Amplification of seismic demands in inter-storey-isolated buildings subjected to near fault pulse type ground motions

Highlight

• •Seismic demand amplifications in frame-ISI systems by pulse type motions are presented.

• Response-specific amplifications are presented considering frames and bearing typology.

• Effects of pulse period(s), mass ratio and bearing damping are demonstrated.

• FEMA guidelines for estimating bearing displacements is shown to be non-conservative.

• Modification factors are proposed for the FEMA formulae for bearing displacement.

Abstract

Inter Storey Isolation (ISI) shows promise for seismic vibration control in relatively tall buildings. With proper design, an ISI system isolates the upper storey block (USB) and emulates the effect of a nonconventional Tuned Mass Damper (TMD) on the Lower Storey Block (LSB) to reduce the seismic responses considerably. The vulnerability of conventional Base Isolation (BI) under pulse type motions was reported in the past. This study demonstrates the adverse effect of pulse type motions on the performance of ISI systems. An inclusive set of building configurations and bearing typology for the ISI system are employed and are subjected to two suites of unidirectional pulse and non-pulse type ground motions pertaining to varying hazard levels. Extensive nonlinear dynamic analysis of the frame-ISI systems are carried out. Representative statistics of the pertinent responses from each suite are contrasted to demarcate the effect of pulse type motions. The pulse type motions are shown to degenerate the efficiency of an ISI system to significantly amplify the seismic demands (and associated uncertainties) by allowing transmission of low frequency components (reminiscent of pulse(s)) to the response time histories. Response-specific amplification factors for the demands are assessed. The FEMA suggested guideline is shown to severely underestimates the bearing displacements, which is potentially dangerous. This deficiency is addressed by suggesting a modification factor in the FEMA formulae. The influence of the pulse period, mass ratio, damping in the bearing and hazard levels on the amplification of demands are discussed. However, this investigation considers a limited number of framing systems and isolation bearings subjected to a limited number of ground motions, which may be expanded in future study.

Introduction

Base Isolation (BI) is an alternative to the conventional ductility based seismic design methodology. Past research efforts established [[1], [2], [3]] the BI as an efficient, reliable and cost effective technology for seismic protection of structures (buildings and bridges generally) subjected to high seismic hazard. The BI is often preferred over the ductility based seismic design in critical structures (e.g. Nuclear power plant) to avoid inelastic deformations. However, few issues concerning the BI technology still deserve attention. The BI is quite effective in short period structures due to the available disparity between the vibration period of an isolation bearing (typically ranging from 2 to 4s) and a short period structure (typically ranges from 0.3 to 0.7s) [[1], [2], [3]]. The efficiency of BI was witnessed during 1994 Northridge earthquake in California, USA, 1995 Kobe earthquake in Japan and 1999 Chi-Chi earthquake in Taiwan. The isolated structures greatly survived these earthquakes. However, the margins of period disparity reduce in tall structures (e.g. tall buildings and long span bridges) due to their relatively long periods. Nevertheless, tall structures are naturally immune to far field ground motions as their long periods are away from the energetic higher frequency bands of the far field motions.

Evidences of pulse type ground motions were noted in recent past [4]. A near fault zone is typically extended within 20 km from the fault. The ground motions in this zone are significantly affected by the rupture mechanism and direction of slip relative to a site. The proximity of the fault rupture velocity to the velocity of shear wave propagation results constructive interference among the seismic waves in the near-fault region to result in spike(s) in the ground velocity time history. This phenomena is referred as the forward directivity effect, which causes high amplitude long period pulse(s) in the velocity time histories. The energy in the pulse type ground motions are significantly concentrated in a low frequency (or long period) band. Several earthquakes (1994 Northridge, 1966 Parkfield, 1971 San Fernando and 1995 Kobe earthquakes) in the past caused significant structural damage in their near fault regions due to significantly amplified inelastic displacement demands [5]. Tall structures are vulnerable to pulse type motions due to the proximity of the long period pulse(s) to their long periods.

Jangid and Kelly [5] and Shen et al. [6] demonstrated that the BI shows compromised efficiency under pulse type motions due to proximity of the pulse period (typically ranges from 2 to 8 s) to the period of isolation (typically ranges from 2 to 4s). An isolation bearing also experiences huge displacement, which may lead to pounding to adjacent structure/non-structural unit in absence of adequate seismic gaps. Pant and Wijeyewickrema [7] and Agarwal et al. [8] reported instances of pounding in base isolated structures and resulting damage(s). Bhandari et al. [9] compared the performances of a base isolated structure from the near and far field ground motions to show significant inelastic displacement demand under near fault motions. Gunes and Ulucan [10] demonstrated the effect of pulse characteristics on a 44 storied RC core wall building. A number of studies were also conducted by Mazza and his research group [[11], [12], [13], [14], [15]] to assess the performance of base isolated buildings under unidirectional near fault ground motions. Oncu-Davas and Alhan [16] assessed the probabilistic behaviour of isolated buildings under pulse type motions by employing a semi-active mechanism in conjunction with the BI system. Both the fling step and forward directivity effects in near fault motions were demonstrated on the isolated buildings by Bhagat et al. [17]. The influence of near fault ground motions on high rise building [[18], [19], [20], [21]], wind turbine [22,23], large transmission tower [24], bridge [25] and hydraulic dam [26] were also investigated in recent past.

Given the seismic vulnerability and infeasibility of BI in tall buildings; a potential alternative has been proposed by placing the isolation bearings at an intermediate storey of choice, so that, the portion above that storey (referred as the USB) is isolated, whereas, the portion below (referred as the LSB) experiences the effect of a Tuned Mass Damper (TMD). The mass in the USB contributes to the TMD mass. The bearings at the intermediate storey must be designed to optimize the effect of isolation as well as the action of TMD. Because of the comparable masses in the LSB and the USB, the mass ratio of such TMD becomes significantly higher (as much as 100%). This is unlike conventional TMDs, in which the mass ratios are typically restricted to few percent (2–5% mostly). Such TMD is thus referred as non-conventional TMD. The system is referred to as the Inter Storey Isolation (ISI) [27,28] system.

Although the ISI technology is relatively recent, analogous systems were described in alternative forms. An optimal energy-based design methodology for non-conventional TMD was proposed by Reggio and Angelis [27]. Loh et al. [28] employed a system identification methodology to a mid-storey isolated building under ambient and seismic vibration. Ryan and Earl [29] presented an ISI system with nonlinear device for controlling seismic vibrations in a multi-storied building. Ziyaeifar and Noguchi [30] proposed partial mass isolation system for tall buildings. Simplified analysis for mid-storey isolation system was presented by Wang et al. [31]. Experimental behavior of ISI system were investigated by this group [32]. Anajafi and Medina [33] provided a comparative assessment of a partial mass isolation system with a conventional TMD and a BI system. A “mega-configuration” was described by Feng and Mita [34]. Tan et al. [35] investigated an ISI system experimentally. Saha and Mishra [36] proposed a methodology for optimizing the performance of a non-conventional TMD implemented by an Adaptive Negative Stiffness Device (ANSD). All these studies addressed various aspects of ISI systems to facilitate their use in seismic vibration control of a class of buildings by effectively resolving the shortcomings in conventional BI systems.

The effect of near fault motions on long period structures is thus a serious concern similar to the buildings with BI. Equally important are the performances of ISI systems under pulse motions. However, this aspect is not attempted in required details, partially due to relatively recent development of the ISI systems. Thus, the need of detail investigation of ISI system subjected to pulse type motions is emphasized. The effects of unidirectional pulse type ground motions on ISI systems (including frame and bearing typology) are studied herein to summarize their implication in design. The typology includes six moment resisting frames and four isolation bearings. A suite of near-fault pulse and non-pulse type ground motions are considered as input in the analysis. The suites pertain to (or scaled to match) two distinct hazard levels, the Design Basis Earthquake (DBE) and the Maximum Considered Earthquake (MCE). Extensive numerical simulations based on nonlinear dynamic analysis of the frame-bearing systems are conducted. Whereas previous studies assume linear structural behavior (which may not be reasonable under pulse type motions), the present analysis takes into account the inelastic deformation of structures into account according to the FEMA 356 [37] guidelines. In anticipation of the amplified displacements in the bearings, the post-yield stiffening/hardening and the secondary hardening (neglected in previous studies) are adequately incorporated while modeling the force-deformation hysteresis of the isolation bearings. In presence of wide variability among the responses, relevant statistics (and factor of amplification) are extracted to quantify the pulse induced demand amplification by contrasting with their non-pulse counterpart. Modification factors are suggested for assessing the amplified bearing displacements in conjunction with the FEMA 450 [38] provision. A summary of database for the pulse induced amplification is provided to serve as a future reference for preliminary design of an ISI system.