REMOTE IDENTIFICATION OF IMPACT FORCES ON LOOSELY SUPPORTED TUBES - PART 2 - COMPLEX VIBRO-IMPACT MOTIONS

Citation
J. Antunes et al., REMOTE IDENTIFICATION OF IMPACT FORCES ON LOOSELY SUPPORTED TUBES - PART 2 - COMPLEX VIBRO-IMPACT MOTIONS, Journal of sound and vibration, 215(5), 1998, pp. 1043-1064
Citations number
54
Categorie Soggetti
Acoustics
ISSN journal
0022460X
Volume
215
Issue
5
Year of publication
1998
Pages
1043 - 1064
Database
ISI
SICI code
0022-460X(1998)215:5<1043:RIOIFO>2.0.ZU;2-A
Abstract
In a previous paper, techniques were presented, based on response meas urements at remote locations, for the experimental identification of t he flexural wave-guide propagation parameters and for recovering the i mpact forces. Numerical simulations and experiments were presented, fo r simple isolated impacts. Those basic results showed that such an inv erse problem can be successfully attempted, and a good agreement was f ound between direct measurements and the remotely identified impact fo rces. However, when subject to flow-induced vibrations, the loosely su pported tubes display very complex rattling motions-with the impact-ge nerated primary waves completely immersed in countless wave reflection s travelling between the tube boundaries. As a consequence, the multip le-impact patterns of tube-support interaction are much more difficult to identify than isolated force spikes. In this paper, the authors mo ve a step further towards the identification of impacts for realistic tube vibrations. To deal with complex vibro-impact regimes, a signal-p rocessing technique is presented for separating the multiple wave sour ces? which uses the information provided by a limited number of vibrat ory transducers. This technique can be applied to both non-dispersive and dispersive waves and is therefore useful for all kinds of beam mot ions. Such a method is instrumental in separating the primary impact-g enerated flexural waves from severe background contamination. This ena bles the straightforward identification of complex rattling forces at a loose support. Extensive results are given in order to assert the nu merical conditioning of the technique used to identify the impact forc es, the optimal location of the transducers used in the identification procedure, and the sensitivity of the identification method to noise contamination. Overall, results are quite satisfactory, as the complex identified impact forces compare favourably with direct measurements. (C) 1998 Academic Press.