Dookie pres By bee 4 3/99 Report 3 8 ——— Research Project on . Study of Seismic Design Codes for Highway Bridges A Proposed Draft For IRC:6 Provisions On Seismic Design Of Bridges :: Code and Commentary Sponsored by Ministry of Surface Transport (Roads Wing) Government of India New Delhi by C. V. R. Murty and Sudhir K. Jain Department of Civil Engineering Indian Institute of Technology Kanpur Kanpur 208016 March 1997Damping : The effect of internal friction, imperfect elasticity of materia, slipping, sliding, etc., in reducing the amplitude of vibration and is expressed as a percentage of critical damping, Design Seismic Force : It is the seismic force prescribed by this standard for each bridge component, that shall be used in its design. It is obtained as the maximum elastic seismic force divided by the appropriate response reduction factor specified in this standard for each component. Ductility : Ductility of a structure, or its members, is the capacity to undergo large inelastic deformations without significant loss of strength or stiftness. Ductile Detailing : It is the preferred choice of location and amount of reinforcement in reinforced concrete structures to provide for adequate ductility in them. In steel structures, it is the design of members and their connections to make them adequately ductilecentre of gravity ofits base Seismic mass martix of the bridge structure Modal participation factor of mode k of vibration Pressure due to fluid on submerged superstructures Response Reduction FactorDraft for IRC:6 Provisions on Seismic Design of Bridges :: Code & Commentary page Sof 44 ",72,73Force resultants due to full design seismic force along two horizontal directions s Se and along the vertical direction, respectively Soil Profile Factor Seat length of the superstructure on the substructure (or of the suspended portion of the superstructure on the restrained portion) Natural Period of Vibration Natural Period of Vibration of mode k Vertical force at support due to seismic force Lateral Shear Force Maximum elastic force resultant at a cross-section of a bridge component ic Design seismic force resultant in any component of the bridge due to all modes considered Seismic weight, which includes full dead load and part live load as discussed in 323, Widths of seating at bearing supports at expansion ends of girders. Weight of water in a hypothetical enveloping cylinder around a substructure Height of water surface from level of deepest scour (in m) Seismic zone factor Horizontal seismic coefficient Horizontal seismic coefficient Basic horizontal coefficient Ratio of natural frequencies of modes / and j {ox} Mode shape vector of the bridge in mode k of vibration Net response due to all modes considered Response in mode & of vibration. Coefficient used in combining modal quantities of modes i and j by CQC Method Natural frequency of mode & of vibration Modal damping ratio The existing version of the IRC code, ie, IRC:6-1966, considers variation in seismic risk in different parts of the country through “horizontal seismic coefficient at.” On the other hand, the !S code, i.¢., |5:1893-1964, uses the “basic horizontal coefficient: aig” for the same parameter. Hence, in the draft provisions, a new parameter “seismic zone factor” has been defined to distinguish from the earlier parameters and has been assigned the symbol “Z” Symbol “A” has been assigned to represent the elastic acceleration spectrum arrived at after considering the relevant factors such ae selsmic zone factor Z, importance factor |, bridge flexibility factor C, and soil profile factor S. This spectrum value A is to be finally used for design of a bridge independent of ‘the method of analysis to be used (i, static or dynamic). General Principles Lt Scope This standard is applicable for the seismic design of new bridges and the seismic evaluation of existing bridges. Bridges and portions thereof shall be designed and constructed, to resist the effects of design seismic force specified in this standard as a oaDraft for IRC:6 Provisions on Seismic Design of Bridges :: Code & Commentary page 6 of 44 The designers may use this draft code both for design of new bridges and for seismic. evaluation of existing bridges in the process of their seismic upgradation. The designer may choose to design bridges for seiomic forces larger than those specified in this code and but not less. 1.2. The intention of this standard is to ensure that bridges possess at least a minimum strength to withstand earthquakes. The intention is not to prevent damage to them due to the most severe shaking that they may be subjected to during their lifetime, Actual forces that appear on portions of bridges during earthquakes may be greater than the design seismic forces specified in this standard, However, ductility arising from material behaviour and detailing, and overstrength arising from the additional reserve strength in them over and above the design force, are relied upon to account for this difference in actual and design lateral loads The earthquake codes provide design forces which are substantially lover than what a structure is expected to actually experience during strong earthquake shaking. Hence, it is important that the structure be made ductile and that it be redundant to allow for alternate load transfer paths. Ductile design and detailing enables a designer to use a lower design force (ie, a higher value of response reduction factor R) than for an ordinarily-detailed structure. 1.3 The reinforced and prestressed concrete components shall be underreinforced so as to cause a tensile failure. Further, they should be suitably designed to ensure that premature failure due to shear or bond does not occur. Ductility demand under seismic shaking is usually not a major concern in bridge superstructures, However, the seismic response of bridges is critically dependant on the ductile characteristics of the substructures, foundations and connections. Provisions for appropriate ductile detailing of reinforced concrete members given in IS:13920-1993 shall be applicable to substructures, foundations and connections. Provisions for ductile design and detailing for reinforced concrete structures are provided in 1S:13820-1993. However, provisions for ductile detailing of prestressed concrete, steel and prefabricated structures are not yet available in the form of Indian Standards. IF euch structures are to be designed for high seismic zones of the country, it is expected that the designer will ensure suitable ductility following the practices of countries with advanced seismic provisions, 6g. USA, New Zealand and Japan. 1.4. Masonry and plain concrete arch bridges with spans more than 10 m shall not be built in the severe seismic zones IV and V. Designers are prohibited to consider masonry and plain concrete arch bridges of spans more than 10 m as structural systems for bridges in high seismic zones, since these systems are known to have a very poor behaviour under strong ground shaking 1.5 Ground Motion The characteristics (intensity, duration, e/c.,) of seismic ground vibrations expected at any location depends upon the magnitude of earthquake, the depth of focus, distance from the epicenter, characteristics of the path through which the seismic waves travel, andraft for IRC:6 Provisions on Seismic Design of Bridges :: Code & Comment ze T of 44 the soil strata on which the structure stands. The random earthquake ground motions, which cause the structures to vibrate, can be resolved in any three mutually perpendicular directions, Situations arise where earthquake-generated vertical inertia forces need to be specifically considered in design. These situations include bridges with large spans, those in which stability is a criterion for design, design of vertical hold-down devices at supports or for overall stability analysis of bridges. Reduction in gravity force due to vertical component of ground motions can be particularly detrimental in cases of prestressed horizontal girders and of cantilevered components. Hence, special attention should be paid to the effect of vertical component of the ground motion on them. ‘The upward seismic forces produce stresses that are usually not accounted for in the gravity design of horizontal prestressed girdere and cantilevered componente. The 1994 Northridge earthquake in USA has clearly shown the vulnerability of horizontal prestressed girders subjected to vertical ground motions. To check the girder for vertical component ground motions, it may be sufficient to consider the girders, except in case of large span bridges, as rigid for vertical vibrations and subjected to zero-period vertical accelerations but with no response reduction factor R (ie, the seismic coefficient as 0.67215, ote eipeel since the vertical accelerations to be taken for ‘the purposes of design are O.67 times that of the horizontal accelerations specified in this code). ln the seiemic design of bridges, vertical ground motions are particularly important. Vertical seiomic forces may cause jumping of girders, and additional stress resultants and displacements, particularly in long span bridges. For this reason, this draft recommends that wherever applicable, vertical seismic forces shall be considered. Also, in the overall stability check of bridges, in the stability of superetructures or portions thereof that are not monolithic with the substructure, and in the design of vertical hold-down devices at supports, vertical seismic forces shall be considered. 1.6 The response of a structure to earthquake shaking is a function of the nature of foundation soil; materials, form, size and mode of construction, and characteristics and duration of ground motion. This standard specifies design forces for structures standing on soils or rocks which do not settle or stide due to loss of strength during shaking This clause warns designers that the provisions contained in this draft code do not provide safeguard against situations where soll underlying the structure may undergo instability due to large settlements, sliding or liquefaction. 1.7 Assumptions The following assumptions are made in the earthquake-resistant design of bridges: (a) Earthquake causes impulsive ground motions, which are complex and irregular character, changing in period and amplitude, and each lasting for a small duration. Therefore, resonance of the type as visualized under steady-state sinusoidal excitations, will not occur as it would need time to build up such amplitudes. (b) Earthquake is not likely to occur simultaneously with wind or maximum flood or maximum sea waves. (©) The value of elastic modulus of materials, wherever required, may be taken as for static analysis unless a more definite value is available for use in such condition.Draft for IRC:6 Provisions on Seismic Design of Bridges :: Code & Commentary page 8 of 44 The clastic modulus of concrete is difficult to specify. The value varies with the stress level, loading conditions (static or dynamic), material strength, age of material, etc. Hence, there tends to be a very large variation in the value of elastic modulus specified by different design codes even for the same grade of concrete under static conditions. For instance, ACISI6(1989) recommends modulus of elasticity E as 41700) F; (MPa), while in 16456-1978, E is calculated as 5700/fix (MPa). Here, ff ie the 26-day cylinder strength and fy is the 28-day cube strength; further, fZ = 0.8 fz. Thus, the value of E given by the IS code io about 1.4 times that given by the ACI code for the same grade of concrete. Further, the actual strength of concrete is more than the 28-day strength; it shows an increase with time. There are further difficulties in choosing the value of the modulus of elasticity for concrete for seismic analysis. The value of E given in codes, such as ACIBIB and |S:456, is often the secant modulus; its value ie prescribed with a view to obtain a conservative estimates of deflections, ie. lower stiffness. On the other hand, the dynamic modulus of elasticity of concrete refers to almost pure elastic effects and Is equal to the initial tangent modulus, which is appreciably higher than the secant modulus. When a structure is new and ie subjected to low amplitude of ground motion, the dynamic modulus of elasticity to be used in the analysis has two opposite implications on seismic design. For calculation of the design seismic force, it is unconservative to have low stiffness given by low value of modulus of elasticity; this leads to a high natural period and lower design seismic coefficient. However, for the deflection calculations, it ie unconservative to make a high estimate of stiffness. Hence, there are no easy answers to the question of what value of modulus of elasticity to use for seismic analysis. Considering the enormous variations, this clause allows the designer to use elastic modulus as for a static condition. 2.0 Design Criteria In the current IRC and IS codes, the design seismic forces for bridges are directly specified; thie was often misunderstood as the maximum expected seismic force on the bridge under design seismic shaking. In line with the worldwide practice in thie regard, the draft: code now distinguishes the actual forces appearing on each bridge component: during design earthquake shaking if the entire bridge structure were to behave linear elastically, from the design seismic force for that component. The draft code makes it clear to the designer that the design seismic forces on superstructure, substructure and foundations are only a fraction of the maximum elastic forces that would appear on the bridge. Onlly in connections, the design seismic forces may be equal to (or more than) the maximum elastic forces that would be transmitted through them. However, if capacity design provisions discussed under 9. become applicable, the connection design forces__Draft for IRC:6 Provisions on Seismic Design of Bridges :: Code & Commentary page 9 of #4 may also be less than the maximum elastic forces. This is in stark contrast with the design forces for any other loading type. For instance, in case of design for wind effects, the maximum forces that appear on the structure are designed for: no reductions are employed. The draft code achieves this by the following step-wise procedure: (@) Obtain the horizontal elastic acceleration coefficient due to design earthquake, which is same for all components; (b) Obtain the seiemic weight of each components (©) Obtain the seismic inertia forces generated in each component by multiplying quantities in (a) and (b) above; (4) Apply these inertia forces generated in each of the components (from (c) above) at the centre of mass of the corresponding component, and conduct a linear elastic analysis of the entire bridge structure to obtain the stress resultants at each cross-section of interest; (@) Obtain the design stress resultants in any component by dividing the elastic stress resultants obtained in (d) above by the response reduction factor prescribed for that component. Thus, first the maximum elastic seismic forces are estimated and then these are divided with the response reduction factors to obtain the design eelemic forces. 2.1 Seismic Zone Map For the purpose of determining design seismic forces, the country is classified into four seismic zones as shown in Figure 1 The seismic zone map is under revision by the concerned Map Sub-Committee of the Sectional Committee on Earthquake Engineering (CED:59) of the Bureau of Indian Standards, New Delhi. However, it is already agreed upon that the new zoning map of India shall have only four seismic zones. Ae an interim measure till ‘the new zoning map becomes available, for the purpose of determining seismic forces as per this draft code, the current seismic zone map as given in IS:193- 1984 (shown in Figure 1) is used with seismic zone | merged upwards with seismic zone Il. The current IRC:6-I966 uses the same seismic zone map as in IGAB93- 1984, 2.2 Methods of Calculating Design Seismic Force The seismic forces for bridges may be estimated by either one of the two methods, namely (a) the Seismic Coefficient Method described in 3.0, or (b) the Response Spectrum Method described in 4.0. For all bridges in seismic zones IV and V, and also for irregular bridges as defined in 2.2.1 in seismic zones HI, the Response Spectrum Method shall be adopted. Linear static analysis of the bridge shall be performed for the applied inertia forces to obtain the force resuitants (e.g., bending moment, shear force and axial force) at the different locations in the bridge. For this purpose, the analytical model of the bridge must appropriately model the stiffnesses of superstructure, bearings, piers or columns (ie., substructure), foundations and bridge ends Special seismic analysis and design studies shall be performed for regular bridges with span more than /00 m and for all irregular bridges in seismic zones IV and V.Draft for IRC:6 Provisions on Seismic Design of Bridges :: Code & Commentary page 10 of #4 Both IRC:6-1966 and |S:1693-1984, currently follow a very simplistic design force calculation procedure which does not qualify under either the Seismic Coeffcient Method or the Response Spectrum Method ae these methods are generally understood for buildings in the context of IS:s1B95-1984. In these codes, the seismic design force computation does not include consideration of flexibility of the bridge. This impiies that all bridges in a seismic zone, irrespective of their span and structural system, have the same accelration coefficient in ‘the design; this is not considered appropriate. This draft code includes the eect of bridge flexibility in its design force computation. Further, it permite the use of both the Seismic Coefficient Method (ie, equivalent static method) and the Reeponse Spectrum Method (i.e, dynamic analysis method). However, it is felt that the latter method is superior in arriving at the distribution of forces in the bridge structure. The Seismic Coefficient Method described in the commentary under 3. assumes that (a) the fundamental mode of vibration has the most dominant contribution to seismic force, and (b) mass and stiffness are evenly distributed in the bridge resulting in a regular mode shape. However, in long span bridges, higher modes may be important. And, in irregular bridges, the mode shape may not be regular. Hence, this clause requires multi-mode analysis, namely Response Spectrum Method, for such bridges. The draft code also prescribes that all bridges in the high seismic zones (|e, V and V) shall be analysed as per the multi-mode (dynamic) method. This is again motivated by the fact that the a better distribution of forces is achieved by thie method. ln both the methods, the accurate modelling of the bridge structure is essential, because unlike in the case of buildings where the empirical natural period lo based on actual measurements of buildings, no euch benchmark is available for bridge structures. The large scatter in the bridge geometry, structural system, and the loading conditions makes the determination of an empirical benchmark for natural period of bridges very difficult The draft code recognises that bridges (even if they are regular) of spans around 100 m or more and all irregular bridges in high seismic zones IV and V, require a more detailed engineering with the help of the state-of-the-art analysis and design methods. 2.2.1 Regular and Irregular Bridge 2.2.1 Regular Bridge A regular bridge has no abrupt or unusual changes in mass, stiffhess or geometry along its span and has no large differences in these parameters between adjacent supports (abutments excluded). A bridge shall be considered regular for the purposes of this standard, if ‘Smaller (a) Itis straight or describes a sector of an are which subtends an angle of geeaves than 90° at the centre of the arc, and (b) The adjacent columns or piers do not differ in stiffness by more than 25% (Percentage difference shall be calculated based on the lesser of the two stiffhesses as reference)... Draft for IRC:6 Provisions on Seismic Design of Bridges :: Code ce Commentary page 11 of dd 2.2.1.2 Irregular Bridge All bridges not conforming to 2.2.1.1 shall be considered irregular. The classification of bridges into the two categories, namely regular bridges and irregular bridges, included in the draft code is adopted from the AASHTO code of USA. While this classification is only meant to be used as a guide, the responsibility of identifying other irregularities in the chosen bridge structure still reste with the designer. 2.3. Vertical Motions The seismic zone factor for vertical motions, when required, may be taken as two- thirds of that for horizontal motions given in Table 2. The existing codes IRC:6-1966 and \S:1893-1984 prescribe that the vertical accelerations be taken as one-half of the horizontal accelerations for the purposes of design. However, studies on recorded strong ground motion records in the past earthquakes Indicate that the peak ground accelerations (PGA) in the vertical direction is generally about two-thirde of that in the horizontal direction. Thus, the factor of two-thirds is considered more appropriate. Now, the draft building provisions, being discussed in the Earthquake Engineering Sectional. Committee (CED:39) of the Bureau of Indian Standards, include that ‘the seismic zone factor (which reflects the PGA in the seismic zone) for vertical motions be taken as two-thirds of that for horizontal motions. The same provision ie now included for the seismic design of bridges here. 2.4 Live Load The design live loads shall be as specified in the relevant standards. 2.4.1 For Calculation of Magnitude of Seismic Forces Only The live load shall be ignored while estimating the horizontal seismic forces along the direction of traffic. The horizontal seismic force in the direction perpendicular to traffic shall be calculated using 50% of design live load (excluding impact) for railway bridges, and 25% of design live load (excluding impact) for road bridges. The vertical seismic force shall be calculated using 100% of design live load (excluding impact) for railway bridges, and 502% of design live load (excluding impact) for road bridges. The above percentages are only for working-out the magnitude of seismic force. By the live load acting on the span, one usually refers to vehicular traffic. Seismic shaking in the direction of traffic causes the wheels to roll once the frictional forces are overcome. The inertia force generated by the vehicle mass in this case is smaller than that produced if the vehicle mass were completely fastened to the epan. Further, the inertia force generated by the vehicle mass due to friction between the superstructure deck and wheels, is assumed to be ‘taken care of in the usual design for braking forces in the longitudinal direction. Thus, live load is ignored while estimating the seismic forces in the direction of traffic. On the contrary, under seismic shaking in the direction perpendicular to that of traffic, the rolling of wheels is not possible. Thus, live load is included forDraf {C6 Provisions on Seismic Desi idges :: Code & Comment e 12 of d4 shaking in this direction. Here, it is assumed that at the time of the earthquake, 100% of design live load is present on rallway bridges and 50% of design live load is present on road bridges. Further, since live load is friction supported on the rail or on the deck, only a portion of the live load could contribute to the seisme forces; thie Ie taken as 50% of the live load considered. Thus, (a) 50% of design live load in case of rallway bridges, and (b) 25% of design live load in case of road bridges, is recommended, When computing the vertical seismic forces, the entire live load, which is considered to be present on the bridge at the time of the earthquake (as discussed in the above paragraph), is taken. 2.4.2 For Calculation of Stresses Due to Live Load, but to be Combined Stresses due to Seismic Forces For calculating the stresses due to live load to be combined with those due to seismic forces, 100% of design live load (including impact) for railway bridges, and 50% of the design live load (including impact) for road bridges shall be considered at the time of the earthquake. As discussed in the commentary under 2.4.1, it is assumed that at the time of the earthquake, 100% design live load is present on the span in case of railway bridges and only 50% in case of road bridges; the clause reflects the same. 2.5 Seismic Load Combinations 2.5.1 The seismic forces shall be assumed to come from any horizontal direction. For this purpose, two separate analyses shall be performed for design seismic forces acting along ‘two orthogonal horizontal directions, The design seismic force resultants (i.e., axial force, bending moments, shear forces, and torsion) at any cross-section of a bridge component resulting from the analyses in the two orthogonal horizontal directions shall be combined as below: (a) 47) £0.31 (b) 20.3, try where 1 = Force resultant due to full design seismic force along the first horizontal direction, ry = Force resultant due to full design seismic force along the second horizontal direction. 2.5.2 When vertical seismic forces are also considered, the design seismic force resultants at any cross-section of a bridge component shall be combined as below: (@) #7) £0.31) £0.37; (©) £03r) £72 £0.373 (0) 40.371 £0.3r2 £73 where ry and ry are as defined in 2.5.1, and 13 = Force resultant due to full design seismic force along the vertical direction, The design ground motion can occur along any direction of a bridge. Moreover, ‘the motion has different: directions at different time instants. The earthquake ground motion can be thought of in terms of ite components in the two horizontal directions and one vertical direction. For bridges that are termed regular, the two orthogonal horizontal directions (say x- and y-direction) are usually the longitudinal and transverse directions of the bridge. For such bridges, it is sufficient to design the bridge for seiomic forces (\.e, ELx and ELy)Draft for IRC:6 Provisions on Seismic Design of Bridges :: Code & Commentary page 13 of 44 acting along each of the x- and y~directions separately. During earthquake shaking, when the resultant motion Is in a direction other than x and y, the motion can be resolved into x- and y-components, which the elements in the two principal directions are normally able to withstand. However, in case of bridges which are irregular, and particularly in those with skew, design based on considering seismic force in x- and y-directions separately, leads to underdesign of the bridge components. In such a case, the bridge should also be designed for earthquake forces acting along the directions in which the structural systems of the substructures are oriented. One way of getting around this without having to consider too many possible earthquake directions is to design the structure for (a) full design force along x-direction (ELx) acting simultaneously with 30% of tthe design force in the y-direction (Ely); Le., (ELx+0.3ELy), and (b) full design force along y-direction (Ely) acting simultaneously with 30% of the design force in the x-direction (ELx); Le, (O.BELx+ELy). This combination ensures that the components (particularly the substructure) oriented in any direction will have sufficient lateral strength. In case vertical ground motions are also considered, the same principle is then extended to the design force in the three principal directions. 2.6 Increase in Permissible Stresses 2.6.1 Increase in Permissible Stresses in Materials ‘When earthquake forces are considered along with other normal design forces, the permissible stresses in material, in the elastic method of design, may be increased by one- half. However, for steels having a definite yield stress, the stress be limited to the yield stress; for steels without a definite yield point, the stress will be limited to 80 percent of the ultimate strength or 0.2 percent proof stress, whichever is smaller, and that in prestressed concrete members, the tensile stress in the extreme fibers of the concrete may be permitted so as not to exceed two-thirds of the modulus of rupture of concrete. 2.6.2 Increase in Allowable Pressure in Soils ‘When earthquake forces are included, the allowable bearing pressure in soils shall be increased as per Table 1, depending upon type of foundation of the structure and the type of soil The increases in permissible stresses In these clauses are the same as in 161893-1984,Table 1 : Percentage of permissibl Draft for IRC:6 Provisions on Seismic Design of Bridges :: Code & Commentary page 14 of 14 crease in allowable bearing pressure of soils S.No.| Foundation | Type of Soil Mainly Constituting the Foundation ‘Type I - Rock or |Type II Medium Type III Soft Hard Soils : Well|Soils : All soils|Soils : All graded gravel and|with N between 10|soils other than sand gravel and 30, and SP* with N < 10 mixtures with or |poorly graded without clay sands or gravelly binder, and sands with little clayey sands or no fines (SP*) poorly graded or |with N > 15 sand clay mixtures (GB, CW, SB, SW, and SC)* having N** above 30, where N is the standard penetration value 1. [Piles passing 50 50 so through any soil but resting on soil type I 2. |Piles not - 25 25 covered under| item 1 3. Raft 50 50 50 Foundations 4. |Combined isolated RCC 50 25 2s footing with tie beams 5. [Isolated RCC footing 50 25 - without tie jbeams, or unreinforced strip foundations 6. | Weil 50 25 25 foundationsDraft for IRC:6 Provisions on Seismic Design of Bridges :: Code & Commentary page 15 of 44 Notes on Table 1 Note 1: The allowable bearing pressure shall be determined in accordance with IS:6403-1981 *** or IS: 1888-1982 ****, Mote 2; If any increase in bearing pressure has already been permitted for forces other than seismic forces, the total increase in allowable bearing pressure when seisaic force is also included shall not exceed the limits specified above Note 3; Desirable minimum field values of N are as follows: If soils of smaller N-values are met, compaction may be adopted to achieve these values or deep pile foundations going to stronger strata should be used. Seismic Depth below Zone ground level | N Values Remark (in metres) III, Iv = 5 15, and V = 10 25 For values of depths between 5 metres I and II <5 10 and 10 metres, (for important = 10 20 linear interpolation structures is recommended. only) : Note 4: The piles should be designed for lateral loads neglecting lateral resistance of soil layers liable to liquefy. * See IS:1498-1970 Classification and Identification of Soils for General Engineering Purposes (first revision). ** See IS:2131-1981 Method of Standard Penetration Test for Soils (first revision) Code of Practice for Determination of Bearing Capacity of Shallow Foundations (first revision). Method of Load Tests on Soils (second revision)Draft for IRC:6 Provisions on Seismic Design of Bridges :: Code & Commentar page 16 of 44 3.0. Seismic Coefficient Method 3.1 Elastic Seismic Acceleration Coefficient A The Elastic Seismic Acceleration Coefficient A due to design earthquake along a considered direction shall be obtained as A=ZICS where Z= Zone Factor, given in Table 2 for horizontal motion. For vertical motion, refer to 2.3. I = Importance Factor, given in Table 3, T< 40sec C= Bridge Flexibility Factor along the considered direction T> 40sec S= Soil Profile Factor, given in Table 4, and T= Fundamental natural period of the bridge (along the considered direction), However, the bridge flexibility factor C need not exceed 2.5 irrespective of soil type. A plot of CS versus Tis given in Figure 2 Table 2 : Zone Factor Z for horizontal motion. Seismic Zone |_Z 11 0.10 i 0.16 1V 0.24 Vv 0.36 Table 3 : Importance Factor / for different bridges. Use 7 Important Bridges (¢.g., Bridges on National and State Highways) | 1.5 Others 1.0 Table 4 : Soil Profile Factor S for different soil profile types at the Soil Profile Type Ss Type I_:: Rock or Hard Soils 10 ‘Type II :: Medium Soils 1.2 | Type Hil :; Soft Soils 15 ‘Note :: The soil types are classified in Table 1 of I8:1893-1984. Several changes have been Incorsorated in this new elastic seismic acceleration spectrum: (a) The basic horizontal seismic coefficient ay Is replaced by the seismic zone factor Z, and the soll-foundation system factor B has been replaced by a sol-profile factor 5, “enmmmenensenasetessesirsommtersoumiae ES RENNER While the values for | have been retained the same, the expression fcr C has been revised. (b) The term Z now reflects realistic values, as fraction of the acceleration due to gravity, of the expected peak ground acceleration in different selemic zones. For instance, the draft code ezecifies zone IV for areas which are likely +0 sustain shaking of intensity Vill ca the Modified Mercalli Intensity (MMI).. Draft for IRC:6 Provisions on Seismic Design of Bridges :: Code & Commentary page 17 of 44 scale. The value of Z (=0.24) for zone V gives the value of peak ground acceleration as 0.249 which may be reasonably expected in shaking intensity Vill, Adoption of realistic values of peak ground acceleration as the seismic zone factor Z has also rationalised the relative values of design seismic force for different seiemic zones. Data from past earthquakes show that as the intensity of shaking goes up one level on the MMI scale (say from VI to VIl, or from YI to Vill), the peak ground acceleration almost doubles. In the existing Indian codes [1S1895-1984; IRC:6-1966], this is not duly reflected since the seismic force in different zones varied in the ratio 1: The draft code uses a factor of about 15, reaulting in the ratio 116:2.4:3.6, (6) Another change introduced in the draft code ie that the soll-foundation system factor Phas been removed and the soil-profile factor S included. The factor B, depending on the type of soll and the type of foundation, was intended to increase the design force for systems that are more vulnerable to differential settlements. However, in real earthquake situations, bridges do not experience higher earthquake-induced inertia forces on account of vulnerability to differential settlement. Also, the problem of differential settlement cannot be addressed by increasing the design seismic force on the bridge; instead it has to be addressed by a proper choice of the foundation. On thé other hand, records obtained from past earthquakes clearly show that the average acceleration spectrum tends to be different for sites with different soil profiles. The new soll-profile factor S considers this variation. The classification of soll ae given in 1S:1893-1984 is used in this draft code. The values of S are taken from AASHTO code. The product of terms Cand S shown in Figure 3 of the draft code represents ‘the shape of the design spectrum with peak ground acceleration scaled to the value of 1.0. This shape is same as the average shape of the acceleration response spectrum, except in the range O - 0.1 sec. In this range, the value of CS is constant as against the response spectrum which varies from 1.0 to ‘the maximum value (equal to 2.5 In this case) at a period of about O.1 sec. The shape of the response spectrum is modified for design purposes in this range in view of the fact that ductility does not help in reducing the @ maximum forces on the tiff structures with fundamental period in the range 0-0. sec. In developing this C versus T spectrum, 5% damping is implicitly assumed. (e) The fundamental natural period T of the bridge along the considered direction of lateral force is required to obtain the bridge flexibility factor C. The expression proposed for Cin the draft code is taken from the AASHTO code, In case of buildings, experimental measurements are made on existing buildings and empirical expressions are arrived at for the fundamental natural period T of typical building structures. However, in case of bridges, there is a significant variation in the parameters of the bridge even withinDraft for IRC:6 Provisions on Seismic Design of Bridges :: Code & Commentary page 18 of #4 the same structural eystem, Thus, an empirical natural period cannot be arrived at. Hence, recourse to analytical methods becomes essential. 3.1.1 The fundamental natural period T of the bridge along a horizontal direction, may be estimated by the expression T= 20, in which D = Dead load reaction of the bridge in kN, and F = Horizontal force in KN required to be applied at the centre of mass of the superstructure for one mm horizontal deflection of the bridge along the considered direction of horizontal force, For the purposes of the seismic coefficient method, a simple procedure based on static analysis is recommended to obtain the fundamental natural period. The bridge is assumed to behave like a single degree of freedom eystem in the considered direction of shaking and the natural period is obtained by the expression Im VE Here, the mass m of the bridge is obtained from its dead load D (kN, say) by dividing with the acceleration due to gravity g. Alco, in order to obtain the - stiffness k in kN/imm, a force F is applied in the direction of the considered lateral force at the centre of mass of the bridge system such that the displacement along that direction io 1 mm (See Figure Cl). Thus, k= F/I=F And the expression for T modifies to [Dig T= 2 To keep the units consistent, g has to be in mm/sec’, ic, 9810 mm/sec”. Thus, ‘the equation reduces to D Pa 28 I0F Simplifying, D P= 2NTo90F” where D F Dead load reaction of the bridge in kN, and Lateral force in kN required to be applied at the centre of mass of the superstructure for one mm ‘semis deflection of the bridge along the considered direction of lateral force. 3.2. Maximum Elastic Forces and Deformations The inertia forces due to mass of each component or portion of the bridge as obtained from 3.2.1 shall be applied at the centre of mass of the corresponding component or portion of the bridge. A linear static analysis of the bridge shall be performed for these applied inertia forces to obtain the force resultants (e.g., bending moment, shear force and axial force) and deformations (e.g., displacements and rotations) at different locations in1 IRC:6 Provisions. ic ‘ode de Comment 19 of 44 the bridge. The stress resultants ”° and deformations so obtained are the maximum elastic force resultants (at the chosen cross-section of the bridge component) and the maximum elastic deformations (at the chosen nodes in the bridge structure), respectively. The inertia force is generated at the locations of the mass. This clause suggests that the entire inertia force generated in a bridge component be applied as a concentrated load at the centre of mass of that component. Clearly, when the mass is distributed along the dimension of the bridge component, the above approach may result in the incorrect estimation of force resultants due to inertia forces. Designers may require to subdivide such bridge components into smaller segments and evaluate the Inertia force for each of these segments separately. Of course, In such a case, the inertia force generated by the mass of each segment may be proportionally distributed at the end nodes of that segment. in fact, this ie already in practice in the AASHTO code, which requires that: (2) the superstructure should, as a minimum, be modelled as a series of plane frame members with nodes at span quarter pointe, and joint elements. The lumped mass inertia effects should be properly distributed at these locations; and (&) the substructure should be modelled as a series of plane frame members and joint elements. In case of short stiff columns having lengths less than one- third of either of the adjacent span lengths, intermediate nodes are not necessary. However, long flexiblé columns should be modelled with intermediate nodes at the third pointe. The criteria for earthquake resistant design is complete only when all of the following are included: (i) the load factors and allowable stresses, (il) the design acceleration spectium, including the method of obtaining the natural period T, (ii) the damping ratio, and (Iv) the method of analysis. The response reduction factors R to reduce the maximum elastic forces to the design forces, are calibrated keeping in mind these factors. Thus, this clause specifies that linear analysis be conducted to obtain the bending moment, shear force and axial force at different locations in the bridge. 3.2.1 Inertia Force Due to Mass of Each Bridge Component The inertia force due to the mass of each bridge component (e.g., superstructure, substructure and foundation) under earthquake ground shaking along any direction shall be obtained from F°=AW, where A = Elastic Seismic Acceleration Coefficient along the considered direction of shaking obtained as per 3.1, and W= Seismic weight as discussed in 3.2.3. The inertia force due to the mass of a bridge component under earthquake ground shaking in a particular direction depends on the elastic seismic acceleration coefficient: computed for shaking along that direction. Clearly, thisDraft for IRC-6 Provisions on Seismic Design of Bridges :: Code & Commentary page 20 of 14 acceleration coefficient will be different along different directions for the same mags owing to different natural periods along those directions. 3.2.2. Elastic Seismic Acceleration Coefficient for Portions of Foundations below Scour Depth For portions of foundations at depths of 30m or below from the scour depth (as defined in 6.2), the inertia force as defined in 3.2.1 due to that portion of the foundation mass may be computed using the elastic seismic acceleration coefficient taken as 0.54, where A is as obtained from 3.1 For portions of foundations placed between the scour depth and 30m depth below the scour depth, the inertia force as defined in 3.2.1 due to that portion of the foundation mass may be computed using the elastic seismic acceleration coefficient value obtained by linearly interpolating between A and 0.54, where A is as specified in 3.1 The propagation of waves within the body of the earth is modified at the surface of the earth owing to the wave reflections at the boundary surface. For this reason, It is generally accepted that the shaking is relatively more violent at ‘the surface, than below the ground. Hence, the draft code permits reduction in the elastic eelemic acceleration coefficient A for portions of foundations below scour depth. 3.2.3 Seismic Weight The seismic weight of the superstructure shall be taken as its full dead load plus appropriate amount of live load specified in 2,4.1. The seismic weight of the substructure and of the foundation shall be their respective full dead load. Buoyancy and uplift shall be ignored in the calculation of seismic weight. The dead load of the superstructufe also includes the superimposed dead load that is permanently fastened or bonded with its structural self weight. Since there is a limited amount of friction between the live load and the superstructure, only a part of the live load is included in the inertia force caiculations. It le clear that the seismic forces on a bridge component are generated due to its own mass, and not due to the externally applied forces on it. The presence of buoyancy and uplift forces does not reduce its mass. Thus, the clause requires that buoyancy and uplift forces be ignored in the seismic force calculations. 3.3. Design Seismic Force Resultants for Bridge Components The design seismic force resultant V at a cross-section of a bridge component due to earthquake shaking along a considered direction shall be given by a =F, where V°= Maximum elastic force resultant at the chosen cross-section of that bridge component due to earthquake shaking along the considered direction as obtained from 3.2, and R_ =Response Reduction Factor for the component as given in Table 5Draft for IRC:6 Provisions on Seismic Design of Bridges :: Code & Comment of 4 Table 5 : Response Reduction Factor R for Bridge Components and Connections. R Component Superstructure 6 Substructure (a) Reinforced Concrete with special ductile detailing 4 with ordinary detailing 3 (b) Masonry 2 Foundation 2 Connection between Adjacent sections of Superstructure 08 Superstructure and Substructure :: Hinge | 0.8 Superstructure and Substructure :: In-situ | /.0 Substructure and Foundation 10 The basic philosophy of earthquake resistant design io that a structure should not collapse under strong earthquake shaking, although it may undergo some structural as well ae non-structural damage. Thus, a bridge is designed for much less force than what would be required if it were to be necessarily kept elastic during the entire shaking, Clearly, structural damage is permitted but should be such that the structure can withetand the large deformations without collapse. Thus, two Issues come into picture, namely (a) ductility, i. the capacity to withstand deformations beyond yield, and (b) overstrength. Overstrength Is the total strength including the additional strength beyond the nominal design strength considering actual member dimensions and reinforcing bars adopted, partlal safety factors for loads and materials, strain hardening of reinforcing steel, confinement of concrete, presence of masonry infils, increased stength under cyclic loading conditions, redistribution of forces after yield owing to redundancy, etc, (Jain and Navin, 1995]. Hence, the response reduction factor R used to reduce the maximum elastic forces to the design forces reflects these above factors. Clearly, the different bridge components have different ductility and overstrength. For example, the superstructure has no or nominal axial load in it, and hence Its basic behaviour is that of flexure. However, the substructure which ie subjected to significant amount of axial load undergoes a combined axial load- flexure behaviour. It is well-known that the latter system is less ductile than the former. Also, the damage to the substructure is more detrimental to the post earthquake functioning of the bridge than damage to the superstructure span. In the second case, the span alone may have to be replaced, while the first requires an overall rethinking of the use of the bridge: minor modifications may not help. Thus, the R factors for superstructures are kept at a higher value than ‘those for substructures. A similar argument can be given for the R values of foundations which are even lower values than those for substructures. ‘An important issue is that of connections, which usually do not have any significant: post-yield behaviour that can be safely relied upon. Also, there is noDraft for IRC:6 Provisions on Seismic Design of Bridges :: Code & Comment page 22 of 14 redundancy in them. Besides, there is a possibility of the actual ground acceleration during earthquake shaking exceeding the values reflected by the seismic zone factor Z. In view of these aspects, the connections are designed for ‘the maximum elastic forces (and more) that are transmitted through them. Thus, the R factors for connections take values less than or equal to 1.0. For quite sometime now, countries with advanced seismic provisions have been using this approach of obtaining the design forces from the elastic maximum forces. For example, the CALTRANS code uses a factor Z, called the adjustment factor (similar to the response reduction factor R used in this draft code); values of the same are shown in Figure C2. Similarly, the AASHTO code uses a factor R, called the response modification factor, whose values are shown in Table Ct below. 3.4 Multi-directional Shaking ‘When earthquake ground shaking is considered along more than one direction, the design seismic force resultants obtained from 3.3 at a cross-section of a bridge component due to earthquake shaking in each considered direction, shall be combined as per 2.5. 3.5 Combination of Seismic Design Forces with Design Forces Due to Other Effects The design seismic force resultant at a cross-section of a bridge component given by this draft code, shall be combined with those due to other forces, e.g., dead load, live load, wind load, and wave load Table C1 : Response Modification Factor R.as per AASHTO code [AASHTO, 1992]. Substructure! Connections Wall-Type Pier” Superetructure to Abutment Reinforced Concrete Pile Bents Expansion Joints Within a Span a. Vertical Piles Only of the Superstructure b.One or more Batter Piles Columns, Piers or Pile Bents Single Columns to Cap Beam or Superstructure” Steel or Composite Steel Columns or Piers to Foundations” and Concrete Pile Bento a. Vertical Piles Only One or more Batter Piles "the R-Factor is to be used for both orthogonal axes of the substructure. * A wall-type pier may be designed as a column in the weak direction of the pier provided all the provisions for columns required for ductile detailing are followed. The R-factor for a single columh can then be used. ° For bridges classified as SPC C and D, it is recommended that the connections be designed for the maximum forces capable of being developed by plastic hinging of the column bent as specified in the code. These forces will often be significantly less than those obtained using an R-factor of 1.Draft for IRC-6 Provisions on Seismic Design of Bri 4.0 Response Spectrum Method The Response Spectrum Method requires the evaluation of natural periods and mode shapes of several modes of vibration of the structure. This method will usually require usage of a suitable space frame dynamic analysis computer program Code & Commenter 23 of 4 4,1 Elastic Seismic Acceleration Coefficient 4, in Mode k The elastic seismic acceleration coefficient 4, for mode k shall be determined by Ay = ZIC,S, where Z, / and S are as defined in 3.1, and C, is the bridge flexibility factor for mode k given by the following expression: 25 i < 40sec 1% Cee , Ty > 4.0 sec (a where 7, is the natural period of vibration of mode & of the bridge. However, the bridge flexibility factor Cz for mode & need not exceed 2.5 irrespective of soil type. For modes other than the fundamental mode, the bridge flexibility factor Cy in mode & for Ty $0.1 sec may be taken as Cy 14 1ST. A plot of CyS versus 7; is given in Figure 3 Typical shape of the acceleration sesponse spectrum when plotted with natural period on the x-axis, is shown in Figure C3(a). It etarts at the value of Peak Ground Acceleration (PGA) at zero period, rises to about 2.5 times (for 5% damping) the PGA value at: a period of about 0.1 sec, and then remains at that value upto about O3 sec period. However, seismic design codes usually assume ‘the design spectrum shape to be horizontal for the range from 0.3 sec all the way upto zero period (le, the codes igncre the fact that the spectrum has lower values of acceleration in the range of O-0.1 sec, as shown in Figure C3(b)). There are several reasons for this conservatism. For instance, ductility does not help in reducing the maximum forces if natural period in this range of O-O.1 sec [Riddel et al, 1989]; hence, one needs to raise the level of spectrum in this range. Also, since the acceleration response soectrum has‘ a very steep slope in the range O-0.1 sec; a small underestimation of the natural period T may lead to a significant reduction in the seismic force. However, in multimode analysis this dratt allows the designer to use the ascending part of the spectrum in the range O-0.1 sec but only for the higher modes of vibration. Since, the fundamental mode makes the most significant contribution to the overall response ard the contribution of higher modes is relatively small, this is now permitted by several codes [¢.g, AASHTO, 1992}, 4.2 Inertia Force due to Mass of Bridge at Node i in Mode k The vector {re of inertia forces to be applied at different nodes in mode k of vibration due to earthquake shaking along a considered direction shall be obtained as,Pe A g IRC:6 Provisions on Seismi of Bridges :: Code & Comment {rt\=tmi fo} des, seismic mass martix of the bridge structure, as defined in 4.2.1, Mode shape vector of mode & of vibration of the bridge structure obtained from free vibration analysis, = Modal participation factor of mode & of vibration of the bridge structure for a given direction of earthquake shaking, = Elastic seismic acceleration coefficient for mode & as defined in 4.1, and = Acceleration due to gravity The above expression is part of the routine solution procedure for analysis of elastic structures subjected to seismic ground motion represented by ite pseudo-acceleration response spectrum. The mathematical model of the bridge structure should properly account for all stiffnesses and masses. A suitable number of intermediate nodes are required for each bridge component to properly estimate the stress resultant caused by the seismic inertia forces generated. In doing 60, guidance may be sought from current AASHTO code practices already discussed in the commentary under 3.2. Rotational moment of Inertia of certain masses in the bridge structure may become important particularly in case of joint elements; the same may be incorporated in the matrix of seismic weights as mass moment of inertia times acceleration due to gravity. 4.2.1. Seismic Mass Matrix The seismic mass matrix of the bridge structure shall be constructed by considering its seismic weight lumped at the nodes of discretisation. The seismic weight of each bridge component shall be estimated as per 3.2.3, and shall be proportionally distributed to the nodes of discretisation of that bridge component. ‘The seiomic weight of each bridge component is proportionally distributed to its end nodes and intermediate nodes ae lumped masses considering its geometry. These lumped masses are used to form the matrix of seismic weights keeping in mind that the mass lumped at a node contributes to all the translational degrees of freedom at that node. 4.2.2. Number of Modes to be Considered ‘The number of modes to be considered in the analysis shall be such that at least 90% of the seismic mass of the structure is included in the calculations of response for earthquake shaking along each principal direction. This clause indirectly requires that all modes that contribute significantly to the response be included in the analysis. And, the book-keeping is done through the modal masses. Clearly, the modes with low participation in the dynamics of tthe bridge for earthquake shaking along a chosen principal direction, will have very small modal maes and the dynamic force carried by these modes would alo be small. The clause suggests that at least 90% of the total seismic mass (as defined in 4.2.1 and 3.23) shall be included through the modes that are considered.Draft for IRC:6 Provisions on Seismic Design of Bridges :: Code & Commentary page 25 of #4 4.3 Maximum Elastic Forces and Deformations The maximum elastic seismic forces in mode k obtained from 4.2 shall be applied on the bridge and a linear static analysis of the bridge shall be performed. The maximum elastic force resultants Fy (e.g., bending moment, shear force and axial force) and the maximum elastic deformations (e.g., displacements and rotations) in mode k at different locations in the bridge for a considered direction of earthquake shaking, The maximum elastic force resultants F<, and the maximum elastic deformations, due to all modes considered, for the considered direction of earthquake shaking, shall be obtained by combining those due to the individual modes by either (a) the Complete Quadratic Coefficient (CQC) Method, or (b) the Square Root of Sum of Squares (SRSS) Method described in 4.3.1 and 4.3.2, respectively. 4.3.1 CQC Method Let the modal response quantity due to ith mode of vibration be 4. Also, let r be the number of modes considered. Then, the net response quantity 4 may be estimated as: where Pare 53,120) Se BP 781+ BY? oj B =), and a =Modal damping ratio. Here, itis assumed that the modal damping ratio is same for all modes considered; else, the above expression shall be replaced by appropriate equations. 4.3.2 SRSS Method Let the modal response quantity due to ith mode of vibration be 4, and let r be the number of modes considered. Then, the maximum response 2. due to all modes considered may be estimated as 5 De The modal response quantities (e.g, bending moment, shear force, axial force, dieplacements and rotations at any locationof the bridge) in each mode k need to be combined to obtain the maximum response due to all modes considered. Studies on modal response combinations show that when modal frequencies are well-separated, the Square Root of Sum of Squares (SRSS) Method provides reasonable estimates. if two modal frequencies are separated from each other by 10% (or less) of the smaller one, then the two modes may be termed as closely-spaced modes. However, when modal frequencies are closely-spaced or nearly closely-spaced, the SRSS method gives poor results. In fact, the Complete Quadratic Coefficient (CQC) Method provides, in general, reasonably good estimates of net response, irrespective of whether the modal frequencies are closely-spaced or well-separated. However, the COC method as stated in 4.3.1 assumes that the modal damping ratio is same for all considered modes ofDraft for IRC:6 Provisions on Seismic Design of Bridges :: Code & Commentary page 26 of 44 vibration. In case it is not 90, reference shall be made to literature [e.g Chopra, 1995] for suitable expressions for modal response combination. 4.4 Design Seismic Force Resultants in Bridge Components The design seismic force resultant V,,¢, at any cross-section in a bridge component for a considered direction of earthquake shaking shall be determined as, re py, —fnet ‘net = Re where the maximum elastic force resultant Fj, due to all modes considered is as obtained in 4,3, and the Response Reduction Factor R of that component of the bridge is as per Table 5 As discussed in the commentary under 3.3, the various components of the bridge do not enjoy the same level of ductility and overstrength. Hence, the level of design seismic force vie-d-vie the maximum elastic force that will be experienced by the component if the entire bridge were to behave linear elastically, varies for different bridge components. The values of the response reduction factor R given in Table 5 reflect the same. 4.5 Muiti-directional Shaking When earthquake ground shaking is considered along more than one direction, the design seismic force resultants obtained from 4.4 at a cross-section of a bridge component due to earthquake shaking in each considered direction, shall be combined as per 2.5. 4.6 Combination of Seismic Design Forces with Design Forces Due to Other Effects The design seismic force resultant at a cross-section of a bridge component given by this draft code, shall then be combined with those due to other forces, e.g., dead load, live load, wind load, and wave load 4.7 Site-Specific Spectrum In case design spectrum is specifically prepared for a structure at a particular site, the same may be used for design. However, the bridge structure shall still comply with the minimum requirements specified in this standard. To ensure at least a minimum strength in the bridge structure, this clause prevents the designer from using a site-specific spectrum that reeulte in unduly ‘small design force resultant in comparison with those given by this draft code. Superstructure 5.1 The superstructure shall be designed for the design seismic forces specified in 3. or 4., plus the other loads appearing on it, ¢.g., dead load, live load, wind load, and wave load. 5.2 Under simultaneous action of horizontal and vertical accelerations, the superstructure shall have a factor of safety of at least /.5 against overturning in the transverse direction. Since the supporting width of the span in the transverse direction is relatively email in comparison with that in the longitudinal direction, overturning of superstructures (that are rested on the substructure and not monolithically connected with it) in the transverse direction may be possible under theDraft for IRC:6 Provisions on Seismic Design of Bridges :: Code & Commentar page 27 of td combined action of seismic forces along transverse and vertical directions. OF course, in these calculations, the direc taken 60 as to produce the worst effect, ion of vertical seismic force shall be 5.3. The superstructure shall be secured to the substructure, particularly in seismic zones IV and V, through vertical hold-down devices and/or horizontal linkage elements as specified in 5.4 and 5.5, respectively. These vertical hold-down devices and/or horizontal linkage elements shall also be used to secure the suspended spans, if any, with the restrained portions of the superstructure. However, the frictional forces shall not be relied upon in these calculations. This clause makes it mandatory in hign seismic regions to have the following devices provided between the supers ucture (that ie not monolithically connected with the substructure) and suzstructure, and between the suspended spans, if any, and restrained portion of tre superstructure: (2) vertical hoid-down devices to prevent che cuperstructure from lifting off from its supports atop the substructure particularly under vertical selsmic forces combined with the transverse seismic “orces, and (b) horizontal linkage elements to prevent excessive relative deformations between portions of the euperstructre or between the superstructure and substructure. 5.4. Vertical Hold-Down Devices Vertical hold-down devices shall be provided at all supports (or hinges in continuous structures), where vertical seismic force U due to the maximum elastic horizontal seismic lateral force opposes and exceeds 50% of the dead load reaction D. 5.4.1 Where vertical force U, due to the maximum elastic horizontal seismic force, opposes and exceeds 50%, but is less than 100%, of the dead load reaction D, the vertical hhold-down device shall be designed for a minimum net upward force of 10% of the downward dead load reaction that would be exerted if the span were simply supported. 5.4.2. If the vertical force U, due to the horizontal seismic force, opposes and exceeds 100% of the dead load reaction D, then the device shall be designed for a net upward force of 1.2(U-D); however, it shall not be less than 10% of the downward dead load reaction that would be exerted if the span were simply supported Vertical hold-down devices are consicered essential in the draft provisions to minimise the potential of adverse effeccs of vertical seismic excitation. The provisions for design force of vertical ho'a-down devices have been adapted from the AASHTO code. 55 Horizontal Linkage Elements Positive horizontal linkage elements (e.g., high tensile wire strand ties, cables and dampers) shall be provided between adjacent sections of the superstructure at supports and at expansion joints within a span 5.5.1 The linkages shall be designed for, at least, elastic seismic acceleration coefficient A times the weight of the lighter of the two adjoining spans or parts of the structure, 5.5.2 If the linkage is at locations where relative deformations are designed to occur, then sufficient slack must be allowed in the linkage so that linkages start functioning only when the design relative displacement at the linkage is exceeded= Draft for IRC:6 Provisions on Seismic Design of Bridges :: Cade de Commentary page 28 of 4 5.5.3 When linkages are provided at columns or piers, the linkage of each span may be connected to the column or pier instead of to the adjacent span. The design seismic force for each bridge component is only fraction of the maximum elastic force that can be sustained by it, if it were to completely remain elastic during earthquake shaking. However, the deformations calculated from the linear analysis of the bridge subiected to these design forces are much smaller than the actual deformations experienced during seismic shaking. Unseating of superstructure from the substructure or of the suspended span from the restrained portion are the possible consequences if the actual deformations are not accounted for in the design of the supports at these interface points. Sometimes, the two portions that move relative to each other are securely fastened by devices called the positive horizontal linkage elements. ‘These devices are usually either high tensile wire strand ties, cables or dampers. For the purposes of the design of these devices, the experience from the AAGHTO code is used. The design forces are stated to be reasonably conservative to provide increased protection at a minimum increased cost. 5.6 Submersible Bridges 8.6.1. The hydrodynamic pressure p (in N/m”) on a submersible bridge superstructure shall be determined by p=8750A\yH, where A= Elastic seismic acceleration coefficient given by 3.1, y= Depth of section below water surface (in m), and H = Height of water surface from level of deepest scour (in m). 5.6.2 The total horizontal shear Vj, (in N/m) and moment M, (in Nim/m) per meter width about the centre of gravity of the base at any depth y due to hydrodynamic pressure are given by 2 Vy, =2 py n= 5Py 402 M,=+. and My = 75 PY Both 5.6.1 and 5.6.2 are retained as given in the existing 11893-1964 except for the following change. In 5.6.1, “A” has taken the place of “a,”. Clearly, A may be much larger than dy, ie,, the existing expression in 1S:1822-1984 uses the “reduced” accelerations while tris draft code uses the maximum accelerations expected during earthquake shaking, This implies that the hydrodynamic pressures calculated as rer the expression in this draft code will be significantly higher. Again, thie drat: code clearly distinguishes the actual forces appearing on the bridge from thoze used in its design. 6. Substructure 6.1 Design earthquake forces and forces due to maximum flood shall not be considered to occur simultaneously. The loads specified in this standard cover general conditions. The designer shall also provide for other loads where they might be critical, e.g., vehicle or ship impact on substructure..- Draft for IRC:6 Provisions on Seismic Design of Bridges :: Code & Commentary page 29 of 4 6.2 Scour Depth Earthquake forces on the substructure shall be calculated based on the depth of scour caused by the discharge corresponding to the annual mean design flood. In the absence of detailed data, this depth shall be taken as 90% of the maximum scour depth. The clause 6.2 is retained as given in IRC:6-1966. 6.3 Design Seismic Force The design seismic forces for the substructure shall be the obtained as the maximum elastic force on it (as defined in 6.3.1) divided by the appropriate response reduction factor given in Table 5 6.3.1. Maximum Elastic Seismic Forces The maximum elastic seismic force resultants at any cross-section of the substructure shall be calculated considering all of the following forces on it: (@) Maximum elastic seismic forces transferred from the superstructure to the top of the substructure through bearings (Figure 4) (b) Maximum elastic seismic forces applied at its centre of mass due to the substructure’s own inertia forces. Reduction due to buoyancy and uplift shall be ignored in the calculation of seismic weight. (©) Hydrodynamic forces owing to stream flow acting on piers, and modification in earth- pressure due to earthquake acting on abutments, 6.3.1.1 When the substructures are oriented normal to the direction of the traffic and along the direction of stream flow, two separate load cases, namely seismic forces acting parallel (a) to the current direction, and (b) to the traffic directions, shall be considered ‘And, when the substructures are oriented skew either to the direction of traffic or to the direction of current, the load combination as given in 2.5, shall be considered. 6.3.2 While considering the stability of the substructure against overturning, the minimum factor of safety shall be 1.5 under simultaneous action of maximum elastic seismic forces in both horizontal and vertical directions during the earthquake. ‘These clauses under 6.3 are retained as given in IS1B93-1984. 6.4 Hydrodynamic Force For the submerged portion of the pier, the total horizontal hydrodynamic force along the direction of ground motion is given by F=CAW., where C, is a coefficient given by Table 6, depending on the height of submergence of the pier relative to that of the radius of a hypothetical enveloping cylinder (Figure 5); and A is the elastic seismic acceleration coefficient as per 3.1; and W, is the weight of the water in the hypothetical enveloping cylinder. The pressure distribution due to the hydrodynamic effect on the pier is given in Figure 6; the coefficients C),Cp,C3 and Cy in Figure 6 are given in Table 7. This clause is retained as given in ISA893-1984, except that “A” replaces “ay,”. Again, as stated in the commentary under 6.6, “A” is different: from “o1y.” Hence, the hydrodynamic forces calculated as per this code will be much higher than those estimated as per the existing code IS1893-1984.Drafi for IRC6 Provisions on Seismic Design of Bridges :- Code & Commentary page 30 of 14 Table 6 : Values of C,. Height of Submerged Portion of PierH | C, Radius of Enveloping Cylinder 10 0.39 2.0 0.58 3.0 0.68 40 0.73 Table 7 : Pressure Distribution Coefficients C,, C,, C, and C,. CG G G CG o1 | 0470 [0.026 | 0.9345 02 | 0.673 [0.093 | 0.8712 03 | 0.832 [0784 | 0.80/3 04 | 0.922 | 0.289 | 0.7515 05 | 0.970 [0.403 | 0.6945 0.6 | 0.990 | 0.521 | 0.6390 08 | 0.999 | 0.760 | 0.5320 1.0 | 4.000 | 7.000 [0.4286 7, Foundations 7.1 In loose sands or poorly graded sands with little or no fines, the vibrations due to earthquake may cause liquefaction or excessive total and differential settlements. Founding of bridges on such sands shall be avoided in seismic zones III, IV and V, unless appropriate methods of compaction or stabilisation are adopted for soils. ‘This clause ie retained as given in IRC:6-1966, 7.2, When substructures terminate on a footing which rests on rock or on piles, they may be considered rotationally fixed. Foundations on soft material may be modelled using equivalent linear spring coefficients. Also, well foundations may be analysed assuming soil springs as lateral supports. 7.3 Seismic Zones IV and V The foundations of bridges in seismic zones IV and V shall be designed to resist smaller of the following: (a) Design seismic forces obtained from 3.3 or 4.4, and (b) Forces developed when overstrength plastic moment hinges are formed in the substructure, as described in 9. Damages to foundations have vey serious implications from structural safety considerations. Also, foundation repairs are very expensive as it is very difficult to access and to make alterations in them. Hence, it is required to ensure that these are not damaged. “ris clause is intended to achieve the objective that in case of severe ground shaking, the foundation Is not damaged. This is done first by requiring 2 much lower value of response reduction factor for foundation than for the substructure, ie, a much higher design seismic coefficient for foundation that for the substructure. However, this is qualified through the concept of capacity design (e.g, Paulay and Priestley, 1992; and Jain and Murty, 1996].. Draft for IRC:6 Provisions on Seismic Design of Bridges :: Code & Commentary Grodiner pe poge Since the seismic forces are inertia induced, the foundation can never experience a seismic force higher than what the substructure is capable of ‘transmitting to it. In section (b) here, the attempt is to obtain this upper-round force that can be transmitted by the substructure by calculating the overstrength plastic moment capacity of the substructure. The code requires the lower of (a) and (b) to be used in design of the foundation. 8. Connections 81 Design Force for Connections within Superstructure and between ‘Superstructure and Substructure 8.1.1 Seismic Zones I, It and IIT The connections between adjacent sections of the superstructure or between the superstructure and the substructure shall be designed to resist at least a horizontal seismic force in the restrained directions equal to 0.20 times the vertical dead load reaction at the bearing, irrespective of the number of spans. In low seismic regions, the effort in the seismic design of the bridges is reduced to some extent by this clause by requiring only a simple design force calculation for the restrained supports (2g. rocker or elastomeric bearings). The clause has been borrowed from the AASHTO code and is considered to provide a somewhat overestimate of the design force. 8.1.2 Seismic Zones IV and V The connections between the superstructure and substructure, and the substructure and foundation shall be designed to resist the smaller of the following: (a) Maximum elastic horizontal seismic force transferred through it in the restrained directions divided by the appropriate Response Reduction Factor R applicable to connections, which are given in Table 5, and (b) Maximum horizontal force that develops when overstrength plastic moment hinges are formed in the substructure. The most common cause for earthquake disasters in case of bridges is the failure of connections, particularly those between superstructure and the substructure. Hence, extra caution is needed to ensure the safety of connections. This is done in this draft code by requiring the value of response reduction factor for bridges as O.8 or 1.0 (See Table 6); this implies that the design force for connections obtained by (a) above is equal to (or more than) the maximum expected elastic force. However, by allowing the designer to use the lower of (a) and (b) above for design of connections, the code brings in the capacity design concept. Force obtained by (b) above provides an upper-bound on the inertia force that can be developed in the superstructure before the substructure becomes plastic. Once the substructure becomes plastic, the bridge will not be able to sustain higher inertia forces. 82 Provisions to Account for Displacements at Connections where Motions are Permitted 8.2.1 Separation Between Adjacent Units When relative movement between two adjacent units of a bridge are designed to occur at a separation joint, sufficient clearance shall be provided between them, to permitDraft for IRC:6 Provisions on Seismic Design of Bridges -» Code & Commentar page 32 of 4d the calculated relative movement under design earthquake conditions to freely occur without inducing damage, Where the two units may be out of phase, the clearance to be provided may be estimated as the square root of the sum of squares of the calculated displacements of the two units under maximum elastic seismic forces given by 3.2 or 4.3. When two adjacent units are designed such that relative movement between ‘them is expected to occur at their separation joint, then adequate clearance is necessary between them to avoid pounding and the consequential damage. To provide the cumulative um of the displacements of the two units as the separation would be too conservative. Thus, this clause proposes that the square root of the sum of squares of the calculated displacements of the two units under the earthquake forces may be provided as the clearance. 8.3 Minimum Width of Seating at Supports of Superstructure on Substructure, or of the Suspended Span Portion on the Restrained Portion of the Superstructure The widths of seating W (in mm) at supports measured normal to the face of the abutment/pier/restrained portion of superstructure from the closest end of the girder shall be the larger of the calculated displacement under the maximum elastic seismic forces estimated as per 3.2 or 4,3, and the value specified below S00+15L+6H for seismic zones I, Mand II 800+25L+10H for seismic zones IV and V where . = Length (in meters) of the superstructure to the adjacent expansion joint or to the end of superstructure, In case of bearings under suspended spans, itis sum of the lengths of the two adjacent portions of the superstructure. In case of single span bridges, it is equal to the length of the superstructure, and For bearings at abutments, H = Average height (in meters) of all columns supporting the superstructure to the next expansion joint. It is equal to zero for single span bridges. For bearings at columns or piers, H = Height (in meters) of column or pier. For bearings under suspended spans, H = Average height (in meters) of the two adjacent columns or piers. Graphical representation of seating widths are shown in Figure 7. The connections between superstructures and substructures are designed for forces specified under 8.1. Even though these are conservative values, there still will remain possibilities of the actual seismic force in the connections exceeding the actual strength of the connections. Also, In bridges the substructures are liable to large displacements due to dynamic earth-preseures. Under these conditions, it is possible that the superstructure span may be separated from the connection. At this instance, if adequate width Ie available on top of the substructure for the superstructure span to rest (despite being separated from the connections), then at least the superstructure span is prevented from being dislodged from its support. Clearly, if the superstructure span is still resting atop the substructure, the cost of repairing the connection and restoring the superstructure span to its desired position is far moreDraft for IRC:6 Provisions on Seismic Design of Bridges :: Code & Commentary page 33 of 44 economical than having to rebuild the superstructure afresh if it falls off from the substructure. Hence, this clause attempts that even under maximum expected deformations, possibility of collapse or loss of span are minimised through conservative provisions of minimum seating widths, The values of seating widths recommended for high seismic regions are about 6% higher than those for low seismic regions; this is because of higher potential of connection failures in high seismic zones. The above provision for minimum seating widths W, (in mm) is similar to that adopted in the the AASHTO code, given by jp +1.67L+6.66H for low seismic performance categories W= . [s05+26.+10H for high seismic performance categories where Land h are as defined in the draft code. Clearly, this draft: code requires a higher seating width than the American practice. This is motivated by the Japanese practice; the Japan code (JRA, 1990] requires that the seat length Sg (In mm) from edge of superstructure to the edge of the substructure shall be longer than the vaiue estimated by the expression | 700+6L L<100m Se= [eoo+at L>100m where L represents the span length (in ri). 9. Capacity Design of Bridge Components The design seismic force for bridges is lower than the maximum expected seismic force on them, However, to ensure good performance at low cost, the difference in the design seismic force and the maximum expected seismic force shall be accounted for through additional cautions. The provisions given under 9. shall be applicable to seismic zones IV and V only, This clause requires some additional provisions which ensure that brittle failure modes do not precede the ductile fallures. In a structure having both brittle and ductile e'ements, if it can be ensured that the ductile elements will yield prior to failure of brittle elements, the post-yield behaviour of the structure will be ductile. The concept of capacity design is used to ensure post-yield ductile behaviour of a structure having both ductile and brittle elements. In this method, ‘the ductile elements are designed and detailed for the design forces. Then, an upper-bound strengti of the ductile elements is obtained. it is then expected that if the seismic “orce keeps increasing, a point will come when these ductile elements wil reach cheir upper-bound strength and become plastic. Clearly, we now need to ensure chat even at that level of seismic force, the brittle elements remain safe. This procedure is referred to as the capacity design procedure [Paulay and Priestley, 1992}Dra IRC:6 Provisions on Seismic Design of Bridges :: Code tenta sof 44 9.1 Single-Column or Single-Pier Substructure Provisions given in IS:13920-1993 for the ductile detailing of RC members subjected to seismic forces shall be adopted for such components of the bridge. In particular, the design shear force for a single column type substructure (as against the frame type substructure) shall be the higher of the following, (a) Maximum elastic shear force at the critical section of the bridge component divided by the response reduction factor for that components as per Table 5, and (b) Maximum shear force that develops when the substructure has maximum moment that it can sustain (i.e., the overstrength plastic moment capacity as per 9.4) ‘The locations for the critical sections are at the bottom of the column or pier in a single- column substructure as shown in Figure 8a. ‘The clause is meant to ensure ductile behaviour of the columns or piers. In RC. members, flexural failure can be ductile if the member is detailed appropriately. On the other hand, shear failure is brittle. Hence, the columns are designed arid detailed for flexure firot. Then, using the principle of capacity design, one calculates how much is the maximum possible earthquake force that this column can sustain in the event of strong shaking, Since the shear failure is a brittle failure, shear design for columns is carried out for this upper bound load, Note that a similar provision for shear design of beams in RC frame buildings is already in practice for many years for buildings, eg. clause 7.2.5 in 15:4326-1976 and clause 6.3.3 in 1S:13920-1998. 9.2 Multiple-Column or Multiple-Pier Substructure Provisions given in IS:13920-1993 for the ductile detailing of RC members subjected to seismic forces shall be adopted for such components of the bridge. In particular, the design shear force for a multi-column frame-type or multi-pier substructure shall be the higher of the following (@) Maximum elastic shear force at the critical section of the bridge component divided by the response reduction factor for that components as per Table 5, and (b) Maximum shear force that is developed when plastic moment hinges are formed in the substructure so as to form a collapse mechanism. Here, the plastic moment capacity shall be the overstrength plastic moment capacity as per 9.4. The locations for the critical sections are at the bottom and/or top of the columns/piers in rmult-column frame-type substructures or multi-pier substructures as shown in Figure 8b. This clause attempts to achieve the same objectives as discussed earlier in commentary to clause 9.1, but for the multiple-column or multi substructures. le-pler 9.3 Design Force for Connections Connections at the restrained ends shall be designed for the lower of the following: (a) Maximum elastic shear force transferred through them at the critical section of the bridge component divided by the response reduction factor for that components as per Table 5, and (b) Maximum shear force that develops when the substructure has maximum moment that it can sustain (i.., the overstrength plastic moment capacity as per 9.4),Draft for IRC:6 Provisions on Seismic Design of Bridges :: Code d& Commentary page 35 of 14 9.4 Overstrength Plastic Moment Capacity 9.4.1 Limit State Method of Design The overstrength plastic moment capacity at a reinforced concrete section shall be taken as 1.4 times the ultimate moment capacity based on the usual partial safety factors for materials and loads, and on the actual dimensions of members and the actual reinforcement detailing adopted. The factor arises from the following. The partial safety factor used in limit state design (15:456-1978] for reinforcing steel is 1.15, i. the yield stress in oteel used in design is 0.878, Further, in estimating the overetrength capacity, it io assumed that 25% increase in steel stress is possible owing to strain- hardening in it; thus, the maximum stress in reinforcing bars for design purposes can be up to 1.25f,, as against 0.676, used in calculating moment capacity as per IS:456. Since sections are necessarily under-reinforced for ductile behaviour, the ultimate moment carrying capacity is influenced primarily by the stress in steel, and only very marginally by the grade of concrete. Thus, the ultimate moment capacity of the section can be scaled-up proportional to the ratio of ‘the maximum stress in steel in the two cases, to obtain the plastic moment hinge capacity. Thus, 1.25 X (0.87) = 1.437, which is rounded off to 1.4. Similar factor of 1.419 in practice in the ductile deatiling provieions of IS:13920-1993 for reinforced concrete members 9.4.2. Working Stress Method of Design The overstrength plastic moment capacity at a section shall be based on the permissible stresses mentioned below, and on the actual dimensions of the members and the actual reinforcement detailing adopted. The permissible stresses applicable for materials under gravity load conditions shall be first multiplied by 1.5 (as specified in 2.6.1, to account for the instantaneous application of the maximum earthquake forces) and then by 1.4 to account for overstength in materials, The entire framework oF capacity design Is really applicable only to the limit state design. It is difficult to extend it to the working stress method of design. However, considering that the Indian professionals will continue to use working stress method for design of bridges for many years to come, this clause provides a crude way to implement the capacity design concept for such situations. The factor 1.4 is kept same as in limit state method. 10. Summary and Conclusions ‘A draft proposal for the seismic design of bridges is presented for the next revision of IRC:6, Many of the issues raised in the earlier reports on the performance of bridges in India during past earthquakes [Murty and Jain, 1996] and on the state-of-the-art review of IRC:6-1966 provisions [Jain and Murty, 1996] have been incorporated. The following is a brief summary of some major and important modifications made in this proposal: © Relative values of seismic zone factor have been changed; these are the same as the ‘ones included in the draft provisions of IS:1893, which is under revision. © Two methods, namely Seismic Coefficient Method and Response Spectrum Method, are given for estimating design seismic forces, this is in line with the draft Indian code for buildings and codes of some other countries for seismic design of bridges. TheseDraft for IRC:6 Provisions on Seismic Design of Bridges :: Code & Commentary page 36 of 44 methods replace the existing methods in the IRC:6-1966 and IS:1893-1984, which do not recognise the flexibility of the bridge. + The concept of ductility and overstrength is brought into the code explicitly, by introducing the response reduction factors in place of the performance factor. ‘© Different response reduction factors have been proposed for the different components of the bridge, depending on the expected ductility and overstrength in them. «The design force level for bridges has been raised from the existing levels and brought in line with the current international practices. © The concept of capacity design is introduced in the design of connections, substructures and foundations. ‘+ The soil-foundation system factor is dropped. A soil profile factor depending on the soil profile has been introduced for obtaining the design spectrum. * Design for displacements in the structure is introduced. * Use of vertical hold-down devices and horizontal linkage elements to account for the large displacements generated during seismic shaking, is made mandatory in certain bridges. + A minimum width of seating of superstructure over substructures to avoid collapse of spans from the atop the substructures, is required for all bridges. The proposed draft includes significant improvements over the IRC:6-1966 and the 1S:1893-1984. However, there are still a number of areas that needs to be further improved. These include detailed clauses on the design and detailing of individual components of foundations and abutments, of all structural steel bridge components, and of all reinforced concrete bridge components. 11. References ‘AASHTO 1992, 1992, Standard Specifications for Seismic Design of Highway Bridges, ‘American Association of State Highway and Transportation Officials, Washington, D.C., USA. CALTRANS 1991, 1991, Bridge Design Specifications, California ent of Transportation, Sacramento, C.A., USA. CalviM,, and Priestley,MIN,, 1991, Seismic Design and Retrofittting of Reinforced Concrete Bridges, Proceedings of the International Workshop on Seismic Design and Retrofittting of Reinforced Concrete Bridges, 2-5 April 1991, Bormio, Italy. Chopra,AK., 1995, Dynamics of Structures : Theory and Applications to Earthquake Engineering, Prentice Hall, USA. IRC:6-196, 1985, Standard Specifications and Code of Practice for Road Bridges Section II : Loads and Stresses, Indian Roads Congress, New Delki 18:1893-1984, 1984, Indian Standard Criteria for Earthquake Resistant Design of Structures, Bureau of Indian Standards, New Delhi. 1S:4326-1976, Indian Standard Code of Practice for Earthquake Resistant Design and Construction of Buildings, Bureau of Indian Standards, New Delhi, 18:13920-1993, 1993, Indian Standard Code of Practice for Ductile Detailing of Reinforced Concrete Structures Subjected to Seismic Forces, Bureau of Indian Standards, New Delhi Jain,S.K., and Murty,C.V.R., 1996, “Review of the State-of-the-Art in Seismic Design of Bridges,” Report submitted to Ministry of Surface Transport, New Delhi, October. Jain,$.K., and Navin,R., 1995, “Seismic Overstrength in Reinforced Concrete Frames,” Journal of Structural Engineering, ASCE, Vol.121, No.3, March 1995, pp 580- 585.Draft for IRC:6 Provisions on Seismic Design of Bridges :: Code é& Comment e 37 of 44 JRA, 1990, Design Specification for Highway Bridges, Part I: General, Part II: Steel Bridges, Part III: Concrete Bridges, Part IV: Foundation, Part V: Seismic Design, Japan Roads Association, February. Murty,C.V.R., and Jain,S.K., 1996, “Seismic Performance of Bridges in India During Past Earthquakes,” Report submitted to Ministry of Surface Transport, New Delhi, March. NZS:3101-1982, Code of Practice for the Design of Concrete Structures, Standards Association of New Zealand, Standards Association of New Zealand, Wellington, New Zealand NZS:4203-1991, Code of Practice for General Structural Design and Design Loadings ‘for Buildings, 2/DZ 4203/2 Draft for Comment, Standards Association of ‘New Zealand, Wellington, New Zealand, . Paulay,T., and Priestley,M.IN., 1992, Seismic Design of Reinforced Concrete and ‘Masonry Buildings, Jobn Wiley & Sons, Inc., New York, USA. Riddel.R., Hidalgo,P., Cruz,E., 1989, “Response Modification Factors for Earthquake Resistant Design of Short Period Buildings,” Earthquake Spectra, EERI, CA, USA, Vol.5, No.3., pp 571-590. TINZ Bridge Manual, 1991, Bridge Manual : Design and Evaluation, Transit New Zealand, Wellington, New Zealand. Acknowledgments The authors are thankful to the Ministry of Surface Transport, New Delhi, for the financial support, which made this work possible... Draft for IRC:6 Provisions on Seismic Desi Code & Comment 5 38 of Hd List of Tables Table 1 : Percentage of permissible increase in allowable bearing pressure of soils Table 2 : Zone Factor Z for horizontal motion Table 3 : Importance Factor J for different bridges. Table 4: Soil Profile Factor S for different soil profile types at the site. Table 5 : Response Reduction Factor R for Bridge Components and Connections. Table 6 : Values of C,, Table 7 : Pressure Distribution Coefficients C,, C, C; and C, Table C1 : Response Modification Factor R as per AASHTO code [AASHTO, 1992]. List of Figures Figure 1 : Seismic zone map of India. Figure 2 : Plot of CS versus natural period T to be used in the Seismic Coefficient Method Figure 3 : Plot of CS versus natural period 7; in mode k of the bridge to be used in the Response Spectrum Method Figure 4 : Transfer of Forces from Superstructure to Substructure. Figure 5 : Hypothetical Enveloping Cyclinders to Estimate Hydrodynamic Forces on Substructures. Figure 6 : Hydrodynamic Pressure Distribution on the Substructure due to Steam Flow. Figure 7 : Minimum Width of Seating of Spans on Supports, Figure 8 : Potential location of plastic hinges in (a) single column substructures, and (b) double column substructures, Figure C1 : Deformed geometry of bridge deck (a) when earthquake force is normal to the direction of traffic, and (b) when earthquake force is along the direction of traffic. The full circle shows the location of the centre of mass of the bridge system before application of the force F, and the hollow circle that after the application of the force F Figure C2: CALTRANS adjustment for ductility and risk assessment factor Z. Figure C3 : Acceleration response spectrum (a) actual (but smoothened), and (b) idealised for design purposes.Draft for IRC:6 Provisions on Seismic Design of Bridges :: Code & Comment page 39 of 44 cancrox, Figure 1 : Seismic zone map of India.Draft for IRC:6 Provisions on Seismic Design of Bridges :: Code & Commentary page 40 of J4 2 Soil Type I a 1 Gus a 1 0 1 2 3 4 5 Natural Period T (sec) Figure 2 : Plot of CS versus natural period T to be used in the Seismic Coefficient Method For modes other than 0.5 findamental mode only o+— a 0 1 2 3 4 5 Natural Period T (sec) Figure 3 : Plot of CaS versus natural period 7 in mode k of the bridge to be used in the Response Spectrum MethodDraft for IRC:6 Provisions on Seismic Design of Bridges :: Code & Commentary page 41 of 44 L Direction of Seismic Force Figure 5 : Hypothetical Enveloping Cyclinders to Estimate Hydrodynamic Forces on Substructures | GF rr (Resultant pressure a on CH) CH +“ a py = 1.2F/H Figure 6 : Hydrodynamic Pressure Distribution on the Substructure due to Steam Flow:Draft for IRC:6 Provisions on Seismic Design of Bridges :: Code & Commentary page 42 of 44 Wi ee VW) (6) Column or Pier wo (6) Suspended Span on Restrained Portion of Superstructure Figure 7: Minimum Width of Seating of Spans on Supportsraft for IRC:6 Provisions on Seismic of Bridges Code & Commenter ye $3 of t4 nage B00 care | Seame Servers Sey a7 Chemie jaan Potennat plastic Ea innge repens ie “al Sect a @) ange anase, aa" eer — at Porennat plastic Conmns hinge regions 4 (b) Secon 4 Figure 8 : Potential location of plastic hinges in (a) single column substructures, and (b) double column substructures. Direction of Traffic Seismic / Force F * Seismic Force F~ Direction of Traffic (b) Figure C1 : Deformed geometry of bridge deck (a) when earthquake force is normal to the direction of traffic, and (b) when earthquake force is along the direction of traffic, The full circle shows the location of the centre of mass of the bridge system before application of the force F, and the hollow circle that after the application of the force F.Draft for IRC:6 Provisions on Seismic Design of Bridges :: Code & Commentary page £4 of 44 ieee RO qn somes 20) ‘Acjustment Factor (Z) Cte 1 Nes dra oor bara t= 08) 0 07 04 ce ch 10 20 30 Period of Structure (Sec) T Figure C2 : CALTRANS adjustment for ductility and risk assessment factor Z. 25 PGA 2.5 PGA PGA : Natural PeriodT ° Natural Period T (a) (b) Figure C3 : Acceleration response spectrum (a) actual (but smoothened), and (b) idealised for design purposes,Enclosure A1. Introduction The behaviour of bridge structures under strong seismic conditions is very different from that under normal gravity loads (dead loads and moving live loads). Further, the seismic design of bridges draws great significance since bridges come under the category of “life-line structures.” The implications of disruption in the transportation network due to distress or collapse of bridge systems, particularly during the immediate post-earthquake relief and rehabilitation operations, demand that special attention be paid in their design and construction. The seriousness of the matter has been recognized by many countries over the last two and a half decades. Major initiatives, in the form of comprehensive revision of design codes, stricter quality control at construction, and an exhaustive retrofit program for existing deficient bridge structures, have been undertaken in those countries, India is, indeed, an earthquake country. More than 50% of the country lies in seismic zones III, [V, and V; while, the remaining area also cannot be considered aseismic as was very graphically illustrated by the 1993 Killari (Latur) earthquake in seismic zone I. It is therefore important to evaluate how the modem bridges being built in the country have been performing, in the past earthquakes with a view to modify the design and construction procedures appropriately. This paper reviews the damages incurred by bridges in India during past earthquakes and lists the lessons learnt from them, 2. Background on Earthquake Magnitude and Inten: When reviewing the performance of bridges in the past earthquakes it is important to have the correct perspective with regard to the strength of shaking that the bridge concerned experienced during the earthquake. In a damaging earthquake, if a major bridge is located in the area which did not sustain strong shaking, then no-damage performance of the bridge is of no consolation as far as review of seismic performance of bridges is concerned, It is therefore important that when reviewing the performance of a bridge during the earthquake, the intensity of shaking in the area be kept in perspective. Considering the prevalent confusion onaccount of intensity and magnitude. in this section background is provided on earthquake magnitude and intensity. Earthquake magnitude is simply a measure of the size of the earthquake reflecting the clastic energy released by the earthquake. It is usually referred by a certain real number on the Richter scale (e.g., magnitude 6.5 earthquake). On the other hand, earthquake intensity indicates the extent of shaking experienced at a given location due to a particular earthquake. It is usually referred by a Roman numeral on the Modified Mercalli Intensity (MMI) scale. more recently on the MSK scale (e.g., intensity VIII on MMI scale). To draw a parallel. consider a /00W light bulb used for illumination purposes. It dissipates energy of 100 Watts, say, However, the brightness of light at different distances from it, referred in candielights, is different. The former is magnitude and the later is intensity. Whether or not there is adequate light at a given location to enable one to read a book, depends on what is the intensity of light at that location; intensity itself depends on the magnitude of bulb (i.e., Watts) and the distance of the bulb from the location under consideration, Similarly, intensity of shaking at a location depends not only on the magnitude of earthquake, but also on the distance of the site from the earthquake source and the geology / geography of the area, Isoseismals are the contours of equal earthquake intensity The Indian seismic code (IS:1893-1984) divides the country into five seismic zones based on expected intensity of shaking in future earthquakes. The five seismic zones I, II, II. IV and V correspond to areas that have potential for shaking intensity on the MMI scale of V or less, VI, VII, VIII, and IX or more, respectively. Insofar as benchmarking the seismic performance of bridges is concerned. it is not the magnitude of the earthquake but the intensity at the location of the bridge that must be used. For example, during an earthquake of magnitude 6.5, the maximum intensity area experiences a shaking of VIII, say, but has no bridges. The bridges of interest may be located in areas that lie in seismic zone IV but thatsustained shaking intensities of VI and VII. Then, the performance of these bridges will only answer the following question: how did the bridges designed for seismic zone IV perform when subjected to earthquake shaking similar to what is expected in seismic zones IT and III. respectively 3. Jabalpur Earthquake of 1997 ‘An earthquake of Richter magnitude 6.0 occurred on 22 May 1997 at 04:22 hours (IST) centered about 8 km southeast of the city of Jabalpur (Madhya Pradesh). The maximum intensity of shaking experienced during the earthquake was VIII on the MSK scale at the villages of Kosamghat and Kudaria in the Jabalpur district. The intensity of shaking in the city of Jabalpur varies between VII and V. The area affected by this earthquake lies in seismic zone IIT as per the Indian codes [IR:6-1966, IS:1893-1984], and hence the level of shaking was indeed expected. The strong earthquake shaking was experienced over a relatively large area covered by the districts of Jabalpur, Mandla, Sivni and Chhindwada. There are a good number of highway and railway bridges in the area. All the bridges except one major railway bridge performed very well (Jain er al, 1997]. The two lane bridge across the Narmada river at Mandla, 95 km southeast of Jabalpur, consists of two-span prestressed concrete superstructure supported on non-prismatic RC piers founded on base rock. The bridge has RC restrainers to prevent the superstructure from dislodging transversely from the elastomeric bearings from the pier top This bridge performed very well (Figure 1). The damage sustained by the buildings in the neighbourhood suggests an intensity of shaking around VI. Another, nine-span bridge across the river Gaur near Kosamghat sustained no damage. The damaged bridge was a steel through-truss bridge consisting of six simply- supported spans, across the river Narmada between Tilwara and Gwarighat in the Nagpur Division of the South Eastern Railway (Figure 2). In the roller and rocker bearings of this bridge, one 32 mm diameter pin connects the knuckle pin to the saddle cover plate (Figure 3).3 These pins were fractured at several of the supports, and the spans were dislodges transversely by about /00 mm. Fortunately, none of the spans was dislodged from the piers. The fractured pins were replaced and the bridge was restored to working condition in about two days. The weight of the superstructure span of this steel bridge is about 200 fonnes. The tensile strength of each of these 32 mm diameter mild-steel pins is about 20 ‘onnes. The transverse load of one span is resisted by four such pins. Hence, the total transverse load carrying capacity is around 80 tonnes. The pin failures indicate that, atthe top of the piers, the peak acceleration may have exceeded 0.4g. 4 lari Earthquake of 1993 The quake measuring 6.4 on the Richter scale struck at 03:53 hours (IST) on September 30, 1993, and was centred around village Killari in Latur district (Maharashtra) Gain er al, 1994]. The maximum intensity of about IX on the MMI scale was experienced in villages Killari and Talni, Figure 4 shows the isoseismals for this earthquake in the worst affected region. The entire area affected by this quake falls in the seismic zone I as per IRC code [IRC:6-1966] and BIS code [IS:1893-1984] (i.e., area having potential for shaking intensity of V or less on the MMI scale), ‘An insportant feature of this earthquake was that the affected area was very small: even less than about 20 km x 20 km sustained severe shaking. This is a common feature for earthquakes having very low focul depth. The meizoseismal area (area of strongest shaking) had no major bridges, except for one aqueduct. Only culverts or small bridges on RC. T= beams or R.C. slabs are in use in the area; most of them sustained no damage or very minor cracking at the abutments. The area also has no rail network; hence, the performance of railway bridges cannot be inferred from this earthquake. The RC aqueduct bridge for the left bank canal over a tributary to the Tima river near the village Ganjankhed suffered the most significant damage. The intensity in this area may be estimated to be around IX on the MMI scale. The aqueduct consisted of 7 spans, each of /6,5‘metres (Figure 5). The bearings over the first pier towards village Talni were found damaged. ‘The damage consisted of shearing and pulling-out of the anchor bolts connecting the steel bearings to the pier (Figure 6). Also, the bottom plates of the bearings were damaged. The post-earthquake location of the bearings indicates that the pier has moved away from the abutment by about 4 to 5 cm. Fortunately, the spans were not unseated from the bearings and dislodged on to the piers. The bearings (sliding and rocking) are welded to steel plates at top and bottom, and these plates are then bolted to the soffit of the girder and to the concrete pedestal on top of the pier. The details of these stee! plates used are shown in Figure 7(a). The configuration of the plate with reduced size and thickness at the location of bolt holes and the lack of welding on the second side between the plates of different thicknesses, demonstrate poor detailing, Figure 7(b) shows a schematic of the failure at these bearing plates due to the horizontal displacements generated during the earthquake. A reinforced concrete box-culvert (Figure 8) supporting the above canal and located between villages Ganjankhed and Talni, very close to extension of the observed fault scarp towards the west side, was itself not damaged. However, the wing walls of the embankment at the box culvert tilted outwards by about 5 to 7 cm (Figure 9). Reports from two independent surveys of bridges in the neighbouring districts of Sholapur, Osmanabad and Beed provide the following observations. A survey of 15 major bridges, in the Sholapur district, of varying age (20-125 years) and construction types (masonry arches, RCC arches, RCC T-beam, RCC solid slab) concludes that no significant distress was observed which could be attributed to the 1993 Killari earthquake [Ghosh, 1996] It may be recalled that this region was subjected to shaking intensity V and VI (on the MMI scale) during the earthquake. In another survey of bridges [Tandon, 1996], three masonry arch bridges in the Osmanabad district and one masonry arch bridge in the Beed district were found to have cracks in the arch barrels, separation of spandrels walls, and displaced stones. TwoRCC solid slab bridges in the Osmanabad district, one /28 m long (/4 spans @ about 9 m) and another 90 m long (/0 spans @ about 9 m), had no bearings provided under the spans, The expansion joints were found to have malfunctioned. In another bridge in the Osmanabad district, a 45 m long RCC T-beam and slab bridge (3 spans of about /5 m), the elastomeric bearings were found embedded in concrete. Fine cracks in webs of girders were also recorded. This bridge is located in the region which experienced shaking of intensity around VI or VIL. §, Uttarkashi Earthquake of 1991 ‘The earthquake occurred at 02:53 hours (IST) on October 20, 1991 in the Garhwal Himalayas in northern India [Jain ef al., 1992]. Strong shaking was experienced in the districts of Uttarkashi, Tehri and Chamoli of the state of Uttar Pradesh. The quake measured 6.6 on the Richter scale [GSI, 1992]. Tehri and Chamoli are in the seismic zone V as per the current code, while Uttarkashi comes under seismic zone IV. The intensity of shaking was estimated to be about VIII in Uttarkashi area, and about VII in Tehri and Chamoli. Peak ground acceleration (PGA) of about 0.3g each along the horizontal and vertical directions have been recorded by strong motions instruments at Uttarkashi [Chandrasekaran and Das, 1992] ‘The transportation system (predominantly a road network) of the area was throttled due to extensive damage to roads, slopes, retaining walls and bridges. The area has a number of single-span RC T-beam bridges of up to 20 meter span, and steel truss bridges of larger span, In a few bridges, cracks appeared between abutments and wing walls. There is no rail network in the affected area and hence the performance of railway bridges cannot be inferred from this earthquake, The rather important Uttarkashi-Harsil-Nelong road link was paralyzed for several days due to a large number of landslides and collapse of a major bridge. The Gawana bridge, located 6 km north of Uttarkashi towards Maneri, was a 56 m span steel truss bridge built in 1974. Under the shaking of intensity VIII or EX experienced in the area, the entire bridge span was unseated from its abutments and fell into the river below (Figure 10); road access to areasnorth of Uttarkashi was cut-off from the rest of the country. No permanent displacement or damage to the bridge abutments was noticed, The superstructure was supported on two rocker bearings at one end and two roller-cum-rocker bearings at the other. The anchor bolts of one of the roller-cum-rocker bearings were found. sheared-off and the base plate was found rotated from its original position (Figure 11). However, the anchor bolts at the other bearings were found to be in place and intact. The top and bottom plates of the bearings were separated; the top plates were stil attached to the superstructure. The evidence at site suggests that inadequate design of bearings and anchor bolts, as well as lack of positive ‘mechanisms to hold the dislodged spans from falling off the abutments seats, were responsible for this failure The Uttarkashi-Lumgaon road link was lost due to collapse of embankments on the approach road to the bridge near Kishanpur. The approach to the bridge is on a 8 meter high embankment with retaining walls in “banded” stone masonry. Under the intensity VIII shaking experienced in the area, these walls on both sides collapsed resulting in the failure of the embankment, The reduced road width enabled the transit of pedestrian traffic only, vehicular traffic was suspended for over 10 days ‘The area has a number of pedestrian suspension bridges pro ing access to villages located across the river Bhagirathi, The main tower and anchor blocks of these bridges are in unreinforced stone masonry. Table | shows the list of damages sustained in some of them. The failure of the anchor block of the suspension bridge near Harsil (about 73 Am north of Uttarkashi) may be attributed to poor quality of construction rather than the intensity of shaking (Figure 12); the intensity of shaking in this region was only about VI on the MMI scale. ar-Nepal Earthquake of 1988 On August 21, 1988, a quake measuring 6,6 on the Richter scale rocked the states of Bihar, West Bengal and Sikkim in India and the neighbouring country of Nepal in the morningat 04:39 hours (IST). The quake centred in Nepal, but close to the Indian, caused significant damage to life and property in both the countries (Jain et af., 1991], It occurred in the same region which was visited by the great earthquakes in the past (Bihar-Nepal earthquakes of 1833 with M 7.0 to 7.5 and of 1934 with M 8.4). The isoseismal map of the region shows ‘maximum intensities upto IX in northern Bihar and upto VIII in Sikkim (Figure 13). In the state of Bihar, damage to road bridges was generally reported from all areas but specific information on bridge performance is not available area wise (e.g, Thakkar et al. 19), There were many small bridges of simple spanned steel beam supported on piers of steel pipes. Such bridges seemed to have behaved well. The 5.6 km long bridge over the Ganges river near Patna (shaking intensity of VI, constructed for seismic zone IV which corresponds to intensity VIII) did not show any apparent damage. The damage to oridges along the rural roads were in the form of cracks in the piers, abutments, wing walls, and RC deck slab. At some places, subsidence in the embankment caused damage to culverts. In Sikkim, two bridges which experienced intensity VII shaking were affected due so failure of hill slopes. The problem was compounded by incessant rains before and after the earthquake, making the hill slopes more prone to damage [Jain and Tripathi, 1989]. The then newly constructed 67 m span prestressed concrete Legship Bazar bridge (Figure 14) on river Kale| Khola near Legship Bazar town (about 93 kn from Gangtok) was damaged due to landslides, The bill slope on the Legship Bazar end failed and soil including huge boulders fell on the bridge deck smashing a portion of the deck slab near the abutment (Figure 15). The bridge however was opened for traffic after clearing the debris. There was no apparent gamage to piers and abutments, About 3 km upstream from the Legship Bazar on the Legship- ‘ishiding road, the 135 m span Tashiding suspension bridge across river Kalej Khola wes under construction; the towers and cable anchors were completed by the time of the eatthquake. The RC abutment-cum-suspension tower on the Legship end was situated just by0 the side of a high hill-cliff, During the quake, a huge soil-cum-boulder mass from the cliff came down due to slope failure and washed away the whole construction at this end, The other end was however intact. An old suspension bridge with wooden decking situated very near to the Tashiding bridge, was not damaged The worst-affected area of the earthquake does not have major broad-gauge railway line, but meter and narrow gauge lines. A number of railway bridges were damaged: information could be obtained on the performance of railway bridges in the Samastipur sion of N.E.Railway [NERly, 1989]. Eight of these bridges were restored to normal working condition after minor repair work. The most significant damage was to the five-span (two of 12.2 m and three of 30.5 m) bri between Samastipur and Muktapur in the ‘Samastipur-Darbhanga section. A shaking intensity around VII or VIII was experienced in this area, The rocker bearings on pier 1 and 3 broke and fell down. Seven numbers of these damaged saddles of the rocker bearings had to be replaced. Another bridge of seven spans (four of 12.2 m and three of 30.5 m) in the same area between Hayaghat and Thalwara experienced a longitudinal movement of the girder between piers 4 and 5 by about 4 cm. The girder had to be pulled back to its original position. A fifteen-span (each of /2.2 m) bridge in the Manasi-Saharsa section was also seriously affected. The girder of the fourth span shifted laterally by 2 cm, Eight piers developed horizontal cracks. The wing walls of a bridge, between Naya Nagar and Ruseraghat stations, bulged out, Piers of a three-span (each of 6.1 m) steel girder bridge between Janakpur Road and Bajpatti stations, developed cracks, A six-span (each of 6.1 m) stee! girder bridge between Kamtaul and Jogiara stations, received severe damages. The bed blocks below the girders cracked and the masonry piers and abutment of the bridge developed wide cracks. The wing walls and the return walls of a three-span (each of 6.1 ‘m) steel girder bridge between Kakarghatti and Trasarai stations, developed cracks. Some portions of the wall masonry cracked and fell down causing the abutment masonry to bulgeu ‘The parapet wall of approaches of a ten-span (each of 6.1 m) bridge between Manigachi and Lohna Road stations, cracked and collapsed. Another seven-span (each of /2.2 m) bridge between Dhamaraghat and Koparia stations received damages. The existing cracks in the bed biock of two piers widened. One pier of a three-span (each of 12.2 m) bridge in Jhanjharpur- Laukaha Bazar section tilted and was found to be out of plumb by 5 ci, immediately after the earthquake. Beyond the above twelve bridges, approaches of two other major bridges located in intensity VIII area (in the Forbesganj-Saharsa and Saharsa-Pumnia sections) subsided and caved-in leading to distortion of the tracks 7. Burma-India Border Earthquake of 1988 ‘A strong earthquake of magnitude 6.8 (some sources attribute the magnitude as 7.3) rocked the north-eastern states on August 6, 1988 at about 6 AM. The quake was centred close to the Indian border in Burma. Some damages were reported in India [Kumar, 1992] where maximum intensities of upto VIL were experienced in Imphal, Kohima, Jorhat and Tejpur. The stations at Berlongfer and Diphu recorded the highest peak ground accelerations of about 0.34g (Chandrasekaran and Das, 1990]. Damage survey also indicated maximum intensity in this area. ‘Three bridges, namely the Kaliabhomora bridge across river Brahmaputra near Tejpur {Prajapati, 1990], the single-span Metong bridge on National Highway 37 between Jorhat ané Sibsagar, and the prestressed concrete box girder bridge on river Barak near Silchar [Prajapati. 1989], are reported to have been damaged during this earthquake. Information on the damage to the Metong bridge is incomplete. ‘The Kaliabhomora (Tejpur) bridge is 30/5 m long, with 24 typical spans of 120 m each and two shore spans of 67.5 m each. It is 7.5 m wide with 1.5 m footpaths on either side. (Figure 16). The typical span consists of two balanced cantilevers 52.5 m each from the supports with a central suspended span of 75 m. The suspended span is a girder-slab system in RCC, while the cantilever span is a segmental tapered box girder system in prestressedb concrete, The north end of the suspended spans are fixed to the cantilever span, and the south end are let free for longitudinal translation, The bridge was about one year old at the time of the earthquake. Due to the earthquake, pier 3 from the north end was observed to have tilted by about 6-8 cm [Prajapati, 1990; Sharma, 1990], It is suspected that differential settlements under the pier may have been responsible for the tilt. The elastomeric bearing pads at the fixed end were displaced and deformed, Cracks were also recorded on the free surface of these pads. About /23 anchor bolts holding the steel plates on either side of the bearings are recorded to have been bent and pulled out. This is attributed to the fouling of the projected portions of the bolts with the movement of the steel plates. Over 26 concrete rail posts at the ends of the suspended spans were found to have been damaged due to the relative longitudinal ‘movement along the bridge spans. The footpath portion adjoining these rail posts were either cracked or broken. The prestressed concrete box girder bridge across river Barak at Sadarghat, Silchar. was about 20 years old at the time of the earthquake. The 234.7 m long 7.5 m wide bridge has two central piers and two end trestles (Figure 17), with a hinge at the centre of the bridge. Cracks were noted at the trestle supports on the top and bottom faces of both the lower and upper flanges of the box girder [Prajapati, 1989] Dharmsala Earthquake of 1986 The hill town of Dharmsala, Himachal Pradesh, was rocked by a quake (M 5.7) at 13:05 hours (IST) on April 26, 1986. The earthquake drew considerable attention because of the proximity of Dharmsala to Kangra, the seat of the great Indian earthquake in 1905 (M 8). “The quake caused a maximum intensity of VI on the MMI scale in the area. The affected area classifies under seismic zone V. Even though significant damage is reported on buildings, no mention of performance of bridge structures is available [Arya er al., 1986] 9, Cachar Earthquake of 19: The northeast region of India was shaken by a magnitude 5.6 tremour at 23:33 hours13 ‘on December 31, 1984. The epicentre of this shallow focus earthquake was centred around the banks of the Sonai river 2 Am east of Baramuni village in Assam. The affected area falls in seismic zone V of the country. Two, somewhat similar, well-designed and well-constructed RC bridges were within about /0 km focal distance. These bridges, namely the Sonaimukh bridge and Rukni bridge. suffered different levels of repairable damage (Agarwal, 1986]. Unfortunately, information on the imensity of shaking in these areas is not available. The reinforced concrete Sonaimukh bridge is /00 m long 4.8 m wide supported on six 8.5 mto 12 m high piers, which rested on well foundations of 12.8 m to /6.3 m height (Figure 18). The bridge consists of three identical supported spans of about 18 m and two suspended spans of about 9 m, The supported spans resting on the piers had rocker and roller bearings The damage incurred by the bridge included the following, One well cap developed cracks and was relatively transversely displaced with respect to the well. The bridge deck was no longer straight in plan after the shaking (Figure 18). The final displaced locations of the five pieces of the bridge deck suggests that there was a relative transverse displacement (right lateral) between the two abutments. Some superstructure spans were dislodged from the bearings resulting in upto 45 cm relative horizontal relative displacement between the superstructure and the pier. The dislodged spans were precariously held over the piers/supported spans. The clear width of seating of the spans at the bearings was insufficient, A slightly increased shakin: intensity or duration may have manifested into the collapse of, at least, a few spans. The Rukni bridge has an identical design as the Sonaimukh bridge, but with a reduced length of 65 meters. The bridge is composed of two fixed spans on four piers and one central suspended span. The abutments had RC retaining walls to support the embanked soil mass, The damage was in the form of pounding of the longitudinal superstructure span against the retai 12 wall at the abutments on the north end of the bridge, and separation at the south end.4 Further, at the intermediate junctions between the fixed and suspended spans. the pounding effect was observed, Residual longitudinal displacements upto 5 cm have been measured along the direction of the bridge at these locations. 10. Implications of Past Performances on Seismic Design of Bridges The seismic shaking intensity experienced by the various bridges discussed above can be only termed moderate, And even under these moderate amounts of shaking, the performance in the bridges is scattered over a wide range, from minor cracks to complete collapses (Table 1). A number of lessons are obvious from these performances which have significance insofar as seismic design of bridges is concemed. The following sub-sections present a summary of the same 10.1 Level of Desi; The Gawana bridge episode during the 1991 Uttarkashi earthquake is a landmark in the seismic history of bridges in India. The peak ground accelerations of about 0.3g recorded in the region, that is classified to be in seismic zone IV. is in contrast with the design seismic coefficient of 0.05g to 0.075g specified for bridges in zone IV. Considering the typical shape of response spectrum with 5% damping (Figure 19) implies a response reduction factor of about /0 to 15. This is too large for bridges where the beneficial effects of ductility and redundancy are small. The experiences from the 1971 San Fernando earthquake in U.S.A immediately lead to a sharp increase in the design seismic force for bridges in the United States of America, However, the level of design seismic force for bridges is too low in the Indian code, 10.2 Inadequate Connections Between Superstructures and Substructures The failures of the railway bridge bearings during the 1988 Bihar-Nepal earthquake, the Gawana bridge bearings during the 1991 Uttarkashi earthquake, and the aqueduct bridge bearing during the 1993 Killari earthquake are clear indicators of the inadequate design provisions in the Indian seismic code for bridges vis-d-vis the bearings, A giance at Table 1 clearly shows the poor performance of bearings in almost all the past earthquakes. This is amajor problem that needs to be tackled for Indian bridges. The issue is explained in the following. Different components in a bridge have different capacities to undergo inelastic deformations. The ductility and redundancy in structures is usually relied upon in reducing the design seismic force from the maximum elastic forces. Amongst the bridge components (superstructure, substructure, foundation and connection between the superstructure and substructure), the connections have little or no capacity to undergo inelastic deformations. In earthquake-resistant design of the bridge superstructure and substructure, usually the design force is only a fraction of the maximum elastic forces; response reduction factors of 3 or more are normally used. However, in the design of bearings which transfer earthquake forces, necessarily the full maximum elastic force, or even more, has to be considered. This is essential because bearings are determinate elements with no or marginal reserve strength beyond yield. In fact, international practice is to design the connection elements (bearings), with a response reduction factor of about 0.3, i.¢., for 25% more than maximum elastic forces. 10.3 Deformation Design at Connection Regions The damages to the railway bridge bearings in the Samastipur Division of N.E Railway during the 1988 Bihar-Nepal earthquake, the Kaliabhomora bridge bearing damage at Tejpur during the 1988 Burma-India Border Region earthquake, and the Rukni and Sonaimukh bridge bearing damage during the 1984 Cachar earthquake, are but some examples of the need for controlling the displacements along and transverse to bridge spans. The connections between the superstructure and substructure and berween the adjacent sections of superstructure (¢.g., suspended span and supporting span elements), undergo relative motions during seismic shaking. These connections must be so detailed that they can indeed allow the required deformations without themselves being damaged. There are three major aspects to this Longitudinal and lateral translational movements of the superstructure facilitated by the16 connection elements © Stability of superstructure against uplifting and possible overturning * Adequate widths atop the substructure or supporting span to permit large movements of superstructure components without dropping-off from the supports 10.4 Substructure and Foundation Design Bridge substructures of the old bridges in India are usually of masonry type. Lately there is trend towards the use of reinforced concrete substructures and foundations. The aqueduct bridge pier during the 1993 Killari earthquake and the Kaliabhomora bridge pier during the 1988 Burma-India Border region earthquake are two examples of tilting of foundations, Since the damages in foundations usually remain undetected, it is not easy 10 evaluate the performance of foundations during the past earthquakes. Further, there are numerous bridge pier, abutment and wing-wall damages reported during earthquakes discussed above. In all, there is a clear need for more scientific design and detailing of foundations and substructures, 11. Concluding Remarks ‘Most of the recent earthquakes in India have been of moderate size and have caused only moderate intensity of shaking. Moreover, the meizoseismal areas (worst-affected areas) in most of these earthquakes did not have many major bridges. Nevertheless, damage to bridges in these earthquakes provides very clear pointers to what one would expect considering that the Indian bridge codes are not in line with the current seismic design philosophy. Our country has clear potential of great earthquakes (of magnitude larger than 8.0) causing significantly higher level of shaking intensity (upto XII on MMI scale). Figures 20 shows the performance of railway lines during the Great Assam earthquake of 1897; this is a grim reminder of what can be expected of our major road and rail lines in the event of a strong earthquake taking place at just the right location (i., wrong location for the society). 12. Acknowledgements The work reported here was carried out with financial support from the Ministry of7 Surface Transport, Government of India; the authors are grateful for this support. The authors also sincerely thank Mr. Mahesh C, Tandon and Mr. A. R. Ghosh for generously sharing information on damage to bridges in the Latur earthquake 13. References Agarwal,P.N., 1986, “Damage to Two RCC Bridges During December 31, 1984 Cachar Earthquake, Northeast India,” Bulletin of the Indian Society of Earthquake Technology, Vol.23, No.1, March 1986, pp 1-16. AryaAS, Gupta,S.P, Lavania,B.V.K, and Kumar,A., 1986, Report on Dharmasala, Himachal Pradesh, Earthquake April 26, 1986, Department of Earthquake Engineering, University of Roorkee, Roorkee, July 1986. Chandrasekaran, A.R., and Das,J.D., 1990, “Strong Motion Arrays in India: Characteristics of Recent Recorded Events.” Bulletin of the Indian Society of Earthquake Technology, Vol.27, No.1, pp 1-66. Chandrasekaran,AR., and Das,J.D., 1992, “Analysis of Strong Motion Accelerograms of Uttarkashi Earthquake of October 20, 1991,” Bulletin of the Indian Society of Earthquake Technology, Vol.29, No.1, pp 35-55 Ghosh,A.R., 1996, personal communication GSI, 1992, Usarkashi Earthquake October 20, 1991, Special Publication No.30, Geological Survey of India, Calcutta GSI, 1993, Bihar-Nepal Earthquake August 20, 1988, Special Publication No.31, Geological ‘Survey of India, Calcutta GSI, 1995, A collection of papers presented at the Workshop on the 30th September, 1993 Killari Earthquake, Maharashtra, 24 December 1993, N.G.R.I., Hyderabad, Special Publication No.27, Geological Survey of India, Calcutta IRC:6-1966, Standard Specifications and Code of Practice for Road Bridges, Section II: Loads and Stresses, The Indian Roads Congress, New Deihi, 1985. IS:1893-1984, 1984, Indian Stemdard Criteria for Earthquake Resistant Deaign of Structures, Bureau of Indian Standards, New Delhi Jain,S.K., TripathiRP., and Agarwal,A.K., 1991, “Geotechnical Damage Due to Bihar Earthquake of August 1988,” Proceedings of Second International Conference on Recent Advances in Geotechnical Earthquake Engineering and Soil Dynamics, March 11-15, 1991, St. Louis, Missouri, U.S.A. pp 519-524 Jain,S.K., and Tripathi, P.. 1989, “Damage to Roads and Bridges in Sikkim and West Bengal,” Proceedings of Workshop on Bihar-Nepal Earthquake, December 28-29, 1988, LLT-Kanpur, Kanpur. Jain,S.K., Singh,R.P., Gupta.V.K., and Nagar,A., 1992, “Garhwal Earthquake of October 20, 1991,” EERI Special Earthquake Report, EERI Newsletter, Vol.26, No.2, February 1992, pp 1-4. Jain,S.K,, Murty,C.V.R., ChandakN., Seeber,L., and Jain.N.K., 1994, “The September 29, 1993, M6.4 Killari, Maharashtra, Earthquake in Central India,” EERI Special Earthquake Report, EERI Newsletter. Vol.28, No.1, pp 1-8. Jain,SK., Murty,C.V.R., Arlekar,JN., JainC.K., SinhaR., and Goyal,A, 1997, “Some Observations on Engineering Aspects of the Jabalpur Earthquake of 22 May 1997,” EERT Special Earthquake Report, EERI Newsletter, Vol.31, No.8, August 1997, pp 1-8 Kumar,B., 1992, “Isoseismals of Burma-India Border Region Earthquake of August 6, 1988,” Bulletin of the Indian Society of Earthquake Technology, Vol.29, No.1, pp 57-67 NE. Railway, 1989, personal communication. Oldham,R.D., 1899, Report on the Great Earthquake of 12th June 1897, Memoirs of theRv-IRe AS 4g Geological Survey of India, Vol.XXIX, Calcutta. Prajapati,G.L, 1989, “Damage to Barak River Bridge at Sadarghat, Silchar, Assam Due to Burma-India Border Region Earthquake of August 6, 1988,” Proceedings of National Seminar on Concrete Admixtures and Repair Material, December 28-29, 1989, Department of Civil Engineering, University of Roorkee, Roorkee, pp 48-57. Prajapati,G.., 1990, “Performance of Kaliabhomora Bridge on River Brahmaputra During Burma-India Border Region Earthquake of August 6, 1988,” Bulletin of the Indian Society of Earthquake Technology, Vol.27, No.3, September, pp 151-159, Sharma,V.C., 1990, “Tejpur Bridge Across River Brahmaputra : Special Features of Superstructure, Earthquake Effects and Present Thinking,” Proceedings of the National Seminar on Bridge Superstructure, Bombay, Vol.1, pp 121-134. Tandon,M.C., 1996, personal communication. Thakkar,S.K., Paul,D.K.. Mukerjee.S., Bandopadhyay,S., Kumar,A., and Lavania, B.V.K.. 192, Damage Survey Report on Bihar-Nepal Earthquake of August 21, 1988. Department of Earthquake Engineering, University of Roorkee, Roorkee.Table 1 Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12 Figure 13, Figure 14 Figure 15 Figure 16 Figure 17 Figure 18 Figure 19 Figure 20 19 List of Tables Summary of Damages to Bridges in India During Past Earthquakes List of Figures View of the RCC bridge across river Narmada at Mandla showing the RC seismic restrainers at the piers. View of the steel through-truss bridge across river Narmada between Tilwara and Gwarighat in the Nagpur Division of the S.E,Railway, which sustained damage View of (a) the rocker, and (b) the roller bearings, showing the knuckle pin and the saddle cover plate fastened together with a 32 mm mild steel pin. Isoseismals for 1993 Killari earthquake in the worst-affected region [GSI, 1995]. Aqueduct near Ganjankhed along left bank canal of Tima river over a tributary. Damage to the steel bearings at pier 1 of the aqueduct (a) pulling of the anchor bolts, and (b) shearing-off of the anchor bolts Details of the steel bearing plate used at the supports in the aqueduct, Box culvert supporting the earthen embankment for the canal, Tilting of the wing wall at the box culvert during 1993 Killari earthquake View of the collapsed Gawana bridge during 1991 Uttarkashi earthquake. View of the displaced bearing plates at the supports of the collapsed Gawana bridge during 1991 Uttarkashi earthquake ‘View of the severely damaged anchor blocks of the suspension bridge near Harsil during 1991 Uttarkashi earthquake. Isoseismal map of the 1988 Bihar-Nepal earthquake [GSI, 1993] View of the Legship Bazaar bridge in Sikkim after the 1988 Bilar-Nepal earthquake [Jain and Tripathi, 1991} View of the damaged deck of the Legship Bazaar bridge in Sikkim during the 1988 Bihar-Nepal earthquake (Jain and Tripathi, 1991] Details of typical cantilever span of Kaliabhomora (Tejpur) bridge across Brahmaputra river [Prajapati, 1990] Line diagram showing elevation of the prestressed concrete box girder bridge on the river Barak at Sadarghat, Silchar (Prajapati, 1989] Details of the Sonaimukh bridge damage during 1984 Cachar earthquake [Agarwal, 1986] Current Design Spectrum vis-d-vis Recorded Motion in Uttarkashi Earthquake (Seismic Zone IV) View of the distorted railway lines at a bridge severely damaged during the Great Assam earthquake of 1897 (Oldham, 1899]20 Table 1 :: Summary of Damages to Bridges in India During Past Earthquakes Earthquake Earthquake Comments Information 1997 Jabalpur | Magnitude 6.0 __| Good number of highway and railway bridge: Max Intensity VIL (ina very small area) the area perfomed well *One major railway steel through-truss bridge between Tilwara and Gwarighat- on S.E.Railway sustained damage to its bearings = Nine simply supported spans on rocker and roller bearings, with one 32 mm diameter pin connecting the knuckle pin and saddle cover | plate at each bearing i = 32 mm diameter pins fractured at many supports | ~ spans transversely displaced by about 100 mm | ~ pins replaced and bridge restored in 2 days {1993 Killari Magnitude 64 Max Intensity IX * Very small meizoseismal area ‘o rail network in affected area *Bearing failure and movement of pier in RCC | aqueduct bridge No other major bridge in meizoseismal area 1991 Uttarkashi Magnitude 6.6 Max Intensity IX feizoseismal area in hilly terrain No rail network in affected area Collapse of 56m span steel truss bridge at Gawana - Failure of bearings ‘Severe damage to anchor blocks of suspension bridges 1988 Bihar-Nepal Magnitude 6.6 Max Intensity IX ‘* Damage to 12 railway bridges - Failure of bearings ~ Excessive transverse movements at supports « Damages to bridges in Sikkim due to landslides \ ~ Cracks in abutments / wing walls | i 1988 Burma-India | Magnitude 6.8 | ¢ Movement of pier and bearing failure in Tejpur Border Region | Max.Intensity VII_| bridge across Brahmaputra river 1986 Dharmsala [Magnitude 5.7 | *No information available on seismic Max.Intensity VII_| performance of bridges 1984 Cachar Magnitude 5.6 | «Two RC girder-slab bridges damaged Max Intensity ?? 1 - Failure of bearings | - Excessive relative displacement along and | transverse to span direction between span and piers, and between suspended and supported spansFigure 1 » View of the RCC bridge across river Narmada at Mandla showing the RC seismic restrainers at the piers. Figure 2 : View of the steel through-truss bridge across river Narmada between Tilwara and Gwarighat in the Nagpur Division of the S.E.Railway, which sustained damage.Figure 3 » View of (a) the rocker, and (b) the roller bearings, showing the knuckle pin and the saddle cover plate fastened together with a 32 mm mild steel pinYt 7 Jawalga © Lamj jan L ~S \ 7 Nilanga os 7 Ligbaladau No Kumpti leuri ° Kanegaon 7 Nendurga Ba \ sano cael Ganjankhe : oSarni 7 vagal ~ ae aki es SSE awtra / : Karasgaon a a 18 S “Rajegaon Mi Sifter TE 528 inch HE, wiser sari / Chincholi Kate 0 x Narangwadi / Hasturi | tausigarh 4 Hull 2 8 HZ Chakur ° vm NS — Sanavinets o/s ‘ Salegaon, —— SSmygrahal Bet Jawalga / oy, Ze .| oramba 4 17 o-“ Hunsat . Kaddara 58 rr Karl —“o Ekarga \ ° © Kaldeonimbala o | ° —~ Belsur \ 2 = \N oa u “w--— ors? @ MAIN stock L 1 76°30" 76°45" Figure 4 : Isoseismals for the 1993 Killari earthquake in the worst affected region [GST, 1995]2 TO GANJANKHED TO TALNI BEARING ou 16.5m Typical, _— 1 i 7 CANAL v — i PER EMBANKMENT seen Auovenen f fovewent |} i eee TBUTARY] PS agutment Vth WALL AQUEDUCT- BRIDGE ALONG LEFT BANK CANAL OF TIRNA RIVER OVER A TRIBUTARY 7 SAT i OUTSIDE DIAPHRAGMS AT BEARINGS ton PEDESTAL PIER ROCKING BEARING SLIDING BEARING Figure 5 » Aqueduct near Ganjankhed along left bank canal of Tima river over a tributary.Figure 6 Damage to steel bearings at pier 1 of the aqueduct (a) pulling of the anchor bolts, and (b) shearing-off of the anchor bolts.412 bolt hole fillet ay : = wt] ‘ no welding ' Elevation 4 “735 25 4 150 25 “P25 25 200 25 Plan (Note: All dimensions are in mm and approximate.) Figure 7 : Details of the steel bearing plate used at the supports in the aqueduct.27 CANAL y) o 7 7 EARTHEN EMBANKMENT —- | cuLveRT 1 A 4 EARTHEN EMBANKMENT Tommie 2S 40mm CULVERT =e 4 =:60mm wwing A) waue wine wat EARTHEN EMBANKMENT CANAL EARTHEN EMBANKMENT suaZV wine CULVERT ane TRANSVERSE Figure 8 ; Box culvert supporting the earthen embankment for the canal MOVEMENT.Figure 9; Tilting of the wing wall at the box cuivert during 1993 Killari earthquake.Figure 11 : View of the displaced bearing plates at the supports of the collapsed Gawana bridge during 1991 Uttarkashi earthquake. »30 Figure 12. View of the severely damaged anchor blocks of the suspension bridge near Harsil during 1991 Urtarkashi earthquake.31 Figure 13 : Isoseismal map of the 1988 Bihar-Nepal earthquake (GSI, 1993]Figure 14: View of the Legship Bazaar bridge in Sikkim after the 1988 Bihar-Nepal earthquake [Jain ef al, 1991]. Figure 15: View of the damaged deck of the Legship Bazaar bridge in Sikkim during the 1988 Bihar-Nepal earthquake (Jain and Tripathi, 1991]Howse moan weve ie Sm Bie wr, oS RUTuat cam an pen capa in sense T ee PIER SHAFT TS 1 WEL STSO SPAN ‘SAND FILL MIXED. rec renner | i maomsCl]~ wine orTU en ae | wou car 23k hd Bu, ae | ™ 5 eu. sremme | not waa bee Waren PILL sey scour EL 3.03m 4 230 sano mu, =< read Lame WELL curs _1}: secre a 5.00m vorge =e Figure 16 : Details of typical cantilever span of Kaliabhomora (Tejpur) bridge across Brahmaputra river [Prajapati, 1990], CENTRAL HINGE ane ae mene, suman exo - “s Figure 17: Line diagram showing elevation of the prestressed concrete box girder bridge on the river Barak at Sadarghat, Silchar [Prajapati, 1989].3 fi timacr arioce rn = sae reoanants te PLAN View fe PSS p fl q 1o ani mot 10 SCALE ca) SECTION RLL oMensions ps] Figure 18: Details of the Sonaimukh bridge damage during 1984 Cachar earthquake [Agarwal, 1986] Typical 5% Spectrum of /~ Recorded Ground Motion Design Spectrum for — Seismic Zone IV ‘Natural Period T (sec) Figure 19: Current Design Spectrum for Seismic Zone IV Vis-a-vis Typical 5% Spectrum of Recorded Ground Motion.38 Tsrchng Survey Inia Own Carona one Figure 20: View of the distorted railway lines at a bridge severely damaged during the Great ‘Assam earthquake of 1897 [Oldham, 1899].Enclosure BA state-of-the-art review on seismic design of bridges - Part! : Historical development and AASHTO code Sudhir K. Jain and C.V.R. Murty ‘The basic philosophy of seismic design is the same for all structures. However, there are certain significant and neces- sary differences in the design calculations for bridges as against those for buildings. For instance, the American codes employ different response reduction factors for different bridge components. The distinctly different calculations have arisen after evaluating the performance of the bridges during past earthquakes in the USA, Japan and other countries, In this paper, the historical development of the American seismic code provisions for design of bridges is reviewed to highlight the departure from the method of calculations usually adopted {for buildings and the origin of the special calculations for bridges. Further, the paper shows how the seismic perform ance observed in the bridges in USA has been translated into code provisions in the AASHTO code of USA. Rather poor performance of bridges inthe 1971 San Femando earthquake in California, USA, and the 1978 Miyaj-Ken Oki earthquake in Japan clearly revealed that the usual seismic design procedures applicable to buildings cannot be applied to bridges. Bridges pose their own unique problems vis-a-vis seismic performance. As result, in the last twenty years the state-of-the-art of earthquake-esistant design of bridges has ‘undergone significant changes and major modifications have taken place in the bridge codes of USA, Japan and New Zea- land, (On the other hand, the provisions on the seismic design of bridges in the Indian codes'? continue to remain rather Dr Sudhir un Proteus Deparanent of Gi of ecology Kanu 218016 DrCVR Murty, Asitant Profesoe, Deparamentof Ci Engineering Indian Inst of Technology Kanpur 208016, agineerng dian nsatte simplistic and in line with whatis perhaps adequate for build- ings. This is reflected in the poor performance of bridges in India™. In this paper, the current provisions in AASHTO code’, USA are reviewed. A companion paper‘ presents a re- view of provisions in another American code, namely CALTRANS code’ together with those in the draft New Zea- land code? and the current Indian codes. Background on earthquake-resistant design Earthquake-resistant design is fundamentally very different from the design for other dynamiceffecs, such as wind loads and vehicle loads. This section reviews some of the basic is sues involved in seismic design Since the size of a future earthquake and shaking inten- sity expected at a particular site cannot be determined accu- rately, the seismic forces are difficult to quantify forthe pur- ‘poses of design. Farther, the actual forces that can be gener- ated in the structure during an earthquake are very large. and designing the structure to respond elastically against these forces makes the structure too expensive. Therefore, in the earthquake resistant design, postyield inelastic behaviour is usually relied upon to dissipate theinput seismic energy. Thus, the design earthquake force may be only a fraction of the maximum (probable) forces generated ifthe structure is to remain elastic during the earthquake. For instance, the de- sign seismic force may at times be say, 8 percent ofthe maxi- mum elastic seismic force. Thus, earthquake resistant design and construction does not aim to achieve a structure that will not get damaged in a strong earthquake having low prob- ability of occurrence; it aims to have a structure that wil per- form appropriately and without collapse inthe event of such a shaking. a Febraary 1998 * The indian Caner Journal