Analyzing Influence Factors of Transverse Cracking on LTPP Resurfaced Asphalt Pavements through NB and ZINB Models

Abstract

The negative binomial (NB) and zero-inflated negative binomial (ZINB) models were employed to simulate the development of pavement transverse cracks on asphalt overlays and to evaluate the influence of different designs of asphalt overlays on crack development. Pavement transverse crack data were collected from 15 long-term pavement performance (LTPP) SPS-5 test sites. Analyzed factors include traffic level, overlay thickness, mixture [using reclaimed asphalt pavement (RAP) or virgin], intensity of surface preparation (mill or no mill) before overlay, total thickness of pavement, and freeze index. Analysis results indicate that the NB and ZINB models were effective in simulating the development of pavement transverse cracks by addressing the overdispersion. In addition, the ZINB model outperformed the NB model by explaining the excess zeros in the cracking count data to capture both the initiation and propagation of cracking. The regression analysis indicates that mill before overlay is effective in retarding the initiation of cracks, but not the propagation of cracks. Thicker overlay appears to reduce transverse cracking. High traffic level or using RAP is likely to increase the number of transverse cracks. Total thickness of pavement and freeze index are not significant for the development of transverse cracks.