Seismic Design of Buildings to Eurocode 8 in Singapore 4-Day Workshop at BCA Academy on 22"¢ April 2015 Professor Koh Chan Ghee Dept. of Civil and Environmental Engineering National University of Singapore Email: [email protected] Course Outline Fundamentals of Seismic Analysis + Fundamentals of structural dynamics + Vibration of single-degree-of-freedom systems and multiple-degree-of-freedom systems + Characteristic of earthquake motion + Seismic design philosophy Eurocode 8 + Key requirements by Eurocode 8 and Singapore National Annex + Overview of Guidebook for Design of Buildings in Singapore + Determination of basic parameters for Eurocode 8 and Singapore National Annex + Evaluation of structural regularity in plan and elevation + Analysis methods: Lateral Force Analysis and Modal Response Spectrum Analysis Design Check DOL =n yee! ofebiley eg eireiren + + Desirable features of earthquake resistant buildings Jd + Seismic design and detailing rules + Foundation design to Eurocode 8 + Limitation of structural drift and structural separation for buildings + Summary of procedures in BC3 Demonstration of seismic analysis using computer analysis software (£7 <' ) = Modelling issues related to computer analysis software + ilustration of seismic analysis results by computer analysis software e © G Koh 2015 2Fundamentals of Structural Dynamics [Fundamentals | * Dynamic loading: magnitude, direction and/or position varies with time = Newton's Law of Motion: F=ma = mais known as the inertia force * Force > Acceleration > Change in velocity & displacement with time + Response (e.g. displacements, forces, bending moments) are function of time. * Much more difficult and time consuming than solving static problems. Examples * Vibration of buildings due to wind and earthquakes * Vibration of bridges and railways due to traffic load * Defence structures against blast and impact * Offshore/marine structures -- current, waves. * Floor systems due to machines and human footfall * Robustness of buildings by sudden column removal, etc ©CGKon 2015 3 | Fundamentals Types of Dynamic Analysis fundamental Deterministic Dynamics: + Load can be defined (though they may be highly irregular such as earthquakes), + Complete time histories can be determined for the given dynamic load * Fora prescribed input, the response is uniquely determined. * For seismic analysis, it is not sufficient to consider one earthquake record as input. Stochastic Dynamics: + Load can only be defined in a statistical or probabilistic way. + Earthquakes are highly uncertain and should be defined in probability form + Hence responses can only be obtained in probability form * More rational approach because of randomness, but more difficult to do. * _ Inseismic design codes, probabilistic approach is used in certain aspects (defining ground motion) but the dynamic analysis is deterministic. ©coKon 2015, ‘SDOF | System Single Degree-of-Freedom (SDOF) System Spring-mass-dashpot system (‘oscillator’) Displacement Spring represents structural stiffness u(t) rk p(t) Dynamic force “Dashpot” accounts —* © for energy loss Consider a linearly elastic system, i.e. m, c and k are constant. Equation of motion oo. mii+cu+ku= p(t) ©cG Kon 2015, | spor | | System Vibrations Caused by Support Motion Vibration of sensitive instrument Vibration of building vibration due to floor vibration due to ground motion u,(t) Total Ground Relative displ. displ. displ. u,(t) = Ug(t) + u(t) Ground displ. G Koh 2018.SDOF | “Effective Force” Due to Ground Motion System | Equation of motion: Damping force and elastic force depend on relative (not total) motion N | cu+ku =0 netorence , Uelt) Z u(t) Inertia force = Mass x Total Accel = mu+cu+ ku Effective force due to support motion is inertia-driven. Not a physical force (unlike wind Uu ( t) Load or blast load, for example). 9 ©cGkon 2015 7 i NCEE ros deen) Sree} \ } Damping * To account for eneray loss (in a simplistic way) > Thermal effects of repeated straining 7 Friction at connections, between structure and non-structural elements > Opening and closing of microcracks, etc + Inpractice, difficult to model energy loss exactly = Need to resort to experimental/field observations Different damping models: viscous, hysteretic, Coulomb friction, fractional derivative, etc. The most commonly used is linear viscous damping _ Cc © 2mo > Damping ratio (dimensionless): ¢ > Natural frequency in radian/s: o=Vkim > Natural frequency in Hertz (cyclels): f= w/(2r) > Natural period in second: T=4F ©cGkon 2015, 8[Damping | Energy Dissipation and Damping Ratio + Free vibration - use the decay rate of amplitudes to estimate damping ratio. = Plot In(peak displacement) versus cycle and do linear fit * Cyclic loading test - damping is related to energy dissipation which is the area of the hysteresis loop of force-deformation cycle. Force Deformation = Other methods: Half-power (bandwidth) method, Nyquist plot method, random decrement method, etc © CG Koh 2015 a Damping | Typical Damping Ratios for Buildings Serviceability | Welded steel, prestressed concrete, RC with | 1-3% (Damping can be | slight cracking lower for low-level vibration, e.g. for | RC with considerable cracking 3-5% human-comfort consideration) _| Bolted/riveted steel 5-7% [Atornear yield | Welded steel, prestressed concrete without | 5-7% point complete loss in prestress | | RC structures, prestressed concrete with 10% | | complete loss in prestress | Bolted/riveted steel, or wood structures with 10-15% nailed/bolted joints For structural systems with passive dampers or base isolation bearings, damping Is usually higher and depends on the materials and devices. Usui, SP-C 0-05) 2 ©cGkohPeriodic Loading * Harmonic — a pure sine (ideal) a * Complex — e.g. most M&E equipments, regular footfall P(t) CG Kon 2015 " Response due to Harmonic Loading P(t) miui+ cu+ ku = py sin(wt +) Excitation Phase angle ; frequency For ®=0, the solution is Transient v where @) = ii — unit impulse + The response is called the unit impulse response, /(t), or known as the “Green function” h(t) = ——expl- Coot sin(wot) Mop, ©cGKon 2015, Analytical Method mii+ cu+ ku = P(t)=-muig + The dynamic load can be represented as a sequence of impulses Duhamel integral: j PO uit) [ play htt—s)ar oe = Unit impulse response: t hit ~ x) = —-exp(- colt -2))sinlop(t ~ 9) Response mop op =o for one impulse = Damped natural frequency tot + This approach is mathematically elegant but computationally inefficient ©cGkon 2015,Numerical Integration Method + Time-stepping, or step-by-step method * Numerically integrate equation of motion in an approximate way. * E.g. Newmark method, Wilson method, HHT method, Central Difference method, Euler method, Runge-Kutta method, etc. = Advantage: Can be extended to nonlinear systems. * Disadvantage: Time consuming for multi-DOF systems if used in full analysis. * Mode superposition method would help significantly by decoupling the full system to multiple SDOF equations. ©CGKon 2015 Frequency Domain Method Method * Basis for spectrum analysis, e.g. for ground motion & machine specifications. = Convert time-domain problem into frequency domain. «Extension of Fourier series method — decompose time signals to harmonics of various frequencies. = Use an efficient algorithm called the Fast Fourier Transform (FFT) to obtain the spectrum. Time history of measured acceleration FFT based on 4 s with 1024 data ©cG kon 2015, celeron ns FFT Amplitude Frequency te) 0Response Spectrum Analys' (RSA) > Very important concept ~ for seismic analysis, but ften misunderstood = * Based on linear dynamic analysis, this ~~ i ' method aims to obtain peak (maximum absolute) response for a given earthquake u ground motion fa * Consider a linearly elastic SDOF system | = The equation of motion in terms of relative displacement (structure relative to ground) is \ | By k,c mii+cu+ ku= -mii,(t) * Divide by m 7 k - ui+—u+—u = -U,(t) m m h Original Displaced ground ground position _ position ©c.6 kon2018 a URSA Structural Parameters Needed — * Since c o=vkim $= the equation of motion becomes Minus sign is not important as earthquake motion direction is uncertain U+ 2cwu+ w*u =U, (t) * Only two structural parameters are needed u(t) (a) Natural frequency (w) or natural period (7 = 277/ w) 2D and (b) Damping ratio (2) = Base shear (elastic force) is given by F,(t) = ku(t) = moult) F(t) ©cGkor 2015, 22RSA Defin * Given an earthquake ground motion, we can obtain the peak displacement Saisp(T.6) =max| u(t)| ion of Response Spectrum (Elastic) (relative to ground) Structure: + If we plot peak displacement versus natural period (7) for a given value of damping ratio, we have the displacement response spectrum: + Similarly, by getting peak velocity (relative) and peak acceleration (total), we can obtain velocity response spectrum & acceleration response spectrum * Let's define Pseudo Spectral Acceleration as ~> SacelT.S) = 0 Saisp(T, 5) * This is approximately the true spectral acceleration for small damping ratio. ©cG Kon 2015 Ey Maximum Base Shear BSA * Max. base shear is fF, 3 = mo*Umax = Ma Sgisg(T.C) = Foman=MSacc(T. 2) + Ignore subscript max and divide the above by structural weight (W= mg): fy SuellS) Fy=C,W where @, = Sel + The dimensionless value C, is called the base shear coefficient or seismic response coefficient * In other words, (Max) Base Shear = Coefficient x Weight * This leads to an approximate static method called the Lateral Force Method (LFM) whereby the base shear is obtained as a fraction of structural weight. ©CG Koh 2015 4Examples of Ground Acceleration Records eee pee ae Mexico City earthquake (M8.1) >10,000 deaths es Epicenter >350 km away eae it. KS SCT SIE Previously no historic record of such a strong earthquake in Mexico. Ref: “Dynamics of Structures” by A\K. Chopra, 2012 2 [ Ground ] The El Centro Earthquake (18 May 1940): | motion Time history of NS Component 150 Acceleration | in/sec/sec | Velocity infoee Displacement Time, sec Ref: http://nisee.berkeley.edu/elibrary/Data/E0001 2Ground Acceleration > Elastic Response Spectrum | Taner Define damping ratio spor , u(t) - ¢ System | l e.g. 0.05 Define natural period T I Solve Equation of Motion WTig | Displacement history u(t) u,(t) —_—- | | Tae Select an earthquake acceleration record as input, eg. El Centro NS Component Obtain peak value ie. max | u(t)| This gives 1 data point: Sysp(T.¢) + = Repeat for all other T values ©CG Koh 2015 a Response Spectra for El Centro Earthquake (NS Component) | RSA Succ = (2RIT) Syej = (20/TY? Seisy 100 30 Damping ratio 0% (top curve) 2% 5% 10% (bottom) 3-way log plot Ref: “Dynamics of Structures" by AK. Chopra, 2012 0. ‘002 005 0.1 02 OS 12 5) 10 20. 50 Natural Period T (sec) cy[rsa | Smoothed Response Spectrum + Use of a single earthquake record is suitable for checking in time domain analysis. * But earthquake motion is random in nature & does not repeat itself. = Asingle earthquake record gives rise to jagged appearance of peaks > peak response is sensitive to the natural period of structure + Smoothed response spectrum is needed and should be obtained from an ensemble of possible earthquake motions. * Statistical analysis is used to find the mean & standard deviation, e.g. mean plus one standard deviation would imply 15.9% probability of exceedance + Furthermore, for ease of use, segments of the spectrum are linearised or represented by simple equations. 34 x cont g ©CG Kon 2015, Example: Newmark-Hall Design Spectrum for PGA = 0.2g [Rsa | Ref, “The Seismic Design Handbook’ Ed. F. Naeim, Kluwer Academic Publisher, 2E (2001) const. v For design, we will plot spectral acceleration in the vertical axis and also linear scale for both axes. Structural Period (s) 20MDOF ‘System Dynamics of Multi-DOF System [System «= SDOF model gives an approximate solution -- the accuracy depends on whether the motion can be adequately described by a single shape "More refined analysis would require MDOF systems. * Finite element method (FEM) is usually adopted for its versatility * Proper structural modelling by FEM is essential * Equations of motion: mii + c+ ku = p(t) + Stifness matrix: Same as static problem. + Mass matrix + Consistent mass approach — same shape functions are used in computing mass matrix & stiffness matrix. + Lumped mass approach — divide the mass equally among the element nodes for translational DOFs; no mass for rotational DOFs. ©CG Koh 2015 a Damping Matrix ez * Difficult to establish analytically = Assume proportional (classical) damping matrix, e.g Rayleigh damping c=aym+a,k 2" stittness in be related to di -— proportional Can be related to damping ratios of proportional any two modes as follows 1] 10; ©; Jfag| _ {Si 2|10, 0, ae gj Damping ratio Z, Mass proportional _— damping (a, = 0) Natural frequency w, ©0G koh 2015 2Effective Forces Due to Ground Motion [System | Let u;, u, and u; be the relative displacements of my Us floors with respect to the ground. ky, C3 I mi + cu + ku m, | |m™ I | ky. ¢ = Poglt) = eG a | |ms| m0, 14 Ground acceleration =k, , Cy | u, KI Original Displaced position position ©CGKoh 2015 . : ih at | Eigenvalue | Eigenvalue Analysis to Obtain Vibration Modes | Analysis + This gives the modal frequencies (w, , 2, ...) and mode shapes (0). which are needed to compute other modal properties. 100019 L 1.0000 1.0000 + 0000 / 06088 Eos re 6700 aoe mm Mode 4 Mode 2 Mode 3 wea s 2311 w= 48.1 rad/seo = The amplitude of mode shape is arbitrary. The mode shape can be scaled by any non-zero constant. Hence modal properties are not uniquely defined (depending on the scale used). Py ° o Kon 2016©cGkon (1) > (2) 3) (4) ©cGkon Eigenvalue Importance of Eigen Solutions me) The natural frequencies and mode shapes are important characteristics of structures in dynamic analysis. They should be obtained first to check the structure — physical (design) and numerical (computer): ¥ Fundamental period should be reasonable (use empirical formula or rule of thumb to check) Y Identify disconnected or improperly restrained members & rectify the computer modelling problem v Identify local modes and suppress them v Severity of torsional mode (use animation in computer software) 2015 35 Mode Superposition Method ‘Superposition (for MDOF Linear Systems) : Solve eigenvalue problem (corr. to undamped free vibration): mu+ku=0 natural frequencies (w,, w,,...) and mode shapes (, . 0, . ) Transform physical coordinates (u) to the eigen-space by letting N U(t)=> o, UMt) = Olt) kat Assume proportional damping matrix, e.g. Rayleigh damping c=am+ak Use orthogonality properties of mode shapes, i.e.. T r % Mo, = o ko, = 0 for kej 2015 36. . im Modal Equations of Motion woe | Superposition > Modal equation of motion (scalar, not matrix) for the k-th mode: MiG + CxGe + Kuda = Pe (t) __ Modal Properties (5) Each modal equation is solved as a SDOF system Mode 1 : ©cGkon 2015, 37 1 . Mode Superposition of Modal Solutions Superposition (6) The computed modal solutions g,(t) are then superposed to give displacement response history (see equation in Step 2). (7) Response quantifies such as shear forces & moments are calculated from their relationship to displacements via structure/element stiffness matrices Notes + The mode superposition method transforms a MDOF system to many independent SDOF systems (involving mode shapes) * Usually contribution of high modes can be neglected. + The main advantage is that the user can include only selected modes, say first J modes out of a total of N modes. For multi-storey buildings due to horizontal seismic action, normally J << N. - Edehs (wade chfeuld = Equation for each mode can be solved very efficiently = Overall this is much faster than direct analysis of full system. eG Kon 2015, 38Mode Solving Modal Equations Superposition * Each modal equation can be solved by: {a) Response history analysis in time domain (via step-by-step method) (b) Response spectrum analysis (suitable for estimating peak response, widely used for seismic design) (c) Frequency domain analysis (with the help of Fast Fourier Transform) + Caution: Do not confuse (b) and (c). Response spectrum method is not a frequency domain method. * Method (a) gives response time history of each mode considered (q,). The total response history is then obtained by superposition via mode shapes: J u(t)= > 6, Qp(t) where Jis the number of modes nm” included in the analysis pone? Yypyene VAY ©cexenaos ys 7 39 Combining Peaks of Different Modes Mode Superposition = If only the peak response is needed, Method (b) can be used. "The peak response value of each mode considered is obtained the same way as in response spectrum analysis of SDOF system. «For each mode, the corresponding mode shape is used to obtain the corresponding peaks at different DOFs. * But the peak values for different modes do not necessarily occur at the same time * Furthermore, response spectrum analysis gives maximum value without + sign = Simply adding the peak values from different modes (“Absolute Sum Rule’) tends to be too conservative. E.g, base shear: J a vere = S°1Vi 1°" combining modes 1 to J a * Alternative combination rule is required to give a more reasonable estimate of the total peak response from modal peak values (to be explained later) ©cG Kon 2015, 40"Mode Alternative to Eigen Vectors | Superposition | = Numerical effort required to compute eigen vectors can be time consuming for large systems. * Eigen vectors correspond to free vibration mode shapes, without paying any regard to the spatial distribution of loading + Many eigen vectors that are orthogonal to loading do not participate in dynamic response * Hence the convergence may not be efficient in terms of the number of modes required to reach certain mass participation ratio. = Alternatively, it is possible to compute a set of stiffness & mass orthogonal Ritz vectors with less modes needed than eigen vectors for convergence * Reason: Load-Dependent Ritz vectors are generated by taking into account the spatial distribution of dynamic (seismic) load. ©CGKon 2015 “ Direct | Direct Analysis of MDOF Systems Integration | (For Linear and Nonlinear Systems) = The matrix equations of motion for the full system can be integrated directly in time domain, usually by numerical integration methods (step-by-step). * Advantage: Applicable to nonlinear systems e.g. large displacement, nonlinear material, damage, and non-classical damping = Solution accuracy depends on the time step (relative to the natural period of the system) — the smaller the more accurate. + Some methods are conditionally stable, i.e. solution will blow up if time step is not small enough, e.g. Central Difference Method * Disadvantage: Computationally expensive for large systems * Also, the solution is not accurate for high modes + Some methods introduce numerical damping, e.g. Wilson-8 method, which would be good to damp out inaccurate high modes. @cGken 2015 2{Direct | Step-By-Step Methods | Integration + Solve equations of motion one step at a time in an approximate way by making certain assumption of the solution in each time step. + Step-by-steps can be extended to solve nonlinear equations of motion incrementally, but need iterative methods to satisfy equilibrium (and hence more time consuming) * Newmar Method: (a) Constant (Average) Acceleration Method -- Assume acceleration is constant in each time step. (b) Linear Acceleration Method: Assume acceleration varies linearly in each time step, i.e rep a(t) =a, hares (a.,-a) fort, equilibrium is based on un-deformed (original) configuration + Second-order analysis takes into account displacement (if significant) caused by loading. Equilibrium is updated based on deformed configuration = In most cases, axial forces on vertical members remain approximately the same after doing P-A analysis once. No iteration is required + Incases where axial forces do change significantly due to relatively large A, iterations are required, i.e. repeat P-A analysis based on updated stiffness until convergence. = Iterative solutions are more accurate but more time consuming. * For practicality, a predominant load case is usually defined for computation of stiffness matrix accounting for P- A effect, and this is then used for all other cases. ©cG Koh 2015, 4 Characteristics of Earthquake Motion * Anearthquake is the result of a sudden release of energy in the earth crust that creates seismic waves. = = These waves radiate outward from the source ("focus") & shake the ground = A fault is a large fracture in rocks, across which the rocks have moved. It can be up to thousands of km long and tens of km deep. + There are many types of earthquakes which have different characteristics (duration, frequency content, depth, etc) Sumatra strike-slip fault Sea Eurasian Indian- Sy Plate Australian Ss te ‘Sunda Arc subduction fault 50Sumatra Faults + The Indonesian region is one of the most seismically active zones in the world. = The closest fault to Singapore is the Sumatra fault ~~ and the closest point to Singapore is ~400 km. + The next fault is 700-800 km away: Sunda Arc - subduction fault at the Java trench (where the Indo- Australian plate subducts under the Eurasian plate at a rate of 60-70 mm per year) References: + Balendra et al (2002) “Analysis of long- distance earthquake tremors and base shear demand for buildings in Singapore" Engineering Structures, V24, pp. 99-108. + Megawati and Pan (2009) “Regional seismic hazard posed by the Mentawai segment of the Sumatran Megathrust’, Bulletin of the Seismological Society of America, V99, pp. 566-584 + Pappin et al (2011) “An approach for seismic design in Malaysia following the Principles of Eurocode 8", Jurutera, Institution of Engineers Malaysia, pp. 22-28 st Australian Plate Earthquake Magnitude and Intensity Magnitude (at source) = Richter Scale is a measure of earthquake magnitude by measurement of ground motion (via seismometer) + E.g. Ground movement is 10 um for a quake of M = 1 at 100 km from the epicenter, and 100 um for M=2, etc. * One magnitude higher > Amplitude of ground shaking is an increase of 10 times in displacement and 31.6 times in energy + For medium to large earthquakes, use moment magnitude. Intensity (at target) + Intensity measures the strength of shaking at a certain location + Example: Modified Mercalli Intensity (MMI), Japan Meteorological Agency (JMA) + Intensity becomes weaker away from the epicenter (attenuation) due to wave scattering and energy loss. + Attenuation relations are empirical in nature ©cGkon 2015, 2Modified Mercalli Intensity (MMI) [Earthquake | An intensity scale based on human perception and observation of damage (12 scales) |. Not felt except by a very few under especially favourable circumstances Il. Felt only by a few persons at rest, especially on upper floors of buildings. Delicately suspended objects may swing IV. Felt indoors by many, outdoors by few during the day. At night, some awakened. Dishes, windows, doors disturbed; walls make cracking sound. Sensation like heavy truck striking building. Standing motor cars rocked noticeably. VI. Felt by all; many frightened. Some heavy furniture moved; a few instances of fallen plaster. Damage slight. VIII. Damage slight in specially designed structures, considerable in ordinary substantial buildings with partial collapse. Damage great in poorly built structures. Fall of chimneys, factory stacks, columns, monuments, walls. Heavy furniture overturned - X. Some well-built wooden structures destroyed; most masonry and frame structures destroyed with foundations. Rails bent. XIl. Damage total. Lines of sight and level distorted. Objects thrown into the air. ©CGKon 2015 3 Very Approximate Relationship for Peak [Earthauake | Ground Acceleration (PGA) PGA(g) |] Duration PGA (g) - (sec) <0.03 os) 2 Vv 0.03 - 0.08) ons fe vi 0.08 - 0.15 0.22 aa vit | 015-025 0.29 ie) vill 0.25-0.45 0.37 2a) IK 0.45 - 0.60 0.45 30 x 0.60 - 0.80 8.0 05 34 xl 0.80 - 0.90 8.5 08 37 Xi 2 0.90 + The above gives a very rough correlation mainly based on Califoriia earthquakes. There were exceptions, e.g. 1989 Loma Prieta earthquake magnitude was 7.1 but PGA was as high as 0.659 * MMI is based on observed damage which depends on not just PGA but also many other factors, e.g, structural design, soil condition and construction quality, 15 54Japan Meteorological Agency (JMA) Scale [Eartueke | + _JMA scale is a measure used in Japan to indicate seismic intensity, in units of shindo (literally "degree of shaking”). * Historically determined by human perception (like MMI) but in early 1990s moved towards instrumental seismic intensity. * Computation from measured accel. time histories to JMA scale is complicated, taking into account three directions, time duration and frequency content. * No simple and linear relationship between JMA scale and PGA. The table below gives only approximate correlation + Ref: http:/www.jma.go.jp/ima/en/Activities/intsummary.pdf IMA Scale [ PGA(mis*) | JMAScale | PGA (mis) [ | 9 oo) < 0,008 | S-lower(4.5-4.9)| 0.80-140 1(0.5-1.4) | 0.008-0.025 | S-upper(5.0-54)| 1.40-2.50 2(1.524) | 0025-008 | étower(5.5-59)| 250-315 3125-34) | 008-025 | 6-upper(6.0-64)| 315-400 403544) | 025-080 | 7165-up) > 4 mist ©cGKon 2015 . ee ing Design Earthquake — type, magnitude, geology, depth, duration, frequency content, distance and attenuation > Ground motion parameters * Site/soil effects — local soil properties that affects ground motion frequency with possible amplification > Ground type (or site classification) * Structural dynamics of buildings — structural system, configuration & properties (e.g. mass, damping & stiffness) > Analysis and design + Soil-structure interaction not easy, usually not considered except for buildings with large basements/foundation systems Seismic Analys From Source to Bi , Buildings Le yo Surface wre } Epicentre sacs ) oe } ee sot Continental crust « oe ate 7 (bedrock) @ ©CGKoh 2015Lessons from “Unexpected” Earthquakes | Seismic Design Earthquake Consequence Remarks 1976 Tangshan EQ | Death doll > 240K, _| Previously considered a low seismicity (M75) severely wounded —_| region > Major upgrade of national >160K seismic code in China 1989 Newcastle EQ | 13 killed and 50,000 _| No earthquake requirement previously (M5.6) buildings damaged. _| This was introduced in Australian codes Financial cost ~ AS4B | in 1993. 2010-2011 186 killed The fault was not known until it ruptured Christchurch EQ | Widespread damage _| in 2010. (7.1, M6.3) in city. Rebuilding cost ~NZS40B 2011 Virginia EQ | Damage cost Surprisingly felt over a large area (M5.8) in eastern | ~US$300M. including Canada > Revision of ground USA’ motion prediction equation 2011 Japan Tohuku | Triggered tsunami and | This was the largest earthquake EQ (M9.0) nuclear accident (unprecedented) in Japan's history and Death/missing >19K. | was not anticipated in Tohuku region. A reminder of uncertainties of earthquake. ©CGKoh 2015 37 Seismic Design Codes * Seismic codes are mainly about “survivability” (life safety) and not “serviceability” except for critical structures** in major earthquakes. + The intent is to provide adequate strength and stability so that the structure will not have life-threatening collapse in maior earthquakes. + Itis economically prohibitive to design structures in elastic range in major earthquakes, Thus, damage and large deformations are usually allowed * Seismic codes provide “reasonable” but not complete assurance of protection of life + Level of protection can be increased by increasing the design lateral force ductility (energy dissipation capacity) and redundancy + Each major earthquake has taught new lessons. Thus seismic codes are regularly revised (e.g. every 3 years in USA) For critical structures (e.g. hospitals and fire stations), a higher requirement of design force is imposed via importance factor. ©CGKon 2015 58Typical Earthquake Design Philosophy Intensity Non-structural ‘Structural (Frequency) | Damage Damage Performance Check Minor N _ (Frequent) No lo Damage limitation Mederatsl (seismic serviceability) for (Occasional) Yes (limited) | No, or insignificant | Teturn period ~100 years Major May be significant No collapse ae Yes but no local or | for return period ~500 (Rare) global collapse years We are dealing with low-probability high-consequence events. Seismic design codes usually require explicit check of two performance levels, i.e. ultimate limit state and limitation limit state. By satisfying the above 2 performance levels and also observing good principles in earthquake design, a third performance level is implicitly satisfied, i.e. to prevent global collapse during a very rare and catastrophic earthquake (return period ~2500 years). Damage may be severe beyond economical repair ©CG Kon 2015 59 Earthquake Probability & Return Period Design If P, is the probability that an earthquake ground motion of a given intensity** will be exceeded in any given year, then the (mean) retum period is T= 1/P, If a building is designed for N years, the probability that an earthquake ground motion of the given intensity will be exceeded during its lifespan is } = -ero{- %) if N and T are large IFN = 50, T= 475, then 614? Psp = 1-|1-= =| =0.1=10% \ 475) IfN = 50, T = 2475, then =) 0" 2478 ) The ground motion intensity corresponding to a return period (say 475 years) depends on the seismicity of location E.g. A return period of 475 years could correspond to ground motion of 0.4g in high seismicity region and only 0.05 g in low seismicity region Peg = =0.02 =2% ©cGKon 2015 60Seismic Hazard Levels Considered Exceedance Time | Return Period | Hazard Levels Considered Probability (P,) | Span (N) (7) In Seismic Design Codes 20% 10 years | 45 years 10% 10years | 95 years “Damage Limitation Requirement’ in EC8** 20% 50 years | 225 years 10% SO years | 475 years “No-Collapse Requirement” in EC8** Design EQ in UBC97 Design EQ in 1BC2012"* 5% 50 years | 975 years 10% 100 years | 950 years Max. Capable EQ in UBC97 2% 50 years | 2475 years | Max. Considered EQ in IBC2012 ** Based on the recommended value of 10% for Pog and Pryce in ECB ** Approximate equivalence (DE = 2/3 * MCE in IBC2012) ©cGkon 2015 6 Overview of Eurocodes SS EN 1990 | Basis of structural design SS EN 1991 Actions on Structures (EN = Euro Norm) SS EN 1992 SS EN 1994 Concrete Steel Composite SS EN 1993 SS EN 1997 | Geotechnical EN 1995, EN 1996 EN 1999 } SS EN 1998 | Earthquake Timber Masonry Aluminium . scnnnenenene This requires understanding of + structural dynamics | + Each standard is accompanied by National Annex (NA) and possibly some other documents e.g. "Published Document" (non-contradictory complimentary info), product standards and execution standards.Ec8 Eurocode 8 Design of Structures for Earthquake Resistance The EN 1998 (2004) or EC8 contains six parts. Part 1: General rules, seismic actions and rules for buildings Part2: Bridges Part 3: Assessment and retrofitting of buildings Part 4: Silos, tanks and pipelines Part 5: Foundations, retaining structures and geotechnical aspects Part 6: Towers, masts and chimneys cocoodg = Each member state has to produce its own National Annex to define Nationally Defined Parameters (NDP) and seismic hazard map. * In Singapore, only Part 1 has National Annex ©CGKon 2015 6 SS EN 1998-1 Ec8 Design of Structures for Earthquake Resistance — Part 1: General rules, seismic actions and rules for buildings * Section 1: General * Section 2; Performance requirements and compliance criteria * Section 3: Ground conditions and seismic action * Section 4: Design of buildings * Section 5: Specific rules for concrete buildings + Section 6: Specific rules for steel buildings * Section 7: Specific rules for composite steel-concrete buildings * Section 8: Specific rules for timber buildings [No national annex in Singapore) * Section: Specific rules for masonry buildings [No national annex] = Section 10: Base isolation of structures [No national annex] = Annex A: (Informative) Elastic displacement response spectrum (For info only] = Annex B: (Informative) Determination of the target displacement for nonlinear static (pushover) analysis [May be used] + Annex C: (Normative) Design of the slab of steel-concrete composite beams at beam-column joints in moment resisting frames © CG kon 2015 oDocuments for Earthquake Resistant nv | Design of Buildings in Singapore BC3 | + The Singapore Standard for earthquake design is SS EN 1998-1 (2015) “Eurocode 8 : Design of structures for earthquake resistance. Part 1, General rules, seismic actions and rules for buildings” (herein referred to as ECB) + The required National Annex is NA to SS EN 1998-1: 2015 “Singapore National Annex to Eurocode 8: Design of structures for earthquake resistance ~ Part 1 General rules, seismic actions and rules for buildings” (herein referred to as NA) + The national annex defines the nationally determined parameters and decisions on the two informative annexes of EN 1998-1 * Certain clauses refer to the following BS! “Published Document’ for guidance PD 6698:2009 “Recommendations for the design of structures for earthquake resistance to BS EN 1998” + BCA has published the following reference documents as design quide (available at www.bca.gov.sg/publications/publications.html) 1. BC3:2015 - Guidebook for Design of Buildings in Singapore to Requirements in SS EN 1998-1 (herein referred to as BC3) 2. Example Calculations to the Requirements of BC3:2015 ©CGKon 2015 65 Summary of Methodology in BC3 Bes * BC3: 2015 is a guidebook for design of buildings in Singapore to EC8 requirements. * A flowchart is provided to summarize the recommended methodology. Sag height i cemined acccdng e Gane 2 Building height He aonees?| 77 % y tt cava aaenan cesar | Ground type | _ Sen Tye winning lpi atrminad accurate use 2 Building | «c-sssr suises type | choice of | method Analysis & design check (ULS and DLS) 6sE Fundamental Requirements (1) 24 rement (NCR) ollapse Ret * To withstand the design seismic action without local or global collapse, thus retaining its structural integrity and a residual load bearing capacity after the seismic events. = Design seismic action is expressed in terms of. (a) reference seismic action associated with a reference probability of exceedance, Prog, in 50 years (or reference return period Tyo), and (b) importance factor y, of building (‘reliability differentiation”) + Recommended Pyca= 10% in 0 years corresponds to Tyce = 475 years — this is adopted in Singapore National Annex (NA) + Reference peak ground acceleration (agq) corresponds to the reference return period Tycp Of the seismic action for the no-collapse requirement — ga = 0.175 mis? for Singapore ©CGKoh 2015 67 Dama + To withstand a seismic action having a larger probability of occurrence than the design seismic action, without the occurrence of damage and the associated limitations of use, the costs of which would be disproportionately high in comparison with the costs of the structure itself. * The seismic action for DLR has a probability of exceedance, Pp... in 10 years and a return period Toi. + Recommended Pp,,= 10% in 10 years corresponds to Typ = 95 years — this is adopted in Singapore NA + Ashorter return period (95 years) means less severe earthquake than the design seismic action with return period of 475 years. ©CGKon 2015 68wm So 8 Compliance Criteria for Seismic Design * To satisfy the fundamental requirements, two limit states shall be checked: (1) Ultimate limit state (ULS) — associated with collapse or other forms of structural failure which might endanger safety of people. (2) Damage limitation state (DLS) — associated with damage beyond which specified service requirements are no longer met * To limit uncertainties and promote good behaviour of structures under seismic actions more severe than the design seismic action, some specific measures are taken as per 2.2.4 (e.g. simple and regular forms, ductile behaviour, good detailing, appropriate analysis model, etc). * For low seismicity, simpler rules can be applied for certain structures (low ductility class), e.g. detailing follows EC2, EC3 and EC4. = For very low seismicity, EC8 needs not be observed + Singapore NA: Lowseismicity _if aS < 0.98 m/s? Very low seismicity if a,S < 0.39 m/s? where a, = design ground acceleration on type A ground, and S = soil factor. ©cGKoh 2015, 6 Comparison with Design Earthquakes [ For | Defined in UBC-97 & IBC-2012 = UBC 1997: Maximum Capable Earthquake (MCE) — Probability of exceedance =10% in 100 yrs (950-year return period) Design Basis Earthquake (DBE) — Probability of exceedance = 10% in 50 years (475-year return period) + IBC 2012: Maximum Considered Earthquake (MCE) ~ Most severe earthquake considered by the code (a) Probabilistic MCE — Probability of exceedance = 2% in 50 years (2475- year return period) (b) Deterministic MCE - based on characteristic earthquakes in the region (ASCE 7-05 Ch 21) Design Earthquake (DE) - 2/3 of MCEEc8 Classification of Ground Types + The influence of local soil conditions on seismic response of structures are accounted for by defining ground type using soil parameter P in the upper 30 m 2 jag P=d, f 2/2 where d, = thickness of layer) and P, = soil parameter of layer i + The soil parameter P could be (a) v,: Shear wave velocity (v, = \G/ where G = shear modulus and » = density ) (b) Nspr: Blow counts for 30 cm in standard penetration test, or (c) c,: Undrained shear strength (Shear wave velocity is preferred.) + EC8 Table 3.1 shows 7 ground types: A to E and two special types S, and S, + In Singapore NA, only 5 ground types are relevant and given. ©CG Kon 2015 n Ground Types for Singapore [Annex Naza | Ground Description of stratigraphic profile ha ee , (kPa) 'A | Rock or other rock Ike geological formation, Including at most 5 m of weaker material atthe surface. Deposits of very dense sand, gravel or very stif clay, at B least several tens of metres in thickness, characterised by | 360-800 >50 >250 4 gradual increase of mechanical properties with depth, Deep deposits of dense or medium-dense sand gravel or | stiff clay with thickness from several tens to many 10-360 | 15.50 70-260 hundreds of metres. Deposits of loose-to-medium cohesioniess soll with or D without some soft cohesive layers), or of predominantly <180 <15 <70 soft-to-firm cohesive sol ‘Deposits of very loose cohesioniess soll (with or without 8, _|some soft cohesive layers), or of predominantly very soft | <100 <6 10-20 soll [See footnote’) * If more than one of the three soil parameters are available, the most onerous Ground Type shall be adopted * Ifthere are different ground types spatially distributed within the footprint of a building, the most onerous Ground Type shall be adopted. * The rules shall apply regardless of whether the building is founded on piles that extend to hard soil stratum or not. * The above description of type S, in Singapore Annex is slightly different from Table 3.1 of EC8 (which is also used in BC3, involving plasticity index). The description here is simpler to use. ©c6Kon2018 n >800 NAL NA[ecs 32) PGA for Design Seismic Action | aaa «Earthquake events are random in nature, with many uncertainties associated with intensity, location, duration, attenuation, etc. + Probabilistic seismic hazard analysis is usually adopted to determine peak ground acceleration (PGA) for a given return period at a given site. + Reference PGA on type A ground, denoted as agp , defines the local hazard (obtained based on seismic hazard assessment) + Singapore NA: The whole of Singapore is under one seismic zone and [_aqn = 0.175 mis? = Design ground acceleration on type A ground: ag = aga X Importance Factor (yi) where the importance factor depends on the consequences of failure. ©CEKon 2015 sc | Ece Importance Class of Bui | 42.5 | ings = Different importance factors are given for different classes, to reflect the different levels of “reliability” assigned + In EC8, buildings are classified into four importance classes (| to IV), depending on the consequences of collapse for human life, etc: = Singapore NA: Only two classes are defined as follows. Importance Class Factor y, | |” Hospitals, fire stations, civil defence Special | _ installations, ministry offices and 14 institutional buildings Buildings other than those classified as “Special” buildings Ordinary 1.0 ©CGKon 2015 4Response Spectrum (Horizontal) 0 T; To sT< ot ~ 0 Tr ot sTs T>4s O 5.5 : EC8 Recommended Parameters to Define S, Ground type s Ts (s) T.(s) Tp (s) A 1.0 0.15 0.4 2.0 B 1.2 0.15 05 20 Cc 7.15 0.20 06 2.0 D 1.35 0.20 08 2.0 E 1.4 0.15 OS 2.0 Type 2 — for low magnitude earthquakes of M, <5 Ground type Ss Ta (s) T, (s) Tp (s) A 1.0 0.05 0.25 12 B 1.35 0.05) 0.25 12 Cc 15 0.10 0.25 12 D 18 0.10 0.30 4.2 E 16 0.05 0.25 12 M,= Surface wave magnitude which is a scale of earthquake based on Rayleigh surface waves travelling in top layers. ©CGKoh 2015 is damping correction factor 76£c83222 | eee | For Info: Elastic Response Spectra Based on EC8 Recommendation i & 500, g | 2 400 } 3.00 Penod Ts Type 2 Normalized acoaleration S.'a,, 8 Singapore Annex: Parameters to Define Ss. Due to the far field effect of Sumatra earthquake, the EC8-recommended parameters are not suitable for Singapore BCA conducted an earthquake vulnerability study, leading to the following Nationally Determined Parameters (NDP) for Singapore Ground type s Ta (8) Ts) | Tos) c 1.6 0.4 44 10.4 D 25 09 | 1.6 46 Sy 3.2 1.6 24 24 Furthermore, seismic design is only required for (a) new buildings and (b) existing buildings undergoing major A&A works, that are above 20 m in height", and founded on certain ground types: + Ground types D or S, for “ordinary” buildings (y, = 1-0) + Ground types C, D or S, for “special” buildings (y; = 1.4) * Building height is taken from the foundation or top of a rigid basement to the topmost habitable structural floor level. See illustrations in Figure 2 of BC3 © GKon 2015Elastic Response Spectrum for Singapore [sa [3 * Damping ratio of 5% is used in the elastic response spectrum. * The elastic response spectrum assumes that the buildings behave linearly elastic. This is usually not adopted in seismic design for cost reasons. * Vertical response spectrum is not needed in Singapore * Illustration below is for a, = agg = 0.175 mis? (i.e, ordinary building with y, = 1.0). 15 COCO eo 8 L | For horizontal seismic action only To 7 i 6 10 3 5 et As oe o 5 3 2 aoedae. 1 0 0 0 1 2 3 4 5 6 7 8 ©c.Gkon 2015 Period T (s) 7° i i { eee Design Response Spectrum (Horizontal) Boas «Earthquake design philosophy permits structures to behave in inelastic Manner, so as to reduce the design lateral force (and hence cost). * This design approach would require ductility so that the structure would deform in the inelastic range to the expected extent without fracture. + To take advantage of ductility, the elastic spectrum (S,) is converted to the design spectrum (S,) by a behavior factor q. Design spectrum: Example: Ordinary Buildings on Ground Type $1 2.7 O Low-dissipative V/ DOM = Ductiity Class “Medium” (q up to 4.0) } camtecuresd DCH = Duetility Clegg “High”, (q> 4.0) DCL + Non-seismic detailing can be adopted in low-seismic regions as in Singapore as per EC2, EC3 and EC4 for concrete, stee! and composite steel-concrete buildings, respectively = Foundation design follows EC7 with no seismic requirement DCM and DCH * Seismic detailing follows Sections 5, 6 and 7 of EC8 for concrete, steel and composite steel-concrete buildings, respectively * Foundation design follows EN1998 Part 5 in addition to EC7 EC GKoh 2015Column Detailing Requirement (DCL vs DCM) Example: | Stitical confinement region: hoop spacing min (642: 175 Bo) rnin, dimension of concrete core (tothe ‘Senterine of the Hoops) = min longtusnal bar oa } | Beam-cotuma Joint Limit sizeof main rebarto ensure suficent bond 44 |2. Atleast iiemedate TUT] |? scosrbetteen cones ot +H} columns CITT Jo. oop mst conte Urreuced trough it Example’ [DCL] ©cGxon 2018 ae Legend: Critical confinement region —no splice allow'd (Detailing to EC8) Reduced confinement region — splice allow'd (Detailing to EC2) Special confinement for beam-column joint (Detailing to ECB) confined unconfined >. oy Ecs | Shear Wall Detailing Requirement (DCL vs DCM) E 2A | 5. eased) Legen Critical confinement region (Detailing to EC8) Reduced confinement L_! region (Detailing to EC2)Low-Dissipative vs. Dissipative Structures feces Low-Dissipative Structures * Do not account for any hysteretic energy dissipation (though some inherent ductility exist even based on non-seismic detailing) + Thus, a low behaviour factor is used to account for overstrength + Para 3.3 of BC3 specifies 1.5 as the minimum q value for all building types. ssipative Structures + Designed to make use of hysteretic energy dissipation in the inelastic range so as to reduce the design seismic forces = There is a price to pay for the reduced design forces. For example, detailing and connections are more elaborate than non-seismic requirements * Also, the drift and displacements would be larger (for damage limitation check). = Depending on the structural system and its dynamic characteristics, seismic action may or may not be the most controlling action in low seismicity regions. + If seismic action controls, it would be reasonable to consider the option of designing a dissipative structure (which should result in smaller vertical members but require seismic detailing) G Koh 2015. Design Response Spectrum (Horizontal) for Singapore (low ductility) | Example for q Ordinary Buildings (y, = 1.0) Special Buildings (y, = 1.4) 15 15 I | I f= f-- S,— Ground | 0 2 Ta 6 : 8 ©cGkon 2015Design Response Spectrum by BC3 eH) * BCA’s Guidebook (BC3: 2015) provides a slightly different alternative. + The elastic spectrum is divided by q and multiplied by y, to obtain S, as follows: 3,(7)= 2) qs n +S, is defined based on a, = agg = 0.175 mis? in BC3. + y; is not involved in defining a, but used in the above equation + The end result for S, is slightly different than the EC8 formula in the low period range 0 < T< T, (but no difference for q = 1.5). * Also, the lower bound factor (8) is not applied in the very long period range (very tall buildings, for which wind action is likely to govern rather than seismic action). ©CGKoh 2015 a7 Ecs Guiding Principles for Conceptual Design is Structural simplicity — less uncertainty in prediction of seismic behaviour Uniformity, symmetry and redundancy — better distribution of action effects and widespread energy dissipation; reduce stress concentration, large ductility demand and eccentricity effect. irectional resistance and stiffness — ensure similar resistance and stiffness characteristics in both main directions as earthquake can be in any direction. Torsional resistance and stiffness — torsional motions stress different structural elements in a non-uniform way. Diaphragmatic behaviour at storey level — floor diaphragm action ensures that vertical structural systems act together in resisting the horizontal seismic action; see 4.3.1(4), Adequate foundation — provide sufficient rigidity and ties to ensure that the whole building is subjected to a uniform seismic excitation ©CGKoh 2015 28. Ec8 Structural Regularity y J\ 423 Structural regularity will affect the structural model, analysis method and q. Regularity Allowed simplification Behavior factor Plan |Elevation| Model Linear-elastic analysis (for linear analysis) Yes | Yes | Planar (2D) Tateral force Reference Yes No | Planar (2D) | Modal Response Spectrum | Decreased (by 20%) No Yes _| Spatial (3) Lateral force Reference No No _| Spatial (3D) | Modal Response Spectrum | Decreased (by 20%) Regularity in plan: The criteria are given in Sect. 4.2.3.2 of EC8. When the building is irregular in plan, a spatial (3D) model should be used Regularity in plan may enhance the q value via an overstrength factor - see 5.2.2.2.(5) of EC8 for concrete buildings, 6.3.2(3) for steel buildings, and 7.3.2(3) for composite steel-concrete buildings Regularity in elevation: The criteria are given in Section 4.2.3.3 of EC8. - When building is irregular in elevation, the modal response spectrum analysis is used (not lateral force method), and the value of q is decreased by 20% After any enhancement or reduction, a minimum value of q = 1.5 can be used for all building types (per BC3 3.3) ©CGKon 2015 T-shaped plan L-shaped plan U-shaped plan Cruciform plan Complex shape Setback Interruption of columns Interruption of beams 4 ‘Multiple towers Split levels Ref: "Recommended Lateral Force Requirements & Commentary’, Seismology Committee, Structural Unusually high storey Unusually low storey Engineers Association of California (1999) 90Criteria for Regularity in Plan (1) * Approximately symmetrical in plan w.rt. two orthogonal axes, in lateral stiffness and mass. * Compact plan configuration, i.e. each floor is, delimited by a polygonal convex line. Avoid plan set-backs (re-entrant comers or edge recess) Total set-back area ‘5% of floor area ne vibrate in phase. * Rigid diaphragm — sufficiently large in-plan stiffness of floors (compared to lateral stiffness of vertical elements) so that the deformation of floor has a small effect on the distribution of forces among vertical elements. + Slenderness ratio of plan Lmmax! Lmin $4 ©cG Kon 2015 Criteria for Regularity in Plan (2) ox = For the y direction of analysis, Lain Cox 0.30%, } Egy $0.30 5 Ky 2Isi ty 2ls €,, = distance between centre of stiffness and centre of mass, measured along the x direction F, = square root of [ torsional stiffness / lateral stiffness in y direction } |, = radius of gyration of floor mass in plan = square root of [ polar moment of inertia of floor mass in plan w.rt centre of mass / floor mass] * Similar definitions of e,, and r, for the x direction of analysis. * Para 3.5 of BC3 “Example Calculation” provides guidance on how to compute the above terms. ©CGkon 2015 + = Centre of mass + = Centre of stiffness The two wings may not 8 423.2 Ecs | 4.2.3.2 = Ateach level and for each direction of analysis (x and y), eccentricity e, and torsional radius r satisfy 2Criteria for Regularity in Elevation hoa + No interruption for all lateral load resisting systems (e.g. cores, walls, frames) from their foundations to the top of building or the top of relevant zone of building » Cb (etch scouts sone 05H * Lateral stiffness and mass of ‘ individual storeys remain constant or reduce gradually without abrupt y - \ changes from base to top. | vt = In framed buildings, the ratio of the actual storey resistance to the omen fo resistance required by the analysis «eos suse should not vary disproportionately between adjacent storeys. = When elevation setbacks are = present, check regularity as shown Cotenon for 93 loo Dof v _ Criteria for Rigid Diaphragm ad 5.10 * Arigid diaphragm is one, when modelled with its actual in-plane flexibility, its horizontal displacements due to seismic action nowhere exceed those resulting from the rigid diaphragm assumption by more than 10% of the corresponding absolute horizontal displacements in the seismic design situation * L,C, Hand X plan shapes should be carefully examined ~ the stiffness of lateral branches (wings) should be comparable to that of the central part * Asolid RC slab may be considered as a rigid diaphragm, if it has a thickness of not less than 70 mm and is reinforced in both horizontal directions with at least the minimum reinforcement specified in EC2 * The above point is also applicable to cast-in-place topping on precast floor if connected through clean and rough interface or shear connectors. * The diaphragm should be free of large openings, especially in the vicinity of main vertical structural elements. If unsure, use the above 10% check ©CEkoh 2015 9— 7 Ec Methods of Seismic Analysis | 4. A. Linear static analysis > Lateral force method (LFM) Ku =F B. Linear dynamic analysis > Response spectrum analysis (RSA) Mi(t)+ Cu(t) + Ku (t) = F(t) C. Non-linear static analysis > Pushover analysis Ku + Fy, =F D. Non-linear dynamic analysis > Time history analysis Mi(t)+ Cu(t) + Ku (t)+ Fy, (t) = F(t) 9 Structural Modelling The model of the building used in structural analysis shall follow 4.3.1 In particular, to account for cracked elements, the elastic flexural and shear stiffness of concrete elements may be taken as one-half of the corresponding stiffness of the uncracked elements. The storey weights (associated with the inertia effect) are calculated from the gravity loads in the following combination of actions (per 3.2.4 of EC8) XG," Live, Qs where ve, =9¥2, rt \ Permanent action (dead load) Variable action (live load) pz, is to account for quasi-permanent nature of variable actions (not all are present at the same time. See Table NA.a.1 of NAto SS EN1990) @ is to account for the fact that some variable live load (e.g, furniture) may not be rigidly connected to the structure and may not vibrate in phase with the structure during a seismic event. See Table 4.2 of ECB Table 2 of BC3 reproduces the values of p,, and @ taken from Singapore NA. ©CGKon 2018 96Inherent and Accidental Torsion BS + Inherent torsion: When horizontal seismic forces are applied, eccentricity (between centre of mass and centre of stiffness) will result in torsional response. Accidental torsion: To account for uncertainties in the location of masses, accidental torsion has to be accounted for by considering 5% eccentricity (per EC84.2.2) 9, =+0.051, where e,,= accidental eccentricity of storey mass / from its nominal location, applied in the same direction at all floors floor dimension perpendicular to the direction of the seismic action + 3 independent cases are considered in each direction * For example, in the X direction (as shown in figure) 1. Lateral forces in the X-direction 2. Lateral forces in the X-direction with +5% ECC 3. Lateral forces in the X-direction with -5% ECC = Similarly there are 3 independent cases in the Y- direction. + = Centre of mass. ©CKoh 2015 97 LrM| (A) Lateral Force Method (LFM) * Limitation of the method: Only applicable to buildings with (a) elevation regularity and (b) fundamental periods T, < 47, and T, <2 s in the two main directions (for Singapore, just check T, < 2s) + The response is assumed to be governed by the fundamental mode + Base shear in the horizontal directio Sa S,(T;) = design spectrum value at T, 7 T ia = Fundamental period of vibration in the horizontal direction considered, obtained by eigenvalue analysis or empirical formula m (or W)= total mass (or weight) of building above foundation or a rigid basement. A = correction factor = 0.85 if T;s 27, and more than 2 storeys; otherwise 4.0 (i.e. no correction). Note: Am represents the “effective modal mass” of the fundamental mode. ©CEKoh 2015Approximate Fundamental Period (T, in s) * For buildings with heights of up to 40 m: T,=C,H°* where H = height in m Structure type Cc, Moment resistant space steel frames 0.085, Moment resistant space concrete frames 0.075 Eccentrically braced steel frames 0.075 All other structures 0.050 * Alternative formulae are given in 4.3.3.2.2 of EC8 = Computer software would give a more accurate value of natural period, provided that the model adequately represents the structural mass and stiffness. ©cGkon2015 99 [LF OM) Summary: Base Shear by icons Lateral Force Method - Seismic Soil_ Earthquake Recor investigation type assessment Parameters Ground type | — Seismic (S. To, Tor To) hazard map Design spectrum Ss, = 5 Design Reference Fi, = S4(T,)-m- A) , PGA (ay) PGA (age) T Ly Importance Behaviour oa factor (q) Build Fund. period (T;), aod mass (m), correction factor (A) ©cekon 2015 100Distribu' n of Seismic Forces Horizontal seismic force F; acting on storey / S,-m, ds,-m, where s|= displacement at storey / of mass m,in the fundamental mode shape. F =F, The fundamental mode shape is assumed to be linear along the height of the building. Thus, where z, = height of mass im, above the base (i.e. foundation or top of rigid basement), The seismic forces are then applied at each storey. F, Do this for the main directions X and Y independently. D ©CGKoh 2015 101 (B) Response Spectrum Analysis (RSA) | asa + 3D structural model is usually used + All modes of vibration contributing significantly shall be taken into account. + Number of modes depends on taser es (a) Sum of effective modal masse62 90% oftotal mass, or (b) _ Include all modes with effective modal mass > 8% of total mass. * Alternatively, «23. /numberof storeys and T, <02s Yime. my "Effective modal mass for n-th mode is M. vA where ®, efficient at the i-th floor of the n-th mode shape (obtained by eigenvalue analysis of the structural dynamic model). * Effective modal mass reflects the contribution of a particular mode of vibration to the base shear. * Sum of effective modal masses for all modes = Total mass of building. ©cG Kon 2015 102For mode k: * Based on the natural period (T,= 217/ w,), obtain the value (S,), from the design response spectrum: Response of Individual Mode * Base shear due to mode kis + Do this for all modes that are needed to satisfy the mass participation requirement. ©ceKoh 2015 RSA| Distribution of Seismic Forces in Each Mode In each mode, distribution of seismic force is determined by the mass distribution and corresponding mode shape (which could reverse in sign). Example: 1000 kg Mo Fay Fay Fag up Fay Fp Fs F F, uy a 12 Fes _— — Fos Fr Fig Mode 1 Mode 2 Mode 3 ©CGKoh 2015 ‘©CGKoh 2015Combination of Modal Responses Note that the response obtained for each mode is the peak quantity without + sign (Le. absolute max.) To combine contributions from all J modes considered, it would be too conservative to simply add the modal peak values (Absolute Sum Rule). Two common combination rules: > SRSS (Square Root of Sum of Squares, or vector sum) 412 [=] Correlation coefficients between modes / & j — important when two frequencies are close to (within 10% of) each other The above equation is for base shear as an example. The same combination rules are similarly applied to other seismic action effects (bending moment, torsion, displacement, etc) When in doubt, use CQC rule since SRSS is a special case of CQC. ©cG Kon 2015, 105 RSA Modal Combining Peaks — Caution! Combination Be careful in computing the peak value of one response quantity from the peak values of other response quantities. E.g. to compute storey drift between level 2 and level 3 from the horizontal displacements of levels 2 and 3: Agh®®* = y,Pe* Peak x + The above computation is wrong since the u, and u; peak values do not necessarily happen at the same time + Furthermore, the peak u(t), (absolute max.) values for u, and u, have lost their + sians due to the combination rule. G Kon 2018 108[RSA] Combining Peaks — The Correct Way oe = Peak computation has to be done for each mode: eee errr YA = (us peak — (uy peal k-th mode ypeak (a3 fe ypeal where level 2: (uhh me8 ea’ — max | g(t) mer| x (gRMMA® eee 4 ‘Same peak at same time captured in level 3: cnmate pak — may [g(t exe (ghihmode,“— mode shape r E.g. for mode 2: 9 up nate gy bitrate, gctmese (ty 2ndmode max absolute value = peak (g2nd mode y,) = Then apply combination rule to combine the modal peak values of A, by e.g. SRSS: 42 gPett = {tcagl® me poke agzmmowe peak? | © Gkon 2015, 107 RSA Z a Ec8 Accidental Torsion 43333 * [tis difficult to include accidental torsion (due to +5% eccentricity) in RSA in a rigorous way. + An approximate way is as follows (dynamic + static) 1. Compute RSA result in the X direction without eccentricity (dynamic) 2. Compute accidental torsion = (Seismic force in X direction) - ey for each above- ground storey. The seismic forces are obtained based on the LFM (static) 3. Apply accidental torsions (one torque at each level*) and compute the static result. Combine the static result** with the dynamic result 5. Repeat the above steps to obtain the results in the Y direction. Rigid diaphragm is normally assumed in the above procedure. For non-rigid diaphragm, the applied torque at the storey level is distributed to all joints in proportion to the joint mass. ** There are two independent sets of static results: one for positive eccentricity and the other for negative eccentricity. Each set is combined with the RSA result independently, and the envelope is taken for the two combined sets of results, © CG Kon 2015 108badd Summary: Response Spectrum Analysis Modal analysis I Natural periods Mode shapes Effective modal masses Other modal properties Decide # of modes to include (cumulative 90% contribution, or all For each mode Obtain spectral acceleration S, from design spectrum Ts. modes with 5% contribution) ‘Compute modal base shear 35 Fin = STM, 38 = STi My |E [ Combine the seismic eftects ooo \$ a due to seismic actions in the main direction (X or Y) \6 8) two main directions 1 T Compute the seismic effects For all modes cor ered, due to seismic action inthe /——__use sRSS orGac ys’ direction considered compute total base shear F, || ec oKonais Combination of Horizontal Components anges | of Seismic Action + Horizontal seismic action is described by two orthogonal components which are assumed as independent and represented by the same response spectrum. * Structural response to each horizontal component is evaluated separately. * Directional Combination: To combine the effects of the two horizontal components, use (a) SRSS rule — Square Root of Sum of Squared values of action effect due to each horizontal component: 2 go uz a Eq Cea + Few (b) Percentage combination rule: 100% for one direction "+" 30% for the other direction (4 cases) Feg Esa "*030E G4, and Egy =0.30E gg, "+" Bea, YY 0384, 0.3E 4, O.3E a, Case 2 >| Case 3 Case4 |< 110ECs Displacement & Drift Calculation = To reduce the design lateral force, the structure is allowed to deform in the inelastic range and take advantage of ductility. = The price to pay for reduction in the design force is the large displacement + The (inelastic) displacement induced by the design seismic action (d,) is given by dy = 494 s = Ia Fe where + d,= displacement computed by a linear analysis based on the design response spectrum. + qq= displacement behaviour factor, assumed equal to q unless otherwise specified. Force low stiffness > = Torsional effects (including accidental torsion) of large displacement seismic action have to be included. ©CGKon 2015 1m Safety Verification * Two limit states per EC8 4.4.2 and 4.4.3 and specific measures per 2.2.4 are considered. * For ordinary buildings, the verifications prescribed in 4.4.2 and 4.4.3 may be considered satisfied if both of the following conditions are met: a. Total base shear due to seismic design situation with behaviour factor for low-dissipative structures < Total base shear due to other action combinations (based on linear elastic analysis), and b. Specific measures per 2.2.4 are considered, except 2.2.4.1(2)-(3) * Specific measures cover the following: a. Design - eg. simple and regular forms both in plan and elevation, and adequate structural model b. Foundations — transmit seismic actions from the superstructure to the ground as uniformly as possible cc. Quality system plan ©CG kon 2015 2N Design Check: ULS and DLS Ecs 44.2 - Ultimate Limit State (ULS) 443 Resistance condition: All structural elements including connections and the relevant non-structural elements satisfy E,< R, where E, = design value of action effect duet to the seismic situation (ECO 6.4.3.4) including second-order effects (if necessary), R, = design resistance of element per relevant Eurocode documents for member design per EC2, EC3 and EC4 for DCL, plus EC8 Sections 5 to 7 (for DCM and DCH) Other conditions per EC8 4.4.2.3 to 4.4.2.6. In particular, 4.4.2.6 is imposed to avoid earthquake-induced pounding from adjacent buildings (structural separation per BC3 Para 8) . Damage Limitation State (DLS) The damage limitation (serviceability) requirement is satisfied for seismic action with a larger probability of occurrence (return period of 95 years) than the design seismic action, by imposing limitation of inter-storey drift per 4.4.3.2. ©cGKon 2015 113 © © K Combination of Actions for Member Design In the member design stage for Ultimate Limit State (ULS), the combination of actions for seismic design situations is given in EN1990 (ECO) XS i "e" PONE" Aggy "+" Lv 2 Qi iz nominal value of permanent action j (dead load), prestressing action (if any), design seismic action (including accidental torsion effects) nominal value of variable action i (live load), Ye , = quasi-permanent coefficient given in Normative Annex A1 of EN1990 Load factor of 1 for Ag, reflects that earthquake load (with return period of 475 years) is specified at ultimate limit level Geometric imperfection effects (GIE) are to be accounted for in the design state according to the respective codes (EC2, EC3 and EC4) BC3 Para 8 gives a slightly different presentation Combination of Actions at ULS = 1.0 x Storey Weight + 1.0 x Seismic Action 7 + 1,0 x Geometric Imperfection Effects n 2015 14Second Order P-A Effects * Second order P-A effects need not be included if the following condition is fulfilled in all storeys: Prot Or Vior A < 0.10 where @ = inter-storey drift sensitivity coefficient d, = design inter-storey drift, evaluated as the difference of the average lateral displacements d, at top and bottom of the storey (in accordance with EC8 4.3.4, i. inelastic drift is used here). P= total gravity load at and above the storey in the seismic situation (total = sum for all vertical elements) Vi = total* seismic storey shear fh = inter-storey height * If > 0.2, second-order effects have to be included. For 0.1<@ $0.2 second-order effects may approximately be included by multiplying the seismic action effects by 1/(1-6). * @should not exceed 0.3 ©CGKoh 2015 5 Resistance of Foundation System For dissipative structures, the action effects for foundation elements are evaluated according to capacity design considerations accounting for development of possible overstrength The action effects need not exceed those corresponding to the response under the seismic design situation assuming elastic behaviour (q = 1) Capacity design method is to obtain the hierarchy of resistance of various components and failure modes necessary for ensuring a suitable plastic mechanism and for avoiding brittle failure modes. ‘An alternative to avoid capacity design for foundations supporting individual vertical elements is given in 4.4.2.6 (4)-(8) For low-dissipative structures, the design actions effects are simply obtained from the elastic analysis (i.e. no capacity design is needed) ©CGkon 2015 16which may also be written as A Drift Limitation Check by EC8 The damage limitation is verified by limiting the design inter-storey drift as follows. d,v < 0.005 h oY < 9,005 h =. storey height design inter-storey drift, evaluated as the difference of average lateral displacements d, at top and bottom of storey (d, is computed from d, factored by q, which may be taken as q in accordance with EC8 4.3.4). = reduction factor to account for a smaller seismic action for DLR (return period of 95 years) than NCR (475 years) Singapore NA: v= 0.5 for ordinary buildings cs v= 0.4 for special buildings. The above limiting value of 0.005 (or 1/200) is for buildings having non-structural elements of brittle materials attached to the structure. This value is increased to 0.0075 (or 1/133) for buildings having non-structural elements of ductile materials attached to the structure. Koh 2015 a tent . | BC3 | Drift Limitation Check by BC3 | Para7 | + The EC8 formula is modified in BC3 by substituting d, = qd, to give d,< 0 005 h vay where d,= design inter-storey drift evaluated as the difference of the average lateral displacements d, (not factored by q) at the top and bottom of the storey under consideration. This may be re-written as aca < 0.005 Caution: @CEKon 2015 Note the difference between EC8 inequality (using inelastic displacement) and BC3 inequality (using elastic displacement) in defining d, If inelastic drift is used as d,, then q has already been factored and thus does not appear in the inequality check (EC8) If elastic drift is used as d,, then q appears in the above inequality check (BC3). 18