Vehicle dynamic load is a key parameter for design of new and rehabilitated pavements. In this paper, the Hilbert-Huang Transform (HHT) was utilized to analyze the dynamic axle load data obtained using an instrumented 5-axle semi-trailer equipped with an air suspension in the drive axels and a rubber suspension on the trailer axles. The experimental data was collected using three operating speeds (i.e., 38, 60.5, and 79 km/h) on five pavement sections with roughness international roughness index (IRI) ranging from 0.88 to 3.17 m/km. The size of the dataset allowed 11 levels of decomposition with test speed dependent frequencies ranging up to 4 cy/m. For each levels of decomposition, the Hilbert energy was used to quantify the extent of variations in dynamic axle load. The frequency subbands primarily affecting vehicle dynamics range from 0.4 to 2.5 cy/m and 0.5 to 1 cy/m forthe air and rubber suspensions, respectively. The rubber suspension impacts much higher energy per unit length traveled than the air suspension andis significant for the frequency subband between 0.32 and 2.5 cy/m. In addition, the Hilbert normalized energy of the dynamic axle loads shows highcorrelation with the conventional pavement roughness IRI at each vehicle speed tested. In summary, analyzing dynamic axle loads using the HHT method by considering the effect of pavement roughness, vehicle speed, and suspension types provided a highly accurate interpretation of pavement-vehicleinteraction. Simply stated, the HHT method is a preferable tool compared to the other frequency domain methods currently utilized for analysis of pavement-vehicle interaction
Abstract