| With the development of the high-rise bulidings and underground structures,cons- truction of deep foundation pit appear unceasingly.with the rapid growth of the nunmber of deep excavation works,the experience in the design and consteuction has be accumulated gradually at the same time,pit accidents have also continued to ocuur.according to incomplete statistics,as of december 2007,the 166 cases major pit accidents have been occurrenced in 37 large and medium cities in the country. duo to the accidents, the national economy and people`s lives,property have been suffered great loss.there are several main forms about the accidents,supptorting structure damaged by large displacement, earthfall and large-scale landslide,creaking and collapse of the around road,deformation and destruction of the contiguous undergound facilities,cracking and collapse of the near building and so on .according to statistical analysis,every accident has a directing bearing to the ineffective monitoring and the inaccurate peril forecasting.therefore,they are necessary to monitoring accurately,to process date scientificly and timely,to practice the information construction for the deep pit construction.Focusing on the monitoring of deep pit,deformation control,and informative construction,the full paper mainly carried out the following work:First of all,the paper introduced the main centens of the pit monitoring and technical methods to be involed.made a brief introduction about the work,including the content of monitoring,the principles and the usage of the instruments,notes,dates processing and so on.the paper presents to apply the BP neural network algorithm to monitoring data and forecast,using the high nonlinear mapping ability of the BP algorithm,done an effective data processing.,Second,after introducte basic theory of artificial neural netwok algorithm,focusing on the BP algorithm and the BP neural network model.there are some difficulties about the engineering applications of the BP network:it is important to determine the network structure,but ,there are some complx issues,suchas,the decision of the number of input layer units and the means of each unit,the selection of param nhide,data normalization methods and large amount of computing and so on .in this paper ,to address these issues combined with engineering practice,select the approprite parameters,set up an pit deformation predicton models,and achieved data processing in Matlab.Then,combining the fulfillment of a pit engineering in the Guangzhou Metro line 5,done a practical study about the foundation pit construction monitoring and the information construction,made a theoretical and practical research on the implementation of the information construction in the pit,achieved good results.Finally,the major work and conclusions of the dissertation are summarized and the prospect for the research in future is forecasted. |