Aseismic Design of Multi-storied RCC Buildings

A.R. Chandrasekaran + and D. S. Prakash Rao ++

ABSTRACT

Reinforced concrete multi-storied buildings are very complex to model as structural systems for analysis. Usually, they are modeled as two-dimensional or three-dimensional frame systems using finite beam elements. For evaluating stiffness matrix of beam elements, the properties required are – area of section, effective area in shear, and moment of inertia about the axis of bending and for three-dimensional analysis, in addition, torsional moment of inertia. Also, the modulus of elasticity and rigidity are to be estimated. However, no guidelines are available for the rational computation of sectional properties incorporating the effects of reinforcements in concrete members and the analysis is full of approximations. A case history of a RC structure, where the first author was involved in the study, is briefly cited in the paper.

The current version of the IS: 1893 - 2002 requires that practically all multistoried buildings be analyzed as three-dimensional systems. This is due to the fact that the buildings have generally irregularities in plan or elevation or in both. Further, seismic intensities have been upgraded in weaker zones as compared to the last version IS: 1893-1984. It has now indirectly become mandatory to analyze all multistoried buildings in the country for seismic forces. This paper appraises briefly the significant changes in the current version of the code compared to the previous version.

Some of the poor planning and construction practices of multistoried buildings in Peninsular India in particular, which lead to irregularities in plan and elevation of the buildings are also discussed in this paper. At present, there is too wide a variation in the modeling of buildings. This paper emphasises the need for guidelines in order to limit the range of assumptions to a narrow range. This is necessary to certify the analysis and design, or in case legal disputes arise later regarding the procedure adopted.

INTRODUCTION

Earthquakes are occasional forces on structures that may occur rarely during the lifetime of buildings. It is also likely that a structure may not be subjected to severe earthquake forces during its design lifetime. Reinforced Concrete Multi-Storied buildings (RCMS) are supposed to be of engineered construction in the sense that they might have been analysed and designed to meet the provisions of the relevant codes of practice and building bye-laws; the construction might have been supervised by trained persons. In such cases, even if earthquake forces have not been considered precisely, the structures would have adequate in-built strength and ductility to withstand some level of earthquake intensity.

During the past 50 years, only two events – Jabalpur earthquake of May 1997, and Kutch earthquake of January 2001 have caused significant damages to RCMS buildings with some spectacular damages during latter event. Compared to the large number of RCMS buildings in existence in the country, the number of damaged structures due to earthquakes is

______

+ Retired Professor, Department of Earthquake Engineering, IIT, Roorkee 247667

++ Professor of Civil Engineering, University College of Engg., Osmania University, Hyderabad 500007

indeed very small. It is usually assumed that damages are due to poor quality of construction, and builders are usually blamed for greed. However, it is possible that codes might have been inadequate in some cases.

The specifications of IS 1893 : 1984 for seismic resistant design have been inadequate for peninsular India. The Koyna earthquake of December 1967 as well as the Killari-Latur earthquake of September 1993 prove this point as far as seismic zoning is concerned. The current version, IS:1893-2002, made substantial upgradation of seismic zones in order to avoid such surprises in future.

Even IS 13920 – 1993 does not incorporate explicitly the specifications for strong-column weak-beam concepts, which are essential for structural survivability (Rai and Seth, 2002). The codes of practice do not include specifications for the design of beam-column joints; guidelines are available in literature, however (Rao, 1995).

The 2002 version of IS 1893 has more clearly defined the irregularities (vertical and horizontal) in the configuration of buildings than the earlier version. The current specifications would imply that most of the RCMS buildings in the country have irregular configurations, and have to be analysed as three-dimensional systems. There are a number of commercial software packages, which have the ability to analyse three-dimensional systems. However, the main problems are with modeling of the structure and member section properties. The Code provides no guidelines on these aspects leading to a wide variation in the results of the analyses. This paper highlights these aspects, and suggests that guidelines be made to rationalise the analysis and design processes.

COMPARISON OF 1984 AND 2002 VERSIONSOF IS 1893

The values of base shear coefficient vb, where base shear Vb = vb W, for a few cities in different seismic zones are compared for the current and previous versions of the code. It is assumed that I = 1.0 and β = 1.0; the normalized shape of “C” in Figure 3 (IS 1893 – 1984) is more or less similar to the shape of spectral acceleration of the 2002 version (it is the same in short period range). Two cases are considered for comparison, namely case I wherein K = 1.6 (1984 version) and R = 1.5 (current version), and case II wherein K = 1.3 (1984 version) and R = 3.0 (current version). The base shear coefficients are listed in Table 1 and 2 for cases I and II, respectively.

Table 1 Comparison of base shear coefficients for a few cities (case I)

CITY / IS : 1893 – 1984 / IS 1893 (Part 1) : 2002
Bhuj
Delhi
Kolkata, Mumbai
Chennai
Hyderabad, Bangalore / K = 1.6
Vb = 0.128 [Zone V ]
Vb = 0.080 [Zone IV]
Vb = 0.064 [Zone III]
Vb = 0.032 [Zone II ]
Vb = 0.016 [Zone I ] / R = 1.5
vb = 0.300 [Zone V ]
vb = 0.200 [Zone IV]
vb = 0.133 [Zone III]
vb = 0.133 [Zone III]
vb = 0.083 [Zone II ]

Table 2 Comparison of base shear coefficients for a few cities (case II)

CITY / IS : 1893 – 1984 / IS 1893 (Part 1) : 2002
Bhuj
Delhi
Kolkata, Mumbai
Chennai
Hyderabad, Bangalore / K = 1.3
Vb = 0.104 [Zone V ]
Vb = 0.065 [Zone IV]
Vb = 0.052 [Zone III]
Vb = 0.026 [Zone II ]
Vb = 0.013 [Zone I ] / R = 3.0
vb = 0.150 [Zone V ]
vb = 0.100 [Zone IV]
vb = 0.067 [Zone III]
vb = 0.067 [Zone III]
vb = 0.041 [Zone II ]

vb=W/Vb

The Tables indicate an across the board increase in the values of the 2002 version. However, it is more pronounced for the cities upgraded to more severe zone such as Chennai, Bangalore and Hyderabad. The increase in those cases varies from 300 to 500 percent depending on the type of structural systems.

However, there is no change in the values in the two versions in the case of Mumbai and Kolkata, , if reduction factor R = 5 and K = 1 (for ductile shear wall with Special Moment Resisting Frame or Special RC Moment-Resisting Frame).

Obviously, as per IS 1893 : 2002, no place in India is deemed to be free from earthquake hazards.

BEHAVIOUR OF RCMS BUILDINGS DURING KUTCH EARTHQUAKE

Most of the RCMS structures in Kutch, especially in Ahmedabad and Bhuj, comprised 4 – 11 stories with stone or concrete masonry walls of 150 - 250 mm thickness. The heavy partition walls must have resulted in large seismic forces, beyond the capacity of the structural systems. The partition walls suffered extensive diagonal cracking not only in the ground floor but also in the upper two to three floors.

Several structures with soft ground story (with less lateral strength than upper storeys) suffered extensive damage, and, in extreme cases, collapsed. Figure 1 indicates such a structure that suffered extensive damage in all the ground floor columns, but has not completely collapsed.

Lack of adequate number of well-designed ties at the beam-column junctions resulted in the failure of concrete. The concrete cracked extensively and, in some cases, was totally crushed. The premature failure of beam-column joints may have led to large lateral deformations and consequent collapse in several cases. In the case of structures that survived, extensive damage at the beam-column joints was discernible. Figure 2 indicates a typical column with inadequate ties (IS 13920:1993), poor compaction of concrete and possible cold joint in the column below the beam. Further, inadequate ties in the columns appear to have resulted in weak columns with poor confinement (Chien-Hung Lin and Feng-Sheng Lee 2001, and Hong Mei et al 2001) leading to large scale collapse of structures.

Most of the structures had water tanks, supported on columns, located over the staircases. These appendages were subjected to large lateral deformations, independent of the main structure, and developed extensive damage. In several cases, the damage to the columns supporting water tank was more severe than that to the main structure (Figure 3).

Even the staircase slabs suffered extensive damage due to improper detailing of the junction with the landing slab. Close proximity of structures was another peculiar feature, particularly in Bhuj (Figure 4). Adequate spaces around the structures might not have damaged some of these structures.

INADEQUACIES OF IS:1893 - 1984 AND THE KUTCH EARTHQUAKE

Under Clause 4.2.1 of IS : 1893 - 1984, the so called seismic coefficient method can be used for all the buildings less than 40.0 m in height in all the zones. Most of the RCMS buildings in the country fall in this category. Further, seismic coefficient method was permitted even for buildings up to 90.0 m in height in zones III, II and I. This meant a simple evaluation of equivalent static (mainly horizontal) forces to represent earthquake effects. The building is then analysed as several plane frame systems. This approach would not prevent the type of damages experienced during Kutch earthquake. The IS:1893 - 2002 seeks to prevent this over simplified approach by forcing a three dimensional analysis in most cases.

Clause 4.2.1 also permits “ modal analysis using response spectrum method”. This clause has been misused, intentionally or unintentionally, by assuming parameters to predict a flexible, long period system, thereby lowering the forces required to be taken in the analysis. The 2002 version has somewhat prevented this loophole by specifying a minimum value to be considered for base shear.

TYPICAL FEATURES OF RCMS BUILDINGS AND BUILDING BYE LAWS

The Building Bye Laws generally govern the lay out and planning of structures, especially in urban areas with dense population and closely located multi-storied structures. The Floor Space Index (FSI) and the maximum permissible coverage specified by various Gazetted Orders (GO) govern the plan dimensions and height of structures in general (Durga Prasad, 2000).

The FSI varies from 0.50 (general and special industries outside the municipal limits) to 1.75 (service industries outside the municipal limits); the usual value being 1.50 for residential and commercial buildings. Similarly, the maximum permissible land coverage varies from 30 percent (institutional buildings outside the municipal limits) to 80 percent (residential and commercial buildings within the municipal limits) depending upon the size of the plot.

Building Bye Laws specify minimum clearances around the structures (set backs) depending upon the plot area, road width and population zones; the minimum permissible clearances vary from 1.0 m to 6.0 m in the front, 1.0 m to 3.0 m at the rear, and 1.0 m to 5.0 in the sides. However, the regulations seldom mention unambiguously that the clearances pertain to any part of the structure. It appears that the builders take advantage of this lacuna to extend the structures at upper floors up to the property lines in all the directions in order to obtain the maximum possible saleable area (Figure 1 and 5). In most of the cases, the specified clearances are satisfied at the ground floor level, but not for the upper floors (Figure 5). The extension of beams beyond the column faces with the column shifted to the edge of cantilever is a common feature in most of the RCMS structures in the country; such columns are referred as ‘floating columns’ and are not conducive to resistance of structures to earthquakes (Figure 6).

It would appear that restricted FSI coupled with misinterpretation of the regulations pertaining to the minimum clearances around structures has led to the undesirable practices of ‘floating columns’. The close proximity of high-rise structures is also undesirable in regions prone to earthquakes.

FEATURES OF IS 1893 : 2002 (PART 1)

This fifth version of the code has given clearer definitions of irregularities in the vertical (elevation) and horizontal (plan) directions in the configuration of buildings. They are briefly as follows:

  • Plan irregularities causing torsion are re-entrant corners, diaphragm discontinuity, out of plane offsets and non-parallel systems.
  • Vertical irregularities are caused by variations in lateral stiffness, mass, vertical geometry, in-plane discontinuity in vertical elements resisting lateral force and discontinuity in capacity like weak story.

Most of the RCMS buildings in the country would have some irregularity or other in elevation or plan or both. This would imply that three-dimensional analysis has to be carried out for all such buildings. Since the practice now a days is to use commercial software packages, the analysis would not pose any problem but results will vary because of modeling and assumption of properties.

The current version also proposes new criteria for the dynamic parameters such as the number of modes and combination of modes, which are based on theoretical vibration mechanics. All these requirements are taken into account in the software packages based on the pioneering work done at the University of California at Berkeley, USA.

The new version has an important stipulation regarding minimum base shear, which is computed using empirical formulae. The likely effect of this stipulation is that it will also be the maximum value of base shear as designers may choose convenient modeling parameters. This is dealt with in a subsequent section.

CASE HISTORY OF A RCC TURBO-GENERATOR FOUNDATION AND RELEVANCE TO MODELING OF RCMS BUILDINGS

A TG Foundation was apparently designed to satisfy IS:2974-1992 (unpublished report). This code has more advanced features than IS 1893 in the sense that it proposes mathematical modeling of the structural system. However, after the equipment was commissioned, it was observed that one of the cross beams supporting a bearing developed resonance. The first author was involved in the System Identification studies based on amplitude-frequency measurements when the system was running at various loads as well as by using a mechanical exciter. The studies indicated that IS 2974 : 1992 provisions were inadequate to predict the behaviour of the actual structure. The cross beam was not resting directly on the columns but had an off-set and the bearing load was also eccentric to the beam. The cross beams are quite deep and modeling of the system by beam elements as proposed by this code was inadequate and hence it could not predict that the system could have one of its natural frequencies close to the operating frequency.

For the study, the TG foundation was modeled with three-dimensional solid finite elements. Gross section of concrete was assumed as effective, and the influence of steel reinforcement was ignored. The modulus of elasticity was the parameter that was varied in order to find a match of the theoretical resonant frequency with that from the experiments on the prototype.

Incidentally, the mode shape of the cross beam matched the observed displacements at that resonant frequency. It was noted that when the matching took place, the modulus of elasticity was not the actual static or dynamic modulus of elasticity but was much less than the static modulus.

RCMS buildings are far more complicated than a TG foundation and it is most likely that any method of analysis with assumed modeling and properties of section will predict a dynamic behaviour different from experimentally measured behaviour after the structure is built.

MODELING OF RCMS BUILDINGS

Uncertainties in the Forces Specified in IS 1893 : 2002

The minimum (probably, it will also be maximum in practice) value of base shear depends up on the zone and importance factors, spectral acceleration based on the empirical value of period, and the response reduction factor R. The intensity of motion is assumed constant in a seismic zone whereas during an actual earthquake, it would never be so. The empirical value of period is not based on the data of the RCMS buildings in India. It is likely that designers may use the highest value of R to get lowest force for design. The value of 5 has never been achieved in any experimental testing on RC structures. Further, even IS : 13920–1993 on “Ductile Detailing ..” (clause 1.1.1) exempts structures in Zone II from seismic analysis. Therefore, the base shear is just a notional figure to give some degree of safety during an earthquake.

Uncertainties in Properties – not covered in IS 1893 : 2002

Considering a RC frame, the columns may be subject to bi-axial bending in addition to axial load and shear force. The floor beams integral with slabs will have T-beam action at mid-span and as a rectangular section at supports. The amount of reinforcement may vary along the length of these elements. At present, there is no rational method available for taking reinforcement into account in the evaluation of properties of section (area, shear area, bending moment of inertia and torsional moment of inertia), which are required in the analysis of beam elements. It is, therefore, assumed that only the gross section of the concrete is effective and the reinforcement is ignored in evaluating the properties. The modulus of elasticity and rigidity are also required in the analysis. Some codes suggest the use of static modulus and others dynamic modulus.