Abstract
The trapdoor problem is a useful model for providing a clearer understanding of stress distribution around basic geotechnical engineering structures such as tunnels, conduits and anchor plates. The interest is mainly directed towards the determination of soil load on the trapdoor, which can be substantially different from the initial geostatic loads, when the trapdoor is moved even slightly. Development of shear bands around the yielding trapdoor is also of interest. A series of centrifuge and 1g model tests were conducted to study active arching in dry granular soil on circular trapdoors. The study was undertaken mainly because of two reasons: (i) limited success in earlier centrifuge modeling studies of a trapdoor problem, and (ii) a lack of understanding of the load-displacement characteristics under axisymmetric conditions. A trapdoor assembly and in-flight precompression technique were developed to perform a series of tests involving different overburden soil thickness on circular doors of different diameters. Correct initial geostatic loads were measured. Greater confidence in the experimental results was obtained because modeling of models type experiments were also successful. A parametric study involving different overburden soil thickness to trapdoor diameter ratios (H/D) ranging from 0.67 to 6 was conducted. The pattern of shear bands in sand above the trapdoor observed in a centrifuge model under a high gravity field differed considerably from that in a 1g model. Maximum arching was completely mobilized at movements of only 1.5% of trapdoor diameter. The minimum load on the trapdoor became constant at H/D equal to 5 regardless of the initial overburden pressure or height of the soil model.
