Ative dmping devies for eroelsti models D. Rohi 1, T. Blduzzi 1, G. Czzulni 1 nd T. Argentini 1 1 Deprtment of Mehnis, Politenio di Milno, Vi L Ms 1, Milno, Itly. dniele.rohi@polimi.it Abstrt The dmping ontrol of eroelsti models for wind tunnel testing is key feture influening the qulity of the obtined results. The pper presents two devies to tively ontrol the dmping prmeter using piezoeletri tutors. The former llows to diretly modify the struturl dmping of the eroelsti model while the ltter llows to modify the dmping through n tive tuned mss dmper. The devies re speifilly developed to perform n esy nd fine modifition of the dmping prmeter for vortex indued vibrtion investigtions. The devies were initilly hrterised in lb nd then used to perform wind tunnel tests on bridge tower model. 1 Introdution A rule of thumb for wind tunnel testing on eroelsti models is to keep the model s dmping s low s possible. The dmping of the full sle struture is not preditble t the design stge nd is usully ssumed equl to other similr onstrutions ording to the designer experiene. Even in se of wind tunnel testing on n existing struture, the rel struturl dmping vlue is not known unless full sle monitoring is dopted. For ske of sfety the design nd the onstrution of n eroelsti model is therefore refully performed trying to obtin very low struturl dmping level in order to better highlight possible eroelsti phenomen. Espeilly, when the vortex indued vibrtions of strutures re studied, this represents n unvoidble speifition sine the dynmi response of the models is minly influened by dmping in lok-in ondition. The Sruton number is the prmeter ontrolling the mximum vibrtion mplitude tht struture my reh under the vortex shedding exittion in lok-in ondition, but sine the eroelsti model is expeted to orretly represent the mss distribution of the rel struture, the dmping is the dominting prmeter. A struturl designer is therefore interested to define, through wind tunnel tests, the dependene of the mximum vibrtion mplitude in lok-in ondition on the Sruton number in order to evlute if the struture my withstnd the wind lods during its expeted life period. Wind tunnel result lso drives the dimensioning of dditionl externl dmpers when the struturl dmping is not suffiient to keep the vibrtion mplitude under the sfety threshold. The more prtil wy to vry the Sruton number of the eroelsti model during wind tunnel tests is by modifying the struturl dmping using externl dmpers. This prtise hs some drwbks due, on one hnd, to the possible modifition of the dynmi hrteristis of the eroelsti model due to the ddition of externl devies nd, on the other hnd, to the very low model dimensions nd to the orresponding low fores sking for very fine ontrol of the tion imposed by the devies tht hs to be miniturised to be inserted inside the model. In the present study two devies hve been designed, built nd hrterised before being pplied to referene eroelsti model for wind tunnel testing. The two devies ddress different experimentl issues: the former ims to modify the struturl dmping of the model through n tive ontrol logi pplible to single or to multiple modes of the struture, the ltter ims to reprodue Tuned Mss Dmper (TMD) positioned inside the struture. Both the devies exploit pth piezoeletri tutors 1
6 th Europen nd Afrin Conferene on Wind Engineering 2 nd rel time ontrol logis. The detils of the devies will be illustrted in the pper together with their pplition to wind tunnel test se using referene erodynmi model representing bridge tower subjeted to vortex shedding exittion. 2 Dmping devie 1: tive struturl dmping ontrol As previously introdued, the proposed solutions ims t providing the possibility of fine tuning of the struturl dmping of the eroelsti models. Due to the reltively little displements tht our during wind-tunnel tests, the struture behviour n be ssumed liner. As onsequene, its dynmi behviour n be represented, using modl-spe pproh, s the sum of the ontribution of eh vibrtion mode q s Mq Rq Kq Λ F F (1) T d where M, R nd K re the digonl modl inertil, dmping nd stiffness mtries, F represents the known fores pplied on the struture nd the kinemti reltionship between the point of pplition of those fores nd the system oordintes ; F d represents the generi unknown disturbne fores ting on the struture. Consequently, in order to identify the effet of dmping on vortex-indued vibrtions, tool ble to modify the dmping ssoited to one or more modes (without modifying the other ones) is of gret interest for wind tunnel tests. In ft, in this wy, it beomes possible to investigte those modes hving ritil ontribution on vortex indued vibrtions. Moreover, for eh mode, the minimum struturl dmping needed to keep vortex-indued vibrtions under the desired vlue n be evluted. The present work proposes the pplition of modl-spe ontroller (Inmn, 2001), whih is ble, under some ssumptions (Rest et l., 2010), to hieve the desired effet. In ft, Independent Modl Spe Control (IMSC) (Czzulni et l., 2012) provides ontrol fores F proportionl to the ontribution of the onsidered modes q T 1 F Λ G q R (2) where G R is the digonl ontrol gin mtrix, while q is the vetor ontining the modl veloities. This ontribution, being not diretly mesurble, n be lulted strting from the dynmi model of the system nd from the vilble mesurements. Typilly, for eroelsti models, the vibrtion mesurement is relized using elerometers or strin guges embedded inside the struture. For the bridge tower model used for implementing the proposed tive ontrol logi, the mesurements re performed using elerometers pled t different heights long the tower. The ontrol fores re provided through piezoeletri pth tutors pplied on the internl spine of the model. Depending on the number of modes involved in vortex indued vibrtions, different number of tutors n be bonded to the struture. In order to keep the dmped modes unoupled, the number of tutors must be t lest equl to the number of ontrolled modes. Considering eq. (2), vrying the gin mtrix G R it is possible to hieve different dmping vlues, obtining preise regultion. Moreover, if negtive gins re pplied, it is possible to impose to the struture dmping tht is lower thn the struturl one. This ondition, whih nnot be obtined
6 th Europen nd Afrin Conferene on Wind Engineering 3 with pssive devies, llows to emphsize vortex indued vibrtions when the struturl dmping of the model is too high to highlight them. 3 Dmping devie 2: tive TMD The dmping devie presented in the previous setion llows to study the vortex-indued vibrtion phenomenon, defining the reltionship between the system dmping nd the vibrtion mplitude. In mny prtil sitution, when the initil struturl dmping of the rel struture is not suffiient, it my be equipped with Tuned Mss Dmpers (Tmur 1998; Co et l., 1998; Csiti & Giulino 2007). These devies re designed in order to reh the required Sruton number for the oupled struture-dmper system. This pssive devie modifies the wind-struture intertion sine, depending on its the dmping, two resonnt peks our. For this reson, it beomes interesting to reprodue the behviour of the struture with n pplied mss dmper in the wind tunnel, in order to experimentlly relte the mss dmper prmeters to the vibrtion mplitude nd, s onsequene, to orretly tune them. Relling eqution (1), onsidering only the eqution of the i-th mode ontrolled by the TMD, nd supposing to ple the TMD t the top of the tower, the oupled eqution beomes 2 mi i, topm i, topm qi ri 0 qi ki 0 qi i, topm m y 0 r y 0 k F y d (3) where m, r nd k re the mss, dmping nd stiffness of the TMD, i, top represents the i-th mode eigenvetor evluted t the tower top nd y is the reltive displement between the TMD mss nd the tower top. The present work proposes mss dmper, whose hrteristis n be tively modified in rel-time. The mss dmper is relized through hrmoni steel br (whih sets the TMD stiffness) equipped with piezoeletri tutor. Figure 1 shows the working sheme of the tive TMD, highlighting its min elements. Figure 1: An imge of the tive TMD mounted t the top of the tower
6 th Europen nd Afrin Conferene on Wind Engineering 4 The mss hs been hosen so tht the mss rtio of the TMD is equl to 3%, while the br stiffness is hosen to optimize the frequeny rtio. This devie, providing n dequte ontrol lw, n tively modify its dynmi hrteristis nd, in prtiulr, the dmping introdued by the TMD on the struture. The tutors re driven by the following ontrol lw F G y (4) tt where G is the tive TMD gin. Also in this se, the opportunity to modify the dmping prmeter, inresing or deresing it by hnging the ontrol lw, llows to esily perform fine tuning of the TMD. 4 Results Both the dmping ontrol devies were initilly librted in order to define the reltionship between the totl dmping rehed by the ontrolled struture nd the vlue of the gin prmeters. To this purpose, 1:100 wind tunnel eroelsti model of ble styed bridge (Figure 2) hs been tken s referene se. The model hs been equipped with the two dmping ontrol devies nd with 6 elerometers distributed long the struture to mesure its vibrtions. In ddition, one elerometer hs been pled on the TMD mss to mesure its movement with respet to the tower top. Figure 2: () The eroelsti model of bridge tower; (b) modl shpe of the 1st bending mode long the y diretion Due to its hrteristis, the tower hs deoupled bending modes long two min diretions. Moreover, previous tests on the eroelsti model (Belloli et l., 2011) showed tht only the first mode in eh diretion is interested by vortex-indued vibrtions. Consequently n IMSC ontrol onsidering only the first two modes hs been implemented on the struture. For the sme reson, the tive TMD hs been tuned to work on the first mode long the y diretion. Figure 3 shows the reltionship between the ontrol gin nd the totl struturl dmping on the first "y" mode for the two proposed ontrol devies.
6 th Europen nd Afrin Conferene on Wind Engineering 5 Figure 3: Gin dmping reltionship for the two tive devies long the y diretion: () tive dmping ontrol; (b) tive TMD At lst, wind tunnel experimentl mpign hs been performed to test the dmping ontrol devies under usul wind tunnel testing onditions. During the experimentl mpign the tower behviour under lok-in ondition ws investigted for different levels of Sruton number obtined by hnging the tive ontrol prmeters without swithing off the wind. As n exmple, Figure 4 reports the mximum vibrtion mplitude mesured t top tower in the lok-in rnge versus the normlized wind speed expressed by U rid U f B (5) where U is the wind speed, f is the struturl osilltions frequeny nd B is hrteristi length (in this se, the setion width). The top vibrtions re expressed s normlized displement (rtio between the displement nd B ). The results referred to the devie 1 (tive dmping ontrol) show tht vortex-indued vibrtions re ompletely neled out with gin equl to 0.35 (Figure 4 - yn line, orresponding to first mode's dmping rtio equl to 0.8%). The result is onfirmed by the test with the devie 2 (tive TMD), where vibrtions re stopped with TMD gin equl to 0 (Figure 4b - blue line, orresponding to the sme dmping rtio vlue, 0.8%).
6 th Europen nd Afrin Conferene on Wind Engineering 6 Figure 4: Wind tunnel test results with the tive dmping ontrol () nd with the tive TMD (b) 5 Conlusions In the present work, two devies for the dmping regultion of eroelsti models for wind-tunnel tests hve been proposed. The first solution onsists of n tive ontrol of the dmping rtio ssoited to the vibrtion modes exited by the wind-struture intertion. The tive ontrol is performed through number of piezoeletri pth tutors bonded to the model itself. It llows to independently regulte the dmping rtios of the different modes, providing preise dmping regultion. The seond solution is n tive TMD embedded in the model. The TMD is designed to reprodue the dynmi behviour of the full-sle TMDs nd it is equipped with piezoeletri tutors. An tive ontrol lw llows the regultion of the TMD dmping effet. The proposed solutions hve been experimentlly tested on n eroelsti model of bridge tower. At first, the reltionship between the ontrol gins nd the dmping rtios hieved on the struture hs been identified. Then, the two devies hve been used for the wind-tunnel tests.
6 th Europen nd Afrin Conferene on Wind Engineering 7 Referenes Belloli M., Fossti F., Gippino S., MuggisS., Villni M. 2011. On the erodynmi nd eroelsti response of bridge tower. Journl of Wind Engineering nd Industril Aerodynmis, 99, 671-800. Co H., Reinhorn A. M., Soong T. T. 1998. Design of n tive mss dmper for tll tv tower in nnjing, Chin. Engineering Strutures, 20, 134-143. Csiti F., GiulinoF. 2007. Performne of multi-tmd in the towers of suspension bridges. Journl of Vibrtion nd Control, Czzulni G., Rest F., Ripmonti F., Znzi R. 2012. Negtive Derivtive Feedbk for vibrtion ontrol of flexible strutures. Smrt Mterils nd Strutures, 21, rt. n. 075024 Kim Y. M., You K. P., Kim H. Y. 2008. Wind-indued exittion ontrol of tll building with tuned mss dmpers. Struturl Design of Tll nd Speil Buildings, 17, 669-682. Inmn D. J. 2001. Ative modl ontrol for smrt strutures. The Royl Soiety - Mthemtil,Phisil nd Engineering Siene, 359, 205-219. Rest F., Ripmonti F., Czzulni G., Ferrri M. 2010. Independent modl ontrol for nonliner flexible strutures: An experimentl test rig. Journl of Sound nd Vibrtion, 329, 961-972 Tmur Y. 1998. Applition of dmping devies to suppress wind-indued responses of buildings. Journl of Wind Engineering nd Industril Aerodynmis, 74-76, 49-72