In this thesis we propose a novel methodology to perform modal analysis starting from acoustic measurements using near-field acoustic holography. To do this, our technique combines the Equivalent Source Method (ESM) with the advantages of searching the solution in a sparse domain using some a priori knowledge. In particular, a sparse solution is searched in a space spanned by a pre-computed dictionary. The entries of the dictionary are the equivalent source weights. Therefore, the solution to the problem is obtained searching the better linear combination of the equivalent source weights in the dictionary, according to the acoustic measurements. Using a dictionary based approach makes possible to significantly reduce the number of required microphones, therefore it is possible to reduce the cost of the microphones acquisition system, to reduce the wiring complexity and the problem related to the microphones standing. Moreover, the dictionary approach intrinsically makes the technique robust against measurement noise. Thanks to this our techniques does not require that the acoustic measurements have to be acquired into an anechoic chamber. To validate our method we use real acoustic pressure measurements acquired in the near-field of a vibrating aluminium rectangular plate. The measurements are acquired in a non-ideal room (i.e., non anechoic), therefore some amount of reverberation and background noise are present. Experimental results confirm that our technique can reconstruct the vibration modes of the plate using a limited number of microphones. Moreover the methodology shows a robust behaviour in the presence of measurement noise. We also present simulative results related to a vibrating violin top plate, to test the suitability of the method for analysing complex surfaces. Results are promising as they show that our method outperforms the classical ESM technique, even using a limited number of microphones and in the presence of measurement noise.
In questa tesi proponiamo una nuova metodologia per effettuare analisi modale partendo da misure acustiche mediante olografia acustica in campo vicino. Per fare questo, la nostra tecnica combina il metodo delle sorgenti equivalenti (ESM) con i vantaggi dati dal cercare la soluzione in un dominio sparso sfruttando una conoscenza pregressa. In particolare, viene cercata una soluzione sparsa in uno spazio identificato da un dizionario pre-calcolato. Le componenti del dizionario sono i pesi delle sorgenti equivalenti. Pertanto, la soluzione al problema è ottenuta individuando la miglior combinazione lineare possibile di pesi delle sorgenti equivalenti presenti nel dizionario, in accordo alle misure acustiche effettute. Un approccio basato su dizionario permette di ridurre significativamente il numero di microfoni necessari, pertanto è possibile ridurre i costi del sistema di acquisizione audio, ridurre la complessità del cablaggio e ridurre i problemi relativi al supporto sul quale ancorare i microfoni. Inoltre, questo tipo di approccio, intrinsecamente rende la tecnica proposta robusta rispetto al rumore di misura. La nostra tecnica non richiede quindi che le misure vengano effettuate in una camera anecoica. Per la validazione del nostro metodo usiamo misure di pressione acustica, acquisite in campo vicino, relative ad un pannello vibrante rettangolare di alluminio. Le misure sono state acquisite in un ambiente non ideale (i.e., non anecoico), pertanto un certo livello di riverberazione e rumore di fondo sono presenti. I risultati sperimentali confermano che la nostra tecnica riesce a ricostruire i modi di vibrazione del pannello usando un numero limitato di microfoni. Inoltre la metodologia proposta mostra un comportamento robusto in presenza di rumore di misura. Infine presentiamo i risultati ottenuti da simulazioni effettuate su una tavola di violinio vibrante, per testare l'idoneità del nostro metodo per l'analisi di superfici complesse. I risulati sono promettenti, infatti dimostrano che il nostro metodo dà prestazioni migliori rispetto al metodo ESM, sia usando un numero limitato di microfoni, sia in presenza di rumore di misura.
Modal analysis of arbitrary surfaces through dictionary-based near-field acoustic holography
VARINI, MASSIMO
2015/2016
Abstract
In this thesis we propose a novel methodology to perform modal analysis starting from acoustic measurements using near-field acoustic holography. To do this, our technique combines the Equivalent Source Method (ESM) with the advantages of searching the solution in a sparse domain using some a priori knowledge. In particular, a sparse solution is searched in a space spanned by a pre-computed dictionary. The entries of the dictionary are the equivalent source weights. Therefore, the solution to the problem is obtained searching the better linear combination of the equivalent source weights in the dictionary, according to the acoustic measurements. Using a dictionary based approach makes possible to significantly reduce the number of required microphones, therefore it is possible to reduce the cost of the microphones acquisition system, to reduce the wiring complexity and the problem related to the microphones standing. Moreover, the dictionary approach intrinsically makes the technique robust against measurement noise. Thanks to this our techniques does not require that the acoustic measurements have to be acquired into an anechoic chamber. To validate our method we use real acoustic pressure measurements acquired in the near-field of a vibrating aluminium rectangular plate. The measurements are acquired in a non-ideal room (i.e., non anechoic), therefore some amount of reverberation and background noise are present. Experimental results confirm that our technique can reconstruct the vibration modes of the plate using a limited number of microphones. Moreover the methodology shows a robust behaviour in the presence of measurement noise. We also present simulative results related to a vibrating violin top plate, to test the suitability of the method for analysing complex surfaces. Results are promising as they show that our method outperforms the classical ESM technique, even using a limited number of microphones and in the presence of measurement noise.| File | Dimensione | Formato | |
|---|---|---|---|
|
2016_07_Varini.pdf
Open Access dal 08/07/2019
Descrizione: Testo della tesi
Dimensione
5.18 MB
Formato
Adobe PDF
|
5.18 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/123430