You are on page 1of 6
Analysis for seismic retrofitting of buildings | Sudhir K. Jain and T. Srikant | ‘The issue of seismic retrofitting of multistoreyed buildings hhas received considerable attention after the Bhuj earthquake (of 2001. To classify vulnerable buildings and to carry out a rational retrofitting project, proper analysis is needed, which takes into account the post-elastic response ofthe building. A static non-linear analysis procedure (push-over analysis) is being increasingly used for this purpose inthe profession. The paper describes the context of the push-over analysis and illustrates its utility with the results of analysis on a hypothetical example building. ‘The numerous collapses of multistoreyed buildings in the Bhuj earthquake of 2001 have clearly shown the seismic vulnerability of a large building stock in our country. Hence, a large number of buildings in our country need seismic retrofitting wherein the building is provided additional strength, stiffness and ductility to ensure acceptable performance in a future earthquake. A building may be seismically deficient for a variety of reasons. For instance, at the time of building construction, there may not have been adequate concern for seismic safety, tr the engineer concerned with design or construction may have lacked aclequate expertise to execute an earthquake resistant building. There could also be upgradation of the codal requirements in view of better understanding of the seismic hazard or of the building performance. Further, the owner may decide to seek better performance from the building in case of a future earthquake. And finally, there may have been deterioration in the materials with time or changes in the structure or its occupancy, making the building, vulnerable, Considering the worldwide concern on this issue, a number of excellent publications are now available, giving in-depth coverage of the analytical tools for design of proper seismic retrofitting schemes (for example, ATC40-1996a, ‘ATC40-1996b, FEMA273-1997, FEMA274-1997, FEMA306- 1999, FEMA307-1999, FEMA308-1999}”. This paper discusses the concept of “Push-Over Analysis” that is becoming a ‘popuilar tool in the profession for (i) design of new buildings, (i) seismic evaluation of existing buildings, and (ii) developing appropriate strategy for seismic retrofitting of buildings. It is shown how this analytical technique can be ‘useful in deciding seismic retrofitting strategy and technique Expected seismic performance Traditionally, seismic design objectives are stated in the following terms, ‘+ The structure should be able to endure minor (and frequent) shaking intensity without sustaining any damage; thus leaving the structure serviceable after the event. ‘+ The structure should withstand moderate levels of ‘earthquake ground motion without structural damage, but possibly with some non-structural (but not life- threatening) damage. © The structure should sustain severe earthquake ground motion without collapse of the structural framework, but possibly with some structural as well as non-structural damage. This limit state may correspond to an earthquake intensity equal to the strongest either experienced or forecast at the site. 1c3 COMPILATION ‘aioe tealy— same "0 cme sal) Te 1 ¥ ogiS ams 00s) 00m) 00m ous MVE 0082 0064 0.061 0046 0.053 0.058 m vom 0132 aaa a0 ear Tea Ger nd Riker, 1956 01M 026024 023030 a8 os 0st 04s osz_ am om Marpy and Oren 137) 6 Mpa snd O'Bxen 1977 2. Newnan, 1954 2 Tina ond Bry, 1975; 4 Time and Bay, 1977 id by Mary and Oem, 177) However, such criteria are no longer considered specific ‘enough. With advancements in the understanding of ‘earthquake performance of buildings, more specific answers are being sought by the owners about the expected performance of buildings in the event ofa strong earthquake, For instance, an owner may ask: what kind of damage to ‘expect in my building if it is shaken by a ground motion ‘which has a 10 percent probability of occurrence in the next 50 years? To answer such a question, several issues must be addressed, for example: (seismic hazard assessment, that is, what is the ‘ground motion that has a 10 percent probability of occurrence in 50 years, (ii) what are the force and ductility demands that such a ground motion will impose on the building, (ii) how much are the strength and ductility available on the building, and (io) as the building undergoes inelastic response what damage will occur in it? Fig 1 Typical acceleration response spectrum (5 percent damping) used in design Ground motion versus seismic design force The Indian seismic code (IS:1893-1984) divides the country into five seismic zones I to V". These zones imply that the ‘expected modified Mercalli shaking intensity (MMI) could be V (or less), VI, VIL, VII, and IX (and above) in these zones, respectively. Table 1 gives peak ground acceleration in the horizontal direction corresponding to different levels of MM intensity. Since these values are empirical, there tends to be a considerable variation in the ground motion prediction. Thus, for seismic zone IV (MM intensity VIII), peak ground acceleration (PGA) of about 0.15g to 0:30g ean be expected. Fig 1 shows atypical elastic acceleration response spectrum for 5 percent damping, Note that short period structures (T in the range of 0 sec to about 05 sec) experience a maximum horizontal acceleration of about 25 times PGA ifthey remain clastic. Hence, linear elastic structure of fundamental period around 03 s located in seismic zone IV may experience a maximum acceleration in the range of 0.38 to 0.75. However, 1S:1893 requires that an ordinary building with fundamental period of 0.3 s in zone IV be designed for a seismic coefficient as low as 0.05g. The huge difference in the ground motion and the design level is accounted for by the following (@ Some amount of damage is acceptable in the structure. If the structure is ductile it will be able to absorb considerable amount of energy without collapse, and hence, the ductility is effective in reducing the design force requirement on the building, (ii) Considerable amount of overstrength is introduced in the building during design and construction, thats, the real strength of the building is much higher than the design strength due to numerous factors of safety and other factors, and (Gi) The redundant systems allow for redistribution of forces after yielding starts ‘Thus, it is clear that the earthquake resistant building is ‘expected to perform satisfactorily even when subjected to earthquake loads much higher than the code-specified design forces. It also means that in the event of design level earthquake shaking, the building may undergo some damage and hence its inelastic response is important. Push over analysis building may be deficient in terms of lateral force capacity, ductility, lateral stiffness or structural configuration. For instance, one side ofthe building may be much stiffer than the other side, forcing the building to go into torsion modes of vibration during ground shaking: such a deficiency too needs to be addressed during relroiting, Let us consider a beam fixed at both ends, Fig 2, carrying uniformly distributed load. The figure also shows the bending moment diagram caused by this load, Let us say that the beam is already built and the moment capacity of the beamis as shown. Notice that the beam is deficient in hogging, moment capacity at the two ends, but has more than the EARTHQUAKE ENGINEERING ae red ad eran er tending nonent Capac fr bending moment Fig 2 Example fixed-flxed beam and bending moment demand and capacity required capacity at mid-span. The question is: does the beam. require retrofitting for the given loading? This question can be best addressed by considering the failure mechanism of the beam. It is seen that the collapse load capacity of the beam is more than the collapse load corresponding to the load pattern, and hence, in this ease no retrofitting is required. Thus, we have resorted to a plastic analysis (non-linear analysis) to decide on the need for retrofitting, Let us extend the above argument to seismic loading on a building and ask: how does one decide if an already built building is adequate for seismic loads — particularly when it is expected to undergo damage in the event of a strong. shaking? Since itis a dynamic problem and the building will undergo non-linear response, it seems desirable to carry out ‘a non-linear dynamic analysis. However, such an analysis is too sophisticated and requires numerous parameters that are difficult to specify accurately (for example, ground motion characteristics and properties of the structure). Hence, the utility ofa highly accurate analysis based on very approximate parameters will be rather limited in professional practice. Therefore, the profession is now moving towards non-linear static analysis (often termed as push-over analysis). In this method, a static horizontal force profile (usually proportional tothe design force profile specified in the codes, ‘or corresponding to the first mode of vibration) is applied on the structure. The force profile is then incremented in small steps and the structure is analysed at each step. As the loads are increased, the building undergoes yielding at a few locations: every time such yielding takes place, the structural properties are modified appropriately to reflect the yielding. ‘The analysis is continued till the structure collapses, or the building reaches a certain level of lateral displacement. Typical example of push-over analysis results ‘As an example, let us consider a four-storeyed flat slab butlding, Fig, located in Mumbai. Let us further assume that the building was designed for horizontal wind loads (which works out to about 28 percent ofthe building weight) but not for earthquake loads. The building is deficient in terms of lateral force capacity. Also, the flat slab system is likely to show brittle behavior that is most undesirable. The objective is to evaluate different options for seismic retrofitting of the building ‘A number of computer programs dealing with structural analysis now include push-over analysis (for example, SAP2000, ETABS)*”. Herein, computer program SNAP-2DX has been used for the push-over analysis”. Moment- curvature relationship of each element is idealised as elasto- plastic, that is, linear upto yield point and constant moment thereafter. The building is loaded with full dead load plus 50 percent design live load. Thereafter, push-over analysis of ‘one interior frame ofthe building is carried out. Dovich and Wight have proposed effective slab widths to be considered for lateral analysis". These are different for strength and stiffness computations: [2 for strength and Iy8 for stiffness, where fis the centre-to-centre slab width transverse to the direction of moments. In the present case, an average value as Icy compnxtioN of ,/25 is used for effective width Cf slab element. The cross-sectional dimensions of the column head clements are taken by averaging their cross-section. Lateral loads are applied to different floors in an inverted triangular distribution, that i, the load at a floor is proportional to the height of the floor from the foundation. This lateral load profile isincreased in small increments and the structural response is observed. in terms of base shear versus roof displacement. Itis assumed for the purpose of this analysis that the building does not undergo punching shear failure during the earthquake ground motion, Fig 4 (a) shows base shear-

You might also like