Scholarlink

<p>Such emerging demands for an environment-friendly form of construction materials led to the development of the so-called alternative construction material-Geopolymer Concrete (GPC), to overcome the current limitation of its conventional counterpart-the Portland cement concrete. However, such a developed mix design, coupled with vulnerability towards environmental aspects (including exposure to sulphates and chlorides), requires much more advanced techniques for predicting such performances. The proposed research introduces an Integrated Predictive Framework (IPF) to optimize GPC mix design. The combined power of Linear Regression (LR), Random Forest (RF), Gradient Boosting Machines (GBM), and Deep Neural Networks (DNNs) is exploited for enhancing the accuracy of predictions made on the mechanical properties of GPC, specifically on its resistance to sulphate and chloride attacks. Further model performance improvement is achieved using a new hybrid technique called Predator-Annealing Optimization (PAO), which will be used to optimize the hyperparameters. In PAO, the Marine Predators Algorithm (MPA) and Simulated Annealing Optimization (SAO) are fused to effectively navigate the search space and improve the model’s accuracy. Experimental results show that PAO-based IPF performs excellently and surpasses traditional models. Its R<sup>2</sup> value is about 0.98, while the Root Mean Square Error (RMSE) is about 0.015, and the Mean Absolute Error (MAE) is about 0.01, showing strong predictive accuracy and robustness. This work emphasizes the efficiency of integrating machine learning (ML) and hybrid optimization methods in predicting and optimizing GPC performance. It also represents a practical tool for designing sustainable and durable mixes to construct structures in sulphate and chloride environments.</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

Optimization of geopolymer concrete mix design using machine learning for enhanced sulphate and chloride resistance

  • 期刊论文

  • 发表日期：

    2025年8月12日

查看全文
  • Akshay Dhawan,
  • Manvendra Verma

摘要

Such emerging demands for an environment-friendly form of construction materials led to the development of the so-called alternative construction material-Geopolymer Concrete (GPC), to overcome the current limitation of its conventional counterpart-the Portland cement concrete. However, such a developed mix design, coupled with vulnerability towards environmental aspects (including exposure to sulphates and chlorides), requires much more advanced techniques for predicting such performances. The proposed research introduces an Integrated Predictive Framework (IPF) to optimize GPC mix design. The combined power of Linear Regression (LR), Random Forest (RF), Gradient Boosting Machines (GBM), and Deep Neural Networks (DNNs) is exploited for enhancing the accuracy of predictions made on the mechanical properties of GPC, specifically on its resistance to sulphate and chloride attacks. Further model performance improvement is achieved using a new hybrid technique called Predator-Annealing Optimization (PAO), which will be used to optimize the hyperparameters. In PAO, the Marine Predators Algorithm (MPA) and Simulated Annealing Optimization (SAO) are fused to effectively navigate the search space and improve the model’s accuracy. Experimental results show that PAO-based IPF performs excellently and surpasses traditional models. Its R2 value is about 0.98, while the Root Mean Square Error (RMSE) is about 0.015, and the Mean Absolute Error (MAE) is about 0.01, showing strong predictive accuracy and robustness. This work emphasizes the efficiency of integrating machine learning (ML) and hybrid optimization methods in predicting and optimizing GPC performance. It also represents a practical tool for designing sustainable and durable mixes to construct structures in sulphate and chloride environments.