Foundation Pit Deformation Research Of Dalian Research Institute Of First Heavy Industries Group Co., Ltd R&D Building

The condition of deep foundation pit itself is complex and its stability is influenced by various factors, which is prone to associated with surrounding environment. In the excavation process, we not only need to guarantee the stability of the pit, but also to ensure the safety of the foundation pit surrounding environment. Deformation monitoring is the basic means of ensuring the safe operation of the foundation pit and surrounding environment, through analyzing of deformation monitoring data, we master the deformation regular of foundation pit, and make predictions of foundation pit deformation, in order to found that changes in a timely manner to guard against.This paper mainly studies the BP neural network theory, by means of using L-M algorithm to optimize traditional SDBP algorithm, wavelet analysis combined with BP neural network theory, two BP neural network optimization model are established. Then we study the stability of Foundation Pit Deformation Research of Dalian Research Institute of First Heavy Industries Group Co., Ltd R&D Building. The main research contents and results are as following in this paper:(1) For deformation of foundation pit engineering characteristics, influencing factors and mechanism were analyzed, and do the research of foundation pit engineering monitoring technology. We illuminate the foundation pit monitoring project, monitoring methods, in order to accumulate experience in deformation monitoring.(2) Through the design of input layer and output layer, the selected method of the nodes of hidden layer and the selection of network function, we make a design for the structure of foundation pit deformation prediction model.(3) Combined with a foundation pit monitoring instance that Foundation Pit engineering monitoring of Dalian Research Institute of First Heavy Industries Group Co., Ltd R&D Building, the horizontal displacement of pile top, internal force on anchor cable, and the surface subsidence are selected as the research object. The improved BP neural network deformation prediction algorithm is designed by selecting the monitoring data, establishing data sample, dealing with sample data, and setting up the network parameters.(4) MATLAB R2012 a software is used to write the optimized algorithm program of BP neural network, through the debugging of model node training, we establish two kinds of deformation prediction model: the BP neural network that based on L-M algorithm and wavelet neural network. The predictive results of the wavelet neural network and BP neural network prediction results are analyzed, and comment on the stability of foundation pit. Results show that both two methods can do well reaction of short-term deformation regular of foundation pit, but the wavelet neural network model has more advantages than BP neural network model in convergence, computing speed and accuracy, which is a kind of good prediction method of the foundation pit displacement. It will have a good prospect development and application in the field of foundation pit engineering.