Journal of Structural Engineering Seismic Simulation of Integrated Ceiling-Partition Wall-Piping System at E-Defense, Part 1: Influence of 3D Structural Response and Base Isolation --Manuscript Draft-Manuscript Number: STENG-3259R1 Full Title: Seismic Simulation of Integrated Ceiling-Partition Wall-Piping System at E-Defense, Part 1: Influence of 3D Structural Response and Base Isolation Manuscript Region of Origin: UNITED STATES Article Type: Technical Paper Section/Category: Seismic Effects Abstract: The seismic response of a full-scale, five-story steel moment frame building in baseisolated and fixed-base configurations with an integrated suspended ceiling-partition wall-sprinkler piping system that was shaken at E-Defense is critically assessed. Horizontal floor accelerations were constrained by the isolation systems to relatively low levels, which allowed observation of damage to the integrated system that was directly related to the vertical component of input acceleration. The floor slabs exhibited single mode vibration at their natural periods with widely varying effective damping. Peak vertical accelerations were amplified by an average factor ranging from 3 to 6 from the table to the middle of the floor slabs, where amplification factors increased as slab vibration periods lengthened. Damage to the ceiling-partion-piping components initiated at slab accelerations of about 2g, and became extensive for slab accelerations exceeding 5g. These metrics establish target vertical accelerations for achieving desired performance objectives. Corresponding Author: Keri L Ryan, PhD University of Nevada, Reno Reno, NV UNITED STATES Corresponding Author E-Mail: [email protected] Order of Authors: Keri L Ryan, PhD Siavash Soroushian Emmanuel Manos Maragakis Eiji Sato Tomohiro Sasaki Taichiro Okazaki Suggested Reviewers: Tara Hutchinson University of California, San Diego [email protected] Hutchinson led another recent test program of a full-scale building outfitted with nonstructural components and has desirable expertise for all aspects of the companion papers. However, Hutchinson is also a Co-PI on the Grand Challenge project, and so it might be considered a conflict of interest, although she was not involved in this aspect of the project. Amir Gilani Miyamoto International [email protected] Gilani has published experimental research on seismic response of ceiling systems and has expertise to evaluate structural and nonstructural responses observed in this research. Claudia Marin Howard University [email protected] Marin has expertise in seismic isolation and was involved in the test program on fullscale building outfitted with nonstructural components that was led by Hutchinson. Matthew and Hoehler Powered by Editorial Manager® ProduXion Manager® from Aries Systems Corporation Hilti Corporation [email protected] Opposed Reviewers: Additional Information: Question Response Is the article being considered for more No than one journal? The Journal of Structural Engineering does not review manuscripts that are being submitted simultaneously to another organization or ASCE journal for publication. Is this article already published? No Material that has been previously published cannot be considered for publication by ASCE. A manuscript that has been published in a conference proceedings may be reviewed for publication only if it has been significantly revised. If you answer YES, please explain. Have all the authors contributed to the Yes study and approved the final version? All authors must have contributed to the study, seen the final draft of the manuscript, and accept responsibility for its contents. It is unethical to list someone as a coauthor who does not want to be associated with the study and who has never seen the manuscript. Was an earlier version of the paper previously considered and declined by ASCE? Declined manuscripts are sent through the review process again. If your manuscript has been submitted to us before under a different title, please provide that title in the space provided below. It is our policy to inform an editor that a manuscript has been previously reviewed, even when it has been reviewed by a different Division, Institute, or Council within ASCE. No Do your table titles/figure captions cite No other sources? If you used a figure/table from another source, written permission for print and online use must be attached in PDF format. Permission letters must state that permission is granted in both forms of media. If you used data from another source to create your own figure/table, the data is adapted and therefore obtaining permission is not required. Does your paper exceed 10,000 words? If Yes YES, please provide justification in your Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation cover letter. If you need help estimating word length, see our sizing worksheet at this link: Sizing Worksheet Estimates for color figures in the printed No journal begin at $924. Cost increases depend on the number and size of figures. Do you intend for any figure to be printed in color? If YES, how many and which ones? Please provide a total count and also list them by figure number. Is this manuscript a companion to one Yes, Part 1 of 2 already submitted/or being submitted? If yes, please note whether this is part I, II, or III. Please make sure all related papers are uploaded on the same day and provide the date of submission, title, and authors of each. Is this manuscript part of a Special Issue? No If yes, please provide the Special Issue title and name of the guest editor. To read ASCE's Data Sharing Policy, please click on the "Instructions" link associated with this question. According to this policy, you are required to report on any materials sharing restrictions in your cover letter. Are you restricted from sharing your data & materials? If yes, did you report on these in your cover letter? No Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation Cover Letter Click here to download Cover Letter: Cover Letter_Revision 1.pdf College of Engineering University of Nevada Reno Oct. 26, 2014 Sherif El-Tawil, Ph.D., P.E., F.ASCE, Dept. of Civil and Env. Engineering University of Michigan Ann Arbor, MI 48109-2125 Ph (734) 764-5617 Fax (734) 764-4292 [email protected] Managing Editor of ASCE Journal of Structural Engineering Dear Professor El-Tawil: On behalf of the NEES/E-Defense collaborative research project on base-isolation and nonstructural components, I hereby submit revised versions of STENG-3259 and STENG-3260 for further consideration as Technical Papers in Journal of Structural Engineering. The titles of the manuscripts have been revised to “Seismic Simulation of Integrated Ceiling-Partition Wall-Piping System at E-Defense, Part 1: Influence of 3D Structural Response and Base Isolation”, and “Seismic Simulation of Integrated Ceiling-Partition Wall-Piping System at E-Defense, Part 2: Evaluation of Nonstructural Damage and Fragilities”. We wish the revised manuscripts to be reviewed as companion papers. The reviewer requested additional information in Part 1 so that the study could be better understood without referring to other documents. Therefore, the length of the submitted manuscript for Part 1 has increased from 9 to 10 pages according to the sizing worksheet estimate. We believe the additions have led to a more readable paper, and the additional length is justified. The length of Part 2 remains at an estimated 10 pages. To reiterate from the original submission, the manuscripts contain color figures, but all figures can be understood in black and white. Thus, our intention is for figures to be published in color electronically and in black and white for the printed journal. We look forward to learning the outcome of the manuscript peer review process. If there are any problems with the submission, please let me know. Sincerely, Keri L. Ryan, Ph.D. Associate Professor Department of Civil and Environmental Engineering University of Nevada, Reno/258 Reno, Nevada 89557-0152 (775) 784-6937 office (775) 784-1390 fax Manuscript Click here to download Manuscript: E-Defense companion paper 1 - Revision 1 - No Figures.docx 1 Seismic Simulation of an Integrated Ceiling-Partition Wall-Piping 2 System at E-Defense, Part 1: 3D Structural Response and Base 3 Isolation 4 Keri L. Ryan,a) Siavash Soroushian,b) E. "Manos" Maragakis,c) Eiji Sato, d) 5 Tomohiro Sasaki, e) Taichiro Okazaki f) 6 Abstract 7 The seismic response of a full-scale, five-story steel moment frame building in base-isolated 8 and fixed-base configurations with an integrated suspended ceiling-partition wall-sprinkler 9 piping system that was shaken at E-Defense is critically assessed. Horizontal floor accelerations 10 were constrained by the isolation systems to relatively low levels, which allowed observation of 11 damage to the integrated system that was directly related to the vertical component of input 12 acceleration. The floor slabs exhibited single mode vibration at their natural periods with widely 13 varying effective damping. Peak vertical accelerations were amplified by an average factor 14 ranging from 3 to 6 from the table to the middle of the floor slabs, where amplification factors 15 increased as slab vibration periods lengthened. Damage to the ceiling-partion-piping components 16 initiated at slab accelerations of about 2g, and became extensive for slab accelerations exceeding a) Assoc. Prof, Dept. of Civil and Environmental Engineering, University of Nevada, Reno, MS 0258, Reno, NV 89557-0258 b) Post-doctoral Scholar, Dept. of Civil and Environmental Engineering, University of Nevada, Reno, MS 0258, Reno, NV 89557-0258 c) Dean of Engr., Dept. of Civil and Environmental Engineering, University of Nevada, Reno, MS 0258, Reno, NV 89557-0256 d) Dr. Engr., National Research Institute for Earth Science and Disaster Prevention, 1501-21 Nishikameya, Mitsuta, Shijimi-cho Miki, Hyogo, Japan 673-0515 e) Dr. Engr., National Research Institute for Earth Science and Disaster Prevention, 1501-21 Nishikameya, Mitsuta, Shijimi-cho Miki, Hyogo, Japan 673-0515 f) Assoc. Prof., Graduate School of Engineering, Hokkaido University, Kita 13, Nishi 8, Kita-ku, Sapporo, Hokkaido, Japan, 060-8628 17 5g. These metrics establish target vertical accelerations for achieving desired performance 18 objectives. 19 Keywords: seismic isolation, nonstructural components, seismic response, vertical ground 20 motion, floor slab vibration, shake table testing 21 Introduction 22 The performance of critical facilities such as hospitals and fire stations during an earthquake 23 depends not only on the structural systems, but also on the functionality of nonstructural 24 components. The shaking intensities that can cause damage to the nonstructural components are 25 typically lower than those that induce structural damage (Miranda, 2003), and the influence of 26 vertical shaking on nonstructural components has not been well studied. Several component and 27 subsystem level experiments of nonstructural components such as suspended ceilings, partition 28 walls, and sprinkler piping, have been conducted in recent years (e.g. Badillo-Almarez et al., 29 2007; Gilani et al., 2010; Zaghi et al., 2012; Tian et al., 2012; Retamales et al., 2012). However, 30 these component and subsystem level experiments may not accurately reflect the following 31 influences of real buildings on nonstructural response: realistic input excitation, interaction 32 between different types of nonstructural components, realistic boundary conditions, and floor 33 system vibration. 34 Mitigation strategies like seismic base isolation are often chosen when the seismic 35 performance objective is post-earthquake continued functionality (Taylor and Igusa, 2004). Base 36 isolation has been shown to effectively protect structural systems from damage through the 37 substantial reduction of story drifts, but its ability to eliminate nonstructural damage has not been 38 proven conclusively. Nonstructural components in base-isolated buildings are subjected to a 39 different proportion of horizontal to vertical accelerations than prescribed by seismic 2 40 qualification tests, and so the observations from component tests referenced above may not 41 strictly apply. Specifically, ACC 156 prescribes that the horizontal acceleration is amplified over 42 the building height while horizontal acceleration in a base-isolated building is significantly 43 reduced relative to the ground acceleration (ICC 2010). 44 Observations of the response of numerous base-isolated buildings in past earthquakes, mostly 45 favorable, have been reported (e.g. EERI, 1996; Saito, 2006; Moroni et al., 2012; Gavin and 46 Nigbor, 2012; Kasai et al., 2013). Although problems with expansion joints and some disruption 47 of contents (e.g. sliding or toppling of furniture, items falling from shelves) have been observed, 48 significant nonstructural component damage has not been reported. For instrumented buildings, 49 authors confirmed that the isolation system generally attenuated the horizontal accelerations 50 recorded at the floor level relative to the ground, especially when the input ground acceleration 51 was significant enough to activate the isolation systems (Kasai et al., 2013). It does not appear 52 that any isolated building, constrained to low horizontal floor accelerations, has been subjected 53 to significant vertical accelerations; however, most reports do not mention vertical response. 54 Among the studies we examined, the peak vertical acceleration recorded in any isolated building 55 was 0.76g, which was observed in the base-isolated Nursing Home Building of the Niigata-ken 56 Cheutsu Earthquake of 2004 (Saito, 2006), for which no damage was reported. In the 2011 57 Christchurch Earthquake, free field vertical accelerations of 0.6g were recorded near the 58 Christchurch Women’s Hospital (Bradley and Cubrinovski, 2011), which was not instrumented. 59 Nonstructural damage was reportedly “limited to minor cracking of partitions around window 60 openings” (Mayes et al., 2012). 61 Although the accelerations were not large, significant amplification of vertical accelerations 62 from ground to roof level in a few buildings has been documented. Bozorgnia et al. (1998) 3 63 examined 12 buildings – including base-isolated and conventional with varying levels of vertical 64 instrumentation – that were shaken in the 1994 Northridge Earthquake. Three of these buildings 65 contained vertical accelerometers both at the ground level and on the roof slab away from a 66 column. The ratios of peak vertical accelerations recorded in the roof to those at the ground level 67 were 3.75, 6.4, and 2.4. In addition, Lew and Hudson (1999) concluded, by examination of 68 several instrumented buildings, that vertical accelerations transmitted from the ground to the 69 building were not affected by the presence of an isolation system. 70 Besides field observation, the attenuation of input horizontal accelerations at upper levels by 71 base isolation has also been demonstrated numerous times by shaking table tests (Kelly et al., 72 1980; Al-Hussaini et al., 1994; Clark et al., 1997; Wolff and Constantinou, 2004; Fenz and 73 Constantinou, 2008; to name a few). The vast majority of these have been conducted on bare 74 frame structures with added mass, and thus could not explicitly address the performance of 75 nonstructural components. Only recently, state-of-the-art large capacity shake table facilities 76 have provided opportunities to test full-scale base-isolated buildings constructed with realistic 77 floor systems and nonstructural components, such as the tests of a full-scale 4-story reinforced 78 concrete medical building at Hyogo Earthquake Engineering Research Center/E-Defense (Sato et 79 al., 2011). In a specific investigation of the effect of vertical excitation in these tests, Furukawa 80 et al. (2013a) reported that the peak vertical acceleration recorded at the table (0.5g) was 81 amplified to about 4g at the center of the roof slabs during a 3-dimensional (3D) El Centro 82 ground motion. Disruption and damage to building contents and equipment (the main focus of 83 the study) was significant, but damage to the Japanese-designed suspended ceilings, piping and 84 sprinkler systems, and partition walls/sliding doors was not observed. 4 85 This paper focuses on another major test of a full-scale base-isolated building incorporating 86 nonstructural components. As part of a collaborative research project between Network for 87 Earthquake Engineering Simulation (NEES) and the National Research Institute for Earth 88 Science and Disaster Prevention (NIED) of Japan, system-level full-scale shaking experiments of 89 a 5-story steel moment frame building were conducted at E-Defense. The building was tested in 90 three different configurations: 1) base-isolated with triple pendulum (TP) bearings (TP 91 configuration), 2) base-isolated with a hybrid isolation system (hybrid configuration), and 3) 92 fixed-base configuration. An integrated ceiling, partition wall, and fire sprinkler piping (CPP) 93 system was installed in the building for each test configuration, as led by the NEES Grand 94 Challenge project “Simulation of the Seismic Performance of Nonstructural Systems”. This is 95 the first of two related papers. In this paper, we evaluate the 3D structural response in all 96 configurations, and identify the relation between CPP damage and observed horizontal floor 97 accelerations and vertical slab accelerations. In the follow-up paper (Soroushian et al., 2013a), 98 the response of the integrated ceiling, partition, and fire sprinkler piping systems is discussed, 99 damage states are identified, and fragility functions are developed. 100 Experimental Setup 101 Building Testbed 102 The CPP system was installed in a 5-story testbed building (Fig. 1) that was utilized for this 103 collaborative NEES/E-Defense project. This building was approximately 16 m (53 ft) tall, 104 asymmetric in plan with dimensions of 10 m by 12 m (33 ft by 40 ft), 2 bays by 2 bays, and 105 weighed about 5220 kN (1174 kip) (Fig. 2(a)). Lateral resistance was provided by steel moment 106 frames in both directions. The columns were made from 350 mm x 350 mm (13.8 in x 13.8 in) 107 hollow square sections with thicknesses that varied from story to story. The beams were either 5 108 rolled or built-up I-sections. The primary beams, which were connected to the columns via fully 109 restrained moment connections, consisted of 3 x 400 mm (15.7 in) deep segments that were 110 bolted together at the approximate inflection points determined from gravity loading. The end 111 segments were haunched to improve their bending strength and the beam-to-column connection 112 strength. Secondary beams spanned between the primary beams as configured in Fig. 2(a); the 113 beams – with depths ranging from 200 to 400 mm (7.9 in to 15.7 in) – were bolted to the primary 114 members (idealized as pin connections). 115 The floor system in floors 2-5 consisted of 80 to 155 mm (3.15 to 6.10 in) thick concrete 116 slabs cast on 1.2 mm (0.047 in) thick corrugated metal decking. Typical slab reinforcement was a 117 single layer of 10 mm (0.39 in) diameter bars placed at the mid surface of the slabs, and spaced 118 150 mm (5.90 in) in both directions. The roof slabs were composed of 150 mm (5.90 in) thick 119 concrete slabs cast on a 1.2 mm (0.047 in) flat steel deck. Two layers of 13 mm (0.51 in) 120 diameter bars spaced at 200 mm (7.9 in) provided reinforcement in each direction. On average, 121 the roof slab was thicker than the floor slabs, to carry additional weight as described below. 122 Shear studs, which were covered by concrete for protection, connected the concrete slabs to the 123 primary beams to provide composite beam-slab behavior. 124 In addition to the weight of structural components, concrete and steel weights were installed 125 to simulate a realistic live load. Concrete blocks, which contributed an additional 175 to 257 kN 126 (39.3 to 57.8 kip) per floor, were installed on floors 2 to 5 with representative dimensions and 127 position as shown in Fig. 2(b). Where the blocks intruded the effective width of the composite 128 slab, polystyrene foam was inserted between the block and the slab so that the blocks did not 129 contribute to composite bending of the floor system. Steel plates weighing 535 kN (123 kips) 130 were installed on the east side of the roof to simulate the weight from a combination of roof 6 131 mounted equipment (e.g. air conditioner system or water tanks) and a penthouse (Fig. 2(c)). 132 Each weight included either 7 or 8 steel plates with dimensions 2.1 x 4.3 x 0.025 m (6.9 x 14.1 x 133 0.08 ft). The weight at the roof was altered from the testbed’s original design configuration to 134 introduce a significant mass eccentricity. 135 Two independent isolation systems were designed, incorporated beneath the building, and 136 tested as part of this experiment. Each isolation system design was developed by the project team 137 in conjunction with the bearing supplier to meet distinct objectives. The TP isolation system 138 incorporated 9 identical TP bearings (Fig. 3(a)), one beneath each column depicted in the plan 139 layout in Fig. 2(a). The properties of these bearings were chosen to accommodate the largest near 140 fault ground motions that could be replicated by the E-Defense shake table. The bearings were 141 1.4 m (4.6 ft) in diameter, with a displacement capacity of 1.13 m (3.7 ft) at a base shear 142 coefficient V/W = 0.275, where V = isolator shear force and W = axial force due to building self- 143 weight. The inner pendulum mechanism (period T = 1.84 s) is engaged in minor shaking, with 144 transition to the outer pendulum mechanisms (period T = 5.57 s) when V/W first reaches about 145 0.08. The isolation system reached a peak displacement of 0.7 m during testing. The TP bearings 146 were very stiff in compression, while providing no resistance in tension. The bearing vertical 147 stiffness was assumed to correspond to a rigid mass vertical frequency of 30 Hz (T = 0.03 s). 148 Given the history of investigation of elastomeric bearings for nuclear applications, the 149 hybrid isolation system evolved from the need to verify the stability and load-carrying capacity 150 of elastomeric bearings at displacements representative of the nuclear industry’s extended design 151 basis ground motions. The hybrid isolation system featured 4 lead-rubber (LR) bearings (Fig. 152 3(b)) for lateral resistance, each 0.7 m (2.3 ft) in diameter with a 102 mm (4 in) lead core and a 153 shape factor S = 29. The manufacturer provided nominal bearing properties were: elastic 7 154 stiffness = 6.5 kN/mm (37 kip/in), post-yield stiffness = 0.65 kN/mm (3.7 kip/in), characteristic 155 strength = 65.7 kN (14.8 kip), and nominal vertical stiffness in compression = 1500 kN/mm 156 (8566 kip/in). The design of the system was also heavily influenced by the building’s light 157 weight; to provide the desired isolation period shift while simultaneously accommodating the 158 overturning moments, the LR bearings (located beneath edge columns) were supplemented with 159 5 cross-linear (CL) bearings (Fig. 3(c)), which were located beneath center and corner columns. 160 The CL bearings were composed of two sets of perpendicular rails with top and bottom guided 161 plates separated by nearly frictionless ball bearings. These low friction sliders share the vertical 162 load with the LR bearings without increasing the horizontal stiffness or base shear of the 163 isolation system. Furthermore, the CL bearings enhance the global system stability by allowing a 164 redistribution of axial force between LR and CL bearings, while also providing tension 165 resistance. The CL bearings were stiff relative to the LR bearings in the vertical direction, 166 with manufacturer-supplied nominal vertical stiffnesses of 3471 kN/mm (19821 kip/in) in 167 compression and 245 kN/mm (1399 kip/in) in tension. 168 The LR bearings reached a peak displacement of 0.55 m during testing. The nominal 169 isolation system properties provided horizontal post-yield period Td = 2.6 s and characterized 170 strength Qd/W (summed over 4 bearings) = 0.050. Based on the sum of the nominal vertical 171 stiffnesses, the rigid mass vertical frequency of the hybrid isolation system was 30 Hz (T = 0.03 172 s). Additional details about the testbed, isolation system design, and overall setup are available in 173 Dao (2012) and Ryan et al. (2013a) for the interested reader. 174 Nonstructural Systems 175 A partition-ceiling-sprinkler piping subassembly was designed and installed in nearly 176 identical configuration over two complete floors of the testbed building. These components were 8 177 installed below the 5th and roof floor slabs, which were expected to draw the maximum floor 178 accelerations. The following components were included on each floor: approximately 83.6 m2 179 (900 sf) of lay-in-tile suspended ceiling designed per ASTM E580/E580M-11ae1 (ASTM, 2011), 180 a standard Schedule 40 piping system conforming to NFPA 13 (2011), and approximately 90 m 181 (300 ft) of typical light gauge steel studded gypsum partition walls with individual lengths 182 varying from 1.5 to 9.8 m (5 to 32 ft). Detailing variations were considered, such as: suspended 183 ceiling sections with or without seismic bracing; straight drops, armover drops, or flexible hose 184 drops for the piping sprinkler heads; and slip track or full connection details for partition walls. 185 Further details of the CPP components can be found in Soroushian et al. (2013a). 186 Instrumentation 187 The table accelerations and the responses of structural and CPP components were monitored 188 by nearly 400 sensor channels (not including the isolation system response, when applicable) 189 recorded at a sampling frequency of 1000 Hz. A 4-pole low-pass Butterworth filter with a cutoff 190 frequency of 25 Hz was applied to all recorded responses, unless otherwise mentioned. 191 Figure 4 shows the layout of accelerometers on a typical floor that are applicable to the 192 responses presented here. Achieved table accelerations were measured by triaxial accelerometers 193 mounted at the 4 corners of the shake table (Fig. 4). Floor accelerations (2 horizontal and vertical 194 components) were measured using triaxial accelerometers installed at the southeast (SE), 195 northeast (NE) and northwest (NW) corners of every floor (Fig. 4). Each accelerometer was 196 attached to the corresponding column face just above the floor slab. Vertical accelerations at 197 intermediate locations on the floor slab and at the center column were also recorded by uniaxial 198 accelerometers attached to the bottom of the slabs. Details of the instrumentation used to 199 measure the response of CPP components are provided in Soroushian et al. (2013a). All data 9 200 discussed in this paper is archived and publicly accessible through the NEES Project Warehouse 201 (Ryan et al. 2013b,c,d). 202 Earthquake Simulation Plan 203 The building was subjected to various ground motions over three days of testing for the TP 204 configuration, two days of testing for the hybrid configuration, and one day of testing for the 205 fixed-base configuration. Out of 41 total earthquake simulations, 23 targeted 3D motion 206 including a vertical component. The achieved table accelerations encompassed a wide range of 207 shaking intensities and frequencies, which allowed the CPP system vulnerability to be critically 208 addressed. Table 1 lists a subset of motions that includes all 3D motions and any other motions 209 specifically mentioned in the papers. The majority are multi-component acceleration histories 210 recorded in previous earthquakes. Included for each motion is earthquake and station where 211 recorded (as applicable), basis for selection, theoretical scale factor (as applicable, described 212 below), system configuration, actual scale factor applied to the recorded motion during the test, 213 target acceleration or PGA, and ratio of recorded table acceleration to target acceleration. The 214 latter three statistics are provided separately for X/Y and Z components. 215 Different considerations factored into the selection of input ground motions for each test 216 configuration (Dao et al., 2012; Ryan et al., 2013a). This discussion focuses on the TP 217 configuration, which was the basis for many of the vertical ground motions applied throughout 218 the test program. For the TP configuration, a target spectrum was developed for a high 219 seismicity, Los Angeles site on site class D soil with SDS = 1.18g and SD1 = 0.71g. Ground 220 motions recorded in similar site conditions were selected to represent the service, design and 221 maximum considered earthquake (MCE) at the U.S. site (the first six shaded records of Table 1). 222 Amplitude scale factors (Theoretical Scale Factors in Table 1) were developed to minimize the 10 223 least square error between 1.3 times the 5% damped target spectrum and the SRSS of the 5% 224 damped spectra of the two horizontal components of motion, over a period ranging from 0.5 to 225 1.25 x Teff, where Teff = effective period of the isolation system. During testing, the theoretical 226 scale factors were applied to the service and design level motions, but MCE motions were 227 generally applied at a scale factor of 1.0 or the largest scale factor that could safely be 228 accommodated. Besides the motions for the U.S. site, additional motions were selected to 229 represent Japanese design practice, and to investigate the response of a base-isolated building 230 during long duration, long period subduction motions like those recorded in the 2011 Great East 231 Japan (Tohoku) Earthquake, which were of special interest to Japanese collaborators. Many of 232 these same ground motions were also applied to the hybrid configuration and the fixed-base 233 configuration to provide direct comparisons. Horizontal components applied to the fixed-base 234 configuration were generally applied at reduced scale to ensure the structural system did not 235 yield. In addition, two synthetic motions representative of the nuclear power industry’s extended 236 design basis for a Central and Eastern U.S. site (Vogtle) and design basis for a Western U.S. site 237 (Diablo Canyon) were selected and applied at various scale factors to the hybrid configuration. 238 While a vertical design spectrum was not developed for the test, the recorded components of 239 vertical ground motions were generally scaled proportionally to the horizontal components. In 240 some simulations, the vertical component was omitted (i.e. the target vertical acceleration was 241 zero) due to table limitations or to preserve the integrity of the CPP system. To interpret the 242 vertical motions that were imposed to the system, a vertical design spectrum and MCE spectrum 243 was developed to match the horizontal spectrum for the Los Angeles site mentioned earlier, on 244 the basis of NEHRP recommendations (FEMA, 2009). The NEHRP recommendations are based 245 on the findings of many studies investigating vertical to horizontal (V/H) spectral ratios such as 11 246 Bozorgnia and Campbell (2004). The 5% damped acceleration spectra of the realized (not target) 247 vertical motions applied during the test program are compared to the vertical design and MCE 248 spectra in Fig. 5. The realized vertical motions are mostly enveloped by the MCE spectrum with 249 the exception of the Northridge-Rinaldi ground acceleration, applied at a target scale factor of 250 88% (RRS88), which was replicated similarly in all three configurations. The realized RRS88 251 motion may be outside of what would normally be considered in design; however, larger vertical 252 components have been recorded in the near-fault region during prior earthquakes; for instance, 253 near the population center during the relatively moderate magnitude Christchurch Earthquake 254 (Bradley and Cubrinovski, 2011). 255 Testbed System Identification 256 White noise motions were applied to the building in the fixed-base configuration; these 257 motions included unidirectional (X or Y-direction) white noise at the beginning and end of the 258 experiment (2 repetitions), and 3D white noise before and after every earthquake simulation (10 259 repetitions). The normalized frequency responses of the recorded floor accelerations with respect 260 to the table accelerations (also known as transfer functions) were analyzed to estimate the 261 vibration periods and damping ratios of the natural modes of the testbed building in the 262 horizontal direction. The transfer functions were smoothed using a periodic Hamming window 263 with 50% overlap ratio. The periods and damping ratios corresponding to the fundamental 264 response modes were evaluated by curve fitting theoretical transfer functions to the measured 265 transfer functions using a least squares algorithm. 266 The average (over the repetitions) periods and damping ratios of the first three modes in each 267 horizontal direction, computed from the transfer functions of the average horizontal floor 268 acceleration (recorded at 3 columns, Fig. 4) relative to the average horizontal input acceleration 12 269 (recorded at 4 locations on the table), are listed in Table 2. These horizontal structural vibration 270 modes together comprise the majority of the mass participation in the horizontal direction. Slight 271 differences in the fundamental periods determined from 3D white noise compared to 272 unidirectional white noise were observed. 273 System identification in the vertical direction was performed using vertical accelerations 274 recorded at the center of 3 of the 4 slab quadrants (ASlabSE, ASlabNE and ASlabNW in Fig. 4), which 275 are referred to hereafter by location as SE, NE and NW. The transfer functions of these 276 accelerations (slab transfer functions) during 3D white noise were computed with respect to the 277 nearest accelerometer at the shake table level (ATSE, ATNE, ATNW in Fig. 4). Figure 6 illustrates 278 representative transfer functions. A single dominant peak can be observed in the transfer function 279 for the roof SE slab quadrant (Fig. 6(a)), which suggests that its response was dominated by 280 vibration in a single mode. However, multiple vibration modes contributed to the response at 281 some locations, as demonstrated by the multiple peaks in the transfer function for the 5th NE slab 282 quadrant (Fig. 6(b)). 283 The transfer function for each slab quadrant was processed using the previously described 284 techniques to identify a dominant participating modal period and damping ratio, but the 285 evaluation of vertical mode shapes was not attempted. This simplified approach was expected to 286 provide reasonable accuracy when the response is primarily single mode (Fig. 6(a)) but have 287 limitations for multi-mode response (Fig. 6(b)). The median (λ) and dispersion (β) – over the 288 white noise repetitions – of the natural period and damping ratio of each slab quadrant are 289 presented in Table 3. These results suggest that the first two vertical modes were localized 290 primarily to the roof and 5th floor east side slab quadrants, with periods of 0.13 sec (7.7 Hz) and 291 0.1 sec (10 Hz), respectively. This finding is consistent with Furukawa et al. (2013b), who 13 292 evaluated the vertical modes of this building using more precise techniques. The SE and NE roof 293 slab quadrants had a lengthened vibration period due to the substantial added mass on the east 294 side of the roof level. The cause of the increase for the 5th floor slab relative to lower floors is 295 unclear. The dominant period for the remaining slab quadrants was either 0.09 sec or 0.08 sec, 296 which suggests that one or two system modes contributed mainly to the slab vibration over the 297 rest of the building. Note that the dispersion in the observed vibration periods was close to zero, 298 meaning that little variation was observed from one white noise repetition to the next. 299 The slab damping ratios found by the approximate technique were more difficult to interpret, 300 with greater variation in the median damping in slab quadrants contributing to the same mode as 301 well as greater variation over the repetitions (based on the dispersion). The estimated damping 302 ratios were 1 – 1.5% for the first mode at 0.13 sec and 2 – 3.25% for the 2nd mode at 0.10 sec. 303 The estimated damping ratios for the vertical system mode(s) at 0.08 – 0.09 sec varied widely 304 from 3% to 10%, which suggests that the approximate technique was not very reliable. With 305 caution against drawing conclusions from this data, we observe a possible correlation between 306 slab damping ratio and modal period (damping ratios were lower for more flexible modes). 307 Overall, the observed periods of slab vibration modes ranged from about 0.08 to 0.13 sec (7.7 308 to 12.5 Hz). In comparison, the slab vibration frequencies of composite floor systems ranged 309 from 4 to 12 Hz in several newly constructed buildings, as measured through low level forced 310 vibration with an eccentric mass shaker (Hicks, 2004). Murray et al. (1997) and Allen and 311 Pernica (1998) suggested that natural frequencies of composite steel floor systems range from 5 312 to 9 Hz and 3 to 10 Hz, respectively, based on expert opinion. Boice (2003) found this range to 313 be 3 to 13 Hz based on floor response measured by heel drops, walking, and ambient vibration 314 tests for 103 case studies. Although the supplementary mass at the roof level may be viewed as 14 315 an unusual design configuration, the range of frequencies observed in this experiment is 316 consistent with the range of normal established in prior studies. 317 The slab vibration properties were also assessed during earthquake simulations. Because the 318 transfer function data was more difficult to interpret for earthquake simulations than for white 319 noise motions, a different procedure was used. First, the 5% damped response spectral 320 acceleration of the measured vertical acceleration at the middle of the slab quadrant, or 321 “measured slab spectrum” was computed. Next, the slab quadrant (Fig. 7(a)) was idealized as a 322 single degree-of-freedom (SDOF) oscillator (Fig. 7(b)) with period and damping ratio to be 323 determined. The response history and 5% damped response spectral acceleration of the idealized 324 SDOF system subjected to average vertical column acceleration (recorded at the center and 325 nearest corner column accelerometers, Fig. 4) were computed iteratively, with a search period 326 increment of 0.01 sec and damping ratio increment of 1%. (Since no sensor was installed at the 327 center column for the 2nd, 3rd and 4th floor slabs, the center column sensor at the 5th floor slab was 328 used instead.) The period and damping ratio of the idealized SDOF system that matched the 329 recorded slab acceleration were identified by a least squares fit of the measured slab spectrum to 330 the idealized SDOF system spectrum. This SDOF idealization procedure was expected to suffer 331 from similar limitations as the white noise transfer function procedure; nevertheless, the “best 332 fit” idealized SDOF system acceleration history and spectrum generally corresponded adequately 333 to the measured versions, where the correspondence was closer for slab quadrants with known 334 localized dominant modes at the 5th and roof level (e.g. Fig. 8(a)) compared to those with system 335 modes or multi-mode response (e.g. Fig. 8(b)). 336 The above process, which was computationally intensive, was applied to a subset of the slab 337 quadrants for four earthquake simulations. The selected simulations – all from the hybrid 15 338 configuration – included the synthetic Vogtle site motion at target scale factors of 75%, 125% 339 and 175% (VOG75, VOG 125, VOG175) and the 88% Northridge at Rinaldi (RRS88), which 340 represent the range of vertical intensity observed in the test program. In particular, only intensity- 341 related variation was present in the Vogtle simulations since the same motion was scaled to 342 different intensities. The resulting periods and damping ratios determined by the aforementioned 343 process are listed in Table 4. 344 Recognizing that a different methodology was applied and a different configuration was used 345 for earthquake excitation, the best fit slab vibration periods (Table 4) did not deviate from those 346 determined for white noise motions (Table 3) by more than 0.01 sec. Therefore, the ground 347 shaking intensity and the building configuration were concluded not to affect the slab vibration 348 period. The estimated damping ratios in most floor slabs increased for earthquake simulations 349 compared to white noise motions. This increased energy dissipation during earthquake excitation 350 could have resulted from mild nonlinearities (cyclic opening and closing of cracks in the 351 concrete slabs) during vertical vibration. Nonetheless, the trends for relative damping in the 352 various modes based on white noise were upheld, so that the damping ratios for earthquake 353 excitation also appeared to be lowest on the East side of the 5th floor and roof slabs, where the 354 most vulnerable components of the CPP system were located during the experiment. The nature 355 of the recorded structural response during 3D shaking is explored next. 356 Structural Response 357 Horizontal Floor Acceleration 358 As is well known, base isolation attenuates horizontal floor accelerations throughout the 359 structure at great benefit to nonstructural components. Representative floor acceleration profiles 360 (peak vector sum acceleration vs. height) recorded for each system configuration are shown in 16 361 Fig. 9(a), and in Fig. 9(b) the peak floor accelerations at each level have been normalized by the 362 peak table acceleration (PGA). Figure 9 confirms that horizontal accelerations were attenuated in 363 the isolation configurations relative to the ground acceleration, while they were amplified over 364 the height in the fixed-base configuration. The range of peak horizontal floor accelerations 365 recorded at the 5th and roof level was 0.12 - 1.12g for the isolated configurations and 0.46 - 1.22g 366 for the fixed-base configuration. The largest floor accelerations in all configurations were 367 recorded during the RRS simulation, where the isolation configurations were subjected to 368 horizontal PGA = 1.21g and the fixed-base configuration to horizontal PGA = 0.41g. 369 Other than mild cracking in the concrete slabs, some of which was present in the testbed 370 building prior to this experiment, structural damage was not observed in any of the simulations. 371 Horizontal floor acceleration, generally constrained to peak values well below 1g in the isolation 372 configurations (with the exception of the RRS simulation), is not believed to be the primary 373 cause of damage to the CPP components. Rather, the vertical slab acceleration was more closely 374 correlated to the CPP component damage that was observed (evidence is presented later). Thus, 375 the nature of the vertical slab vibration is discussed next. 376 Vertical Column and Slab Acceleration 377 To provide an overview of how vertical acceleration propagated from the shake table up 378 through the building columns and into the floor slabs, the peak accelerations recorded in the 379 shake table, columns, and middle of the floor slabs for every 3D simulation are investigated. The 380 data in this section has been processed with a modified filter cutoff frequency of 50 Hz to point 381 out higher frequency effects that were observed. Figure 10 presents the peak column 382 accelerations – absolute and normalized by vertical PGA - for the NW column in Fig. 10(a)-(b) 383 and for the SE column in Fig. 10(c)-(d). Column accelerations are also representative of floor 17 384 slab accelerations that would be recorded near the columns. Figure 11 uses the same format to 385 present the corresponding accelerations recorded at the middle of the floor slabs. The normalized 386 accelerations are equivalent to amplification factors of the PGA. The simulations are numbered 387 chronologically, where the first 10 correspond to the TP configuration, the next 10 to the hybrid 388 configuration, and the last three to the fixed-base configuration. The last data point in every 389 subplot is an average over all the simulations. 390 Figure 10 suggests that in general, column acceleration amplification factors were low and 391 insensitive to the height; that is, column acceleration did not significantly increase as the seismic 392 waves propagated from the base to the roof. This behavior was expected since the columns were 393 relatively rigid and could transfer the motion with little distortion. The average column 394 acceleration amplification factor was about 2 for the NW columns (Fig. 10(b)) and 2-3 for the SE 395 columns (Fig. 10(d)). Some exceptions to these general observations were noted; for instance, 396 during a few motions (GM #3 and 4 for both columns, and 7 and 10 for the NW column), the 397 acceleration amplification was notably higher. These instances of amplification, which were 398 limited to the TP configuration, arose due to local uplift/impact excursions that caused high 399 accelerations to propagate through individual columns. Consequently, amplification factors were 400 larger and increased significantly over the height. 401 The trends for floor slab acceleration amplification factors (relative to the shake table 402 acceleration) were quite different for the NW and SE floor slabs (Fig. 11). The average slab 403 amplification factor for the NW slabs was about 3 and did not increase much over the height of 404 the building (similar to what was observed in the NW column). The amplification factors for the 405 SE slabs, on the other hand, were notably higher, and increased steadily over the height of the 406 building (Fig. 11(d)). The average amplification factor varied from about 3 on the 2nd floor to 18 407 about 6 on the roof. As mentioned earlier, an amplification factor exceeding 6 was recorded in 408 one building in the Northridge Earthquake (Bozorgnia et al., 1998). The differences in response 409 of the NW and SE slabs appear to be related to period and damping trends identified earlier. For 410 instance, the NW slab dynamic properties – lower period and larger damping ratios that did not 411 vary much with height – were more favorable to vibration suppression, while the SE slab 412 dynamic properties – period increasing with height and damping ratio decreasing with height – 413 closely corresponded to the intensification of acceleration amplification factors in the upper 414 floors of the building. 415 To provide greater insight into the differences between the two slab locations, the spectral 416 accelerations computed from accelerations recorded in the SE and NW slabs at the 2nd through 417 roof level (with reference to the spectral acceleration for the recorded table acceleration) are 418 plotted for 3 representative input accelerations in Fig. 12: VOG175 and RRS88 that were 419 mentioned earlier, and 80% of 1978 Tabas at Tabas Station (TAB80). For VOG175 (GM #16) 420 and TAB80 (GM #9), the shake table spectral acceleration was maximized between 0.1 and 0.2 421 sec, a trend that was reflected in the majority of the ground motions used in the test program. As 422 noted earlier, the natural vibration periods of the NW slabs were consistently near 0.07 sec, while 423 the periods of the SE slabs increased from 0.07 sec at the 2nd floor to 0.13 sec at the roof. Thus, 424 the SE slabs at upper floors, with periods more aligned to the maximum spectral content of the 425 earthquakes, were generally more vulnerable to the ground input. This is reflected in Fig. 12, 426 where the spectral peaks for the SE slabs were generally shifted to the right (Fig. 12(a)-(b)) 427 compared to the NW slabs for the same motions (Fig. 12(d)-(e)), especially at the roof level. The 428 shake table spectral acceleration for RRS88 (GM #3) was approximately constant in the period 429 range of 0.08 to 0.2 sec. This uniformly large intensity shaking produced similar intensity 19 430 spectral peaks for many of the floors – regardless of vibration period (Fig. 12(c), 12(f)) – and 431 likewise uniform peak acceleration amplification factors ranging from 4-6 in the SE slab and 2-4 432 in the NW slab (GM #3 in Fig. 11(b), 11(d)). In summary, slab acceleration amplification factors 433 were somewhat dependent on the frequency content of the ground acceleration, where flexible 434 slabs with vibration periods exceeding 0.1 sec were more vulnerable than stiffer slabs. 435 Influence of Isolation System on Vertical Amplification Factors 436 For low intensity vertical table accelerations, column acceleration amplification factors were 437 generally insensitive to the presence of the isolation system (Fig. 10). This is consistent with 438 Lew and Hudson (1999), who identified three examples where the vertical accelerations recorded 439 below and at columns above the isolation system were essentially the same. However, during 440 high intensity vertical accelerations, column acceleration amplification factors were affected by 441 the presence of the isolation system. The effect is best illustrated by comparing the column 442 accelerations recorded during RRS88 in each of the three configurations. The RRS88 table 443 acceleration was characterized by a large near fault fling pulse that appeared in the vertical 444 direction at about 8.6 s, producing peak vertical table accelerations between 1.06 and 1.26g (Fig. 445 13(a)). This vertical motion was transmitted differently in the three configurations, as evidenced 446 by accelerations recorded in the 5th story column near the roof level (Fig. 13(b)). For instance, 447 the table acceleration was transmitted directly to the 5th story column with little amplification for 448 the fixed-base configuration. On the other hand, the table acceleration was amplified by nearly a 449 factor of 7 in the 5th story columns for the TP configuration. During this simulation, every 450 bearing uplifted, causing high frequency acceleration spikes from the bearing impact to 451 propagate through the building. Somewhat lower amplitude high frequency spikes were also 20 452 observed in the column accelerations for the hybrid configuration, suggesting some looseness in 453 the rails before the CL bearings engaged in tension. 454 The slab acceleration amplification factors, on the other hand, were not much affected by the 455 presence of the isolation systems. Figure 13(c) shows that during RRS88, the peak accelerations 456 recorded in the middle of the roof slab – where much of the damage was observed - were 457 approximately the same for each building configuration (variation was less than 20%). In 458 addition, the high frequency signals appearing in the column accelerations were not transmitted 459 to the slabs. Rather, the slab accelerations are characterized by a relatively lower frequency 460 vibration at the slab fundamental frequency, which is approximately 7-8 Hz according to Table 461 4. Thus, the slab amplification factors were basically unaffected by system configuration; that is, 462 normalized slab accelerations were approximately the same for the TP configuration (GM #1- 463 10), the hybrid configuration (GM #11-20), and the fixed-base configuration (GM #21-23). Since 464 the CPP component damage was closely related to the accelerations recorded in the center of the 465 floor slabs (demonstrated in the next section), the vertically rigid base isolation systems 466 considered in this experiment neither helped nor hurt the system response in the vertical 467 direction. This observation is consistent with the conclusions 468 Ceiling-Partition-Piping Component Damage vs Structural Response 469 While seismic design engineers have generally assumed that nonstructural components will 470 be protected from damage in an isolated building, these tests (similar to Furukawa et al., 2013a) 471 showed that vertical acceleration need also be considered. Thus, an understanding of the target 472 horizontal and vertical acceleration demands that induce nonstructural component damage, and 473 hence the demands that should be targeted to prevent such damage, is needed. The earthquake 21 474 simulations from this test series, which induced a varied combination of horizontal floor 475 accelerations and slab accelerations, produced useful data in this regard. 476 We qualitatively evaluated the damage to CPP components by inspection of all available 477 video footage, and correlated the observed damage states to recorded peak demand parameters 478 for every earthquake simulation. The damage mechanisms and the extent of damage were very 479 similar for the three system configurations because, as discussed previously, similar peak 480 demands were observed in the three system configurations. Thus, the damage observations are 481 presented and discussed without further mention of the system configuration during which they 482 were observed. Note that, similar to other related studies (e.g. Ryu and Reinhorn, 2012), CPP 483 damage was not inspected or repaired after every earthquake simulation. Generally, repairs were 484 applied to the suspended ceiling and sprinkler piping at the end of each test day, but these 485 components were not restored to their original strength after RRS88 in the TP configuration (3rd 486 3D and 5th overall of 41 simulations), which inflicted significant damage. Furthermore, damage 487 to partition walls was not considered in the evaluation, for the following reasons. Partition wall 488 damage states are generally drift-sensitive, but the typical drift sensitive damage states were not 489 observed. New vertical induced damage states were observed, which are described in Soroushian 490 et al. (2013a), but these damage states could not be directly observed from the videos, and thus 491 could not be associated with a particular earthquake simulation. 492 For the purpose of damage evaluation, we developed three general damage ratings, each 493 classified by several behaviors that occurred alone or in combination. The behaviors associated 494 with the damage ratings are described in Table 5. Ceiling panel equivalent fallen areas were 495 based on Gilani et al. (2013), wherein the total ceiling area suspended from the 5th and roof 496 levels was about 83.6 m2 (900 sf) each. Partially dislodged ceiling panels were not considered 22 497 part of the equivalent fallen area. Soroushian et al. (2013b) defined piping damage to be 498 extensive when 15% of pipe hangers fail - equivalent to 1 pipe hanger in this experiment. 499 Likewise, Soroushian et al. (2013b) rated damage as extensive when the permanent rotation 500 across a grooved or threaded connection exceeds 2 to 4 degrees. Rotations were not measured in 501 these experiments, but since 2 to 4 degrees is very small, any visible permanent rotation was 502 considered to indicate permanent damage. Finally, note that the assigned damage rating was 503 generally determined by the most severe rating when the observed behaviors overlapped multiple 504 ratings. 505 Based on video inspection, damage to the ceiling-sprinkler piping systems suspended from 506 the 5th and roof floors was evaluated independently and assigned a damage rating from Table 5 507 for each earthquake simulation. All simulations that applied XY (horizontal only) input 508 acceleration to the building in an isolation configuration were excluded since these simulations 509 never induced any damage. The results of this inspection and ranking are plotted in Fig. 14; 510 distinct markers for each damage rating are plotted against peak horizontal floor acceleration (X- 511 axis) and peak slab acceleration recorded at the center of the NE or SE quadrant (Y-axis). As 512 shown in Figs. 10 and 11, the vertical accelerations at the slab centers were substantially larger 513 than those at the corners, and thus the vertical acceleration was not uniform throughout the floor. 514 However, the accelerations at the slab centers reasonably represented the observed damage in 515 these experiments since, for example, ceiling panels tended to become dislodged in large 516 concentrations near the slab centers (see Soroushian et al. 2013a for further information). 517 Figure 14 illustrates that both Slight and Moderate damage ratings occurred for horizontal 518 accelerations ranging from about 0.4g–1.25g (variation by a factor of 3); but for vertical 519 accelerations ranging from about 2g–4g (variation by a factor of 2). Furthermore, individual 23 520 damage rating data points overlapped the intensity measures in the horizontal direction more 521 than the vertical direction. These observations suggest that the damage ratings were more closely 522 correlated to the measured vertical accelerations than horizontal accelerations, because the 523 horizontal floor accelerations were constrained to relatively low levels, which is a typical 524 objective of base isolation. If comparable intensity accelerations had been measured in the 525 horizontal direction (e.g. 5g and above), damage to nonstructural components would surely have 526 been pervasive; however, such intensities are generally unlikely to be realized in the horizontal 527 direction due to yielding in the structural system. 528 Damage to the CPP components generally initiated at vertical accelerations in the range of 2- 529 3g. As a notable exception, moderate damage to the ceiling system was observed in the fixed- 530 base configuration during the Iwanuma (XY) simulation, which generated peak horizontal floor 531 accelerations of about 0.9g on the 5th floor and 1.15g on the roof with vertical accelerations very 532 near zero (Fig. 14). The Iwanuma motion, which was recorded during the 2011 off the Pacific 533 coast of Tohoku Earthquake, applied continuous strong shaking for several minutes that caused 534 eventual fatigue to the ceiling system. Aside from this outlier, the vertical acceleration ranges 535 associated with the initiation of the discrete damage ratings overlapped slightly. In summary, for 536 the system considered in this test, Slight damage initiated at vertical accelerations between 2-3g, 537 Moderate damage at vertical accelerations between 3-5g, and Extensive damage at vertical 538 accelerations above 5g. Therefore, the data preliminarily suggests that vertical slab accelerations 539 should be limited to 2g for continued post-earthquake operation of typical seismically detailed 540 suspended ceiling and piping systems. This conclusion should be validated with further data. 541 Significant differences in the response of the ceiling systems suspended from the 5th and roof 542 floors were observed. In particular, very few ceiling panels fell from the 5th floor and the 24 543 equivalent area of fallen panels rarely reached even 5% (for Slight damage) despite recording 544 similar intensity accelerations at the 5th and roof levels. The technical explanation for the 545 discrepancy in the ceiling response over the two floors is given in Soroushian et al. (2013a). 546 However, Moderate and Extensive damage ratings at the 5th floor were assigned based on 547 sprinkler head-ceiling panel pounding interaction, damage to the perimeter seismic clips/wall 548 molding, and damage to the piping system such as permanent rotations. 549 Discussion 550 Since CPP damage in this experiment has been directly associated with vertical slab 551 vibration, we examine whether this slab vibration was representative of realistic floor systems. 552 The slab acceleration amplification factors averaged 5 or 6 at locations directly above the 553 concentration of CPP damage, compared to 2 or 3 elsewhere in the structure (Fig. 11(b), 11(d)). 554 The slab vibration periods lengthened in the areas with greater amplification factors, wherein the 555 period lengthening was justified, and the vibration periods were within the range of normal. 556 However, as discussed, the data related to effective damping in the experiment was inconclusive. 557 Thus, further evidence is needed to conclude that the trends for slab damping ratios and related 558 acceleration amplification factors observed in the experiment was representative of realistic 559 systems. 560 With regard to current design, the current ICC-AC156 code (ICC, 2010) can be interpreted as 561 follows: nonstructural components are designed/qualified based on a vertical spectral 562 acceleration equal to 2/3 of the horizontal spectral acceleration. Flexible components are 563 subjected to a component amplification factor ap that represents the dynamic amplification of the 564 component relative to the attached structure. Since z/h is taken to be 0 in the vertical direction, 565 nothing in the code accounts for the amplification of the vertical acceleration from the ground 25 566 (table) to the middle of the floor slabs by a factor of 3 to 6, as was observed in the flexible-side 567 floor slabs in this experiment. However, selection of alternative levels of amplification for design 568 is difficult to justify until the factors that contribute to such amplification (slab vibration periods, 569 modal properties, and effective damping as a function of vibration intensities) are better 570 understood. 571 Conclusions 572 573 The major findings of this experiment are summarized below: 574 575 Vertical ground acceleration can be a significant source of damage to integrated ceilingpartition wall-piping (CPP) components. During seismic input, the floor slab vibrations were dominated by single mode response, 576 with periods ranging from 0.07 to 0.13 sec. For many of the slabs, the slab spectral 577 acceleration was accurately estimated using a calibrated equivalent single degree-of- 578 freedom system. Because the slab response was determined by its vibration properties, 579 the associated CPP performance was insensitive to whether the building was base- 580 isolated. 581 The peak vertical acceleration was amplified by an average factor ranging from 3 to 6 582 from the table to the middle of the floor slabs. Current design/qualification procedures for 583 nonstructural components do not account for any such amplification, and are thus 584 unconservative. 585 The amplification factor increased as the slab vibration period increased. Slab vibration 586 periods were within normal ranges found by previous studies. The amplification factor 587 was also correlated to effective damping in the slab, but the interpretation of damping 588 was difficult and requires further validation. 26 589 From qualitative evaluation of available video footage, damage ratings for suspended 590 ceiling and piping systems were assigned for each 3D earthquake simulation in this 591 experiment. Because the horizontal floor accelerations were constrained to relatively low 592 levels by base isolation, the damage ratings were more closely correlated to vertical slab 593 acceleration than horizontal floor acceleration. CPP component damage initiated at slab 594 accelerations of about 2g. Damage was minimal for accelerations from 2-3g, moderate 595 for accelerations from 3-5g, and extensive for accelerations exceeding 5g. These metrics 596 establish vertical acceleration targets for achieving desired performance objectives. 597 Acknowledgements 598 This material is based upon work supported by the National Science Foundation under 599 Grants No. CMMI-1113275 and CMMI-0721399 and the National Institute for Earth Science 600 and Disaster Prevention (NIED) in Japan. Any opinions, findings, conclusions or 601 recommendations expressed in this document are those of the authors and do not necessarily 602 reflect the views of the sponsors. The authors recognize and thank the following companies for 603 providing product donations and technical support: Earthquake Protection Systems, Dynamic 604 Isolation Systems, Aseismic Design Company, USG Building systems, Victaulic, Tolco, Hilti, 605 Allan Automatic Sprinkler and CEMCO Steel. 606 References 607 Al-Hussaini, T.M., Constantinou, M.C., Zayas, V.A. (1994). “Seismic isolation of multi-story frame 608 structures using spherical sliding isolation system”, Technical Report NCEER-94-0007, National 609 Center for Earthquake Engineering Research, State University of New York at Buffalo, Buffalo, NY, 610 USA, 1994. 27 611 612 Allen, D. E. and Pernica, G. (1998). “Control of floor vibration”, Construction Technology Update, No. 22, NRC Institute for Research and Construction, National Research Council Canada. 613 American Society for Testing and Materials (ASTM), (2011). E580/E580M-11ae1: Standard Practice for 614 Installation of Ceiling Suspension Systems for Acoustical Tile and Lay-in Panels in Areas Subject to 615 Earthquake Ground Motions. ASTM International, Volume 04.06. 616 617 Badillo-Almaraz, H., Whittaker, A., Reinhorn, A. (2007). “Seismic fragility of suspended ceiling systems”, Earthquake Spectra, Earthquake Engineering Research Institute, 23(1):23-40. 618 Boice, M. D. (2003). “Study to improve the predicted response of floor systems due to walking”, M.S. 619 Thesis, Department of Civil Engineering, Virginia Polytechnic Institute and State University, 620 Blacksburg, VA. 621 Bozorgnia, Y. and Campbell, K. W. (2004). “The vertical-to-horizontal response spectral ratio and 622 tentative procedures for developing simplified V/H and vertical design spectra”, Journal of 623 Earthquake Engineering, 8:175-207. 624 Bozorgnia, Y., Mahin, S. A., Brady, A. G. (1998). “Vertical response of twelve structures recorded during 625 the Northridge Earthquake”, Earthquake Spectra, Earthquake Engineering Research Institute, 626 14(3):411-432. 627 628 Bradley, B. A. and Cubrinovski, M. (2011). “Near-source strong ground motions observed in the 22 February 2011 Christchurch Earthquake”, Seismological Research Letters, 82(6):853-865. 629 Clark, P.W., Aiken, I.D., Kelly J.M. (1997). “Experimental studies of the ultimate behavior of seismically 630 isolated structures”, Report No. UCB/EERC-97/18, Earthquake Engineering Research Center, 631 University of California, Berkeley, 1997. 632 633 Dao, N. D. (2012). Seismic response of a full-scale steel frame building isolated with triple pendulum bearings under 3D excitation, PhD Dissertation, University of Nevada, Reno. 28 634 Dao, N. D. and Ryan, K. L. (2013). “Computational simulation of a full-scale fixed-base and isolated- 635 base steel moment frame building tested at E-Defense”, J. Struct. Eng., Preview Manuscript, doi: 636 10.1061/(ASCE)ST.1943-541X.0000922. 637 Earthquake Engineering Research Institute (EERI). (1996). “10. Miscellaneous Building Types”, in 638 Northridge earthquake of January 17, 1994, Reconnaissance Report, Special Issue of Earthquake 639 Spectra, 12(S1):229-278. 640 641 Federal Emergency Management Agency (FEMA), (2009). “NEHRP recommended seismic provisions for new buildings and other structures”, FEMA P-750. Building Seismic Safety Council. 642 Fenz, D.M., Constantinou, M.C. (2008). “Development, implementation, and verification of dynamic 643 analysis models for multi-spherical sliding bearings”, Technical Report MCEER-08-0018, 644 Multidisciplinary Center for Earthquake Engineering Research, State University of New York at 645 Buffalo, Buffalo, NY, USA, 2008. 646 Furukawa, S., Sato, E., Shi, Y., Becker, T., and Nakashima, M. (2013a). “Full-scale shaking table test of a 647 base-isolated medical facility subjected to vertical motions”, Earthquake Engineering and Structural 648 Dynamics, 42:1931-1949. 649 Furukawa, S., Sasaki, T., Sato, E., Okazaki, T., Ryan, K. L. (2013b). “Comparison of vertical dynamic 650 response characteristics of two base-isolated buildings based on full-scale shaking table test”, Proc., 651 13th World Conference on Seismic Isolation, Energy Dissipation and Active Vibration Control of 652 Structures, Japan Society for Seismic Isolation and Anti-Seismic Systems International Society, 653 Sendai, Japan. 654 Gavin, H. P., Nigbor, R. L. (2012). “Performance of the base-isolated Christchurch Women’s Hospital in 655 the Sep. 4 2010 Darfield Earthquake and the Feb. 22 2011 Christchurch Earthquake”, 20th Analysis 656 and Computation Specialty Conference, ASCE Structures Congress, Chicago, IL. 29 657 Gilani, A. S. J., Reinhorn, A. M., Glasgow, B., Lavan, O., Miyamoto, H. K. (2010). “Earthquake 658 Simulator Testing and Seismic Evaluation of Suspended Ceilings”, Journal of Architectural 659 Engineering, ASCE, 16(2):63-73. 660 Gilani, A. S. J., Takhirov, S., and Tedesco, L., (2013). “Seismic Evaluation of Suspended Ceiling 661 Systems Using Static and Dynamic Procedures”, Proceeding of 44th Structures Congress, ASCE/SEI, 662 Pittsburg, PA. 663 664 665 666 Hicks, S. (2004). “Vibration characteristics of steel-concrete composite floor systems”, Progress in Structural Engineering and Materials, 6:21-38. ICC Evaluation Service (2010). AC 156: Acceptance Criteria for Seismic Certification by Shake Table Testing of Nonstrucutral Components. 667 Kasai, K., Mita, A., Kitamura, H., Matsuda, K., Morgan, T. A., Taylor, A. W. (2013). “Performance of 668 seismic protection technologies during the 2011 Tohoku-Oki Earthquake”, Earthquake Spectra, 669 29(S1):S265-S293. 670 Kelly, J.M., Skinner, M.S., Beucke, K.E. (1980). “Experimental testing of an energy absorbing seismic 671 isolation system”, Report No. UCB/EERC-80/35, Earthquake Engineering Research Center, 672 University of California, Berkeley. 673 674 Lew, M., Hudson, M. B. (1999). “The effects of vertical ground motion on base-isolated building systems”, Earthquake Spectra, 15(2):371-375. 675 Mayes, R. L., Brown, A. G., Pietra, D. (2012). “Using seismic isolation and energy dissipation to create 676 earthquake-resilient buildings”, Paper No. 093, Proc., 2012 NZSEE Conference, Christchurch, NZ. 677 Miranda, E. (2003). ”Building Specific Loss Estimation for Performance Based Design”, Pacific 678 679 680 Conference on Earthquake Engineering. University of Canterbury, Christchurch, New Zealand. Moroni, M. O., Sarrazin, M., Soto, P. (2012). “Behavior of instrumented base-isolated structures during the 27 February 2010 Chile Earthquake”, Earthquake Spectra, 28(S1):S407-S424. 30 681 682 683 684 Murray, T. M., Allen, D. E., Ungar, E. E. (1997). “Floor Vibrations due to Human Activity”, Steel Design Guide Series 11, AISC. National Fire Protection Association (NFPA), (2011). NFPA 13: Standard for the Installation of Sprinkler Systems." National Fire Protection Association, 2010 Edition, Quincy, MA. 685 Retamales, R., Davies, R., Mosqueda, G., Filiatrault, A. (2012). “Experimental Seismic Fragility of Cold- 686 Formed Steel Framed Gypsum Partition Walls”, Journal of Structural Engineering, ASCE, 139, 687 Special Issue: NEES 2: Advances in Earthquake Engineering, 1285-1293. 688 Ryan, K. L., Coria, C. B., Dao, N.D. (2013a). “Large scale earthquake simulation of a hybrid lead rubber 689 isolation system designed with consideration of nuclear seismicity”, CCEER Report No. 13-09. 690 Center for Civil Engineering Research, University of Nevada, Reno. 691 Ryan K, Sato E, Sasaki T, Okazaki T, Guzman J, Dao N, Soroushian S, Coria C (2013b). "Full Scale 5- 692 story Building with Triple Pendulum Bearings at E-Defense", Network for Earthquake Engineering 693 Simulation (database), Dataset, DOI:10.4231/D3X34MR7R. 694 Ryan K, Sato E, Sasaki T, Okazaki T, Guzman J, Dao N, Soroushian S, Coria C (2013c). "Full Scale 5- 695 story Building with LRB/CLB Isolation System at E-Defense", Network for Earthquake Engineering 696 Simulation (database), Dataset, DOI:10.4231/D3SB3WZ43. 697 Ryan K, Sato E, Sasaki T, Okazaki T, Guzman J, Dao N, Soroushian S, Coria C (2013d). "Full Scale 5- 698 story Building in Fixed-Base Condition at E-Defense", Network for Earthquake Engineering 699 Simulation (database), Dataset, DOI:10.4231/D3NP1WJ3P. 700 Ryu, K.P., Reinhorn, A.M. and Filiatrault, A. (2012). “Full Scale Dynamic Testing of Large Area 701 Suspended Ceiling System”, Proc., 15th World Conference on Earthquake Engineering, Lisbon, 702 Portugal. 703 Saito, T. (2006). “Observed response of seismically isolated buildings”, In Response Control and Seismic 704 Isolation of Buildings, Eds. Higashino, M. and Okamoto, S., Taylor and Francis, New York, NY. 31 705 Sato, E., Furukawa, S., Kakehi, A., Nakashima, M. (2011). “Full-scale shaking table test for examination 706 of safety and functionality of base-isolated medical facilities”, Earthquake Engineering and 707 Structural Dynamics, 40(13):1435–1453. 708 Soroushian S., Maragakis, E. M., Ryan, K. L., Sato, E., Sasaki, T., Okazaki, T., Mosqueda, G. (2013a). 709 “Seismic simulation of integrated nonstructural systems at E-Defense, Part 2: Evaluation of 710 nonstructural damage and fragilities”, Under Review in J. Struct Eng. 711 Soroushian, S., Zaghi, A. E., Maragakis, E. M., Echevarria, A., Tian, Y., Filiatrault, A. (2013b). 712 “Analytical Seismic Fragility Analyses of Fire Sprinkler Piping Systems with Threaded Joints”, 713 Earthquake Spectra, In Press, doi: 0-dx.doi.org.innopac.library.unr.edu/10.1193/083112EQS277M. 714 Taylor, A. W. and Igusa, T. (2004). Primer on Seismic Isolation, ASCE Task Committee on Seismic 715 Isolation, American Society of Civil Engineers, Reston, VA. 716 Tian, Y., Filiatrault, A., Mosqueda, G., 2013. “Experimental Seismic Study of Pressurized Fire Sprinkler 717 Piping Subsystems”, Technical Report MCEER-13-0001, Multidisciplinary Center for Earthquake 718 Engineering Research, State University of New York at Buffalo, Buffalo, NY, USA, 2013. 719 Wolff, E.D., Constantinou, M.C. (2004). “Experimental study of seismic isolation systems with emphasis 720 on secondary system response and verification of accuracy of dynamic response history analysis 721 methods”, Technical Report MCEER-04-0001, Multidisciplinary Center for Earthquake Engineering 722 Research, State University of New York at Buffalo, Buffalo, NY, USA, 2004. 723 Zaghi, A. E., Maragakis, E. M., Itani, A., and Goodwin, E. (2012). “Experimental and analytical studies 724 of hospital piping subassemblies subjected to seismic loading”, Earthquake Spectra, Earthquake 725 Engineering Research Institute. 28(1):367-384. 32 726 727 728 729 Tables Table 1. Subset of Ground Motions Selected for Testing, along with Scale Factors, Target Peak Acceleration (g), and Realized to Target Acceleration Ratios Reason for Earthquake Station Selection 1979 Imperial Westmorland U.S. Service Valley, U.S. 1994 Northridge, U.S. Rinaldi Receiving Station (RRS) 1989 Loma Prieta, U.S. Theoretical System Test Scale Target PGA Scale Factor Config. Factor (X/Y, Z) g (X/Y, Z) 0.8 TP, Hybrid, 0.8, 0.8 0.17, 0.17 Fixed Recorded to Target PGA Ratio (X/Y, Z) 1.14, 0.83 U.S. Design 0.88 TP, Hybrid TP, Hybrid Fixed Fixed Fixed 0.88, 0 0.88, 0.88 0.35, 0 0.35, 0.35 0.35, 0.88 0.73, 0 0.73, 0.72 0.29, 0 0.29, 0.35 0.29, 0.72 1.63, NA 1.66, 1.75 1.37, NA 1.40, 1.21 1.41, 1.48 Los Gatos Pres. Ctr U.S. MCE 1.09 TP 0.7, 0.7 0.42, 0.64 1.07, 1.07 Sylmar Hospital U.S. MCE 1.22 TP 1.0, 1.0 0.87, 0.52 1.32, 1.05 U.S. MCE 1.03 TP TP TP TP 0.5, 0.5 0.8, 0.8 0.9, 0 1.0, 0 0.45, 0.33 0.72, 0.52 0.81, 0 0.90, 0 1.30, 1.09 1.21, 1.13 1.15, NA 1.15, NA 1999 Chi-Chi, TCU065 Taiwan U.S. MCE 0.89 TP TP TP 0.5, 0 0.7, 0 0.8, 0 0.41, 0 0.57, 0 0.65, 0 1.11, NA 1.13, NA 1.14, NA 1940 Imperial El Centro Valley, U.S. 1995 Kobe, JMA Japan 1995 Kobe, Takatori Japan Japan design motion Japan design motion Japan nearfault NA TP, Hybrid 1.3, 1.3 0.28, 0.26 1.08, 1.07 NA TP 1.0, 1.0 0.60, 0.34 1.14, 1.20 NA TP TP 1.0, 1.0 1.15, 1.0 0.75, 0.29 0.86, 0.29 1.06, 0.90 1.09, 0.97 2011 Tohoku, Iwanuma Japan Long duration subduction NA TP, Hybrid 1.0, 0 Fixed 0.7, 0 0.42, 0 0.29, 0 1.41, NA 1.28, NA Synthetic Vogtle Central/Eastern U.S. Extended Design Basis NA Hybrid Hybrid Hybrid Hybrid Hybrid 0.75, 0.75 1.0, 1.0 1.25, 1.25 1.5, 1.5 1.75, 1.75 0.33, 0.22 0.44, 0.29 0.55, 0.36 0.66, 0.43 0.77, 0.50 1.19, 1.00 1.19, 1.04 1.25, 1.03 1.30, 1.02 1.34, 0.98 Synthetic Diablo Canyon Western U.S. Extended Design Basis NA Hybrid Hybrid 0.8, 0.8 0.95, 0 0.78, 0.46 0.93, 0 1.17, 0.99 1.20, NA 1994 Northridge, U.S. 1978 Tabas, Iran Tabas Sta. (TAB) 730 33 Table 2. Fixed-Base Building Natural Periods and Damping Ratios in the Horizontal Direction Mode 1 X Mode 2 X Mode 3 X Mode 1 Y Mode 2 Y Mode 3 Y White noise X Period Damping (s) ratio (%) 0.652 3.30 0.204 1.62 0.112 3.31 n/a n/a n/a n/a n/a n/a White noise Y Damping Period (s) ratio (%) n/a n/a n/a n/a n/a n/a 0.677 2.54 0.211 1.65 0.113 2.64 White noise 3D Damping Period (s) ratio (%) 0.677 4.09 0.205 1.95 0.112 3.74 0.686 3.49 0.212 1.93 0.113 3.61 731 732 733 Table 3. Median (λ) and Dispersion (β) of the Natural Periods and Damping Ratios of the Individual Slabs, Computed from the Transfer Functions of White Noise Motions SE Floor Level NE NW Period Damping Period Damping Period Damping (s) Ratio (%) (s) Ratio (%) (s) Ratio (%) Roof λ 0.13 β 0.00 λ 1.00 β 0.22 λ 0.13 β 0.00 λ 1.50 β 0.25 λ 0.08 β 0.00 λ 7.50 β 0.95 5th 0.10 0.00 2.00 0.00 0.10 0.00 3.25 0.34 0.08 0.00 4.50 0.63 4th 0.09 0.00 3.00 0.69 0.09 0.00 4.00 0.38 0.08 0.00 5.00 0.69 3rd 0.08 0.00 6.00 1.32 0.08 0.00 2.75 0.30 0.08 0.00 9.50 0.80 2nd 0.08 0.00 7.50 1.32 0.08 0.00 4.75 0.47 0.08 0.00 10.00 0.00 734 34 735 736 737 738 Table 5. Ceiling-Piping Damage Ratings and Associated Behaviors Damage Rating Slight Moderate Extensive Behavior Description Ceiling panels up to 5% equivalent area fall. Slight damage to panels at sprinklerheads due to pounding interaction (hole enlarged by 1 inch in any direction). Pipe hanger surge clips pops out. Ceiling panels between 5 and 20% equivalent area fall. More significant damage to panels at sprinkler heads due to pounding interaction (hole enlarged by 1-2 inches in any direction). 1 or 2 cross tees may fail, and a ceiling hanger wire may break. Damage to perimeter seismic clips and wall molding is visible. Ceiling panels exceeding 20% equivalent area fall. Large sections of the ceiling grid are compromised, due to buckling, misalignment or connection failure. Very significant damage to panels at sprinkler-heads due to pounding interaction (hole enlarged by more than 2 inches). 1 or more pipe hangers break or are permanently deformed. Permanent rotation of armover pipes is visible. 739 35 740 Figures 741 Figure 1. 5-story steel moment frame testbed set on triple pendulum bearings 742 Figure 2. (a) Plan and elevation view of structural framing, (b) supplemental concrete weight on 743 floors 2-5, and (c) supplemental steel weight on roof (dimensions in mm) 744 Figure 3. (a) TP bearing, (b) LR bearing, (c) CL bearing 745 Figure 4. Typical layout of accelerometers at the floor levels 746 Figure 5. 5% damped spectral acceleration for the vertical components of 3D motions compared 747 to design and MCE spectra based on NEHRP provisions 748 Figure 6. Representative slab transfer function for: (a) roof SE slab, (b) 5th NE Slab 749 Figure 7. (a) Depiction of single floor slab between columns, and (b) associated single-degree-of 750 freedom (SDOF) system idealization 751 Figure 8. Measured and idealized SDOF slab acceleration history and spectrum for (a) Roof SE 752 Slab for VOG 75 and (b) 3 SE Slab for RRS 88 753 Figure 9. Representative peak (vector sum) horizontal floor acceleration profile (acceleration vs. 754 level) for TP, hybrid, and fixed-base buildings for XY (horizontal only) and 3D input motions 755 Figure 10. (a) Peak acceleration in table and NW column, (b) normalized NW column 756 acceleration, (c) peak acceleration in table and SE column, (d) normalized SE column 757 acceleration; for each floor level and every 3D simulation in the NEES/E-Defense test 758 Figure 11. (a) Peak acceleration in table and NW floor slab, (b) normalized NW slab 759 acceleration, (c) peak acceleration in table and SE floor slab, (d) normalized SE slab 760 acceleration; for each floor level and every 3D simulation in the NEES/E-Defense test rd 36 761 Figure 12. 5% damped spectral acceleration at the ground through roof levels for (a) SE slab – 762 VOG 175, (b) SE Slab – TAB80, (c) SE Slab – RRS88, (d) NW Slab – VOG175, (e) NW Slab – 763 TAB80, (f) NW Slab – RRS88 764 Figure 13. Recorded vertical accelerations at (a) table, (b) NE column near roof, and (c) NE roof 765 slab in each building configuration during RRS88. 766 Figure 14. Damage rating for select XY and every 3D simulation plotted against peak vector 767 horizontal acceleration in X and peak vertical acceleration in Y for (a) 5th floor and (b) roof floor 37 Figure Click here to download high resolution image Figure Click here to download high resolution image Figure Click here to download high resolution image Figure Click here to download high resolution image Figure Click here to download high resolution image Figure Click here to download high resolution image Figure Click here to download high resolution image Figure Click here to download high resolution image Figure Click here to download high resolution image Figure Click here to download high resolution image Figure Click here to download high resolution image Figure Click here to download high resolution image Figure Click here to download high resolution image Figure Click here to download high resolution image Copyright Agreement Click here to download Copyright Agreement: Copyright Transfer - Paper 1.pdf Sizing worksheet (.xls) Click here to download Sizing worksheet (.xls): Sizing Worksheet_Paper1.xls Journals Sizing Worksheet ***Please complete this form for all new manuscripts*** October 26, 2014 This worksheet will automatically calculate the total number of printed pages your article will occupy in the journal. Please fill in all fields in green below. If you do not know your Manuscript Number, you may leave that field blank. Technical Paper/Case Study = 8 pgs. Technical Note = 3 pgs. Manuscript number: Journal name: Corresponding author name: Email address: Length Limits: Forum = 4 pgs. Discussion/Closure = 2 pgs. STENG-3259 Journal of Structural Engineering (ASCE) Keri Ryan [email protected] Information on the maximum allowed length for each article type can be found online at: http://www.asce.org/Content.aspx?id=29559 Number of pages in your manuscript: 31 Number of figure pages: 8 - Please include figure captions when indicating the size of your manuscript. Number of table pages: 3 - Manuscripts should use 12 pt. font, double spaced, with 1 inch margins. Estimated article pages: 10 Note: The total displayed above is only an estimate. Final page count will depend on a number of factors, including the size of your figures and tables, and the number of display equations in your manuscript. Additional author resources can be found online using the ASCE Author Guide located at: http://www.asce.org/Content.aspx?id=18107 = total ms pages = # of figs = # of tables 10 10 10 0 10 8 2.8 1 12 7 1.8 1.05 10 7 2.8 1.05 11 0 11 0 0 0 0 0 1 0 0 0 1 0 0 Response to Reviewers Comments Click here to download Response to Reviewers Comments: Response to Reviewers - Paper 1.pdf Manuscript STENG-3259: Seismic Simulation of Integrated Nonstructural Systems at E-Defense, Part 1: Influence of 3D Structural Response and Base Isolation Authors: Keri L Ryan, PhD; Siavash Soroushian; Emmanuel Manos Maragakis; Eiji Sato; Tomohiro Sasaki; Taichiro Okazaki The authors thank the reviewers for their comments, and have made a sincere effort to address each comment. We believe that the revisions generated by these comments have improved the quality of the paper. Each comment is addressed in turn. Reviewer #2: This paper discusses a full scale shake table study of a building equipped with an integrated ceiling, partition, and fire sprinkler piping system. The building is tested as 1) isolated with triple friction pendulum bearings, 2) isolated with a combination of lead plug and cross linear bearings, and 3) fixed base. The study focus heavily on the effect of the vertical excitation. The paper is well written and presents interesting and valuable experimental information from full scale shake table testing. Nonetheless, before the reviewer can recommend publication, various issues need to be addressed by the authors. Technical Comments 1. The term "Integrated Nonstructural Systems" in the title of the paper is too broad. There are innumerable types of nonstructural components in a building, but the present study focuses on a very specific one: an "an integrated suspended ceiling-partition wall-sprinkler piping system." The title should be revised to reflect exactly what is discussed in the paper. Similarly, the broad term "nonstructural components" in the abstract, body, and conclusions of the paper should be replaced with a more descriptive one. The reviewer is correct that to generalize the response of the integrated ceiling-partition-piping system to all nonstructural systems is misleading. Where used in the context of the specific system tested in this paper, we have replaced the term “nonstructural component” or “nonstructural system” with “ceilingpartition-piping” or the abbreviation CPP, which is defined in the last paragraph of the introduction. 2. The reviewer does recognize that this is a study with a large scope; however, the fact that this paper constantly refers to other papers by the authors for further test setup details and findings makes is very inconvenient and frustrating to read. At times, it seems that this paper is an overview that discusses the entire project thinly, rather than a research article that discusses a specific aspect of the study in depth. The reviewer urges the authors to take this into consideration towards a more focused and improved revised paper. The reviewer’s impression has been carefully considered. We hope the reviewer appreciates the challenge of both overviewing the experiment and discussing the specific aspect in depth within a reasonable length paper. Somehow, the request to add more general details while at the same time suggesting the paper reads like an overview rather than an in depth discussion was a little confusing. In the original submission, we attempted to include all details necessary to understand the study. It would have been helpful if the reviewer identified what details seemed to be missing. The reviewer has identified some issues in later comments, which we have addressed, and we hope will also improve the readability of the paper. Specifically, additional details have been added to describe the isolation systems (Lines 146 to 170) and the earthquake simulation plan (Lines 199 to 233) so that the description of the experiment may stand on its own. Additional discussion of the uplift observed in the triple pendulum system and its significance has been added (Lines 436 to 454). The discussion of horizontal-vertical coupling has been removed, as it is not essential to the hypothesis of the paper (see response to comment 11). Details of the testbed building configuration, structural system, and composite floor system provided in the original submission are believed to be sufficient. Details of the ceiling-partition-piping system are appropriately deferred to the companion paper, as the link to the 2nd paper is clear and our expectation is that most readers will consult the papers jointly. Generally, we left in place references directing the reader to other sources of information; these are provided as a courtesy and not considered a necessity to understand the study. 3. "The shaking intensities that can cause damage to the nonstructural components are typically lower than those that induce structural damage, and nonstructural components may be more sensitive than structural components to vertical shaking." Provide references. A reference has been added for the first part of the statement. Since we do not know of a good reference for the second part, we have modified the statement, so that it now reads “The shaking intensities that can cause damage to the nonstructural components are typically lower than those that induce structural damage (Miranda, 2003), and the influence of vertical shaking on nonstructural components has not been 4. Line 36. Unsubstantiated. Has this been shown elsewhere? or is it a conclusion of the present study? If the latter, then this sentence probably does not belong in the introduction. Line 36 states: “Nonstructural components in base-isolated buildings are subjected to a different proportion of horizontal to vertical accelerations than prescribed by seismic qualification tests, and so the observations from component tests referenced above may not strictly apply.” We believe this statement can be made in confidence based on the present state of knowledge, independently of the data later presented in the paper. To clarify our thinking, we have added the following “Specifically, AC 156 prescribes that the horizontal acceleration is amplified over the building height while horizontal acceleration in a base-isolated building is significantly reduced relative to the ground acceleration (ICC 2010)” (Lines 36 to 41). The vertical acceleration as prescribed by AC 156 should not be significantly altered by the presence of an isolation system. 5. Line 62 points to the work of Lew and Hudson (1999) that found that the vertical accelerations transmitted to the building from the ground were not affected by the presence of base isolation system. Compare and contrast the findings of the present study to that study. Not much more can be said about the Lew and Hudson study. Three different examples were identified – all characterized by small amplitude vertical shaking - suggesting that vertical accelerations recorded at the foundation and at the columns above the isolation units were essentially the same. This is consistent with our observations at low intensities, but not high intensities. However, our more important observation is that the presence of the isolation system did not affect the slab vibration and thus the nonstructural response. The buildings considered by Lew and Hudson did not have vertical accelerometers located anywhere on the floor slabs. The following statement has been added (Lines 430-435): “For low intensity vertical table accelerations, column acceleration amplification factors were generally insensitive to the presence of the isolation system (Figure 9). This is consistent with Lew and Hudson (1999), who identified three examples where the vertical accelerations recorded below and at columns above the isolation system were essentially the same. However, during high intensity vertical accelerations, column acceleration amplification factors were affected by the presence of the isolation system.” The statement appears at the beginning of the new section titled “Influence of Isolation System on Vertical Amplification Factors”. 6. Line 74. It is stated that a study by Furukawa et al. (2013a), also at E-Defense, and also full scale, found that despite the large vertical acceleration amplification observed at mid-slab, no damage to suspended ceilings, piping and sprinkler systems was observed. Clarify the novelty of the present study, in comparison to that by Furukawa et al. Also, offer an explanation as to why the study by Furukawa et al. observed no damage to these nonstructural systems, while the present study did. The study by Furukawa discussed a different building that incorporated Japanese-designed ceilingpartition wall-piping systems. Japanese design practice is very different from U.S. design practice and so the response of the two types of systems is not comparable. Furthermore, the response of the CPP system was not a focus of the study by Furukawa et al, which focused on anchored and unanchored medical equipment. No details about the CPP system is given; it is only mentioned in passing. In addition, as mentioned in the original submission, the building studied in Furukawa et al. is a reinforced concrete building, which has different floor vibration characteristics than a steel frame building. We feel that the study by Furukawa et al is quite clearly of different focus than our manuscript, and further discussion of the distinction or novelty of our study is not warranted. However, the above points are clarified in the revised statement “Disruption and damage to building contents and equipment (the main focus of the study) was significant, but damage to the Japanese-designed suspended ceilings, piping and sprinkler systems, and partition walls/sliding doors was not observed.” (Lines 79-82 of the revised manuscript). We believe that damage was not observed to CPP in the Furukawa et al study due to differences in the design of U.S. and Japanese CPP systems. However, without knowing details about the design of the Japanese style CPP systems, it is not appropriate to speculate. 7. Line 176. "Many observers of the test program felt that this level of vertical input acceleration was extreme or unrealistic." The reviewer is not convinced by the argument that follows this statement, i.e. discussion on the vertical PGA and spectral accelerations of two Christchurch motions. The seismic vulnerability of a system needs to be evaluated based on a number of motions for a specific hazard level and site. What is the rationale for the selection of the ground motions used in this study? For example, the Rinaldi Receiving Station motion from the 1994 Northridge earthquake is a near fault motion that is well known for its high destructive potential. Fig. 4 shows vertical spectral accelerations in the order of 3g over the 0.1-0.2 s period range; consequently, it would be a big surprise if this motion didn't result in extensive damage to the ceiling system, which has its fundamental period in this range. But how realistic is it to base conclusions regarding the seismic vulnerability of nonstructural components on an input motion like this? Is there a realistic design vertical spectral acceleration that the authors are working with to select/scale motions? The authors have yet to establish that the selected group of ground motions are realistic/representative since this heavily bears upon the validity of the conclusions drawn in the study. The authors may find the work of Bozorgnia and Campbell on v/h ratios and vertical response spectra relevant. The realism of the selected ground motions, especially the Rinald Rec. Station motion, has been a point of concern among project investigators and stakeholders since the completion of the tests. This concern was obviously manifested in some of the comments in the original submission, which perhaps came across as “defensive”. We have completely reworked the presentation of the vertical ground motions taking into account the reviewers’ comments. In the revised section, we have: provided a basis for the selection and scaling of ground motions in the study; on the basis of NERHP provisions (FEMA, 2009), developed vertical design and MCE spectra corresponding to the horizontal design and MCE spectra that were used for the ground motion selection; and showed how the vertical ground motions realized in the experiment compared to the vertical design and MCE spectra. Please see the response to comment 8 for discussion regarding drawing conclusions to unrealistic ground motions. 8. The selection of realistic ground motions is one thing; the ability of the shake table to accurately reproduce these is another. If the table does not track properly, the resulting motion on the table may be significantly larger (and, consequently, even more unrealistic) than the input. For example, what is the input vertical PGA of the RRS88 motion which resulted in measured vertical PTA ranging from 1.05 to 1.25g? or, how does the output vertical spectrum for RRS compare to the input one? Discuss the tracking performance of the E-Defense shake table and its ability to reproduce the selected ground motions. Realized to target peak acceleration ratios are provided in Table 1 (a new table) for completeness. However, the ability of the shake table to reproduce the input motions is not the subject of this paper, and should not be a criterion for evaluation of the technical merit of the paper. The question of whether the realized motions were realistic may be relevant if conclusions are based on responses observed during those motions. We feel that our revised presentation of the ground motion establishes that the realized vertical ground motions were realistic except for Rinaldi. Regarding the Rinaldi motion, we have provided the relevant facts and presented a neutral perspective, allowing the reader to draw his/her own conclusion. The remaining question is whether the conclusions in the paper depend on the potentially unrealistic Rinaldi motion. This is not the case, since the conclusions are drawn on the basis of observations over the range of horizontal and vertical input intensities, not on a single ground motion. For instance, damage to CPP systems was observed in several motions aside from the Rinaldi motion. The slab amplification factors were consistent across the range of shaking intensities. Furthermore, the damage thresholds and damage ratings were established at intensities well below the Rinaldi motion. 9. Lines 201-203. Was table rocking observed but ignored? Or, was table rocking not observed at all? “Kasai et al. (2011) observed that rocking of the shake table can affect the natural frequencies and mode shapes; however, the effect of rocking was ignored in the analysis presented here.” We believe the comment about table rocking is misleading and therefore have opted to remove it. To clarify, the paper by Kasai presented a new method to analyze the natural frequencies and mode shapes of the building that accounted for the rocking of the shape table. Kasai’s method was first demonstrated on the same testbed building examined in this paper, since the testbed was also used in an earlier experiment by Kasai and collaborators. However, we chose to present the experimental frequencies and modes of the building determined by traditional methods, because they more closely matched the frequencies/modes determined by eigenvalue analysis of the calibrated building computational model. To answer the reviewer’s question about rocking, the research team initially investigated table rocking in detail as a possible source of “unusual” structural response, but ultimately dismissed it as insignificant. Specifically, the recorded table rocking was input to the building computational model through multiple support excitation, in addition to the horizontal/vertical shaking. No difference was observed in the computational response of the building with and without table rocking input. 10. Discuss the two isolation types in a bit more detail, focusing specifically on the aspects of their design that effects the vertical response of the columns and slabs. For example, the TFP is practically rigid in compression but provides zero tensile resistance, which, as the authors noted, can result in uplift and pounding, and consequent acceleration spikes that propagate up the columns. How about the hybrid lead plug + cross linear bearing system? What are these cross linear bearings like? Do they provide any tensile resistance? If so, is their stiffness in tension equal to that in compression? How stiff are they in comparison to the lead plug bearings? When the lead plug bearings displace laterally, they also move downward; how do the cross linear bearing respond? Do they follow this motion, do they restrict it, etc.? Additional details have been provided about the isolation systems (Lines 146 to 170), including those that will clarify the vertical response characteristics. Furthermore, we have strengthened our justification that for these vertically stiff isolation systems, the vertical accelerations transmitted to the CPP system through the floor slabs was not affected by the presence of an isolation system – an issue that was raised in the review of the companion paper. Specifically, we have demonstrated that high frequency accelerations propagating through the columns were not transmitted to the floor slabs (lines 448-454 and new Figure 13). With regard to the compliance between the LR bearings and the CL bearings, this is a complex issue that we believe was inconsequential to the overall vertical response of the isolation system. Therefore, we choose not to address it specifically in the paper. 11. Lines 310-312. "These sources of coupling have been investigated (Dao, 2012; Ryan 310 et al., 2013a); they are not strictly an artifact of the test setup and they can be predicted by numerical simulation (Dao and Ryan, 2013)". Could these sources of coupling be due to irregularities in the distribution of mass and/or stiffness of this specific design? To the best of the reviewer's knowledge, other authors have not noted such a coupling before. Upon further consideration, we have decided to remove any mention of horizontal-vertical coupling from this paper. While it is helpful to mention the horizontal-vertical coupling in the context of the overall big picture of the experiment, the hypothesis of this paper is that the damage to the integrated CPP system principally resulted from direct vertical shaking, and the coupling was of minor or no consequence to this outcome. Since the coupling cannot be described, demonstrated, and validated satisfactorily in this paper, it is better not to confuse the issue. Another paper focused on the horizontal-vertical coupling is currently in review. The description of the observed horizontal acceleration has been modified accordingly (Lines 353 to 370 of the revised manuscript). 12. It is noted that damage to the integrated ceiling system was not evaluated or repaired after each motion was run on the table, but rather at the end of each day. Furthermore, the system was never restored to its original strength after RRS was run (which was very early on, based on Figs 9 and 10). If so, then how can the authors know that it was a particular motion that caused damage, and it was not compounding damage over the course of several tests? After all, the authors do point out the case of the Iwanuma motion where fatigue and damage was caused to the ceiling system due the prolonged shaking—even if the intensity of the shaking was very low. The reviewer is correct that we cannot be sure that the damage states observed after each motion were not influenced by cumulative damage over several tests. However, because of the effort and expense involved in fabricating nonstructural systems for experimental testing, it is common to perform a series of shaking tests where the initial state of the system is not restored after each test (e.g. Ryu and Reinhorn, 2012). We feel that it is appropriate to report and interpret observations from sequential tests as long as the conditions are clearly stated, which they have been here. The passage that states these important conditions has been slightly reworded to state: “Note that, similar to other related studies (e.g. Ryu and Reinhorn, 2012), CPP damage was not inspected or repaired after every earthquake simulation. Generally, repairs were applied to the suspended ceiling and sprinkler piping at the end of each test day, but these components were not restored to their original strength after RRS88 in the TP configuration (3 rd 3D and 5th overall of 41 simulations), which inflicted significant damage.” Regarding the statement about the Iwanuma motion, the authors believe that fatigue was a factor in the Iwanuma motion, but the intensity was not low. The paper states: “The Iwanuma motion, which was recorded during the 2011 off the Pacific coast of Tohoku Earthquake, applied continuous strong shaking for several minutes that caused eventual fatigue to the ceiling system.” The duration of strong shaking in all of the other motions applied was very short by comparison. 13. Did the authors investigate other intensity measures, besides horizontal and vertical acceleration (Fig. 12), to see if those correlate better with damage to the ceiling system? It is not clear to the reviewer what the failure modes of the ceiling system are. The reviewer raises a valid question regarding the desire to find the best response measure to correlate with the ceiling system response. Other intensity measures that we might have considered are: peak spectral acceleration, spectral acceleration at the natural period of structure, peak floor velocity, etc. However, we have several reasons for restricting the consideration to peak horizontal and vertical accelerations recorded in the floors: 1) The component fragility functions that have been developed for performance-based design and for use in the Performance Assessment Calculation Tool (PACT) work with very general response measures, generally floor accelerations and story drifts, for simplicity. 2) We wanted a general response measure, and not a response measure like period specific spectral acceleration that would only be valid for this specific system. 3) We were trying to show, in a simple way, that the CPP damage is more closely correlated to vertical shaking than to horizontal shaking. The failure modes of the ceiling systems (damage states) are clearly stated in Table 4. 14. The establishment of damage states (Line 459) based on a very limited sample appears somewhat arbitrary and unreliable. Please comment. Are there any other studies with data that could be used to enhance confidence in the set damage state limits? The reviewer is correct that it is premature to present conclusive damage states based on the limited data considered in this paper. The text starting from Line 459 (now 528) has been reworded as follows: “In summary, for the system considered in this test, Slight damage initiated at vertical accelerations between 2-3g, Moderate damage at vertical accelerations between 3-5g, and Extensive damage at vertical accelerations above 5g. Therefore, the data preliminarily suggests that vertical slab accelerations should be limited to 2g for continued post-earthquake operation of typical seismically detailed suspended ceiling and piping systems. This conclusion should be validated with further data.” Unfortunately, we do not think that other existing data sets can validate the vertically-induced damage states at this time. The data set discussed in this paper appears to be unique with respect to the intensity of vertical accelerations that were observed and/or the independence of horizontal and vertical accelerations. For instance, for the data set discussed in Ryu et al. (2013), the specimen reached maximum frame vertical accelerations of only 1.54g at mid-bay of the roof level. The data set discussed in Gilani et al. (2010) did include larger vertical accelerations (e.g. nearly 6g recorded in the frame) but they were accompanied by large horizontal accelerations, since horizontal and vertical accelerations were scaled up proportionally (i.e. data points that would appear as a straight line in Figure 12 (now Figure 14)). 15. Paragraph starting on Line 482. "rigid components are designed/qualified for a vertical acceleration equal to the vertical PGA, and flexible components are designed/qualified for a vertical acceleration equal to 2.5 times the vertical PGA." The authors are urged to re-examine carefully Section 6.5.1 of AC156 and revise this paragraph accordingly. We hope that we have correctly interpreted the reviewer’s comment. There is no direct reference to vertical PGA in AC156, but rather the horizontal spectral acceleration is the basis for deriving both horizontal and vertical input motion. The statement has been revised to “nonstructural components are designed/qualified based on a vertical spectral acceleration equal to 2/3 of the horizontal spectral acceleration. Flexible components are subjected to a component amplification factor ap that represents the dynamic amplification of the component relative to the attached structure. Since z/h is taken to be 0 in the vertical direction, nothing in the code accounts for the amplification of the vertical acceleration from the ground (table) to the middle of the floor slabs by a factor of 3 to 6, as was observed in the flexibleside floor slabs in this experiment.” (now lines 553-559) 16. Paragraph starting on Line 482. "Since 2.5 is a spectral amplification factor, nothing in the code accounts for the amplification of the vertical acceleration from the ground (table) to the middle of the floor slabs by a factor of 3 to 6, as was observed in the experiment. Thus, the current design/qualification procedures for nonstructural components appear to be unconservative." As this conclusion leans on the ground motions used in this study, the reviewer urges the authors to exercise caution. We have removed the very strong statement “The current design/qualification procedures for nonstructural components appear to be unconservative.” However, we wish to preserve the comment that current codes do not account for amplification of vertical acceleration in the floor slabs. The amplification factors of 3 to 6 that were observed on the flexible side of the building were not limited to one or two motions of strong intensity; rather they were observed consistently over the diverse set of ground motions with regard to horizontal and vertical intensity, duration, and frequency characteristics. 17. Table 4. Why were the damage states defined as such? Is there a monetary correlation to these damage states? We believe the rationale for the selection of the damage states is clearly described in the original submission. “Ceiling panel equivalent fallen areas were based on Gilani et al. (2013), wherein the total ceiling area suspended from the 5th and roof levels was about 83.6 m2 (900 sf) each. Partially dislodged ceiling panels were not considered part of the equivalent fallen area. Soroushian et al. (2013b) defined piping damage to be extensive when 15% of pipe hangers fail - equivalent to 1 pipe hanger in this experiment. Likewise, Soroushian et al. (2013b) rated damage as extensive when the permanent rotation across a grooved or threaded connection exceeds 2 to 4 degrees.” Basically, the damage states for our CPP system are based on precedents established for individual components in previous experiments. We chose not to use the commercial damage states in the PACT software, because they were not wellcorrelated to the range of observed damage in the experiment. For instance, the PACT extensive DS is outside of the range of what we observed. In addition, there is a precedent to fine tune the damage states to fit the patterns observed in the experiment. (e.g. Gilani et al. (2013) defined the transition to extensive ceiling damage at 20% equivalent area of fallen ceiling panels, which matches well with their data). 18. Figure 2. Check the dimensions in the E-W direction. They don't add up. The correction to the figure has been made. Editorial Comments Line 229. "The SE and NE roof slab quadrants were more flexible due to the substantial added mass on the east side of the roof level." Is it the flexibility or the period that the added mass increases? The sentence has been reworded “The SE and NE roof slab quadrants had a lengthened vibration period due to the substantial added mass on the east side of the roof level.” References: Federal Emergency Management Agency (FEMA), (2009). “NEHRP recommended seismic provisions for new buildings and other structures”, FEMA P-750. Building Seismic Safety Council. Miranda, E. (2003). ”Building Specific Loss Estimation for Performance Based Design”, Pacific Conference on Earthquake Engineering. University of Canterbury, Christchurch, New Zealand, 2003. Ryu, K.P., Reinhorn, A.M. and Filiatrault, A., (2012). “Full Scale Dynamic Testing of Large Area Suspended Ceiling System”, 15th World Conference on Earthquake Engineering (15WCEE), Lisbon, Portugal. Reviewer #3: Seismic Simulation of Integrated Nonstructural Systems at E-Defense, Part 1: Influence of 3D Structural Response and Base Isolation This paper is the first of two papers that summarize an impressive experimental study to investigate dynamic response of nonstructural components. It is well written and comprehensive. A few questions and comments are listed below: * Was there a specific reason as to why the typical iterative feedback control algorithms were omitted while replicating the target ground motions? We thank the reviewer for the overall assessment of the series of papers. Regarding this first question, during the test program, we were limited to 6 total days of testing and about 7 trials per day. We had to choose between slowly ramping up one or two ground motions to reach a final target, or applying several different motions at their target intensity. However, upon reworking this paper we have decided to remove the statement about the feedback control algorithms as we believe it is unnecessary. Rather we have chosen to provide additional data from which the reader can evaluate how the target motions were replicated by the table. * How was the scaling applied to obtain the given vector sum of horizontal accelerations? Was there any scaling applied to the vertical component? A common scale factor was applied to all components of the ground motion (including vertical) to achieve desired targets in the horizontal direction. In the revised manuscript, the ground motion selection and scaling is discussed in further detail, and the scaling procedure is clarified in lines 217-221. * Please check if in line 406 it is meant "structural" or "nonstructural damage" We confirm that the intent is “non-structural damage” (now changed to “CPP damage”.) * A discussion about the structural damage, if any, observed in the experiments will be useful. The following statement has been added for clarification: “Other than mild cracking in the concrete slabs, some of which was present in the testbed building prior to this experiment, structural damage was not observed in any of the simulations.” (lines 364-365) * Was there any correlation between the sudden acceleration peaks due to uplift and damage? No, there was not correlation because the damage to CPP systems was determined by the slab vibration, which was not affected by the uplift in the isolation system. An expanded discussion of the comparative response of the three configurations to the largest vertical acceleration input, which includes the uplift excursion in the TP configuration, has been added (lines 436 to 454).
* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project