Use of Falling Weight Deflectometer Multi-Load Level Data for Pavement Strength Estimation.

Abstract

The objective of this study is to describe a mechanistic-empirical approach to developing an analysis method for assessing pavement layer conditions and estimating the remaining life of flexible pavements using multi-load level Falling Weight Deflectometer (FWD) deflections. A dynamic finite element program, incorporating a stress-dependent soil model, was developed to generate the synthetic deflection database. Based on this synthetic database, the relationships between surface deflections and critical pavement responses, such as stresses and strains in each individual layer, have been established.A condition assessment procedure for pavement layers using multi-load level FWD deflections is presented in this study. The results indicate that the proposed procedure can estimate the base and subgrade layer conditions. However, large variations were observed in the relationships between the DBCI and desg values and the subgrade CBR values for aggregate base pavements. A FWD test with a load of 53 kN or less does not result in any apparent nonlinear behavior of the subgrade in aggregate base pavements. With regard to the condition assessment of the asphalt concrete (AC) layer, the AC layer modulus and the tensile strain at the bottom of the AC layer are found to be better indicators than the deflection basin parameter.The procedures for performance prediction of fatigue cracking and rutting are developed for flexible pavements. The drastically increasing trend in fatigue cracking with time may not be predicted accurately using the proposed procedure. Such trends may be due to the environmental effects and the inconsistent distress measurements. Predicted rut depths using both single and multi-load level deflections show good agreement with measured rut depths over a wide range of rutting potential. However, the procedure using single load level deflections consistently underpredicts the rut depths. This observation demonstrates that the rutting prediction procedure using multi-load level deflections can estimate an excessive level of rutting quite well and, thus, improve the quality of prediction for rutting potential in flexible pavements.