Abstract
There are many miles of low volume roads in cold regions within the USA and other countries, which are often built with minimal sub‐base and surface treatments. These roads are especially vulnerable to damage from heavy traffic loading during the late winter and early spring, when the previously frozen roadways begin thawing. To minimize costly roadway damage during the thaw-weakened period, spring load restriction (SLR) policies, which limit the axle loads of heavy trucks, are often implemented. The SLR protocol that is currently used by many agencies in the United States is the Minnesota Department of Transportation (MnDOT) protocol. That protocol is based upon cumulative air freezing and thawing indices (CFI and CTI, respectively), and recommends use of a variable reference temperature in the CTI calculations to account for increases in solar radiation that are observed during the early spring. The MnDOT protocol and reference temperatures work well for regions located at about the same latitude as Minnesota; however, for sites at different latitudes, local calibration may be required. Therefore, a major goal of this research was to evaluate and suggest modification to the MnDOT protocol based upon an extensive set of subsurface temperature data obtained from numerous test sites in New England (NE), the mid-west and Alaska. Additionally, a new SLR application protocol was developed, based on a model for estimating pavement surface temperatures, and was found to work much better at higher latitudes than the MnDOT protocol. To further assess and refine both the MnDOT and the new protocol, data obtained from falling weight deflectometer (FWD) tests was utilized. The FWD is an in-situ test that measures the deflection response of pavement by placing displacement sensors at several radial distances from a load plate. Increasing deflections result from weaker pavement conditions. FWD tests performed throughout the year and at frequent intervals during the spring were analyzed to capture the thaw weakening and recovery behavior of several roadway test sections in NE. Modulus values for the pavement layers were estimated using both back calculation and forward-calculation approaches, and were confirmed via finite element analyses. Mechanistic analyses were conducted to more accurately study the effect of frost-thaw processes on strains within the pavement layers and to provide guidance on SLR timing.