Date of Award


Publication Type

Doctoral Thesis

Degree Name



Civil and Environmental Engineering


Engineering, Civil.




Soil-steel structures are flexible conduits made of corrugated metal sheets and buried in a well compacted granular media. Classical methods of design for conventional sized flexible buried conduits are simple and not suitable for long span structures, especially under relatively shallow cover. Herein, the live load effect cannot be considered secondary as in the case of conduits under deep cover. Also, field experience has shown that failure of the structure could be initiated by soil failure. The present codes avoid the problems associated with shallow cover by requiring a minimum depth of cover. This requirement is empirical and does not differentiate between the different shapes of conduits. A plane strain finite element analysis is developed and used to study long span soil-steel structures of different shapes under any depth of cover and loading condition with due consideration given to the soil state of stress. The structure wall is replaced by conventional beam elements. Soil is simulated by a combination of constant strain and linear strain compatible elements. Storage and time requirements for the numerical solution are kept to the lowest level without sacrificing the desired accuracy. Two-noded spring type interface elements are introduced between soil and culvert elements. Nonlinear stress-dependent hyperbolic relationships are used to model the soil and shear behaviour of interface elements. Construction procedure is simulated by automatically generating the finite element mesh layer by layer and by including the effect of compaction in the incremental analyses. Live load is applied concentrically or eccentrically, in increments. Initiation and propagation of soil failure are detected based on a curved Mohr envelope, and applying a stress transfer technique. A finite element program is developed on the basis of the foregoing analytical procedure to predict the state of stresses in the system due to dead and live loads, as well as the live load causing soil failure. The analytical results are compared with experimental results for live load tests conducted on existing long span soil-steel structures under shallow cover. The results of laboratory tests carried out until failure of soil are also compared with the analytically predicted values. Reasonable agreement is found in both cases. An analytical study establishes the parameters governing soil failure above the conduit. These parameters include the depth of cover, magnitude and position of loading, and the size and geometric shape of the conduit.Dept. of Civil and Environmental Engineering. Paper copy at Leddy Library: Theses & Major Papers - Basement, West Bldg. / Call Number: Thesis1982 .H244. Source: Dissertation Abstracts International, Volume: 43-03, Section: B, page: 0806. Thesis (Ph.D.)--University of Windsor (Canada), 1982.