摘要
通过有限元模拟软件积累足够量的多层黏土中圆形锚板抗拔承载力数据库,基于机器学习中的集成学习开展数据库的训练与预测,建立了多层黏土中圆形锚板抗拔承载力预测模型,并且验证了集成学习模型在均匀黏土和多层黏土地层条件下的承载力预测准确度。验证结果显示,集成学习模型对于复杂土层下的圆形锚板承载力预测具有较高的准确性,预测值与实测值基本吻合。结合有限元模拟结果和集成学习的参数重要性分析,发现了锚板以上4倍锚板直径内的黏土平均抗剪强度是预测其抗拔承载力的关键参数,其次是距离锚板2倍至5倍直径范围内的黏土平均抗剪强度。该集成学习模型成功解决了以往经验公式与理论公式在多层土环境中预测准确度低的问题,为后续研究提供了新思路。
Based on the machine learning model theory,a dataset,and machine learning model for predicting the uplift capacity factor of circular plate anchors in undrained gravity-free clay soils were developed using finite element software to simulate multilayer clay soil cases.The accuracy of the machine learning prediction model was analyzed,and its predictive capability was compared for uniform clay,linear clay,and multilayered clay cases.The analysis results demonstrated the high accuracy of the machine learning prediction model,with predicted values closely matching the measured value curves.The parameter importance analysis identified the average shear strength of the soil above the anchor plate as the key factor in predicting the pullout bearing capacity.Among the parameters,the average shear strength of the clay at 4 times the anchor plate diameter from the anchor plate was found to be the most influential,followed by the average shear strength of the clay at 2 times and 5 times the anchor plate diameter from the anchor plate.The machine learning prediction model successfully addressed the limitations of previous empirical formulas,which were less accurate and overly complex for complex soil conditions,and presented a novel direction for future research.
作者
黄蕴晗
叶礼强
祁昊
宗钟凌
HUANG Yunhan;YE Liqiang;QI Hao;ZONG Zhongling(School of Civil and Ocean Engineering,Jiangsu Ocean University,Lianyungang 222005,China;Lianyungang Power Supply Branch,State Grid Jiangsu Electric Power Co.,Lianyungang 222004,China)
出处
《江苏海洋大学学报(自然科学版)》
2025年第3期35-43,共9页
Journal of Jiangsu Ocean University:Natural Science Edition
基金
江苏省重点研发计划(社会发展)项目(BE2021681)。
关键词
集成学习
圆形锚板
有限元分析
抗拔承载力
ensemble learning
circular plate anchor
finite element analysis
uplift capacity factor
