Robust Optimization of the Steel Single Story Frame

Abstract

In contemporary design practices, building structures are expected to not only meet safety requirements but also be optimized. However, optimal designs can be highly sensitive to random variations in model parameters and external actions. Solutions that appear effective under nominal conditions may prove inadequate when parameter randomness is considered. To address this challenge, the concept of robust optimization has been introduced, which extends deterministic optimization formulations to incorporate the random variability of parameter values. In this study, we demonstrate the applicability of robust optimization in the design of building structures using a simple orthogonal frame as an example. The static-strength analysis is conducted based on the displacement method, utilizing second-order theory. To assess the safety level of the steel frame, a preliminary evaluation is performed by determining the reliability index and failure probability using the Monte Carlo Method. Robust optimization is then employed, leveraging the second-order response surface. Experimental designs are generated following an optimal Latin hypercube plan. The proposal of a mathematical-numerical algorithm for solving the optimization problem while considering the random nature of design parameters constitutes the innovative aspect of this research.