Comparative assessment of computational models for the effective tensile strength of nano-reinforced composites

Some of the most important industries, such as aerospace, automotive, among others, have stipulated new requirements for materials behavior that include high specific, mechanical, and thermal properties. According to this, nanocomposites have emerged to satisfy these requirements. However, manufacturing these nanocomposites implies cost and time-consuming problems that do not allow their use in technological applications; additionally, the lack of knowledge about the prediction of their mechanical properties is an obstacle to its technological implementation. Therefore, several studies have focused on the development of computational models to predict the mechanical behavior of nano-reinforced composites.  In the present work, a comparative assessment of the main computational models for predicting the tensile strength of nanocomposites is carried out. Firstly, a new taxonomy of these models is proposed, which allows identifying the main experimental variables, model evolution, and precision. With the categorization, computational algorithms are developed for these models for predicting the tensile strength of nanocomposites, accomplishing a comparative analysis of accuracy, robustness, and time-cost among them. The precision of these models is evaluated by deeming benchmark experimental works focused on the tensile strength of nanocomposites. The results obtained demonstrated a minimum relative error of 44.7%, 10.1%, and 10.6% for First-Generation, Second-Generation, and Third-Generation models, respectively. Moreover, linear and non-linear behaviors were found in the evaluated models, being coherent with the number and kind of parameters required for the assessment.