Sunteți pe pagina 1din 6

Stessa 2009 Mazzolani, Ricles & Sause (eds) 2009 Taylor & Francis Group, London, ISBN 978-0-415-56326-0

Design approach for the seismic strengthening of an existing RC building with buckling restrained braces
L. Di Sarno
University of Sannio, Benevento, Italy

G. Manfredi
University of Naples, Federico II, Italy

ABSTRACT: Structural designers may adopt either traditional interventions, such as strengthening of members and connections or global strategies, or utilize innovative schemes for seismic retrofitting of existing Reinforced Concrete (RC) structures designed for gravity loads. The latter schemes may include base isolation or supplemental damping and a combination of them. Bracing systems is very cost-effective for seismic retrofitting of RC multi-storey frames. In particular, innovative Buckling-Restrained Braces (BRBs) are a viable option because of their stable energy dissipation capacity especially under moderate-to-high magnitude earthquakes. The present paper focuses on the seismic performance of a RC framed school building retrofitted by means of BRBs. The latter enhance the energy dissipation capacity of the RC structure designed for gravity loads only and increase significantly its damping capacity. Detailed non linear static analyses were carried out in order to investigate the inelastic performance of the structure and to validate the value of the force-reduction (or behaviour-) factor adopted in the spectral analyses. The results of the performed analyses demonstrate that the use of BRBs is extremely cost-efficient. Notwithstanding, the design of such structural components is not straightforward. A step-by-step procedure, compliant with the performance-based (force- and displacementbased) framework is outlined. The outcomes of the present preliminary work demonstrate that further numerical studies are required to assess the reliability of the proposed design formulation and to validate the values of equivalent damping factors and force-reduction factors. 1 INTRODUCTION softening due to the Bauschinger effect, the hysteretic behaviour of CBFs with traditional steel braces is unreliable. Alternatively, unbonded braces (UBs), or buckling-restrained braces (BRBs), may be employed as diagonal braces in seismic retrofitting of steel and RC frames designed for gravity loads only. Such braces exhibit compressive strength which is about 1015% greater than tensile; the global buckling is inhibited (Iwata et al., 2000). Frames with UBs are being used for new and existing structures worldwide (e.g. Bozorgnia and Bertero, 2004; Di Sarno, 2006; among many others), especially for damage controlled structures as shown pictorially in Figure 2 and initially formulated by Wada et al. (1997). The global response of the inelastic structural system can be assumed as the sum of the elastic frame (also termed primary structural system) and the system formed by the diagonal braces (secondary system) that absorbs and dissipates large amount of hysteretic energy under earthquake ground motion. The primary system is capable to withstand vertical loads and behaves elastically under earthquake loads. The secondary system includes the dissipative members and is thus designed to damp the seismic lateral actions and deformations. Dissipative members, such

Framed systems are extensively used for building structures in earthquake-prone regions because of their seismic performance. However, a number of existing reinforced concrete (RC) framed building structures, were designed for gravity loads only and hence do not possess adequate lateral stiffness and resistance; seismic detailing is also lacking. It is, therefore, of paramount importance to retrofit such existing building framed structures and enhance their seismic performance. A number of intervention schemes, either traditional or innovative, are available (e.g. Soong and Spencer, 2002; Christopoulos and Filiatrault, 2006), as shown pictorially in Figure 1. Framed structures may be retrofitted by using diagonal braces, either traditional steel or innovative (Di Sarno and Elnashai, 2002). Braced systems exhibit high lateral stiffness and strength under moderate-to-large magnitude earthquakes. The most common structural configurations for lateral-resisting systems are concentrically brace frames (CBFs), which possess a lateral stiffness significantly higher than that of unbraced frames, e.g. moment resisting frames. Nevertheless, due to buckling of the metal compression members and material

841

TRADITIONAL STRUCTURE F v

DAMAGE CONTROLLED STRUCTURE F v F v F v

BARE FRAME+DAMPERS (COMPLETE STRUCTURE)

BARE FRAME (PRIMARY PRIMARY)

DAMPERS (SECONDARY SYSTEM)

Figure 3. Earthquake response of traditional framed (top) and damage controlled (bottom) structural system.

2
Figure 1. Seismic retrofitting intervention schemes.

UNBONDED BRACES

Complete Structure

Framed System (Elastic)

Energy Dissipation System

Figure 2.

Damage controlled structure.

as UBs, may be installed in the exterior frames of multi-storey buildings and can be thus easily replaced in the aftermath of a devastating earthquake. Primary and secondary systems act as a parallel system; the lateral deformation of the structure as a whole corresponds to the deformation of both primary and secondary systems. Figure 3 compares the earthquake response of a traditional frame and damage controlled structural system. The response is expressed in terms of cyclic action-deformation relationships. When controlled damage strategy is adopted, the primary structure shows a linear elastic response both under moderate and high magnitude earthquake. The energy dissipation is localized merely in the diagonal braces acting as dampers. Conversely, traditional framed systems dissipate seismic energy either within all members of the structure or in the beams, if the capacity design rules are employed. This paper assesses the seismic structural performance of a RC framed school building retrofitted with UBs. The results of comprehensive non linear analyses showed that the use of BRBs is extremely cost-efficient. Notwithstanding, the design of such structural components is not straightforward. A step-by-step procedure, compliant with the performance-based (force- and displacement-based) framework is outlined hereafter. A brief discussion of the pros and cons of the use of the UBs is presented in the next paragraph.

The disadvantages of traditionally braced frames may be prevented whether the occurrence of buckling for the metallic braces in compression is inhibited, e.g. using unbonded braces (UBs) or buckling-restrained braces (BRBs). The energy dissipation capacity of traditional metal braces is limited by the occurrence of buckling and hence stiffness reduction and strength degradation may occur. Conversely, UBs exhibit large and stable hysteretic dissipation even at large amplitudes. Unbonded braces may be employed within damage controlled structures (see also Figure 2). Under high magnitude earthquake ground motion the UBs exhibit large residual deformations and should be replaced. A typical UB is shown in Figure 4, it consists of a steel ductile core designed to yield both in tension and compression. The core is placed within a hollow section member, filled with either mortar or concrete. The outer tube prevents the occurrence of the buckling of the brace and, in turn, the hysteretic dissipation is augmented. The confinement of the outer tube may also increase the compressive resistance of the braces (e.g. Bozorgnia and Bertero, 2004). The transfer of the axial load of the core to the outer tube is minimized by the frictionless material employed between the inner metal plate and the mortar. The onset of brace buckling may be prevented by assuming (Watanabe et al., 1988): Pe 1, 0 Py (1)

where Py is the yielding resistance of the inner core (yielding and restrained elements displayed in Figure 4) and Pe is the Eulerian critical load of the outer tube, given as follows: Pe = 2 E I OT LOT 2 (2)

where E is the Young modulus, IOT is the moment of inertia of the outer steel tube and LOT is the distance between two adjacent contra-flexure points of the diagonal brace. Eqn. (2) does not account adequately for the effect of the mortar. The latter

842

5mm

10mm

150mm

200 Axial Foce (kN) 100 0 -100 -200 -300


-16 -12 -8 -4 0 4 8 12 16
Displacement (mm)

The estimation of the optimum parameters for the dissipative braces by using simplified methods; The application of capacity design checks for all members of the structure under the expected ultimate force induced by the dissipative braces, e.g. the yielding force of the UBs; The verification of the design performance, preferably through nonlinear response history analyses. Nonlinear dynamic analyses utilizing a suite of spectrum-compatible records, either artificial or natural, are yet not suitable for design office use. Thus linear dynamic or non linear static analyses are employed hereafter. The proposed design methodology for damage controlled structures is an iterative strategy based on response spectra and an equivalent viscous damping () used to quantify the effective hysteretic global response of the earthquake-resistant (see Figures 2 and 3). The selected damping can be utilized to estimate both design spectral accelerations and displacements. The design method formulated herein is a mixed force- and deformation-based scheme employing an equivalent inelastic static approach. The evaluation of the spectral displacements is essential to ensure the elastic response of the structure to be retrofitted. The step-by-step design procedure is as follows: 1. Determine the seismic base shear (Vb) using the 5% damped acceleration spectra; 2. Distribute the seismic horizontal forces along the building height. For ordinary low-to-medium rise frames the distribution can be assumed linear (Kim and Seo, 2004), e.g. compliant with the code-based formulation: Fi = Vb z i Wi z j Wj (5)

Figure 4. Typical unbonded braces: layout (left) and response (right) to cyclic loading at increasing amplitudes.

enhances the flexural capacity of the steel outer tube; however, such beneficial effects are conservatively not included in eqn. (2). Additionally, under cyclic loading, an increase of about 30% in the compression resistance of the brace should be included. As a result, the design of UBs should comply with the following relationship: Pe P 1, 0 or e 1, 3. 1, 3 Py Py (3)

Existing UBs are chiefly patented systems and their layout may vary according to the manufacturer. Dissipative braces used as UBs may, for instance, include a traditional metallic diagonal member in series with a hysteretic buckling-restrained member, e.g. the buckling-restrained axial damper (BRAD) manufactured by FIP Industriale in Italy. The BRAD devices comprise of a steel hollow section tube with a low yield metal dissipative core. The device ends with a flanged bolted connection and a pinned or a gusset plate. The detailing of the gaps between the joint components is of paramount importance for the effective response under earthquake ground motions. The manufacture provides the geometrical and mechanical properties of the dissipative device. The latter may be utilized to compute the yield deformation (br,y) of the in-series system formed by the BRAD and the traditional steel brace through the following relationship: br , y = br , y L br cos (4)

where br,y is the yield deformation of the brace, Lbr is the length of the brace and is the angle of the diagonal with respect to the horizontal beam. The procedure used for the design of the braces is presented hereafter. 3 SEISMIC RETROFITTING STRATEGY

where W is the seismic mass and z is the storey height with respect to the foundation level. Alternatively a modal analysis can be carried out for the existing structure; 3. Determine the axial forces (Fbr) in the diagonal braces assuming that the existing frame has pinned beam-to-column and base-column connections and the braces are effective to resist earthquake loads: Fbr = 1 Vi n cos (6)

The design of new and existing structures with hysteretic dampers generally comprises (Christopoulos and Filiatrault, 2006):

where Vi is the seismic storey shear at the i-th floor, n the number of storeys and is the angle of the braces with respect to the horizontal beams; 4. Determine the cross-section of the core of the dampers at each storey:

843

Acore =

Fbr fy

(7)

CASE STUDY

where fy is the steel yield stress of the core element of the UBs. Material overstrength may also e accounted for; 5. Perform non linear static analyses of the existing frame with the added braces and determine the capacity curves of the earthquake-resistant system. The maximum displacement demand is also computed by employing the equivalent viscous damping: = i + h (8)

where the damping h accounting for the hysteretic behaviour is: h = Ep 4 Es = Ep 1 2 d max Fmax (9)

6. 7. 8. 9.

10.

with Ep the hysteretic energy dissipated in a cycle and Es the elastic energy stored in the system. The value of Ep can be computed by performing a cyclic pushover and assuming a target displacement of the control point equal to the maximum lateral displacement demand of the structure retrofitted with the UBs at collapse limit state. As a rule of thumb, the latter displacement can be assumed equal to a global drift of 0.5%0.6% of the building height. The initial damping i accounts for all sources of dissipation in a structure which do not include hysteretic dissipation; it may be assumed equal to 5%; Determine the maximum displacement demand, using the displacement response spectrum and the equivalent damping; Check that the existing frame responds elastically for the computed displacement demand; alternatively iterate steps 1) to 6); Compute the seismic forces by using the acceleration response scaled through the equivalent damping in eqn. (8); Use the displacement demand estimated in 7) and the capacity curve in 5) to compute the effective base shear in the existing frame and the added braces. The latter may be re-designed and optimised by iterating the steps 1) to 7); Check the maximum axial loads in the braces at damageability limit state to prevent yielding under service loads.

The case study illustrated in the present work is a typical two-storey RC school building located in the South of Italy. The plan layout of the building is irregular; it consists of three main blocks: two T-shape block and a rectangular block. The T-shape unit is about 31 m long, it has a width of 28.5 m, the web thickness of 14.7 m and two 7 m-long offsets. The discussion hereafter focuses on the seismic performance assessment of the above T-shape unit. Seismic joints were utilized to separate the three main blocks of the school building. The ground floor of the structure is 3.08 m high; the first and second floors are 3.65 m high. The top floor has an inclined tiled roof: its height varies between 0.2 m and 1.90 m. The ground floor has rigid RC walls along the perimeter and it is assumed to provide fixity to the upper floors; thus the structural model assessed herein includes only two floors and it is fixed at the base. Additional design data of the sample school building may be found in Di Sarno et al. (2007). The existing framed building exhibits high seismic vulnerability and hence a retrofitting intervention was adopted. The latter consists of BRBs placed along the perimeter of the existing frames. The seismic response of the bare and the retrofitted RC frames is analysed and compared in the next paragraphs. 5 SEISMIC PERFORMANCE ASSESSMENT

To estimate the expected plastic mechanisms and the distribution of damage in the sample framed buildings, non linear static (pushover) analyses were carried out both for the as-built and retrofitted structural systems. Two horizontal force patterns were employed for the seismic structural assessment: A modal pattern, proportional to lateral forces consistent with the lateral force distribution in the direction under consideration determined in the elastic analysis; A uniform pattern based on lateral forces that are parallel to mass regardless of elevation (uniform response acceleration). It is worth mentioning that for the existing structure, the non linear static analysis was not compliant with the provisions implemented in the seismic codes of practice (e.g. EC8, 2006; DM, 2008). The participation mass of the mode of vibration along the x-direction is indeed lower than 75% (67,48%). Conversely, the retrofitted structure possesses participation masses higher than 85% along both x- (85,50%) and y-directions (89,54%). Further details on the dynamic response of the sample structures can be found in Di Sarno and Manfredi (2009). Figure 5 provides the response curves of the as-built structure; the performance points at damage (DLS), life safety

The design methodology illustrated above quantify the dissipation through the equivalent viscous damping, used to scale the acceleration and displacement response spectra. Alternatively, adequate response modification factors (R- or q-factors) may be employed; values of R ranging between 4.5 and 6.5 have been proposed (Sabelli et al., 2003).

844

25
Base Shear / Seismic Weight (%)

Table 2. Inter-storey drifts of the as-built structure, at different limit states (in percentage). Limit state

20

15
X M XM DLS

10

Force pattern XM
2.00 2.50

Floor level 1st 2nd 1st 2nd 1st 2nd 1st 2nd

Damage 0.247 0.656 0.344 0.410 0.312 0.334 0.372 0.326

Life safety 0.468 1.923 0.839 1.346 0.840 1.057 1.060 0.976

Collapse prevention 0.485 2.251 0.982 1.766 0.971 1.265 1.820 1.477

LSLS CPLS

0 -2.50 -2.00 -1.50 -1.00 -0.50 0.00

0.50

1.00

1.50

YM XM

Lateral Displacement / Height of the Control Point (%)

Figure 5. Capacity curves of the as-built structure (X-direction). Key: DLS = damage limit state; LSLS = life safety limit state; CPLS = collapse prevention limit state. Table 1. Lateral drifts of the as-built structure, at different limit states (in percentage). Limit state Force pattern XM YM XM YM Damage 0.328 0.370 0.324 0.348 Life safety 0.961 1.085 0.949 1.019 Collapse prevention 1.215 1.371 1.199 1.287

YM

60 Base Shear / Seismic Weight (%) 50 40 30 20 10


XM XM DLS LSLS CPLS

0 -2.50 -2.00 -1.50 -1.00 -0.50 0.00 0.50 1.00 1.50

2.00 2.50

Lateral Displacement / Height of the Control Point (%)

(LSLS) and collapse prevention (CPLS) limit states are also included. The computed results show that for the modal load distribution the displacement demands relative to LSLS and CPLS cause structural instability. As a result, seismic strengthening was aimed at enhancing the global lateral stability of the building. When subjected to the uniform load pattern, the frame is severely damaged, but the system is stable. The displacement demand for the as-built structure are summarised in Table 1; the estimated values are similar for both positive and negative loading directions, along the x- and y-directions. The ductility demand is about 3.7. The values listed in Table 1 show that at DLS the maximum drift, i.e. 0.370% is lower than 0.5%. Nonetheless, the computed global lateral displacements do not ensure that the damage in the as-built structure is prevented. The assessment of storey capacity curves demonstrates that the inter-storey drifts (see Table 2) of the second floor exceed the 0.5% threshold at DLS and becomes greater than 2.0% at CPLS, thus leading to progressive collapse (storey mechanism). The horizontal displacement patterns highlight a significant variation of the lateral stiffness at the second floor of the existing frame, especially at LSLS and CPLS. As a result, high inelastic demand under severe earthquake ground motion is concentrated at the second storey of the building.

Figure 6. Capacity curves of the retrofitted structure (X-direction). Key: DLS = damage limit state; LSLS = life safety limit state; CPLS = collapse prevention limit state.

The use of UBs enhances significantly the lateral stiffness and global strength. The design of the added braces was carried out in compliance with the procedure illustrated earlier. The elastic lateral stiffness of the retrofitted frame increases of about 150% along X-direction and 140% along y-direction. The stiffness of the selected dissipative braces is effective to regularize the lateral displacement of the existing frame under earthquake-induced horizontal forces. The inelastic demand at the second floor is thus prevented and enhanced ductile response achieved. The capacity curves for the retrofitted system are provided in Figure 6. The displacement demands at DLS, LSLS and CPLS are also included in the plots. The comparison of the capacity curves shown in Figures 5 and 6 proves the efficiency of the use of UBs to enhance the seismic response of the as-built structure. The displacement demand is halved. Moreover, the retrofitted frame exhibits a regular response height-wise. The values of the inter-storey drifts computed for the uniform and triangular lateral load distribution in the retrofitted building and provided in Table 3 are much lower than those estimated for the existing structure (see Table 2).

845

Table 3. Inter-storey drifts of the retrofitted structure, at different limit states (in percentage). Limit state Force pattern XM YM XM YM Floor level 1st 2nd 1st 2nd 1st 2nd 1st 2nd Damage 0.201 0.179 0.237 0.229 0.141 0.177 0.181 0.239 Life safety 0.533 0.529 0.656 0.706 0.279 0.580 0.387 0.853 Collapse prevention 0.666 0.652 0.839 0.882 0.347 0.761 0.457 1.105

values ranging between 4.5 and 6.5. Further analytical work based on detailed inelastic response history analyses is ongoing to assess the reliability of the proposed design scheme and the suggested values of response modification factors for buildings with UBs.

REFERENCES
Bozorgnia, Y., & Bertero, V . (2004). Earthquake Engineer.V ing. From Engineering Seismology to Performance-Based Engineering. CRC Press, Boca Raton, Florida, USA. Christopoulos, C., & Filiatrault, A. (2006). Principles of Passive Supplemental Damping and Seismic Isolation. IUSS PRESS, Pavia, Italy. Di Sarno, L. (2006). Bracing systems for seismic retrofitting of steel frames. Proceedings of the 5th International Conference on Behaviour od Steel Structures in Seismic Area, Yokohama, Japan, 821826. Di Sarno, L., & Elnashai, A.S. (2002). Seismic Retrofitting of Steel and Composite Building Structures. Mid-America Earthquake Center Report, CD Release 02-01, University of Illinois at Urbana-Champaign, IL, USA. Di Sarno, L., & Manfredi, G. (2009). Seismic response analysis of an existing RC school building. Proceedings of the 3rd International Conference on Modeling, Simulation and Applied Optimization, Sharjah, UAE, CD-ROM. Di Sarno, L., Manfredi, G., & Acanfora, M. (2007). Design Approach for the Seismic Strengthening of an Existing RC Building with Buckling Restrained Braces. ASSISi 10th World Conference on Seismic Isolation, Energy Dissipation and Active Vibrations Control of Structures, Istanbul, Turkey, CD-ROM. DD.MM.LL.PP. (2008). Norme tecniche per le costruzioni NTC (in Italian). Eurocode 8 (2006). Design provisions for earthquake resistance of structures. Part 1.3: General rules. Specific rules for various materials and elements. Eur. Comm. for Standardisation, Brussels, Belgium. Iwata, M., Kato, T., & Wada, A. (2000). Buckling-restrained braces as hysteretic dampers. Proceedings of the 3rd International Conference on Behavior of Steel Structures in Seismic Areas (STESSA 2000), Montreal, Canada, 3338. Kim, J., & Seo, Y. (2004). Seismic design of low-rise steel frames with buckling-restrained braces. Engineering Structures, 26(5), 643651. Sabelli, R., Mahin, S., & Chang, C. (2003). Seismic demands on steel braced frame buildings with buckling-restrained braces. Engineering Structures, 25(5), 655666. Soong, TT., & Spencer, B.F. Jr. (2002). Supplemental energy dissipation: state-of-the-art and state-of-practice. Engineering Structures, 24(3), 243259. Wada, A., Iwata, M., & Huang, Y.H. (1997). Seismic design trend of tall steel building after Kobe earthquake. Passive energy dissipation and control vibration structures, Taormina, Italy, 251269. Watanabe, A., Hitomi, Y., Yaeki, E., Wada, A., & Fujimoto, M. (1988). Properties of brace encased in bucklingrestrained concrete and steel tube. Proceedings of the 9th World Conference on Earthquake Engineering, TokyoKyoto, Japan, 4(6), 719724.

Table 4. Global overstrength, translational ductility and response modification factor of the retrofitted structure. Load pattern Response parameter q XM 2.54 2.17 5.51 YM 2.03 2.36 4.79 XM 2.13 2.07 4.40 YM 2.14 2.27 4.86

The all-encompassing response factor (R- or q-factor), computed as q = ., is thus nearly equal to 5.0; the latter values is implemented in many seismic codes world-wide (e.g. EC8, 2006; DM, 2008) for ordinary capacity-designed moment-resisting frames. The computed values of the response factor are similar to those proposed for the design of new steel framed systems with UBs (Sabelli et al, 2003). The retrofitted system shows values of overstrength () and translational ductility () higher than 2.0 as shown in Table 4. 6 CONCLUSIONS

The results of extensive non linear analyses carried out on a sample reinforced concrete (RC) school retrofitted with unbonded braces (UBs) were presented. It is shown that UBs are cost-efficient for the seismic retrofitting of existing RC framed school buildings designed for gravity loads only. A step-bystep procedure, compliant with the performance-based (force- and displacement-based) framework was also illustrated and applied to the sample structure. Equivalent viscous damping or response modification factors may be adopted for the design of new steel frames with UBs or for existing RC frames retrofitted with UBs. The computed values of response modification factors for retrofitted RC with UBs are in agreement with those estimated for steel framed structures, i.e.

846

S-ar putea să vă placă și