Foundation pit multi-point displacement RBF monitoring model and application key points

摘要

To know deep foundation pit state and ensure its safety, multi monitoring points data are used to establish model in a quick and useful way. Considering grading excavation effect, time effect and synchronous multi-point displacement monitoring information, the frame of Radial basis function artificial neural network (RBF) multi-point monitoring model has been structured. The concise effect factors base on Genetic Creep Theory are studied to forecast multi-point displacement. After analyzing the action of RBF centers in this network, preselected centers considering the grading excavation and curve characteristic of displacement have been used in calculation. Monitoring data from a deep foundation pit are used to establish multi-point model base on the multi-point model frame, concise effect factors and preselected centers. Analyses show that the results of instance are very good. Application suggestions are also presented combining RBF characteristics and monitoring practice.