Multi-objective optimization assisting design of sandwich panels

Abstract

In the scope of a R&D project, the design of three new lightweight flooring panels with distinct characteristics was performed, namely: i) two steel face sheets, steel webs and polyurethane (PUR) foam as core system; ii) steel face sheets, glass fiber-reinforced polymer (GFRP) webs and PUR foam core system; iii) two steel face sheets and two steel outer webs enclosing a balsa wood core. The design of these panels included optimization procedures. The optimization method selected was a multi-objective optimization genetic algorithm (GA) which has proven to be well suited to solve this class of problems. The multi-objective function includes the minimization of the i) cost, ii) weight and iii) environmental impact. The objective functions are generally conflicting meaning that the minimization of one of the functions prevents the simultaneous minimization of the other ones. The definition of the multi-objective GA is presented and the implementation of its essential features, namely selection method, diversity maintenance and elitism is discussed. Boundary conditions are imposed so that the population will represent a feasible solution to the problem. These boundary conditions consist in the analytical formulation of serviceability, ultimate limit states and thermal transmittance verifications required by the building codes used in the design of the panels. The present paper deals with the introduction to all the aspects of the problem, then introduces, as example, the panel with two steel face sheets, steel webs and PUR foam as core system.