Otimização de projeto de viadutos de múltiplas longarinas considerando critérios ambientais e econômicos

Abstract

Brazil lacks investments in road infrastructure, as it is responsible for more than half of the volume of cargo transportation in the country. When addressing the topic of road infrastructure, one of the main variables is the construction and maintenance of viaducts. The design of such structures is traditionally carried out through a trial-and-error process that is no longer justified given the available computational resources and optimization methodologies. Traditionally, the main objective of structural design has been to minimize construction costs while ensuring structural efficiency. However, concerns about building a more sustainable future have shifted societal priorities, making sustainability a fundamental aspect as well. Ecological and durable structures are increasingly demanded. From this perspective, heuristic optimization methods provide an alternative to experience-based pre-dimensioning. The development of new construction materials, design techniques, and sustainable criteria motivates the need for a methodology that can automatically and accurately consider all these aspects. This research proposes a methodology for the efficient design of viaducts composed of pre-stressed and precast I-girder, rectangular cross beams, circular columns, and deep foundations using caisson-type piles, considering both structural efficiency and sustainability criteria. This research integrates cost, environmental impact, and durability in viaduct design. The proposed methodology provides a set of multi-objective compromise solutions among the involved variables, returning the best geometry, concrete type, and prestressing level. The performance of the MOPSO (Multiple Objective Particle Swarm Optimization), NSGA-2 (Nondominated Sorting Genetic Algorithm II), and SPEA2 (Strength Pareto Evolutionary Algorithm 2) metaheuristics was tested, and MOPSO proved to be the most effective in addressing the nature of the problem at hand. The parameters of MOPSO were calibrated using the Taguchi method, resulting in an average hypervolume increase from 2.211⋅10^9 to 2.222⋅10^9 in a proposed problem. By applying the developed method to two case studies, improvements were achieved compared to the original design, including up to a 23.2% reduction in construction costs, a 23.5% decrease in CO2 emissions, and a 540.7% increase in project lifespan when evaluating the individual performance of each objective function. This methodology has been implemented in a computer program available at https://www.optimusviaduto.ufscar.br, along with its source code.