Reliability-Based Pavement Roughness Progression Modeling Using Bayesian Approach

Abstract

This paper aims to develop reliability-based roughness progression models for bituminous pavements across different climatic zones in India. The methodology involves dividing the Indian geographical region into six climatic zones with specific climatic characteristics that influence pavement deterioration. This delineation is achieved using K-means clustering and considers parameters such as temperature, rainfall, and humidity. The roughness progression models and reliability interpretations are developed through a Bayesian regression framework. Initially, International Roughness Index (IRI) progression models are created for three climatic zones: hot and dry, warm and humid, and moderate. These models are based on time series data from 2015 to 2016, gathered from 130 pavement sections in each zone. To validate the models, additional time series data from 2016 to 2017 are utilized. The forecast results indicate that the hot and dry zone exhibits the highest IRI progression rate for specified causal variable values, followed by the warm and humid zone and the moderate zone. The adoption of the Bayesian regression framework provides probabilistic parameter distributions for model coefficients, enabling the assessment of model-level reliability. The study demonstrates that, by adjusting the quartile of causal variables, the overall reliability of the model can be improved, thereby reducing the deviation from actual IRI values. Moreover, it is possible to evaluate variable-level reliability by examining the influence of individual variables on roughness progression.