ON THE ACCURACY REQUIRED OF EXPERIMENTAL DATA FOR FINITE ELEMENT MODEL UPDATING

Abstract

In this paper, the determination is considered of the required experimental accuracy that must be attained when updating finite element models with use of measured vibration test data. A theoretical basis is developed for FRF-based updating techniques, as these use measured data directly. It is shown that a well-defined relationship, that can be expressed as a characteristic function, exists between the system's properties, the correction matrices and the actual amount of experimental noise. The formulation is then applied to the standard response function updating formulation, where the element mass and stiffness matrices are corrected by using a single multiplier, the so-called p-value. In the presence of noise, the convergence of the updating algorithm is shown to be dependent on a number of conditions which arise from two distinct cases: one convergent and the other divergent. The findings are illustrated in the case of a 3-D space frame, and the efficacy of the proposed characteristic function is discussed in some detail. Finally, a way of selecting the optimum excitation frequency values is presented as a means of relaxing the minimum experimental accuracy.