Health Monitoring Of National Highway Network Using HDM-4 and Genetic Programming To Develop Road Maintenance Management System

Abstract

With the increasing traffic loads on National highways of India, pavements are deteriorating at a faster rate leading to premature failure. In addition to this lack of scientific road management system leads to lower levels of serviceability and unreliable road network in the long term. Therefore, in order to maintain the highway network in good condition the road administration should focus on long lasting and economical road maintenance solutions. The research study focuses on the development of road management system for high volume roads using calibrated HDM-4 model. Long term pavement performance under various maintenance strategies has been measured in terms of roughness progression using HDM. The most appropriate maintenance strategy has been identified using the life-cycle cost analysis based on NPV/COST ratio parameter. GP system has been configured to develop four distress prediction models i.e., roughness, ravelling, cracking and rutting. Adequacy of GP models has been measured using simple linear regression analysis. Statistical significance of roughness model has been evaluated using student’s T-Test. Variability in the output results of the two deterministic models i.e., HDM and GP has been computed by comparing the difference between predicted and observed roughness behaviour. Prediction models play a crucial role in development of RMMS for systematic technical as well as economic appraisal of road projects. Future modelling of pavement behaviour and cost streams related to various maintenance activities will assist the highway planners and road agencies in timely monitoring and conditioning of roads by adopting suitable management framework.