A study is undertaken to formulate and solve the equations governing the time dependent response of gassy soils. The formulations have been implemented into a finite element program capable of analyzing stress state and flow phenomenon under a variety of boundary conditions. The validity of such a program has been verified by comparing its results with closed form solutions. This program has then been applied to simulate the processes involved in depleting oil sand reservoirs in order to give some insight into the mechanism causing fluid flow and sand production.

In order to compute the change in pore pressure in gassy soils, both under undrained and transient states, the concept of a homogenized compressible phase is introduced which is used to treat a multiphase soil system as a two phase material. Such a hypothesis is found to be highly akin to the procedure normally followed in finite element analysis since it replaces the compressibility of fluid and solid phases by one phase which occupies the entire soil volume. Assuming that the gases are present only in the form of bubbles within the liquid phase, the compressibility of the fluid phase is obtained by giving due consideration to the mixture of liquid and gas phases using Boyle's law and Henry's law and taking account of the surface tension effects. Under undrained conditions the pore pressure is computed by invoking volumetric compatibility between the soil skeleton and the compressible phase. Under transient conditions, the pore pressure is calculated by using Biot's theory of consolidation and modifying it to account for a soil with an incremental stress-strain law and a compressible fluid phase. Formulations are derived to compute the change in pore pressure and effective stress as a result of changes in temperature and a methodology is proposed for implementing these effects into finite element analysis. Various numerical techniques are incorporated for increasing the accuracy, efficiency, and stability of the finite element procedure. The computer program based on these formulations is verified by comparing the computational results with known solutions for several problems.