Pavement Performance Prediction Modelling for Urban Roads in Coimbatore City
Urban roads deteriorate from cracks, potholes, and rutting due to traffic, rain, and poor materials, making timely repairs expensive if not predicted early. This project studies Coimbatore city roads to calculate current condition scores (Pavement Condition Index - PCI) and predict future performance using simple math models—this is academic modeling based on field surveys, not city-wide implementation.
Teams surveyed road stretches for distress types (transverse/longitudinal cracks, rutting, potholes, raveling, depressions, patchwork) and rated severity (low/medium/high) per IRC 82, ASTM D6433, and MTO SP-024 standards. PCI (0-100 score) calculated via deduct curves, distress density weighting, or C-language/Excel codes combining severity, density, and weights. Linear regression and ARIMA models used historical PCI data over time to forecast decline.
Results classified sample roads as “Fair” (PCI 41-60) with common issues like 3-6mm cracks (15% density) and minor rutting; models predicted “Poor” (21-40) in 6-12 months without maintenance. Codes handled inputs like distress % to output PCI categories (Very Good to Very Poor) and future values, factoring traffic, age, climate, and repairs. Graphs showed steady PCI drop over semesters.
Findings prove simple models can prioritize fixes (e.g., repair at PCI<41), cut costs via proactive maintenance, and improve safety—but limited to surveyed stretches and basic data. Real use needs more years of data, sensors for live monitoring, AI for complex patterns, and GIS mapping. Future: integrate drones/ML for city-scale predictions linking PCI to budget needs.
← Back to Projects