滇池湖相粉土的工程性质

Engineering Properties of Lacustrine Silty Soil in Dianchi Lake

滇池湖相粉土为典型的高原湖新生代沉积物,具有厚度大、规律多变、不确定性显著等特征。依托昆明巫家坝大型城市综合体项目,在现场地质勘察的基础上,利用机器学习算法和数理统计方法深入研究了滇池湖相粉土的工程性质。首先,利用K-means算法对土体样本进行二级分类,并通过计算各参数的基本统计特征,验证了上述二级分类的合理性和必要性。然后,利用Bootstrapping等数理统计方法,进一步探究了土性参数的深度趋势特性、统计分布特征以及多元相关性等,结果表明滇池湖相粉土参数的不确定性显著,参数间存在明显的交叉依赖性。研究结果将为相关岩土工程设计和风险分析提供数据参考。

The Dianchi Lake lacustrine silty soil,a typical Neogene deposit in the plateau lake,has characteristics such as large thickness,changeable laws,and significant uncertainties.Based on the field investigation in the Kunming Wujiaba large-scale urban complex project,machine learning algorithms and mathematical statistics were used to investigate the engineering properties of Dianchi Lake lacustrine silty soil in-depth.Firstly,the K-means algorithm was used to classify the soil samples into two levels,and the basic statistical characteristics of each parameter were calculated to verify the rationality and necessity of the above classification.Then,using mathematical statistics methods such as Bootstrapping,the depth trend characteristics,statistical distribution characteristics,and multivariate correlations of soil parameters were further explored.The results show that the uncertainty of the property parameters of Dianchi Lake lacustrine silty soil is significant,and there is evident cross-dependency among parameters.This study can provide data references for related geotechnical engineering design and risk analysis.