SEISMIC VULNERABILITY OF HILL BUILDINGS

Abstract

The structural configurations of buildings in hilly regions are significantly different as compared to flat terrain counterparts, due to topographic constraints. The buildings in hilly regions generally follow the natural slope of the ground, resulting in foundations of a building stretching over different heights. These ‘hill buildings’ have short-columns on uphill side, and mass and stiffness irregularities in plan as well as along the height. Currently, none of the seismic design codes world-over provides any specific seismic design guidelines for hill buildings. In addition, slope-stability and topographic amplification effects are also crucial, which lead to a further increase in seismic vulnerability of hill buildings. However, the scope of the present thesis is limited to assessing the impact of different structural configurations on seismic vulnerability of hill buildings. To identify the prevalent structural configurations of buildings in hilly regions, extensive field surveys are conducted in two popular tourist destinations in the Indian Himalayas, viz. Mussoorie and Nainital, both located in seismic zone IV, as per current seismic zoning map of India. Based on the details collected from the field surveys, a building typology classification scheme (considering the various structural configurations of hill buildings) and a building stock inventory database are developed. The typology classification scheme takes into account a building’s structural configuration, building height, material of construction, load-bearing system, and roof type. It is identified that, reinforced-concrete (RC) buildings with regular (denoted as ‘SC A’), split-foundation (SF; denoted as ‘SC B’) and step-back (SB; denoted as ‘SC C’) structural configurations are most common and cover approximately 50% of the building stock in both the test beds. A majority of these RC hill buildings are ‘pre-code’ buildings whereas only few buildings with ‘moderate-code’ design level could be observed. These hill buildings are predominantly low- and mid-rise, and high-rise buildings are not observed in the selected test beds. These RC hill buildings have been studied in detail by conducting a numerical study. A statistical analysis of the plan details of RC buildings surveyed in Mussoorie is carried out to select a representative building plan for numerical study. Three different design levels representative of pre-code (designed for gravity loads alone), moderate-code (designed for gravity loads and earthquake forces, without conforming to strong-column weak-beam design) iv and high-code (designed for gravity loads and earthquake forces, and also conforming to strong-column weak-beam design) buildings are considered. Incremental dynamic analysis (IDA) is conducted on the considered buildings using near- and far-field ground-motion record suites, identified in FEMA P695. The effect of the seismic design level and the building height on collapse fragility is studied. The effects of near-field site and seismic zone on collapse fragility are also studied. It is observed that the period of vibration of SC B and SC C hill buildings is controlled by the number of storeys above the uppermost foundation level. In case of pre-code buildings, SC C buildings (having the highest torsional effects) have the least median collapse capacity whereas regular buildings have the highest collapse capacity. On the other hand, in case of moderateand high-code buildings, the regular buildings have the highest median collapse capacity whereas SC B buildings have the least median collapse capacity, though, all the buildings (SC A, SC B and SC C) were designed for identical base shear coefficients. The least median collapse capacity in case of SC B buildings designed for moderate- and high-code design level can be attributed to increased torsional effects in the inelastic range. It is observed that the average spectral acceleration Sa,avg (0.2T-3T, 5%), as a collapse intensity measure, captures the effects of higher modes of vibration as well as spectral shape of the ground-motion records. It results in collapse capacities of a building, nearly independent of the ground-motion record suites with a significantly reduced record-to-record variability, in case of moderate- and high-code buildings. This observation is not only valid in case of regular buildings, but also in case of torsionally irregular SC B and SC C hill buildings. On the other hand, in case of pre-code buildings, Sa (T, 5%) results in lower record-to-record variability than Sa,avg (0.2T-3T, 5%), due to very limited ductility capacity of such buildings. The damage patterns obtained from the numerical investigations in the present study suggest that the storey just above the uppermost foundation level is the most vulnerable location in the SC B and SC C hill buildings. This observation is found to be in good agreement with the observed damage in an SC B building after Sikkim earthquake of 2011. It is also observed that even after designing the buildings for a strong-column weak-beam factor of 1.40, the column hinging cannot be avoided, and there is scope for further enhancement of this factor. v Pre-code buildings result in unacceptably high probabilities of collapse (upto 95%) for Maximum Considered Earthquake (MCE) hazard. For all the investigated low- and mid-rise buildings, designed for moderate- and high-code design levels, the collapse probability has been found to be well within 10% for all the considered sites, conditioned on the occurrence of MCE. On the other hand, in case of high-rise SC B and SC C buildings, designed for high-code design level, the collapse probability for MCE hazard has been found to be significantly higher in a near-field site located in seismic zone IV and a far-field site located in seismic zone V. The floor acceleration demands for regular and irregular hill buildings have also been studied for performance-based design of non-structural components (NSCs). It is observed that peak floor acceleration (PFA) demands reduce with increase in period of vibration as well as inelasticity of the supporting structure. In case of SC B and SC C structural configurations, the PFA demands are controlled by a higher mode of vibration for building portion below the uppermost foundation level, whereas it is controlled by the fundamental mode in the building portion above the uppermost foundation level. The floor response spectrum (FRS) is observed to be better correlated to ground response spectrum (GRS) rather than peak ground acceleration (PGA) as used in current seismic design codes. Further, the spectral amplification factors along the height approximately follow the elastic mode shapes, for both elastic and inelastic supporting structures. In case of SC B and SC C structural configurations responding elastically, the torsional acceleration amplification in floor response is observed to be proportional to the torsional displacement amplification. Based on the observations and results obtained from numerical study, comprehensive spectral amplification functions are developed and validated for SC A, SC B and SC C structural configurations. The developed spectral amplification functions can be used with a code-based design response spectrum as well as a site-specific response spectrum to construct the floor spectrum. The developed spectral amplification functions are more comprehensive in predicting the floor acceleration demand than currently available models as these take into account the ground-motion characteristics, the dynamic characteristics of the supporting structure (both periods and mode shapes), the level of inelasticity expected in the supporting structure, and the period of vibration of the NSC.