Mechanical Performance and Mechanistic-Empirical Evaluation of Tropical Soil Stabilized with Reclaimed Asphalt Pavement for Flexible Pavement Base and Subbase Layers

The increasing generation of reclaimed asphalt pavement and the consumption of virgin aggregates in road construction have encouraged the development of alternative materials for pavement base and subbase layers. However, the mechanical response of unbound mixtures containing tropical soils and reclaimed asphalt pavement, particularly under repeated loading, remains insufficiently documented for regional pavement design. This study evaluated a residual tropical soil derived from migmatitic paragneiss and three mixtures containing soil, coarse-grained gneiss, and reclaimed asphalt pavement. Mixture M1 consisted of 30% soil and 70% coarse-grained gneiss, M2 contained 30% soil, 35% coarse-grained gneiss, and 35% reclaimed asphalt pavement, and M3 consisted of 30% soil and 70% reclaimed asphalt pavement. Resilient modulus tests were conducted on the soil and all three mixtures. Based on the resilient response and the objective of maximizing waste incorporation, M3 was selected for permanent deformation testing and Shakedown analysis together with the natural soil. The laboratory parameters were subsequently applied in comparative mechanistic-empirical simulations using MeDiNa. Among the evaluated mixtures, M2 had the highest mean resilient modulus (292.60 MPa), followed closely by M3 (291.48 MPa). Because this difference was only approximately 0.38%, M3 was selected for permanent deformation testing and Shakedown analysis based on its comparable resilient response, complete replacement of virgin coarse aggregate, and maximum RAP incorporation. The permanent deformation results subsequently showed that M3 accumulated less deformation than the natural soil under the investigated stress conditions. For the lower traffic level, placing M3 in the base layer resulted in predicted rutting and cracked area values that were 11.29% and 28.38% lower, respectively, than those obtained when M3 was used in the subbase layer. These results represent relative trends between the simulated structures rather than absolute field-performance predictions because the MeDiNa transfer functions were not calibrated using local monitoring data. This study contributes regional mechanical parameters for a non-lateritic sandy soil and demonstrates the combined use of resilient modulus, permanent deformation, Shakedown analysis, and mechanistic-empirical simulation to assess soil–RAP mixtures while maximizing the replacement of virgin aggregate.