GSA Annual Meeting in Seattle, Washington, USA - 2017

Paper No. 193-9
Presentation Time: 10:40 AM

IMPLEMENTATION OF A DYNAMIC CAPILLARY SUCTION MODEL FOR NUMERICAL MODELING OF UNSATURATED SYSTEMS WITH CHANGING POROSITY


JOHNSON, Peter1, STAUFFER, Philip H.2, BOURRET, Michelle2 and ZYVOLOSKI, George A.2, (1)EES-16, Earth & Environmental Sciences Division, Los Alamos National Laboratory, Mailstop T003, Los Alamos, NM 87545; Department of Geology, The University at Buffalo, 126 Cooke Hall, Amherst, NY 14260, (2)EES-16, Earth & Environmental Sciences Division, Los Alamos National Laboratory, Mailstop T003, Los Alamos, NM 87545, pjjohnson@lanl.gov

Many geological situations involve fluid flow through the unsaturated zone during which porosity of the medium changes due to chemical, thermal, or mechanical processes. This effect can be seen in a variety of scenarios, such as hydrothermal systems, soil development, diagenesis, and others. In some cases, opening or closing of pore space can be extensive, resulting in local porosity that changes greatly over time. This can present a problem for numerical simulations of flow in these systems. One concern in such simulations pertains to functions calculating capillary pressure, which are commonly specified for initial conditions or some average value during the course of a model run. Neglecting the effects of porosity on capillary suction can lead to unreasonable model results if porosity varies greatly over time from initial conditions.

In support of experiments and studies relating to disposal of high level radioactive waste in salt, we have conducted numerous model runs using the coupled heat/stress/mass flow model Finite Element Heat and Mass (FEHM, https://fehm.lanl.gov). Capillary forces tend to pull brine towards heat sources where evaporating brine causes element undersaturation. Precipitation of salt as brine evaporates decreases porosity, which should result in greater capillary forces for comparable saturation values than the initial specified conditions. Vapor condensation elsewhere causes local dissolution of salt, increasing porosity, and should decrease the importance of capillary effects in those areas. This caused problems during model runs where increasingly porous areas had unphysically high suction effects which kept them more saturated than expected and led to a further dissolution of the salt. As a result, portions of the model domain erroneously show porosity approaching 1 while retaining high saturation. To address this, we have implemented capillary pressure functions that calculate capillary pressure for each node within each timestep as a function of both porosity and saturation. Residual saturation and the maximum saturation above which capillary pressures go to zero are calculated as end members and a fit applied between them. This approach allows for dynamic capillary suction calculations that ameliorate problematic pore-capillary behaviors in the model.