Avaliação de modelos matemáticos de comportamento para o módulo de resiliência de materiais de pavimentação

Abstract

Currently, Brazil boasts one of the largest road networks globally. Therefore, an effective study of the materials comprising the pavement layers is of paramount importance for improved design, durability, and maintenance of highways. To better understand pavement design, it is crucial to be acquainted with the properties of the materials involved, such as the resilient modulus. Thus, the main objective of this study is to analyze mathematical models for the resilient modulus applied to pavement materials. The repeated triaxial load test is the primary method for obtaining resilient modulus values for these materials. Data from confining stress, deviator stress, and calculated resilient modulus are utilized, originating from Ferreira's (2008) database and tests conducted at the UFJF Pavement Laboratory. Fourteen mathematical models for resilient modulus were applied to 341 samples from Ferreira's (2008) optimized database, meaning without outliers, with the respective analysis of three statistical parameters to assess their suitability: R-square, adjusted R-square, and Mean Squared Error. Similarly, the same procedure was carried out in the Pavement Laboratory database, totaling 33 test samples, and in a third database obtained from the second. Ultimately, a ranking was established among the models. It is concluded that, for Ferreira's (2008) database, the Tam and Brown (1988) model demonstrated the best performance, emphasizing its relevance and suggesting greater adherence by Brazilian designers in their analyses. Furthermore, in the Pavement Laboratory test database, the Witczak and Uzan (1988) model showed superior performance, also suggesting its possible applicability in pavement design projects.