Ground Settlement Analysis of Rectangular Pipe Jacking Construction Considering Conditions Random Field of Soil Parameter

Abstract:

Considering the insufficient utilization efficiency of geotechnical data when using the random field theory to describe the variability of soil parameter space, by combining the Kriging theory and the data-driven random field theory based on Bayesian compressive sensing and Karhunen-Loeve(KL) series expansion method, a data-driven three-dimensional conditional random field generation method was proposed. This method was applied to characterize the variability of soil parameter space in a certain rectangular pipe-jacking project site in Suzhou City. [JP+1]A three-dimensional numerical model considering the variability of soil parameter space was established, and the settlement law caused by pipe-jacking construction was analyzed. The results were compared and verified with the on-site measured data, with the maximum error being less than 10%. The research results show that the generated data-driven conditional random field can fully utilize the survey data in the actual engineering site, achieve the optimization simulation of the conditional random field of soil parameter space variability, complete the constraint of the survey data on the known sampling points, and reduce the uncertainty. Compared with deterministic calculation, uncertain calculation can fully represent the trend and fluctuation range of surface settlement caused by the rectangular pipe-jacking construction, providing a reference basis for the design and construction of complex geotechnical engineering.