外文翻譯-----地面運動沖擊作用下調諧質量阻尼器的有效性（英文）

1、Effectiveness of Tuned Mass Dampers against Ground Motion PulsesEmiliano Matta, Ph.D., P.E.1Abstract: It is known that the effectiveness of tuned mass dampers (TMDs) decreases as the input duration shortens. As a result,

2、 their use is commonly discouraged against short-duration, pulse-like ground motions, such as those occurring in near-field (NF) zones in the presence of forward directivity or fling-step effects. Yet a systematic assess

3、ment of such control impairment is still missing. In this paper, a recent analytical model of ground motion pulses is applied to the design and evaluation of TMDs against impulsive earthquakes. Based on this model, first

4、, a new optimization method is introduced as an alternative to the classical H‘ approach. Then, the two strategies are tested on single- and multi- degrees-of-freedom linear structures subject both to analytical pulses a

5、nd to a large set of NF records possessing pulse-like features. The resulting statistical evaluation, expressed by percentile response spectra, shows the pros and cons of a pulse-oriented TMD design and improves the gene

6、ral understanding of TMD performance under impulsive ground motions. DOI: 10.1061/(ASCE)ST.1943-541X.0000629. © 2013 American Society of Civil Engineers.CE Database subject headings: Earthquake engineering; Structur

7、al control; Seismic design; Damping; Earthquake resistant structures; Ground motion; Impulsive loads; Response spectra.Author keywords: Earthquake engineering; Tuned mass damper; Optimal design; Near-field earthquake; Gr

8、ound motion pulse model; Percentile response spectra.IntroductionPassive tuned mass dampers (TMDs) are widely used in civil en- gineering to mitigate vibrations induced by quasi-stationary dy- namic loads (winds, sea wav

9、es, pedestrians), but their seismic performance is known to depend on ground shaking properties (Kaynia et al. 1981). Requiring the motion of the primary structure to react with, TMDs prove effective against long-duratio

10、n, narrow- band ground motions, but may fail in reducing the peak response to impulsive quakes. The literature acknowledges this inconvenience. Sladek and Klingner (1983) showed that a TMD under a short-duration input ma

11、y not have time to produce a significant control action. Using sinusoidal excitations of finite duration, Tsai (1995) proved that a TMD with 5% mass ratio is useless during the first cycle of the response. Sinha and Igus

12、a (1995) developed closed-form expres- sions for TMD performance under short-duration, wide-band ground motion, concluding that their effectiveness is negligible for very short-duration inputs and quadratically approache

13、s its stationary value as the duration increases. Chen and Wu (2001) observed that under impulsive ground motion the efficacy of a TMD is not fully developed when the structure experiences its peak re- sponse, and that t

14、he heavier the TMD, the slower it reaches its full potential. Lukkunaprasit and Wanitkorkul (2001) and Pinkaew et al. (2003) excluded near-field (NF) impulsive records from theiranalyses because they would be hardly cont

15、rollable using TMDs. Taniguchi et al. (2008), testing a TMD on a base-isolated structure, showed that TMD effectiveness is 10% less for NF records than for far-field (FF) records. Leung et al. (2008) proposed a new TMD o

16、ptimization against nonstationary ground motions modeled as evolutionary stationary random processes, but they did not test it under real accelerograms. These studies demonstrate that the pos- sible deficiencies of TMDs

17、against pulse-like earthquakes are generally acknowledged. However, the limited number of real records used in the analyses (none of these studies uses more than seven impulsive records, and the whole of them use a total

18、 of 16 only) shows that such an issue is far from finding a conclusive, statistically consistent, systematization. One typical example of pulse-like earthquakes is given by NF ground motions. Ground shaking near a fault

19、rupture may be charac- terized by a short-duration impulsive motion that exposes structures to high-input energy at the beginning of the record. This pulse-type motion can be explained through the concepts of forward dir

20、ectivity andfling-stepeffects(Bray and Rodriguez-Marek 2004). The former is because of the fault rupture propagating toward the site at a velocity close to the shear-wave velocity, causing most of the seismic energy to r

21、each the site within a short time, in the form of a large energy pulse at the beginning of the record, mostly oriented in the fault-normal direction. The latter is because of permanent ground displacements accumulating a

22、t the site as a result of tectonic movements. Both effects may result in large-amplitude, long-period pulses in the ve- locity and displacement time histories, which are particularly chal- lenging for the structural safe

23、ty of long-period structures. Many studies have been devoted to improving the performance of structures exposed to NF ground motion. The use of simplified analytical models of ground motion pulses may prove a valid tool

24、for the systematic design and assessment of seismo-protective systems. Several models are available for this purpose. Makris and Chang (2000) classified NF velocity pulses into types A, B, and Cn to investigate the perfo

25、rmance of various damping devices. Alavi and1Postdoctoral Researcher, Dept. of Structural and Geotechnical Engi- neering, Politecnico di Torino, 10129 Turin, Italy. E-mail: emiliano. matta@polito.it Note. This manuscript

26、 was submitted on July 27, 2010; approved on April 24, 2012; published online on April 26, 2012. Discussion period open until July 1, 2013; separate discussions must be submitted for individual papers. This paper is part

27、 of the Journal of Structural Engineering, Vol. 139, No. 2, February 1, 2013. ©ASCE, ISSN 0733-9445/2013/2-188–198/ $25.00.188 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / FEBRUARY 2013J. Struct. Eng. 2013.139:

28、188-198.Downloaded from ascelibrary.org by University of Liverpool on 07/18/15. Copyright ASCE. For personal use only; all rights reserved.and (3), posing v 5 0 and t0 5 0 s, and scaled to have a unitary maximum value; t

29、herefore, the input signal depends exclusively on the pulse frequency, vp, and on the pair of buildup and decay parameters, N and zp. Then, for any assigned pair N and zp, the pulse transfer function, TFp, is evaluated f

30、or the combined system, reporting the pulse frequency vp in the abscissas and the corre- sponding peak structural displacement in the ordinates. The new Hp design consists in numerically finding the optimal TMD parameter

31、s, r and z, which minimize the maximum of TFp, for the assigned values of zs, m, N, and zp. Still assuming zs 5 0:02 and m as between 0.001 and 1.00, and further taking N 5 1 and zp 5 0:4 (as somewhat representative of t

32、he average pulse, as will be explained next), the results of such numerical optimization are plotted as white squares in Fig. 1, the response ratio R being expressed in terms of TFp. As for the H‘ design, a close-form ap

33、proximation (continuous lines in Fig. 1) may be derived asropt ¼ ropt,0 1 þ p1 ffiffiffi ffi m p þ m3½p2 2 p3=ð1 þ p4m5Þ? andzopt ¼ p5m4 2 p6m3 þ p7m2 2 p8m m3 2 p9m2 2 p10m &

34、#254; p11ð6Þwhere coefficients pj are listed in Table 1.Eqs. (5) and (6) and Fig. 1 may be valuable tools for TMD design. Fig. 1 also helps clarify similarities and differences between H‘- and Hp-designed TMDs.

35、 With both methods, ropt decreases and zopt increases for increasing m. Yet the Hp design presents a more ir- regular trend (because of the transient character of the excitation), and above all, provides much lower frequ

36、ency and damping ratios, particularly in the case of small mass ratios. Furthermore, with both methods, Ropt tends to decrease as m increases, but there is little advantage obtained in increasing m beyond 10–20% (in the

37、Hp case, Ropt actually reaches a minimum around m 5 45%). The perfor- mance of the TMD appears strongly influenced by the impulsive character of the ground motion, and Ropt results are sensibly smaller in the H‘ case (ha

38、rmonic input) than in the Hp case (pulse-like input). In both cases, for large m values, ropt decreases so much that the classical concept of tuning, which implies frequency closeness (r ? 1), needs to be revised, and ra

39、ther intended as the mere ad- justment of the frequency ratio to its optimum value, no matter how far from unity. With this in mind, the terms tuning and TMDs are still used independently of the values taken by m or ropt

40、. Assuming for instance m 5 10%, the two design methods are compared in Fig. 2 on a SDOF structure of period T 5 1 s. The numerical optimization provides the following TMD parameters: r 5 0:873 and z 5 0:194 for the H‘ d

41、esign, and r 5 0:856 and z 5 0:026 for the Hp design. In Fig. 2(a), the steady-state transfer functions TF are plotted for, respectively, the uncontrolled structure, the H‘-controlled structure, and the Hp-controlled str

42、ucture. In Fig. 2(b), the corresponding pulse transfer functions TFp are reported. In Fig. 2(c), the displacement time history of the SDOF structure is reported under the effect of an impulsive ground motion with a pulse

43、 frequency of 0.8 Hz. A trade-off clearly emerges between the two design methods: the H‘ design, optimal against sine inputs, loses effectiveness under pulse-like inputs; and the Hp design, optimal for pulse mitigation,

44、is severely impaired in steady-state terms. The reason is that, to minimize the peak response to the pulse load, the Hp design decreases the TMD damping ratio with respect to the H‘ design [as already observed in Fig. 1(

45、b)] to such an extent that the thick curve in Fig. 2(a) looks more like Frahm’s undamped vibration absorber than like Den Hartog’s TMD (Warburton 1982). The re- duced damping allows the TMD to more rapidly respond to the

46、 impulsive load, but hinders it in reducing the two peaks of the steady- state transfer function. Also, the reduced damping has a second in- convenience, clearly apparent in Fig. 2(c), of diminishing TMD capabilitytocont

47、rolthepostpeak,free-decayresponseofthestructure. The two mentioned drawbacks, together with the need to rely on predictions of pulses features, pose the question of the trueFig. 1. Optimal frequency ratio (a), damping ra

48、tio (b), and response ratio (c) as a function of the mass ratio, according to the H‘ design (black) and the Hp design (white): numerical results (squares) and their closed-form approximation according to Eqs. (5) and (6)

49、 (continuous line)Table 1. Numerical Values of Coefficients in Eqs. (5) and (6)Coefficient Valuei1 0.04758 i2 0.03056 i3 0.002715 i4 0.007453 p1 0.2926 p2 0.2301 p3 75.34 p4 30185 p5 0.7269 p6 0.3934 p7 0.07388 p8 0.0011

50、94 p9 0.2978 p10 0.01214 p11 0.01407190 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / FEBRUARY 2013J. Struct. Eng. 2013.139:188-198.Downloaded from ascelibrary.org by University of Liverpool on 07/18/15. Copyright AS