Short-Period Building Performance Paradox Resolution of the Short-Period Building Seismic Performance Paradox and Recommendations for Improving Seismic Codes and Design Collection of FEMA P-695 Collapse Results (NIST GCR 12-917-20) Why do analytical models of code-compliant designs predict high probabilities of collapse for short-period buildings contrary to damage observed in actual earthquakes and the judgment of earthquake engineers? Charlie Kircher (Kircher & Associates) Kelly Cobeen (Wiss, Janney, Elstner Associates) 2 Why does this matter? ATC-116 Project Seven-year project (2013 – 2020) Funded by FEMA Conducted by the Applied Technology Council (ATC) Four FEMA P-2139 reports Short-period buildings (T ≤ 0.5 seconds) constitute the vast majority of buildings Do short-period buildings meet the collapse safety objective of seismic codes? Are improvements in seismic building codes, design, and assessment methodologies warranted? Overarching Findings, Conclusions, and Recommendations (Vol. I) Light-Frame Wood Buildings (Vol. II) Reinforced Masonry Buildings (Vol. III) Steel Special Concentrically Braced Frame Buildings (Vol. IV) Reports available at: https://femap2139.atcouncil.org/ Credit: KlausVonVilver, WikimediaCommons 3 4 1 Project Objectives Identify causes of the short-period building paradox and develop solution concepts Improve and validate numerical modeling methods for short-period buildings Develop recommendations for improving seismic codes and engineering practice and for future research initiatives 5 Observed Response and Collapse Performance Post-Earthquake Surveys 1971 San Fernando 1978 Miyagi ken-oki (JP) 1987 Whittier 1994 Northridge 1995 Kobe (JP) 2011 Christchurch (NZ) Northridge Meadows Apartment Collapse Reinforced-Masonry Test Structure – UC San Diego Full-Scale Shake Table Testing E-Defense (Miki, Japan) University of Buffalo and UC San Diego 7 6 Selection of Systems Analysis of low-rise occupancies from HAZUS inventory of about 2 million buildings in Northern California Wood light-frame: W1, W1A, W2 (77%) Reinforced masonry: RM1, RM2 (7%) Steel braced frame: S2 (1.3%) More than 85% of square footage of low-rise occupancies S3 S4 C1 C2 PC S1 S2 RM W2 W1A W1 8 2 Approach of Analytical Studies (each SFRS) Identify Representative Archetypes Develop Designs of Archetypes Develop Numerical Models of Archetype Designs Analyze Numerical Models (P-695) Archetypes Light-Frame Wood 2.33 g Archetype (1.2 x 1.94 g) SMT = 1.5 g System 3.1 g Wood light-frame walls with wood structural panel sheathing (wood) Single-Family Dwelling (SFD) ACMR = 3.1 g/1.5 g = 2.07 Reinforced Masonry Archetype Pushover Analysis (Vmax/W) Incremental Dynamic Analysis (IDA) Evaluate Analysis Results (P-695) Dynamic Response Collapse Statistics Evaluate Collapse Performance (P-695) ACMR = SSF x SCT / SMT P[Collapse|SMT] = Lognormal (median, bTOT) Steel SCBF ST [T=0.25s] (g)s Archetype Figure Figure4-16 4-17of ofFEMA FEMAP-2139-2 P-2139-2(Wood (WoodReport) Report) showing median collapseand showing IDA results and (collapse fractions) acceleration (SCT =curve, 1.94 g) and collapse collapse fragility median, SCT = 3.1 g, and displacement capacity (drift ratio 8.17%) lognormal standard deviation of b=TOT = 0.5. No. of Stories Response Modification Coefficient (R ) Design Seismicity 1, 2 6.5 High, Very High Multi-Family Dwelling (MFD), Commercial (COM) 1, 2, 4 6.5 High, Very High Special reinforced masonry shear walls (masonry) Commercial (COM) 1, 2, 4 5 High, Very High Steel special concentrically braced frame (SCBF) (steel) Commercial (COM) 1, 2, 4 6 High, Very High Where, High: SMS = 1.5g; Very High: SMS = 2.25g 9 Archetype Designs Engineered designs prepared using ASCE 7-10, assuming Risk Category II Designs represent typical modern practice in areas of significant seismicity Configurations selected to be realistic and representative of actual buildings in size and proportion Occupancies 10 Wood Archetypes Two-story Masonry Foundation Plan Single-Family Dwelling (SFD) 32 ft x 48 ft plan One and two stories OSB shear walls and diaphragms Two-story Masonry Floor Framing Plan 11 12 3 Wood Archetypes Wood Archetypes Multi-Family Dwelling (MFD) Commercial (COM) 48 ft x 96 ft plan One, two and four stories Two-story: four-unit row houses/townhouses One- and four-story: apartments with central corridor OSB shear walls and diaphragms 48 ft x 96 ft plan One, two and four stories One-story: retail or repair shop Two- and four-story: office OSB shear walls and diaphragms 13 Masonry Archetypes 14 Steel Archetypes Commercial (COM) Commercial (COM) 48 ft x 96 ft plan One, two and four stories Office building Rigid floor diaphragms, flexible roof diaphragms Load bearing, fully grouted, reinforced hollow-unit concrete masonry with cantilever shear walls 90 ft x 150 ft plan One, two and four stories Office building Rigid floor diaphragms, flexible roof diaphragms Steel SCBF with square HSS braces 15 16 4 3D Nonlinear Numerical Models Parametric Studies Baseline Configuration Steel SCBF Model Best-estimate of each modeling parameter, used to compare against benchmark collapse rates Variations in Backbone Curve Shape for Wood Models Variant Configurations & Modeling Parameters Light-Frame Wood Model Collapse displacement capacity (wood, masonry) Soil structure interaction and foundation flexibility (wood, masonry, steel) Nonstructural wall finishes (wood) Backbone curve shape (wood) No redundancy (steel) No reserve moment frame (steel) Reinforced Masonry Model 17 18 U.S. Earthquake Losses – 1964 – 20141,2 Earthquake Observed Response and Performance Benchmarks Name Fatalities Building Loss (billions) Date Magnitude Total Building EQ Date 1964 Anchorage 3/28/1964 M9.2 128 10 $0.3 Est. 2014 $5.0 1971 San Fernando 2/9/1971 M6.6 65 53 $1.5 $20.0 1989 Loma Prieta 10/17/1989 M6.9 63 12 $6.5 $17.5 1994 Northridge 1/17/1994 M6.7 57 20 $25.0 $55.0 Other Earthquakes (23) 33 14 $3.7 $7.4 Total All Earthquakes 346 109 $37.0 $104.9 1. Date and magnitude of selected large-magnitude (> M5.5) U.S. earthquakes, 1964 - 2014, and associated fatalities and building-related economic losses. 2. Approximately 2 building-related fatalities per year (109 fatalities/50 years). 19 3. Estimated as many as 3,400 (day time) fatalities for a repeat of the 1906 San Francisco earthquake (Kircher et al., “When the Big One Strikes Again,” Earthquake Spectra, 2006) 20 5 Summary and Comparison of Earthquake Losses Sources of Earthquake Data Used by this Project Sources of Damage Data on SFRSs Considered by the Project Earthquake Date Name Magnitude Wood Light- Reinforced Frame -Masonry 1971 San Fernando 02/09/1971 M6.6 1978 Miyagi-ken-okI (JP) 06/12/1978 M7.4 Steel SCBF 6 7 3 1987 Whittier 10/01/1987 M5.9 1994 Northridge 01/17/1994 M6.7 1 1995 Kobe (JP) 01/17/1995 M6.8 2 2011 Christchurch (NZ) 02/22/2011 M6.2 4 8 9 5 1. OES, 1995; HUD, 1994, ATC, 2000; Schierle, 2003; 2. AIJ, 1995a; Yamaguchi and Yamazaki, 2000; Yamazaki and Murao, 2000; 3. Hart et al., 1988; 4. OES, 1995; TMS, 1994; 5. Dizhur, 2011. 6. Simpson, 2017; NOAA, 1973; 7. Simpson, 2017; Tanaka et al., 1980; 8. OES, 1995; Krawinkler, 1995, Tremblay et al., 1995; 9. AIJ, 1995a, AIJ, 1995b; Tremblay, 1996. 1994 Northridge Earthquake 1995 Kobe Earthquake • 60 fatalities (20 due to building collapse – 4 wood buildings) • 1,044 hospitalized injuries • 11,088 displaced households • 14,500 Yellow/Red tagged buildings: • Less than 1% of buildings with Red Tags in strongest (MMI IX) shaking • $26 - $40 billion of total direct economic loss (1994 dollars) • $18.5 - $25 billion of building-related economic loss (1994 dollars) 6,340 fatalities (most due to collapse of smaller buildings) 25,000 serious injuries 300,000 homeless 150,000 collapsed/destroyed bldgs: More than 20% of buildings collapsed within 5 km of fault rupture $100 -$200 billion of total direct economic loss (1995 dollars) $80 - $150 billion of building-related economic loss (1995 dollars) 21 1994 M6.7 Northridge Earthquake Single-Family Dwelling (SFD) Damage – Wood Frame Buildings 22 1994 M6.7 Northridge Earthquake Multi-Family Dwelling (MFD) Damage – Wood Frame Buildings Woodland Hills Apartment Complex Older 2-story home with cripple walls (not in the study) Post-1960 2-story home Northridge Meadows Apartment Complex 23 24 6 1995 M6.8 Kobe Earthquake Severe (Heavy) Damage and Collapse of Short-Period Buildings Observations of Collapse from Shake Table Testing Examples of Large (> 10%) Drift Displacement Capacity Severe damage (incipient collapse) of a 2-story mixed-use building 1-story reinforced masonry coupled T-wall test structure (UC, San Diego) Modern Japanese 2-story wood house (E-Defense, Miki, Japan) Collapse of a 2-story house with weak walls and heavy clay-tile roof 25 Observed Performance – Key Findings for Modern Short-Period Buildings Collapse Probability Quantified with Red-Tag Data Post-1960 Wood Buildings - 1994 Northridge Earthquake Collapse Probability or Red Tag Percentage 3.0% 2.5% First-story collapse failure mode MCER Ground Motion Intensity Red Tag % - 186 Post-60 Census Tracts Red Tag % - 22 Equal Count Groups Best Fit (MLE) of Red Tag % (1st 21 Grps) P[Col] - 'Newer' W1 Bldgs. (FEMA P-155) Large lateral displacements at incipient collapse. Example Statistic Census Tract 115200 Buildings Count w/Red Tag 23 Total 905 2.0% 26 Low probability of collapse for MCER ground motions (where SMS ≤ 1.5g) 1.5% Benchmark collapse probabilities derived from Red-Tag data: 1-story buildings - P[Collapse|SMS = 1.5g] ≤ 2 percent 1.0% ≥ 2-story buildings - P[Collapse|SMS = 1.5g] ≤ 5 percent 0.5% 0.0% 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 0.3-Second Response Spectral Acceleration (g) 27 28 7 Three Seismic Force Resisting Systems (SFRSs) 1. Wood light-frame (aka Wood) 2. Special reinforced masonry Archetypes, Parametric Studies and Findings 2-story wood light-frame MFD 3. Steel SCBF Representative short-period (T ≤ 0.5 sec) building archetypes of each SFRS: 2-story special reinforced masonry COM Modern code-compliant construction. Height (one-, two-, and four-story). Occupancy (COM, SFD and MFD). 2-story steel SCBF COM Design level (High, Very High Seismic) 29 Two Seismic Ground Motion Design Levels High Seismic archetypes: Very High Seismic archetypes: Short-period MCER spectral response acceleration adjusted for site class effects SMT = SMS = 2.25g. Not required by FEMA P-695. Used in to investigate collapse performance under 1.5 x MCER ground motions that could occur in areas of very high seismicity (e.g. sites located close to fault rupture). Parametric Studies of Building Archetypes 2.0 Spectral Acceleration (g) Short-period MCER spectral response acceleration adjusted for site class effects SMT = SMS = 1.5g. SDC Dmax per Table 5-1A of FEMA P-695. 1.8 MCE SDC D (maximum) 1.6 MCE SDC D (min) or SDC C (max) 1.4 MCE SDC C (min) or SDC B (max) 1.2 MCE SDC B (minimum) 1.0 0.8 0.6 0.4 0.2 0.0 0 0.5 1 1.5 30 2 2.5 3 3.5 4 Period (seconds) Figure 5-2 FEMA P-695 Objective Investigate the effects of variations of archetype configuration and modeling parameters on the collapse potential of each of the three SFRSs. Nine Parametric Studies: Some common to various SRFSs. Some for specific SFRS. 31 32 8 Key Findings of High Seismic Baseline Studies Generically Applicable to all Three SFRSs Numerical models representing best-estimate response behavior and collapse performance. Failure mechanism of all high seismic baseline archetypes characterized by P-Delta failure at large displacement of the first story. Important to consider P-Delta effects in numerical models, particular for taller archetypes. Baseline archetypes incorporated all elements of real buildings e.g. nonstructural wall finishes for wood building archetypes. 33 40% Common Failure Mechanism 30% Improved Collapse Performance In general, the MCER collapse probabilities of onestory and two-story wood light-frame, special reinforced masonry and steel SCBF archetypes of this project comply with benchmark metrics. Exception for COM wood light-frame archetypes for which MCER collapse probabilities are much greater than benchmark values. Directly related to relatively low strengths of these archetypes. Wood MFD (1, 2, 4-story) Wood SFD (1 and 2-story) Masonry COM (1, 2, 4-story) Steel COM (1, 2, 4-story) 35% MCER Collapse Probablity Investigated variation in response behavior and collapse performance of representative archetypes of three short-period SFRs. 25% 20% 4-story 15% ASCE/SEI 7-16 10% 2-story 5% ≥ 2-story BM 1-story 1-Story BM 0% 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Period, T (seconds) 40% Wood High h/b (1 to 5-story) 2-story 35% Wood Low h/b (1, 2, 3-story) Masonry HG (1, 2, 4-story) 30% MCER Collapse Probablity Baseline Configuration Parametric Study Steel (2-story and 3-story) 25% 2-story 20% 15% 1-story 4-story 2-story 10% 5-story ASCE/SEI 7-16 3-story 4-story 5% 3-story ≥ 2-story BM 1-Story BM 0% 0 0.1 0.2 0.3 0.4 0.5 0.6 Period, T (seconds) 0.7 34 Key Findings of Very High Seismic Baseline Studies Generically Applicable to all Three SFRSs Key Findings of High Seismic Baseline Studies Generically Applicable to all Three SFRSs Key findings for High Seismic archetypes apply also to archetypes designed for Very High Seismic (e.g., same failure mechanism) Strength (Vmax/W) Strength found to be the most important parameter influencing collapse performance of the high seismic baseline archetypes. Collapse performance is worse for an archetype designed and evaluated for Very High Seismic loads than that of the same archetype designed and evaluated for High Seismic loads. Consistent trend in MCER collapse probability with strength for all archetypes of this project and, in general, those of prior FEMA P-695 studies Seismic loads for Very High Seismic regions are 50% higher than those for High Seismic regions. Actual strength for a Very High Seismic archetype is typically less than 50% greater than that of same archetype designed for High Seismic loads (e.g., less over-strength). “…the stronger the archetype the better the collapse performance…” Reasons for differences in over-strength between Very High Seismic and High Seismic archetypes are different for each SFRS. 35 36 9 Reasons for Differences in Over-Strength of Very High Seismic and High Seismic Archetypes Collapse Displacement Capacity Parametric Study Wood light-frame baseline archetypes: Only structural wood shear designed for 50 percent greater strength. Nonstructural walls (interior gypsum and exterior stucco) - similar for high seismic and very high seismic archetypes, hence combined strength of structural and nonstructural walls of very high seismic archetypes less than 50 percent higher than that of corresponding high seismic archetypes. Special reinforced masonry baseline archetypes: Wall cross-sectional area – More efficient designs for higher loads Investigated effects of collapse displacement capacity on response behavior and collapse performance. Motivated by results of shake table and pull tests of full-scale structures showing collapse displacements at story drift of 10% or greater without loss of stability. Significantly greater than collapse displacement capacity of nonlinear models of prior FEMA P-695 studies. Variations in the collapse displacement capacity of numerical models of archetypes were represented by different modeling assumptions of postcapping residual strength. Steel SCBF baseline archetypes: Brace Section – More efficient designs for higher loads 37 38 Modeling of Collapse Displacement Capacity Modeling of Collapse Displacement Capacity Wood light-frame archetypes: Special reinforced masonry archetypes: Displacement capacity expressed in terms of a residual strength based on a percentage of model peak strength. Displacement capacity expressed by slope of postpeak base shear – vs. – story drift curves. Affects more negatively collapse performance of weaker models (due to P-Delta effect). In reality, residual strength ratio of a weaker model should be larger than that of a stronger model. Lower collapse probabilities with increased collapse displacement capacity or increased post-capping residual strength. Residual Strength Plateau 39 Properties of detailed and simplified nonlinear models emulate post-peak behavior Strong trend of lower collapse probabilities with increased collapse displacement capacity. 40 10 Soil-Structure Interaction (SSI) and Foundation Flexibility Parametric Study: Key Findings of Collapse Displacement Capacity Parametric Studies Investigated SSI inertial and kinematic effects and foundation flexibility for two soil conditions (stiff and soft sites) on response behavior and collapse performance. SSI inertial effects modeled with a distributed set of discrete nonlinear soil springs and dashpots below flexible foundation elements. Kinematic SSI effects were evaluated by modifying frequency content of ground motion records (filtered records) used for response history analysis. Response and collapse results compared with those of corresponding archetype models on a rigid foundation excited by using unfiltered records. Displacement capacity at point of incipient collapse found to be the second most important parameter influencing collapse performance of wood light frame and special reinforced masonry high seismic baseline archetypes. Why? – For wood, displacement capacity resists the effects of P-Delta on side-sway collapse of the archetype model Specific value of displacement capacity at point of incipient collapse to incorporate in numerical models of a given building archetype is not well understood. Root Cause - Lack of test data at very large displacements 41 42 Key Findings of SSI and Foundation Flexibility Parametric Studies In general, modeling of SSI and foundation flexibility is not required for accurate calculation of short-period building response and collapse performance. Generic Collapse Performance Exception: Rocking of spread footing below single-bay of steel SCBF Results suggest that reduction in design base shear of short-period buildings due to SSI and foundation flexibility included in ASCE/SEI 7 is not justified except for unusual cases such as buildings with very large plan dimensions and/or deep foundations. 43 44 11 Short List of Prior Analytical Studies of Bilinear (SDOF) Models Equal Energy Criterion of an Elasto-Plastic SDOF Model Force Veletsos and Newmark, 1960, 2WCEE ”Effect of Inelastic Behavior on the Response of Simple Systems to Earthquake Motions” FE Miranda and Bertero, 1994, Earthquake Spectra, EERI “Evaluation of Strength Reduction Factors for Earthquake-Resistant Design” Equal areas according to equal energy criterion Improvement of Nonlinear Static Seismic Analysis Procedures, FEMA 440, June 2005 FI Effects of Strength and Stiffness Degradation on Seismic Response, FEMA 440A, June 2009 Tentative Framework for Development of Advanced Seismic Design Criteria for New Buildings, NIST GCR 12-917-20. NIST, 2012, Section 3.1 “Study of Short-Period Systems” 45 Comparison of Strength Reduction Factors from Various Studies dY Areas common to both elastic and inelastic responses dE dI Displacement Concepts by Veletsos and Newmark, 1960, 2WCEE Artwork by Andre Filiatrault, 2020, FEMA P-2139-1) 46 FEMA P-695 collapse results for bilinear SDOF systems with collapse displacement capacity 10 times yield displacement (NIST GCR 12-917-20) (Figure 8, “Evaluation of Strength Reduction Factors for Earthquake-Resistant Design”, Miranda and Bertero, 1994, Earthquake Spectra) 47 48 12 Comparison of Results of the SDOF study of FEMA P-2139-1 with those of Prior SDOF Studies of NIST GCR 12-917-20 Collapse Criteria of the SDOF Study Analysis methods and models: Same set of earthquake records and IDA methods Similar SDOF models Collapse displacement criteria: Similar ductility-based limits Different drift-based limits - 1st-story failure Different drift-based limits – multi-story failure “1-Story” Models Peak inelastic response results: Same trends and similar values “4-Story” Models Collapse results (e.g., ACMR/ACMR10%): Ductility-based – essentially the same trends and values Drift-based – very different trends and values 49 Plots of ACMR and the ratio of ACMR/ACMR10% as a function of the model period of SDOF models (V = 0.4W) Ductility-Based Collapse ( = 8.0) Drift-Based Collapse (7.5 Percent) Elastic Period of SDOF Model (sec.) Yield Displ. (in.) Collapse Displ. (in.) Yield Displ. (in.) Collapse Displ. (in.) Yield Displ. (in.) Collapse Displ. (in.) Model Height (ft.) Collapse Displ. (in.) Vmax = 0.2W Vmax = 0.4W Vmax = 0.8W 1st-Story Failure1 Multi-Story Failure2 Model Collapse Height3 Displ. (ft.) (in.) 0.1 0.02 0.16 0.04 0.31 0.08 0.63 10 9.0 10 9.0 0.15 0.04 0.35 0.09 0.71 0.18 1.41 10 9.0 10 9.0 0.20 0.08 0.63 0.16 1.25 0.31 2.5 10 9.0 10 9.0 0.25 0.12 0.98 0.25 1.96 0.49 3.9 10 9.0 12.3 11.1 0.30 0.18 1.41 0.35 2.8 0.71 5.6 10 9.0 15.7 14.2 0.35 0.24 1.92 0.48 3.8 0.96 7.7 10 9.0 19.3 17.4 0.40 0.31 2.5 0.63 5.0 1.25 10.0 10 9.0 23.1 20.8 0.45 0.40 3.2 0.79 6.4 1.59 12.7 10 9.0 27.0 24.3 0.50 0.49 3.9 0.98 7.8 1.96 15.7 10 9.0 31.1 28.0 1. 1st-story failure assumes that all inelastic story drift occurs at the 1st-story of a multi-story building represented by the SDOF model, where the height of the 1st-story is H = 10 feet. 2. Multi-story failure assumes that inelastic story drift is uniformly distributed equal over the height of a multi-story building. 3. Model height, H, represents the effective height which is assumed to be 2/3 of total height (hn) of a multi-story building, where the total height (hn) is calculated from the elastic period (T) of the model using the approximate fundamental period formula of Section 12.8.2.1 of ASCE 7-16, i.e., H = 2/3 exp[ln(T/1.4(0.02))/0.75]. 50 Summary of SDOF Model Comparisons Provide a theoretical basis for understanding why: short-period buildings have been observed to perform well in past earthquakes and improved numerical models of this project have found better collapse performance of archetypes with shorter periods. But ..... Collapse performance of theoretical bilinear elasto-plastic SDOF models may not be realistic representations of the collapse performance of the SFRS of interest. Ductility-based trend same as Fig. 3-1 NIST 51 52 13 Plots of notional collapse surfaces for SDOF models with periods, T = 0.15 s and T = 0.45 s, assuming 1st-story failure Notional Collapse Surfaces During the course of the ATC-116 project, the concept of a “collapse surface” was developed as a valuable tool to describe the interaction of primary building response properties affecting collapse performance of structures. Data from the generic collapse performance investigation of SDOF models was used to develop a series of notional collapse surfaces. Notional collapse surfaces (based on SDOF models) are conceptual and collapse surfaces of a given SFRS would need to be developed from collapse (ACMR) results of FEMA P-695 analyses (IDAs) of the SFRS of interest. 53 54 Collapse Surface Utilization Collapse surfaces describe the amplitude of a collapse metric (e.g., ACMR) as a function of key building response properties, i.e., strength, displacement capacity, elastic period and failure mode, that significantly influence the collapse performance of the system of interest. Paradox Solved For the system of interest, collapse surfaces could be used to either: Estimate collapse performance given values of building response properties, or, conversely Estimate the value of a building response property (e.g., strength) required to achieve a specific collapse objective. 55 56 14 The Short-Period Building Performance Paradox Resolution of the Short-Period Paradox Observed good performance: Why do analytical models of code compliant designs predict high probabilities of collapse for short-period buildings contrary to damage observed in actual earthquakes and the judgement of earthquake engineers? “Show me the bodies” Very few collapses of modern short-period buildings (e.g., evidenced by only 109 U.S. building-related earthquake fatalities in the last 50 plus years) Collapse trends reversed due to: Representative (realistic) archetypes Improved numerical models Theoretical insight (gained from the SDOF study) Bill Holmes MCER collapse probabilities (low) Collapse performance of improved nonlinear models of representative archetypes is generally consistent with observed good performance 57 58 Recommendations for Improving Seismic Design Recommendations for Improving Seismic Design For Improved Seismic Design Codes Intended primarily for seismic-code-development committees 59 60 15 Recommendations for Improving Seismic Design Recommendations for Improving Seismic Design For advanced seismic design and analysis practice For enhanced modeling, testing and data collection Intended primarily for engineering practitioners Intended primarily for research engineers and academicians Figures: FEMA P-2139-3 61 Figures: FEMA P-2139-3 62 Seismic Design Codes and Standards Seismic Design Codes and Standards Develop performance-based design criteria (collapse surfaces) to inform development of seismic design 63 Figures: FEMA P-2139-1 Update: ATC-154 Project is currently exploring 64 16 Seismic Design Codes and Standards Seismic Design Codes and Standards Consider actual strength and deformation capacity Recognize that future refinements in design and construction can affect undesigned overstrength and therefore seismic performance Figures: PEER 2020/20 65 Figures: FEMA P-2139-3, FEMA P-2139-4, CUREE W-29 66 Seismic Design Codes and Standards Review current ASCE 7 checks for deformation compatibility of Non-SFRS components Address increased collapse potential with very high seismic demand Very High 29.0 % Very High 29.0 % Update: ATC-154 Project is currently exploring Very High 26.0 % R COL I MCE ] P [COL IP [ MCER] Seismic Design Codes and Standards Very High High High 7.3% Very High 15.7 % High 19.0 % 13.4 % 67 17.2 % High 1.0% Figures: FEMA P-2139-3 High MODEL COM1 B COM2B COM3B MODEL COM4 B COM5B COM6B COMMERCIAL 5.5% 1.8% MFD1B MFD2B MFD3B MFD4B MFD5B MFD6B MULTI-FAMILY RESIDENTIAL Figure: Based on FEMA P-2139-2 Table 6-1 Very High 7.9% Very High 8.4% 2.6% SFD1B SFD2B SFD3B SFD4B SINGLE-FAMILY RESIDENTIAL 68 17 Seismic Design Codes and Standards Seismic Design Codes and Standards Undertake research to address targeted collapse probability: Undertake research to address targeted collapse probability: 1. Should the target for conditional probability of collapse be revised to be more in line with empirical experience? If so, what changes would be needed in the FEMA P-695 methodology to accommodate the revised target collapse probabilities? 2. Should the methodologies for estimating collapse probability be revised to explicitly consider strength from elements not a formal part of the seismic-force-resisting system, and, if yes, how would their contribution be controlled in design? 3. Should the target conditional probability of collapse be different for very high-seismic (i.e., near fault) sites than for other sites? If no, are more rigorous design rules needed for very high-seismic sites in order to achieve the target conditional probability of collapse? 4. Do the observations identified in item 4 and item 5 above hold true for longer-period structures? Figures: FEMA P-2139-1 69 Figures: FEMA P-2139-1 Seismic Design Codes and Standards Seismic Design Codes and Standards Update FEMA P695 seismic criteria including resolution of the following discrepancies: Revisit current ASCE 7 provisions for soilstructure interaction to determine whether current SSI reductions in design forces for short-period structures lead to worse collapse performance The seismic criteria of FEMA P695 are about 90% of those in ASCE 722 for short-period buildings and 60% of those in ASCE 7-22 for the velocity domain The seismic criteria of FEMA P-695 ignore higher levels of ground shaking typical of site closer to major faults, thereby implicitly permitting probabilities of collapse in excess of 10%,which is at odds with the stated intent of NEHRP and ASCE 7 seismic design provisions Figures: FEMA P-2139-1 71 Figures: FEMA P-2139-3 70 72 18 Seismic Design Codes and Standards Revisit current ASCE 7 provisions for soil-structure interaction to determine whether current SSI reductions in design forces for shortperiod structures lead to worse collapse performance Figures: FEMA P-2139-3, Figure 5-12 Advanced Seismic Design and Analysis 73 Advanced Seismic Design and Analysis 74 Advanced Seismic Design and Analysis Evaluate collapse safety issues associated with foundation rocking occurring prior to development of mechanisms in vertical elements Recognize overstrength not included in seismic design can create higher that anticipated element seismic forces, inconsistent with building code capacity design principals. A SHEARWALL STRONGER THAN NECESSARY MAY LEAD TO NEAR=ELASTIC FORCES, INCREASE LOADS TO DIAPHRAGMS AND THEIR CONNECTIONS Figures: FEMA P-2139-3 75 76 19 Enhanced Modeling, Testing and Data Collection Enhanced Modeling, Testing, and Data Collection Develop cyclic testing protocol up to and including incipient collapse or collapse 77 Figures: FEMA P-2139-3 Enhanced Modeling, Testing and Data Collection Enhanced Modeling, Testing and Data Collection Emphasize shake table testing to large displacements including insipient collapse, as well as bi-directional ground motions Emphasize testing of assemblies with realistic boundary conditions over component testing Figure: Ron Gallagher 79 Figures: PEER 2020/20 78 80 20 Enhanced Modeling, Testing and Data Collection Closing & Thank you Put plans in place for rapid and thorough collection of post EQ data PDH certificates Provided for participants of live webinar (not the recording) Distributed via email within four weeks Q&A Distributed via email within four weeks Figures: FEMA P-2139-1 81 82 21