Friday, 14 July 2006 - 2:05 PM

Finite Element Analysis of Fluid Transfer in Soil Using CT.

Thomas R. Elliot1, Richard J. Heck1, Dan Reynolds2, and Bahram Gharabaghi1. (1) University of Guelph, Guelph, ON N1G2W1, Canada, (2) Agriculture & Agri-Food Canada, 2585 County Road 20, Harrow, ON N0R1G0, Canada

This study combines methods employed in soil science research and computational fluid dynamic engineering through the use of Computed Tomography (CT). By utilizing CT for identifying soil features, specifically macro and mesopores for this study, the digital imagery can be manipulated for the purpose of numerical, or Finite Element (FE), modeling. FE models, when applied in conjunction with fluid dynamics, can numerically simulate the potential fluid flow through a porous media. While a majority of hydrologic transfer models are designed for 1D or 2D flow simulation, there is a growing number of FE computer programs, such as FLUENT, I-deas or ADINA, capable of 3D flow calculation. The functional data input for finite element models include the physical characteristics (porosity, heat conductance, resistance to applied force, etc.) and arrangement of the porous media in space. The primary objective is a method of soil feature segregation and numerical model generation based on micro-CT imagery. Secondary objectives include the calibration parameters of the numerical model for the specific soil type as well as identifiable primary flow channels within the numerical simulation of fluid transfer. Sample acquisition requires a custom core sampler utilizing a 65mm Lexan cylinder, as the standard aluminum sample containers induce artifacts in CT imagery. Core sampling of a gleysol under long term crop rotation at the AAFC Harrow Research station, Ontario Canada, reveals strong structure and visible macropores in the intact soil column. No post-acquisition processes were applied to the samples before CT imaging. Using a second generation cabinet CT scanner the sampled intact soil column, measuring 100 mm vertical extent and 65 mm diameter, is imaged at 15 micrometer resolution. The imagery is based on the attenuation of X-rays by soil constituents, measured on an atomic density scale called Hounsfield Units (HU). 3D image processing segregates the pore structure within CT imagery by isolating voxels comprised of a single constituent (i.e. a pure voxel). A primary threshold is then identified based on the peak distribution of pure voxel HU values in a histogram. A secondary threshold integrates 3D context by utilizing the number of pure voxels in the neighbourhood of any given voxel as the determinant factor for feature segregation. The I-deas FE analysis software is used to import point cloud data of the segregated pore structure. Using the object surface creation and shell mesh functions of I-deas a FE, specifically boundary element, model of pore structure is generated. Assuming an ideal fluid, absolute boundary conditions will be employed to ignore the conductance of water through the pore walls for computational efficiency and ease of calculation. Material properties will be assigned assuming no friction occurs between the fluid and the pore space as well as standard viscosity and compression characteristics of water. Governing differential equations are applied to the FE model for fluid transfer experiments using an approximation of the Navier-Stokes equations for an incompressible Newtonian fluid, with Reynolds averaging. When considering choked flow, this model will dismiss the Venturi effect for ease of calculation. A number of FE fluid transfer models are generated, with each model being a product of how boundary condition and governing differential equations are employed based on certain assumptions of fluid transfer. The various models calibrated will be examined for their relative accuracy across all core samples. Deliverables from this study will include an intact soil core FE modeling method using CT, as well as the parameters and accuracy of the FE model developed for the sampled soil. The impact of this study will be that fluid transfer in soil can be examined in a numerical simulation for consulting, remediation and optimization practices.

Back to 2.1B Soil Hydrology, Structure, and Micromorphic Propoerties (Soil Porous System) - Oral
Back to WCSS

Back to The 18th World Congress of Soil Science (July 9-15, 2006)