Enhancing the seismic performance of building using damage-avoidance shear wall hold-downs

. Concrete coupled wall is one of the outstanding seismic resisting systems for mid to high-rise building structures located in a high seismic region. In this system, the link beam provides coupling action between the beam and the adjacent wall panel, which significantly increases the wall system's lateral stiffness and eventually reduce the footprint for the seismic resisting system. However, once the link beam is damaged due to a severe earthquake, it is difficult, costly, and time-consuming to be repaired, which results in business disruption and the increased building life cycle cost. To deal with these issues, the Resilient Slip Friction Joint (RSFJ) has recently been introduced and employed in the New Zealand construction industry. This novel technology does not only aim to provide “ life safety ” but also “ immediate occupancy ” criteria. The flag-shaped hysteresis of the RSFJs provides the required seismic performance, including a self-centring behaviour. This paper addresses the enhanced seismic performance of the structures using this system. An eight-storey reinforced concrete building is designed using the conventional forced-based design method, and its seismic performance is evaluated by non-linear static pushover and nonlinear dynamic time-history simulations. The results showed that this system can provide a high level of structural ductility while providing fully self-centring behaviour.


Introduction
Studies showed that the concrete coupled wall is one of the most effective seismic resisting systems for mid to high-rise building structures, particularly for those located in an active seismic region.In this system, the link beam provides coupling action between the beam and the adjacent wall panel, which significantly increases the wall system's lateral stiffness and eventually reduces the footprint for the seismic resisting system.However, once the link beam is damaged due to a severe earthquake, it is difficult, costly, and timeconsuming to repair, which results in business disruption and increased building life cycle costs.
To comply with the "life-safety" and "immediate occupancy" criteria, the researchers conducted many studies to develop the low damage concept for concrete structures, where they still have the benefits of concrete shear walls.The general principle of the low-damage design concept is to control damage by using rocking connections, which are typically combined with supplemental damping devices, to dissipate energy.These damping devices act as sacrificial fuses which can be easily repairable or replaceable after moderate to severe earthquakes.Housner [1] initially introduced the rocking concept by undertaking an experimental study on a rigid rocking block.He reported that the rocking motion had a considerable contribution in enhancing the structural stability against the over-turning moment.Aslam et al. [2] studied the rocking response of rigid *Corresponding author: luhur.budi86@yahoo.comstructures and stated rocking resistance of the structure can be improved by anchoring it to the ground.Priestley et al. [3] developed a hybrid post-tensioned connection on a concrete shear wall and reported that the prototype has an extremely satisfactory performance when subjected to severe seismic force.Yielding, friction, and viscous damping are proposed, tested, and used in rocking shear walls [4].
The above studies have advanced the concept of low-damage design for concrete structures, but there remains a significant literature gap.The poor repairability of the sacrificial elements requires the element replacement after a severe earthquake which is costly and causes business disruption for the building.In addition, as the residual capacity of these sacrificial elements will not be sufficient to resist severe aftershocks, the structure will still be vulnerable until getting fully repaired by replacing the damaged fuses.Furthermore, regular inspection is required to deal with the creep issue in post-tensioning steel strand cables and their supporting members.Therefore, there is still a remaining step to achieve an ideal maintenance-free rocking wall system which provides an efficient and high-performance lateral load-resisting system and excellent self-centring mechanism during earthquakes and aftershocks.
To respond to the above challenges, Zarnani and Quenneville [5] introduced a new technology, known as the Resilient Slip Friction Joint (RSFJ).It provides essential energy absorption and self-centring capability, and other excellent features, including secondary fuse activation, easy repairability, and free maintenance, to ensure life "safety" and "immediate occupancy" criteria post-earthquake.This research project proposed a concrete coupled wall structure using the RSFJ to produce an efficient and high-performance lateral loadresisting system.
The objective of this research is to analyse and assess the implementation of the RSFJ device in concrete coupled wall systems.The basic equations and the hysteresis curve of the RSFJ are introduced.The specific configuration of the RSFJ will be proposed, and the procedure to design RSFJ specifications will be developed.The effect of the RSFJ damping device in increasing the seismic performance of the building structure will be further investigated.Finally, the numerical modelling is developed, and its structural analysis result is verified.

Analytical model
The RSFJ technology is a novel friction joint that provides energy absorption and self-centring behaviour in one compact package.The main component of the RSFJ comprises an elongated holes steel plate, metal cap plate, high strength bolts, and disc springs, as shown in Fig. 1(a).In RSFJ, the restoring force comes from specific steel grooved plates which are tied through high-strength bolts and disc springs.By slipping of grooved plates, the input energy is dissipated through frictional resistance.Based on the free body diagrams presented in Fig. 1(c), the design procedure is developed for the prediction of the performance of the RSFJ.The slip force (Fslip) and residual force (Fres) can be determined by Equation 1-2, respectively: Where nb=number of bolts on each splice, θ=groove angle, Fb.pr is the clamping force of prestressing and the μs and μk are the static and kinetic coefficient of friction respectively, while considered μk=0.85μs[8].The general hysteresis behaviour of RSFJ is illustrated in Fig. 1(b).Fult,loading and Fult,unloading are the system forces at the maximum disc springs displacement and bolts force.
Fult,loading and Fult,unloading is derived by replacing the bolt forces in Equations 1-2 by Equation 3, and μs, μk with μk, μs, where Ks and Δs is the stiffness of the stack of Belleville springs on the bolt and the maximum possible deflection of Belleville springs after initial prestressing, respectively.

RSFJ as the hold-downs
Fig. 2(a) shows the proposed rocking wall system.The free-body diagram of a rocking shear wall with RSFJ and the hysteretic behaviour of the system is shown in Fig. 2(b).The deflection of the joints is related to their lever arm, which is the horizontal distance between the RSFJ centre to the rotating edge of the shear wall.Therefore, ΔRSFJ,1/ΔRSFJ,2=L1/L2, where ΔRSFJ and L are, the expansion of the RSFJ and the associated lever arms, respectively.
Hashemi and Quenneville [8] explained that the horizontal force acting at the top of the wall (Ftop) can be calculated from the moments of the centre of rotation (the base of the wall) [6].In this way, Ftop can be further obtained using the following Equation 4.

Fb Fb Fb Fb
Fb Fb Fb Fb In this equation, H is the height of the wall, W is the vertical loads, LW is the horizontal distance from the vertical load to the centre of rotation, and FRSFJ,slip is the slip force of the RSFJ.After the slip stage, the force within the RSFJ corresponds to the deflection within them.Therefore, Equation 1 can be used to determine the lateral strength of the wall, FRSFJ,1 and FRSFJ,2 are the forces within the tensioned and compressed RSFJ, respectively.During the loading of the wall, Equation 2 can be employed to develop a correlation between FRSFJ,1 and FRSFJ,2.Then, Equations 1-2 can be combined to determine the overall load-deformation behaviour of the wall.

RSFJ as the shear link
In the coupled walls system, the coupling beam could be replaced by RSFJ (as the shear links).Similar to the single shear wall, the horizontal displacement on all columns and walls is equal because all columns and walls is connected to the floor.Sahami et al. (2019) proposed a formula to calculate the rocking base moment (Mrock) and rocking stiffness (krock) on the coupled wall system (Equations 5-6) [7].
where nw is the number of coupled walls and  is the number of dampers on each side of the walls, respectively.Meanwhile, W is the seismic weight of the wall, B is the width of the wall and d is the length of the shear link.The numerical analysis results showed that the coupled wall system using RSFJ as the shear link could achieve the capacity of two identical single walls.

The design of the case study building
An eight-storey reinforced concrete building structure with a concrete coupled wall as its seismic resisting system is proposed in this research project to evaluate the seismic performance of the system and the adopted design philosophy.The concrete coupled walls are used in the middle of the building to minimise the torsional effects.Meanwhile, the RSFJs are used as the holddown at the bottom corner of the wall and as the shear link in the middle of the coupling beam.To satisfy the requirement of NZS1170.5:2004, the following criterion is set for the design philosophy: (a) the Ultimate Limit State (ULS) lateral drifts are kept under 1.4 % to minimize the damage to structural and nonstructural components (b) the structure remains elastic, the dampers are not activated and the lateral drift ratios are kept 1.5 % and 0.33 % for Ultimate Limit State (ULS) and Serviceability Limit State (SLS) cases, respectively (c) the structure has zero or negligible residual displacement at the end of the seismic event.
The building plan is 35 m by 35 m and is symmetrical about the two main axes.Along each axis, a concrete coupled wall equipped with RSFJ is used as the Lateral Load Resisting System (LLRS).A novel type of shear key is installed at the wall's base to provide displacement compatibility and transfer the shear forces from the wall to the foundation [8].The total building height is 28 m and has an identical storey height of 3.5 m.The building is designed for office application (i.e., importance level 3), is located in Christchurch, and is built on the soft soil layer (i.e., soil class type D).Fig. 3. illustrates the developed model used in this research project.
The loads applied during the preliminary design are a self-weight (for the frame) of 6 x 10 -3 kg/cm 2 , a superimposed dead load of 5 x 10 -3 kg/cm 2 , a floor weight of 3 x 10 -2 kg/cm 2 , a cladding weight of 10 -2 kg/cm 2 , a floor live load of 3 x 10 -2 kg/cm 2 and a roof live load of 2.5 x A ductility factor of μ=4.0 and a structural performance factor of Sp=0.7 are adopted for the design.The value of Sp is linked to the design ductility/level of detailing used and the value of 0.7 is appropriate for structures designed under NZS3101:2006.However, for other structural systems that use damage avoidance technologies, this value might be unconservative and may need to be re-evaluated.The period of the structure is determined as T1=0.43 seconds using the empirical equation provided in NZS1170.5:2004.Note that for periods equal to or more than 0.7 seconds, kμ=μ.The base shear of the structure is specified as Vb =9.11 x 10 5 kg and Vb = 1.584 x 10 6 kg for SLS and ULS cases, respectively.This base shear is distributed in the structure respecting the storey weights and heights as per the recommendations in NZS1170.5:2004.

Numerical analysis
In this section, a structural analysis using the SAP2000 program is performed to investigate the performance of the structure and to check the seismic lateral drift.A 3D model of an eight-storey building with concrete coupled walls and RSFJs is considered for the numerical modelling.The beam-column connections and base connections are designed as pinned joints while the column sections are continuous as required by NZS3101.1:2006.A rigid diaphragm is assigned to each floor to constrain the horizontal displacement of the beams.The gravity loads have been applied to the beam and the seismic weights have been assigned to the nodes in each elevation.
The "Damper-Friction Spring" element in SAP2000 is adopted to represent the load-displacement behaviour of the RSFJs as the shear link and hold-down.The accuracy of using this link element for the RSFJs has previously been verified by comparing the experimental data with numerical results [9].The shear force demand and axial force demand on the wall's toe from the elastic numerical model are used to determine the shear link and hold-down stiffness value, respectively.To achieve the rocking behaviour of the wall, the rocking foundation model which is represented by the "Gap" element in SAP 2000 is used.
The Nonlinear Push-over Analyses (NPA) are carried out to estimate the shear force for a specified lateral drift and calculate hysteresis damping values.The numerical model and analysis procedure are considered valid if the base shear and maximum roof displacement obtained from nonlinear pushover analysis is less than 5 % than that calculated using Equivalent Static Method (ESM) formula.Besides, the cyclic pushover analysis curve should represent the flag-shape hysteresis curve, which is the main characteristic of the self-centring structure.Furthermore, the nonlinear time history analysis is carried out to validate the initial ductility factor used in the design process and investigate the structure's behaviour.Seven seismic events are selected and scaled for ULS based on method described in NZS 1170.5 for the given location and soil type.The 'peak of three' or the 'average of seven' records may be considered when designing using time-history analysis [10].The numerical model and the design process are considered valid if the discrepancy between ESM estimation and NLTHA outcomes (i.e., base shear and roof displacement) are less than 5 %.

Nonlinear pushover analysis
Overall, the analysis shows that the maximum base shear and lateral displacement in the developed model are consistent with the initial estimation (i.e., ESM formula).The base shear generated from pushover analysis is 9.99 x 10 5 kg, which is slightly lower than that of the ESM calculation (1.0235 x 10 6 kg).Meanwhile, the lateral drift from pushover analysis is 1.5 % (i.e., 420 mm), which satisfies the target displacement.Moreover, the flag-shaped hysteresis curve, which is the main characteristic of the selfcentring system can be reasonably achieved.
The pushover curve shows a bi-directional performance resulting from the flag-shape hysteresis of the RSFJs.It is observed from the hysteresis curve that the structure remains elastic up to 0.2 %.After this point, the RSFJs start to activate, resulting in the observed bilinear pushover curve.It should be noted that since not all RSFJs are activated at the same time, a transition zone from the linear elastic zone to the geometrically nonlinear zone is observable in the pushover curve.Nevertheless, the entire structure remains elastic, and no material nonlinearity is expected within the transition zone.Fig. 4. depicts the final hysteresis curve obtained from pushover analysis for the proposed structural system.Fig. 4. The final hysteresis curve obtained from non-linear pushover analysis.

Nonlinear dynamic time history analysis
Overall, the base shear and roof displacement obtained from seven ground motions are consistent with those estimated with the Equivalent Static Method (ESM).The NLTHA results are about 4 % lower than that calculated with the ESM.In this study, the average of seven ground motions is considered for analysis and an equivalent ductility factor of μ=2.9 is used for the developed model.Fig. 5. illustrates the base shear and roof displacement obtained from NLTHA.NLTHA H1 and NLTHA H2 refers to analysis for x-axis and y-axis, respectively.
The chart shows that the highest base shear is obtained from the Christchurch CBGS and Chi-Chi TCU event which are accounted at 1.048 x 106 kg and 1.045 x 106 kg, respectively.Similarly, the highest roof displacement is generated from the Christchurch and Chi-Chi TCU event, which are recorded at 43.1 cm and 42.6 cm.The average base shear is accounted for at 9.824 X 105 kg, which is slightly lower than that of the ESM estimation (10.235 x 105 kg).Meanwhile, the average roof displacement is calculated at 1.46 %, which is 0.04 % lower than the target drift (i.e., 1.5 %).If the mean of the recorded drifts is taken, it reasonably satisfies the target drift and confirms the predicted seismic performance of the developed structural model.It should be emphasized that if the average base shear of the records is considered, it well below the design base shear, which confirms the acceptable seismic performance.If the maximum of all records is considered (which might be overly conservative), it is slightly above the design base shear.However, it should be noted that RSFJ has the secondary fuse feature that can tolerate up to 1.25 times of the design base shear without any damage except yielding of the clamping rods in the RSFJs.This result also confirms the efficiency of the seismic design.
As previously mentioned in section 4, the analysis result can be considered valid if the discrepancy between NLTHA results (i.e., base shear and Therefore, both the average base shear and average roof displacement from the final iteration can be reasonably accepted.

Inter-storey drifts
It can be seen in form Fig. 6. that the maximum interstorey drift of the proposed model is about 1.5 %, which is significantly lower than the requirement in the NZS 1170.5.It means that the configuration of the RSFJ and the lateral resisting system can prevent the "soft storey" formed at any building level.The highest and the lowest inter-storey drift is recorded for the Loma Prieta LGP (1.61%) and the Imperial 79 BCR (1.37 %), respectively.These values are well close to the target lateral drift, which shows the accuracy of the proposed design and analysis method.

The residual drift
Bruneau and MacRae [11] in stated that if the residual drift is larger than 0.3%, the building structure will suffer from permanent damage and need to be demolished.McCormick et al. [12] noted that this large residual drift larger than 0.3% will results in a situation where realignment of the components is required (which normally is costly), and rebuilding would be a more efficient option.Therefore, to achieve the full selfcentring behaviour, the maximum residual drift at the end of NLTHA shall be limited at 0.2% and 0.25% for the design level intensity earthquake and for the maximum considered earthquake (MCE) intensity, respectively [13].Fig. 7 shows the residual drift from Chi-Chi and Christchurch event which is less than 8mm (i.e., less than 0.2%).The residual drift from other events is less than 6mm.Therefore, these residual drifts will result in the negligible structural damage and minor cosmetic repairs after earthquake.It confirms the fully self-centring behaviour (i.e., with-out relying on any supplementary devices such as the post-tensioned strands), which can be attributed to the RSFJs used in the LLRS. (

Response of the individual RSFJ
Overall, each RSFJ performs within its capacity and achieve its desired lateral displacement.According to the analysis results, the internal forces (i.e., shear force for the shear links and axial force for the hold-downs) are lower than their ultimate force (Fult,loading).Similarly, their lateral displacements are slightly lower than the specified displacement.In other words, the secondary fuse activation is not needed for the proposed structural model.Furthermore, these individual responses of the RSFJs show the efficiency of the pro-posed structural wall system and design procedure in controlling lateral drift.It means that the self-centring behaviour does not rely on any external mechanism (i.e., gravity resisting system), which is particularly advantageous for the structural systems in which the lateral load resisting system is separated from the gravity load resisting system.Fig. 8(a) describes that the general hysteretic behaviour of the joint is consistent with the results of the quasi-static analysis conducted by Hashemi (2017).The maximum displacements are 15 mm in compression and 55 mm in tension, which the latter is close to the specified displacement for the joint (i.e., 59 mm).Fig. 8(b) illustrates that the RSFJ device at roof level has a higher ultimate force (Fult_loading = 67500 kg) than the RSFJ at the first level (Fult_loading = 49900 kg), which is relevant to the design parameter specified in section 5. 4  compression and tension, respectively.The maximum displacement in tension is slightly lower than the predetermined displacement for the joint (i.e., 59 mm).These above facts strongly imply that the proposed structural prototype has the potential to be further developed as a resilient and damage avoidant seismic solution.Moreover, the analysis results confirm that the design procedure stated in section 3. can be efficiently implemented for the mid-storey building structure when RSFJs are used as the shear links and the hold-downs.Fig. 9. depicts the load-deformation history response of the RSFJ shear link at the fourth level for the four ground motions (i.e., Christchurch, Chi-Chi, Imperial 79, and Darfield).It can be observed that the shear links performed well within the expected specifications, and the link elements could successfully represent the predicted flag-shaped hysteresis behaviour, which offers the self-centring behaviour and an excellent rate of energy absorption.It is evident that the bounded area between the hysteretic loops, which represents the dissipated energy, is increasing at a constant rate.Furthermore, it shows the shear link can maintain its strength and stiffness after numerous loading and unloading cycles.
Regarding the Christchurch and Chi-Chi event, the recorded base shears are slightly lower than the ultimate force (Fultimate_loading).Similarly, the obtained displacements are well below the specified displacement.In terms of Imperial 79 and Darfield ground motion, the base shear is about 7% lower than the ultimate force (Fultimate_loading).However, the displacement recorded from these ground motions is relatively close to the specified requirements.In summary, the results of the time-history simulations indicated that the proposed structural prototype satisfy the force and displacement demands.Also, it illustrates that the structural model is properly designed and the RSFJs are properly sized.

The achieved equivalent ductility factor
The finding of this study shows that an equivalent ductility factor (µ) of 2.9 is applicable for the proposed structural building prototype.This equivalent ductility factor (µ) is considerably higher than that obtained by a conventional coupled wall system (i.e., µ of 2.0), which means that the application of the RSFJs result in a more ductile behaviour of the structure.The RSFJs provide a higher energy dissipation than the conventional shear wall system while remaining control the lateral drift.This fact strongly reveals that the application of RSFJ results in a lower seismic base shear and a more economical design.
Regarding the design procedure, the above result indicates the efficiency and the accuracy of the proposed procedure, which is described in section 4. The proposed equations developed in section 2 can easily calibrate the key parameters of the RSFJs and produce the flagshaped hysteresis curve.By following this procedure and equation, a low number of iterations are required to generate an equivalent ductility factor for the seismic resilient structure with RSFJs.In this paper, an initial equivalent ductility factor (µ) of 2 is adopted at the start of the design because the ultimate force (Fult_loading) is defined from the base shear at ULS case (with µ=2).In this way, it requires only four iterations to obtain the optimum equivalent ductility factor.
The above design procedure is developed by following the "equivalent ductility" method and ESM formula.The primary purpose of this procedure is to define the optimum force/displacement capability of RSFJ based on the given structural performance and adjust the equipment size of the RSFJs accordingly.This procedure generally is compatible with most of the international building standards.

Conclusions
This paper discusses the seismic behaviour of building using damage-avoidance shear wall hold-downs.In this system, the application of self-centring friction damping on the building structure is proposed and its performance is analytically and numerically observed.This innovative RSFJ system is incorporated in a coupled concrete wall system with the aim of achieving damage avoidance.
To investigate the performance of the novel concept on the system level, an eight-storey concrete building structure with concrete coupled wall as Lateral Resisting System is proposed in this study.The design philosophy used was to limit the inter-storey drift ratios to 0.33 % and 1.4 % for SLS and ULS, respectively.The following conclusion is observed after the NLTHA of the structural system: 1.The developed structural model can produce an ideal flag-shaped hysteresis curve, which is a key feature for a low damage design structure.It means that the damping devices employed in the structure do not only have a good energy absorption but also an excellent self-centring behaviour.Therefore, the proposed prototype can be considered as an efficient alternative to traditionally high-damage lateral resisting systems to minimise and localise damages.In this way, the life-safety and immediate occupancy criteria can be achieved.2. The proposed RSFJs configuration can achieve an equivalent ductility factor of µ equal to 2.9, which is significantly higher than the ductility coefficient for the conventional structures (i.e., µ=2).It means that the application of RSFJs in the lateral resisting system results in a more economical design than the conventional structural design.3. The proposed configuration of the RSFJs generates a reasonably uniform lateral displacement profile for the building structure.It can be achieved by tuning the key parameters of the RSFJs.In fact, this feature enables the designers to easily specify the RSFJs parameters and story drift threshold according to the project's requirements.4. The analysis results confirmed the efficiency of the introduced design procedure for designing the building structure and sizing the RSFJs.This design procedure requires a low number of iterations to obtain the optimal equivalent ductility factor.Furthermore, this design procedure is compatible with most of the international seismic standards.

Recommendation for future research
According to the discussion in section 5 and the conclusion in section 6, the following investigations should be carried out: 1. Numerical study on the high-rise buildings structure (i.e., the number of storey is more than 20) using the same configuration of the lateral resisting system to confirm the benefit of the RSFJs damping device in the concrete coupled wall.2. Research study on the same structural building prototype with different configurations and properties of the RSFJs device to investigate its seismic performance under the MCE NLTHA.3. Numerical analysis on the same building prototype and LLRS using RSFJs device and other damage avoidance devices to investigate these damping devices' benefit in a seismic resilient building structure.

Fig. 8 .
Fig. 8. Response of individual RSFJ: (a) Hysteresis on the First Level; (b) Hysteresis on the Roof Level.