A probabilistic approach for modelling deterioration of asphalt surfaces

Abstract

The following paper details findings from New Zealand Transport Agency’s research project “Pavement Deterioration Models for Asphalt-Surfaced Pavements in New Zealand“ (Henning and Roux, 2008). It forms part of the overall New Zealand Long-term Pavement Performance (LTPP) programme. Earlier work completed between 2005 and 2007 resulted in pavement deterioration prediction models mostly for thin flexible chip sealed pavements. This paper documents the development of prediction models for dense graded asphalt (AC) surfaces and open graded porous asphalt (OGPA) surfaces. The research was successful in developing pavement deterioration models for crack initiation on AC surfaces and ravelling initiation on OGPA surfaces. Continuous probabilistic models were developed for both crack and ravelling initiation in order to predict the probability of the defect occurring. Fundamental and mechanistic prediction models, most commonly developed internationally, use details of both the bitumen and asphalt mix properties to forecast the long-term behaviour of the asphalt. Models developed during this research however, use data which is readily available on network level databases and can therefore be applied to asset management applications such as the New Zealand (NZ) dTIMS system (NZ’s nationally adopted Pavement Management System (PMS)). The models were tested on the NZ network data and had a significant success rate (up to 75%) in correctly predicting the behaviour of the respective surfaces. Further work required includes refining of the models based on the LTPP data that include bitumen property data. Although the developments are solely based on NZ data, there are a number of aspects applicable to the South African (SA) context. Firstly it presents a novel way of modelling the performance of asphaltic surfaces. When calibrated to SA conditions it is expected that these models would have similar success rate on SA highways. Secondly it demonstrates some practical implications of maintenance practices that are sometimes consider for SA conditions.