City Research Online

Abstract

During earthquake induced strong ground motion (GM) adjacent buildings with inadequate clearance will interact/collide resulting in the development of pounding forces at locations of contact. Typically, forces due to earthquake induced seismic pounding (EISP), and their consequences, are not accounted for in the seismic design of buildings as contemporary codes of practice for earthquake resistance specify minimum clearance among neighbouring structures regarded as adequate to minimise EISP occurrence/consequences likelihood at least for the nominal design earthquake level. However, field observations in congested cities in the aftermath of several recent major seismic events suggest that considerable seismic loss is due to EISP as code-prescribed clearances are not implemented in practice.

These observations triggered significant research efforts since the late 1980s to develop efficient finite element (FE) modelling schemes capturing EISP, to study the influence of EISP in seismic demands of colliding structures, and to propose methods of mitigating EISP consequences. Nevertheless, to date, most relevant computational-based research works adopted simplified structural models used as proxies of the colliding buildings, such as planar multi degree-of-freedom (MDOF) frames arranged in series or single degree-of-freedom (SDOF) pounding oscillators, to study EISP via nonlinear response history analyses (NRHA). Further, uncertainty quantification due to record-to-record GM variability to inelastic seismic demands under EISP has not been addressed within modern probabilistic performance-based earthquake engineering (PBEE) context.

To this end, this thesis aims, first, to assess the influence of EISP to inelastic demands at structural member level in a case-study real-life building block and, second, to quantify EISP influence to fragility curves of commonly adopted simplified structural models (i.e., inelastic SDOF oscillators and inelastic planar MDOF frame structures) as a measure of seismic vulnerability of colliding structures in a statistical framework accounting for record-to-record variability. The thesis focuses on reinforced concrete (rc) code-compliant building structures and treats exclusively slab-to-slab interaction/pounding assuming that no significant local failure occurs at locations of collision.

The first aim is addressed by developing detailed three-dimensional lumped-plasticity FE models of three adjacent irregular in-plan rc structures with coupled frame-wall lateral load resisting systems in an L-shaped arrangement and with unequal number of floors. Series of NRHA is conducted for a pair of spectrum-compatible GMS with increasing intensity (i.e., incremental dynamic analysis-IDA) acting along two horizontal perpendicular axes for FE models with and without EISP. Variations of inelastic demands across all building floors for different types of structural members (i.e., beams, columns, and walls) are reported due to EISP for different GM intensities. Considerable floor-wise spread of differences of inelastic demands due to EISP is found in all 3 structures and types of members. This novel finding suggests that EISP influence to local member inelastic demands may not be accurately quantified through simplified planar FE MDOF models which cannot capture the response of complex building blocks colliding bi-directionally and accounting for torsional response. Therefore, it is recommended that detailed spatial FE models are adopted for seismic vulnerability assessment of existing case-specific structures subject to EISP in several directions.

The second aim is pursued by putting forth a performance based seismic assessment (PBSA) approach which can readily account for record-to-record variability, following standard PBEE steps, through application of IDA for a suite of judicially selected GMs to simplified inelastic FE models capturing EISP. In doing so, a novel intensity measure (IM), namely the geometric mean of the spectral acceleration at the fundamental natural period of the pounding/interacting structures, avgSa, is proposed. It is proved numerically that avgSa is much more efficient than peak ground acceleration (PGA) which is exclusively used as the IM in all EISP studies found in the literature. This is established by noting that avgSa reduces significantly the spread of IDA curves compared to PGA, gauged via the standard deviation of log-normal distributions fitted to the IDA curves data at different limit states, for several different pairs of colliding inelastic SDOF oscillators used as proxies to 5 different 8-storey and 12-storey benchmark rc multi-storey frame structures design to the current Eurocode 2 and 8 subject to a suite of 72 GMs. Moreover, novel probabilistic models in terms of fragility curves of adjacent rc structures are presented and discussed derived for both the above inelastic SDOF oscillators and for the detailed MDOF lumped-plasticity models of the planar multi-storey frame structures. Sensitivity analyses is undertaken to quantify the influence of various pounding model parameters to inelastic demand statistics (i.e., shape of fragility curves) indicating that stiffness and damping properties of the pounding model is not as influential as clearance between structures. Lastly, mean and standard deviation of IDA curves data obtained by interacting MDOF models and their equivalent (i.e., derived through pushover analysis) inelastic SDOF oscillators are compared. It is found that interacting SDOF proxies capture accurately record-to-record variability expressed through the standard deviation of fitted log-normal distributions to IDA curves but tend to underestimate peak inelastic demands in the mean sense compared to the MDOF models. Thus, it is again concluded that caution need to be exercised in adopting simplified models for capturing EISP.

Overall, the PBSA tools developed in this thesis and the numerical data furnished shed new light on the influence of EISP to different levels of sophistication in structural modelling of building structure and to the uncertainty in inelastic seismic demands due to record-to-record variability. These tools together with foreseeable extensions pave the way for seismic risk analyses in congested urban environments accounting for EISP phenomena to improve the accuracy of seismic loss predictions.

Publication Type:	Thesis (Doctoral)
Subjects:	T Technology > TA Engineering (General). Civil engineering (General) T Technology > TH Building construction
Departments:	Doctoral Theses School of Science & Technology > School of Science & Technology Doctoral Theses School of Science & Technology School of Science & Technology > Engineering

Export

Downloads