Closed-form solution based Genetic Algorithm Software: Application to multiple cracks detection on beam structures by static tests

摘要

In this paper a procedure for the static identification and reconstruction ofconcentrated damage distribution in beam-like structures, implemented in adedicated software, is presented. The proposed damage identification strategyrelies on the solution of an optimisation problem, by means of a geneticalgorithm, which exploits the closed form solution based on the distributiontheory of multi-cracked beams subjected to static loads. Precisely, theadoption of the latter closed-form solution allows a straightforward evolutionof an initial random population of chromosomes, representing different damagedistributions along the beam axis, towards the fittest and selected as thesought solution. This method allows the identification of the position andintensity of an arbitrary number of cracks and is limited only by the amount ofdata experimentally measured. The proposed procedure, which has the greatadvantage of being robust and very fast, has been implemented in the powerfulagent based software environment NetLogo, and is here presented and validatedwith reference to several benchmark cases of single and multi-cracked beamsconsidering different load scenarios and boundary conditions. Sensitivityanalyses to assess the influence of instrumental errors are also included inthe study.