A SIMPLE PROCEDURE FOR ESTIMATING PRECONSOLIDATION PRESSURE FROM SOILCOMPRESSION CURVES

Classical graphics and regression procedures have been used to estimat e preconsolidation pressure (sigma(p)) from soil compression curves, b ut none of these procedures is easy to use and they often involve subj ective judgement. This paper presents a simple procedure for estimatin g sigma(p) from uniaxial compression tests for either saturated or uns aturated soil conditions. We evaluated five methods for estimating sig ma(p) from standard soil compression curves for an applied stress sequ ence of 25, 50, 100, 200, 400, 800, and 1600 kPa. Four methods estimat ed sigma(p) as the intersection of two lines: (a) the regression line obtained for the first two, three, four or five points of the applied stress sequence in the secondary compression portion of the compressio n curve and (b) the extension of the virgin compression line determine d from the points associated with applied stress of 800 and 1600 kPa. Method 5 consisted of the Schmertmann method. The sigma(p) determined for each method was compared to sigma(p) estimated using the graphical procedure of Casagrande for 288 soil compression curves from three so ils in Michigan and from values reported in the literature. Methods 1 and 5 fit our data best at low sigma(p) (high soil water content) whil e methods 2 and 3 fit the data better at high sigma(p) (low soil water content). Based on a low RMSE (18), a high R(2) (0.92), and closeness of fit to the 1:1 line, a combination of methods 1 and 3 was selected as the best estimation procedure. For data from the literature, metho ds 1 and 2 provided the best estimate based on lowest RMSE of 5 to 9, R(2) of 0.98 to 0.99, and the closest fit to the 1:1 line. The combine d methods were not tested for published data since matric potentials f or measured values were unknown. The final procedure, combined methods 1 and 3, was programmed into a computer spreadsheet provided in an Ap pendix. This procedure provides a fast and reliable estimation of sigm a(p) for saturated and unsaturated soil conditions and eliminates subj ective judgment associated with classical graphical procedures.