Multifactorial coupled asphalt aging processes: Evaluation and predictive analysis

The performance of asphalt deteriorates with aging. Analysis and prediction of asphalt aging under multifactorial environmental conditions have become a critical research focus in pavement engineering. Predicting the degree of asphalt aging is essential for determining the optimal utilization strategy of recycled reclaimed asphalt pavement (RAP) materials. This investigation employs an innovative approach that combines climate-specific accelerated aging tests with an artificial neural network (ANN). An indoor accelerated aging scheme for simulating complex environments (temperature, humidity, and radiation) was developed, and six climate sub-regions were delineated. The experiments place particular emphasis on the phase angle change rate as an aging indicator. A backpropagation neural network (BPNN) architecture was engineered to establish predictive correlations between environmental parameters and the change rate of phase angle (CRPA) using 228 asphalt samples. In addition, the aging mechanism was characterized by GPC, FTIR and four-component analysis. The test results demonstrate that high temperatures, high irradiance, and low humidity contribute to more severe aging, with asphalt aging primarily influenced by irradiance, followed by temperature and humidity. Phase angle is representative of rheological parameters and can be selected as a sensitivity index. The BPNN method was used to predict the degree of asphalt aging, and the R 2 value reached 91.72%. This study proposes using the phase angle as a judgment index for rapid assessment and prediction of aging, and achieves the prediction of aging degree under certain specific environmental conditions via an artificial neural network.