ARTICLE IN PRESS

International Journal of Pressure Vessels and Piping 83 (2006) 672685 www.elsevier.com/locate/ijpvp

Optimum X-plate dampers for seismic response control of piping systems

S.V. Bakrea, R.S. Jangida,, G.R. Reddyb
a

Department of Civil Engineering, Indian Institute of Technology Bombay, Powai, Mumbai 400 076, India b Reactor Safety Division, Bhabha Atomic Research Center, Anushaktinagar, Mumbai 400 058, India Received 20 April 2005; received in revised form 2 May 2006; accepted 20 May 2006

Abstract In a vibrating system, the most effective mechanism to dissipate energy is the inelastic strain of supplemental metallic elements with plastic deforming characteristics. An X-plate damper (XPD) is one device that is capable of sustaining many cycles of stable yielding deformation resulting in a high level of energy dissipation or damping. The present paper focuses on a numerical study to investigate the seismic effectiveness of an XPD for piping systems in industrial units (e.g. chemical and petrochemical industries) and utilities such as thermal and nuclear power plants. The seismic performance of piping systems is investigated under important parametric variations of the damper properties (i.e. height, width and thickness of the XPD) under arbitrary ground motions. Investigations are reported for an industrial piping system equipped with an XPD and the response quantities of interest are the relative displacements, absolute accelerations and support reactions of the piping system. The response quantities of the controlled (with XPD) piping system are compared with the corresponding uncontrolled (without XPD) piping systems, to establish the seismic effectiveness of the XPD. Seismic energy dissipation in the piping system, which is represented by the hysteretic energy of the XPD, is also evaluated and compared. It is observed that the XPDs are very effective in reducing the seismic response of piping systems. Moreover, for a given piping system and ground motion, it is difcult to arrive at the optimum properties of an XPD from the parametric variation of the properties of the XPD and by monitoring the responses of the piping system. Therefore, use of hysteretic energy dissipation by an XPD is proposed to obtain the optimum properties of the XPD. Furthermore, the effects of the properties of an XPD on the free vibration characteristics of the piping system are also presented, which is crucial for the design of piping systems with XPDs. r 2006 Elsevier Ltd. All rights reserved.
Keywords: Piping system; X-plate damper; Hysteretic; Seismic; Energy; Parametric; Optimum parameters

1. Introduction Structural control using energy dissipating devices is an appealing alternative to the traditional earthquake-resistant design approaches. In this approach, a substantial portion of the vibration energy is absorbed or consumed at selected locations within a structure through protective devices especially designed for this purpose. In particular, passive control devices offer various advantages over functionally complex active and semi-active control devices. Devices in this class have the ability to dissipate the earthquake input energy by virtue of their nonlinear
Corresponding author. Tel.: +91 22 2572 2545; fax: +91 22 2572 3480.

E-mail address: rsjangid@civil.iitb.ac.in (R.S. Jangid). 0308-0161/$ - see front matter r 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.ijpvp.2006.05.003

behavior. Since these protective devices are separated from the main structure, they act as structural fuses that can be replaced, if damaged, after the occurrence of a severe seismic event. For piping systems, these devices should satisfy the basic requirement of thermal expansion without generation of undesired stresses. During strong earthquakes, it should be ensured that these devices dissipate most of the earthquake input energy and thereby reduce the forces transferred to the piping system. At present, snubbers are used in nuclear power plants to reduce the seismic forces in the piping system. However, snubbers are very expensive and are associated with problems of oil leakage (in the case of hydraulic snubbers) and locking (in the case of mechanical snubbers) and as a result, require frequent inspection. Hence, Olson and Tang [1] and Cloud

ARTICLE IN PRESS
S.V. Bakre et al. / International Journal of Pressure Vessels and Piping 83 (2006) 672685 673

Nomenclature a b g t sy a b n t A post-to-pre-yield stiffness ratio Wens model parameter Wens model parameter Wens model parameter yield stress of an XPD half the height of an XPD width of an XPD Wens model parameter thickness of an XPD Wens model parameter

E Ed Eh EI F Fy H Kd q xp p x _p x

modulus of elasticity percentage energy dissipated by an XPD hysteretic energy in an XPD input energy to the piping system force in an XPD yield force of an XPD rate of hardening initial stiffness of an XPD yield displacement of an XPD displacement of piping system absolute acceleration of piping system relative velocity of piping system

et al. [2] proposed the reduced use of snubbers and proposed seismic stops instead. It is very difcult to frequently inspect snubbers in the high-radiation conditions that exist in nuclear power plants and there are huge costs required for yearly maintenance of snubbers. Moreover, malfunctioning of snubbers develops undesired loads [3] on the piping systems thus questioning their safe functioning. A variety of passive devices have been proposed for the structural control of piping systems including the viscoelastic damper, the compact dynamic absorber, friction damper and X-plate damper (XPD) [4,5]. The XPD consists of an assembly that holds either single or multiple components of thin metallic or layered plates of X or V shape. It utilizes the plastic deformation characteristics of the steel components to damp the input seismic energy. The XPD can sustain many cycles of stable yielding deformation without fatigue thus dissipating the input seismic energy in the form of hysteretic deformation. Kelly et al. [6] were the rst to propose the use of XPDs for seismic energy dissipation in structures, and this work was extended by Skinner et al. [7] and Tyler [8]. Proposals for the use of XPDs in piping systems was rst presented, again by Kelly et al. [9]. Schneider et al. [10] performed a series of experimental tests on a complex spatial piping system equipped with XPDs. Kobayashi [4] reported studies on composite laminated plates for a triangular plate damper. Later, several experimental and analytical studies were reported on a piping system equipped with metallic energy absorbers [1117]. More recently, Parulekar et al. [18] and Bakre and Jangid [19] performed several component tests on X-plate metallic damper and on piping systems with an XPD. In all the aforementioned studies, XPDs were found to be very effective in seismic control of structures. However, comparatively detailed studies are not yet available on parametric variations of the properties of XPD, which play an important role in the seismic analysis and design of piping systems equipped with XPD. Moreover, the stiffness being added to the piping structure in the form of an XPD signicantly alters the vibration characteristics of the system. Hence, it is important to study the

effect of the damper properties on the free vibration characteristics of the piping system. The present preliminary research focuses on a numerical study to investigate the seismic effectiveness of an XPD as a seismic protective system for industrial piping systems. The seismic performance of a piping system is studied under important parametric variation of the damper properties for an industrial piping system under real earthquake ground motions. The damper parameters considered are height, width and thickness. It is observed that the optimal XPD properties are very difcult to obtain by simply monitoring the piping responses. Hence, use of hysteretic energy dissipation by the XPD is proposed to obtain the optimal properties of an XPD for a given piping system and ground motion. Lastly, the effect of the properties of an XPD on the free vibration characteristics of the piping system is also studied.

2. Mechanism of XPD XPDs are made of thin metallic plates that dissipate energy through their exural yielding deformation. They can sustain many cycles of stable yielding deformation [16,17], resulting in high levels of energy dissipation or damping. The X shape of the damper is adopted so as to have a constant strain variation over its height, thus ensuring that yielding occurs simultaneously and uniformly over the full height of the damper. A typical XPD with the holding device used in the present work and its application to a piping system is shown in Fig. 1(a). A series of experimental tests was conducted at Bhabha Atomic Research Centre (BARC) [18] and IIT Bombay [19] to study the behavior of these dampers. The following observations are noted from the forcedeformation characteristics of the XPD shown in Fig. 1(b): (i) it exhibits smoothly nonlinear hysteretic loops under plastic deformation, (ii) it can sustain a large number of yielding reversals, (iii) there is no signicant stiffness or strength degradation and (iv) it can be accurately modeled by Wens hysteretic model [20] or as a bilinear elasto-plastic material. A

ARTICLE IN PRESS
674 S.V. Bakre et al. / International Journal of Pressure Vessels and Piping 83 (2006) 672685

Connecting Lugs

X-plate F

where Kd is the initial stiffness, Fy is the yield load and q is the yield displacement of the XPD; E and sy are elastic modulus and yield stress of the damper material, respectively; a, b and t are the height, width and thickness of the XPD as shown in Fig. 1(a). The properties of the plastically deformed XPD are expressed as   sy b Ht3 2 4y2 P 3 t H E , (4) 0 12Ea y0 where P is the plastic force in the XPD due to given displacement d; H is rate of strain hardening and y0 is the elastic depth given by sy a2 . (5) Ed It is to be noted here that using the above equations, the properties of the XPD, i.e. Kd, Fy, q and a, could be obtained for a particular combination of a, b and t of an XPD. These properties are required in Wens hysteretic model. y0 3. Modeling of piping system with XPD The piping system considered for the present study is made of carbon steel having Youngs modulus (E) of 192.2 GN/m2 and Poissons ratio 0.214, and is supported on guides with its ends anchored. Fig. 2 shows a schematic and FE model of the industrial piping system with XPD.

a x z (a) b x y

600

Damper Force (N)

0
2.7 4m

Z Y X Restraint Anchorage Pipe diameter=165 mm Pipe thickness=5.5 mm Elbow radius=225 mm

1.24 m

-600 Experimental Wen's (10 mm peak) -20 (b) -15 -10 Wen's (15 mm peak) Wen's (5 mm peak) 10 15 20

XP

D 2m

2.4

0.83 m

1.55 m

4m 2.5

3.0

5m

-5 0 5 Displacement (mm)

(a) Rx

3.1

2m

4.1

Fig. 1. (a) X-plate damper and (b) forcedeformation behavior of X-plate damper.

comparison of experimental and Wens representation of the hysteresis loop is also shown in Fig. 1(b). Using beam theory, the elastic properties of the XPD are expressed as Kd Ebt3 , 12a3 (1)
69 kg Kd F 69 kg

sy bt2 Fy , 6a 2sy a2 q , Et

(2)
(b)

69 kg 69 kg

(3)

Fig. 2. Schematic and FE model of a piping system with an XPD.

ARTICLE IN PRESS
S.V. Bakre et al. / International Journal of Pressure Vessels and Piping 83 (2006) 672685 675

All the bends in the piping system are 901 with bend radius of 225 mm. Fig. 2 also shows the location of the lumped masses and the XPD in the piping system. The following assumptions are made for seismic analysis of a piping system with XPDs: (1) The straight members in the piping system are modeled as 3D Beam elements and the bends are modeled as 3D Elbows having six degrees-of-freedom at each node. (2) The mass of each member is assumed to be distributed between its two nodes as a point mass. In addition to the mass of the piping system, the externally lumped masses are assumed to be effective in the three translational degrees-of-freedom. (3) Only the XPD behaves nonlinearly whereas the piping system remains linear. (4) The forcedeformation behavior of the XPD is considered as hysteretic, based on the nonlinear model proposed by Wen [20].

interesting to investigate the effects of these (geometric) properties on the seismic performance of a piping system. Moreover, the seismic energy in the piping system is absorbed by the hysteretic forcedeformation behavior of the XPD. The control of the seismic energy being transmitted to the piping system is predominantly governed by the hysteretic characteristics of the XPD, which depend on the properties of the XPD. The response of the piping system is thus signicantly inuenced by the hysteretic loss of input energy. The rate of the hysteretic energy dissipated in any XPD in the piping system at time T is given by dE h _p 1 aK d Z x (8) dT and the total hysteretic energy dissipated by an XPD at time T is given by Z T _ p dT . E h 1 a K d Zx (9)
0

4. Hysteretic modeling of XPD XPDs, when subjected to cyclic loading, show stable hysteresis loops dissipating signicant amounts of energy. As noted from Fig. 1(b), the transition of this hysteresis loop from its initial value, called the initial stiffness (Kd), to its limiting value, called the post-yield stiffness, is very smooth. For such smoothly varying hysteretic behavior, Wen has proposed a mathematical model that can be used to represent the force in the XPD. Based on the versatility of Wens smooth hysteretic model to achieve various forcedeection characteristics, it is adopted to represent the hysteretic force in the damper. The restoring force (fd) in the XPD is given by f d aK d xp 1 aK d qZ , where Z is expressed as dZ _ p bjx _ p jjZ jn1 gx _ p jZ jn , Ax (7) dt in which A, b, g and n are dimensionless Wens model parameters; and a is the post- to pre-yield stiffness ratio for the XPD, Z is a function (whose value varies from 1 to _ p are, respectively, the +1) governed by Eq. (7), xp and x displacement and velocity of the piping system at the XPD location. For the present study, Wens model parameters are obtained by a trial and error method to t the experimental hysteresis loops shown in Fig. 1(b). The parameters obtained are A 0.545, b 0.22, g 0.22 and n 1. The shape of the hysteresis loop is controlled by adjusting Wens model parameters. It is to be noted that Wens model parameters will only affect the shape of the hysteresis loop whereas the seismic performance of the piping systems will be controlled by the geometrical properties of the XPD (i.e. a, b and t). It will be (6)

The input energy to the piping system is given by Z T _ gT M frgu g dT , fu EI

0

(10)

where [M] is the mass matrix of the controlled piping g is the system, {r} is the inuence coefcient vector and u earthquake ground acceleration. The percentage energy dissipated in the piping system is expressed as Ed 100E h . EI (11)

It will also be interesting to study the effect of the properties of XPD on the percentage seismic energy being dissipated. 5. Governing equations of motion The equations of motion of a piping system equipped with an XPD, under a uni-directional component of ground motion, are expressed in the following matrix form: g C fu _ g K fug U fF g M frgu g , M fu (12)

fug fx1 ; y1 ; z1 ; x2 ; y2 ; z2 ; x3 ; y3 ; z3 ; . . . . . . ; xN ; yN ; zN gT , (13) where [C] and [K] represents the damping and stiffness matrix, respectively, of the piping system of order 6N 6N, g; fu _ g and {u} represent where N is the number of nodes; fu acceleration, velocity and displacement vectors, respectively; [U] is the location matrix for the restoring force of the XPD; {F} is the vector containing the restoring force of the XPD; and xi, yi and zi are the displacements of the ith node in the piping system in X-, Y- and Z-directions, respectively. The mass matrix has a diagonal form. The stiffness matrix of the piping system with friction support is constructed separately and then static condensation is carried out to eliminate the rotational degrees-of-freedom.

ARTICLE IN PRESS
676 S.V. Bakre et al. / International Journal of Pressure Vessels and Piping 83 (2006) 672685

With the rst two natural frequencies of the piping system known, and the damping ratio obtained from the test model, the damping matrix is obtained using Rayleighs method. 6. Incremental solution technique Because of the hysteretic behavior of the XPD, the governing equations of motion are solved in the incremental form using Newmarks time-stepping method assuming linear variation of acceleration over a small time interval, Dt. The equations of motion in incremental form are expressed as g C fDu _ g K fDug U fDF u g M frgDu g , M fDu (14) where {DFu} is the incremental restoring force vector of the XPD. Following the assumption of linear variation of accel g and fDu _ gare eration over the small time interval, DT, fDu given as g a0 fDug a1 fu _ T g a 2 fu T g, fDu _ g b0 fDug b1 fu _ T g b 2 fu T g, fDu (15) (16)

where dT is the time interval at which the ground motions are recorded and Nd are the number of divisions adopted for convergence of the structural responses due to the nonlinearity of frictional forces, which is based on the ground motion considered and lies in the range of 140160 for the present study. 7. Numerical study The seismic response of the piping system with an XPD is investigated under uni-directional excitation of four components of real earthquake ground motions. The specic components of these ground motions are indicated in Table 1. The response quantities of interest for the piping system under consideration are the relative dis p of the piping placements (xp), absolute accelerations x system at the XPD location and the support reaction (Rx) as indicated in Fig. 2(b). In addition, the percentage energy dissipated (Ed) by the XPD (as expressed by Eq. (11)) is also noted. The relative displacements and the absolute accelerations of the piping system are crucial from a design point of view of the XPD and the piping system and the reactions at the support are directly proportional to the forces exerted on the piping system. In contrast, the percentage energy dissipated reects the effectiveness of the XPD for seismic control of the piping system. In Table 2, the peak response quantities of the piping system with and without an XPD are compared under all ground motions. The results are tabulated for an XPD of size a and b 60 mm and for four different thicknesses of the XPD (i.e. 25 mm). The percentage energy dissipated in the piping system is also compared for the XPD of various thicknesses. Reduction in the peak responses of the piping system is noted under all ground motions for higher thicknesses of the XPD. However, only in the case of the Northridge earthquake is the response of the piping system with a 2 mm thick XPD marginally increased. This is because of the typical frequency contents of the Northridge earthquake motion. The time variation of the relative displacement, absolute acceleration at the XPD location and the support reaction at the end support of the controlled piping system under the Northridge earthquake are shown in Fig. 3 for t 2 mm and 4 mm along with the corresponding

where a0 6=DT 2 , a1 6=DT , a2 3, b0 3=DT , b1 3 and b2 DT =2, and the superscript denotes the time. Substituting Eqs. (15) and (16) in Eq. (14), we get ^ fDug fDP ^ g U fDF u g, K where ^ a0 M b0 C K U K f , K ^ M rfDu g g M a1 fu _ T g a 2 fu T g DP _ T g b 2 fu T g. C b1 fu 19 After solving for incremental displacement vector from Eq. (17), the incremental acceleration and velocity vectors are obtained from Eqs. (15) and (16), respectively. Finally, the displacement and velocity vectors are obtained using Eqs. (20) and (21), respectively, as given below: f u T DT g f u T g f D u g , _ T DT g f u _ T g fD u _ g. fu (20) (21) (18) (17)

Eq. (14) can be solved by an iterative technique. The iterations in each time step, DT, are required due to dependence of the incremental damper force, fd, on the response of the system (see Eqs. (6) and (7)). The solution of the rst-order nonlinear differential Eq. (7) for evaluation of incremental damper force is carried out using the RungeKutta method. The DT is expressed by dT DT , Nd (22)

Table 1 Peak ground acceleration of various ground motions Earthquake Recording station Component Peak ground acceleration (g) N00E N00E N00E N90E 0.348 0.843 0.57 0.629

Imperial Valley, 1940 El-Centro Northridge, 1994 Sylmar Converter Station Loma Prieta, 1989 Loas Gatos Presentation Center Kobe, 1995 JMA

ARTICLE IN PRESS
S.V. Bakre et al. / International Journal of Pressure Vessels and Piping 83 (2006) 672685 Table 2 Peak response quantities for piping system without and with X-plate damper under various ground motions (for a 60 mm and b 60 mm) Piping system Imperial Valley, 1940 Northridge, 1994 Loma Prieta, 1989 Kobe, 1995 677

xp (mm) Rx (kN) Ed (%) xp (mm) Rx (kN) Ed (%) xp (mm) Rx (kN) Ed (%) xp (mm) Rx (kN) Ed (%) Without damper With X-plate damper, With X-plate damper, With X-plate damper, With X-plate damper, 11.397 t 2 mm 10.019 t 3 mm 9.189 t 4 mm 6.306 t 5 mm 3.912 1.561 1.362 1.275 0.918 0.631 42.76 70.59 77.39 86.52 13.813 13.983 11.744 11.046 7.913 2.255 2.274 1.91 1.754 1.193 43.48 71.19 82.93 87.28 15.663 14.9 13.327 12.746 12.17 2.246 2.167 1.96 1.915 1.853 40.6 67.02 77.39 79.69 24.47 21.517 16.321 13.16 7.596 3.509 3.16 2.562 2.191 1.417 39.78 71.51 84.52 89.67

Fig. 3. Time variation of displacement, acceleration and support reaction of the piping system under the 1994 Northridge earthquake (for a 30 mm and b 80 mm).

responses of the uncontrolled piping system. It is evident from Fig. 3 that there is a signicant reduction in the displacements, accelerations and reactions for the piping system with an XPD. This implies that XPDs are effective in reducing the seismic response of the piping system. Moreover, the forcedeformation hysteresis and the time variation of the hysteretic energy of the XPD are shown in Figs. 4 and 5, respectively, under various earthquakes. It is noted from Figs. 4 and 5 that XPDs are very effective in

dissipating large amounts of the input seismic energy being transferred to the piping systems under all ground motions. The hysteretic energy decreases with increase in thickness of the XPD. Thus, the hysteretic energy decreases as the piping system becomes rigid. To study the effect of the properties of the XPD (i.e. a, b and t) on the seismic response of the piping system, the controlled piping system is analyzed by varying the properties of the XPD, in the practical range of its sizes

ARTICLE IN PRESS
678 S.V. Bakre et al. / International Journal of Pressure Vessels and Piping 83 (2006) 672685

Fig. 4. Hysteresis loops of the XPD for the piping system under various real-earthquake ground motions.

Fig. 5. Time variation of the hysteretic energy for various thicknesses of the XPDs for the piping system under various real-earthquake ground motions.

ARTICLE IN PRESS
S.V. Bakre et al. / International Journal of Pressure Vessels and Piping 83 (2006) 672685 679

Fig. 6. Effect of a, for various thicknesses of the XPDs, on the seismic responses of the piping system under various real-earthquake ground motions.

Fig. 7. Effect of b, for various thicknesses of the XPDs, on the seismic responses of the piping system under various real-earthquake ground motions.

ARTICLE IN PRESS
680 S.V. Bakre et al. / International Journal of Pressure Vessels and Piping 83 (2006) 672685

Fig. 8. Variation of the displacements of the piping system against properties of the XPD (a and b) for t 2 mm for the piping system under various realearthquake ground motions.

for different thicknesses (t) of the XPD (ranging from 2 to 5 mm). For each of the above combinations, the controlled piping system is analyzed under all ground motions and the peak response quantities are plotted against the properties of the XPD, a and b, respectively in Figs. 6 and 7. It is noted that the seismic response of the piping system is signicantly affected by changes in the properties of the XPD. The responses of the piping system are reducing for a particular range of lower values of a and remains unaffected for further increase in the values of a. Similarly, the responses of the piping system are reducing for a particular range of b and remains unaffected for further increase in the value of b. Moreover, the responses are reducing with increase in the thickness of the XPD. This is expected because increasing the thickness of the XPD makes the piping system more rigid thereby reducing the response values. Fig. 8 shows variation of the displacement response of the piping system against both a and b under various earthquakes. The displacement responses are more sensitive to a than b with reduction in the responses observed at

lower values of a and higher values of b. Moreover, the responses are unaffected with no further reduction in the response values for higher values of a and lower values of b. Thus, it can be predicted that further reduction in a leads to impractical values and increase in b is of no use as the responses are very little sensitive to b. Moreover, it can also be predicted that further variation in the XPD properties will not lead to any optimal combination of the XPD properties that will result in the minimum responses of the controlled piping system implying that this criterion of the response reduction of the piping system is not valid to numerically obtain the optimum properties of XPDs that yield the minimum response quantities. Hysteretic energy dissipated by any nonlinear system can be effectively used as a criterion to obtain the optimal properties of the nonlinear system. In the present problem, the optimality criterion can be based upon the hysteretic energy dissipated by the XPD. For each combination of the XPD properties, the input energy and the corresponding hysteretic energy is obtained and the total energy dissipated by the XPD (Ed, as expressed by Eq. (11)) is plotted in

ARTICLE IN PRESS
S.V. Bakre et al. / International Journal of Pressure Vessels and Piping 83 (2006) 672685 681

Fig. 9. Effect of a on the hysteretic energy for various thicknesses of the XPDs for the piping system under various real-earthquake ground motions.

Figs. 9 and 10 against a and b, respectively. The dissipated energy is plotted for four thicknesses of the XPD. It is noted that the hysteretic energy is signicantly affected by changes in the damper properties, a and b. Unlike the variation of the responses of the piping system, the plot of energy dissipated shows a particular combination of the properties of the XPD for which the percentage energy dissipated is a maximum. It is observed that for each of the thicknesses of the XPD, the plot of percentage energy dissipation against a (Fig. 9) shows a particular value of a for which the maximum percentage energy dissipation in

the controlled piping system is obtained. Similarly, the plot of percentage energy dissipation against b (Fig. 10) shows a particular value of b for which the maximum percentage energy dissipation is observed. The present piping system has a static stiffness of 388198 kN/m at the damper location. For this piping system under the Imperial Valley earthquake, Fig. 9 shows maximum energy dissipation of 84.05% for an XPD of size a 30 mm and b 60 mm (i.e. having properties Kd 2488 kN/m, Fy 1.173 kN and a 0.0258). This implies that there exists a combination of the properties of XPD, which results in maximum energy

ARTICLE IN PRESS
682 S.V. Bakre et al. / International Journal of Pressure Vessels and Piping 83 (2006) 672685

Fig. 10. Effect of b on the hysteretic energy for various thicknesses of the XPDs for the piping system under various real-earthquake ground motions.

dissipation in the controlled piping system. Moreover, it is observed that the energy dissipation in the piping system is more sensitive to the variation in a than b. Fig. 11 shows the plot of percentage energy dissipated in the controlled piping system against the properties of the XPD. It is conrmed from Fig. 11 that there exists a certain combination of the properties of the XPD for which the maximum energy dissipation is obtained. However, it is to be noted here that the optimal combination of the properties of XPD obtained is not the global optimal solution and varies with respect to the given piping layout. Thus, there exist different solutions dependent on the type,

layout and the earthquake excitation used for analyzing the piping system. Application of XPDs in a piping system increases the stiffness of the entire system depending on the number of XPDs used. Though the stiffness of an individual XPD may be small compared to the stiffness of the piping system, it affects the free vibration characteristics of the piping system. Moreover, using a large number of XPDs in a piping system will signicantly affect the free vibration characteristics of the piping system. Therefore, to study the effect of the properties of the XPD on the free vibration characteristics of the piping system, the fundamental

ARTICLE IN PRESS
S.V. Bakre et al. / International Journal of Pressure Vessels and Piping 83 (2006) 672685 683

Fig. 11. Variation of the hysteretic energy in the XPD against properties of the XPD (a and b) for the piping system under various real-earthquake ground motions.

frequency of the piping system is plotted in Fig. 12 against the damper properties, a and b. The frequencies are plotted for two sizes of XPD and for four thicknesses. It is observed that the frequency of the piping system is notably affected by change in b and is very sensitive to a, where signicant reduction in the frequency is noted with increase in a. This is expected as increasing a makes the piping system more exible and reduces the natural frequency whereas increase in the thickness of the XPD makes the piping system rigid thereby increasing the natural frequency. The effect of a and b on the natural frequency of the piping system is crucial as it may bring the system into the zone of high acceleration amplitudes of the input ground motions, consequently attracting more earthquake forces. 8. Conclusions A numerical study is presented in this paper that investigates the seismic effectiveness of the X-plate damper (XPD) for piping systems in industrial installations. The seismic responses of a spatial piping system are then studied under important parametric variation of the

damper properties under real earthquake ground motions to obtain the optimum properties of the XPD. The damper properties considered are height, width and thickness of the XPD. The effect of damper parameters on the response quantities of the piping system is studied for damper width and height in the practical range of 20100 mm for four different thicknesses of the XPD in the range of 25 mm. However, it is observed that the criterion to obtain minimal response quantities with variation in the damper properties does not yield any optimal solution. Therefore, the role of the hysteretic energy dissipated by the XPD is also studied by using the percentage energy dissipation in the piping system as a criterion to decide the optimal combination of the properties of the XPD. The effect of the damper properties on the natural frequency of the piping system is also investigated. Based on the trends of the results, the following conclusions are drawn. (1) XPDs are very effective in reducing the seismic response of piping systems. (2) The effectiveness of the XPD increases as the percentage energy dissipated by the XPD increases, implying

ARTICLE IN PRESS
684 S.V. Bakre et al. / International Journal of Pressure Vessels and Piping 83 (2006) 672685

Fig. 12. Effect of the properties of the XPD (a, b and t) on the natural frequency of the piping system.

that the dissipated energy controls the effectiveness of the XPD in the controlled piping system. (3) There exist optimal combinations of the properties of an XPD for which maximum energy is dissipated by the XPD in the controlled piping system. The energy dissipated in the piping system is dependent on the thickness of the XPD and the input ground motion. (4) The percentage energy dissipated by the XPD in the controlled piping system is higher for XPDs having lower values of a (half the height of the XPD), and higher values of b (width of the XPD). (5) The natural frequency of the system based on the initial stiffness of the XPD is signicantly affected by changes in a and b. Moreover, it is observed to be very sensitive to a. The frequency of the piping system signicantly reduces with increase in a, as the piping system becomes more exible and increases with increase in the thickness of the XPD.

References
[1] Olson DE, Tang YK. Decreasing snubber inservice inspection costs through snubber reduction and improved test limits. Nucl Eng Des 1988;107(12):18399. [2] Cloud RL, Anderson PH, Leung JS. Seismic stops vs. snubbers, a reliable alternative. Nucl Eng Des 1988;130(3):41133. [3] Jonczyk J, Gruner P. Loads of piping systems due to malfunctions of snubbers. Nucl Eng Des 1991;107(1-2):20513.

[4] Kunieda M, Chiba T, Kobayashi H. Positive use of damping devices for piping systemssome experiences and new proposals. Nucl Eng Des 1987;104(2):10720. [5] Shimuzu N, Suzuki K, Watanabe T, Ogawa N, Kobayashi H. Large scale shaking table test on modal responses of 3D piping system with friction support. Seismic Eng ASME PVP 1996;340:26975. [6] Kelly JM, Skinner RI, Heine AJ. Mechanisms of energy absorption in special devices for use in earthquake resistant structures. Bull NZ Soc Earthquake Eng 1972;5(3):6388. [7] Skinner RJ, Kelly JM, Heine AJ. Hysteresis dampers for earthquakeresistant structures. Earthquake Eng Struct Dyn 1975;5(3):28796. [8] Tyler RG. Tapered steel energy dissipators for earthquake resistant structures. Bull NZ Nat Soc Earthquake Eng 1978;11(4):28294. [9] Kelly JM. The design of steel energy-absorbing restrainers and their incorporation into nuclear power plants for enhanced safety, vol. 2; development and testing of restrainers for nuclear piping systems. Report no. UCB/EERC 80/21, Earthquake Engineering Research Center, University of California, Berkeley, California, 1980. [10] Schneider SS, Lee HM, Godden WG. Piping seismic test with energy absorbing devices. EPRI NP-2902, Electric Power Research Institute, 1983. [11] Bergman DM, Goel SC. Evaluation of cyclic testing of steel-plate devices for added damping and stiffness. Report no. UMCE 87-10, University of Michigan, Ann Harbor, MI, 1987. [12] Chiba T, Kobayashi H, Aida S. Application of elasto-plastic dampers to piping systems. Seismic Eng ASME PVP 1988;133:1016. [13] Namita Y, Yoshinga T, Shibata H, Kunieda M, Hara F, Suzuki K, et al. Development of the energy absorber and its application to piping systems in nuclear power plant. Seismic Eng ASME PVP 1991;211 p. 51. [14] Shibata H, Kunieda M, Hara F, Suzuki K, Ichihashi I, Fukuda T, et al. Development of the elasto-plastic damper as a seismic support for piping systems in nuclear power plants. Seismic Eng. ASME PVP 1991;211 p. 63.

ARTICLE IN PRESS
S.V. Bakre et al. / International Journal of Pressure Vessels and Piping 83 (2006) 672685 [15] Whittaker AS, Bertero VV, Thompson CL, Alonso LJ. Seismic testing of steel plate energy dissipation devices. Earthquake Spectra 1991;7(4):563604. [16] Aiken AD, Nims DK, Whittakar AS, Kelly JM. Testing of passive energy dissipation systems. Earthquake Spectra 1993;9(3):33570. [17] Namita Y, Yoshinga T, Shibata H, Kunieda M, Hara F, Suzuki K, et al. Development of energy absorber and its application to piping system in nuclear power plants. J Press Vessel Piping Div ASME 1997;211:517. 685 [18] Parulekar YM, Reddy GR, Vaze KK, Kushwaha HS. Elasto-plastic damper for passive control of seismic response of piping systems. BARC Internal Report no. BARC/2003/E/028, Reactor Safety Division, BARC, Mumbai, 2003. [19] Bakre SV, Jangid RS. Simplied seismic analysis of piping systems with energy dissipating devices. Project no. 2002/36/7-BRNS/469, Indian Institute of Technology, Bombay, 2004. [20] Wen YK. Methods of random vibration for inelastic structures. Appl Mech Rev ASME 1989;42(2):3952.