137 views

Uploaded by Muhammad Yasin

- Art 19 Ardelean u
- California Seismic Hazard MapReport
- Time-History Analysis on Seismic Stability of Nuclear Island Bedrock with weak Interlayer
- Reliability-Based Structural Design Optimization for Nonlinear Structures in OpenSees
- Urban Lamadrad
- Preliminary Seismic Hazard Map of Peninsular India
- Comparative Study of Typical R.C. Building using Indian Standards and Euro Standards under Seismic Forces
- NGA Models Version1 4
- 14_05-02-0067
- ATC-40
- Geohazard Earthquake
- Bsds Seismic
- 2KL17CSE10
- 01-1057.pdf
- Performance Based Design
- 211CE2238
- Zha 2008
- Earthquakes
- Tunnel Form
- Improving the Seismic Resilience of Existing Braced-Frame Office Buildings

You are on page 1of 138

Muhammad Yasin

DEPARTMENT OF CIVIL ENGINEERING BANGLADESH UNIVERSITY OF ENGINEERING AND TECHNOLOGY

May 2007

1

**TABLE OF CONTENTS 1 Introduction............................................................................................11
**

1.1 Backgrounds and Present State of the Problem:.........................................................11 1.2 Objectives with specific aims......................................................................................11 1.3 Outline of Methodology:.............................................................................................12

2 Literature Review..................................................................................14

2.1 Seismic Performance Assessment of Structures.........................................................14 2.2 Seismic Hazard Analysis ............................................................................................14 2.3 Development of Fragility Curves................................................................................29

**3 Seismic Hazard Analysis of Bangladesh..............................................50
**

3.1 Methodology................................................................................................................50 3.2 Database.......................................................................................................................51 3.3 Site Specifications:......................................................................................................53 3.4 Identification of Earthquake Sources for Analysis.....................................................53 3.5 Deterministic Seismic Hazard Analysis......................................................................55 3.6 Probabilistic Seismic Hazard Analysis.......................................................................56 3.7 Development of PGA Contour Map for Bangladesh..................................................62 3.8 Concluding Remarks...................................................................................................72

**4 Fragility Analysis of Structures............................................................73
**

4.1 Methodology................................................................................................................73 4.2 Defining Limit States from Hazard Curve..................................................................83 4.3 Fragility Analysis of a Typical Three Storey Reinforced Concrete Frame Structure 85

5 Conclusion...........................................................................................111

5.1 Seismic Hazard Analysis of Bangladesh...................................................................111 5.2 Fragility Analysis of Structures.................................................................................111

References..............................................................................................112 Appendix A............................................................................................119 Appendix B............................................................................................125 Appendix C............................................................................................126 Appendix D............................................................................................135

2

LIST OF FIGURES

Figure 1 Normalized Response Spectra Recommended for use in Building Codes (NEHRP, 1991); soil profile types are as described in NEHRP 1991.............................................................................................17 Figure 2 Four steps of a deterministic seismic hazard analysis.............20 Figure 3 Four steps of a probabilistic seismic hazard analysis..............22 Figure 4 Examples of variations of source-to-site distance for different source zone geometries. The shape of the probability distribution can be visualized by considering the relative portions of the source zone that would fall between each of a series of circles (or spheres for threedimensional problems) with equal differences in radius........................23 Figure 5 (a) Gutenberg-Richter recurrence law, showing meaning of a and b parameters; and (b) application of Gutenberg-Richter law to worldwide seismicity data. (After Esteva, 1970)....................................25 Figure 6 Typical fragility curves.............................................................31 Figure 7 Hypothetical Static Pushover curves for the two orthogonal directions of a building. Legend: BS stands for base shear (Bazzurro et al. 2004).....................................................................................................34 Figure 8 Hypothetical NSP curves for structure in the intact condition and at different levels of damage (i.e., DS2 to DS5) (Bazzurro et al. 2004)..........................................................................................................36 Figure 9 16%, 50%, and 84% fractile incremental dynamic analysis (IDA) curves for T=0.92 s oscillator with the displayed backbone, estimated using both IDA and static pushover 2 incremental dynamic analysis (SPO2IDA) software (Vamvatsikos and Cornell 2005)..........38 Figure 10 Hypothetical NSP and IDA curves for the building in the intact condition. In this case the abscissa represents the global ductility ratio, μ (namely, the roof drift divided by the roof drift at first yielding, i.e., at DS1). The ordinate R is equal to BS/BSy for the NSP curve and to Sa(T1)/Say(T1) for the IDA curve. Note: the dashed line is an example of quadrilinear fit of the pushover curve, as explained (Bazzurro et al. 2004)...............................................................................39 Figure 11 Estimate of the residual median capacity (Ša,cap)3 for damage state DS3 (after accounting for the expected, or measured, residual offset after the mainshock). As in Figure 10, the ordinate R is

3

... 2004)... (After Bazzurro et al..................... 2004).................48 Figure 18 Earth’s crustal plate (courtesy of www...........................equal to BS/BSy for the NSP curve and to Sa(TDS3)/Say(TDS3) for the IDA curve (Bazzurro et al........................................ 2004). and for collapse of the building (After Bazzurro et al..................................... occurrence period 1869 to 2000.................................42 Figure 14 Tagging of different damage states within the illustrative example.................. and red tags................52 Figure 19 Location of earthquake epicenter...................................................cap)1 at the building site is about ≈ 2..........................................55 4 ......... or measured....41 Figure 12 IDA curves for the intact structure and for the structure at different levels of damage (after accounting for the expected.......2x10-4 (Bazzurro et al..... Yasin and Ansary 2005)........................................................ (b) Average relationship for loss of ground motion capacity and rate of increase in mean annual frequency of exceedance of ground motion for coastal California sites for which the absolute value of the (log-log) slope of an average ground motion hazard curve in the surroundings of 10-3 annual frequency of exceedance is about three (After Bazzurro et al...... Yasin and Ansary 2005)...... 2004)...45 Figure 15 Median spectral acceleration capacity associated with all the structural limit states except the green tag state (After Bazzurro et al..................................... at and near Bangladesh......... 2004)................41 Figure 13 (a) Graphical representation of the recommended tagging criteria...........com).................... All the IDA’s have been de-normalized and scaled to the spectral acceleration at the same oscillatory period before including them in the same plot (Bazzurro et al.......................... 2004).......... The circles represent the global collapse residual capacity of each case........................54 Figure 21 Effect of number of division on calculated probability (after Noor.46 Figure 16 Recommended (default) values for βR.... 2004)......................................................................... green................................................accuracyingenesis........ residual offset after the mainshock and adjusting for the behavior observed in dynamic analyses of damaged structures)......47 Figure 17 Fragility curves for onset of damage..... The yellow tagging given to DS3 assumes that the P0 of exceeding the spectral capacity (Ša.... yellow..................................................52 Figure 20 Earthquake sources identified for analysis (after Noor......................

....................61 Figure 25 Seismic Hazard Curve for Chittagong city using Duggal (1989).....69 5 .................... (1993)..................... (1993) acceleration attenuation expression)...............178x – 5..................64 Figure 29 PGA (in g) Contour Map for Bangladesh Based on two Percent Probability of Exceedence in 50-Years Period (using Boore et al.........66 Figure 31 PGA (in g) Contour Map of Bangladesh Based on five Percent Probability of Exceedence in 50-Years Period (using McGuire (1978) acceleration attenuation expression)....65 Figure 30 PGA (in g) Contour Map of Bangladesh Based on ten Percent Probability of Exceedence in 50-Years Period (using McGuire (1978) acceleration attenuation expression)...... McGuire (1978) and Duggal (1989) attenuation expression.................................................................9214)....... (1993) acceleration attenuation expression).. McGuire (1978) and Boore et al......................... (1993).............67 Figure 32 PGA (in g) Contour Map of Bangladesh Based on two percent probability of exceedence in 50-years Period (using McGuire (1978) acceleration attenuation expression).......57 Figure 23 Hazard curves of Dhaka for source one (A1) using Boore et al................................................................................................Figure 22 Gutenberg-Richter b-line for source A1 ( y = -1.............................. (1993) McGuire (1978) and Duggal (1989) attenuation expression 60 Figure 24 Seismic Hazard Curves using all source data for Dhaka using Boore et al.63 Figure 28 PGA (in g) Contour Map for Bangladesh Based on five Percent Probability of Exceedence in 50-Years Period (using Boore et al. McGuire (1978) and Duggal (1989) attenuation expression............... (1993) acceleration attenuation expression)......................................................................62 Figure 27 PGA (in g) Contour Map for Bangladesh Based on ten Percent Probability of Exceedence in 50-Years Period (using Boore et al...............................................68 Figure 33 PGA (in g) Contour Map of Bangladesh Based on ten percent probability of exceedence in 50-years Period (using Duggal (1989) acceleration attenuation expression)................................ (1993) attenuation expression 61 Figure 26 Seismic Hazard Curve for Sylhet city using Boore et al..

.......... green....................................... and for collapse of the building..............75 Figure 37 Steel01 Material ........................ 1989 acceleration attenuation expression.........................91 Figure 50 Incremental dynamic analysis curves for the intact structure 92 6 ............................................. yellow....................................... elevation and column and beam sections of the typical three storey building frame...........71 Figure 36 NSP curves for the intact structure and damage states......Figure 34 PGA (in g) Contour Map of Bangladesh Based on five percent probability of exceedence in 50-years Period (using Duggal (1989) acceleration attenuation expression).......78 Figure 40 Quadrilinear approximation of DS2 NSP curve (with and without offset)........77 Figure 39 NSP curve for damage state.......................................80 Figure 42 Fragility curves for onset of damage............. 2005)................................. DS2.......................... 2005)..70 Figure 35 PGA (in g) Contour Map of Bangladesh Based on two percent probability of exceedence in 50-years Period (using Duggal (1989) acceleration attenuation expression)....................79 Figure 41 Normalized NSP curve and incremental dynamic analysis curves for the intact structure (here R is BS/BSy for NSP and Sa/Say for IDA).....88 Figure 47 Non linear static pushover curve for damage state DS2...................................................84 Figure 44 (a) Graphical representation of the proposed tagging criteria.............................................Material Parameters of Monotonic Envelope (after Mazzoni et al.................83 Figure 43 Seismic Hazard Curves for Dhaka city using Duggal (1989) attenuation expression....................86 Figure 46 Non linear static pushover curve for the intact structure.................................................................................................90 Figure 49 Quardrilinear approximation of DS2 NSP curve................................ (b) Average relationship for loss of ground motion capacity and rate of increase in mean annual frequency of exceedance of ground motion for Dhaka city (longitude 90............................................................. with and without offset .................90 Figure 48 Non linear static pushover curve for damage state DS3.. and red tags...................................37 E and latitude 23..........73 N) developed from hazard curve using Duggal............85 Figure 45 The plan...Material Parameters (after Mazzoni et al................................76 Figure 38 Concrete01 Material ....

.....................................102 Figure 59 Loss of Sa capacity at different spectral acceleration.... obtained from the NSP curves using the SPO2IDA software for the intact structures with concrete of three values .......................................................109 Figure 64 Fragility curves for different limit states of the building when system demand is peak ground acceleration........98 Figure 55 The Incremental dynamic analysis (IDA) curves....................96 Figure 54 Non linear static push over curves obtained using three percentile values of concrete compressive strength..........107 Figure 63 Fragility curves for different limit states of the building when system demand is spectral acceleration................................................... obtained from the NSP curves using the SPO2IDA software for the intact structures with three percentile values of column cover.....104 Figure 60 Non linear static push over curves obtained using three percentile values of column cover..................................................................................105 Figure 61 The Incremental dynamic analysis (IDA) curves.Figure 51 Incremental dynamic analysis curves for damage state DS293 Figure 52 IDA curves for the intact structure and for the damaged structure in damage states DS2 and DS3....106 Figure 62 Loss of Sa capacity at different spectral acceleration.........100 Figure 57 Non linear static push over curves obtained using three percentile values of column depth.................................................101 Figure 58 The Incremental dynamic analysis (IDA) curves....94 Figure 53 Loss of Sa capacity at different spectral acceleration....................................... obtained from the NSP curves using the SPO2IDA software for the intact structures with three percentile values of column depth.............99 Figure 56 Loss of Sa capacity at different spectral acceleration..................110 7 ........

...........................95 Table 11 Median spectral acceleration......................103 Table 23 Percent loss of Sa capacity at four structural damage states 103 8 ............101 Table 20 Roof drift values of four damage states...................................................................................................98 Table 15 Spectral acceleration values of four structural damage states99 Table 16 Residual dynamic capacities of damaged structures....93 Table 9 Un-normalized (median) IDA curves...................................................97 Table 14 Roof drift values of four damage states...........................................88 Table 7 Point of normalized quadriliniear NSP curves................93 Table 10 Dynamic capacity and loss of dynamic capacity of the structure at four structural damage states.........91 Table 8 Normalized (median) IDA curves obtained via the SPO2IDA spreadsheet.............54 Table 2 Calculation of design ground motion parameter.................................56 Table 3 Summary of calculation for b-line...........................100 Table 19 β1 values....................................102 Table 22 Residual dynamic capacities of damaged structures.......58 Table 5 Input parameters of the frame structure:......96 Table 12 Aleatory uncertainty (βR) for four structural limit states of the structure.. median roof drift and corresponding main shock ground motions causing the onset of four structural limit states............................................................................................................97 Table 13 Uncertainty of three major structural parameters......................................................102 Table 21 Spectral acceleration of four structural damage states..99 Table 17 Percent loss of Sa capacity at four structural damage states 100 Table 18 Spectral acceleration and roof drift causing the onset of four structural limit states.........................................57 Table 4 Magnitude range and corresponding probability.....................................LIST OF TABLES Table 1 Parameter of the earthquake sources.........................................86 Table 6 Identification of damage states...............................................................................................................................................................................................................

...................................................104 Table 26 Roof drift values of four damage states......................................................................................................................108 Table 33 Probability of attaining different limit states of the building at PGA of 0.......................................104 Table 25 β2 values........110 9 ............107 Table 31 β3 values..........................................................................................................2g........................................................................106 Table 29 Percent loss of Sa capacity at four structural damage states 107 Table 30 Spectral acceleration and roof drift causing the onset of four structural limit states......................106 Table 28 Residual dynamic capacities of damaged structures.........107 Table 32 Combination of β-values...........................Table 24 Spectral acceleration and roof drift causing the onset of four structural limit states............................................105 Table 27 Spectral acceleration of four structural damage states..................

List of Symbols 10 .

and Sylhet have population exceeding several millions. The results of this research work will provide a rational support to the engineer inspecting the facility after an earthquake on whether to permit the occupancy of a building soon after the occurrence of a damaging earthquake. fire brigade building and digester shelter is essential for earthquake digester management. The history of earthquakes. Prediction of the post-earthquake functionality of the important structures such as hospital building.2 OBJECTIVES WITH SPECIFIC AIMS Objectives of the present research study are to Develop probabilistic seismic hazard curve for Dhaka. The seismic vulnerability of building structures depends on the construction practice in a city and is related to quality of the existing building stock. 2001) and the 2003 Rangamati Earthquake (Ansary et al. Develop a methodology for the assessment of the seismic performance of structures using the developed hazard map for Bangladesh. The seismic risk of different building categories. 2001) by combining the information on seismic hazard and structural vulnerability.. 11 . 2001). has recently been estimated (Ansary. The population increase around this region is at least 50 times than the population of 1897 and cities like Dhaka. The process is a crucial step in evaluating the likelihood that the structure will not be able to provide facility after an earthquake. in Bangladesh is sufficient enough to require their careful consideration in the design of structures and facilities..1 BACKGROUNDS AND PRESENT STATE OF THE PROBLEM: The assessment of the seismic performance of existing structures means prediction of the post-earthquake functionality of structures. engineered and non-engineered. it is very important to identify the vulnerable buildings and estimate the level of damage under a probable earthquake. 1. 2001. The 1999 Moheskhali Earthquake (Ansary et al. Chittagong and Sylhet city from probabilistic seismic hazard analysis. Retrofitting of these existing structures is very much needed to make those less vulnerable to earthquakes. 2003) caused damages in cities and villages in southeastern part of Bangladesh. In previous studies (Ansary. Results of the analyses predicted risk of catastrophe with thousands of casualties. For this purpose. Chittagong. In 1897.0 caused serious damages to buildings in the northeastern part of India (including Bangladesh) and 1542 people were killed.1 INTRODUCTION 1. Develop contour maps of PGA (in g) for Bangladesh. an earthquake of magnitude 8. a methodology has been developed to combine seismic hazard and vulnerability to determine the seismic risk of future earthquakes affecting Dhaka. Sharfuddin.

1990) is adopted in analysis. location and time of occurrence are explicitly considered. The NSP analysis should be stopped when the value of global ductility obtained in the analysis becomes unattainable. The nonlinear static pushover analysis involves the monotonic application of incrimental lateral forces or displacements to a model of the building until a target deformation is reached. namely until either severe lateral force degradation or local loss of vertical capacity of at least one structural component (e. In this application the pushover analysis should include element strength degradation and should be conducted to failure. In probabilistic seismic hazard analysis (PSHA).6.g.. uncertainties in earthquake size. or more. Coupling the fragility curves with probabilistic seismic hazard analysis will permit assessments of the vulnerability of seismically induced structure. Nonlinear static behavior of a specific building subjected to incremental lateral loads is inferred from its nonlinear dynamic response expected for different levels of ground motion severity. beam.) are reached. McGuire (1978) and Duggal (1989) acceleration attenuation expressions are used in the analysis. (1993). key node(s) of the structure.3 OUTLINE OF METHODOLOGY: Earthquake sources capable of producing significant ground motion at the site are identified and characterized from historical earthquakes.. roof displacement divided by the height of the building) of the structure are the main products of the NSP analysis. Seismic performance of structures will be cast in terms of fragility curves for structural limit states directly related to after earthquake building occupancy status. etc. A four-step process (Reiter. For this purpose a structural analysis software-OpenSees1.1. Peak ground acceleration (PGA) is used to characterize the seismic hazard. A limit-state fragility curve provides the conditional probability that the specified limit state will be reached or exceeded as a function of the severity of the future ground motion. A key feature of OpenSees is the interchangeability of components and the ability to integrate (Mazzoni et al 2005). connection. This can be achieved by including strength degradation at the element level and by limiting at each step of the analysis the incremental displacement at one. OpenSees (Open System for Earthquake Engineering Simulation) is an object-oriented framework for finite element analysis. The lateral load is selected to follow a predetermined or adaptive pattern (Kunnath and Balram. Nonlinear static pushover analysis of the selected structure is conducted using a displacement-controlled approach that includes P-delta effects and strength degradation of structural elements. The uncertainty inherent in building response and capacity for different ground motion levels due to variability in construction and to uncertainty in structural evaluation process is used to obtain the desired fragility curves for the different structural limit states. 1999) that approximately represents the inertial forces at the locations of the significant masses.e. Boore et al. The NSP (Non-linear Static Pushover) curves expressed in terms of base shear versus roof drift (i.2 is used. The engineer identifies on the NSP curves the major inelastic events that occur in the structure 12 . column. This procedure requires that the applied lateral load be allowed to drop (FEMA-273). Source characterization includes definition of each source's geometry and earthquake potential.

These major inelastic events sometimes cause significant drops in the base shear or changes in the global stiffness of the building which in turn translate into changes in slope of the NSP curve These inelastic events are to be associated with specific damage states.along with associated roof drift levels. More formally. This study describes damage states in terms of structural damages. frame structures has recently been prepared (Noor and Manzur 2005) by defining damage states in terms of roof drift value. DSi is therefore defined by (a) a roof drift value. the occurrence of the ith major inelastic event (or a set of events at approximately the same deformation level) identifies the ith damage state. including whether any element has reached ultimate vertical capacity.C. The damage description can be valuable to engineers to compare with observable damage during the inspection of a facility before deciding on the possible building occupancy restriction after an earthquake. A methodology on development of fragility curves for R. and (b) a detailed description of the structural damage associated with that event. DSi.C. 13 .

velocity. as when a particular earthquake scenario is assumed. If greater levels of damage are tolerable. The motion parameter may be acceleration. 2. and duration of strong ground motions. That level of shaking is often described by a design ground motion. 2004). for many structures. Much of the difficulty in design ground motion specification results from its unavoidable reliance on subjective decisions that must be made with incomplete or uncertain information. and time of occurrence are explicitly considered.. of course. Obviously. It is also consistent with the broader approach to Performance-Based Earthquake Engineering put forward by the Pacific Earthquake Engineering Center (e. Seismic hazards may be analyzed deterministically. a relatively strong level of shaking must be designed for. some describe only one of these characteristics. 14 . The process is a crucial step in evaluating the likelihood that the structure will not be able to provide facility after an earthquake and the output will give a rational support to the engineer inspecting the facility after an earthquake. Moehle and Deierlein. Amplitude Parameters The most common way of describing a ground motion is with a time history. quantitative form.2LITERATURE REVIEW 2.1 SEISMIC PERFORMANCE ASSESSMENT OF STRUCTURES The assessment of the seismic performance of existing structures means prediction of the post-earthquake functionality of structures. lower design levels of shaking may be considered and the resulting design will be less expensive. The specification of design ground motion parameters is one of the most difficult and most important problems in geotechnical earthquake engineering. A limit-state fragility curve provides the conditional probability that the specified limit state will be reached or exceeded as a function of the severity of the future ground motion. These decisions largely revolve around the definition of the boundary between acceptable and excessive damage. in which uncertainties in earthquake size. and the measures required resisting that shaking can be quite expensive. while others may reflect two or three. and location of future earthquakes.g.1 Ground Motion Parameters (Kramer 2003) Ground motion parameters are essential for describing the important characteristics of strong ground motion in compact. of more general applicability. Coupling the fragility curves with probabilistic seismic hazard analysis will permit assessments of the vulnerability of seismically induced structure.2. location.2 SEISMIC HAZARD ANALYSIS Seismic hazard analyses involve the quantitative estimation of ground-shaking hazards at a particular site. and uncertainty in the size. Seismic performance of structures can be cast in terms of fragility curves for structural limit states directly related to after earthquake structure occupancy status. may never be realized) of earthquake-induced damage. there are trade-offs between the short-term cost of providing an earthquakeresistant design and the potential long-term cost (which. time. which can be characterized by design ground motion parameters. frequency content. or displacement. 2. If very little damage is acceptable. The procedure developed in this research work is. or probabilistically. or all three may be displayed. Many parameters have been proposed to characterize the amplitude.

The PHA can also be correlated to earthquake intensity (e. By taking the vector sum of two orthogonal components. Duggal. McGuire. 1989 acceleration attenuation expression for alluvial soil.5g but caused no significant damage to structures because the peak accelerations occurred at very high frequencies and the duration of the earthquake was not long. As bed soil of Bangladesh is alluvium McGuire and Duggal expression can be applicable.74g was measured between the Imperial and Brawley faults in the 1979 Imperial Valley earthquake.e. Horizontal accelerations have commonly been used to describe ground motions because of their natural relationship to inertial forces. Murphy and O'Brien. Boore expression can be used with amplification factor for alluvium. The controlling earthquake that is expected to produce the strongest level of shaking is described in terms of its size (usually expressed as magnitude) and distance from the site. the largest dynamic forces induced in certain types of structures (i. indeed. The ratio of PVA to PHA. 1977.g. Public Works Research Institute (PWRI) in Japan proposes McGuire. Trifunac and Brady. it provides no information on the frequency content or duration of the motion. Very high peak accelerations that last for only a very short period of time may cause little damage to many types of structures.. For engineering purposes. 1987). 1993 (equation 3) acceleration attenuation expression is used to predict PGA. a PVA of 1. the maximum resultant PHA (the direction of which will usually not coincide with either of the measured components) can be obtained. Boore expression is developed for rock and McGuire expression is developed for both rock and alluvium and Duggal expression is developed for alluvium. but not always. Vertical accelerations have received less attention in earthquake engineering than horizontal accelerations. very stiff structures) are closely related to the PHA. it can be very useful for estimation of PHA when only intensity information is available. Ground motions with high peak accelerations are usually. Peak vertical accelerations can be quite large. The probability of occurrence of the controlling earthquake is assumed to be one at the points in each source zone closest to the site and zero elsewhere. 1989 (equation 2) and Boore et al. Although peak acceleration is a very useful parameter. 1982). Peak ground acceleration (PGA) is used to characterize the seismic hazard. consequently. A number of earthquakes have produced peak accelerations in excess of 0. primarily because the margins of safety against gravity-induced static vertical forces in constructed works usually provide adequate resistance to dynamic forces induced by vertical accelerations during earthquakes.. 15 . 1978 (equation 1). The PHA for a given component of motion is simply the largest (absolute) value of horizontal acceleration obtained from the accelerometer of that component. as in the cases of earthquakes that occurred before strong motion instruments were available (preinstrumental earthquakes). The use of intensity-attenuation relationships also allows estimation of the spatial variability of peak acceleration from the isoseismic maps of historical earthquakes. Abrahamson and Litehiser. Although this correlation is far from precise. the peak vertical acceleration (PVA) is often assumed to be two-thirds of the PHA (Newmark and Hall. however. 1985. 1978 and Duggal. 1989).Typically. A number of intensity-acceleration relationships have been proposed. Peak Acceleration The most commonly used measure of the amplitude of a particular ground motion is the peak horizontal acceleration (PHA). more destructive than motions with lower peak accelerations. Krinitzsky and Chang. 1975a. only one of these quantities is measured directly with the others computed from it by integration and/or differentiation. has more recently been observed to be quite variable but generally to be greater than two-thirds near the source of moderate to large earthquakes and less than two-thirds at large distances (Campbell. it must be supplemented by additional information to characterize a ground motion accurately.

For many real structures.205 M = Surface magnitude of earthquake. log( R 2 + h 2 ) + b6Gb + b7 Gc (3) b1 -0. Since pulses of high acceleration at high frequencies induce little response in most structures. and displacement values are associated with different frequencies (or periods).89 M ( R2 + h2 ) −1. Response Spectra The response spectrum describes the maximum response of a single-degree-of-freedom (SDOF) system to a particular input motion as a function of the natural frequency (or natural period) and damping ratio of the SDOF system. velocity.. Because of this behavior.17 e −0.308 M ( R + 30 ) −1. one that corresponds to a nonlinear force-displacement relationship.0 for alluvium. Spectral accelerations decrease with increasing ductility.216 b3 0 b4 0 b5 -0.777 b6 0. (1993) acceleration attenuation expression (PGA in g) Log(PGA)= b1 + b2 ( M − 6) + b3 ( M − 6) 2 + b4 R 2 + h 2 + b5 . At low frequencies the average spectral displacement is nearly constant.2 (2) Boore et al. can be used to account for the effects of inelastic behavior. R = Epicentral distance in km. 16 . Benjamin and Associates (1988) proposed that an effective design acceleration be taken as the peak acceleration that remains after filtering out accelerations above 8 to 9 Hz. S is chosen equal to 1.158 b7 0.2 S (1) Where. at high frequencies the average spectral acceleration is fairly constant. The shapes of typical response spectra indicate that peak spectral acceleration.McGuire (1978) acceleration attenuation expression (PGA in g) PGA = 0. has been proposed by at least two researchers. response spectra are often divided into acceleration-controlled (high-frequency). however. In between lays a range of nearly constant spectral velocity. velocity-controlled (intermediate-frequency) and displacement-controlled (low-frequency) portions. h = Hypocentral depth of earthquake in km = 30 km for Bangladesh.0306e 0. An inelastic response spectrum (i. Duggal (1989) acceleration attenuation expression (PGA in cm/s2) PGA = 227 ×10 0. Elastic response spectra assume linear structural force-displacement behavior. with different definitions. inelastic behavior may be induced by earthquake ground motions.254 Gb 0 Gc 0 σ log PGA 0. Effective Design Acceleration.e. The-notion of effective design acceleration. Kennedy (1980) proposed that the effective design acceleration be 25% greater than the third highest (absolute) peak acceleration obtained from a filtered time history.038 b2 0. but total displacements increase.

Within hours. However. and even evaluate source parameters. 1991).11 of BNBC 1993 shows the Normalized Response Spectra for 5% Damping Ratio. must be observed and interpreted. all possible sources of seismic activity must be identified and their potential for generating future strong ground motion evaluated. Figure 1 shows normalized response spectra proposed in NEHRP 1991. Fig. The fact that no strong motions have been instrumentally recorded in a particular area does not guarantee that they have not occurred in the past or that they will not occur in the future. and to a lesser extent. In the 1990s. Figure 1 Normalized Response Spectra Recommended for use in Building Codes (NEHRP. 2. it is virtually impossible for a significant earthquake anywhere in the world to go undetected. seismologists are able to determine its magnitude. Application of response spectra for dynamic analysis of structure in Bangladesh is described in Art. The availability of modern seismographs and seismographic networks has made observation and interpretation of current earthquakes rather convenient. These may take the form of geologic and tectonic evidence. The amplitude. soil profile types are as described in NEHRP 1991 2. frequency content. 17 .Response spectra reflect strong ground motion characteristics indirectly.7 of Part 6 of BNBC 1993. duration of the input motion all influence spectral values. particularly when compared with the time scales on which large earthquakes usually occur. It is important to remember that response spectra represent only the maximum responses of a number of different structures. Identification of seismic sources requires some detective work.2. or historical (preinstrumental) seismicity. some of which are obvious and others quite obscure. 6. the response of structures is of great importance in earthquake engineering. locate its rupture surface. and the response spectrum has proven to be an important and useful tool for characterization of strong ground motion.5. since they are "filtered" by the response of a SDOF structure. The current ability to identify and locate all earthquake sources is a relatively recent development. other clues of earthquake activity must be uncovered.2. In the absence of an instrumental seismic record. nature's clues. The occurrence of a large earthquake is now recorded by hundreds of seismographs around the world.2 Identification and Evaluation of Earthquake Sources (Kramer 2003) To evaluate seismic hazards for a particular site or region.

the geologic record may be very complex or it may be hidden by thick layers of recent sediments that have not been displaced by seismic activity. they can also be used to evaluate the rate of recurrence of earthquakes. though often difficult part of a seismic hazard analysis. Since historical records are dated. should be related to the rate of strain energy accumulation and also to the rate of strain energy release (Smith.0089T + 0. The search for geologic evidence of earthquake sources centers on the identification of faults. in Japan and the Middle East it may extend about 2000 years and up to 3000 years or so in China (Ambraseys.. field geology and geomorphology (Adair. 1987. a geographic pattern of historic epicenters provides strong evidence for the existence of earthquake source zones. The written historical record extends back only a few hundred years or less in the United States.Geologic Evidence The theory of plate tectonics assures us that the occurrence of earthquakes is written in the geologic record. Idriss.g. of various strata. Atwater et al. including the review of published literature. and geophysical techniques.. Subsequently. this geologic record is easily accessible and relatively easily interpreted by the trained seismic geologist. Criteria for identification of faults are described in numerous textbooks on structural geology. In some parts of the world.. 1975. Historical Seismicity Earthquake sources may also be identified from records of historical (preinstrumental) seismicity. 1987) along the coasts of Washington and Oregon. therefore. the maximum intensity can be determined and used to estimate the location of the earthquake epicenter and the magnitude of the event. geologic evidence of historical great earthquakes was discovered (e. Tectonic Evidence Plate tectonics and elastic rebound theory tell us that earthquakes occur to relieve the strain energy that accumulates as plates move relative to each other. Alien. interpretation of air photos and remote sensing (e. For major subduction zones. 18 . 1979). primarily in the form of offsets. or seismicity. 1981). Historical accounts of ground-shaking effects can be used to confirm the occurrence of past earthquakes and to estimate their geographic distributions of intensity.134V+ 7. Study of the geologic record of past earthquake activity is called paleoseismology (Wallace. Bolt. 1971. Heaton and Kanamori (1984) used this relationship to suggest that the Cascadia subduction zone off the coasts of Oregon. infrared photograph) imagery. 1976. Woodward-Clyde Consultants. A variety of tools and techniques are available to the geologist. The rate of movement. field reconnaissance including logging of trenches. 1979.96 Where T is the age in millions of years and V is the rate of convergence in cm/yr.g. however. 1978. Although the accuracy of locations determined in this way depends strongly on population density and the rate of earthquake recurrence. In other locations. in particular areas. The identification of seismic sources from geologic evidence is a vital. When sufficient data are available. and British Columbia could be capable of generating great earthquakes of magnitude well above 8.0. test pits and borings. Ruff and Kanamori (1980) related maximum magnitude to both the rate of convergence and the age of the subducted slab according to MW = . Atwater. or relative displacements. 1988). Washington. 1985).

2. Nevertheless. the use of deterministic seismic hazard analysis (DSHA) was prevalent. it provides no information on the likelihood of occurrence of the controlling earthquake. particularly regarding earthquake potential (step 1). The scenario consists of the postulated occurrence of an earthquake of a specified size occurring at a specified location. DSHA provides a straightforward framework for evaluation of worst-case ground motions. The selection is made by comparing the levels of shaking produced by earthquakes (identified in step 1) assumed to occur at the distances identified in step 2. the level of shaking that might be expected during a finite period of time (such as the useful lifetime of a particular structure or facility).2. Identification and characterization of all earthquake sources capable of producing significant ground motion at the site.3 Deterministic Seismic Hazard Analysis (Kramer 2003) In the early years of geotechnical earthquake engineering. depending on the measure of distance of the predictive relationship(s) used in the following step. generally expressed in terms of some ground motion parameter. the shortest distance between the source zone and the site of interest is selected. compared with the average period of time between large earthquakes. DSHA appears to be a very simple procedure. such as nuclear power plants and large dams. Their most significant limitation is the short period of time. the earthquake that is expected to produce the strongest level of shaking). A typical DSHA can be described as a four-step process (Reiter. The controlling earthquake is described in terms of its size (usually expressed as magnitude) and distance from the site. and in many respects it is.e. 2. Source characterization includes definition of each source's geometry (the source zone) and earthquake potential. about 10 earthquakes of Ms > 1 have occurred somewhere in the world each year (Kanamori. The DSHA procedure is shown schematically in Figure 2. and response spectrum ordinates are commonly used to characterize the seismic hazard. Instrumental records from large earthquakes have been available since about 1900. Perhaps most important. peak velocity. Peak acceleration. 3. When applied to structures for which failure could have catastrophic consequences. for which they have been available. The distance may be expressed as an epicentral distance or hypocentral distance. instrumental recordings represent the best available information for the identification and evaluation of earthquake sources. Again. Selection of the controlling earthquake (i. Analysis of aftershocks can also aid in the delineation of earthquake source zones.. Its characteristics are usually described by one or more ground motion parameters obtained from predictive relationships. 4. DSHA involves subjective decisions. that can require the combined expertise and opinions of seis19 . although many from before 1960 are incomplete or of uneven quality. 1990) consisting of: 1. usually in terms of the ground motions produced at the site by the controlling earthquake. However. the likelihood of it occurring where it is assumed to occur. at the site. The hazard at the site is formally defined.Instrumental Seismicity Over the past 80 or 90 years. Selection of a source-to-site distance parameter for each source zone. 1988). In most DSHAs. or the effects of uncertainties in the various steps required to compute the resulting ground motion characteristics. Expressed in these four compact steps. the alignment of instrumentally located epicenters or hypocenters indicates the existence of earthquake sources. A DSHA involves the development of a particular seismic scenario upon which a ground motion hazard evaluation is based.

on the 20 . seismic geologists.4 Probabilistic Seismic Hazard Analysis (Kramer 2003) In the past 20 to 30 years the use of probabilistic concepts has allowed uncertainties in the size. is identical to the first step of the DSHA. Source 1 Site Source 2 Source 1 Site Source 2 M1 M3 Source 3 M2 Source 3 Step 1 Step 2 Ground Motion Parameter. . quantified. as illustrated in Figure 4. The first step. The PSHA methodology described in this section is similar in many respects to the well-established methods developed by Cornell (1968). The DSHA. In most cases. The PSHA can also be described as a procedure of four steps (Reiter. and rate of recurrence of earthquakes and in the variation of ground motion characteristics with earthquake size and location to be explicitly considered in the evaluation of seismic hazards. economists. Probabilistic seismic hazard analysis (PSHA) provides a framework in which these uncertainties can be identified. each of which bear some degree of similarity to the steps of the DSHA procedure. and combined in a rational manner to provide a more complete picture of the seismic hazard. (1982). location. social scientists.6. implying that earthquakes are equally likely to occur at any point within the source zone. These distributions are then combined with the source geometry to obtain the corresponding probability distribution of source-to-site distance. R Step 3 Step 4 Figure 2 Four steps of a deterministic seismic hazard analysis 2. and Algermissen et al. Understanding the concepts and mechanics of PSHA requires familiarity with some of the terminology and basic concepts of probability theory. Y M1 M2 M3 Controlling Earthquake Y1 Y= R1 R3 R2 Y2 . identification and characterization of earthquake sources. uniform probability distributions are assigned to each source zone. except that the probability distribution of potential rupture locations within the source must also be characterized. YN Distance. 1990). risk analysts.2. and government officials. 1.mologists. engineers.

A recurrence relationship. 2.other hand. the seismicity or temporal distribution of earthquake recurrence must be characterized. earthquake size. earthquake size. as DSHAs often do. The uncertainty inherent in the predictive relationship is also considered. 21 . The uncertainty inherent in the predictive relationship is also considered in a PSHA. The recurrence relationship may accommodate the maximum size earthquake. the uncertainties in earthquake location. The ground motion produced at the site by earthquakes of any possible size occurring at any possible point in each source zone must be determined with the use of predictive relationships. Next. and ground motion parameter prediction are combined to obtain the probability that the ground motion parameter will be exceeded during a particular time period. which specifies the average rate at which an earthquake of some size will be exceeded. but it does not limit consideration to that earthquake. 3. 4. and zero elsewhere. the uncertainties in earthquake location. and ground motion parameter prediction are combined to obtain the probability that the ground motion parameter (PGA is used as the parameter in this study) will be exceeded during a particular time period. The proper performance of a PSHA requires careful attention to the problems of source characterization and ground motion parameter prediction and to the mechanics of the probability computations. implicitly assumes that the probability of occurrence is 1 at the points in each source zone closest to the site. The ground motion produced at the site by earthquakes of any possible size occurring at any possible point in each source zone is determined with the use of attenuation expression. Finally. is used to characterize the seismicity of each source zone. Finally.

nonuniform distributions may be used when sufficient 22 . generally originate in zones near the volcanoes that are small enough to allow them to be characterized as point sources. Earthquakes associated with volcanic activity.Site log (# earthquake > m) Source 1 Source 2 1 3 2 Source 3 Magnitude. m Step 1 Ground Motion Parameter. R Parameter Value. or where faulting is so extensive as to preclude distinction between individual faults. of the distribution of earthquake size for each source. for example. can be considered as two-dimensional area sources. Y Step 2 P [ Y > y* ] Distance. For the purposes of a seismic hazard analysis. the source zones may be similar to or somewhat different than the actual source. Earthquakes are usually assumed to be uniformly distributed within a particular source zone (i. Well-defined fault planes.2. Each of these characteristics involves some degree of uncertainty. Areas where earthquake mechanisms are poorly defined.e. and of the distribution of earthquakes with time. Spatial Uncertainty The geometries of earthquake sources depend on the tectonic processes involved in their formulation. can be treated as three-dimensional volumetric sources. earthquakes are considered equally likely to occur at any location). The assumption of uniformity is by no means required.. on which earthquakes can occur at many different locations. y* Step 3 Step 4 Figure 3 Four steps of a probabilistic seismic hazard analysis 2. depending on the relative geometry of the source and site of interest and on the quality of information about the sources.5 Earthquake Source Characterization (Kramer 2003) Characterization of an earthquake source requires consideration of the spatial characteristics of the source and of the distribution of earthquakes within that source.

consequently. a histogram that approximates fR(r) can be constructed by tabulating the values of R that correspond to the center of each element. Figure 4 Examples of variations of source-to-site distance for different source zone geometries. For the linear source of Figure 4b. Since predictive relationships express ground motion parameters in terms of some measure of source-to-site distance. however. Since l 2 = r 2 − rmin the probability density function of R is given by fR ( r) = r 2 L f r 2 − rmin For source zones with more complex geometries. the distance. 2 fL(l) = l/Lf. parts of which may be much closer to the site than the hypocenter. energy is released over the entire fault rupture surface. However. R. Other cases are not as simple. often translate into a uniform distribution of source-to-site distance. the probability that R = rs is assumed to be 1 and the probability that R ≠ rs. that is. Der-Kiureghian and Ang (1977) noted 23 . the probability that an earthquake occurs on the small segment of the fault between L = l and L = l + dl is the same as the probability that it occurs between R = r and R = r + dr. Consequently fR ( r) = fL (l) dl dr If earthquakes are assumed to uniformly distributed over the length of the fault. A uniform distribution within the source zone does not. dividing the irregular source zone of Figure 4c into a large number of discrete elements of equal area. The shape of the probability distribution can be visualized by considering the relative portions of the source zone that would fall between each of a series of circles (or spheres for three-dimensional problems) with equal differences in radius The preceding discussion assumes that all the energy is released at the hypocenter of the earthquake. the spatial uncertainty must be described with respect to the appropriate distance parameter. is known to be rs. it is easier to evaluate fR(r) by numerical rather than analytical methods. f L ( l ) dl = f R ( r ) dr Where fi(l) and fR(r) are the probability density functions for the variables L and R. respectively. The uncertainty in source-to-site distance can be described by a probability density function. For example. zero.information to justify them exists. For the point source of Figure 4a.

the mean annual rate of exceedance of small earthquakes is greater than that of large earthquakes.6 Gutenberg-Richter Recurrence Law Gutenberg and Richter (1944) gathered data for southern California earthquakes over a period of many years and organized the data according to the number of earthquakes that exceeded different magnitudes during that time period. As would be expected. The a and b parameters are generally obtained by regression on a database of seismicity from the source zone of interest. The resulting GutenbergRichter law for earthquake recurrence was expressed as logλm = a-bm Where λm is the mean annual rate of exceedance of magnitude m. Size Uncertainty Once an earthquake source is identified and its corresponding source zone characterized. the database is likely to be relatively sparse. 2. All source zones have a maximum earthquake magnitude that cannot be exceeded. and b (the b value) describes the relative likelihood of large and small earthquakes.that the rupture surface of a large earthquake with a distant hypocenter could release energy much closer to the site. it can be large for some and small for others. The Gutenberg-Richter law is illustrated schematically in Figure 5a. They divided the number of exceedances of each magnitude by the length of the time period to define a mean annual rate of exceedance. The reciprocal of the annual rate of exceedance for a particular magnitude is commonly referred to as the return period of earthquakes exceeding that magnitude. the number of larger magnitude earthquakes decreases compared to those of smaller magnitudes. epicentral intensity has also been used. Temporal Uncertainty To calculate the probabilities of various hazards occurring in a given time period. The strain energy may be released aseismically. When the logarithm of the annual rate of exceedance of southern California earthquakes was plotted against earthquake magnitude. As the b value increases. or in the form of earthquakes. Unless the source zone is extremely active. The distribution of earthquake sizes in a given period of time is described by a recurrence law. Xm of an earthquake of magnitude m. examination of available seismicity records has revealed little evidence (when aftershocks are removed) of temporal patterns in earthquake recurrence. Earthquakes have long been assumed to occur randomly with time.2. and in fact. the seismic hazard analyst's attention is turned toward evaluation of the sizes of earthquakes that the source zone can be expected to produce. and developed methods to account for rupture surface dimensions in PSHA. the source zone will produce earthquakes of different sizes up to the maximum earthquake. the distribution of earthquake occurrence with respect to time must be considered. 10a is the mean yearly number of earthquakes of magnitude greater than or equal to zero. Since the use of both instrumental and historical events is 24 . A basic assumption of PSHA is that the recurrence law obtained from past seismicity is appropriate for the prediction of future seismicity. with smaller earthquakes occurring more frequently than larger ones. The Gutenberg-Richter law is not restricted to the use of magnitude as a descriptor of earthquake size. The assumption of random occurrence allows the use of simple probability models. In general. a linear relationship was observed. Worldwide recurrence data are shown in Figure 5b.

a PSHA is intended to evaluate the hazard from discrete. showing meaning of a and b parameters. (After Esteva.303b.usually required. Fitting a straight line such as that implied by the Gutenberg-Richter law through recurrence data in which the mean rate of exceedance of small earthquakes is underestimated will tend to flatten the line. As a result. If earthquakes smaller than a lower threshold magnitude m0 are eliminated. the database may contain both magnitude (possibly based on different scales) and intensity data. and (b) application of Gutenberg-Richter law to worldwide seismicity data.303a and β = 2. small earthquakes can go undetected for a variety of physical and demographic reasons. independent releases of seismic energy. Although such dependent events can cause significant damage. 1980. the record of seismicity may be distorted by the presence of dependent events such as aftershocks and foreshocks (Merz and Cornell. the actual mean rate of small earthquakes will be underpredicted and the mean rate of large earthquakes will be overpredicted. Figure 5 (a) Gutenberg-Richter recurrence law. Different methods have been proposed (Stepp. 1970) The standard Gutenberg-Richter recurrence law may also be expressed as ln λm = α − β m Where α = 2. the effects of very small earthquakes are of little interest and it is common to disregard those that are not capable of causing significant damage. Therefore. For engineering purposes. dependent events must be removed from the seismicity database and their effects accounted for in separate analyses. from − ∞ to + ∞ . The standard GutenbergRichter law covers an infinite range of magnitudes. The equation shows that the Gutenberg-Richter law implies that earthquake magnitudes are exponentially distributed. 1972. 1990) as λm = νe [ − β ( m−m0 ) ] m > m0 25 . Weichert. 1973). The historical record is usually more complete for large earthquakes than for small earthquakes. In some areas. the mean annual rate of exceedance can be written (McGuire and Arabasz. necessitating the conversion of one measure of size to the other. EPRI. 1986) to correct incomplete records. Completeness of the database must also be considered.

7 Predictive Relationships Predictive relationships are nearly always obtained empirically by least-squares regression on a particular set of strong motion parameter data. even though earthquakes of that size have never been observed. some amount of scatter in the data is inevitable.02 per year (a return period of only 50 years). Reflecting the form of most predictive relationships. the lower threshold magnitude is set at values from about 4. the standard Gutenberg-Richter law predicts nonzero mean rates of exceedance for magnitudes up to infinity.0 since magnitudes smaller than that seldom cause significant damage. increasing the maximum magnitude requires a substantial decrease in the mean annual rate of exceedance of lower magnitude events to account for the extra energy released in large earthquakes. The extent to which actual slip rates vary with time.0 to 5. the mean annual rate of exceedance can be expressed (McGuire and Arabasz. the moment rate is proportional to the slip rate. This implies.2. mmax.Whereν = e ( α − βm0 ) . The probability that a particular ground motion parameter Y exceeds a certain 26 . If it is known or can be estimated. This considerable uncertainty must be accounted for in computation of seismic hazards. Scatter in the data can be quantified by confidence limits (Campbell. the standard deviation of the (natural) logarithm of the predicted parameter is usually computed. that the Circumpacific belt (Figure 5b). 1985) or by the standard deviation of the predicted parameter. In most PSHAs. however. 1990) as λm = ν e [ − β ( m−m0 ) ] − e [ − β ( mmax −m0 ) ] 1 − e [ −β ( mmax −m0 ) ] m0 ≤ m ≤ mmax In the constant moment rate model. would produce a magnitude 10 earthquake at a mean annual exceedance rate of about 0. Some maximum magnitude. Since the seismic moment is proportional to the amount of slip (displacement) that occurs in an earthquake. The CDF and PDF for the Gutenberg-Richter law with upper and lower bounds can be expressed as FM ( m ) = P[ M < m ∩ m0 ≤ m ≤ mmax ] = f M ( m) = 1 − e [ − β ( m−m0 ) ] 1 − e [ −β ( mmax −m0 ) ] βe [ −β ( m−m0 ) ] 1 − e [ − β ( mmax −m0 ) ] 2. Hence the constant-moment-rate model is equivalent to a constant-slip-rate model and can be used when the slip rate is known to be constant. appears to be different for different faults and can even fluctuate with time along the same fault. for example. is associated with all source zones. travel path and site conditions. The resulting probability distribution of magnitude for the Gutenberg-Richter law with lower bound can be expressed in terms of the cumulative distribution function (CDF): FM ( m ) = P[ M < m ∩ M > m0 ] = λm0 − λm λm0 = 1 − e − β ( m−m0 ) or the probability density function (PDF): f M ( m) = d FM ( m ) = β e −β ( m−m0 ) dm At the other end of the magnitude scale. The scatter results from randomness in the mechanics of rupture and from variability and heterogeneity of the source. Despite attempts to remove questionable data and the use of quality-based weighting schemes.

5 would imply a 0. 2. for an earthquake of a given magnitude. one that yields values of a random variable describing the number of occurrences of a particular event during a given time interval or in a specified spatial region. or location of any preceding event..2.+ P[N =∞] = 1 − P[N = 0] = 1 − e − λt 27 .06% probability that the PHA would exceed 2.5g. r. To characterize the temporal distribution of earthquake recurrence for PSHA purposes. Poisson processes possess the following properties: 1. (1979). the spatial applications of the Poisson model will not be considered further. For a Poisson process. The probability of occurrence during a very short time interval is proportional to the length of the time interval.8 Poisson Model The temporal occurrence of earthquakes is most commonly described by a Poisson model. The number of occurrences in one time interval is independent of the number that occurs in any other time interval. The uses of distributions that impose an upper limit on Y have been studied by Kulkarni et al. is given by P[Y>y*│ m. The probability of more than one occurrence during a very short time interval is negligible. In general. the probability of a random variable N.value. Since PSHAs deal with temporal uncertainty.. size. occurring at a given distance. ground motion parameters are usually assumed to be lognormally distributed (the logarithm of the parameter is normally distributed). The value of FY (y) depends on the probability distribution used to represent Y. 2. with no "memory" of the time. The time between events in a Poisson process can be shown to be exponentially distributed. Note that the probability of occurrence of at least one event in a period of time t is given by P[N≥1] = P[N=1] + P[N=2] + P[N=3] +. the Poisson probability is usually expressed as P[ N = n ] = ( λ t ) n e − λt n! Where λ. a hypothetical PHA attenuation relationship that predicts a mean PHA of 0. The Poisson model provides a simple framework for evaluating probabilities of events that follow a Poisson process. Bender (1984). m.. and Zemell (1984). For example. is the average rate of occurrence of the event and t is the time period of interest.r] = 1-FY (y*) Where FY (y) is the value of the CDF of Y at m and r. however. the unbounded characteristics of that distribution can attribute a nonzero probability to unrealistic values of the ground motion parameter. These properties indicate that the events of a Poisson process occur randomly.. 3. representing the number of occurrences of a particular event during a given time interval is given by P[ N = n ] = µ ne−µ n! Where µ is the average number of occurrences of the event in that time interval. y*.5g with σ ln y = 0.

Numerical integration. The process is then repeated for all possible magnitudes and locations with the probabilities of each summed. the total average exceedance rate for the region will be given by [ ] λ y* = ∑ν i ∫∫ P[Y > y * m. A common approach involves the development of seismic hazard curves. respectively. for virtually all realistic PSHAs.10Seismic Hazard Curves Seismic hazard curves can be obtained for individual source zones and combined to express the aggregate hazard at a particular site. r ] f ( m ) f ( r )dmdr M R Where P[Y>y*│m.2. the Poisson model can be combined with a suitable recurrence law to predict the probability of at least one exceedance in a period of t years by the expression P[N≥1] = 1 − e − λmt 2. is to divide the possible ranges of magnitude and distance into NM and NR segments. and distance.When the event of interest is the exceedance of a particular earthquake magnitude. is therefore required. P[Y>y*] = P[Y>y*│X] P[X] = ∫ P[Y > y* ¦ X] fx(X)dx Where X is a vector of random variables that influence Y. The seismic hazard curves can then be used to compute the probability of exceeding the selected ground motion parameter in a specified period of time. R. r ] f Mi ( m ) f Ri ( r ) dmdr i =1 NS The individual components of the equation are. Y. respectively. The average exceedance rate can then be estimated by 28 . frequency. The required calculations are described in the following paragraphs.2. is calculated for one possible earthquake at one possible source location and then multiplied by the probability that that particular magnitude earthquake would 'occur at that particular location. All involve some level of probabilistic computations to combine the uncertainties in earthquake size. which indicate the annual probability of exceedance of different values of a selected ground motion parameter. M. used here for simplicity rather than efficiency. y*. In most cases the quantities in X are limited to the magnitude. that is. 2. sufficiently complicated that the integrals cannot be evaluated analytically. of a ground motion parameter. which can be performed by a variety of different techniques. and effects to estimate seismic hazards. The basic concept of the computations required for development of seismic hazard curves is fairly simple. If the site of interest is in a region of Ns potential earthquake sources. each of which has an average rate ( α − βm ) of threshold magnitude exceedance. The probability of exceeding a particular value. location. the probability that a ground motion parameter Y will exceed a particular value y* can be computed using the total probability theorem. For a given earthquake occurrence.9 Probability Computations The results of a PSHA can be expressed in many different ways. One approach.r] is obtained from the predictive relationship and f M ( m ) and f R ( r ) are the probability density functions for magnitude and distance.ν i = e i 0 . Assuming that M and R are independent the probability of exceedance can be written as P[Y>y*] = ∫∫ P[Y > y * m.

2004) The five limit states identified in increasing order of damage severity are: Green tag. mj. Onset of Damage. Yellow tag. Y: the structure is deemed fit for restricted occupancy. interstory drift ratio. material properties. peak ground motion values. rk. This Equation is then equivalent to λ y* = ∑∑∑ν i P Y > y * m j . 29 . story shear force or loss of dynamic capacity.3 DEVELOPMENT OF FRAGILITY CURVES A limit-state fragility curve provides the conditional probability that the specified limit state will be reached or exceeded as a function of the severity of the future ground motion (Bazzurro et al.1 Limit States (Bazzurro et al. OD: FEMA 356 and HAZUS manual define the onset of significant nonlinear behavior (Immediate Occupancy in FEMA 356 and Slight Damage in HAZUS) for different types of structures. earthquake intensities etc. The access is limited to specialized personnel only. retrofit or rebuilding. structure modeling and analysis procedure.5)( rmax − rmin ) / N R ∆m = ( mmax − m0 ) / N m ∆r = ( rmax − rmin ) / N R This is equivalent to assuming that each source is capable of generating only NM different earthquakes of magnitude. Access is not allowed until completion of detailed engineering evaluation. until detailed engineering evaluation is completed. No limitations on post-earthquake access are implied by this limit state. C: the structure has collapsed or is on the verge of global instability or local collapse. 2. R: the structure is deemed unsafe.λ y* = ∑∑∑ν i P Y > y * m j .5)( mmax − m0 ) / N M rk = rmin + ( k − 0. More refined methods of numerical integration will provide greater accuracy at the same values of NM and NK. 2. Collapse. Fragility curves involve uncertainties associated with dynamic displacements produced by different ground motion records (even though they have same intensity value). The ground motion intensities in the fragility functions can be spectral quantities. rk f Mi ( m j ) f Ri ( rk )∆m∆r NS NM NR i =1 j =1 k =1 [ ] Where m j = m0 + ( j − 0. at only NR different source-to-site distances.3. maximum roof displacement. Red tag. structure geometry. G: the structure is deemed fit for immediate occupancy. rk P[ M = m j ] P[ R = rk ] NS NM NR i =1 j =1 k =1 [ ] The accuracy of the crude numerical integration procedure described above increases with increasing NM and NR. 2004). The limit states in fragilities are related to structure functionality status and may be defined as global drift ratio.

2. The fragility of a component or system defines the conditional probability of its attaining or exceeding a performance limit state. These dominant contributors can be identified through the formalism of a probabilistic safety assessment. P[LS] = ∑ P[LS|D = d] P[D = d] (4) In which D is a random variable (or random vector) describing the intensity of the demand on the system. fragility (or vulnerability) can be described in terms of the conditional probability of a system reaching a prescribed limit state (LS) for a given system demand D = d. The capacity controls the central location of the Fragility Curve (FC) and uncertainty in the capacity and uncertainty in response of the structure to the demand controls the shape (or dispersion) of the FC (Figure 6). For a deterministic system with no capacity uncertainty. The probability P[D=d] defines the hazard and can be obtained from probabilistic hazard analysis. In a system of different environment limit states can be generalized keeping equal safety or equal probability of the demand for the same limit state. and the summation is taken over all possible values of D.Hence. Expressed in this general manner. The conditional probability. Strictly speaking.3.3. P[LS/D = d] = FR(x). 30 . of course.3 Fragility Curve As noted above. such limit states for specific structural components and systems may be either strength or deformation-related. the fragility (or vulnerability) is a function of the system capacity against each limit state as well as the uncertainty in the capacity. or PSA. or conditions in which the system ceases to perform its intended functions in some way. and P[LS/D = d] is the conditional limit state probability. the most severe stage of the red tag condition. the onset of damage limit state lies within the green tag state boundaries while the collapse state is. given the occurrence of a particular operational or environmental demand. or probability of demand P(LS/D = d). performance and reliability of an engineered system. which may range from loss of function to incipient collapse. the risk to a distributed infrastructure systems is quantified. and accordingly must play a central role in PBE. It is distinguished from traditional deterministic approaches to safety assurance by its focus on why and how the system might fail and by its explicit treatment of uncertainties. Demands can be in the form of maximum force. the limit state probability can be expressed as. 2. In a (narrow) structural engineering sense. FC is primarily a property of the system dependent on the limit state. In PBE. displacement caused by earthquake ground motions. Limit states related to structural behavior range from unserviceability to various degrees of damage including incipient collapse. the FC is a step function. This process enables the benefits of alternate seismic risk mitigation strategies to be assessed in terms of their impact on the performance of the built environment during a spectrum of earthquake hazards and on the affected population. With the limit states identified.2 Performance Based Engineering and Probabilistic Safety Assessment Performance-based engineering (PBE) is a new paradigm for seismic risk reduction across regions or interconnected systems (Abrams. over a given period of time. 2002). both in the phenomena and in the analytical tools used to model them. given that D = d. is the fragility. One begins the PSA process by identifying limit states (LS). The variable d is denoted the "control" or "interface" variable. or more generally a prescribed intensity measure of the ground motion. It is clear that components and systems that are dominant contributors to risk should receive the focus of attention in the assessment process underlying PBE. evaluated and managed through an assessment and selective intervention process aimed at selected components of that system. A PSA is a structured framework for evaluating uncertainty.

Pt ( LS ) = ∫ ( P ( LS D = d ) f D ( d ) ) dD 2.3 0.25 0. The mainshock damage may cause either Decreased global lateral collapse capacity. the seismic hazard needs to be included in the consideration.75 0.6 0. 2004 for seismic assessment of structures uses the knowledge of the nonlinear static behavior of a specific structure subjected to incremental lateral deflection to infer its nonlinear dynamic response expected for different levels of ground motion severity. Decreased seismic capacity with respect to loss of vertical resistance. the limit state (LS) probability over the time period. The vulnerability needs to be described in terms of probability of a set of given limit states being reached of a system at a given location over a given period of time (t).. depending on the regional seismicity and ground excitation.3.5 0.4 0.2 0. A unique aspect of this procedure is that the structure safety while in a damaged state is measured in terms of its ability to protect occupants from death or injury due to aftershocks that might occur prior to its repair. yellow or red tagging). Knowing the fragility curve. to gravity and live loads.50 0. t.FLS (Probability of attaining limit state) 1.4 Limit State Probability To tie the vulnerability of a given system to the seismicity of the region. through Equation 5 the fragility curve and the probabilistic demand curve are combined.3.00 Smaller uncertainty fragility curve Deterministic fragility curve Larger uncertainty fragility curve 0. t) can be evaluated from Equation 4. In other words.g. (5) In which fD(d) = the probability density function of the demand over a time period. These restrictions dictate the facility effectiveness of the structure in the immediate aftermath of an earthquake and provide valuable information to estimate expected downtime. Quantitative measures of the implied degradation in structure safety is used to associate each of several post-earthquake structural damage states with an appropriate post-earthquake structural limit state that may imply some degree of occupancy restrictions (e. The uncertainty inherent in building response and capacity for different ground motion levels due to variability in construction and to uncertainty in structural evaluation process is 31 . locally or more widely.1 0. t (P.5 Fragility Analysis of Structures The Guidelines of Bazzurro et al.7 System Demand (PGA in g) Figure 6 Typical fragility curves 2.

Increase the deformation until either the structure base shear has significantly dropped (e. key node(s) of the structure. hence. This step computes the incremental dynamic analysis (IDA) curve for the intact structure in each direction. SPO) of the selected structure (in each orthogonal direction) is conducted in the as-is condition using a displacement-controlled approach that includes P-delta effects and strength degradation of structural elements. post-earthquake damage state) expected under a range of potential (mainshock) ground motions.used to obtain the desired fragility curves for the different structural limit states. Local loss of vertical capacity is potentially life threatening and for this reason this condition is associated here with structural collapse. Each potential damage state is categorized with no loss of local vertical capacity as one of the following five structural limit states: green-tag. STEP 2: NSP for the damaged Structure For each of the damage states that do not include loss of local vertical capacity (STEP 1). Each of the intact structure’s two NSP curves (STEP 1) is converted into estimates of the peak dynamic displacement (and. STEP 3: Inferring dynamic response from static response. conduct analogous NSP analyses assuming the structure is in that damage condition after the mainshock. The procedure as described in Bazzurro et al. Important damage states along the displacement axis of the NSP curve(s) is identified. Local loss of vertical capacity does not necessarily imply that the structure has no global lateral capacity left to resist aftershocks. The tagging is based on the corresponding residual capacities identified in STEP 3 and on the (site-specific) mean annual frequency of exceedance of the (aftershock) ground motion values corresponding to such residual capacities. The ordinate of the highest point on the IDA provides the (dynamic) capacity of the intact building. onset-of-damage. or more. This can be achieved by including strength degradation at the element level and by limiting at each step of the analysis the incremental displacement at one. red-tag. if any. These results will be used (STEP 3) to estimate the residual (dynamic) capacity to resist aftershocks and hence the relative life safety threat of the building when in one of these potential states..g. 2004 has six steps: STEP 1: Nonlinear Static Procedure (NSP) for the intact Structure A conventional nonlinear static procedure (NSP) analysis (also known as static pushover. which is expressed here in terms of (median) ground motion level necessary to induce either global lateral instability of the structure or local collapse anywhere in the structure. 32 . to 20% or 30% or less) compared to its peak value or until at least one of the structural elements or connections has reached its ultimate capacity to withstand vertical loads even in absence of any future ground shaking. STEP 4: Occupancy Status for Damaged Structure. Damage states. for which vertical capacity is lost at least at one structural component are associated with the collapse limit state. or collapse. yellow-tag. using the SPO2IDA spreadsheet tool provided (Vamvatsikos and Cornell. The NSP analysis should also be stopped when the value of global ductility obtained in the analysis becomes unattainable in reality. 2001b). This procedure requires that the applied lateral load be allowed to drop (FEMA-273).

g. the fragility curves associated with the onset-ofdamage. therefore.. connection.3. These major inelastic events sometimes cause significant drops in the base shear or changes in the global stiffness of the building which in turn translate into changes in slope of the NSP curve. 33 . roof displacement divided by the height of the building) for the two main orthogonal directions of the structure are the main products of the NSP analysis. It is important to emphasize that the seismic evaluations should be performed using expected strength values rather than nominal or lower bound strengths. the red-tag state. failure of a significant portion of the roof/wall connection of a tilt-up building.6 Nonlinear Static Procedure (NSP) Curve for the Intact Structure The knowledge of the post-elastic behavior of the building subject to incremental lateral seismic loading is the starting point of this procedure. 2000) that approximately represents the inertial forces at the locations of the significant masses. fracture of at least 10% of all top flanges in a steel moment-resisting frame. 1999. or local collapse of beam-column connection due to shear failure) along with associated roof drift levels.. In other cases their occurrence may only moderately change the global stiffness of the structure and. the red-tag. it is assumed here that the engineer is accustomed to performing nonlinear static procedure (or static pushover) analyses at least for the two main orthogonal directions of the structure. The lateral load is selected to follow a predetermined or adaptive pattern (Kunnath and Balram. namely until either severe lateral force degradation or local loss of vertical capacity of at least one structural component (e.. The engineer identifies on the NSP curves the major inelastic events that occur in the structure (e. FEMA 356. and the collapse states are plot. The NSP curves expressed in terms of base shear versus roof drift (i. 2000). The intent here is to estimate the realistic response of a structure and not to apply conservative procedure to achieve a “safe” building design. The nonlinear static procedure normally involves the monotonic application of lateral forces (or displacements) to a model of the building until a target deformation is reached. may not produce sharp kinks in the pushover curve. The fragility curve corresponding to the green-tag state is equal to unity for all levels of ground motion. 2. the yellow-tag state. beam.e. column.g.g. and the collapse state are identified according to a pre-defined criterion provided in this step. the four (mainshock) ground motion levels expected to cause the structure to enter the onset-of-damage state. Such ground motion levels are identified by their median values and related dispersion measure. In this application the pushover analysis should include element strength degradation and should be conducted to failure. In line with other current Guidelines (e. STEP 6: Computation of the fragility curves. as depicted in Figure 7.) are reached.STEP 5: Ground motion level associated with a structural limit state. Using the results of STEPS 3 and 4. Balram and Kunnath. FEMA 273. Based on the median values and corresponding dispersion measure of the four ground motion levels number provided at STEP 5.. etc. 1997. the yellow-tag.

on whether the ultimate deformation of a ductile element (or connection) or the maximum strength of a brittle member (or connection) have been attained. DSi is therefore defined by (a) a roof drift value. 2006) Modern earthquake engineering utilizes modeling and simulation to understand the behavior and performance of systems during earthquakes. OpenSees is an object-oriented framework for finite element analysis. Each one of these damage states will be later associated with one (and only one) structural limit state. Such ultimate strain and strength values are sometimes available from laboratory tests published in the literature. A key feature of OpenSees is the interchangeability of components and the ability to integrate (Mazzoni et al 2005). the occurrence of the ith major inelastic event (or a set of events at approximately the same deformation level) identifies the ith damage state. 2004) The engineer needs also to assess whether the structural members that are damaged at any stage of the NSP curve are still able to withstand gravity and live loads. OpenSees for short. With the support of the National Science Foundation. Δi. if local loss of vertical capacity at a certain damage state is likely. a steel beamcolumn connection that reaches a curvature sufficient not only to fracture the flanges but also to pull off the shear tab connection may not physically collapse (due. as a software platform for research and application of simulation for structural and 34 . The decision can be based.Figure 7 Hypothetical Static Pushover curves for the two orthogonal directions of a building.6. including whether any element has reached ultimate vertical capacity. If tests are not available either a detailed finite element analysis of the member (or connection) or simply engineering judgement need to be used. DSi. Analysis in OpenSees can be performed with reinforcement. More formally. Legend: BS stands for base shear (Bazzurro et al.2 is used. the Pacific Earthquake Engineering Research Center (PEER) has developed the Open System for Earthquake Engineering Simulation. for example. These inelastic events are to be associated with specific damage states. As mentioned above. For example. OpenSees (Mazzoni et al. The damage description can be valuable to engineers to compare with observable damage during the inspection of a facility before deciding on the possible building occupancy restriction after an earthquake. For this purpose a structural analysis software-OpenSees1. and (b) a detailed description of the structural damage associated with that event. to membrane effects in the slab) but it cannot be considered suitable for withstanding vertical loads of any significance. for example. then there is no need to continue the NSP analysis is stopped.

The assessment of the residual lateral capacity of the building in the DSi damage state is of 35 . Network for Earthquake Engineering Simulation (NEES (http://www. and databases.neesgrid. visualization. that the structure may be in after an earthquake can be identified. all of which meet important NEES objectives.org) System Integration project as the NEES simulation component. The framework is a set of inter-related classes. which allows great flexibility in combining modules to solve simulation problems for buildings and bridges. developers and researchers are using the extensible features of the software architecture to add additional capability. The classes are as independent as possible. and data processing and communication procedures. Finally. along with different approximations of kinematics to account for largedisplacements and P-D effects. including soil and soil-structure-foundation interaction.3. The software architecture and open-source approach for OpenSees provide many benefits to users interested in advanced simulation of structural and geotechnical systems with realistic models of nonlinear behavior. The NEESgrid decision to utilize OpenSees and adapt it to interface with other NEESgrid resources provides an important capability for NEES researchers and users. in particular in aftershocks prior to repair. elements (which are hierarchical). Brown. While in that damaged condition the structure is vulnerable and could experience more extensive and severe damage in future events. As an advanced platform for computational simulation. The modular design of OpenSees means that it can be customized for the integrating physical and computation simulation through data repositories. and hybrid control for advanced experimental methods.org)). this step of the procedure requires obtaining the NSP curves for the building in every damage state. solution algorithms.berkeley. solution procedures. A second advantage is that there is a wide range of solution procedures and algorithms that the user can adapt to solve difficult nonlinear problems for static and dynamic loads. models. DSi.geotechnical systems. OpenSees provides a flexible interface to computer resources. complex systems. OpenSees provides an important resource for the National Science Foundation-sponsored George E. equation solvers.edu (http://opensees. The OpenSees software framework uses object-oriented methodologies to maximize modularity and extensibility for implementing models for behavior. and it has been adopted by NEESgrid (http://www. Jr. This life-safety threat drives the tagging of damaged buildings. the modeling approach is very flexible in that allows selection and various combinations of a number of different element formulations and material formulations. NSP curves for each damage state are needed for characterizing the residual lateral capacity left in the structure after different levels of damage that might have been inflicted on it by the mainshock. and network communication to take advantage of high-end computing systems. or very sophisticated modeling and parameters studies of large. 2. First. integrators. Structural and geotechnical models can be analyzed from desktop PC’s to parallel computers within OpenSees. The residual lateral capacity is the quantity used when assigning a damage state to a structural limit state and therefore to a tagging condition. storage and databases. and most recently including reliability computational modules.7 NSP Curves for the Damaged Structure From NSP analysis a suite of potential damage states. as illustrated in this manual. As an open-source project. and post-processing that can provide simple problem solving capability.edu). To evaluate the post-earthquake residual capacity of the building to withstand future lateral loads. Another feature is that OpenSees has a fully programmable scripting language for defining models. solution methods. such as domains (data structures). The open source software is managed and made available to users and developers through the OpenSees website at http://opensees.berkeley.nees.

Development of NSP curves for the damaged Structure needs to create a model of the structure in the ith damage state to the same NSP analysis described for intact structure. possibly. Obtaining these NSP curves by assuming parallel-to-elastic unloading/re-loading is a simple alternative that can result in comparable estimates of residual dynamic capacity (Bazzurro et al. DS7) displacement. DS2 to DS5) (Bazzurro et al. The effects of the expected (or measured) dynamic residual offsets on the residual lateral capacities are accounted to infer dynamic response. Figure 8 Hypothetical NSP curves for structure in the intact condition and at different levels of damage (i. be reduced by the residual offset... 36 . The extent of this permanent displacement is somewhat an artifact of applying a static procedure to modeling the dynamic response of the structure subject to ground shaking. one for positive and negative directions as well) for each of such damage states. the collapse displacement will. This exercise is to be repeated for all the damage states identified for which the local ability of sustaining gravity and live loads is not compromised.interest provided that at the local level every damaged structural member is still capable of carrying vertical loads.e. two (one per orthogonal direction and. the horizontal shift of the NSP curves for the damaged structures back to the origin does not shift the collapse (e. As illustrated in Figure 8. However. Note that the pushover curve for a structure at the onset-of-damage state (i. The NSP curve for DS6 is missing because it is assumed here that at that damage state some members were so severely damaged that they could not reliably carry vertical loads. no permanent displacement). Unloading the structure from a damage state may create an initial offset in the damaged-structure NSP curve due to the residual permanent displacement in the structure.e.. This process leads to a family of NSP curves. 2004) In practice the NSP curves for the structure in the damaged state can be obtained by “reloading” (re-imposing an increasing deformation on the model of the structure after it has been quasi-statically unloaded from the ith damage state. If there is loss of local member vertical capacity in the DSi damage state the building will be assigned a structural limit state that corresponds to collapse.g. The residual displacement obtained from the NSP can be considered as an upper bound because the structure is not allowed to oscillate and therefore return to a residual offset closer to its original upright position. The NSP curves for the structure in the damaged conditions are assumed to start from the origin of the axes (i.. Figure 8 shows NSP curves for the intact structure and for the damaged structure in damage states from DS2 to DS5 in the E-W direction (Figure 7). 2004). DS1) coincides with that of the intact structure.e. in effect.

Still. Based on the established principle of using SDOF oscillators to approximate MDOF systems. while they featured backbones ranging from simple bilinear to complex quadrilinear with an elastic. Al-Sulaimani and Roessett 1985. almost instantaneously recreating the fractile IDAs in normalized coordinates of strength reduction factor R versus ductility μ. Having compiled the results into the SPO2IDA tool.. Miranda 2000).NSP techniques are applied to the damaged structure in all damage states DSi for which vertical load carrying capacity is not compromised anywhere in the structure. but they use oscillators with bilinear backbones that only allow for elastic perfectly–plastic behavior.. Fajfar and Gaspersic 1996. A family of NSP curves that are associated with different levels of initial structure damage state DSi for which local vertical capacity is preserved everywhere in the structure is obtained. the SPO to IDA connection has been investigated for simple oscillators.g. an engineer user is able to effortlessly get an accurate estimate of the performance of virtually any oscillator without having to perform the costly analyses.3. FEMA 1997). At the core of performance-based earthquake engineering (PBEE) lies the accurate estimation of the seismic demand and capacity of structures. coupled with empirical equations derived for such oscillators to rapidly obtain a measure of the seismic demand (e. moderate. (FEMA 2000)] and offers thorough demand and capacity prediction capability. Fajfar and Fischinger 1988. One of the promising candidates is incremental dynamic analysis (IDA) (Vamvatsikos and Cornell 2002). Unless one of the two main orthogonal directions of the structure is significantly weaker than the other. provide the ability to accurately approximate the full IDA and investigate the connection between the structure’s SPO curve and its seismic behavior. professional practice favors simplified methods. available online (Vamvatsikos 2002).g. when suitably applied to the MDOF SPO. Federal Emergency Management Agency.. and 84% fractile IDA curves which were in turn fitted by flexible parametric equations (Vamvatsikos and Cornell 2005). a task that several methods are being proposed to tackle. which. 50. a hardening and a negative-stiffness segment plus a final residual plateau that terminated with a drop to zero strength. and occasionally positive or negative post yield stiffness (e. empirical relations for full quadrilinear backbones are readily available.8 Incremental Dynamic Analysis (IDA) Fundamentals and SPO2IDA The IDA (Vamvatsikos and Cornell 2002) involves performing a series of nonlinear dynamic analyses of a structural model for multiple records by scaling each record to several levels of intensity that are suitably selected to uncover the full range of the model’s behavior: from elastic to yielding and nonlinear inelastic. The SDOF systems studied were of short. finally leading to global dynamic instability.g. A typical example of applying SPO2IDA appears in Figure 9. The oscillators were analyzed through IDA and the resulting curves were summarized into their 16. by using a series of nonlinear dynamic analyses under suitably multiply-scaled ground motion records. 2. in regions ranging from elasticity to global dynamic instability. 37 . and long periods with moderately pinching hysteresis and 5% viscous damping. mostly using single degree of freedom (SDOF) models that approximate the multi degree of freedom (MDOF) system’s behavior by matching its static pushover (SPO) curve. analyses are performed for both directions. a computer-intensive procedure that has been incorporated in modern seismic codes [e. With the emergence of the SPO2IDA software (Vamvatsikos and Cornell 2005). Nassar and Krawinkler 1991. Such procedures could be extended to reach far into the nonlinear range and approximate the results of IDA.

Figure 9 16%, 50%, and 84% fractile incremental dynamic analysis (IDA) curves for T=0.92 s oscillator with the displayed backbone, estimated using both IDA and static pushover 2 incremental dynamic analysis (SPO2IDA) software (Vamvatsikos and Cornell 2005)

**2.3.9 Inferring Dynamic Response from Static Response
**

The nonlinear dynamic behavior of the structure in state is estimated here using the NSP Curve and a (Microsoft Excel) spreadsheet SPO2IDA (Vamvatsikos and Cornell, 2001b; Vamvatsikos and Cornell, 2002) software. The dynamic response of the intact structure is evaluated for a full range of potential mainshock ground motion levels. The result of this operation, which is the dynamic counterpart of the static pushover curve, is called the Incremental Dynamic Analysis (IDA) curve (Vamvatsikos and Cornell, 2001a). The use of the SPO2IDA tool requires that the NSP’s used as input be normalized in both axes (i.e., base shear, BS, vs. roof drift, Δ) by the corresponding quantities at the onset of damage (or incipient yielding) of the structure. The engineer can use his/her judgment to identify an appropriate normalization point that can be associated with onset of damage or incipient yielding in the structure. Selecting an accurate yielding point is desirable but not crucial. It is, however, important that both the normalizing base shear, BSy, and roof drift, Δy, values be consistent with the same selected point on the pushover curve. However, the IDA’s, which constitute the output of the SPO2IDA tool, have the Y axis expressed in terms of the ratio of a ground motion parameter rather than of base shear. The parameter selected here is the spectral acceleration, Sa(T1), at the fundamental oscillatory period of the intact structure, T1, and the normalizing quantity is Say(T1), the spectral acceleration at incipient yielding. Of course, the two ratios in terms of base shear or of spectral accelerations are numerically equivalent. This change of variables on the Y axis for the IDA case is important in this procedure because the tagging strategy is based on exceedance of ground motion spectral acceleration rather than base shear. To facilitate the comparison of IDAs corresponding to different damage states, their ordinates should be “de-normalized” (namely, multiplied by Say(T), where T is equal to T1 for the intact structure IDA and to TDSi for the IDA of structure in damage state DSi). Under the assumption that the response of structures is dominated by the first mode, Sa(T1) for the intact structure (or Sa(TDSi) for the structure in damage state DSi) can be obtained simply by dividing BS by the structure effective modal mass for the fundamental vibration mode, M1* = M X Ψ1

38

Where Ψ1 = first-mode participation factor M = mass If the contribution to the response from higher modes is significant then more modal participation factors come into play. A detailed procedure on how to transform base shear into spectral acceleration can be found, for example, in the HAZUS manual (1999). The procedure above applied to the structure in different states of damage yields, however, denormalized IDAs that, in general, are expressed in terms of spectral accelerations at different oscillatory periods. Intact structure An IDA curve is generally considered a tool for predicting the dynamic median deformation (e.g., inter-story drift or roof displacement) of a structure for a given level of ground motion severity (here measured by Sa(T1)) (Vamvatsikos and Cornell, 2002). Here, the IDA curve is used in the opposite direction, namely to obtain the mainshock median ground motion level, Sa, that is expected to cause the intact structure to end up in each of the identified damage states, DSi. Such Sa values can be read off the IDA curves at the roof drift values Δi associated with such states. Figure 10 shows a typical output of the SPO2IDA tool where the NSP curve (or, more precisely, its quadrilinear fit, as discussed in the commentary) is the input and the IDA curve is the output. Note that in Figure 10 the deformation measure on the X axis is the roof global ductility ratio, μ, which is the ratio of the roof drift to the roof drift at the onset of damage or incipient yielding. Further the IDA curve defines the global collapse capacity (indicated by a circle in Figure 10) of the intact structure. The basis for IDA curves is described in the commentary. It is important to emphasize that the IDA curves are to be considered as median curves.

Figure 10 Hypothetical NSP and IDA curves for the building in the intact condition. In this case the abscissa represents the global ductility ratio, μ (namely, the roof drift divided by the roof drift at first yielding, i.e., at DS1). The ordinate R is equal to BS/BSy for the NSP curve and to Sa(T1)/Say(T1) for the IDA curve. Note: the dashed line is an example of quadrilinear fit of the pushover curve, as explained (Bazzurro et al. 2004)

39

Damaged structure IDA curves are also part of the procedure applied to the damaged structure. In this case, however, IDA curves are not used for predicting the mainshock median Sa (given a roof drift value, Δi) that will cause the intact structure to end up in damage state DSi, but rather for estimating the median residual dynamic lateral “capacity”, (Ša,cap)i, of the damaged structure in each post-mainshock damage state DSi to resist aftershocks. (Ša,cap)i is the (aftershock) ground motion level that is expected to cause subsequent collapse and, should the building still be occupied, life loss. The same SPO2IDA spreadsheet provides an estimate of (Ša,cap)i as one of the points of the IDA for the damaged structure (Figure 11 shows an example for damage state DS3). This capacity is defined either as the condition where small increases in “load” during aftershocks will cause arbitrarily large increases in deformation (i.e., where the IDA curve becomes effectively flat; this definition was also adopted by the SAC/FEMA 350-352 Guidelines (2000)) or as the aftershock spectral acceleration that causes local collapse anywhere in the structure. The NSP curves for the damaged structures have been shifted to remove the static residual displacement offsets created by the NSP procedure. The IDA curves derived from those NSP curves, therefore, assume no residual offset. As one might expect, though, the residual offset after a mainshock influences the residual capacity to resist aftershocks, (Š a,cap)i (e.g., Luco et al., 2004;). In these guidelines, an IDA curve for each damaged structure that accounts for the residual offset is obtained by tracing the IDA curve that assumes no residual offset up to the point that has a displacement equal to the displacement capacity (defined in the preceding paragraph) minus the expected (or measured) dynamic residual offset (defined in the commentary). Equivalently, the collapse displacement of the NSP curves (e.g., the displacement at DS6 in Figure 8) can be reduced by an amount equal to the expected (or measured) dynamic residual offsets and these NSP curves that account for the residual offset can be input into the SPO2IDA spreadsheet. An example of the resulting IDA curve is shown in Figure 11. Even without considering a residual offset the IDA curves for the damaged structures will generally provide estimates of the residual capacities, (Ša,cap)i, that are not equal to that of the intact structure, namely Š a,cap=(Ša,cap)1 (the estimates of the median capacity for the intact structure and for the damage state DS 1, which defines the Onset of Damage state, are the same). In contrast, dynamic analyses of damaged structures (e.g., Luco et al., 2004) indicate that without a residual offset little or no reduction in the median lateral capacity relative to intact structures is expected, at least for the structures analyzed here. To reflect this observation, an adjustment to (Ša,cap)i is made in these guidelines. The adjustment is achieved simply by multiplying (Ša,cap)i by the ratio of the capacity of the intact structure to the residual capacity assuming no residual offset. Hence, if no residual offset is expected (or measured), the adjusted residual capacity will equal (Ša,cap)1. The NSP curve needs to be obtained for all the post-earthquake damage states in which none of the structural elements has reached its ultimate vertical capacity to withstand gravity and live loads during the mainshock. All the damage states that do show loss of local vertical capacity will be automatically associated with collapse.

40

If loss of vertical capacity is observed. the ordinate R is equal to BS/BSy for the NSP curve and to Sa(TDS3)/Say(TDS3) for the IDA curve (Bazzurro et al.10Occupancy Status for Damaged Structures The criteria that follow are used to recommend the tagging condition (i.Figure 11 Estimate of the residual median capacity (Ša. the dynamic response of the structure in its intact condition is estimated from the corresponding NSP curves. Figure 12 IDA curves for the intact structure and for the structure at different levels of damage (after accounting for the expected. Local collapse potentially represents a life41 . or measured. Typical IDA curve for a structure in the intact condition is shown as red line in Figure 10.. for which no loss of local or global vertical capacity is observed during the mainshock. DSi. then that damage state is associated with collapse.3. As in Figure 10.cap)3 for damage state DS3 (after accounting for the expected. residual offset after the mainshock and adjusting for the behavior observed in dynamic analyses of damaged structures).e. the limit state) to the structure in any specific damage state. or measured. The circles represent the global collapse residual capacity of each case. 2004) 2. All the IDA’s have been de-normalized and scaled to the spectral acceleration at the same oscillatory period before including them in the same plot (Bazzurro et al. residual offset after the mainshock). 2004) Using the SPO2IDA spreadsheet.

The alternative tagging criteria account for the increased probability of collapse from pre. can be obtained using Probabilistic Seismic Hazard Analysis (PSHA) of building-site. a time-varying quantity that decreases with time elapsed from the mainshock and.. the building-site-specific mean annual frequency (MAF) of exceedance of the ground motion corresponding to the median capacity. The figure has two scales for the ordinates. which refers to the pre-earthquake conditions. The proposed primary tagging criteria are displayed in graphical form in Figure 13. are based more explicitly on the aftershock hazard as is more consistent with the intent of the procedure. are expressed in terms of: P0. A damaged building that was made unstable by the mainshock may in fact become progressively unsafe or may (partially or globally) collapse under its own weight and live loads without the occurrence of any aftershock (e. 2004.cap)1. The values of the quantities in Figure 13and in the text are proposed for coastal coastal California sites at Bazzurro et al. The alternative criteria. the percentage of loss in Ša. discussed later in the commentary. it is better computed using an aftershock PSHA (APSHA) approach. Estimates of Ša.cap and the ratio of P/P0 that measures the increase in frequency of exceeding the median residual capacity of the building damaged by the mainshock. therefore. ATC TechBrief2. P0. The proposed criteria for tagging damaged buildings.to post earthquake conditions due both to decreased capacity and to increased (aftershock) seismicity. is the building-site-specific MAF of the aftershock ground motion corresponding to the median capacity. as a proxy for evaluating the post-mainshock or “aftershock hazard”.cap)i for all the damage states DSi’s. (Ša. (b) Average relationship for loss of ground motion capacity and rate of increase in 42 . (Ša.threatening condition. Figure 13 (a) Graphical representation of the recommended tagging criteria.g. which is more familiar and more readily available. such as that presented in Wiemer. P. in whichever damage states they may be. see also Yeo and Cornell. P is. 2004) the primary criteria proposed here implicitly assume that one can use the pre-mainshock hazard or MAF. 1999). of the building in the damage state DSi. 2000and Yeo. of the building in its intact conditions. (For an application of APSHA.cap)i. strictly speaking. 2004 and should be reevaluated for other sites and structures of different importance.cap for the intact building and of (Ša.

We selected low values to implicitly and partially account for the increased aftershock hazard that the damaged building is subject to when the inspection may take place perhaps one or two days after the earthquake.e. under-designed buildings that are potentially unsafe (i. Tagging Y or R an undamaged building would be difficult to accept by owner and occupants. for example. and 40%. MRP. that a rather widespread damage in the building is needed before the capacity drops by these amounts.. If the intact building had been much weaker or in a harsher seismic environment such that its value of P0 were equal. Results from dynamic analyses (Luco et al. 43 . the primary tagging criteria explicitly consider the likelihood. In cases of widespread damage the assessment of the true building capacity is more uncertain and it is conservative to force some restriction of occupancy until further more detailed analyses are performed. then a nominal loss of lateral capacity of only 2% or larger would cause the building to be redtagged. to 1 x 10-3. Conversely a much stronger building with P0 smaller than 2 x 10-4 would need to lose 20% of its original capacity before being tagged Yellow and 40% to be tagged Red. the green. are delimited by horizontal lines drawn somewhat arbitrarily at constant values of capacity loss of 2%. respectively. the tag would be either green or red. and red tag areas in Figure 13 would be oblique bands delimited by straight lines of constant P values. “Weak”. P. or a combination of both. The diagonal bands in Figure 13.e. which represent quantitative measures of acceptable risk.475 years) and 5% in 50 years (MRP of 975 years). we selected P values to be equal to 2% in 50 years (mean return period. A larger value of P0 implies that the building is either relatively “weak”. 20%. Explanation: As stated above. however. for example.mean annual frequency of exceedance of ground motion for coastal California sites for which the absolute value of the (log-log) slope of an average ground motion hazard curve in the surroundings of 10-3 annual frequency of exceedance is about three (After Bazzurro et al. How the color of the tag changes with capacity loss can be found by searching on a vertical line at that specific value of P0. to 3 x 10-4 needs to be damaged severely enough to loose about 5% of its initial capacity before it is tagged Y and about 30% before it is tagged R. No yellow tag could be assigned in this case. however. 2004) have shown. Any building is identified by a particular value of P0 that can be computed during “peace” time before any earthquake has occurred. retrofitted unless some identifiable physical damage occurred in the building. These values. A hardly detectable capacity loss of only 2% encourages Y and R tagging and it is meant here to simply serve as a trigger for action for such buildings. For these guidelines for the boundaries between green and yellow tags and between yellow and red tags. that the damaged-building capacity is exceeded but the computation of P is performed for simplicity using pre-mainshock PSHA. These values may appear too restrictive when compared to building code requirements for new buildings that prescribe life safety as performance objective for a 10% in 50 years (i. in general. or that it is located in an area of higher seismic hazard compared to that considered in the original design. If the proposed tagging criteria were simply based on maximum acceptable collapse risk of the partially damaged building.. should be modified according to the building importance and severity of failure consequences. that increase the Y and R tagging areas of “strong” (smaller P0) structures.. MRP of 475 years) ground motion level. The lines at 20% and 40%. larger P0) even in pre-earthquake condition would not be tagged Y or R and possibly. yellow. Therefore a building whose P0 is equal. The opposite is true for lower values of P0. have been dictated by a different concept. at a later stage. of 2. 2004) The criteria can be summarized as follows.

cap)1<(Ša. Similarly. The illustrative example shown in Figure 14 refers to a structure located at a site for which P0 ≈ 2. yellow. which corresponds to incipient (local and global) collapse limit state is always red tagged.cap)1 and (Ša. because 0. Similarly.85x(Ša. CF. The increased knowledge about building capacity deriving from more detailed inspection and further engineering analyses may also call for the removal of the conservative lines at 20% and 40% capacity drops in the criteria applicable at a later stage. 44 .cap)3<0.65x(Ša. Under certain tenable assumptions the former MAF can be computed by multiplying the latter by a “correction factor”. which is meant to be applicable immediately after the mainshock. For this pre-mainshock hazard level the structure is yellow-tagged if it looses about 15% of the original capacity and it is red tagged if it looses about 35% of the original capacity. Figure 13could still be used as is with this alternative interpretation. Y.cap)2. the only change would be in the value. and R areas at higher values than those displayed in Figure 7. respectively. This entails considering the ground motion capacity as a random variable as opposed to a constant. To mimic the decreasing aftershock hazard with time one could devise criteria that are less stringent as time goes by. which will be larger than the MAF of exceeding the median ground motion capacity currently used. P0 ’. This would be reflected by a change in the constant-P lines that separate green. yellow. As a first approximation. and red tag regions. which shows a residual capacity. of the abscissa to be used to enter the graph in Figure 13. (Sa. For a particular application. CF can be considered to be the same for the capacities in both the intact and the damaged cases and therefore it does not have any impact on the values of the oblique lines of constant P/P0 that divide green. which is within the green tag state boundaries regardless of the value of P0. Criteria applicable.It should be noted that a collection of figures such as Figure 13could be produced to account for the time-varying nature of aftershock hazard.cap)4<0. (Ša.cap)1. larger than one that accounts for the aleatory and epistemic uncertainties in the ground motion capacity (called β in the section to come) and the slope of the hazard curve in the neighborhood of that ground motion level. It is conceptually preferable to develop tagging criteria in terms of MAF of building collapse rather than MAF of median ground motion capacity.2x10-4. is green-tagged. one month after the mainshock may have constant P lines demarcating the G. and red tagging areas. the damage states DS3 and DS4 are tagged Yellow and Red. the damage state DS6.65x(Ša. greater than 85% of the capacity of the intact structure. The damage state DS1 can be associated with the onset of damage (OD) limit state. Such tagging criteria could be applicable after the earthquake at different snapshots in time when more detailed inspections and/or improved capacity analyses may take place. for example.cap)1. In this case the damage state DS2.

are not readily available because the damage states DSi (and.3. 2004) Criteria based on global residual capacity reduction and site specific aftershock ground motion hazard are applied for mapping damage states into structural limit states. for all practical purposes. or Coll. Figure 15 shows how this information can be used to determine the median spectral capacity value.Figure 14 Tagging of different damage states within the illustrative example.e. The values of ŠaY and ŠaR. The process is shown in Figure 15. This drift is that associated with the inception of the limit state that was sought. this IDA-based procedure has identified the median spectral acceleration value corresponding to the onset of a given structural limit state. From Figure 14 it follows that for this illustrative example the damage state DS1 correspond to incipient OD. In this case either the procedure is repeated for one or more intermediate DS’s until the computed spectral acceleration capacity is. red. onset of damage. the value of ŠaG is zero. 2. The yellow tagging given to DS3 assumes that the P0 of exceeding the spectral capacity (Ša.. LS. The onset of R. any level of ground motion will generate a G tag or worse.g. yellow. collapse). Yellow tag (Y). however. denoted ŠaLS for limit state LS (e.cap)1 at the building site is about ≈ 2. For any given P0 the loss in capacity that corresponds to the target boundary between two limit states is known from Figure 13. or. spectral acceleration) at which the limit state (and associated tagging) 45 .2x10-4 (Bazzurro et al. respectively) by reading them off the IDA for the intact structure in Figure 9 at the drift value of ΔOD and ΔC. the interpolation procedure will provide the target drift that corresponds to the desired drop in capacity. therefore.11Ground motion level associated with a structural limit state The proposed tagging approach coupled with the IDA curve for the intact structure leads to the identification of the median spectral capacity values associated with the onset of postearthquake tagging status.. alternatively but less accurately. Therefore. DS1 and DS6. R. respectively. DS4 and DS5 are in Red. Y. and DS6 refers to incipient Collapse. an interpolation scheme is used instead. Red tag (R) and Collapse (C) are obtained. The value of ŠaLS can be found immediately for OD and C whose onset correspond to one of the damage states (here. A one-toone map of each damage state DSi into one of the following four limit states: Onset of Damage (OD). DS2 is in Green. for example. LS equal to OD. ŠaLS. The ground motion intensity (i. DS3 is in Yellow. Identifying the median spectral capacity value for the green tag state is not necessary for reasons that will discussed later. occurs for a damage state in between DS3 and DS4 (Figure 14). for any limit state. Finally. Δi) from the NSP curve in general do not correspond to the inception of the Y and R limit states. As mentioned earlier. reasonably close to the target threshold.

depends on the initial period of vibration of the structure. at some expense. etc.will occur cannot. Aleatory uncertainty is intrinsic in the random. namely short-period (high-frequency) structures show more record-to-record variability in their displacements (or correspondingly. and it is assumed therefore that there is a 50-50 chance that the limit/tagging state (or worse) will be observed if this ground motion occurs at the site. more testing of in-place material properties. one to the Red. The epistemic kind of uncertainty stems. The value just identified is a “best guess”. Further the response dispersion is larger for larger degrees of nonlinearity. Miranda. 2000) that the aleatory portion of the dispersion. this second type of uncertainty can be reduced. one corresponding to collapse. from both the limited accuracy of the selected response analysis approach and the imperfect knowledge of parameter values of the adopted mathematical model of the structure.g. in the variability in the dynamic displacements produced by different ground motion records (even though they may have the same intensity value). especially when the ground motion is near the collapse capacity. more refined models. for example.. Within the limits associated with current scientific knowledge. There is a smaller chance at lower ground motion levels and a larger chance at higher levels. with more detailed investigation of the structure. for example. it is known (e. This difference is captured in the three curves presented in Figure 16. and one to the less severe Onset-of-Damage limit state. 46 . βR. Figure 15 Median spectral acceleration capacity associated with all the structural limit states except the green tag state (After Bazzurro et al. which is a combined measure of two basic kinds of uncertainty: aleatory uncertainty (or randomness) and epistemic uncertainty. 2004) These chances are first quantified by estimating the dispersion. β. in the spectral acceleration at which a given displacement or damage state is first observed). The former kind manifests itself.and Yellow-tag conditions. unpredictable nature of earthquakes and cannot be reduced. Aleatory uncertainty (βR) Based on the results of many nonlinear dynamic analyses. the values of βR follow the trend shown in Figure 16. be predicted perfectly. in fact. On average.

Very limited research has been devoted to date to a systematic estimation of the epistemic uncertainty in the structural nonlinear deformation caused by a given ground motion or. older structures (e.g. reflects the professional confidence that the selected model and the analysis procedure will predict accurate results. For example. βU. in a SMRF a center-line model yields less reliable results than a model with panel zones). When applying the procedure in these guidelines. in the ground motion necessary to induce a given structural deformation.g... the provided SPO2IDA spreadsheet will provide structure-specific βR.. values for each limit state that should be used in place of those shown in Figure 16. they are suitable as βR default values.25 to 0. one will have less confidence) for complex..g.35 for the Immediate Occupancy Performance Level and from 0. Yun et al.g. the SAC project (e.The curves in Figure 16 are based on average results of nonlinear dynamic analyses of many nonlinear oscillators with different NSP backbone curves and to somewhat arbitrary definitions of limit states based on average structural deformation levels. however. (After Bazzurro et al.60 for the Collapse Prevention Performance Level.15 to 0. mill-type buildings) modeled using a simplified model and untested material properties than for a clean. the values of βU for SMRF structures in Table A-1 of FEMA 351 range from 0.. Although limited to SMRF structures.. Hence. see Appendix A in FEMA 351. 47 . 2004) Epistemic Uncertainty (βU) The epistemic part of the uncertainty. In the SAC reference βU is intended to represent mainly the epistemic uncertainty in the estimation of demands and capacities. modern steel moment-resisting frame whose properties have been well determined and whose model has been developed with extreme care to details. the values for βU will be larger (i. 2000) from the uncertainty in each separate component.e. Similarly the confidence in the results increases with the level of effort in the structural modeling (e. Figure 16 Recommended (default) values for βR. Although equations are provided to compute βU values (e. dually. FEMA 351 and FEMA 355F) is the most widely available reference for gathering information on βU values.

0.05. the latter controls its slope (the larger the β value. respectively.65}. for the generic structural limit state LS is determined by plotting the values of probability p = {0. to be used in the determination of the fragility curve (STEP 6) is the SRSS value: 2 2 β = β R + βU 2.65. green. for example.67. the curve’s estimation for the generic structural limit state LS requires two parameter values. or collapse state) provides the annual probability that the intact building will end up in the specified limit state (or worse) given the occurrence at the site of an earthquake ground motion of intensity Sa. Based on the common lognormal assumption. β. This is because. and for collapse of the building (After Bazzurro et al. -0. green. the fragility curve provides the likelihood that the building will be in the given limit state or worse if a ground motion with specified Sa were to occur at the site. 2004) It should be noted that in Figure 17 the fragility curve for the green-tag state is equal to unity for all values of ground motions.g. FLS(Sa). is denoted as FY(Sa).12Computation of the Fragility Curves The fragility curve (Figure 17) for a given structural limit state LS (LS equal to onset of damage.Total Uncertainty (β) Finally the value of net dispersion.5. and red tags. Values of ŠaLS and β are provided in STEP 5. yellow. 0.67 and 1. 0. 1970). β. Sa: S a = S aLS e xβ (3) for the values of x equal to {–1. a median ŠaLS value and a measure of dispersion. yellow. 0. The fragility curve. Figure 17 Fragility curves for onset of damage. Benjamin and Cornell. The fragility curve. red-. the flatter the curve). Because there are no structural limit 48 . Additional values of p and x can be found in any table of the Gaussian distribution function (e. The former is the central value of the curve that correspond to an exceedance probability of 50%. for the yellow-tag state.75 and 0.. 0.95} versus the corresponding values. The former parameter is referred to as the median spectral capacity value of that limit state.3. 0. by definition.25.

The opposite is true for the fragility curve corresponding to the collapse state. it follows that the building will have a green tag or worse with certainty for any level of ground motion. 49 .states less severe than green. Note that fragility curve for onset of damage (OD) is the steepest of the four because the value of β is the smallest.

according to the predictive relationship used in the analysis. This study should help to update the seismic map of Bangladesh National Building Code.1. A four-step process (Reiter. Chittagong and Sylhet city and Peak ground acceleration (PGA) contour map of Bangladesh. Boore expression can be used with amplification factor for alluvium. In deterministic seismic hazard analysis (DSHA). The history of earthquakes. 1990) is adopted in the analysis. 1978 (equation 1). Mechanics of PSHA is 50 . 1989 (equation 2) and Boore et al.1 Deterministic Seismic Hazard Analysis Earthquake scenario assumed in DSHA. uncertainties in earthquake size. location and time of occurrence are explicitly considered This chapter presents the deterministic seismic hazard analysis of Dhaka city and probabilistic seismic hazard analysis of Dhaka. As bed soil of Bangladesh is alluvium McGuire and Duggal expression is more applicable. The objective of earthquake-resistant design is to install a structure or facility that can withstand a certain level of ground shaking without excessive damage. 1993 (equation 3) acceleration attenuation expression is used to predict PGA. quantified and combined in a rational manner to provide a more complete picture of the seismic hazard. location and rate of recurrence of earthquakes and in the variation of ground motion characteristics with earthquake size and location can be identified.3SEISMIC HAZARD ANALYSIS OF BANGLADESH Seismic hazard analyses involve the quantitative estimation of ground-shaking hazards at a particular site. in Bangladesh is sufficient enough to require their careful consideration in the design of structures and facilities. Boore expression is developed for rock and McGuire expression is developed for both rock and alluvium and Duggal expression is developed for alluvium. Source characterization includes definition of each source's geometry (the source zone) and earthquake potential. a particular earthquake scenario is assumed and in probabilistic seismic hazard analysis (PSHA). 3.1 METHODOLOGY 3. The distance is expressed as an epicentral distance or hypocentral distance. “Peak Ground Acceleration” or abbreviated as “PGA”. The controlling earthquake that is expected to produce the strongest level of shaking is described in terms of its size (usually expressed as magnitude) and distance from the site. The probability of occurrence of the controlling earthquake is assumed to be one at the points in each source zone closest to the site and zero elsewhere. 3.2 Probabilistic Seismic Hazard Analysis PSHA provides a frame-work in which uncertainties in the size. Duggal. Earthquake sources capable of producing significant ground motion at the site are identified and characterized from historical earthquakes. consists of the postulated occurrence of an earthquake of a specified size occurring at a specified location. That level of shaking is often described by a design ground motion parameter. The shortest distance between the source zone and the site of interest is selected.1. Two types of quantitative estimation of ground-shaking hazards at a particular site are involved in the seismic hazards analysis – one is deterministic and another is probabilistic. McGuire. Peak ground acceleration (PGA) is used to characterize the seismic hazard.

(1993) relationship separately. 51 . the uncertainties in earthquake location. the spatial uncertainty must be described with respect to the appropriate distance parameter. The ground motion produced at the site by earthquakes of any possible size occurring at any possible point in each source zone is determined with the use of McGuire (1978).00º E and 97. The uncertainty inherent in the predictive relationship is also considered.0 is 1410 and the time range is from 1869 to 2000 (131 years). Characterization of an earthquake source requires consideration of the spatial characteristics of the source and of the distribution of earthquakes within that source. About one thousand of earthquakes occur in this portion of fault line within 100 years.002 annual probability of exceedence are picked and used to plot the PGA Contour Map of Bangladesh for ten percent probability of exceedence in 50-years period.00º N and longitude 85. and of the distribution of earthquakes with time. The PSHA is also done using the procedure of four steps of Reiter. however.11º N and 30.) will be exceeded during a particular time period.based on probability theory.00º E.3 degree are developed. PGA of 0. and ground motion parameter prediction are combined to obtain the probability that the ground motion parameter (PGA is used as the parameter in this study. 3. is identical to those of DSHA. The earthquake database used in the analysis is obtained from Sharfuddin. In PSHA identification and characterization of earthquake sources. Earthquakes are assumed to be uniformly distributed within a particular source zone. Finally. (2001). (1990). The non-uniform distributions may be used when sufficient information to justify them exists. According to the earthquake database. 3. of the distribution of earthquake size for each source. Duggal (1989) and Boore et al. The epicenters of the earthquakes are in between latitude 18. Fault line between Australian plate and Eurasian plate passes by the north and south side of Bangladesh at a distance about 150 km to 200 km from border.001 annual probability of exceedence are picked and used to plot the PGA Contour Map of Bangladesh for five percent probability of exceedence in 50-years period. The uncertainty in source-to-site distance can be described by a probability density function. Since predictive relationships express ground motion parameters in terms of some measure of source-to-site distance.3 Development of PGA Contour Map Hazard curve of different latitude and longitude with grid interval of 0. often translate into a uniform distribution of source-to-site distance. Figure 19 shows the distribution of epicenter of the earthquake considered for this purpose. except that the probability distribution of potential rupture locations within the source must also be characterized. PGA of 0. Each of these characteristics involves some degree of uncertainty. earthquake size. So north and east portion of Bangladesh is at high seismic risk zone. The proper performance of a PSHA requires careful attention to the problems of source characterization and ground motion parameter prediction and to the mechanics of the probability computations. A uniform distribution within the source zone does not.1.2 DATABASE Figure 18 shows the earth’s crustal plates. the number of earthquakes equal to and bigger than magnitude 4.

at and near Bangladesh. occurrence period 1869 to 2000 52 .com) Figure 19 Location of earthquake epicenter.Figure 18 Earth’s crustal plate (courtesy of www.accuracyingenesis.

When sufficient data are available. Analysis of aftershocks can also aid in the delineation of earthquake source zones. will decrease the data set for regression. in particular areas. These may take the form of geologic and tectonic evidence.37 E 91. To estimate the seismic hazard in any particular site within a region requires a selection of earthquakes that affect significantly the value of the hazard output. the alignment of instrumentally located epicenters or hypocenters indicates the existence of earthquake sources. The written historical record extends back only a few hundred years.4 IDENTIFICATION ANALYSIS OF EARTHQUAKE SOURCES FOR To evaluate seismic hazards for a particular site or region. a geographic pattern of historic epicenters provides strong evidence for the existence of earthquake source zones. the maximum intensity can be determined and used to estimate the location of the earthquake epicenter and the magnitude of the event. naturally. earthquakes of magnitude equal to and bigger than five are selected and plotted as larger dot as shown 53 . they can also be used to evaluate the rate of recurrence of earthquakes. On the other hand. for which they have been available. Their most significant limitation is the short period of time.3. Although the accuracy of locations determined in this way depends strongly the rate of earthquake recurrence. To identify earthquake sources for analysis. Uniform probability distributions are assigned to each source zone implying that.84 E Latitude 23. Instrumental recordings represent the best available information for the identification and evaluation of earthquake sources. 2001).73 N 22. Again. A basic assumption of PSHA is that the recurrence law obtained from past seismicity (α and β of b-line) is appropriate for the prediction of future seismicity. a too large area may include earthquakes that do not affect the seismic hazard in the site and are thus useless. This. Since historical records are dated. There is no strict rule for selecting the maximum epicentral distance to the site.86 N Name of the Sites Dhaka Chittagong Sylhet 3. A small area around the site results in a smaller number of earthquakes to be considered and some events outside the zone considered may affect the hazard in the site.32 N 24. Historical accounts of ground-shaking effects can be used to confirm the occurrence of past earthquakes and to estimate their geographic distributions of intensity. or seismicity.3 SITE SPECIFICATIONS: Position Longitude 90. Identification of seismic sources requires some investigative work. compared with the average period of time between large earthquakes.82 E 91. To characterize the seismicity of each source zone. or historical (preinstrumental) seismicity. all possible sources of seismic activity must be identified and their potential for generating future strong ground motion evaluated. earthquakes are equally likely to occur at any point within the source zone. Gutenberg and Richter (1944) law for earthquake recurrence is used. Significant earthquakes are equally likely to occur anywhere in the area of 200 km in radius surrounding the sites under consideration (Sharfuddin.

0 Source 3 A3 76537 70 1093.45 15 4. The sources and their parameters are given in Table 1.7 54 4.3 5.55 622 4.in Figure 20.5 Source 5 A5 105900 112 945.3 Source 7 A7 57774 56 1031. of Earthquake in database Minimum magnitude (Ms) Maximum magnitude (Ms) Source 1 A1 21158 18 1175.0 7. Figure 20 Earthquake sources identified for analysis (after Noor.0 8.0 7.6 Source 4 A4 14434 15 962. Seven area sources are assumed as shown in Figure 20.0 7.0 Source 2 A2 28494 33 863.0 6.5 Source 6 A6 42227 40 1055.0 8. Yasin and Ansary 2005) Table 1 Parameter of the earthquake sources Source parameter Area (sq km) No.4 277 4.5 72 4.6 54 . of division Area of each division (sqkm) No.27 17 4.7 11 4.

Again when the size of division is large the contour is lumped at the center point of division in stead of being evenly distributed throughout the source. The controlling earthquake is defined as the maximum magnitude earthquake that a given source is capable of generating according to the (a) Database (b) Historical earthquakes.The sources are divided into finite number of elements. The sources are so divided that area of each division is limited to 1200 sq km. and fluctuates with negligible value at number of division > 4. The McGuire (1978). Yasin and Ansary 2005) 3.380 E. 2. number of division curve is shown in Figure 21.5 DETERMINISTIC SEISMIC HAZARD ANALYSIS In this paper deterministic seismic hazard analysis for Dhaka city (longitude 90. (1993) acceleration attenuation expression is used to determine the resulting site PGA from each source’s controlling earthquake.780 N) is performed.0E -05 5. Duggal (1989) and Boore et al. The annual probability of PGA>0.). McGuire (1978) acceleration attenuation expression (PGA in g) 55 . The random source is source one (A1) and the site is Dhaka city.0E -05 1. PGA (peak ground acceleration) is used to represent the ground motion.0E +00 0 5 10 N of div o ision 15 20 Figure 21 Effect of number of division on calculated probability (after Noor.2g vs.5E -05 Probability of PGA > 0.2g g 2.5E -05 1. Calculated probability increases with number of divisions up to a limit of: number of division = 4 (approax. latitude 23.0E -06 0. A parametric analysis to select suitable number of division of a source is done.

Using equation (1).061g. Source A1 A2 A3 A7 Nearest Eq.437 25. h = Hypocentral depth of earthquake in km = 30 km for Bangladesh.0 7. 3.025 0. (1993) acceleration attenuation expression (PGA in g) Log(PGA) = b1 + b2 ( M − 6) + b3 ( M − 6) 2 + b4 R 2 + h 2 + b5 .061 0. (1993) relationships is 0.090 0.779 104. Same method is followed for the other sources.216 b3 0 b4 0 b5 -0.6 PROBABILISTIC SEISMIC HAZARD ANALYSIS Probabilistic Seismic Hazard Analysis (PSHA) for each source is done.158 b7 0.052 0.061 0. (1993) 0.7 Rmin (km) 113.254 GB 0 GC 0 σ log PGA 0. Lat 24 N Long 90. R = Epicentral distance in km. PGA for nearest big earthquake is also calculated.174 0.0 7.174g.6 5.142 0. log( R 2 + h 2 ) + b6Gb + b7 Gc (3) b1 -0.3 E Database Mmax 8.PGA = 0.117 0.053 0.47 91.117g and Boore et al.0306e 0. Duggal (1989) acceleration attenuation expression (PGA in cm/s2) PGA = 227 ×10 0.038 b2 0.045 The largest of the resulting PGA using Duggal (1989) relationship is from the area source A1 controlling earthquake and the value is 0. Therefore the design ground motion parameter is PGA=0.6 5.2 (2) Boore et al.17 e −0. 56 .89 M ( R2 + h2 ) −1.174g and that using McGuire (1978) relationships is 0. Brief description of analysis method is given for source one (A1) and for Dhaka city. Table 2 Calculation of design ground motion parameter.043 0.104 0.027 0.05 64. (2) and (3) the peak ground acceleration of the site for the controlling earthquakes and associated site to source minimum distances are calculated and given in Table 2.308 M ( R + 30 ) −1.75 PGA (g) McGuire (1978) 0.0 for alluvium. The sources and their specifications were described in Table 1.106 Boore et al.777 b6 0.205 M = Surface magnitude of earthquake. S is chosen equal to 1.054 Duggal (1989) 0.2 S (1) Where.

0 M>7.9214-1.572 -11.5 7 7.0 M>6.022901 0. Table 3 Summary of calculation for b-line.5 -14 -14.197 -11.5 M>5.449 -13.10687 0.0 M>5.5 -11 -11.15267 0.736 -14.9 1 ← y =-1 7 x –5 2 4 Magnitude (M) Figure 22 Gutenberg-Richter b-line for source A1 ( y = -1.08E-06 1.5 Sum 71 38 20 14 4 3 3 1 Sum/Year 0.5 -15 4 4.1 8 .5 6 6.5 5 5.08E-06 3.736 -13.22E-06 5.44E-06 1.37E-05 7.839 -12.178M 57 .61E-07 ln(N) -10. Table 3 shows a summary of calculation and Figure 22 is the b-line.29008 0.5 Ln of number of earthquakes greater than M -10.178x – 5. Gutenberg-Richter b-line for source 1 (A1) Magnitude (Ms) M>4.9214) Gutenberg-Richter relationship: Ln(N) = α – βM Ln(N) = -5.5 -12 -12.196 -13.54198 0.Earthquake recurrence is expressed by Gutenberg-Richter b-line in natural log.030534 0.5 M>6.835 .0 M>4.5 -13 -13.05E-06 1.5 M>7.56E-05 1.0076336 Sum/Year/Area (N) 2.022901 0.

25)*0.β1 . P ((Mmid1 – ΔM/2) < M < (Mmid1 + ΔM/2)) = ƒ1(Mmid1).178 (8.178* e −1.0 < M < 5.0091 ƒ1(M) = 1.9214 β1 = 1. To calculate the probability that magnitude of a given earthquake falls in a given subinterval is the product of the magnitude probability density function at the center value of the subinterval and the interval increment.13631 P (5.5 │ EQ) 0.178 ( M −4.0 │ EQ) 0.075635 P (6.0 < M < 4.44274 P (4.5 = 0.0 < M < 6.178 ( 4.5 │ EQ) = ƒ1(Mmid1)* ΔM = ƒ1(4.25−4.0091*1.5 < M < 6.0 Magnitude probability density function: ƒ1(M) = c1 .0 ) The magnitude range is divided into subintervals with increments of 0. ΔM P (4.0 │ EQ) 0.5. Magnitude Range Probability P (4.0 ) *0.0 < M < 4.e-β1(M-Mo) c1 = 1 1− e − β1 ( M max − M o ) = 1 1− e −1.5 = 1.5 < M < 5.178 Magnitude limits in evaluating the seismic hazard of the area source is Mo = 4.44274 Table 4 Magnitude range and corresponding probability.0 and Mmax = 8. The probabilities are calculated and given in Table 4.α1 = 5.5 │ EQ) 0.0 ) =1.178** e −1.041967 58 .0− 4.5 │ EQ) 0.24566 P (5.0091*1.

of division of the earthquake source. of magnitude subintervals with increments of 0.5.023286 P (7.0 │ EQ) 0.205 when the constants derived for estimating the peak acceleration for the larger of two horizontal components. log( acc) − log( PGA) P (PGA > acci│ EQ : R. The acceleration range from 0. A Poisson forecasting model is employed to account for the probability that an earthquake occurs. given the occurrence of an earthquake of magnitude M and at distance R. M) = 1 – φ σ log PGA The above equation explains the probability that PGA will exceed an acceleration of interest.5 Annual probability of exceedence at different PGA is calculated and hazard curve of Dhaka city for source one (A1) is obtained as in the Figure 23.5 │ EQ) 0. (1993) attenuation relationship is 0.. ∆M = Magnitude subinterval increments = 0. (1993) acceleration attenuation expression (equation 1 and 2) with the coefficients corresponding to the largest horizontal component of the PGA.01g to 0...05g will be examined now.012921 P (7.56g by an increment of 0. given the occurrence of an earthquake of magnitude M and at distance R..P (6.0 │ EQ) 0. P ( PGA > acci │ EQ ) = ∑∑ P (PGA > acc k =1 j=1 n m i ¦ EQ : M j . Combined probability =1 − (1 − prbsource1 )(1 − prbsource 2 ).0 < M < 7. R k ) × f1 ( M j ) × ∆M / n Where n = No.5 < M < 8. Standard normal distribution forecasting model will be used in this section to determine the probability that the PGA will exceed an acceleration of interest (acci).. Same procedure is repeated for other sources and obtained probabilities are combined using the following equation of statistics. m = No..5 < M < 7. The standard error (σ) of Boore et al.0071695 The peak ground acceleration (PGA) is estimated through The McGuire (1978) and Boore et al.(1 − prbsource7 ) 59 .

(1993) McGuire (1978) and Duggal (1989) attenuation expression 0 10 10 Annual probability of exceedence P (PGA>acc) 10 10 10 10 10 10 10 10 10 HAZARD CURVES FOR DHAKA CITY Duggal (1989) McGuire (1978) Boore et al.6 0.3 PGA (g) 0. 10 10 Annual probability of exceedence P (PGA>acc) 10 10 10 10 10 10 10 10 10 0 -1 Duggal (1989) McGuire (1978) Boore et al.5 0. (1993) -1 -2 -3 -4 -5 -6 -7 -8 -9 -10 0 0.7 Figure 23 Hazard curves of Dhaka for source one (A1) using Boore et al.1 0.Combined probability is calculated and seismic hazard curve for Dhaka city is developed and shown in the Figure 24.6 0.2 0.7 60 . (1993) -2 -3 -4 -5 -6 -7 -8 -9 -10 0 0.3 PGA (g) 0.4 0.1 0.2 0.4 0.5 0.

Figure 24 Seismic Hazard Curves using all source data for Dhaka using Boore et al. McGuire (1978) and Duggal (1989) attenuation expression Following the same procedure Hazard Curve of Chittagong and Sylhet city are developed and shown in Figure 25 and Figure 26.3 PGA (g) 0. (1993).6 0. (1993) attenuation expression 61 .5 0. 10 10 Annual probability of exceedence P (PGA>acc) 10 10 10 10 10 10 10 10 10 0 HAZARD CURVES FOR CHITTAGONG CITY Duggal (1989) McGuire (1978) Boore et al. (1993) -1 -2 -3 -4 -5 -6 -7 -8 -9 -10 0 0. McGuire (1978) and Boore et al.4 0.7 Figure 25 Seismic Hazard Curve for Chittagong city using Duggal (1989).1 0.2 0.

(1993). PGA of 0.2 0.4 0. Figure 32 and Figure 35.10 0 HAZARD CURVES FOR SYLHET CITY Duggal (1989) McGuire (1978) Boore et al.1 0.001 annual probability of exceedence are picked and used to plot the PGA Contour Map of Bangladesh for five percent probability of exceedence in 50-years period.002 annual probability of exceedence are picked and used to plot the PGA Contour Map of Bangladesh for ten percent probability of exceedence in 50-years period. (1993) 10 Annual probability of exceedence P (PGA>acc) -1 10 -2 10 -3 10 -4 10 -5 10 -6 10 -7 0 0.7 Figure 26 Seismic Hazard Curve for Sylhet city using Boore et al.6 0. McGuire (1978) and Duggal (1989) attenuation expression 3. 62 . Figure 31 and Figure 34.7 DEVELOPMENT OF PGA CONTOUR MAP FOR BANGLADESH Hazard curve of different latitude and longitude with grid interval of 0.3 PGA (g) 0. (1993) acceleration attenuation expression is shown in Figure 27 and that developed using McGuire(1978) and Duggal (1989) acceleration attenuation expression is shown in Figure 30 and Figure 33 respectively. The map developed using Boore et al. The maps are shown in Figure 28.5 0. PGA of 0.0004 annual probability of exceedence are picked and used to plot the PGA Contour Map of Bangladesh for two percent probability of exceedence in 50-years period. The maps are shown in Figure 29.3 degree are developed. PGA of 0.

Figure 27 PGA (in g) Contour Map for Bangladesh Based on ten Percent Probability of Exceedence in 50-Years Period (using Boore et al. (1993) acceleration attenuation expression) 63 .

Figure 28 PGA (in g) Contour Map for Bangladesh Based on five Percent Probability of Exceedence in 50-Years Period (using Boore et al. (1993) acceleration attenuation expression) 64 .

Figure 29 PGA (in g) Contour Map for Bangladesh Based on two Percent Probability of Exceedence in 50-Years Period (using Boore et al. (1993) acceleration attenuation expression) 65 .

Figure 30 PGA (in g) Contour Map of Bangladesh Based on ten Percent Probability of Exceedence in 50-Years Period (using McGuire (1978) acceleration attenuation expression) 66 .

Figure 31 PGA (in g) Contour Map of Bangladesh Based on five Percent Probability of Exceedence in 50-Years Period (using McGuire (1978) acceleration attenuation expression) 67 .

Figure 32 PGA (in g) Contour Map of Bangladesh Based on two percent probability of exceedence in 50-years Period (using McGuire (1978) acceleration attenuation expression) 68 .

Figure 33 PGA (in g) Contour Map of Bangladesh Based on ten percent probability of exceedence in 50-years Period (using Duggal (1989) acceleration attenuation expression) 69 .

Figure 34 PGA (in g) Contour Map of Bangladesh Based on five percent probability of exceedence in 50-years Period (using Duggal (1989) acceleration attenuation expression) 70 .

Figure 35 PGA (in g) Contour Map of Bangladesh Based on two percent probability of exceedence in 50-years Period (using Duggal (1989) acceleration attenuation expression) 71 .

1989 acceleration attenuation expression is developed for alluvium soil of Japan. five and ten percent probability of exceedence in fifty years period are developed. or historical (preinstrumental) seismicity. Gutenberg and Richter (1944) law for earthquake recurrence is used. As the soil of Bangladesh is alluvium and almost similar to that of Japan the output obtained using Duggal expression is proposed. Chittagong and Sylhet city are developed and curves developed using Duggal. earthquakes are equally likely to occur at any point within the source zone.18g is proposed. Hazard curves for Dhaka.8 CONCLUDING REMARKS tectonic evidence. Deterministic seismic hazard analysis is done for Dhaka city and PGA of 0. A basic assumption of PSHA is that the recurrence law obtained from past seismicity (α and β of b-line) is appropriate for the prediction of future seismicity. 72 . 1989 expression are proposed.3. Duggal. In this analysis the historical seismicity is used and the other evidences can be incorporated when reliable data are obtained. 1989 expression are proposed. Identification of seismic sources requires some investigative work of geologic and Uniform probability distributions are assigned to each source zone implying that. To characterize the seismicity of each source zone. Contour maps developed using Duggal. Contour maps of two.

2004).. Yellow tag. earthquake intensities etc. The ground motion intensities in the fragility functions are spectral acceleration (Sa) and peak ground acceleration (PGA) incorporated with probability of exceedence. structure geometry. The limit states in fragilities are related to structure functionality status and may be defined as global drift ratio. the knowledge of the nonlinear static behavior of a specific structure subjected to incremental lateral deflection is used to infer its nonlinear dynamic response expected for different levels of ground motion severity.g. locally or more widely. maximum roof displacement. These restrictions dictate the facility effectiveness of the structure in the immediate aftermath of an earthquake and provide valuable information to estimate expected downtime. decreased seismic capacity with respect to loss of vertical resistance. until detailed engineering evaluation is completed. structure modeling and analysis procedure. G: the structure is deemed fit for immediate occupancy. Limit States The five limit states identified in increasing order of damage severity are: Green tag. peak ground motion values.4FRAGILITY ANALYSIS OF STRUCTURES A limit-state fragility curve provides the conditional probability that the specified limit state will be reached or exceeded as a function of the severity of the future ground motion (Bazzurro et al. yellow or red tagging). The mainshock damage may cause either decreased global lateral collapse capacity. OD: the onset of significant nonlinear behavior for different types of structures. Fragility curves involve uncertainties associated with dynamic response and material properties. material properties. The uncertainty inherent in building response and capacity for different ground motion levels due to variability in construction and to uncertainty in structural evaluation process is used to obtain the desired fragility curves for the different structural limit states. Y: the structure is deemed fit for restricted occupancy. A unique aspect of this procedure is that the structure safety while in a damaged state is measured in terms of its ability to protect occupants from death or injury due to aftershocks that might occur prior to its repair.1 Development of Fragility Curves The limit states in fragilities are related to loss of dynamic capacity. The access is limited to specialized personnel only. Quantitative measures of the implied degradation in structure safety is used to associate each of several post-earthquake structural damage states with an appropriate post-earthquake structural limit state that may imply some degree of occupancy restrictions (e.1. The ground motion intensities in the fragility functions can be spectral quantities. Onset of Damage. For seismic assessment of structures. 73 . to gravity and live loads. Fragility curves involve uncertainties associated with dynamic displacements produced by different ground motion records (even though they have same intensity value). interstory drift ratio.1 METHODOLOGY 4. No limitations on post-earthquake access are implied by this limit state. 4. story shear force or loss of dynamic capacity.

The deformation was increased until either the structure base shear has significantly dropped (e. retrofit or rebuilding. The NSP curves.e. The intent here is to estimate the realistic response of a structure and not to apply conservative procedure to achieve a “safe” building design. the onset of damage limit state lies within the green tag state boundaries while the collapse state is. Hence. The tagging criteria for Dhaka city used in this fragility analysis are determined through hazard curve (Figure 43) developed using Duggal. key node(s) of the structure. C: the structure has collapsed or is on the verge of global instability or local collapse. expressed in terms of base shear versus roof drift (i. roof displacement divided by the height of the building) for the weaker direction of the structure are the main products of the NSP analysis.. yellow or red tagging). The mainshock damage may cause either decreased global lateral collapse capacity.g.2 Fragility Analysis of Structures For seismic assessment of structures. of course.6. R: the structure is deemed unsafe. to gravity and live loads.1. OpenSees1. The NSP analysis should also be stopped when the value of global ductility obtained in the analysis becomes unattainable in reality. Important damage states along the displacement axis of the NSP curve(s) are identified. In this procedure the applied lateral load was allowed to be dropped. or more. to 20% or 30% or less) compared to its peak value or until at least one of the structural elements or connections has reached its ultimate capacity to withstand vertical loads even in absence of any future ground shaking. 4.g.. the most severe stage of the red tag condition. Quantitative measures of the implied degradation in structure safety is used to associate each of several post-earthquake structural damage states with an appropriate post-earthquake structural limit state that may imply some degree of occupancy restrictions (e. designed for gravity load is conducted using a displacement-controlled approach that includes P-delta effects and strength degradation of structural elements. Access is not allowed until completion of detailed engineering evaluation.. Collapse. or decreased seismic capacity with respect to loss of vertical resistance.2 (Open System for Earthquake Engineering Simulation) is used for the analysis.Red tag. STEP 1: Nonlinear Static Procedure (NSP) for the intact Structure A conventional nonlinear static pushover analysis (in the weaker direction) of a structure. An analytical software developed for earthquake engineering simulation. On the NSP curves the major inelastic events that occur in the structure (change of slope of quadrilinear approximation of SPO curve) along with 74 . The seismic evaluation is performed using expected strength values rather than nominal or lower bound strengths. The uncertainty inherent in building response and capacity for different ground motion levels due to variability in construction and to uncertainty in dynamic response are used to obtain the desired fragility curves for the different structural limit states. locally or more widely. This was achieved by including strength degradation at the element level and by limiting at each step of the analysis the incremental displacement at one. A unique aspect of this procedure is that the structure safety while in a damaged state is measured in terms of its ability to protect occupants from death or injury due to aftershocks that might occur prior to its repair. 1989 attenuation expression as in and graphically represented in Figure 44. the knowledge of the nonlinear static behavior of a specific structure subjected to incremental lateral deflection is used to infer its nonlinear dynamic response expected for different levels of ground motion severity.

6. then there is no need to continue the NSP analysis is stopped. A fiber section model at each integration point. These major inelastic events sometimes cause significant drops in the base shear or changes in the global stiffness of the building which in turn translate into changes in slope of the NSP curve.2 (Mazzoni et al 2005) OpenSees is an object-oriented framework for finite element analysis. Core components. More formally. which in turn is associated with uniaxial material models and enforces the Bernoulli beam 75 . The inelastic events are to be associated with specific damage states. Figure 36 NSP curves for the intact structure and damage states OpenSees 1. that are the abstract base classes. DSi is therefore defined by (a) a roof drift value.associated roof drift levels are identified. As mentioned above. A typical frame is modeled as a three-dimensional framework of beams and columns. A key feature of OpenSees is the interchangeability of components and the ability to integrate existing libraries and new components into the framework without the need to change the existing code. including whether any element has reached ultimate vertical capacity. Structure Model The current version of OpenSees includes a general model-builder for creating two and three dimensional frame and continuum models using a scripting language called TCL. the occurrence of the ith major inelastic event (or a set of events at approximately the same deformation level) identifies the ith damage state. DSi. and (b) a detailed description of the structural damage associated with that event. Adding new component classes is easier and all that is required can be accommodated. if local loss of vertical capacity at a certain damage state is likely. Δi. The software is developed for the users of research community. Beams and columns are modeled using force-based non-linear beam column elements that consider the spread of plasticity along the length of the element. Each one of these damage states will be later associated with one (and only one) structural limit state. as depicted in Figure 36. define the minimal interface. The integration along the element is based on Gauss-Lobatto quadrature rule.

In this research work to model reinforced concrete frame structure nonlinear beam column element and two materials from OpenSees material library are used.assumption for axial force and bending. The input material properties are $Fy = yield strength $E0 = initial elastic tangent $b = strain-hardening ratio (ratio between post-yield tangent and initial elastic tangent) The parameters are shown in Figure 37. i. They are described below. Centerline dimensions were used in the element modeling for all cases. represents the force-based element. the beam is represented by the section model identified at each integration point. The input material properties are $fpc = concrete compressive strength at 28 days (compression is negative)* 76 .Material Parameters of Monotonic Envelope (after Mazzoni et al. Steel01 Material It is used to construct a uniaxial bilinear steel material object with kinematic hardening and optional isotropic hardening described by a non-linear evolution equation. Nonlinear Beam Column Element This element is used to construct a nonlinearBeamColumn element object. which is based on the non-iterative (or iterative) force formulation. and considers the spread of plasticity along the element. The element is prismatic. Figure 37 Steel01 Material .e. 2005) Concrete01 Material This material is used to construct a uniaxial Kent-Scott-Park concrete material object with degraded linear unloading/reloading stiffness according to the work of Karsan-Jirsa and no tensile strength.

Material Parameters (after Mazzoni et al. NSP curves for each damage state are needed for characterizing the residual lateral capacity left in the structure after different levels of damage that might have been inflicted on it by the mainshock.e. The residual lateral capacity is the quantity used when assigning a damage state to a structural limit state and therefore to a tagging condition. NSP curves for other damage states are obtained simply by assuming parallel-to-elastic unloading/re-loading that can result in comparable estimates of residual dynamic capacity. The assessment of the residual lateral capacity of the building in the DSi damage state is of interest provided that at the local level every damaged structural member is still capable of carrying vertical loads. To evaluate the post-earthquake residual capacity of the building to withstand future lateral loads. If there is loss of local member vertical capacity in the DSi damage state the building will be assigned a structural limit state that corresponds to collapse.. 2005) STEP 2: NSP for the Damaged Structure For each of the damage states that do not include loss of local vertical capacity (STEP 1). DS1) coincides with that of the intact structure. Figure 38 Concrete01 Material . The extent of this permanent displacement is somewhat an artifact of applying a static procedure to modeling the dynamic response of the structure 77 . analogous NSP curves are determined assuming the structure is in that damage condition after the mainshock. Unloading the structure from a damage state may create an initial offset in the damaged-structure NSP curve due to the residual permanent displacement in the structure. this step of the procedure requires obtaining the NSP curves for the building in every damage state. The parameters are shown in Figure 38. The pushover curve for a structure at the onset-of-damage state (i.$epsc0 = concrete strain at maximum strength* $fpcu = concrete crushing strength * $epsU = concrete strain at crushing strength* The initial slope for this model is (2*$fpc/$epsc0).

DS4) displacement. The effects of the expected (or measured) dynamic residual offsets on the residual lateral capacities are accounted to infer dynamic response. DS2 78 .. The NSP curves for the structure in the damaged conditions are assumed to start from the origin of the axes (i. As illustrated in Figure 39.g.e. the horizontal shift of the NSP curves for the damaged structures back to the origin does not shift the collapse (e. the collapse displacement will. However. Figure 39 NSP curve for damage state. be reduced by the residual offset. The residual displacement obtained from the NSP can be considered as an upper bound because the structure is not allowed to oscillate and therefore return to a residual offset closer to its original upright position. in effect. no permanent displacement).subject to ground shaking..

It is. and roof drift. T1. Sa(T1). however. multiplied by Say(T). values be consistent with the same selected point on the pushover curve. have the Y axis expressed in terms of the ratio of a ground motion parameter rather than of base shear. Of course. The result of this operation. Under the assumption that the response of structures is dominated by the first mode. at the fundamental oscillatory period of the intact structure. The parameter selected here is the spectral acceleration.Figure 40 Quadrilinear approximation of DS2 NSP curve (with and without offset). vs. The nonlinear dynamic behavior of the structure in state is estimated here using the NSP Curve and a (Microsoft Excel) spreadsheet SPO2IDA (Vamvatsikos and Cornell. where T is equal to T1 for the intact structure IDA.. which is the dynamic counterpart of the static pushover curve. is the Incremental Dynamic Analysis (IDA) curve (Vamvatsikos and Cornell. An appropriate normalization point that can be associated with incipient yielding in the structure is identified through engineering judgment. Selecting an accurate yielding point is desirable but not crucial. Δy. roof drift. the two ratios in terms of base shear or of spectral accelerations are numerically equivalent. STEP 3: Inferring Dynamic Response from Static Response. 2001b. the spectral acceleration at incipient yielding. which constitute the output of the SPO2IDA tool. BSy. and the normalizing quantity is Say(T1). BS. Vamvatsikos and Cornell. important that both the normalizing base shear. 79 .e. To facilitate the comparison of IDAs corresponding to different damage states. the IDA’s. their ordinates are “de-normalized” namely. base shear. However. Δ) by the corresponding quantities at the incipient yielding of the structure. The use of the SPO2IDA tool requires that the NSP’s used as input be normalized in both axes (i. This change of variables on the Y axis for the IDA case is important in this procedure because the tagging strategy is based on exceedance of ground motion spectral acceleration rather than base shear. 2001a). Sa(T1) for the intact structure can be obtained simply by dividing base shear by the structure effective modal mass for the fundamental vibration mode. The dynamic response of the intact structure is evaluated for a full range of potential mainshock ground motion levels. 2002) software.

namely to obtain the mainshock median ground motion level. de-normalized IDAs that. Figure 41 shows a typical output of the SPO2IDA tool where the NSP curve is the input and the IDA curve is the output.0 5. however. Sa. The spectral accelerations can be replaced with corresponding peak ground accelerations by using the ratio at fundamental oscillatory period of the intact structure (T1). μ.0 4.0 2.5 1. The median IDA curve is considered. μ Figure 41 Normalized NSP curve and incremental dynamic analysis curves for the intact structure (here R is BS/BSy for NSP and Sa/Say for IDA) Damaged structure 80 . are expressed in terms of spectral accelerations at different oscillatory periods.0 10.0 6. 5.0 Static PO IDA-50% capacity IDA-16% IDA-84% ductility.0 7. which is the ratio of the roof drift to the roof drift at the incipient yielding.0 8. Further the IDA curve defines the global collapse capacity (indicated by a circle in Figure 41) of the intact structure. DSi.0 3.5 0.0 4. M1* = M X Ψ1 Where Ψ1 = first-mode participation factor M = mass (self weight) of the frame structure The procedure above applied to the structure in different states of damage yields.5 4.5 R = SA/SAy 3.Effective modal mass. in general.0 1. Such Sa values can be read off the IDA curves at the roof drift values Δi associated with such states. Intact Structure The IDA curve is used in the opposite direction.0 0.0 3.0 9. In Figure 41 the deformation measure on the weaker axis is the roof global ductility ratio.5 2. that is expected to cause the intact structure to end up in each of the identified damage states.0 1.0 2.0 0. from Normalized Response Spectra of Figure 1.

The estimates of the median capacity for the intact structure and for the damage state DS1. All the damage states that do show loss of local vertical capacity will be automatically associated with collapse. If the proposed tagging criteria were simply based on maximum acceptable collapse risk of the partially damaged building. the red-tag state. Local loss of vertical capacity is potentially life threatening and for this reason this condition is associated here with structural collapse. the green. yellow. which defines the Onset of Damage state. Damage states. and red tag areas in Figure 44 would be oblique bands delimited by straight lines of constant P values. red. Aleatory uncertainty (βR) 81 . As mentioned earlier. the four (mainshock) ground motion levels expected to cause the structure to enter the onset-of-damage state. There is a smaller chance at lower ground motion levels and a larger chance at higher levels.cap)i as one of the points of the IDA for the damaged structure.. These chances are first quantified by estimating the dispersion.g. for which vertical capacity is lost at least at one structural component are associated with the collapse limit state. which is a combined measure of two basic kinds of uncertainty: aleatory uncertainty (or randomness) and epistemic uncertainty. spectral acceleration) at which the limit state (and associated tagging) will occur cannot. yellow-tag. if any. LS equal to OD.The same SPO2IDA spreadsheet provides an estimate of (Ša. this IDA-based procedure has identified the median spectral acceleration value corresponding to the onset of a given structural limit state.cap)i. The IDA curves for the damaged structures will generally provide estimates of the residual capacities. that are not equal to that of the intact structure. are the same. and the collapse state are identified according to a pre-defined criterion provided in this step. R. Each potential damage state is categorized with no loss of local vertical capacity as one of the following five structural limit states: green-tag. Identifying the median spectral capacity value for the green tag state is not necessary. 20%. and it is assumed therefore that there is a 50-50 chance that the limit/tagging state (or worse) will be observed if this ground motion occurs at the site. STEP 5: Ground Motion Level Associated with a Structural Limit State Using the results of STEPS 3 and 4. The tagging is based on the corresponding residual capacities identified in STEP 3 and on the (site-specific) mean annual frequency of exceedance of the (aftershock) ground motion values corresponding to such residual capacities. The proposed tagging approach coupled with the IDA curve for the intact structure leads to the identification of the median spectral capacity values associated with the onset of postearthquake tagging status. however. red-tag. The limit states are determined from the Hazard curve of the site. in fact. yellow. The value just identified is a “best guess”. or Coll. onset of damage. The diagonal bands in Figure 44. or collapse. The ground motion intensity (i.e. be predicted perfectly. (Ša. and 40%. Local loss of vertical capacity does not necessarily imply that the structure has no global lateral capacity left to resist aftershocks. however. onset-of-damage. STEP 4: Occupancy Status for Damaged Structure.. denoted ŠaLS for limit state LS (e. Y. the yellow-tag state. β. are delimited by horizontal lines drawn at constant values of capacity loss of 2%. collapse). Such ground motion levels are identified by their median values and related dispersion measure.

the fragility curves associated with the onset-ofdamage. yellow. This difference is captured in the four curves presented in Figure 16. 0. βR. Values of ŠaLS and β are provided in STEP 5. β.and Yellow-tag conditions. For example. the yellow-tag. The fragility curve (Figure 42) for a given structural limit state LS (LS equal to onset of damage. the values of βR follow the trend shown in Figure 16. the red-tag. the values for βU will be larger (i. Further the response dispersion is larger for larger degrees of nonlinearity. and the collapse states are plot. reflects the professional confidence that the selected model and the analysis procedure will predict accurate results. red-. The fragility curve corresponding to the green-tag state is equal to unity for all levels of ground motion. to be used in the determination of the fragility curve (STEP 6) is the SRSS value: 2 2 β = β R + βU STEP 6: Computation of the Fragility Curves Based on the median values and corresponding dispersion measure of the four ground motion levels number provided at STEP 5. one will have less confidence) for complex. one corresponding to collapse. FLS(Sa). On average. Similarly the confidence in the results increases with the level of effort in the structural modeling. and one to the less severe Onset-ofDamage limit state. In this study variability in concrete compressive strength and column depth is used to determine epistemic uncertainty (βU) Total Uncertainty (β) Finally the value of net dispersion. The former is the central value of the curve that correspond to an exceedance probability of 50%. βU. The fragility curve. Sa: S a = S aLS e xβ (3) 82 .05. β. it is known that the aleatory portion of the dispersion. older structures modeled using a simplified model and untested material properties than for a clean. The fragility curve. The former parameter is referred to as the median spectral capacity value of that limit state. especially when the ground motion is near the collapse capacity.Based on the results of many nonlinear dynamic analyses. modern steel moment-resisting frame whose properties have been well determined and whose model has been developed with extreme care to details. one to the Red.5.95} versus the corresponding values. or collapse state) provides the annual probability that the intact building will end up in the specified limit state (or worse) given the occurrence at the site of an earthquake ground motion of intensity Sa. the curve’s estimation for the generic structural limit state LS requires two parameter values. for example.e. is denoted as FY(Sa). green.75 and 0.. for the yellow-tag state. Based on the common lognormal assumption. the latter controls its slope (the larger the β value. 0. 0.25. a median ŠaLS value and a measure of dispersion. namely short-period (highfrequency) structures show more record-to-record variability in their displacements (or correspondingly. in the spectral acceleration at which a given displacement or damage state is first observed). depends on the initial period of vibration of the structure. for the generic structural limit state LS is determined by plotting the values of probability p = {0. the flatter the curve). Epistemic Uncertainty (βU) The epistemic part of the uncertainty.

Here to define limit states from seismic hazard of Dhaka city the hazard curve (Figure 43) developed using Duggal 1989 acceleration attenuation expression is used. which is more familiar and more readily available. of the building in its intact conditions. (Ša.67. Figure 42 Fragility curves for onset of damage. It should be noted that in Figure 42 fragility curve for onset of damage (OD) is the steepest of the four because the value of β is the smallest. P0. as a proxy for evaluating the post-mainshock or “aftershock hazard. yellow.for the values of x equal to {–1. 4. -0.67 and 1. 0. P. and red tags. The primary criteria proposed here implicitly assume that one can use the pre-mainshock hazard or MAF.65}.65. Additional values of p and x are used from table of the Gaussian distribution function (Appendix). of the building in the damage state DSi. in whichever damage states they may be. which refers to the pre-earthquake conditions.2 DEFINING LIMIT STATES FROM HAZARD CURVE The proposed criteria for tagging damaged buildings. 0. is the building-site-specific MAF of the aftershock ground motion corresponding to the median capacity. the building-site-specific mean annual frequency (MAF) of exceedance of the ground motion corresponding to the median capacity. The opposite is true for the fragility curve corresponding to the collapse state.cap)1. and for collapse of the building. green.0.cap)i. are expressed in terms of: P0. respectively. (Ša. can be obtained using Probabilistic Seismic Hazard Analysis (PSHA) of building-site. 83 .

9 14.6 YELLOW 5%/50yrs Loss of Sa cap (%) 20 40 60 30 P/P0 3 20 2 GREEN 1 2%/50yrs 10 10 -5 P0 10 -4 10 -3 4 2 0 84 .Figure 43 Seismic Hazard Curves for Dhaka city using Duggal (1989) attenuation expression The proposed primary tagging criteria are displayed in graphical form in Figure 44. 60 10 50 Capaci ty Loss (%) 2 P/Po 7 6 5 4 RED 40 1.1 2. The figure has two scales for the ordinates. the percentage of loss in Ša.cap and the ratio of P/P0 that measures the increase in frequency of exceeding the median residual capacity of the building damaged by the mainshock.4 5. The values of the quantities in Figure 44 and in the text are proposed for Dhaka city and should be reevaluated for other sites and structures of different importance.

If the intact building had been much weaker or in a harsher seismic environment such that its value of P0 were equal. (b) Average relationship for loss of ground motion capacity and rate of increase in mean annual frequency of exceedance of ground motion for Dhaka city (longitude 90. Therefore a building whose P0 is equal. The opposite is true for lower values of P0. to 3 x 10-4 needs to be damaged severely enough to loose about 5% of its initial capacity before it is tagged Y and about 30% before it is tagged R. The plan. 4. Three dimensional models are created and static displacement controlled pushover analysis is done using Opensees1. that the damaged-building capacity is exceeded but the computation of P is performed for simplicity using pre-mainshock PSHA. P. the tag would be either green or red. the primary tagging criteria explicitly consider the likelihood. If the proposed tagging criteria were simply based on maximum acceptable collapse risk of the partially damaged building. elevation and column and beam sections are shown in Figure 45.(a) (b) Figure 44 (a) Graphical representation of the proposed tagging criteria. 1989 acceleration attenuation expression. No yellow tag could be assigned in this case. The building is designed for gravity load. These values. of 2.73 N) developed from hazard curve using Duggal. Output of SPO analysis is read and statistical calculation for development of fragility curves of different limit states is done in Matlab7. Conversely a much stronger building with P0 smaller than 2 x 10-4 would need to lose 20% of its original capacity before being tagged Yellow and 40% to be tagged Red. MRP. however. for example. which represent quantitative measures of acceptable risk. Any building is identified by a particular value of P0 that can be computed during “peace” time before any earthquake has occurred.3 FRAGILITY ANALYSIS OF A TYPICAL THREE STOREY REINFORCED CONCRETE FRAME STRUCTURE A typical three storey two by two building frame is analyzed.37 E and latitude 23. to 1 x 10-3. In this study for the boundaries between green and yellow tags and between yellow and red tags. P values to be equal to 2% in 50 years (mean return period. As stated above. or a combination of both. the green. The diagonal bands in Figure 43. and 40%. are delimited by horizontal lines drawn at constant values of capacity loss of 2%. respectively.6. for example. 85 . 20%. then a nominal loss of lateral capacity of only 2% or larger would cause the building to be redtagged. however. should be modified according to the building importance and severity of failure consequences.1. How the color of the tag changes with capacity loss can be found by searching on a vertical line at that specific value of P0. A larger value of P0 implies that the building is either relatively “weak”.2. yellow. or that it is located in an area of higher seismic hazard compared to that considered in the original design. and red tag areas in Figure 44 would be oblique bands delimited by straight lines of constant P values. The criteria can be summarized as follows.475 years) and 5% in 50 years (MRP of 975 years) are selected.

where displacement is applied. Node.5" 2-#5 bar Ctrl.00 in 50 percentile = 12. where displacement is applied. Node. Table 5 Input parameters of the frame structure: Parameter Beam depth Beam width Uncertainty No No Quantity 15.25 in 86 Dispersion 10' .15' 3-#5 bar 15" 2. 10' 4-#6 bar 12" 10' COLUMN SECTION DIRECTION OF DISPLACEMENT ELEVATION Figure 45 The plan.00 in 2. The input parameters are stated in Table 5.75 in 75 percentile = 12.6.2 using tcl command language.50 in 25 percentile = 2.00 in 10.00 in 25 percentile = 11.25 in Column width Cover No Yes 12.50 in 50 percentile = 2.5" Ctrl.00 in Column depth Yes 12. elevation and column and beam sections of the typical three storey building frame Frame Structure Model Three dimensional model of the typical frame structure is created in Opensees1. 15' 10" BEAM SECTION 12' 12' PLAN 12" 2.

25 percentile = 3.8 ksi 57 50 percentile = 4.02 4 .0025 0.5 ksi 75 percentile = 4.1 f c′ 0.0 ksi 29000. The solid line in the Figure 46 is the SPO curve and dashed line is the quadrilinear approximation of the 87 .0 ft 10.Parameter Uncertainty Quantity Dispersion 75 percentile = 2.5 ksi ( f c′ ×1000 ) 0.0 ksi dev.75 in Concrete compressive strength Yes ( f c′) Concrete modulus of Yes elasticity (Ec) Concrete strain at No maximum strength Concrete strength crushing at Yes No No of No 4.0 ft ten percent capacity ten percent capacity of of column column Concrete strain crushing strength Steel yield strength Steel modulus elasticity Strain-hardening ratio No of steel Steel area in column Steel area in beam No No Lump mass in each No floor Gravity load on No column of each floor Span Bay Floor to floor height No No No Step 1: Nonlinear Static Pushover Analysis of the Intact Structure NSP is performed on the intact structure with the parameter stated above.4 kip 10.#6 bot. 2-#5 top 3-#5 86.004 40.0 ksi (mean) standard =0.0 kip 2 @ 12.0 ksi 0.0 ft 2 @ 15.

The damage states are identified as in Table 6 and specified on the curve of Figure 46. These choices of damage states are not unique.0322 before failure DS4 (Collapse) Failure of one Δ4 = 0. P. can also be considered as failure. DS1 (Onset of Starting of nonlinear Δ1 = 0. Infill and diaphragm may damage completely. these damage states will be associated to different structural limit states.Δ effect is significant.03). Any other point relating to a change in any structure properties can be chosen as a damage state.0417 structural element 88 . Structural element may be a beam or a column.0060 Damage) behavior First considerable Δ2 = 0. Later.SPO curve. Figure 46 Non linear static pushover curve for the intact structure Table 6 Identification of damage states Damage States Selection criteria Roof drift Remark Yielding point Smaller drops are due to yielding of beam steel and larger drops are due to yielding of column steel. As roof drift is high (>0.0195 drop of base shear DS2 DS3 Last considerable drop of base shear Δ3 = 0. As roof drift is very high.

Output of Step 1 Roof drift at Onset of damage.. 89 ..e.020 ΔDS3 = 0.g. NSP curves for other damage states are obtained simply by assuming parallel-to-elastic unloading/re-loading (Figure 47 and Figure 48) that can result in comparable estimates of residual dynamic capacity. Unloading the structure from a damage state may create an initial offset in the damaged-structure NSP curve due to the residual permanent displacement in the structure. The NSP curves for the structure in the damaged conditions are assumed to start from the origin of the axes (i.042 Step 2: NSP Curves for the Damaged Structure The assessment of the residual lateral capacity of the building in the DS i damage state is of interest provided that at the local level every damaged structural member is still capable of carrying vertical loads. The extent of this permanent displacement is somewhat an artifact of applying a static procedure to modeling the dynamic response of the structure subject to ground shaking. As illustrated in Figure 49.e. ΔDS2 = 0. the horizontal shift of the NSP curves for the damaged structures back to the origin does not shift the collapse (e.006 Roof drift at DS2. Roof drift at Collapse. no permanent displacement). DS4) displacement.ΔOD = 0. If there is loss of local member vertical capacity in the DSi damage state the building will be assigned a structural limit state that corresponds to collapse. The residual displacement obtained from the NSP can be considered as an upper bound because the structure is not allowed to oscillate and therefore return to a residual offset closer to its original upright position.. The effects of the expected (or measured) dynamic residual offsets on the residual lateral capacities are accounted to infer dynamic response.032 ΔColp = 0. DS1) coincides with that of the intact structure. The pushover curve for a structure at the onset-of-damage state (i. Roof drift at DS3.

For the same maximum displacement. 90 . as shown in Figure 18. on average a comparatively much smaller residual offset would be expected in a proper dynamic analysis. We have chosen therefore in these guidelines to remove that offset by shifting the damaged NSP’s back to zero offset.Figure 47 Non linear static pushover curve for damage state DS2 Figure 48 Non linear static pushover curve for damage state DS3 The damaged structure NSP’s display a large initial residual static offset that is an artificial product of the static way in which the structure has been loaded and unloaded to reflect the mainshock response. Note that this shift does not reduce the roof drift at which collapse is predicted.

The normalized IDA that is output is then being “de-normalized” for proper interpretation.72 Point 5 (DS4) μ R 0. Each NSP curve is considered separately. and the ratio.72 1. R.97 Intact DS2 DS3 91 .95 0. The Incremental dynamic analysis (IDA) curves are obtained from the NSP curves using the SPO2IDA software. of the base shear (BS). The ordinates of these points are called BSyi and Δyi. the NSP curves need be approximated by no more than four straight lines. Step 3: Inferring dynamic Response from Static Response. to the base shear at incipient yielding (BSyi). Hence.00 8.54 1. In order to be used in the SPO2IDA spreadsheet.Figure 49 Quardrilinear approximation of DS2 NSP curve. and the point corresponding to first yielding is selected on each of them (DS1. the following normalization step is performed on NSP results from Step 2.00 6. with and without offset Output of Step 2 From Figure 46 to Figure 49. and DS3 marks in Figure 46).19 3. Table 7 Point of normalized quadriliniear NSP curves Point 1 (Origin) μ R 0 0 0 0 0 0 Point 2 (Yield) μ R 1 1 1 1 1 1 Point 3 (DS2) μ R 1. The piecewise linear fit of the NSP curves is also shown in Figure 46 to Figure 49. Using Ri = BSi/BSyi.10 Point 4 (DS3) μ R 0. DS2. and μi = Δi/Δyi.78 5. The SPO2IDA spreadsheet requires normalized NSP curves expressed in terms of global ductility ratio.44 0.41 2. To match the input requirements for this software several simple steps of straight-line approximation of the NSP’s and normalization of the axes are conducted. the input for SPO2IDA is as shown in Table 7.28 1.78 7. The NSP curves in Figure 46 to Figure 49 are in terms of base shear and roof drift. μ.

0 6.0 7. The normalized NSP and IDA curve for the intact structure are shown in Figure 50.0 2. where Say. which is estimated by BSy (58 kip) divided by the product of the mass of the structure (260 kip) and the first-mode participation factor (0.87).0 1.3 second. The first-mode participation factor is determined from eigenvalue analysis of the structure.006 from Point 2 of Figure 46) and show the ordinate transformation.0 0. R relationships (i.0 R = SA/SAy 5. 8.0 4.6.Using the NSP curves in Table 6 as an input to the SPO2IDA spreadsheet produces the dynamic μ vs.0 7.0 4.2 gives the eigenvectors matrix and participation factor is determined from the matrix (appendix).0 6.0 Static PO IDA-50% capacity IDA-16% IDA-84% 1. R = Sa/Say. Opensees1.. Natural period of the structure is also obtained from eigenvalue analysis and is found to be 0. where R is now to be interpreted as Sa/Say.0 8.0 10. by multiplying the μ’s by the yield drifts (0. Table 8 and Table 9 show the results of de-normalizing the abscissa or displacement axis. R is multiplied by Say.0 9.0 ductility.0 3. For the intact structure. Table 8 and Table 9 demonstrate this procedure. the IDA curves in Figure 50 and Figure 51 are in terms of μ (ductility) and R.0 0. is the yield acceleration to be discussed below. Similarly the IDA curves of damaged structures are also determined and shown in Figure 52. Sa. the IDA curves) for the intact structure and the damaged state structures. As stated. A “de-normalization” process is performed to produce more meaningful IDA curves directly in terms of roof drift vs.0 5.0 3.0 2.e. μ Figure 50 Incremental dynamic analysis curves for the intact structure 92 . The SPO2IDA output is in terms of normalized drift (μ = Δ/Δy) and normalized spectral acceleration.

0 0.0 Static PO IDA-50% capacity IDA-16% IDA-84% 1.54 Point 4 (DS3) R μ 5.0 2.0 8.28 2.042 0.0 R = SA/SAy 5.95 Point 5 (DS4) R μ 7.641 0.05 2.006 0.0 5.0 0.44 8.0 6.667 0.0 7.0 1.00 Intact DS2 DS3 Table 9 Un-normalized (median) IDA curves Point 1 (Origin) Drift Sa (g) 0 0 0 0 0 0 Point 2 (Yield) Drift Sa (g) 0.513 Point 5 (DS4) Drift Sa (g) 0.0 4.042 0.032 0.010 0.41 3.0 3.005 Point 3 (DS2) Drift Sa (g) 0.225 0.256 0.820 0.0 7.00 2.60 6.40 1.50 1.042 Intact DS2 DS3 93 .782 0.225 0.20 2.0 4.00 3.0 6. μ Figure 51 Incremental dynamic analysis curves for damage state DS2 Table 8 Normalized (median) IDA curves obtained via the SPO2IDA spreadsheet Point 1 (Origin) R μ 0 0 0 0 0 0 Point 2 (Yield) R μ 1 1 1 1 1 1 Point 3 (DS2) R μ 3.00 1.359 Point 4 (DS3) Drift Sa (g) 0.256 0.0 3.020 0.0 9.8.006 0.019 0.0 ductility.0 2.

For the structure in damage state DS2 spectral acceleration at yield S ay was obtained 0. Output of Step 3 For the intact structure four structural damage states are attained at four different spectral accelerations.78g 0.82g The capacity of the intact structure and residual capacity of damaged structure are obtained and as follows: Structure condition Intact structure Capacity 0. This is due to the procedure of obtaining SPO curve for the damage state DS2 (parallel to elastic unloading-reloading at DS2).26g 0. Like the NSP curves (e.32g. these IDA curves assume no residual offset after the mainshock. So Say of intact structure (0. however.Figure 52 IDA curves for the intact structure and for the damaged structure in damage states DS2 and DS3 The IDA curves for the intact structure and for structure in different damage states expressed all in terms of the spectral acceleration at the fundamental period of the intact structure are shown in Figure 52. Figure 49) they are derived from.g. 0.64g 0.82g 94 . As in Table 9 for the intact structure spectral acceleration at yield Say (0.256g) is considered as Say for DS2.256g) is obtained from analysis. which is larger than that of the intact structure.. The behavior of the structural elements which start to damage before damage state DS2 is not considered (DS1 to DS2 portion of the intact SPO curve). Damage states Onset of damage (DS1) Damage state DS2 Damage state DS3 Collapse (DS4) Spectral accelerations.

26g. what limit or tagging states to associate with each of the damage states are determined. As in Figure 44 for a structure of P0 = 6×10-5 the criteria (loss of spectral acceleration) for structural limit states Yellow Tag and Red Tag are Tagging condition Onset of Yellow Onset of Red Loss of Sa capacity 20% 40% From the results of STEP 3. The simplest case.23g 0.82g.82g Residual Sa capacity 0.82g 0.Damaged structure DS2 Damaged structure DS3 Damaged structure DS4 0. From Normalized Response Spectra of Figure 1. is predicted to occur at a drift of 0.23g.33g. Output of Step 4 Tagging of four structural damage states Structural damage states Onset of damage (DS1) Damage state DS2 95 Tag Green Green . From Figure 44 the damage state DS2 is green-tagged whereas DS3 is red-tagged.26g 0.33g is 6×10-5 (=P0).67g 0.67g 0.00 Loss of Sa capacity 0% 18. The damage state DS4. From seismic hazard curve for Dhaka city of Figure 43 the mean annual frequency of exceedence of PGA 0. is by default tagged red and it is used here as describing the incipient collapse for the purpose of developing fragility curves. for a structure of period 0.5.2% drift that implies collapse at a Sa value of 0. Table 10 Dynamic capacity and loss of dynamic capacity of the structure at four structural damage states Damaged structures Onset of damage DS2 DS3 Collapse (DS4) Sa 0. To determine the tagging states associated with DS2 and DS3 the loss in capacity and the probability of an aftershock (here probability of aftershock is considered same as that of mainshock) causing a collapse of the damaged structure are considered.2 sec period Sa) that are expected to cause (local) collapse of the damaged structures. So corresponding pick ground motion of Sa = 0. which is associated with local collapse.3sec the ratio of spectral acceleration to maximum ground motion is 2.0% 100% These are the aftershock ground motion intensities (as measured in terms of 0.3% 72.64g 0. 0 STEP 4: Occupancy Status for Damaged Structure In this step. the onset of damage limit state.6% and spectral acceleration of 0. From Figure 52 the intact structure will reach the 4.82g is 0. the (median) Sa capacity estimates of the damaged structures and loss of Sa capacity are as in Table 10.78g 0.

8 0.26g 0.72g.Damage state DS3 Collapse (DS4) Red Red STEP 5: Ground Motion Level at Incipient Structural Limit State In this step. the mainshock ground motion intensity that is expected to bring the intact structure to the verge of each relevant structural limit state is identified. median roof drift and corresponding main shock ground motions causing the onset of all limit states are shown in Table 11. main shock ground motions causing the onset of all limit states are calculated and given in the Table 11.6 0.006 0.1 0 0 20 40 60 80 100 Loss of Sa capacity (percent) Figure 53 Loss of Sa capacity at different spectral acceleration From Figure 53 20% Sa capacity loss (onset of yellow tag state) is associated with spectral acceleration of 0.65g and 40% Sa capacity loss (onset of red tag state) is with spectral acceleration of 0. median roof drift and corresponding main shock ground motions causing the onset of four structural limit states Structural limit states Onset of damage Onset of yellow Spectral acceleration 0.25g . Estimated values of the median spectral acceleration.3 0. Dividing the spectral acceleration by the ratio 2. Table 11 Median spectral acceleration.7 Spectral acceleration (g) 0.4 0. 0.5.10g 0.7% respectively.65g 96 Roof drift 0. obtained from Figure 1. These correspond to the median values of the limit state fragility curves.2 0. Figure 53 is drawn using the result of Table 10 obtained in step 4. From Figure 52 corresponding roof drift values are 2.5 0.9 0.1% and 2.021 Ground motion (PGA ) 0.

75 in 75 percentile = 12.42 OC 0. Table 12 Aleatory uncertainty (βR) for four structural limit states of the structure Structural limit states βR OD 0.31 OY 0.00 in 25 percentile = 11.5 ksi 50 percentile = 12.82g 0. = 0. 97 .57 The epistemic uncertainty (βU) is determined considering uncertainty of three major structural parameters.5 ksi 75 percentile = 4.35 OR 0.Onset of red Onset of collapse 0.33g Estimation of Dispersion (β) The values of aleatory uncertainty (βR) are taken from Figure 16 and shown in Table 12. The input percentile values of those parameters are as in Table 13.0 ksi (mean) standard dev.027 0.8 ksi Percentile value 50 percentile = 4.75 in β3 β2 β1 Uncertainty ( f c′) Determination of β1 From step one to step four are executed two times with ( f c′) = 3.042 0. Two SPO curves together with the SPO curve obtained using 50 percentile value are shown in Figure 54.25 in 75 percentile = 2.72g 0.25 in 50 percentile = 2.0 ksi 25 percentile = 3. Table 13 Uncertainty of three major structural parameters Parameter Concrete compressive strength Mean value 4.50 in Cover 2.5 ksi (75 percentile value).00 in Column depth 12.50 in 25 percentile = 2.29g 0.5 ksi (25 percentile value) and ( f c′) = 4. which uncertainties are common in Bangladesh. keeping other parameters unchanged.

05 due to loss of vertical load bearing capacity. obtained are as in Table 15.5 ksi onset of collapse is assumed to occur at a roof drift value of 0.020 0.006 DS2 0.025 0.Figure 54 Non linear static push over curves obtained using three percentile values of concrete compressive strength The damage states are identified for all the SPO curves and the roof drift values are as in Table 14.5 ksi DS1 0.032 0. 98 .5 ksi 3.0 ksi 4.014 DS3 0.006 0.042 For ( f c′) =4.028 DS4 0.027 0. Table 14 Roof drift values of four damage states Damage States ( f c′) 4.042 0. The Incremental dynamic analysis (IDA) curves are obtained from the NSP curves using the SPO2IDA software for the intact structures with concrete of three ( f c′) value and shown in Figure 55 and the spectral acceleration values.050 0.006 0.

5 ksi Damage States DS1 0.820g 0.98g 0. The residual capacities for damage state DS1 are same as those of the intact structures.67g 0.0 ksi 4.5 ksi 0.23g 0.5 ksi 3.782g 0.76g DS3 0.0 ksi 4.265g 0.80g DS2 0.90g 0.Figure 55 The Incremental dynamic analysis (IDA) curves.821g 0.641g 0.820g 0.5 ksi DS1 0. Table 16 Residual dynamic capacities of damaged structures Damage States ( f c′) 4.0 ksi 4.82g 0.708g DS4 0.5 ksi 3.253g DS2 0.5 ksi 3.801g So the dynamic capacities of intact structures with concrete of three follows. obtained from the NSP curves using the SPO2IDA software for the intact structures with concrete of three ( f c′) values Table 15 Spectral acceleration values of four structural damage states ( f c′) 4.470g DS3 0. ( f c′) Capacity 4.29g 0.979g 0.979g 0.256g 0.790g 0.27g DS4 0 0 0 99 .801g ( f c′) values are as Similarly the residual dynamic capacities of damaged structures are obtained and shown in Table 16.

72 0.0 ksi 4.025 0.256 0.80 Drift 0.98 0.3 0.027 Onset of collapse Sa (g) 0.5 ksi Figure 56 is drawn using the result of Table 15 and Table 17.021 0.9 0.3% 8. 1 0.5 ksi Onset of damage Sa (g) 0.3% DS4 100% 100% 100% 4.1 0 0 20 40 60 80 100 fc = 4.5 ksi fc = 3.042 0.006 Onset of yellow Sa (g) 0.5 ksi 3.8 Spectral acceleration (g) 0.80 0.021 Onset of red Sa (g) 0.0% DS3 72.265 0.65 0.050 0.6 0.253 Drift 0.7 0.5 ksi 3.0% 66. the percent loss of Sa capacity for each case is obtained as in Table 17.026 0.82 0.0 ksi fc = 4. Table 17 Percent loss of Sa capacity at four structural damage states Damage States ( f c′) DS1 0% 0% 0% DS2 18.0 ksi 4.006 0.2% 5.5 ksi Loss of Sa capacity (percent) Figure 56 Loss of Sa capacity at different spectral acceleration From Figure 56 and Figure 55 estimated values of spectral acceleration and roof drift causing the onset of all limit states are shown in Table 18.82 0.5 0.027 0.042 β1 values are calculated using equation (3).54 Drift 0. 100 .2 0.4 0.0% 70. Table 18 Spectral acceleration and roof drift causing the onset of four structural limit states Limit States ( f c′) 4.006 0.62 Drift 0.From the results shown in Table 16.

0.266 0.25 in (75 percentile) and 11.67. keeping other parameters unchanged.018 0. S a = S aLS e xβ 1 Sa ⇒ β = ln LS x Sa Where. 12.75 in (25 percentile).037 0.294 OR 0.209 OC 0.052 0.67 for 25 percentile.152 Determination of β2 Similar approach is executed with column depth = 12.194 0.223 0.310 0.0. S a = percentile Sa value S aLS = median Sa value (3) x = −0.0. Table 19 β1 values Limit State Obtained from 25%-tile value 75%-tile value Average OD 0. β1 values are obtained and shown in Table 19. 50 percentile and 75 percentile value respectively. Figure 57 Non linear static push over curves obtained using three percentile values of column depth 101 .277 0. Three SPO curves obtained using defferent percentile values are shown in Figure 57.035 OY 0.0 in (50 percentile).

obtained are as in Table 21.75 in DS1 0.697g 0.019 DS4 0.685g So for the intact structures with three percentile values of column depth the dynamic capacities are as follows.00 in 12. Table 20 Roof drift values of four damage states Damage States Column depth 12.253g DS2 0.256g 0.006 0.050 0.020 0.032 0. obtained from the NSP curves using the SPO2IDA software for the intact structures with three percentile values of column depth Table 21 Spectral acceleration of four structural damage states Damage States Column depth 12.00 in 12.25 in 11.75 in DS1 0.782g 0.042 0.264g 0.832g 0.25 in 11.937g 0.021 0. 102 .033 0.641g 0.The damage states are identified for all the SPO curves and the roof drift values are as in Table 20. Figure 58 The Incremental dynamic analysis (IDA) curves.018 DS3 0.624g DS4 0.006 DS2 0.029 The Incremental dynamic analysis (IDA) curves are obtained from the NSP curves using the SPO2IDA software for the intact structures and median IDA curves are shown in Figure 58 and the spectral acceleration values.619g DS3 0.006 0.820g 0.

75 in DS1 0% 0% 0% DS2 18. the percent loss of Sa capacities are obtained as in Table 23.25 in 11.25g 0.7% 15.25 in 11.67g 0.00 in 12.23g 0.69g DS2 0.685g Similarly the residual dynamic capacities of damaged structures are obtained and shown in Table 22. 103 .Column depth 12.82g 0.94g 0. Table 23 Percent loss of Sa capacity at four structural damage states Damage States Column depth 12.9% DS3 72.937g 0.75 in DS1 0.83g 0.4% 63. The residual capacities for damage state DS1 are same as those of the intact structures.58g DS3 0.25 in 11.00 in 12.25g DS4 0 0 0 From the results shown in Table 22.0% 73.75 in Capacity 0.820g 0.8% DS4 100% 100% 100% Figure 59 is drawn using the result of Table 21 and Table 23. Table 22 Residual dynamic capacities of damaged structures Damage States Column depth 12.3% 11.00 in 12.

3 Column depth = 12.00 in 12.4 0.82 0.006 0.72 0.7 0.006 0. Table 24 Spectral acceleration and roof drift causing the onset of four structural limit states Limit States Column depth 12.1 0.110 OR 0.8 Spectral acceleration (g) 0.032 OY 0.25 in 11.204 0.1 0 0 20 40 60 80 100 Column depth = 12.6 0. Table 25 β2 values Limit State Obtained from 25%-tile value 75%-tile value Average OD 0.018 0.047 0.75 in Loss of Sa capacity (percent) Figure 59 Loss of Sa capacity at different spectral acceleration From Figure 59 and Figure 58 estimated values of spectral acceleration and roof drift causing the onset of all limit states are shown in Table 24.00 in 0.69 Drift 0.125 OC 0.042 0.021 Onset of red Sa (g) 0.021 0.173 0.231 104 .73 0.78 0.264 0.050 0.9 0.027 0.006 Onset of yellow Sa (g) 0.2 0.256 0.024 Onset of collapse Sa (g) 0.63 Drift 0.25 in Column depth = 11.66 Drift 0.13 0.028 0.023 0.253 Drift 0.5 0.046 0.258 0.65 0.94 0.75 in Onset of damage Sa (g) 0.12 0.029 β2 values are calculated from equation (3) and obtained as in Table 25.

75 in 0. keeping other parameters unchanged. Three SPO curves obtained using different percentile values are shown in Figure 60.015 0.006 0.026 0. 105 . 2. Table 26 Roof drift values of four damage states Damage States DS1 DS2 DS3 DS4 Column cover 2. obtained are as in Table 27.032 0.50 in 0.Determination of β3 Similarly β3 is determine from varying column cover = 2.033 The Incremental dynamic analysis (IDA) curves are obtained from the NSP curves using the SPO2IDA software for the intact structures and median IDA curves are shown in Figure 61 and the spectral acceleration values.042 2.25 in 0. Figure 60 Non linear static push over curves obtained using three percentile values of column cover The damage states are identified for all the SPO curves and the roof drift values are as in Table 26.75 in (75 percentile).25 in (25 percentile) and 2.020 0.006 0.016 0.050 2.006 0.50 in (50 percentile).022 0.

67g 0.75 in 2.732g Similarly the residual dynamic capacities of damaged structures are obtained and shown in Table 28. Table 28 Residual dynamic capacities of damaged structures Damage States Column cover 2.75 in 2. The residual capacities for damage state DS1 are same as those of the intact structures.25 in DS1 0.919g 0.75 in 2.67g DS3 0.577g DS3 0.23g 0.50 in 2.27g 0.253g 0.30g DS4 0 0 0 106 .732g So for the intact structures with three percentile values of column cover the dynamic capacities are as follows.Figure 61 The Incremental dynamic analysis (IDA) curves.88g 0.919g 0.73g DS2 0.82g 0. Column cover Capacity 2.641g 0.737g 0.782g 0.50 in 2.267g DS2 0.256g 0.92g 0.674g 0. obtained from the NSP curves using the SPO2IDA software for the intact structures with three percentile values of column cover Table 27 Spectral acceleration of four structural damage states Damage States Column cover 2.50 in 0.820g 0.820g 2.580g DS4 0.25 in 0.25 in DS1 0.

006 0.006 0.7 0.From the results shown in Table 28.021 0.4 0.25 in Onset of damage Sa (g) 0.015 Onset of red Sa (g) 0.72 0.73 0.026 0.25 in DS1 0% 0% 0% DS2 18.253 0.267 Drift 0.027 0.050 0.3 Column cover = 2. 1 0.022 0.65 0.82 0.75 in 2.74 0.042 0.25 in Loss of Sa capacity (percent) Figure 62 Loss of Sa capacity at different spectral acceleration From Figure 62 and Figure 61 estimated values of spectral acceleration and roof drift causing the onset of all limit states are shown in Table 30.256 0.0% 70.50 in 2.50 in 0.57 Drift 0.92 0.5 0. Table 31 β3 values 107 .7% 58.2 0.4% 8.6 0.9% DS4 100% 100% 100% Figure 62 is drawn using the result of Table 27 and Table 29.73 Drift 0.1 0 0 20 40 60 80 100 Column cover = 2.50 in 2.75 in Column cover = 2. the percent loss of Sa capacities are as in Table 29.8 Spectral acceleration (g) 0. Table 30 Spectral acceleration and roof drift causing the onset of four structural limit states Limit States Column cover 2.006 Onset of yellow Sa (g) 0.2% DS3 72.016 Onset of collapse Sa (g) 0.9 0.58 Drift 0.033 β3 values are calculated from equation (3) and obtained as in Table 31.3% 4.75 in 2. Table 29 Percent loss of Sa capacity at four structural damage states Damage States Column cover 2.

65 0.29 0.304 0.42 0.063 0.57 0.018 0.335 0.174 0.50 Onset of red 0.294 0.32 OY 0. and onset collapse states are plot using equation (3).32 Onset of yellow 0.364 0.33 0. S a = S aLS e xβ (3) Where.66 STEP 6: Computation of the Fragility Curves From the median Sa-values and corresponding β-values of the four ground motion levels provided at STEP 5.062 0.52 Onset of collapse 0.032 0.72 0.204 0.26 0. Table 32 Combination of β-values Limit State β β1 β2 β3 βU βR β OD 0.26 0.189 0. Structural limit states Sa (g) PGA (g) β Onset of damage 0.209 0.184 0.52 OC 0. Sa S aLS β =Spectral acceleration value at different probability to reach the limit state =median Sa value to reach the limit state =measure of dispersion 108 . onset of yellowtag.182 0. the fragility curves associated with onset-of-damage.184 OR 0.035 0. Median PGA to reach different structural limit states and βvalues (dispersion) of those limit states are obtained.189 The square-root-of-sum-of-squares (SRSS) rule is used to combine the uncertainty (β values) and shown in Table 32.Limit State Obtained from 25%-tile value 75%-tile value Average OD 0.35 0.125 0.110 0.041 0.231 0.040 OY 0. onset red-tag.323 0.152 0.173 0.66 Output of Step 5 Median (50 percentile) Sa.182 OC 0.50 OR 0.196 0.31 0.82 0.040 0.10 0.

It should be noted that in Figure 63 fragility curve for onset of damage (OD) is the steepest of the four because the value of β is the smallest. The former is the central value of the curve that correspond to an exceedance probability of 50%. peak ground acceleration is used as system demand. The fragility curve corresponding to the green-tag state is equal to unity for all levels of ground motion. When instead of spectral acceleration. The opposite is true for the fragility curve corresponding to the collapse state. The former parameter is referred to as the median spectral capacity value of that limit state. the fragility curves/ for the building is as shown in Figure 64.x =X-value in Gaussian table for different probability (area under the standard normal distribution curve) The resulting fragility curves are shown in Figure 63. Figure 63 Fragility curves for different limit states of the building when system demand is spectral acceleration Based on the common lognormal assumption. 109 . a median Š aLS value and a measure of dispersion. the flatter the curve). β. the latter controls its slope (the larger the β value. the curve’s estimation for the generic structural limit state LS requires two parameter values.

In Figure 64 at PGA of 0.2g the building will be tagged Green and the structure will be deemed fit for immediate occupancy.2g Structural limit state Onset of damage Yellow tag Red tag Onset of collapse Probability of attaining the limit state 98% 30% 25% 22% Probability of attaining Yellow Tag state is thirty percent.Figure 64 Fragility curves for different limit states of the building when system demand is peak ground acceleration 4.3.1 Conclusion From Figure 34 for Dhaka city at five percent probability of exceedance in fifty years period predicted PGA is 0.2g. 110 .2g probability of attaining different limit states of the building is shown in Table 33. So after a ground shaking of PGA 0. Table 33 Probability of attaining different limit states of the building at PGA of 0.

all possible sources of seismic activity must be identified and their potential for generating future strong ground motion should be evaluated. of more general applicability.2 FRAGILITY ANALYSIS OF STRUCTURES A methodology is described to predict post-earthquake functionality of a structure. the uncertainties in earthquake location. 5. The uncertainty inherent in the predictive relationship is also considered. To evaluate seismic hazards for a particular site or region. Duggal (1989) expression is developed for alluvium soil of Japan. Fragility analysis of a typical three story building is performed and post-earthquake functionality of the building after a ground shaking of PGA 0. of course. Coupling the fragility curves with probabilistic seismic hazard analysis will permit assessments of the vulnerability of seismically induced structure. Duggal (1989) and Boore et al. and ground motion parameter prediction are combined to obtain the probability that the ground motion parameter will be exceeded during a particular time period. Finally. earthquake size. 111 . Identification of seismic sources requires some investigative work. In this analysis the historical seismicity is used and the other evidences can be incorporated when reliable data are obtained. The procedure developed in this research work is. Seismic performance of structures is cast in terms of fragility curves for structural limit states directly related to after earthquake structure occupancy status.5CONCLUSION 5. These may take the form of geologic and tectonic evidence. The proper performance of a PSHA requires careful attention to the problems of source characterization and ground motion parameter prediction and to the mechanics of the probability computations. The ground motion produced at the site by earthquakes of any possible size occurring at any possible point in each source zone is determined with the use of McGuire (1978).2g is predicted.1 SEISMIC HAZARD ANALYSIS OF BANGLADESH The history of earthquakes. Duggal (1989) and Boore et al. Fragility curves of a structure can be valuable to engineers to compare with observable damage during the inspection of a facility before deciding on the possible building occupancy restriction after an earthquake. (1993) acceleration attenuation expressions. As the soil of Bangladesh is alluvium and almost similar to that of Japan the output obtained using Duggal expression is proposed. or historical (preinstrumental) seismicity. The process is a crucial step in evaluating the likelihood that the structure may not be able to provide facility after an earthquake and the output will give a rational support to the engineer inspecting the facility after an earthquake. in Bangladesh is sufficient enough to require their careful consideration in the design of structures and facilities. Analysis is done using McGuire (1978). A limit-state fragility curve provides the conditional probability that the specified limit state will be reached or exceeded as a function of the severity of the future ground motion. (1993) acceleration attenuation relationship separately.

Ansary. pp. W. Oriental Geographer.A. U. Hanson. 4. 27 (2). Ansary.A. Bulletin of the Seismological Society of America. Noor. No. 112 . B. G. 2611–2623. Washington. T. 549-580.. 86. Colorado. J..T. Ansary M..J.M.. Organized by Disaster Management Bureau.. Allen. 79. 1999 Moheskhali earthquake: a damage study. 232. Al-Hussaini. J. Sharfuddin and J. Value of Historical Records of Earthquakes. Bender.. Struct.L. S.A. 2003.. Ambraseys. Boulder. Ansary.. Hatheways and C.REFERENCES Abrahamson. 1982.R. M. 8.. M. Probabilistic Estimates of Maximum Acceleration and Velocity in Rock in the Contiguous United States. Dhaka.. 2003. Reviews in Engineering Geology. 48(1). M. October 2004. Government of Bangladesh and Bangladesh Earthquake Society. R. M. Geology in the Siting of Nuclear Power Plants. eds. pp. Vol. 1989. Ministry of Disaster Management and Relief.. Roessett. S. P. Sadek and T. M. Open-File Report 82-1033. McClure. Proc. 111(12).A. 1-16.N.C. Eng. 1979. Design Spectra for Degrading Systems. Vol. Bulletin of the Geological Society of America. M. Algermissen. 375-379. Thenhaus.L.. Geological Society of America. Vol. Agra.M. 27-39. Al-Sulaimani. A.. 2001. Litehiser. 1975. N.. 139-149. N. J. 99. J.R.. pp. pp. 1985. Adair.J. Geologic Evaluation of a Site for a Nuclear Power Plant. 1971. of the Seminar on 2003 Rangamati Earthquake. Al-Hussaini. 2003 Rangamati earthquake – an engineering assessment. Attenuation of Vertical Peak Acceleration. India. Journal of Asiatic Society of Bangladesh.C. Geologic Criteria for Evaluating Seismicity.S.. C. 1041-1057. D. Vulnerability Assessment of Existing Engineered And Nonengineered Structures Of Dhaka City Using RVS And NDT Techniques. Seismic loss estimation of Dhaka for an earthquake of intensity VII. 2004.A.M. in A. D. Vol.A. Geological Survey. pp..... September 4. Perkins. Choudhury. M. Nature. 2004.

. EPRI Report NP-5930. New York. Science. RADIUS Initiative. and Associates (1988). Stanford CA.A. pp. 1993. 1451-1462. Vol. M. Suite 550.A. Aperiodic Holocene Recurrence of Widespread. California. 1468 pp.. 30. 58. J. U.. W. Villacis. Freeman and Company. SCIENCE. Understanding urban seismic risk around the world. C. Applied Technology Council.. Vol. 1999.. C. B. Electric Power Research Institute. Gaur and P. A Criteria for Determining Exceedance of The Operating Basis Earthquake. pp. Atwater. 293. W.. C.A..F.. TechBrief 2. A. Palo Alto. 1999. and Decision for Civil Engineers.. Benjamin. Sep. Cardona.. Bulletin of the Seismological Society of America. 282 pp. Molnar. B. Davidson and C. P. Vol. 759-804. 1984. Probably Coseismic Ssubsidence in Southwestern Washington. New York. 1985. V. R. Redwood City. D. 113 . Menun. 4. Hull. No. B. W.. Cornell. 2004. K. Strong Ground Motions Attenuation Rrelations: A ten-year perspective. C. 2001.A.G. 1970.M... Bolt. 1987. Advanced Seismic Assessment Guidlines.1968. R. Cornell. Inc. Incorporating Acceleration Variability into Seismic Hazard Analysis. Vol. 72 pp.H. Luco. Boore. 2004. Stanford University.B. 555 Twin Dolphin Drive. Vol. Probability. B.. OpenFile-Report 93-509. Mc-Graw Hill. 942-944. Benjamin. 68.. Bazzurro. Estimation of response spectra and peak accelerations from western North America earthquakes: An interim report. 236.. Reston. 1583-1606.R. Motahari. N.Entering Damaged Buildings. K. 1987. Campbell. Bilham. l. J.. 74.ATC. Earthquakes. Himalayan seismic hazard. Cornell..F. Evidence for Great Holocene Earthquakes along the Outer Coast of Washington State. Bulletin of the Seismological Society of America. Statistics. Eos. Summary Report on the Comparative Study of the United nations International Decade for Natural Disaster Reduction.. Bevis. California 94065 Atwater. Geological Survey. Earthquake Aftershocks . Earthquake Spectra. C. Engineering Seismic Risk Analysis.R. Virginia. pp. No. 4. T.K..E. Joyner. Bender. Fumal.A.S. 1988.. pp.

1985-1988. A fault-rupture model for seismic risk analysis. 1944. MIT Press. 34. R. FEMA 155. Luxembourg: Cahiers due Centre European de Geodynamique et de Seismology. 1999. (1998). No. Dyn. Proc. in R. FEMA-154 (ATC-21). Seismic Design of Nuclear Power Plants. Bulletin of the Seismological Society of America. 4. 1173-1194.. Washington. SAC Joint Venture.. Esteva. Sacramento. 4. 67. 1989.. Federal Emergency Management. 2004. F. 1986. 1988. pp. Thesis. ed. Struct. P. 2000. Rapid Visual Screening Buildings for Potential Seismic Hazards: Supporting Documentation Grunthal G. Recommended Seismic Design Criteria for New Steel MomentFrame Buildings (350). July. 111–116. Their Utilization as Design Criteria. 1970. Massachusetts. 23–67. FEMA 350-352.. Building data collection of Dhaka University campus. Recommended Seismic Evaluation and Upgrade Criteria for Existing Welded Steel Moment-Frame Buildings (351). Fajfar. 1997. B. and Richter. California. Rapid Visual Screening of Buildings for Potential Seismic Hazards: A Handbook. Seismic hazard methodology for the central and eastern United States.Der-kiureohian.. (ATC-2002). Personal Communication. HAZUS. Bulletin of the Seismological Society of America. Federal Emergency Management 114 . Earthquake Eng. on Earthquake Engineering. P. 25 (1). Gutenberg. Electric Power Research Institute.. Tokyo. California. Seismic Risk and Seismic Design Decisions.. European Macroseismic Scale.S. editor. P. C. Palo Alto.. L. Cambridge. NEHRP guidelines for the seismic rehabilitation of buildings. 1977. Report NP-4726. Fajfar..J. Natural Hazard Estimation Methodology. A. Recommended Postearthquake Evaluation and Repair Criteria for Welded Steel Moment-Frame Buildings (352). Estimation of Seismic Risk and Damage. University of Tokyo. 9th World Conf.C. EPRI. DU Chief Engineer. Frequency of Earthquake in California. Hansen. Fischinger. N2—A method for non-linear seismic analysis of regular structures. Gaspersic.. Japan. Volume 15. Vol. M Engg.H. FEMA-273.. Duggal. pp.. 2002.. ANG. 1996. Vol. D. A. M. The N2 method for the seismic damage analysis for RC buildings. No.

California. Washington.L. Canonsburg. pp. Maui.. T. Importance of Historical Seismograms for Geophysical Research. Proceedings. “Spectra-compatible Pushover Analysis of Structures”. C. Lee. B. FEMA (http://www. pp. 1993.. pp. BSTI. (1997). 73. Steven L. Chang. Report 25. Geotechnical Earthquake Engineering. Canada. Dynamic Versus Static Computation of the Residual Capacity of a Mainshock-Damaged Building to Withstand an Aftershock. and K. San Diego. Kunnath.` Kramer. Evaluating seismic risk in engineering practice. BNBC 1993. Historical Seismograms and Earthquakes of the World. Bazzurro. H. I. Idriss. 1985. I. B. Bangladesh National Building Code. Ground Motion Parameters Useful in Structural Design. Mississippi. U.K. 1987.S. Vol. 2003.P. R. Army Corps of Engineers Waterways Experiment Station. Gupta. (1980). No.. PA. San Francisco. Proceedings of the llth International Conference on Soil Mechanics and Foundation Engineering... University of Washington.C.. Southpointe. Proceedings of 13th WCEE. Paper 2405. Miscellaneous Paper S-73-1. Inc.. 3. Krinitzsky. Bulletin of the Seismological Society of America. Luco. P. ANSYS Theory Reference Release 5. K. Sudigh. Vol. 90-98.. N. August 1-6.M. B. 43 pp. eds. Heaton.gov/hazus/hazus6. 1999. Cornell. 16-31. 1988.H. F. M.S. ANSYS. 1984. Kohnke. 8th World Conference on Earthquake Engineering..Agency. Probabilistic evaluation of seismic exposure.H. H. Santa Fe. Shimazaki. E. Parameters for specifying intensity-related earthquake ground motions. Idriss. 933-941. H. Meyers. Eidth Edition.K.A. Kanamori. 1979. in W..htm) HBRI.. 13 September.... Vicksburg. Kulkarni. Vancouver. P. U. New Mexico. 1. 115 . S. Hawaii. 255-320.fema. Academic Press. Kanamori. 2004.. presented at the Conference on Evaluation of Regional Seismic Hazards and Risk.4.Japan Workshop on Performance-Based Earthquake Engineering Methodology for Reinforced Concrete Building Structures. Kennedy. San Francisco. Seismic potential associated with subduction in the northwestern United States. R.

Scott. C. pp.. 1. 67. August 1-6. 6. pp. pp. Vol. Inelastic displacement ratios for structures on firm sites. McKenna. in S. Newmark. Journal of the Geotechnical Engineering Division. 2005. July 2006. Ward..A. L.. Merz.H.A. Rosenblueth. 2004. 1991. 1949-2006. N. Miranda.L..H. Cornell. S. 1978. ASCE. Recommended Provisions for the Development of Seismic Regulations for New Buildings. National Earthquake Hazards Reduction Program. No. B. Vol. W.. Washington. Massachusetts. Merz. Krawinkler. 1973b. OpenSees Command Language Manual.. Fenves. G.. No. The John A.. Calif. Building Seismic Safety Council. No. J. 1973a. E. G. Bulletin of the Seismological Society of America. 877-915.. Cambridge. 104.. Deierlein. Seismic demands for SDOF and MDOF systems.. Blume Earthquake Engineering Center. 1977. Vol. J. 2006.A. G. M. O'brien..Mazzoni.333-353. Proceedings of 13th WCEE.. Stanford.R. F. F.. S. 126~10. Rep.. E. Canada.. Bulletin of the Seismological Society of America.A..L. 640 pp. GT4. June 2005. NEHRP. 1990. 481-490. Society of Exploration Geo-physicists. Fenves. 1991. R.A. D. ed. 1150–1159.. 116 . J. 1991. C.M.C.. A Framework Methodology for Performance-Based Earthquake Engineering. H. Aftershocks in Engineering Seismic Risk Analysis. 95. Paper 679. R. pp.. Eng. Geotechnical and Environmental Geophysics. Seismic Ground Motion Parameter Relations.J.. Report R73-25.. Vol. Arabasz. Struct. 73. K. Seismic risk based on a quadratic magnitude-frequency law. The correlation of peak ground acceleration amplitude with seismic intensity and other physical parameters.J.. McKenna. A. Stanford Univ. Cornell. Fundamentals of Earthquake Engineering. 1971. H. OpenSees Command Language Manual. H.G. Murphy. Nassar. Massachusetts Institute of Technology. Department of Civil Engineering.K. New Jersey. Mcguire.. McGuire. 2000.C. Moehle. Prentice Hall. Vancouver.. Englewood Cliffs. An Introduction to Probabilistic Seismic Hazard Analysis.. Mazzoni.

.A. Cornell. BUET.. April. 6. Analysis of completeness in the earthquake sample in the Puget Sound area and its effect on statistical estimates of earthquake hazard. M. Ruff. 1972. BUET. D. Vol. Development of Fragility Curves for R. pp.edu/cgi-bin/spo2ida-mt. Thesis. S. 2. Engg.stanford. Noor. Thesis. K. Manzur. Stepp.. Sharfuddin. D. 3. M.A..Sc. 139-162. 2001b. 240-252. Earthquake Hazard Analysis – Issues and Insights. 14-15. M... M. Seismic Hazard Analysis of Bangladesh. Thesis. S.. 23. Bulletin of the Seismological Society of America.. New York. T.. Cornell. pp. Sabri. Geophysical Research Letters. International Conference on Microzonation. Dhaka. Vol. Dec. GIS based methodologies for seismic hazard and risk analysis for Bangladesh..A. 3. Columbia University Press. Determination of maximum earthquake magnitude. First Bangladesh Earthquake Symposium (BES-1). J. 1980. J. First Bangladesh Earthquake Symposium (BES1).edu/docs/IDA-paper. Engg. Engg. Vol. No.. 1975. Dhaka. 1997.Dhaka.. L. “SPO2IDA: Excel spreadsheet”. Frame Structures: Bangladesh Context. A. Earthquake Engineering and Structural Dynamics. Physics of the Earth and Planetary Interiors. Vol. Smith. Seattle.. (2004). Vol.Noor. Reiter.. 2005. M.pl 117 .C.. C..C. M..C. Noor.A. On the Correlation of Seismic Intensity with Peaks of Recorded Strong Ground Motion. Ansary. 351-354.stanford.pdf Vamvatsikos. Kanamori.W. Trifunac.D. 2005.A. C. URL: http://nausika. BUET. M. 2001. L.. M. Ltd. BUET. 1990.Web implementation. M. D. Brady. 65. A. Proceedings. Thesis. Reappraisal Of Code Provisions For Structures Subjected To Seismic Loading. Yasin. ongoing Ph. 14-15. Dhaka. 2005. John Wiley & Sons.G. Seismicity and Subduction Processes. 254 pp. 897-910. Earthquake intensity-attenuation relationship for Bangladesh and its surrounding region. URL: http://tremble. H.A. Earthquake Hazard Analysis for Bangladesh. Dhaka. 1976. pp. New York.Dhaka. 2001a. 2002.Sc. Vamvatsikos. M. Washington. No. 31. Dec.. Incremental Dynamic Analysis. Sarker. 2005. pp.

Woodward-Clyde Consultants. 2002. San Onofre Nuclear Generating Station. No. Earthquake Eng. 2002. C. Richards. 1981. Weichert. Dyn. in D. R.. 2005. D. eds. A.. Simpson and T. Bulletin of the Seismological Society of America. Practical Estimation of the Seismic Demand and Capacity of Oscillators with Multi-Linear Static Pushovers through Incremental Dynamic Analysis. Active faults. Washington.S. 209-216. 12. SPO2IDA software for short. pp. 1337-1347.edu/nausika/software/spo2ida-allt. 1979. Maurice Ewing Series 4. <http://tremble. Proceedings of 7th U. 4. pp. moderate and long periods.A. Imposed Upper Limit for Probabilistic Seismic Risk. and earthquake hazards in the western United States. Wallace.W. D.E. 1980. Cornell. pp. National Conference on Earthquake Engineering. Report of the Evaluation of Maximum Earthquake and Site Ground Motion Parameters Associated with the Offshore Zone of Deformation. Report to Southern California Edition Company. Vamvatsikos.stanford. Proceedings. American Geophysical Union. 118 . 239-254. D. Direct estimation of the seismic demand and capacity of oscillators with multi-linear static pushovers through incremental dynamic analysis.. paleoseismology. Vol. 1984. July 21-25. D.xls> (Feb. Zemell.G. and C. 70. Estimation of the Earthquake Recurrence Parameters for Unequal Observation Periods for Different Magnitudes. S. C. D.Vamvatsikos. 2004).... San Fransisco.W. Cornell. Vamvatsikos. Boston. San Francisco. Struct. 8th World Conference on Earthquake Engineering. Earthquake Prediction: An International Review. Woodward-Clyde Consultants.

5239 1.5319 1.06 0.05 0.07 0.03 0.8485 2.06 0.09 0.8554 2.0239 0.8599 2.0244 0.0015 0.02 0.92 0.0250 0.8438 2.1611 0.1660 0.0014 0.9812 3.02 0. In order to get each value.5279 1.09 0.02 0.5120 1.03 0.92 2.1635 0.98 1.03 0.1685 0.9808 3.9987 3. from the minus infinite value to X value.9783 3.9778 3.96 0.98 0.94 2.0000 0. find the X value in the table.94 1.APPENDIX A Table of the Gaussian Distribution Function Areas Under the Standard Normal Distribution Curve The following table is used to find the value of the area under the standard normal distribution curve.0256 0.9989 3.95 1.5040 1.8508 2.0016 0.00 0.97 0.0000 0.0018 0.8621 2.07 0.01 0.06 0.08 0.9987 4.9803 3.0000 0.99 0.8531 2.5000 1.96 2.94 0.0013 0.98 2.00 3.97 2.0018 0.98 0.95 2.0014 0.0000 0.5359 1.08 0.0274 0.09 0.05 0. X Área X Área X Área X Área X Área X Área X Área X Área 0.96 0.05 0.00 1.9817 3.1788 0.9798 3.93 0.00 0.5080 1.5199 1.09 0.0000 0.1736 0.5160 1.02 0.95 0.93 2.01 0.0262 0.07 0.1587 0.9788 3.0016 0.04 0.97 0.9987 3.0017 0.93 0.8577 2.01 0.9988 3.9989 3.0000 0.9793 3.08 0.0000 0.0281 0.93 1.99 1.0000 0.92 0.0228 0.92 1.0000 0.0000 0.9772 3.01 0.07 0.00 0.9990 119 .00 0.8461 2.06 0.9989 3.0233 0.96 1.00 0.1814 0.9988 3.04 0.99 2.97 1.8413 2.99 0.0015 0.08 0.95 0.94 0.04 0.05 0.00 2.1762 0.9990 3.03 0.0268 0.04 0.1711 0. and in the right adjacent cell you will find the area under the standard normal distribution curve.

X 3,91

Área

X 2,91

Área

X 1,91

Área

X 0,91

Área

X

Área

X

Área

X

Área

X

Área

0,0000 0,0019 0,0287 0,1841 0,10 0,5398 1,10 0,8643 2,10 0,9821 3,10 0,9990 3,90 2,90 1,90 0,90 0,0001 0,0019 0,0294 0,1867 0,11 0,5438 1,11 0,8665 2,11 0,9826 3,11 0,9991 3,89 2,89 1,89 0,89 0,0001 0,0020 0,0301 0,1894 0,12 0,5478 1,12 0,8686 2,12 0,9830 3,12 0,9991 3,88 2,88 1,88 0,88 0,0001 0,0021 0,0307 0,1922 0,13 0,5517 1,13 0,8708 2,13 0,9834 3,13 0,9991 3,87 2,87 1,87 0,87 0,0001 0,0021 0,0314 0,1949 0,14 0,5557 1,14 0,8729 2,14 0,9838 3,14 0,9992 3,86 2,86 1,86 0,86 0,0001 0,0022 0,0322 0,1977 0,15 0,5596 1,15 0,8749 2,15 0,9842 3,15 0,9992 3,85 2,85 1,85 0,85 0,0001 0,0023 0,0329 0,2005 0,16 0,5636 1,16 0,8770 2,16 0,9846 3,16 0,9992 3,84 2,84 1,84 0,84 0,0001 0,0023 0,0336 0,2033 0,17 0,5675 1,17 0,8790 2,17 0,9850 3,17 0,9992 3,83 2,83 1,83 0,83 0,0001 0,0024 0,0344 0,2061 0,18 0,5714 1,18 0,8810 2,18 0,9854 3,18 0,9993 3,82 2,82 1,82 0,82 0,0001 0,0025 0,0351 0,2090 0,19 0,5753 1,19 0,8830 2,19 0,9857 3,19 0,9993 3,81 2,81 1,81 0,81 0,0001 0,0026 0,0359 0,2119 0,20 0,5793 1,20 0,8849 2,20 0,9861 3,20 0,9993 3,80 2,80 1,80 0,80 0,0001 0,0026 0,0367 0,2148 0,21 0,5832 1,21 0,8869 2,21 0,9864 3,21 0,9993 3,79 2,79 1,79 0,79 0,0001 0,0027 0,0375 0,2177 0,22 0,5871 1,22 0,8888 2,22 0,9868 3,22 0,9994 3,78 2,78 1,78 0,78 0,0001 0,0028 0,0384 0,2206 0,23 0,5910 1,23 0,8907 2,23 0,9871 3,23 0,9994 3,77 2,77 1,77 0,77 0,0001 0,0029 0,0392 0,2236 0,24 0,5948 1,24 0,8925 2,24 0,9875 3,24 0,9994 3,76 2,76 1,76 0,76 0,0001 0,0030 0,0401 0,2266 0,25 0,5987 1,25 0,8944 2,25 0,9878 3,25 0,9994 3,75 2,75 1,75 0,75 0,0001 0,0031 0,0409 0,2296 0,26 0,6026 1,26 0,8962 2,26 0,9881 3,26 0,9994

120

X 3,74

Área

X 2,74

Área

X 1,74

Área

X 0,74

Área

X

Área

X

Área

X

Área

X

Área

0,0001 0,0032 0,0418 0,2327 0,27 0,6064 1,27 0,8980 2,27 0,9884 3,27 0,9995 3,73 2,73 1,73 0,73 0,0001 0,0033 0,0427 0,2358 0,28 0,6103 1,28 0,8997 2,28 0,9887 3,28 0,9995 3,72 2,72 1,72 0,72 0,0001 0,0034 0,0436 0,2389 0,29 0,6141 1,29 0,9015 2,29 0,9890 3,29 0,9995 3,71 2,71 1,71 0,71 0,0001 0,0035 0,0446 0,2420 0,30 0,6179 1,30 0,9032 2,30 0,9893 3,30 0,9995 3,70 2,70 1,70 0,70 0,0001 0,0036 0,0455 0,2451 0,31 0,6217 1,31 0,9049 2,31 0,9896 3,31 0,9995 3,69 2,69 1,69 0,69 0,0001 0,0037 0,0465 0,2483 0,32 0,6255 1,32 0,9066 2,32 0,9898 3,32 0,9995 3,68 2,68 1,68 0,68 0,0001 0,0038 0,0475 0,2514 0,33 0,6293 1,33 0,9082 2,33 0,9901 3,33 0,9996 3,67 2,67 1,67 0,67 0,0001 0,0039 0,0485 0,2546 0,34 0,6331 1,34 0,9099 2,34 0,9904 3,34 0,9996 3,66 2,66 1,66 0,66 0,0001 0,0040 0,0495 0,2578 0,35 0,6368 1,35 0,9115 2,35 0,9906 3,35 0,9996 3,65 2,65 1,65 0,65 0,0001 0,0041 0,0505 0,2611 0,36 0,6406 1,36 0,9131 2,36 0,9909 3,36 0,9996 3,64 2,64 1,64 0,64 0,0001 0,0043 0,0516 0,2643 0,37 0,6443 1,37 0,9147 2,37 0,9911 3,37 0,9996 3,63 2,63 1,63 0,63 0,0001 0,0044 0,0526 0,2676 0,38 0,6480 1,38 0,9162 2,38 0,9913 3,38 0,9996 3,62 2,62 1,62 0,62 0,0002 0,0045 0,0537 0,2709 0,39 0,6517 1,39 0,9177 2,39 0,9916 3,39 0,9997 3,61 2,61 1,61 0,61 0,0002 0,0047 0,0548 0,2743 0,40 0,6554 1,40 0,9192 2,40 0,9918 3,40 0,9997 3,60 2,60 1,60 0,60 0,0002 0,0048 0,0559 0,2776 0,41 0,6591 1,41 0,9207 2,41 0,9920 3,41 0,9997 3,59 2,59 1,59 0,59 0,0002 0,0049 0,0571 0,2810 0,42 0,6628 1,42 0,9222 2,42 0,9922 3,42 0,9997 3,58 2,58 1,58 0,58 0,0002 0,0051 0,0582 0,2843 0,43 0,6664 1,43 0,9236 2,43 0,9925 3,43 0,9997

121

X 3,57

Área

X 2,57

Área

X 1,57

Área

X 0,57

Área

X

Área

X

Área

X

Área

X

Área

0,0002 0,0052 0,0594 0,2877 0,44 0,6700 1,44 0,9251 2,44 0,9927 3,44 0,9997 3,56 2,56 1,56 0,56 0,0002 0,0054 0,0606 0,2912 0,45 0,6736 1,45 0,9265 2,45 0,9929 3,45 0,9997 3,55 2,55 1,55 0,55 0,0002 0,0055 0,0618 0,2946 0,46 0,6772 1,46 0,9279 2,46 0,9931 3,46 0,9997 3,54 2,54 1,54 0,54 0,0002 0,0057 0,0630 0,2981 0,47 0,6808 1,47 0,9292 2,47 0,9932 3,47 0,9997 3,53 2,53 1,53 0,53 0,0002 0,0059 0,0643 0,3015 0,48 0,6844 1,48 0,9306 2,48 0,9934 3,48 0,9997 3,52 2,52 1,52 0,52 0,0002 0,0060 0,0655 0,3050 0,49 0,6879 1,49 0,9319 2,49 0,9936 3,49 0,9998 3,51 2,51 1,51 0,51 0,0002 0,0062 0,0668 0,3085 0,50 0,6915 1,50 0,9332 2,50 0,9938 3,50 0,9998 3,50 2,50 1,50 0,50 0,0002 0,0064 0,0681 0,3121 0,51 0,6950 1,51 0,9345 2,51 0,9940 3,51 0,9998 3,49 2,49 1,49 0,49 0,0003 0,0066 0,0694 0,3156 0,52 0,6985 1,52 0,9357 2,52 0,9941 3,52 0,9998 3,48 2,48 1,48 0,48 0,0003 0,0068 0,0708 0,3192 0,53 0,7019 1,53 0,9370 2,53 0,9943 3,53 0,9998 3,47 2,47 1,47 0,47 0,0003 0,0069 0,0721 0,3228 0,54 0,7054 1,54 0,9382 2,54 0,9945 3,54 0,9998 3,46 2,46 1,46 0,46 0,0003 0,0071 0,0735 0,3264 0,55 0,7088 1,55 0,9394 2,55 0,9946 3,55 0,9998 3,45 2,45 1,45 0,45 0,0003 0,0073 0,0749 0,3300 0,56 0,7123 1,56 0,9406 2,56 0,9948 3,56 0,9998 3,44 2,44 1,44 0,44 0,0003 0,0075 0,0764 0,3336 0,57 0,7157 1,57 0,9418 2,57 0,9949 3,57 0,9998 3,43 2,43 1,43 0,43 0,0003 0,0078 0,0778 0,3372 0,58 0,7190 1,58 0,9429 2,58 0,9951 3,58 0,9998 3,42 2,42 1,42 0,42 0,0003 0,0080 0,0793 0,3409 0,59 0,7224 1,59 0,9441 2,59 0,9952 3,59 0,9998 3,41 2,41 1,41 0,41 0,0003 0,0082 0,0808 0,3446 0,60 0,7257 1,60 0,9452 2,60 0,9953 3,60 0,9998

122

X 3,40

Área

X 2,40

Área

X 1,40

Área

X 0,40

Área

X

Área

X

Área

X

Área

X

Área

0,0003 0,0084 0,0823 0,3483 0,61 0,7291 1,61 0,9463 2,61 0,9955 3,61 0,9998 3,39 2,39 1,39 0,39 0,0004 0,0087 0,0838 0,3520 0,62 0,7324 1,62 0,9474 2,62 0,9956 3,62 0,9999 3,38 2,38 1,38 0,38 0,0004 0,0089 0,0853 0,3557 0,63 0,7357 1,63 0,9484 2,63 0,9957 3,63 0,9999 3,37 2,37 1,37 0,37 0,0004 0,0091 0,0869 0,3594 0,64 0,7389 1,64 0,9495 2,64 0,9959 3,64 0,9999 3,36 2,36 1,36 0,36 0,0004 0,0094 0,0885 0,3632 0,65 0,7422 1,65 0,9505 2,65 0,9960 3,65 0,9999 3,35 2,35 1,35 0,35 0,0004 0,0096 0,0901 0,3669 0,66 0,7454 1,66 0,9515 2,66 0,9961 3,66 0,9999 3,34 2,34 1,34 0,34 0,0004 0,0099 0,0918 0,3707 0,67 0,7486 1,67 0,9525 2,67 0,9962 3,67 0,9999 3,33 2,33 1,33 0,33 0,0005 0,0102 0,0934 0,3745 0,68 0,7517 1,68 0,9535 2,68 0,9963 3,68 0,9999 3,32 2,32 1,32 0,32 0,0005 0,0104 0,0951 0,3783 0,69 0,7549 1,69 0,9545 2,69 0,9964 3,69 0,9999 3,31 2,31 1,31 0,31 0,0005 0,0107 0,0968 0,3821 0,70 0,7580 1,70 0,9554 2,70 0,9965 3,70 0,9999 3,30 2,30 1,30 0,30 0,0005 0,0110 0,0985 0,3859 0,71 0,7611 1,71 0,9564 2,71 0,9966 3,71 0,9999 3,29 2,29 1,29 0,29 0,0005 0,0113 0,1003 0,3897 0,72 0,7642 1,72 0,9573 2,72 0,9967 3,72 0,9999 3,28 2,28 1,28 0,28 0,0005 0,0116 0,1020 0,3936 0,73 0,7673 1,73 0,9582 2,73 0,9968 3,73 0,9999 3,27 2,27 1,27 0,27 0,0006 0,0119 0,1038 0,3974 0,74 0,7704 1,74 0,9591 2,74 0,9969 3,74 0,9999 3,26 2,26 1,26 0,26 0,0006 0,0122 0,1056 0,4013 0,75 0,7734 1,75 0,9599 2,75 0,9970 3,75 0,9999 3,25 2,25 1,25 0,25 0,0006 0,0125 0,1075 0,4052 0,76 0,7764 1,76 0,9608 2,76 0,9971 3,76 0,9999 3,24 2,24 1,24 0,24 0,0006 0,0129 0,1093 0,4090 0,77 0,7794 1,77 0,9616 2,77 0,9972 3,77 0,9999

123

19 1.7852 1.4207 0.18 1.0162 0.9633 2.4483 0.23 Área X 0.13 0.12 2.1170 0.0132 0.13 0.9664 2.9979 3.0008 0.4247 0.9977 3.1271 0.16 0.1210 0.8023 1.15 0.9999 3.21 1.18 2.20 1.86 0.0009 0.20 2.9999 3.9975 3.80 0.14 2.17 0.23 Área X 2.85 0.79 0.9999 3.14 0.9999 3.85 0.12 124 .7995 1.78 0.9686 2.16 0.1112 0.0009 0.81 0.1230 0.0136 0.4325 0.0143 0.7881 1.78 0.9973 3.9999 3.82 0.9693 2.15 0.9999 3.79 0.82 0.87 0.23 Área X Área X Área X Área X Área 0.79 0.84 0.20 0.1314 3.9641 2.81 0.9625 2.22 2.84 0.13 2.9974 3.7823 1.86 0.9671 2.86 0.18 0.0006 0.85 0.15 2.4443 0.0008 0.16 1.4286 0.80 0.4404 0.1151 0.0008 0.0007 0.19 2.23 Área X 1.19 0.82 0.15 1.80 0.17 0.9977 3.83 0.0007 0.9999 3.0150 0.87 0.82 0.0146 0.1251 0.21 2.22 0.0007 0.17 1.19 0.9974 3.21 0.7967 1.22 0.8051 1.16 2.1292 0.87 0.17 2.1190 0.85 0.4364 0.9979 3.83 0.87 0.7910 1.4168 0.78 0.X 3.0158 0.86 0.0007 0.79 0.0008 0.12 1.18 0.9999 3.4129 0.8078 1.9649 2.0139 0.9678 2.83 0.9656 2.9999 3.1131 0.0170 0.7939 1.9978 3.0166 0.14 0.21 0.20 0.81 0.9999 3.13 1.0154 0.14 1.22 1.81 0.9976 3.84 0.83 0.78 0.84 0.80 0.

02807 0.07559 0.02269 125 .04709 -0.11197 0.07559 0.03120 Sum Sum2 0.06186 0.07121 0.07434 -0.06033 -0.08341 Second Mode 0.06033 Third Mode -0.07121 0.17331 0.86534 0.07434 -0.03004 0.02804 0.06186 0.00389 -0.08341 0.06234 0.APPENDIX B Calculation of mode participation factor from eigenvector matrix Eigenvectors Matrix 0.00079 ΣSum2 = 0.03471 Participation Factor 0.04709 -0.02804 0.03120 Calculation of mode participation factor First Mode 0.

3)).2). lat2=a(s). nos=7 %Source Parameter ellipsoid=almanac('earth'.APPENDIX C Developed program in MATLAB 7. sesn=sesn+1. %Area lat1=a(s). %No.pos=load('A4.pos=load('A2.pos=load('A6. %Earthquake eq=load('eq.2)>=lat1 & (eq(esn. %ocurance period.pos(3.dat'). a(1).:).dat').1)<=lon2).1 for Seismic Hazard Analysis of Bangladesh %1 Source Parameter %amplification factor.dat').pos=load('A8.2)<=lat2) & (eq(esn. a(2).pos=load('A3. %No.eq(sesn.dat').pos(1.2).ellipsoid). end end %Mmax & Mmin para(s).:)=eq(esn.eq(:. a(6).ar=areaquad(lat1.'sphere'). if eq(esn.eq(:. a(8).pos(3.dat').3)).dat').pos=load('A7.1).1). para(s). of Source ampf=1.lat2.'ellipsoid'. sesn=1.lon1.dat'). a(3).dat').Mmin=min(para(s). for s=1:nos.0 yrs=131. para(s). [tneq dummy]=size(eq).dat'). lon2=a(s).Mmax=max(para(s). of Div & Area/Length of Div and Pos of CP of Source Div 126 . for esn=1:tneq. a(4).pos(1.lon2. lon1=a(s).1)>=lon1) & (eq(esn.'km'.pos=load('A1. a(5). a(7).pos=load('A5. para(s).

nqmg(2.:). end %b-Line spy=para(s).0 6. para(s).0 7.dar=para(s).sdivlon).ar/para(s). %N 127 . para(s).lnN=log(spypa). para(s).0 8.nqmg(1. %para(s). divX=(lon1+0.p2).eq(snoeq. para(s). para(s).alfa=abs(fit1.np)=noeq-1.0:0. end end [sizeq dummy]=size(para(s). for sdivlon=1:divSX.5].0.latdy]=meshgrid(divX.0 4. %beta [fit1 gof1 out1] = fit(para(s). end end np=np+1.2)=londx(sdivlat.lnN. [londx. [dummy divSX]=size(divX).5:8. % Magnitude Distribution & fmm mRange=[4.beta=abs(fit1.eq).eqm(noeq. para(s).15):0.ar.divY).'ok').1)=latdy(sdivlat.ndiv.nqmg(2.3:lon2. for snoeq=1:sizeq. div=1.para(s). para(s). np=1.5 8.divY=(lat1+0. para(s).5 6.5 5.pos(div. div=div+1. for sdivlat=1:divSY.3:lat2.lnN'.15):0.ndiv=divSX*divSY. if para(s).5 7. para(s).'poly1'). [dummy divSY]=size(divY). noeq=1.:)/yrs.np)=moeq.sdivlon).3)>moeq.para(s). spypa=spy/para(s).p1).nqmg(1. for moeq=4.:)=para(s). noeq=noeq+1.0 5.nqmg(1. %b_plot=plot(para(s).pos(div.:).eq(snoeq.:)'.

beta*exp(-para(s).M(1)). for slon=1:SX.Mmidpara(s). Y=latu:latin:latb.M(2)-para(s).p ara(s). lonin=0.grdp(slat.fmm=para(s).M=[para(s).3.1).slon).M- [dummy sizem]=size(para(s).slon).beta*(para(s).1)].M=[para(s).v1=(exp(para(s). end para(s).beta*para(s).laty]=meshgrid(X. [dummy SY]=size(Y).2). R(slat. [dummy sim]=size(para(s).slon).Mmax))*para(s). X=lonl:lonin:lonr. grdp(slat. for slat=1:SY. latb=20. for s=1:nos. end %%%%%%%%%%%%%%%%%%%%%%%%%%%%% %2 Distance Calculation and meshgrid latu=26.Y).ellipsoid). for j = 1:sizem-1.alfa-para(s).pos(div.M(j)+para(s).r(div)=distance(grdp(slat. end end end 128 .alfa-para(s).para(s). % Average occurrence Rate: %para(s). [dummy SX]=size(X).nqmg).lon.beta*(para(s).[dummy sizmR]=size(mRange).c1=1/(1-exp(-para(s). mExt=mRange(1.Mmid(j)=(para(s).slon).Mmax-para(s). para(s).slon).3.Mmin)).Mmin)exp(para(s).c1*para(s).lat=laty(slat. para(s). for div=1:para(s).prb=para(s).c1*para(s). lonl=88.beta*para(s).s).6. [lonx. lonr=93.6.ar. latin=-0. %para(s).M).ndiv.slon.nqmg(1. grdp(slat.beta*(para(s).M(j+1))/2. para(s).Mmin)). fm=para(s).:) mExt(1.sim+1:1:sizmR).slon).lat.nqmg(1. para(s).lon=lonx(slat.pos(div. para(s).beta*exp(-para(s).:) mExt].Mmin))).fmm*(para(s).

slon.div.s).Mmid(i)-6). h=30. P(slat.(1993).a)*para(s).pr(a)=0.slon.038.s).777. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %step-3: Ground motion estimation: %Boore et.s).div. [dummy sizeacc]=size(acc).slon. +b5*(log10(sqrt((R(slat.205). b4=0.s).pg(i.01:0. for slon=1:SX. for div=1:para(s).slon. PR(slat.5/para(s).fmm(i)*0.s).p(i. al.216.slon. for slat=1:SY.div.158.7].^2)+b4*(sqrt((R(slat. for a=1:sizeacc.pga(i.ndiv.r(div)^2+h^2))))+b6*Gb+b7*Gc.pr(a)=PR(slat.pga(i.r(div)^2+h^2))).slon.end.ndiv.a)=b1+b2*(para(s). PGA(slat.slon.s). for i=1:sizemd. end end end for a=1:sizeacc.slon. b7=0.s). b6=0.a)=1-normcdf((log10(acc(a))log10(PGA(slat.05:0.^PG(slat. Gb=0.s).254.s).s). end 129 . PR(slat. b3=0.Mmid(i)-6)+ (b3*(para(s). acc=[0.div.pr(a) +P(slat. for i=1:sizemd.slon.. for s=1:nos.a)))/0. PG(slat. b5=-0.. [dummy sizemd]=size(para(s).s).a)=ampf*10. b2=0.slon. b1=-0. for div=1:para(s).div. Gc=0.div.pg(i.a).slon.p(i.Mmid).ndiv.

PR(slat. for div=1:para(s).s).pga(i.p(i.s).div. end end end for a=1:sizeacc.s).end end end end end %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %step-3: Ground motion estimation: %McGuire(1978).div.ndiv.a)*para(s). %for alluvium h=30. for s=1:nos.0306*exp(0.a)=1-normcdf((log10(acc(a))log10(PGA(slat.slon.7]. P(slat.slon.div.a)=ampf*0. capS=1.fmm(i)*0. for slon=1:SX.2*capS). for i=1:sizemd.slon.s). end end end end end end 130 .slon. for slat=1:SY.slon.01:0. PGA(slat.pga(i.pr(a) +P(slat.s).05:0.Mmid(i))*(R(s lat.205). for a=1:sizeacc.5/para(s).s). PR(slat.r(div)^2+h^2)^(-1.s).p(i.ndiv.slon.slon.17/2)*exp(-0.s).pr(a)=0. acc=[0. for div=1:para(s).slon. [dummy sizeacc]=size(acc).ndiv.89*para(s). for i=1:sizemd. [dummy sizemd]=size(para(s).div.Mmid).pr(a)=PR(slat.a)))/0.

131 . for a=1:sizeacc.slon.slon.slon.5/para(s). for slon=1:SX.p(i.s).fmm(i)*0.pr(a) +P(slat. for i=1:sizemd. PR(slat.2).308*para(s).s).p(i.slon.slon.slon.pga(i. PR(slat. %for alluvium h=30. for s=1:nos.ndiv.%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %step-3: Ground motion estimation: %Duggal. capS=1.slon.div. end end end end end end %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %4 Hazard Curve Development slon=1:SX.s).pga(i. for slat=1:SY.s). P(slat.s). end end end for a=1:sizeacc. for div=1:para(s). PGA(slat.pr(a)=0.50]. [dummy sizeacc]=size(acc). acc=[0.a)=ampf*227/981*10^(0.slon.div.205). for i=1:sizemd.a)=1-normcdf((log10(acc(a))log10(PGA(slat.01:0.ndiv.a)))/0.r(div)+h)^(-1.s).s).div.Mmid).pr(a)=PR(slat.Mmid(i))*(R (slat.s). for a=1:sizeacc.div. for slat=1:SY.a)*para(s).ndiv. [dummy sizemd]=size(para(s). for div=1:para(s).05:02.

prb(a)*(1- %%%%%%%%%%%%%%%%%%%%%%%%% %5 Contour Grid pt PGA prbty=0.slon).[87 93]).slon).001 for slat=1:SY.2).prb(ac)-cprb(slat. end end %%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %6 Map of Bangladesh % Earthquake epicenter axesm eqacylin.prb(a)=1. cprb(slat.prb(ac)>prbty ac=ac+1.prb(a)=1-cprb(slat. end grdPga(slat.slon)=acc(ac-1)+(acc(ac)-acc(ac-1))/ (cprb(slat. cprb(slat.s). end ac=1.slon).slon. end cprb(slat.slon).slon).'MapLonLimit'. gridm('MapLatLimit'.slon).slon).slon.pr(a)). for s=1:nos. for s=1:nos.[20 27]..slon).prb(a)*(1-PR(slat.cprb(slat.slon).prb(a)=cprb(slat. mlabel('MLabelLocation'..'PLineLocation'. gridm('MLineLocation'. for slon=1:SX.slon). PR(slat.[20 27].prb(ac-1)).slon).'FLonLimit'.s).prb(ac-1)). end cprb(slat. *(prbty-cprb(slat. end end cprb(slat.pr(a)).slon). while cprb(slat.prb(a)=cprb(slat. framem('FLatLimit'.[87 93]).prb(a)=1-cprb(slat.1).prb(a).prb(a)=1.1. 132 .slon). for a=1:sizeacc.prb(a).slon).

1). % River3 riv3=load('river3. plotm(i1lat. eqlat=eq(:. r2lon=riv2(:.i1lon.1).r3lon. 'LineWidth'.0.2).2).1.5).2).'k'. % Site cit=load('city..'YLoc'. eqlon=eq(:. %plotm(eqlat.5). % River riv1=load('river1.'k'.0:10:50). 'LineWidth'. % Island1 isl1=load('island1. r3lon=riv3(:.'XLoc'.. 'LineWidth'.dat'). r2lat=riv2(:.dat'). 'LineWidth'.'k'.2).2).1).1).dat'). r3lat=riv3(:. plotm(r1lat. plotm(r3lat. i1lat=isl1(:. plotm(r4lat.r2lon. % River2 riv2=load('river2.2).%LatLon lat=bord(:.345.. r4lat=riv4(:.dat'). 133 .lon. %Earthquake eq=load('eq. % River4 riv4=load('river4. plotm(lat.'Mino rTick'.'MarkerSize'..'k'.'..'k'.54. northarrow %Bangladesh border bord=load('bordr.'k'. i1lon=isl1(:.k'. 'LineWidth'. r1lat=riv1(:. 'LineWidth'.2).r1lon.r4lon.'MajorTick'.dat').2).1).dat'). plotm(r2lat. r1lon=riv1(:.plabel('PLabelLocation'. scaleruler('Long'.2).eqlon.5).1).5). lon=bord(:.5).90. r4lon=riv4(:.5).dat').dat').1).0:50:200.

nelat] = meshgrid(newlng.klon.cilat=cit(:. rlat=cit(4.1 .1.2)+0.'MarkerSize'.newlat).6).6].1)-0.6). rlon=cit(4.1)-0.8).sylon. newlng=[88:.2)+0. textm(klat. textm(sylat.convec. %%%%%%%%%%%%%%%%%%%%%%%%%%% 134 .6). dlat=cit(1. %clabelm(C.12 .'-k').0.1.2.'*k'.grdPga.2).'FontSize'. set(text_handle.2 . textm(rlat.'SYLHET'. klon=cit(5.2)-0.newPga.'DHAKA'.clon.'LabelSpacing'.1). newPga=interp2(lonx. %%%%%%%%%%%%%%%%%%%%%%%% %7 smoothing contour newlat=[26.1. textm(clat.1)-0.'FontSize'.25 . klat=cit(5.'EdgeColor'.0.'FontSize'.2.3] [C.'KHULNA'.rlon.'CHITTAGONG'. sylon=cit(3.01:20. text_handle = clabelm(C.nelng.dlon.'FontSize'.'black').laty.3.h).2. [nelng.1)+0. textm(dlat.6).h. plotm(cilat.h]=contourm(nelat.6:-0.nelng.2.165).'spline').7).'RAJSHAHI'. %%%%%%%%%%%%%%%%%%%%%%%%%%%% %8 Countour Map at 10% probability of exceedence in 50 years period convec = [.2. clon=cit(2. clat=cit(2.2)+0. sylat=cit(3.2)-0.nelat.'FontSize'.cilon. dlon=cit(1.'rotation'.15 . cilon=cit(:.08 .01:93].1)-0.

APPENDIX D Gutenberg-Richter b-line for each source 135 .

136 .

137 .

138 .

- Art 19 Ardelean uUploaded byHaider Shah
- California Seismic Hazard MapReportUploaded bynotavalidemail1952
- Time-History Analysis on Seismic Stability of Nuclear Island Bedrock with weak InterlayerUploaded byInternational Journal of computational Engineering research (IJCER)
- Reliability-Based Structural Design Optimization for Nonlinear Structures in OpenSeesUploaded bymgrubisic
- Urban LamadradUploaded byAvaluador Edixson Villalobos MSc
- Preliminary Seismic Hazard Map of Peninsular IndiaUploaded byRajeev Dua
- Comparative Study of Typical R.C. Building using Indian Standards and Euro Standards under Seismic ForcesUploaded byGRD Journals
- NGA Models Version1 4Uploaded byFabian Andres
- 14_05-02-0067Uploaded byVikas Biradar
- ATC-40Uploaded byAZ314
- Geohazard EarthquakeUploaded byasrafil
- Bsds SeismicUploaded byMar Martillano
- 2KL17CSE10Uploaded byvaibhav
- 01-1057.pdfUploaded bymy09
- Performance Based DesignUploaded bybinud das
- 211CE2238Uploaded byDeb Pradhan
- Zha 2008Uploaded byGerman Rodriguez
- EarthquakesUploaded byrida adila
- Tunnel FormUploaded byMarshita Ramlee
- Improving the Seismic Resilience of Existing Braced-Frame Office BuildingsUploaded byJoaquim Tchamo
- The New Brazilian Standard for Seismic DesignUploaded byjmcc2
- EarthquakeUploaded byanju suresh
- Seismic Analysis of Cantilever Retaining WallsUploaded byJosé Céspedes
- A Critical Examination of Near-Field AccelerogramsUploaded byshiva
- Israeli Standard SI 413Uploaded byrrrrmmmm1
- Franja Institucional UNIMETA 2012Uploaded byyesthuro
- Pushover 3Uploaded byEnrico Pranoto
- San Andreas Act1Uploaded byMaestro de Grapico
- 2 RT4b Lecture 2016 Lasse Rajala - OKUploaded byalexitomichael
- V4I5-IJERTV4IS051043Uploaded byJoaquim Minusse Tchamo

- Geophysical investigationUploaded byHrishikeshShahane
- [Udias a., Madariaga R., Buforn E.] Source MechanismUploaded byDavid Alejandro Segura Sabogal
- ch06 Earthquakes and Related PhenomenaUploaded byFranc
- basicsofsurveyingUploaded byRomeo Corneliez
- Seismic Analysis for offshore structuresUploaded byBharathiJeganathan
- island chain with no Age ProgressionUploaded byg.ponmudi
- Height Datum & Height Determination Using GNSS in SingaporeUploaded byMarco Conopio
- Eathquake EngineeringUploaded byJimmy Kassady
- The Basics of Earthquake LocationUploaded byvibhore15
- Heuret y Lallemand 2005Uploaded byMarcos
- EpicenterUploaded bymojeebmashal
- Earth Move Link 1Uploaded byRon Hart
- Reservoir Characterization Using Seismic InversionUploaded byMASAGUS MANGKU GAMA
- plate tectonics web quest studentUploaded byapi-264222897
- Plate TectonicsUploaded bySangita
- plate tectonics web quest studentUploaded byapi-330049991
- Tectonic Map Mexico 2013 Final VUploaded byknightrider2010
- gd4207-gdlaut-3-2015Uploaded byGina Andryana
- CUENCAS SEDIMENTARIASUploaded byGeovys Ch
- bumi sebagai sferaUploaded bySelvarani Themathsmaniac
- cvUploaded byapi-359839885
- IntroductionUploaded bySamson Ayeni
- Global Tectonics and Basin FormationUploaded byNick Bale
- OFR 2013-15 Alberta Earthquake Catalogue, Version 1.0: September 2006 through December 2010Uploaded byAlberta Geological Survey
- TDEM_NearSurfaceMappingUploaded byYoqie Ali Taufan
- Photosat Lidar Drones Gps Gm 20oct2016Uploaded byБојан Цвијановић
- lec06.pdfUploaded byCarlos Arturo García
- IsostasyUploaded byKedar Bhasme
- 1976 earthquake report.pdfUploaded byMarcoRimolaGutierrez
- Cross Hole SeismicUploaded byPriscilla Howell