Determination of precompression stress from uniaxial compression tests

Abstract

Soil compaction has been studied for many years due to its implications for crop yield. In addition, precompression stress is important tool in evaluating soil load capacity and its implications for soil structure. This study evaluated whether precompression stress fitted by different methods gave rise to different results when a wide range of soils was tested. Soil cores were collected from different depths at 13 sites in Sweden, Denmark and Brazil and subjected to uniaxial compression tests. Soil precompression stress was determined using five different fitting methods: (1) Casagrande by polynomial fit; (2) Casagrande by van Genuchten equation fit; (3) intercept of the virgin compression line (VCL, straight portion of stress–strain curve) and regression with the first three points of the curve; (4) intercept of VCL and regression with the first two points of the curve; (5) intersection of VCL with the x-axis at zero strain. Method 1 produced the highest precompression stress values, while method 5 differed from other methods for all sites. Methods 1 and 3 did not significantly differ for most sites. Mean precompression stress was 50–130 kPa for the different methods. Polynomial and sigmoidal equations showed a good fit for the compaction curve in our dataset ( R 2 = 0.99), but Casagrande method parameters, e.g. point of maximum curvature, its tangent and slope of VCL, were different, resulting in different precompression stress values. Choice of method proved to have a significant influence on precompression stress values. Although methods 1 and 2 are both Casagrande-based, the results frequently differed. For the stress range tested (up to 1000 kPa), the van Genuchten equation was less suitable than polynomial fitting, since the compaction curve was usually not sigmoidal and the inflection point was sometimes outside observed values. Thus, the method used to fit precompression stress must be carefully chosen to avoid overestimates and underestimates. This restricts use of the absolute precompression stress value as a limit for soil stress in order to avoid soil compaction.