ANN and DoME to Predict the Moisture Damage Resistance of HMA with RCA

Abstract

[Abstract]: The production of hot-mix asphalt using recycled concrete aggregates from construction and demolition debris as raw material could support the circular economy and the development of more environmentally friendly infrastructure. A crucial characteristic of HMA manufactured with RCA is its moisture resistance. Careful research should be done to ensure satisfactory performance. The experimental inquiry might be perfectly complemented by a mathematical approach. To predict the indirect tensile strength value and the tensile stress ratio as a function of the study parameters (wet or dry state, bitumen per-centage, and RCA percentage), three models—linear, artificial neural networks (ANN), and development of mathematical expressions (DoME)—were proposed. It was possible to get mathematical expressions. The key finding of this study is that the DoME approach led to more accurate estimations of the ITS values. DoME’s primary advantage is that it returns a simpler expression.