An iterative program to back-analyze grain-size distribution from a predetermined soil–water characteristic curve

Abstract

Numerous methods for predicting unsaturated soil properties based on basic soil parameters have been introduced to reduce the cost of unsaturated soil testing. This research proposes an unsaturated soil estimation program that uses a predetermined soil–water characteristic curve (SWCC) to predict the grain-size distribution (GSD) using computer iteration. The results indicate back-calculating a GSD from a given SWCC is possible, and that different GSDs can produce the same SWCC. A Monte Carlo approach examining variations of the GSD was conducted and associated packing porosities are provided. The program was tested on coarse- and fine-grained soils to determine the program’s capability and indicate it is appropriate for samples dominated by silty sand and silt but not clay. The back-analysis program described here could bypass the arduous testing phase of multiple soils to find a suitable SWCC for a capillary break layer as it can start with a predetermined SWCC and estimate a suitable GSD. In such efforts, the most important characteristic (i.e., strength or stiffness) of a soil must be determined when considering the GSD required because an infinite number of GSDs with similar properties could produce the same SWCC.