Scholarlink

<p>A parametric study was conducted to investigate the bearing capacity of strip footings on anisotropic soft clay reinforced with granular trenches using finite element limit analysis. The analysis covered 3600 configurations by varying trench width ratio, embedment depth ratio, internal friction angle, normalized undrained shear strength, and anisotropic strength ratio. The results indicated that geometric parameters, especially trench width and embedment depth, strongly influenced the efficiency of the improvement, with deeper trenches providing greater capacity enhancement in weak soils. Failure mechanism analyses revealed that optimized trench geometries mobilized wider passive resistance zones and produced deeper critical failure surfaces, thereby improving load transfer mechanisms. A deep neural network integrated with a novel metaheuristic optimization algorithm was employed for predictive modeling. The trained model achieved high accuracy, with determination coefficients close to unity for both training and testing datasets. Evaluation across multiple performance metrics confirmed the robustness of the predictive framework. External validation further demonstrated its reliability, satisfying established criteria. An interpretability analysis using Shapley values highlighted trench width ratio and embedment depth ratio as the most influential parameters controlling bearing capacity improvement. This study provided a robust methodology for designing shallow foundations with granular reinforcement in anisotropic clay.</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

Bearing Capacity Prediction of Strip Footings on Anisotropic Clays with Granular Trenches Using Optimized Deep Neural Network

  • 期刊论文

  • 发表日期：

    2025年10月28日

查看全文
  • Wittaya Jitchaijaroen,
  • Nitchapat Pinnatsakda,
  • Suraparb Keawsawasvong,
  • Pitthaya Jamsawang

摘要

A parametric study was conducted to investigate the bearing capacity of strip footings on anisotropic soft clay reinforced with granular trenches using finite element limit analysis. The analysis covered 3600 configurations by varying trench width ratio, embedment depth ratio, internal friction angle, normalized undrained shear strength, and anisotropic strength ratio. The results indicated that geometric parameters, especially trench width and embedment depth, strongly influenced the efficiency of the improvement, with deeper trenches providing greater capacity enhancement in weak soils. Failure mechanism analyses revealed that optimized trench geometries mobilized wider passive resistance zones and produced deeper critical failure surfaces, thereby improving load transfer mechanisms. A deep neural network integrated with a novel metaheuristic optimization algorithm was employed for predictive modeling. The trained model achieved high accuracy, with determination coefficients close to unity for both training and testing datasets. Evaluation across multiple performance metrics confirmed the robustness of the predictive framework. External validation further demonstrated its reliability, satisfying established criteria. An interpretability analysis using Shapley values highlighted trench width ratio and embedment depth ratio as the most influential parameters controlling bearing capacity improvement. This study provided a robust methodology for designing shallow foundations with granular reinforcement in anisotropic clay.