RADIAL BASIC FUNCTION-BASED ANALYSIS OF DYNAMIC DEFLECTION OF INVISIBLE FLEXIBLE PAVEMENT LAYER PROFILES

摘要

This study proposes a radial basic function (RBF) neural network model which can simulate the dynamic deflection process of invisible individual layers in the full-scale flexible pavement along with an increase of load repetitions. The training and testing data is formed through empirical and conceptual judgment on the final profiles of the four pavement layers in the test. The independent and dependent variables are defined as the known top and invisible layer deflections respectively. Then, the RBF model produces the numerical results between layer dynamic deflections. Finally, several parameters are suggested to study the response of the invisible pavement layers. The RBF model shows that the implicit dynamic relationship between pavement layer deflections could be modeled by a static state of the flexible pavement. Furthermore, some working features of the pavement might be revealed from its dynamic response.