| With the rapid development of basic construction in China, proportion of tunnelingengineering in traffic engineering will be further expanded. Limited to objective conditionsuch as engineering geology, test funds, investigation may not present mechanicalparameters completely. Through surrounding rock geological situation revealed by tunnelexcavation and deformation of monitoring to feedback, design parameters can be acquiredaccurately, that will modify and optimize plan both tunnel construction and supporting ofrock mass.Based on monitoring of tunnel construction from Zhu-Zang Tunnel in Gu Ctiy-Zhu XiExpressway, the back analysis neural network model of surrounding rock mechanicsparameters is built on the basis of analyzing displacement measurement data andregression analysis model predictions. The major research contents are as follows:(1) For complexity and particularity of tunneling engineering problems, this paperexpounds the necessity and importance of back analysis brought into tunnelingengineering, and discusses the basic theory of back analysis and BP neural network on thebasis of displacement measurement.(2) Considering high discreteness of field displacement measurement data andvolatility of test data which change by time in the process of measurement, regressionanalysis model is used by displacement measurement data processing, in order to obtainmathematical expressions of change curves with surrounding rock displacement and tenseafter tunnel excavation.(3) Based on regression analysis finial deformation values, BP neural network modelis used by nonlinear displacement back analysis with mechanics parameters (deformationmodulusE, cohesionC, internal friction angle) of tunnel surrounding rock anddisplacement. |