Seismic collapse prediction of frame structures by means of genetic algorithms

TL;DR

This paper introduces an automatic, efficient method using genetic algorithms to predict seismic collapse loads and failure modes of planar frame structures, with a new software implementation in Netlogo.

Contribution

It presents a novel application of genetic algorithms for seismic collapse prediction and develops a versatile, user-friendly software tool in Netlogo for structural analysis.

Findings

Genetic algorithms effectively estimate collapse load factors.

The software demonstrates versatility in classical and seismic load scenarios.

Parametric studies reveal influence of design parameters on collapse loads.

Abstract

This paper presents an automatic approach for the evaluation of the plastic load and failure modes of planar frames. The method is based on the generation of elementary collapse mechanisms and on their linear combination aimed at minimizing the collapse load factor. The minimization procedure is efficiently performed by means of genetic algorithms which allow to compute an approximate collapse load factor, and the correspondent failure mode, with sufficient accuracy in a very short computing time. A user-friendly original software in the agent-based programming language Netlogo, here employed for the first time with structural engineering purposes, has been developed showing its great versatility and advantages. Many applications have been performed both with reference to the classical plastic analysis approach, in which all the loads increase proportionally, and with a seismic point of view considering a system of horizontal forces whose magnitude increases while the vertical loads are assumed to be constant. In this latter case a parametric study has been performed aiming at evaluating the influence of some geometric, mechanical and load distribution parameters on the ultimate collapse load of planar frames.