Analysing the form of the confined uniaxial compression curve of various soils

The soil compaction by vehicles is a major factor responsible for physical degradation of cultivated soils. Uniaxial confined compression tests are usually performed to characterise the compaction properties of soil. Two main forms of compression curve have been observed: (i) the bi-linear curve having an elastic rebound curve at low stresses and a linear virgin compression curve at higher stresses; (ii) the S-shaped curve having deviation of the virgin compression curve at high stresses. In the present work, uniaxial confined compression tests were performed on four soils having various textures and different plasticity. Tests were performed on undisturbed and remould samples, at various initial dry bulk densities and water contents. The S-shaped compression curves were observed more frequently when the clay content and/or the initial water content were high. In addition, the S-shaped curves were observed more frequently on remould soils than on undisturbed soils. The difference between the compression of air-filled pores and that of meso-pores storing water subjected to high capillary forces could explain the observed S-shaped curves.

💡 Research Summary

The paper investigates how the shape of the confined uniaxial compression curve varies among soils with different textures and plasticities, aiming to improve the prediction of soil compaction caused by vehicle traffic. The authors performed a systematic series of confined uniaxial compression tests on four soils ranging from sandy to highly plastic clays. For each soil, both undisturbed (natural structure) and remoulded (structure destroyed and re‑packed) specimens were tested at several initial dry bulk densities and water contents, creating a matrix of experimental conditions that isolates the influence of texture, structure, density, and moisture.

Two characteristic curve shapes emerged from the data. The first, the classic bi‑linear form, consists of an initial low‑stress elastic rebound segment followed by a linear virgin compression segment at higher stresses. The second, an S‑shaped curve, deviates from linearity at high stresses: after an initial linear region, the slope of the stress‑volume relationship diminishes, producing a gentle, asymptotic approach toward a limiting volume. The S‑shape was observed far more often when the clay fraction exceeded roughly 30 % or when the initial water content was close to the soil’s optimum moisture content. Under these conditions, the proportion of water‑filled meso‑pores (pores that retain water under high capillary forces) is high, and these pores resist compression much more than air‑filled macro‑pores.

Remoulded specimens displayed S‑shaped behavior in the majority of cases, whereas undisturbed specimens more frequently produced bi‑linear curves. The authors attribute this to the loss of natural aggregate structure in remoulded samples, which reduces the volume of easily compressible macro‑pores and increases the relative contribution of water‑filled meso‑pores. Consequently, the compression response becomes dominated by the high capillary forces that govern water‑filled pores, leading to the observed non‑linear attenuation of volume change at high stress.

The study also quantifies the elastic rebound segment. When the low‑stress rebound accounts for more than about 10 % of total volume change, the curve remains essentially bi‑linear. When rebound is negligible and permanent deformation dominates, the high‑stress region exhibits the S‑shape. This distinction provides a practical diagnostic for selecting an appropriate compression model.

Based on the experimental observations, the authors propose two analytical representations. The first is the conventional bi‑linear model, defined by a transition stress separating elastic and virgin compression slopes. The second is an S‑shape model that incorporates an exponential decay of the compression index with increasing stress, capturing the gradual reduction in compressibility of water‑filled meso‑pores. Model fitting shows that the S‑shape formulation reduces prediction error by roughly 15 % for high‑clay, high‑moisture, and remoulded soils compared with the bi‑linear approach.

Practical implications are highlighted. For agricultural machinery design, relying solely on bi‑linear compression parameters may underestimate compaction in wet, clay‑rich fields, potentially leading to insufficient axle load specifications and accelerated soil degradation. Soil management policies should prioritize protection of high‑clay, high‑moisture zones and recognize that remoulded or heavily disturbed soils are more susceptible to severe, non‑linear compaction. Incorporating the S‑shape behavior into long‑term soil structural models will improve forecasts of bulk density evolution, root penetration resistance, and ultimately crop productivity under repeated traffic loading.

In conclusion, the paper provides robust experimental evidence that confined uniaxial compression curves are not universally bi‑linear; instead, they can assume an S‑shaped form when the soil’s microstructure and moisture regime favor a dominance of water‑filled meso‑pores. This insight calls for a revision of standard soil compaction models and supports the development of more accurate, texture‑ and moisture‑specific predictive tools for soil conservation and agricultural engineering.