Developing New Empirical Formulae for the Resilient Modulus of Fine-Grained Subgrade Soils Using a Large Long-Term Pavement Performance Dataset and Artificial Neural Network Approach

OpenAlex İngilizce

Artificial neural network (ANN) has been successfully used for developing prediction models for resilient modulus (M r ). However, no reliable M r formula derived from these models has been proposed in previous studies, although engineers/researchers need empirical formulae for hand calculation of M r . Therefore, this study aimed to propose reliable empirical formulae for the M r of fine-grained soils using ANN. For this purpose, thousands of ANN models were developed using the long-term pavement performance (LTPP) and external datasets. The input parameters were the percentage of soil particles passing through #200 sieve (P200), silt percentage (SP), clay percentage (CP), liquid limit (LL), plasticity index (PI), maximum dry density ([ρ dry ] max ), optimum moisture content (w opt ), confining pressure (σ c ), and nominal maximum axial stress (σ z ). The ANN models were compared with several constitutive models. The results indicate that the constitutive models failed to predict the M r , and the best M r predictions were obtained by the ANN-C9 (P200, SP, CP, LL, PI, σ c , and σ z ), ANN-C10 (P200, SP, CP, [ρ dry ] max , w opt , σ c , and σz), and ANN-C11 (P200, SP, CP, LL, PI, [ρ dry ] max , w opt , σ c , and σ z ) models. Thus, the structures of these ANN models were formulated and proposed as the new empirical formulae for the M r of fine-grained soils. Sensitivity analysis was also performed on these ANN models. It was determined that (ρ dry ) max is the most influential parameter in the ANN-C10 model, and LL is the most influential parameter in the ANN-C9 and ANN-C11 models. On the other hand, σ c and σ z are the least influential parameters.