Model updating using uncorrelated modes

Abstract

Abstract FE model updating techniques are used to update dynamic FE models of structures in the light of modal test data. Iterative methods of model updating that update a set of chosen parameters of the model, so as to reduce the difference between the natural frequencies and the mode shapes of the FE model and the corresponding quantities obtained through a modal test on the structure, are probably the most widely used methods. Once experimental modal data has been identified, a necessary prior step, before updating can be carried out, is that of establishing the correspondence between the FE model modes and the experimentally identified modes. It is however experienced that, many a times a situation is encountered where not all of the modes identified through an experiment can be correlated with certainty with those predicted by the FE model and some experimental modes may be left uncorrelated. There could be several reasons for this lack of correlation as identified in the paper. But the consequence is that such uncorrelated modes cannot be used in FE model updating using existing iterative methods based on modal data even when they form valid known pieces of information about the structure. This is a disadvantage since it reduces the quantity of experimental data available for model updating and hence makes the updating process less effective in yielding an updated model that is a closer representation of the structure. This paper identifies this as a limitation of the existing iterative methods of model updating based on modal data and puts forward a notion of FE model updating using uncorrelated modes. The paper proposes a solution to overcome this limitation in the form of a new method of FE model updating that accepts both correlated as well as uncorrelated modes for updating. This is in contrast to all the current iterative modal data based methods of model updating that are based on the assumption of availability of correlated mode pairs and hence cannot use uncorrelated mode shapes and corresponding natural frequencies in the updating process. Formulation of the proposed method is described followed by a couple of numerical examples based on a beam structure to validate the method. The robustness of the method in the presence of simulated noise is also studied. Another numerical example of a more complex F-shape structure is also presented. The method is then validated though an experimental study. The proposed method is found to successfully update an FE model yielding correct estimates of the updating parameters in the presence of uncorrelated modes.