Application of weighted average on modal parameters for damage detection algorithms: Case study on steel beam

摘要

This study verifies the use of the proposed weighting method by Fayyadh and Abdul Razak (2011a) for damage detection algorithms applied on cracked steel beam. The results of the proposed weighting method for two damage severity algorithms are presented. That is, one based on the natural frequencies and the other based on the mode shape vectors. In addition, the stiffness deterioration based on the mathematical calculation of the change in the second moment of inertia is presented. In order to demonstrate the significance and capability of this new method, the magnitude of damage was calculated from an experimental model on a simply supported solid steel beam. Modal parameters were first obtained for the undamaged state by performing modal testing on the un-cracked steel beam. Subsequently, damage was created by cutting the beam section at mid-span from the tension zone to induce different degrees of flexural damage. The crack was created with different depth such as 2, 5, 10 and 20 mm with a constant width of 2 mm. At each level of damage, the Eigen value analysis was repeated to obtain the modal parameters relevant to the degree of damage induced. Stiffness change based on frequency and mode shapes were obtained; the proposed weighting method was applied and compared to the normal averaging method. Dynamic algorithms were compared to the change in the flexural rigidity as based on the change in the second moment of inertia of the beam cross-section at the crack zone. Proposed method returns one value of stiffness deterioration based on the considered set of modes, without affecting the sensitivity of damage detection algorithms. Flexural rigidity results showed higher sensitivity compared to dynamic algorithms. © 2011 Academic Journals.