DEVELOPMENT OF TRANSITIONAL PROBABILITY MATRIX MODEL FOR PAVEMENT DESIGN

By

Author

Presented To

Department of Engineering

ABSTRACT
A model of transitional probability matrix for Mechanistic-Empirical Pavement Design was developed with the aim of tarnishing the traditional method obtained from the judgment of an expert panel. The model was developed using fatigue and rutting as the critical factors in pavement failure. Model performance was assessed using first order reliability method (FORM) with the concept of genetic search algorithm developed with Matrix Laboratory(MATLAB). The proposed model allowed the distressed model to be model individually so that immediate measures for correcting deficiencies can readily be identified. The model formulation was followed by applying it into a real data. The data consist of three layered Hot-Mix asphalt (HMA) consisting of 100mm surfacing underlay by 200mm base course, 250mm sub base which is purported to have a life span of 15 years. Result indicate a decrease in system transitional probability of about 3% and 6% for light and heavy traffic considered respectively with age as against the constant homogeneous Markov-chain Matrix. The components probabilities also trends the same. Therefore, the assumption of time-invariance matrix of transitional probability was disregarded. The performance of the original pavement structure (100mm) was compared to 125mm and 150mm surfacing respectively. The result shows that the heavy traffic which are characterized by larger axle load and higher truck percentage has lower transitional probability than the lighter one because of its higher damages on the pavement structures both for the component and system models.