Reducing the environmental cost of concrete constructions

Abstract

Environmental sustainability is seldomly considered at the design phase of concrete structures, which mainly concerns optimizing the structural mechanics properties. In this project we propose and implement a hybrid optimization protocol that solves the structural mechanics optimization along with minimization of environmental costs. These hybrid method consists of two optimization phases: the first gives a fast and crude estimation of the concrete resistance based on a data assimilation technique, the second refines the results with more expensive predictive modelling. After comparing our protocol against various existing optimization methods we conclude that our hybrid optimization leads to more sustainable designs without a compromise on mechanical properties. Surprisingly, benchmarking our technique on several real world datasets revealed that having data-assimilated preconditioner not only improves environmental costs but also leads to designs with better mechanical properties in less time.