Pavement Life Prediction Using MATLAB and Machine Learning

A student-laptop project that predicts how long a flexible (bituminous) road pavement will last and uses the prediction for maintenance planning. A mechanistic-empirical simulator built on IRC:37-2018 (Odemark equivalent thickness with Boussinesq strains, IRC fatigue and rutting equations, cracking, rut depth and roughness month by month for 40 years) generates a synthetic network of 3,000 road sections with 22,064 condition surveys. Machine-learning models then predict the design-stage service life (test error 3.003 years against 11.859 years for the IRC:37 structural life alone), give an 80 % prediction interval, and estimate the remaining life from condition surveys (2.098 years). Survival analysis, Markov chains and a 30-year life-cycle cost analysis turn the predictions into a maintenance programme. A web application with five database options stores sections, surveys, assessments and plans; the same algorithms are also written as MATLAB scripts. All data are synthetic.

Category: Civil Engineering / Highway and Pavement Engineering / Pavement Management / Machine Learning / MATLAB and Python

Civil Engineering Project C08 from ProTechSC Academy, with graphical abstract, video demonstration and project guidance.