Measured and Estimated Suction Indices for Swelling Potential Classification

Abstract

McKeen’s expansive soil classification methodology relies on a parameter referred to as the “total suction-water content index” for describing the slope of the soil–water characteristic curve on a semilog plot. The swelling potential of expansive soils is qualitatively classified (e.g., “low” or “high”) based on the magnitude of the total suction-water content index. This study examines the validity of using a “benchmark intercept simplification” for indirectly estimating the total suction-water content index when complete soil–water characteristic measurements are not available or economical. Suction indices estimated using the benchmark intercept simplification are compared with indices measured directly using the noncontact filter paper technique for 80 undisturbed expansive shale specimens from the Colorado Front Range Corridor. The results show that the suction-water content index is consistently overestimated using the benchmark simplification by amounts ranging from negligible to 50%, and averaging 23%. For 49 of the 80 specimens (61%), the estimated indices fall in different swelling potential categories than the measured indices. In 44 of the 49 cases (90%), the estimated indices fall in higher swelling potential categories than the measured indices. These discrepancies reflect potential errors that may arise from the use of the benchmark intercept simplification in classifying expansive soils.