Investigation of loading features effects on resilient modulus of asphalt mixtures using Adaptive Neuro-Fuzzy Inference System

• Resilient modulus (Mr) of asphalt mixes was studied using proposed ANFIS models.
• The Mr under haversine loading was more susceptive to loading time variations.
• The Mr variation due to change of R/L ratio from 9 to 30 was insignificant.
• The Mr under square loading reduced up to 40% of Mr under haversine one at 40 °C.
• Using ANFIS models, Mr under various loading conditions can be estimated properly.

The aim of this study is to explore the effects of loading features on resilient modulus (Mr) of asphalt mixtures using Adaptive Neuro-Fuzzy Inference System (ANFIS). The resilient modulus of asphalt mixture samples were determined using indirect tensile test (IDT) under haversine and square waveforms at four temperatures (5, 15, 25, and 40 °C). Applied loading times were 50, 100, 300, 600, and 1000 ms with the ratio of rest period to loading time (R/L) equal to 4, 9, and 30. Two ANFIS models were developed for two loading waveforms. In ANFIS modeling, temperature, loading time and R/L are the parameters for the input layer and the resilient modulus is the parameter for the output layer. With application of developed models, the effects of loading parameters on Mr were examined. The results showed that Mr increased by reduction of temperature, decrease of loading time and slightly by reduction of R/L. Besides, it is realized that predicted resilient modulus is closely relevant to the measured one and prediction capability of the ANFIS models was satisfactory that can be prevented from expensive and time-consuming laboratory tests.