Scholarlink

<p>Bentonite–sand/fly ash-based thermal backfill materials are used as a heat transfer medium between the heat sources (e.g., underground power cable) and near-field. The characterization of materials in the laboratory, as a thermal backfill material, often requires extensive field as well as laboratory work, which is quite expensive and time-consuming. Therefore, this paper aims to develop a soft computing (SC) technique to predict the thermal conductivity due to its ability to handle complex, nonlinear and uncertain relationships in soil behavior. To achieve this goal, four SC algorithms, namely artificial neural network (ANN), ANN–PSO (particle swarm optimization), adaptive neural network-based fuzzy inference system (ANFIS) and extreme gradient boosting (XGB), were utilized based on the experimental database considering compaction state and physical properties of backfill as input variables. The performance of different SC techniques was evaluated based on scatter plots, statistical indices, Taylor’s diagrams and rank analysis. The results exhibited that XGB predicts more accurately than ANFIS, ANN and ANN–PSO. Finally, the influence and significance of the input parameters on XGB model performance are highlighted using Shapley additive explanation (SHAP) analysis. The findings demonstrated that the water content, dry density and particle size content had the most significant impact on the thermal conductivity of bentonite–sand/fly ash-based backfill material.</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

Soft computing techniques for predicting thermal conductivity of bentonite–fly ash/-sand composite materials

  • 期刊论文

  • 发表日期：

    2025年8月26日

查看全文
  • Vishakha Bharti,
  • Pawan Kishor Sah,
  • Shiv Shankar Kumar,
  • Bhabani Shankar Das

摘要

Bentonite–sand/fly ash-based thermal backfill materials are used as a heat transfer medium between the heat sources (e.g., underground power cable) and near-field. The characterization of materials in the laboratory, as a thermal backfill material, often requires extensive field as well as laboratory work, which is quite expensive and time-consuming. Therefore, this paper aims to develop a soft computing (SC) technique to predict the thermal conductivity due to its ability to handle complex, nonlinear and uncertain relationships in soil behavior. To achieve this goal, four SC algorithms, namely artificial neural network (ANN), ANN–PSO (particle swarm optimization), adaptive neural network-based fuzzy inference system (ANFIS) and extreme gradient boosting (XGB), were utilized based on the experimental database considering compaction state and physical properties of backfill as input variables. The performance of different SC techniques was evaluated based on scatter plots, statistical indices, Taylor’s diagrams and rank analysis. The results exhibited that XGB predicts more accurately than ANFIS, ANN and ANN–PSO. Finally, the influence and significance of the input parameters on XGB model performance are highlighted using Shapley additive explanation (SHAP) analysis. The findings demonstrated that the water content, dry density and particle size content had the most significant impact on the thermal conductivity of bentonite–sand/fly ash-based backfill material.