Potential vertical rise of lime-treated expansive clays using centrifuge technology
Expansive soils are prevalent throughout Central Texas and can cause extensive damage to pavements and other lightweight transportation projects. Methods of directly quantifying soil swell potential are often prohibitively time-consuming, and indirect methods that correlate swell to soil index properties are often based on limited empirical data and ignore variances in soil mineralogy that can significantly affect swell. Soil stabilization or modification, using additives like Portland cement or lime, is often utilized to mitigate the effects of expansive soil. In particular, the Texas Department of Transportation (TxDOT) recommends lime treatment for highly plastic soils below many transportation projects. However, many common design procedures do not directly measure the swell potential of the lime-treated soils and rather rely on methods like the Eades-Grim pH test to determine dosage. Additionally, design procedures do not often include a method for determining project-specific treatment depths. This research seeks to develop a procedure for directly quantifying the swell of lime-treated soils such that a project-specific treatment depth and dosage may be prescribed. Centrifuge tests performed on Eagle Ford clay were used to assess the effects of the testing procedure on swell results. It was determined that mellowing lime-treated soil samples did not significantly affect their swell potential but did increase soil workability. Additionally, curing of lime-treated soil samples decreased swell for samples treated with 4% by dry mass hydrated lime after curing for at least 6 weeks, but curing for any time did not decrease swell for samples treated with 2% hydrated lime. It was concluded that lime-treated swell testing could be optimally performed on uncured samples that were allowed to mellow for 12-24 hours to maximize workability. The reduced testing procedure for directly calculating an untreated and lime-treated soil potential vertical rise (PVR) recommends obtaining 3 data points across a representative range of stresses for each distinct layer of untreated soil in a given stratum, which is fit to a semi-log linear curve. One data point is used for each lime dosage to be tested, and the approximation of a unique soil swell pressure is used to produce a semi-log linear stress-swell curve from each of these points. The area under each curve for the stress range in question is then used to calculate the PVR of a soil profile. PVR analyses on two sites from San Antonio in Bexar County, Texas show that the assumption of log-linear stress-swell curves and the approximation of a constant swell pressure return PVR calculations that are similar to those calculated from stress-swell curves that require more parameters and more data points to propagate.