Damage Detection in Steel Beams Using Generalized Flexibility Quotient Difference Based Damage Index and Artificial Neural Network.

Background: Failure of the structure may be avoided if, the damage is noticed at an early stage and proper remedial work is performed. In this context, this paper presents a detection of structural damage in steel beams, which are major load carrying members in bridges and buildings. Method: In the present study, the Generalized Flexibility Energy Quotient Difference Index (χ) is determined at various locations of the beam by introducing the damage by means of a cut by 0.5 mm to 2.5 mm at an interval of 0.25 mm. In practice, it is very difficult to capture all the modes of vibration. Hence, the method followed uses the first mode of vibration only for quantifying the magnitude of damage and its location. The validation is done between numerical simulations and experimental works for the estimation of first frequency. The efficiency of the method is tested by inducing the damage very close to the support. In ANN, the depth of cut at various locations of the beam and χ is utilized in the input layer. Results: The predicted depth of cut or severity of damage from the output layer of ANN model is compared with the actual depth of cut induced in the beam. Finally, the polynomial equations are formulated for the damaged elements, which are the major contribution of the present study. These equations can be used as a ready reckoner to estimate the actual depth of cut in the beam. This approach is validated through a beam with fixed ends and two-span beam, and the results indicate that this method is very much beneficial in preventing the structural failure. [ABSTRACT FROM AUTHOR]