On crack detection in curved beams using change of natural frequency

TLDR: In this article, an inverse method for the purpose of identifying the parameters of a single-edge crack in a curved beam is investigated, where the beam is divided into some elements and the governing equations, the continuity and boundary conditions, are exerted on these elements to create an eigenvalue problem which is solved to obtain the required frequencies.

Abstract: In this study, an inverse method for the purpose of identifying the parameters of a single-edge crack in a curved beam is investigated. This method requires both numerical and experimental natural frequencies. The differential quadrature element method (DQEM) is introduced to get the numerical ones. For the purpose of implementing the DQEM the beam is divided into some elements and the governing equations, the continuity and the boundary conditions, are exerted on these elements to create an eigenvalue problem which is solved to obtain the required frequencies. A rotational spring whose stiffness is a function of crack depth is used to model the crack. An objective function with the variables of crack location and depth which is the weighted mean squared error between experimental and numerical frequencies is defined. The artificial bee colony algorithm (ABC) an optimization algorithm inspired by the natural foraging behavior of honeybees is used to minimize the aforementioned objective function and obtain crack parameters. To investigate the accuracy of this non-destructive, fast and easy method, an experimental modal analysis test is conducted on a cracked beam. The results indicate that the position and depth of the crack can be determined successfully using this method. This test is done five times with a different crack depth each time, to show the efficiency of the method for all cracks even small ones. A brief investigation is also done on the effect of crack location and depth on the natural frequencies.