Obtaining soil-water characteristic curves by numerical modeling of drainage in particulate media /

Shoarian Sattari, Amir
The soil water characteristic curve (SWCC) reflects fundamental drainage properties of partially saturated soils by revealing the relation between soil suction and water content. Although during the past decades, various experimental methods have been proposed for obtaining the SWCC, these approaches are generally time consuming, expensive and highly dependent on operator skills. One solution to this has been sought through various empirical or physico-empirical formulations that link the SWCC to common soil properties, through transformation calculations or within available SWCC databases. Another solution is to fit a curve onto partial SWCC data via regression analysis. A more reasonable approach both in terms of accuracy and cost could be to simulate the pore-scale drainage of soil medium with application of computer modeling techniques. This study is aims to model the drainage of soil medium through developing a computer program in MATLAB for simulating the drying SWCC. In the proposed approach, initially, a solid phase of soil structure is generated based on the particle size distribution as well as void ratio (e) of a given soil sample. Next, the volume of pore water is divided into two main portions: (i) bulk water and, (ii) liquid bridges. Assuming the inter-particle region of three adjacent particles with minimum hydraulic radius as a pore throat, the air entry pressure of each pore throat can be computed. To this end, the Young-Laplace equation is solved through a combination of the finite difference approximation and Newton-Raphson (Jacobian) numerical methods. In this simulation the amount of applied suction is gradually increased and resulting drainage scheme of bulk pores is visualized step by step. Eventually the final residual water content of bulk pores and liquid bridges after each suction increment are determined and plotted as SWCC. The usefulness of the developed computer program is verified by comparison of the obtained SWCC to those attained based on experimental and empirical techniques. The application of proposed method is limited to non-plastic soils, and is not suitable for modeling clay particles. The developed method is capable of providing SWCC, with an almost perfect level of conformity to the experimental results, in a timely manner. Moreover, it is shown that the SWCCs obtained by the proposed algorithm are much more accurate than results of the estimation methods found in the literature for a variety of real soils.