模态参数识别的特征系统实现算法:研究与比较

COMPARISON OF DIFFERENT VARIANTS OF THE EIGENSYSTEM REALIZATION ALGORITHM IN MODAL PARAMETER IDENTIFICATION

以 GARTEUR飞机模型的飞行颤振和地面共振试验为例 ,对模态参数识别的特征系统实现算法的 4种变形 :特征系统实现算法 (ERA)、快速 ERA(FERA)、相关 ERA(ERA/ DC)和快速相关 ERA(FERA/ DC) ,进行了深入的理论分析和仿真研究。结果表明 :与 ERA相比 ,ERA/ DC的阻尼识别精度在 1 0 %和 3 0 %噪声下可以提高 6～ 60倍 ;FERA的识别速度可以比 ERA提高 4～ 1 0倍 ,且精度较高 ;而 FERA/ DC的识别速度可以比 ERA/ DC提高 3～ 5倍 。

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.