Preliminary optimization design of inverted asphalt pavement structure using response surface method

Abstract

Inverted asphalt pavement (IAP) has the advantage of maintaining good performance under the condition of thin asphalt concrete (AC) layer compared with conventional flexible asphalt pavement. The stress-dependent property of unbound aggregate base (UAB) plays an important role in mechanistic responses of IAP. The focus of this paper is to quantitatively analyze the function of UAB’s stress-dependent property on the structure combinations design of IAP using response surface method (RSM). In this paper, an improved UMAT subroutine was developed to characterize the stress-dependent stiffness of UAB. The three-dimensional FE pavement model was established and validated against KENLAYER calculation and field FWD test deflection data. The critical mechanical response, maximum principal strain and bulk stress, were selected to evaluate the fatigue performance of AC layer and stress-dependent property of UAB. The significance analysis was conducted to describe the relationship between input variables (thickness and stiffness of structural layers) and the critical mechanical response variables. An optimization design method to balance the fatigue performance and stress-dependent property was proposed to obtain the optimal structure combinations of IAP. Research results demonstrate that the thickness and stiffness of the AC and UAB layers present a clearly significant impact on the maximum principal strain at the bottom of AC layer and the bulk stress in UAB, particularly during the summer season, whereas the thickness and stiffness of CTB and stiffness of subgrade have a minor impact. The responses of bulk stress and maximum principal strain can be empirically predicted by 2FI and Quadratic models. Optimal structure combinations (13.6 cm AC layer and 10 cm UAB) were recommended after quantifying the stress-dependent properties of UAB. The recommended structural combination has good fatigue resistance and can also efficiently minimize AC and UAB layer thickness.