软土地基沉降是土力学中的重要课题之一,通过沉降预测可以为软土地基的工程设计以及下一步施工提供指导性意见。以自组织神经网络为基础,利用其具有的自组织、自适应、容错性和较强的学习、联想能力,结合连徐高速公路软基沉降工程实例进行预测。说明该方法能够简便、精确的预测软土地基的沉降量。 Settlement of soft soil foundation is one of the most important topics in soil mechanics. Through the prediction of settlement, it can provide some guidance for engineering design and construction of soft soil foundation. Based on the self-organizing neural network, the self-organizing, self-adaptive, fault-tolerant and strong learning and associative ability are used to predict the soft subsidence engineering of Lian-Xu Expressway. It shows that this method can predict the settlement of soft ground easily and accurately.