软土地基堆载预压沉降特性及智能化预测研究  

Study on Settlement Characteristics and Intelligent Forecasting with Surcharge Preloading in Soft Soil Foundation

作　　者：余景良[1] 吕言新[2] 李海波[2] 

机构地区：[1]广州航海学院航务工程系,广东广州510725 [2]中国科学院武汉岩土力学研究所岩土力学与工程国家重点实验室,湖北武汉430071

出　　处：《施工技术》2014年第11期112-114,共3页Construction Technology

基　　金：国家973计划项目(2010CB732001)

摘　　要：针对软土地基含水量高、压缩性高及抗剪强度低的特点,采用理论分析与数值模拟相结合的方法,从力学机理探讨了堆载效应与软土地基沉降的关系。软件FLAC3D数值模拟表明,随着地基深度的增加,地基的竖向沉降不断减少,并且其递减幅度随深度的增加而增大,最大水平向位移发生在堆载物边缘以下的深部地基土中;随着水平距离的增加,地基竖向沉降也不断减少,并且其减少的幅度较均匀,当水平距离增加到一定值,其沉降值维持在某一个恒定值。以现场监测资料为基础构建动态模糊神经网络,获得了准确度和可靠度较高的预测结果,表明动态模糊神经网络具有较强的网络稳定性和泛化能力。Based on high water content, high compressibility and low shear strength characteristics in soft soil foundation, combined with theoretical analysis and numerical simulation, this paper introduces the relationship between surcharge effect and settlement in mechanical mechanism respect in soft soil foundation. Software FLAC3D numerical simulation shows that the vertical foundation settlement decreases gradually with the foundation depth increased, and the decreasing rate increases with depth increased. The maximum horizontal displacement occurs in deep foundation soil under the edge of the surcharge object. The vertical settlement also decreases with the horizontal distance increased, and the decreasing rate is relatively uniform, a constant settlement is achieved at certain horizontal distance. Based on field monitoring data,the dynamic fuzzy neural network （DFNN） is constructed, high accuracy and reliable prediction results are achieved. The prediction results show that the dynamic fuzzy neural network is a stable network with high prediction ability.

关 键 词：地基 堆载预压 沉降 软土 预测 数值模拟 

分 类 号：TU472.33[建筑科学—结构工程]