Study of Damage Detection and Quantification in Cantilevered Beams using Modal Strain Energy & Genetic Algorithm

Abstract

Structural health monitoring (SHM) is known to be a demanding area of research which has its applications in large scale load carrying structures like bridges, aircrafts, automobiles, offshore platforms, and submarines. Deteriorating and aging structures may require early detection and quantification of damage to ensure safety and service life. Various techniques based on modal parameters and optimization have been proposed in literature. The drawback of using conventional and deterministic optimization algorithms is that complete set of data and input parameters may be required to proceed. While GA can operate even with incomplete set of data and with minimum number of input modal parameters i-e mode shapes & natural frequencies. So, GA based methodology has the advantage in real life applications over deterministic and conventional optimization methods and therefore considered for damage detection in current study. Estimation of objective function is required for entire set of population of chromosomes at the end of each generation. Therefore, genetic algorithm (GA) is computationally expensive tool. The current study employs a multistage methodology to reduce the solution parameters of the GA. To compare the vibrational properties of the undamaged and damaged structures, modal analysis of the structure was conducted. The ratio of change in modal strain energy (MSECR) was used to localize damage and reduce the size of the solution space of GA. True damage percentage of identified damaged elements was calculated using a GA-based on mode shapes and natural frequencies. Numerical studies of cantilever beam have been carried out to validate the methodology. Numerical studies disclose multistage approach is found to be robust and rapidly convergent at a reduced computational cost.