Space structures dynamic identification through output only approach

Paper number

IAC-06-C2.3.03

Author

Dr. Giuliano Coppotelli, University of Rome "La Sapienza", Italy

Coauthor

Prof. Paolo Gasbarri, University of Rome "La Sapienza", Italy

Year

2006

Abstract

Space missions involve the use of large and light weight structures such as antennas, sunshields, solar arrays etc. One of the main characteristics of large flexible space structures is the large uncertainty in the knowledge of space dynamics. Models of flexible appendages, for instance, are always only partly representative of the true space dynamics. Other characteristics such as damping factors, couplings between axes, number of flexible modes also have a great impact on spacecraft design. As far as control design is concerned, for example, the number of flexible modes, the uncertainties on the knowledge of their characteristics, such as resonance frequencies and/or damping factor motivate the evolution of sophisticated and dedicated control strategies, Refs. [ kubica, hanks].

It is well known that space environment imposes hard constraints in the design and control of rigid-flexible satellite because its dynamic behavior is quite different from its ground behavior. Generally, the study of the physical characteristics of large space structure components indicates that the dynamic modeling is an approximation to the actual system and can only be verified after the structure is in orbit and its response to disturbance behavior can measured. Furthermore due to the fact that complex structures (such as the International Space Station, ISS) could be viewed as a distributed parameter system having, theoretically, an infinite number of degree of freedom, it is necessary to perform a model reduction in order to achieve efficient computation, dynamic analysis and control design. In turn, this reduction introduces a high level of uncertainties in the mathematical model used for control design synthesis both for structural vibrations and attitude dynamic control of the whole spacecraft. In addition, structural parameters could suffer changes in their values during the mission life, modifying both the dynamic and static behavior of the whole structure. In this frame, a suitable approach capable to identify the dynamic model of the structure, in terms of the modal model, during the on-orbit-life time of the spacecraft is proposed. In the developed methodology, the modal parameters - natural frequencies, damping ratios, and mode shapes - are estimated by the analysis of the response of the structure only as a result of its “natural” loading environment such as for instance the one produced by gravity and gravity gradient effects. These modal parameters could be estimated assuming that the output response is due to a white noise excitation in the frequency range of interest. Within this main hypothesis, both time and frequency domain - based techniques have been developed in the recent years. Time domain techniques, based on the Stochastic Subspace Identification (SSI) method, allow the estimation of the modal model through a state-space formulation solved in the time domain by an orthogonal-projection approach, Refs. [ ssi, si]. In the frequency domain techniques, based on the Frequency Domain Decomposition (FDD), natural frequencies and mode shapes are estimated through the analysis of the spectral properties of the power spectral density matrix of the responses only, whereas the damping ratios could be evaluated, in time domain, by the logarithmic decrement technique, Ref. [ Brincker3]. Nevertheless, the previous approaches are not able to identify the whole modal model. Since the input level is not measured, the modal scale factors or generalized masses are not disposable. in recent studies, these participation factors have been obtained using sensitivity-based methods, Refs. [ deweer, Brincker1, coppotelli].

In this paper, a methodology aimed to identify dynamic parameters of a spacecraft structure is presented. The time responses of a large space structure during its normal orbit operation are first evaluated, eventually they are used for a subsequent output-only modal parameter identification. Moreover, the capability to identify structural changes by the measurement of the natural frequency shifts, is also investigated.

References

[kubica]

F. Kubica, 1998. New flight control laws for i large capacity aircraft. Experimentation on airbus A 340, 21th ICAS Congress Proceedings, Melbourne, Australia, 1998.

[hanks]

Hanks, B. R., 1988. NASA’s Control/Structures Interaction (CSI) Program, Proceedings of the Workshop on Computational Aspects in the Control of Flexible Systems, Williamsburg, VA, July 12-14, 1988. NASA TM 101578, Part 1, pp. 21-32.

[ssi]

P. Van Overschee, B. De Moor, 1996. Subspace Identification for Linear Systems, Kluver Academic Publisher.

[si]

Juang, J.N., 1994. Applied System Identification, Prentice Hall, Englewood Cliffs., New Jersey.

[Brincker3]

R. Brincker, Zhang L., P. Andersen, 2000. “Modal Identification from Ambient Response Using Frequency Domain Decomposition”, Proc. of 18 ^th IMAC.

[deweer]

J. Deweer, B. Dierckx,, 1999. Obtaining a Scaled Modal Model of Panel Type Structures Using Acoustic Excitation, Proc. of 17 ^th IMAC, pp. 2042−2048.

[Brincker1]

R. Brincker and P. Andersen, 2003 A Way of Getting Scaled Mode Shapes in Output Only Modal Testing, Proc. of 21 ^st IMAC, pp. 141−145.

[coppotelli]

G. Coppotelli, 2003. Identification of Frequency Response Functions by “Output-Only” Experimental Data, in Proceedings of XVII Associazione Italiana di Aeronautica e Astronautica, Rome, ITALY.

Abstract document

IAC-06-C2.3.03.pdf

Manuscript document

IAC-06-C2.3.03.pdf (🔒 authorized access only).

To get the manuscript, please contact IAF Secretariat.