating Phase Change Material and Ash Blends using Deep Learning for Performance Prediction

  • 期刊论文

  • 发表日期：

    2026年5月30日

查看全文
  • T. Santhoshkumar,
  • R. Ilangovan

摘要

Incorporating phase change materials into concrete may increase thermal conductivity but also decrease mechanical properties and potentially cause leakage problems. This research proposes a framework for predicting the macroscopic properties of sustainable concrete using phase change material aggregates and additional cementitious materials, such as fly ash and rice husk ash, through a data-driven approach. PCM aggregate are fabricated using two-stage encapsulation with epoxy-resin and cementitious coating for thermal stability and compatibility. Concrete mixes with different PCM replacement levels and partial cement substitution using RHA and FA are prepared to balance thermal and mechanical performance. The Deep Adaptive Perceptual Generative Adversarial Network (DAPGAN) is used to predict the mechanical and thermal properties based on mix design parameters, demonstrating high predictive accuracy within RHA-FA-PCM blend domain. The results show that incorporating RHA-FA with PCM aggregate improves thermal performance while reducing strength loss caused by PCM. The proposed model achieves higher prediction accuracy and robustness than traditional models, demonstrating the effectiveness of deep learning based PCM encapsulation in designing energy-efficient and sustainable concrete for building envelope applications. With a low RMSE of 1.14 MPa and compressive strength prediction R2 of 0.982, the results show that DAPGAN performs better than conventional models. According to experimental results, synergistic use of 15% RHA and 20% FA mitigated strength loss by 12.5% while maintaining structural capacity of 32.5 MPa, despite the fact that high PCM replacement levels generally reduce strength. Additionally, the addition of cocktail PCMs led to a notable increase in volumetric heat capacity and 42% decrease in thermal conductivity (to 0.88W/mK).