A POD-based methodology for structural finite element model updating

This paper presents an efficient Proper Orthogonal Decomposition (POD) based methodology for structural finite element model updating problem. In the proposed methodology, the vibrational responses of the system are projected into proper modal coordinates and then the Power Spectrum Density (PSD) Functions of the projected data are used to establish the objective function. Since the new objective function integrates the effects of frequencies and damping of the structure in addition to its Proper Orthogonal Modes (POMs), it performs better than other POD-based objective functions, especially in cases of using ambient vibration data. Also, an effective algorithm is introduced to decide how many and which of the POMs should be selected for defining the objective function to minimize noise effects. Unlike the conventional energy criterion that has been widely used in previous research, the new algorithm provides a good view of noise distribution between POMs. Therefore, in severe noisy conditions, it can be used as a useful tool to detect affected POMs and remove them from the calculations. Several numerical case studies and a full-scale experimental model were used to evaluate the effectiveness and capability of the proposed methodology. The case studies showed that the POD-based methodology gives encouraging results even in cases of using ambient vibration or high noisy vibration data.