Sensitivity analysis of the impact of construction condition control factors on pavement performance

Authors:

Abstract
Non-uniform compaction temperature and asphalt content during construction are among the primary causes of early-stage pavement distress. To address this problem, a literature survey was conducted to characterize the field distributions of compaction temperature and asphalt content. Based on these distributions, an L-shaped experimental design was developed to evaluate the high-temperature stability, low-temperature crack resistance, water stability, and fatigue durability of AC-16 SBS-modified asphalt mixtures across the identified distribution ranges. A bivariate Gaussian surface model was then constructed to predict each performance indicator as a function of both variables, achieving R² of 0.84-0.99 and mean absolute percentage error (MAPE) below 5% for four of the five indicators. Monte Carlo simulations were combined with a global normalized sensitivity index to quantify the sensitivity of each indicator to variations in compaction temperature and asphalt content. Results show that compaction temperature monotonically enhanced all indicators except retained Marshall stability, which peaked at 160°C. Asphalt content exhibited non-monotonic behavior with respect to most pavement performance properties, and the optimal values vary. Sensitivity analysis revealed that raising the lower limit of compaction temperature is the dominant means of improving high-temperature stability, low-temperature crack resistance, and water stability. Reducing the coefficient of variation of asphalt content is the primary means of enhancing fatigue durability. These findings provide a quantitative framework for refined construction quality control of asphalt pavements.

This publication has 15 references indexed in Scilit: