Otimização multiobjetivo aplicada à seleção de projetos de misturas asfálticas porosas utilizando o conceito de dominância de soluções

Abstract

The Low Impact Development (LID) approach to rainwater management in urban areas has been gaining interest from the international community due to the growing urbanization of developing countries, the sealing of urban soil and, mainly, the changes in rainfall patterns caused by climate change. However, debates due to the lack of consensus on its conceptualization reinforce the creation of blind spots in practice. This disconnect between conceptualization and practice is a manifestation of the way in which dominant visions of urban development are imposed from the top down. Despite the criticism of LID, this research studied one of its most traditional techniques, the permeable pavements. The aim was to select designs for porous asphalt mixtures used as a surface coating for permeable pavements, based on mechanical and functional criteria. Fifteen porous asphalt mix designs were conceived, combining five different grain sizes and three different binder contents. Among five gradings, three were from national ranges and two from international ranges. In the first stage, the mechanical parameters studied were restricted of tensile strength, resilience modulus, cantabro loss mass and induced moisture damage tests. In the second stage, the functional performance was assessed with permeability, infiltration and water retention tests. The ponding time of the pavement surface was also evaluated. In the third stage, specific optimization problems, such as reconciling conflicting mechanical and functional objectives, were designed in different scenarios. Each problem was solved using an optimization model based on the concept of solution dominance, which consists of choosing one or more optimal solutions with the best trade offs between the multiple objectives to be optimized. Each problem was solved by selecting the most suitable mixtures for each proposed scenario. Some boundary conditions helped formulate problems with conflicting objectives involving permeable pavements, such as traffic levels, climate, temperature and road drainage conditions. Among the results of the mechanical tests, meeting the minimum tensile strength was an almost systematic problem, while stiffness was more influenced by the grain size of the mixture and mass loss by the binder content. With regard to moisture susceptibility, the mixtures obtained satisfactory responses, mainly due to the type of binder used. In terms of functional performance, the communicating voids in the mix were directly proportional to the permeability and, together with the grain size and binder content, influence its water retention capacity. The binder content and granulometry also influenced the infiltration rate and, consequently, the ponding time, but it is the thickness and anisotropy of the other pavement layers that determine how quickly they reach their potential infiltration capacity, which directly impacts their ponding time. As for the results of the optimizations, not just one, but a set of mixtures can be the solution to the same multiobjective problem and the decision depends on the boundary conditions. In general, porous mixtures with higher binder contents and maximum aggregate diameters of 12.5 mm stood out from the other mixtures available in the decision space.