Non-unique solutions of tunnel characterization using ground settlment profiles.

摘要

The Garner and Coffman (2011) method provides a means to characterize an unknown underground facility (UGF) by performing a back analysis of the measured or allowable ground surface settlement profile. The described method uses both static (volume loss) and numerical (finite element) methods to determine the depth, diameter, and an estimation construction method of an UGF by an analysis of the settlement at the ground surface induced by tunneling.;The static characterization equations used in the Garner and Coffman (2011) method are based on the original work by Peck (1969) and subsequent refinements proposed by Cording and Hansmire (1975), 0' Reilly and New (1982), Aoyagi (1995) and Mair and Taylor (1997). The static methods, when coupled with a statistical analysis of the measured or allowable surface settlement profile, provide an initial estimate of both tunnel depth and diameter by attributing the volume of the ground surface settlement to the compression of the excavated area at depth. These methods rely on empirical coefficients developed from a database of existing tunneling projects from around the world. If multiple tunnels are present then a superposition technique is used to determine the settlement trough induced by each tunnel. After an initial estimation of a UGF's configuration, a numerical model is constructed to confirm and refine the initial estimate. The numerical analysis for each tunnel is performed using two (2) dimensional finite element software. As the simulation of three (3) dimensional processes during tunneling are critical to correctly determine the ground surface settlement induced by tunnel construction. The finite element simulation is then iterated (at different geometries) until the predicted UGF configuration produces the measured or allowable surface settlement profile. The Garner and Coffman (2011) method was calibrated using settlement data from eighteen existing tunneling projects in Bangkok, Taipei, Singapore, London, and Heinenoord. Validation of the Garner and Coffman (2011) method was then conducted using data from sixteen additional tunneling projects at the same sites which were not involved in the calibration process. The Garner and Coffman method (2011) identified a relationship between the groundloss coefficient and the finite element contraction increment (using Plaxis v8.6, 2010). For the validation sites the Garner and Coffman method (2011) predicted facility depths within 12 percent of the actual facility depth, facility diameter within 10 percent of the actual facility diameter, and facility center to center spacing within 19 percent of the actual center to center spacing. The Garner and Coffman (2011) method provides a predicted facility configuration that overlaps the actual configuration of the facility for 88 percent of the validation sites, and a predicted facility configuration that is within the actual configuration for 44 percent of the validation sites.