 |
The aeronautics industry have benefited from the use of numerical models to supplement or replace the costly design-build test paradigm. Dynamic models are commonly calibrated to obtain optimal fidelity to data. In order to improve calibration methods, we wish to take into account the unavoidable compensating effects between the parameters which lead to the nonuniqueness of the responses. An envelope-bound info-gap model will be used to explore the change in predictions as the the parameter values are allowed to vary for different horizons of uncertainty. The calibration is considered robust if an acceptable level of fidelity to data is obtained even in the presence of uncertainty. Our methodology is demonstrated on an example involving the dynamic response of a clamped plate with progressively reduced thickness.
