This study innovatively combines Isogeometric Analysis (IGA) with Machine Learning (ML) to assess strip footing bearing capacity on dual clayey layers. Overcoming limitations of conventional methods with small sample sizes, our research generates a dataset of 10,000 samples, allowing a thorough exploration of diverse soil profiles. Facilitated by ML, 10,000 IGA analyses using upper bound limit analysis unveil intricate patterns and relationships previously obscured. The key innovation lies in harnessing big data and employing advanced data visualization, particularly 2D and 3D Partial Dependency Plots (PDPs). These PDPs visually showcase the impact of factors such as upper layer thickness, cohesion ratios, shear strength profiles, footing depth, and foundation roughness on bearing capacity. Offering intuitive insights, these visualization tools enhance comprehension, aiding informed decision-making in design and construction. Engineers and geotechnical experts receive a precise predictive tool, optimizing strip footing performance on clayey soil layers. Moreover, this research contributes to advancing geotechnical engineering by enriching fundamental knowledge of load-bearing characteristics. In summary, the fusion of big data, advanced visualization, and upper bound limit analysis, exemplified by PDPs, signifies a substantial leap in geotechnical engineering, impacting design, construction, and infrastructure development.

中文翻译：

---

等几何分析和数据驱动建模的协同集成，用于增强两层粘土上的条形基础设计：推进岩土工程实践


这项研究创新地将等几何分析 (IGA) 与机器学习 (ML) 相结合，以评估双粘土层上的条形基础承载力。我们的研究克服了小样本量传统方法的局限性，生成了包含 10,000 个样本的数据集，可以对不同的土壤剖面进行彻底的探索。在 ML 的推动下，使用上限分析进行的 10,000 次 IGA 分析揭示了以前模糊的复杂模式和关系。关键的创新在于利用大数据和采用先进的数据可视化，特别是 2D 和 3D 部分依赖图 (PDP)。这些 PDP 直观地展示了上层厚度、粘聚比、抗剪强度分布、基础深度和地基粗糙度等因素对承载能力的影响。这些可视化工具提供直观的见解，增强理解力，有助于在设计和施工中做出明智的决策。工程师和岩土专家获得了精确的预测工具，优化了粘土层上的条形基础性能。此外，这项研究通过丰富承载特性的基础知识，有助于推进岩土工程。总之，以 PDP 为代表的大数据、高级可视化和上限分析的融合标志着岩土工程的重大飞跃，影响着设计、施工和基础设施开发。