Rheological Performance and Predictive Modeling of Waste Toner–Modified Asphalt Binders Using an Artificial Neural Network

Abstract The use of waste toner (WT) as an asphalt modifier supports sustainable pavement design, but its performance under construction-level thermal exposure remains insufficiently understood. This study evaluated the effect of extended rolling thin-film oven (RTFO) aging on the rheological and thermorheological behavior of VG10 binders modified with 8%–14% WT. Binder performance was assessed through Superpave grading, multiple stress creep recovery (MSCR), linear amplitude sweep (LAS), and time–temperature superposition (TTSP) analyses. WT improved high-temperature stiffness and rutting resistance up to a 12% dosage, increasing the performance grade from PG 58 to PG 70. At 14%, the improvements diminished and deviations from thermorheological simplicity appeared under prolonged aging, indicating limited compatibility at higher dosages. Fatigue evaluation showed that although WT generally reduced fatigue life, the 8% dosage provided the most balanced response. A predictive framework was developed, and a feedforward artificial neural network (ANN) outperformed multiple linear regression in predicting complex modulus and phase angle. The results demonstrate that WT can serve as an effective modifier when used within optimized dosage limits.