Abstract:

Different variants of the Eigensystem Realization Algorithm(ERA), i.e, ERA, Fast ERA(FERA), ERA with Data Correlation(ERA/DC), Fast ERA/DC(FERA/DC), are compared for accuracy, robustness and identification speed. Comparisons are made on the basis of a popular aircraft model, GARTEUR. Ground and flutter tests of the model are simulated in detail. Results suggest that identification quality of the two DC algorithms is clearly superior to the others, with the accuracy of the identified modal damping improved by 6-60 times, and that FERA accelerates the process of ERA 4-10 times and likely renders more accurate results. As to FERA/DC, its speed is 3-5 times that of ERA/DC, while the accuracy has not deteriorated.

Key words: modal parameter identificat ion, eig ensy stem realizatio n, co r relatio n algo rithm, g ro und t est, flutt er test