Field data on frost heave, thaw weakening, and frost penetration acquired over a period of 3 years at the Winchendon field test site are analysed. These data show that simple grain-size criteria such as the percent of particles smaller than 0.075 or 0.02 Mm correlate weakly with field performance. Although soils with few fines (particles smaller than 0.075 Mm) performed well in the field test, some of those with intermediate and many fines performed well and some poorly, with a great deal of scatter. The U.S. Army Corps of Engineers Frost Design Classification System and tabulated freeze test data provided very wide ranges of frost susceptibility compared with the field performance of the 12 test soils. Comparisons of the field frost heave data with data on soils in freezing conditions from the massachusetts department of public works heave stress test on nonplastic soils and with data from the new freeze-thaw test of the U.S. Army Corps of Engineers Cold Regions Research and Engineering Laboratory (CRREL) show excellent correlations. The thaw-weakening California Bearing Ratio (CBR) data from the CRREL freeze-thaw test correlated well with the resilient deflections from repeated plate bearing (RPB) tests in the field. Frost penetration depths calculated by the modified Berggren formula are in general agreement with, but somewhat larger than, the frost penetrations measured at the Winchendon field test site. This may possibly be due to the effects of freezing-point depressions of soil moisture, approximations in the adjustment of air to ground surface temperatures, and other factors not accounted for in the field test.
Abstract