Jtam.dvi JOURNAL OF THEORETICAL AND APPLIED MECHANICS 49, 2, pp. 313-326, Warsaw 2011 EXPERIMENTAL CHARACTERIZATION OF HYBRID NOISE CONTROL SYSTEME ACTING IN ACOUSTIC DUCT Barbara Tiseo, Antonio Concilio, Salvatore Ameduri, Antonio Gianvito C.I.R.A, The Italian Aerospace Research Centre, Capua (CE), Italia e-mail: b.tiseo@cira.it; a.concilio@cira.it; s.ameduri@cira.it; a.gianvito@cira.it In this investigation,a soundabsorbingmaterial is coupledwithanactive surface to reduce sound pressurewithin an acoustic duct. In this way, an enhancement of the passive foam performance can be observed because of further acceleration of the fluid particles within the foam pores. The experimental activity implementsnoise control strategies to increase dampingwithin the cavity through acoustic foamwith a flexible backing plate, acted by piezoelectric actuators. Keywords: soundabsorbingmaterial, acousticduct, foam,active surface, feedback control 1. Introduction There are two fundamental approaches to control acoustic noise: passive and active. The traditional approach to acoustic noise control uses passive tech- niques to attenuate undesired noise. These passive treatments are valued for their high attenuation over a broad frequency range. However, they are relati- vely large, heavy, costly, and ineffective at low frequencies,making the passive approach to noise reduction often impractical. In an effort to overcome these problems, considerable interest has been shown in active noise control. The active noise control system contains an electroacoustic device that cancels the unwanted sound by generating an antinoise of equal amplitude and opposite phase. The successful application of active control is determined on the ba- sis of its effectiveness compared with passive attenuation techniques. Active attenuation is an attractive way to achieve large amounts of noise reduction, particularly at low frequencies,where lower sampling rates are adequate, active control offers real advantages. Active noise control permits significant impro- vements in noise control, often with potential benefits in size, weight, volume, and cost of the system. 314 B. Tiseo et al. However, the main limitation of using active vibration control alone at higher frequencies stems from the fact that extensive computational requ- irements are needed for effective implementation of a proper controller. In addition, other drawbacks such as risk of instability due to control spillover and unreliability due to failure of sensors/actuators reduce the reliability of active control systems. Gentry et al. (1997) introduced an active passive noise control device, consisting of cylindrically curved sections of polyvinyldifluoride material (PVDF) embedded in a partially reticulated polyurethane acoustic foam. Nabil (2004) investigated a new class of smart foam able to control simultaneously structural and acoustic cavity modes over a broad frequency range. Thenail et al. (1994) investigated an active system that included a fibre- glass absorbing layer backedbyanair cavity terminatedwith anactive surface. In the study at hand, a sound absorbingmaterial is coupledwith an active surface to reduce sound pressure within an acoustic duct. In this way, an enhancement of the passive foam performance can be observed because of further acceleration of the fluid particles within the foam pores. 2. Experimental setup Theexperimentalmockupwas a cylindrical duct, closedatboth edges (Fig.1). The length of the duct (L=1500mm) was sufficiently large compared to its cross section dimension (D=150mm) so that the acoustic waves travel along the axis of the duct with planar wave fronts. This assumption enabled us to treat the duct as a one-dimensional system. Fig. 1. Experimental mock up: acoustic duct A loudspeaker was placed at the right end of the duct, while an exchan- geable metallic panel was positioned at the left end. The duct, integrated a longitudinal narrow piece with 40 holes, to host microphones. Experimental characterization of hybrid noise ... 315 3. Acoustic cavity modal analysis The experimental activity started with the evaluation of cavity normalmodes in terms of natural frequencies andmode shapes. Initially, a steel rigid panel, 1cm thickwas pinned at the left side of the duct. A chirp signal (100-1500Hz) was generated, amplified and given as input to the loudspeaker. As an acqu- isition and generation system, the LMS SCADASSystem, integrated with the LMSTest.Lab software was used. Frequency Response Functions (FRF) of microphone signals were evalu- ated in the frequency range of 100-1000Hz. The experimental setup scheme was reported in the following figure (Fig.2). Fig. 2. Experimental setup scheme for normalmode analysis By using LMS Test.Lab PolyMAX modal parameter identification algori- thm, natural frequencies and mode shapes were evaluated (Fig.3). Also the averaged pressure within the duct was calculated and plotted (Fig.4). Fig. 3. Acoustic cavity: mode shapes 316 B. Tiseo et al. Fig. 4. Acoustic cavity averaged pressure versus frequency: naked panel After that, a polyurethane foamwas inserted and placed close to the rigid panel. The foamthickness was chosen to be 15cm, being interested at the 0- 500Hz frequency range. Also in this case, the microphone signals frequency response function, the averagedpressure within the cavity and the normal modes were evaluated. The insertion of the passive foam implied a small shift of the natural fre- quencies (Table 1), no mode shapes modifications and a significant damping increase all over the frequency range of interest, even if this is more evident above the 300Hz frequency (Fig.5). Table 1. Acoustic cavity natural frequencies: naked panel versus panel with foam Natural frequencies [Hz] Naked panel Panel with foam 125 125 236 237 369 329 462 423 576 515 690 639 805 756 919 872 Experimental characterization of hybrid noise ... 317 Fig. 5. Acoustic cavity averaged pressure versus frequency: panel with foam Later, the rigid panel was replaced by a 1mm thick steel flexible one. The same kinds of investigations performed earlier, were repeated. In the case of flexible panel, the system response is superimposed to the one related to the rigid plate, except for the 300-400Hz frequency range. Here, two peaks show the coupling of fluid and structure (Fig.6). Furthermore, the system response is strictly dependent on panel-tube connection conditions, which are realised by clamping the panel between two rigid frames throughout four corner bolts (Fig.7). Fig. 6. Averaged FRF in the case of naked panel. Rigid versus flexible panel The insertion of foam implies a considerable damping augmentation in the low frequency range. Very high damping values evidence a shift of natural frequencies in the middle and high frequency range (Fig.8). 318 B. Tiseo et al. Fig. 7. Panel-duct connection: (a) frontal view, (b) lateral view Fig. 8. Averaged FRFmedia in case of flexible panel: naked panel versus panel with foam 4. Collocated feedback structural control The first control strategy consisted in improving passive foam performance, coupling it with the flexible panel which was ”acted” by using a piezoelectric actuator. Apiezoelectric patchandaphysically collocated accelerometer sensor iden- tify a typical structural collocated feedback control strategy. The panel vibration was measured by the accelerometer whose signal was derived, amplified and fed back to the actuator. The loudspeaker driven by a signal generator was used to excite the system at resonance frequencies. By using a control card, the optimal gain value K (defined as the one which implied the higher damping coefficient and consequently the maximum at- tenuation of peaks) was evaluated, observing the accelerometer signal on a oscilloscope. Experimental characterization of hybrid noise ... 319 Having identified the ”optimal K”, the experimental activity continued with a broadband acoustic excitation and themicrophone signals acquisition. Two cases were considered: single anddual channels control, implying one and two actuator-sensor pairs, Fig.9-13. Fig. 9. Single channel collocated feedback structural control: scheme Fig. 10. Single channel collocated feedback structural control – acceleration signal naked panel – Ctrl OFF vs ON The single channel feedback control strategy affected the system response only around themodemainly influenced by the structural part of the system (Figs.10-12). 320 B. Tiseo et al. Fig. 11. Single channel collocated feedback structural control – averaged FRF – naked panel Fig. 12. Single channel collocated feedback structural control – averaged FRF – panel with foam Fig. 13. Dual channel collocated feedback structural control: scheme Experimental characterization of hybrid noise ... 321 In the second configuration characterized by two pairs of collocated actu- ators and sensors, the piezoelectric actuators were located so that the maxi- mum numbers of structural modes within the frequency range of our interest were excited. An attenuation of 4dB on the accelerometers signals within the whole frequency range was observed, at the optimal gain value (Figs.14 and 15). Fig. 14. Dual channel collocated feedback structural control – acceleration signal 1: naked panel – Ctrl OFF vs ON Fig. 15. Dual channel collocated feedback structural control – accelerometer signal 2: naked panel – Ctrl OFF vs ON The adopted control strategy, in the case of naked panel, also affected the system acoustic response, above 300Hz frequency (Fig.16). When the panel was covered by foam, the influence of control was evident only around 369Hz mode (Fig.17). 322 B. Tiseo et al. Fig. 16. Dual channel collocated feedback structural control – averaged FRF – naked panel Fig. 17. Dual channel collocated feedback structural control – averaged FRF – panel with foam 5. PVDF control In this application, polyvinyldifluoride (PVDF) material was used as the ac- tuator because of its specific features. PVDFs are light weight, very thin, very flexible and readily shaped in complex shapes to cover irregular structures. They can also be easily integra- ted with existing structures, where classical point actuators like shakers or loudspeakers can not be used. The main disadvantage is that its piezoelectric constants are two orders of magnitude lower than those of PZT.Additionally, high driving voltage and limited temperature range have to bementioned as well. Experimental characterization of hybrid noise ... 323 An active foam was realised by incorporating a 0.28µm thick PVDF film within a foam. The filmwas opportunely curved to couple plane deformation due to the piezoelectric effect with vertical motion that is needed to accelerate fluid particles and, hence, to increase damping capacity. A PVDF film can be driven by a maximum input of 500V. Two uncor- related signals were given as the input to the PVDF film and loudspeaker. Two opposite trends occurred: high input voltage, probably increased dam- ping capacity of the foam (with noise reduction, as a consequence). At the same time, the PVDF worked as an additional noise source which increased the noise within the cavity, cancelling the first tendency. To avoid it, a high frequency excitation outside the range of hearing (> 20KHz)was given as the input to the piezo film. It resulted in small vibration amplitudes, incapable to accelerate adequately the fluid particles within the pores, so that to increase the dissipation energy. 6. Feedback structural acoustic control The control strategy consisted in capturing the pressure within the tube by using two adjacentmicrophones and located close to the piezoelectric patches, Fig.18. Fig. 18. Structural acoustic feedback control The pressure signals were subtracted, derived and fedback to the piezo actuators Fig.19. Strictly speaking, in that case, it was improper to speak of a ”collocated” feedback control, since the twomicrophones were positioned at the plate, but not to the same height of the piezoelectric. 324 B. Tiseo et al. Fig. 19. Structural acoustic feedback control: scheme The results obtained show an attenuation of about 3dB, within the range of 0-500Hz, in the case of nakedpanel,Fig.20. In the case of foam, the adopted control strategy effected only the 369Hzmode, Fig.21. Fig. 20. Structural acoustic feedback control – averaged FRF – naked panel Fig. 21. Structural acoustic feedback control – averaged FRF – panel with foam Experimental characterization of hybrid noise ... 325 7. Collocated feedback structural acoustic control The last experimental configuration, a collocated structural-acoustic control, was implemented. A collocated actuator sensor pair was used. It was reali- sed by inserting a microphone within the foam Figs.22-24. Compared to the nocollocated case, no changes were observed. Fig. 22. Collocated feedback control: (a) microphones, (b) piezoelectric actuators Fig. 23. Collocated feedback control: microphones within foam 326 B. Tiseo et al. Fig. 24. Structural acoustic feedback control – averaged FRF – panel with foam References 1. Gentry C.A., Guigou C., Fuller C.R., 1997, Smart foam for applications in passive active noise radiation control, Journal of Acoustical Society of Ame- rica, 101, 4 2. Nabil W., 2004, Smart Foam for Active Vibration and Noise Control, disser- tation Thesis, Department ofMechanical Engineering, University of Maryland 3. Thenail D., GallandM., SunyachM., SunhackM., 1994,Active enhan- cement of the absorbent properties of a porousmaterial, Smart Mater. Struct., 3, 18-25 Eksperymentalna charakterystyka hybrydowego układu redukcji hałasu działającego w akustycznym dukcie Streszczenie W pracy zbadano skuteczność zastosowania materiału absorbującego dźwięk współ- pracującego z aktywnąpowierzchnią przeznaczoną do redukcji ciśnienia akustycznego w dukcie. Zaobserwowano, że takie skojarzenie znacznie poprawia pasywne właści- wości samej dźwiękochłonnej pianki wskutek zwiększenia przyspieszeń cząstek płynu zawartegowporachpianki.Wwynikubadańdoświadczalnychuzyskanostrategiekon- troli hałasu zwiększające tłumienie poprzez użycie akustycznej pianki na elastycznej płycie domykającej dukt i sterowanej elementami piezoelektrycznymi. Manuscript received July 30, 2010; accepted for print September 22, 2010