PERFORMANCE PREDICTION WITH THE MMLS3 AT WESTRACK

The one-third scale Model Mobile Load Simulator (MMLS3) was used to traffic four pavement sections at WesTrack to establish and validate its ability to reliably predict rutting performance under full-scale trafficking. Researchers utilized two different analysis methods in determining rut depth (RD) from transverse profiles under both scaled (MMLS3) and full-scale (WesTrack trucks) loading. Researchers also completed a limited laboratory testing program to complement results contained in the WesTrack database. Researchers ranked performance and conducted statistical analyses using both field and laboratory results to compare the rutting performance of three coarse-graded replacement sections that showed poor performance and one fine-graded section with good performance. They also compared results from both RD analysis methods and both loading conditions. Rankings based on field results using either RD analysis method were in close agreement, and the majority of the rankings based on laboratory results matched those from the field. Two methodologies to predict rutting performance using the MMLS3 were also demonstrated. For both, researchers recommend the critical temperature for permanent deformation over a hot week during the summer of a 30-year period for MMLS3 testing. A minimum of 100,000 load repetitions, three RD measurements along the length of an MMLS3 section, and a standard RD analysis method were suggested to determine a mean RD for (1) comparison with acceptable performance criteria for the first prediction method and (2) use with a theoretical rutting ratio based on stress analyses for the second method. The second prediction methodology involves a quantitative comparative analysis based on the hypothesis that the extent of rutting is dependent on the nature of the vertical contact stress under the tire, the material characteristics and pavement structural composition, and the prevailing environmental conditions prior to and during trafficking. Results from the initial analysis using this second methodology exhibited some apparent inconsistencies that led to a second, more detailed analysis that included additional important factors necessary for successful implementation of this performance prediction methodology. These factors included lateral wander effects, transverse profile measurement errors, misalignment of the MMLS3, and tire contact stresses at elevated temperatures. This more comprehensive analysis involved improving material property estimates, revising RDs, considering deformation throughout the pavement structure, and accounting for differences in lateral wander between the two loading conditions. Results from this second analysis indicated that the hypothesis required in the performance prediction methodology appears to hold for the four independent pavement sections at WesTrack, provided steps are taken to factor in differences in loading and environmental conditions between MMLS3 and full-scale truck trafficking.