Scribd Logo
Upload

Welcome to Scribd!

  • 10 Vol7 3893 PDF

    Uploaded by

    VENKATA RANGANADH
    12 views6 pages

    Original Title

    10_vol7_3893.pdf

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Report this Document

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    12 views6 pages

    10 Vol7 3893 PDF

    Uploaded by

    VENKATA RANGANADH

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    of 6
    EtquakeEngneog, Tenth Weis Conearce 1852 Bakers, Ratecsam. ISBN 9054100605 Seismic pounding effects Survey and analysis. K Kasai, WJeng, PC. Patel & JA. Munshi Minis Institute of Technology: Chicago, lL, USA B.EMaison| {8D Inc. Berkeley Calif, USA ABSTRACT: The results from a pounding damage survey and analytical pounding studies are presented. These Include, a pounding survey from the 1989 Loma Prieta earthquake, development of pounding dynamic analysis computer programs, parameter studies on building pounding response as well a appurtenance response, aspect ‘method to oblain péak pounding responses, acral case stues, a spectum method to determine required Bullding separations to preclude pounding, and a possible pounding miigaion technique, 1, INTRODUCTION Building collision, commonly called ‘pounding’ occurs during an earthquake when, due wo ther ciferent dynamic characteristics, adjacent buildings vibrate out of phase and there is insufficient separation distance between them. any ‘ncidens of seismic’ pouncing "have. been reported 10 date, Pounding of adjacent buildings has made damage worse, and/or caused total collapse ofthe buildings. The earthquake that stuck Mexico Ciy in 1985 has revealed the fact tat pounding was present in lover 40% of 330 collapsed or severely damaged buildings Surveyed, and in 15% of all cases it led to collapse (Rosenblieth and Meli 1986). Ths earnquakeiusrated {he significant seismic hazard of pounding by having the largest numberof buildings damaged by its effect during 4 single earthquake (Bereero 1986) Continued research is urgently needed in order to provide the engineering design profession with practical ‘means to evaluate and mitigate the extremely hazardous effcs of pounding. ‘The following describes the writers" furrent research effort. 2, POUNDING SURVEY ‘The writers surveyed the damage due to pounding inthe San Francsso Bay area during the 1989 Loma Pri. Earthquake (Kasal and Malson 1991). This survey is compiled from data provided by: engineers, government officials and engineers, building owners, and block:by- block inspections performed by the writers. The database Contains the input of about 90 interested parties and ‘ecards more than 200 pounding occurrences involving ‘more than 500 structures, Significant pounding was observed at sites over 90°km from the epicenter thus Indlesting the possible catastrophic camage that may oocur during fare earthquake having closer epicenters. ‘Classification of Pounding Damage. - Pounding damage pauerns are classifieds follows: Type-1, major euctural damage; Type2, fallie and falling of building appurtenances creating alife-safty hazard; Type-3, loss fof building function due to failure of key mechanical, clerical of fire protection systems; and Type, frchltecural and/or minor structural damage. ig. 1 shows an example for Type-1 damage. The 10 story building is consruced of thick masoary walls (13 Inch thickness) combined with 9 steel plane frames. I was built in 1904, This bullding experienced severe ‘pounding with an aijacent massive 5 story building which ‘cups most ofthe ety block, ‘The $ story building is ‘originally & concrete frame building baving a very sft wall atthe 2nd level, and was sesmically upgraded by ‘ding tel races in 1980.” Pounding wat loeated a he ‘7 level in the 10 story building anda the roof level in the 5 story building (Fig. 1). Only 1 to 1.5 inches building separation is present. The 10 story bulding suffered srutural damage above the pounding elevation 4s evidenced by the large diagonal shear cracks in the rasonry piers (Fig. 1 right) Fig. ‘Type-L Pounding Damage Example. Fig. 2 shows an example for Type? and possly ‘Type-1 damage. The building is 10 stories and constructed Teton cnr Ap enone arte 2 lef, bling argh. It yas ball in Bes" “i fatdng pounded wn an elder 7 sary Inlting whose ower ris are compoued of einforesd ere, andi uppe 3 sores ae of sel constuction {Pig 2 bung atl) The balling have about? {rf seperation Tee 10 story buiding sured significant dinage Farber stady is teed t determine wheter 3290 this damage is atributed to pounding, | However, pounding. caused the seismic hazard of falling, bullding febris. The brick veneer at the boundary of the two bulldngs was damaged due to impact, anda large amount of fling debes fellon and through the canopy located at the 2nd floor level Fig, 2: Type-2 Pounding Damage Example. Fig. 3 shows an example combining Type-2, 3, and 4 damage, The building isa large 6 sory modern steel ‘ezucture occupying an ene ety block.” tt was builtin 1981. The building in plan, consists of three segments separated by 4 inch expansion joins oe to the earthquake, these segments pounded at their ‘oor slabs which are at common elevations, produces sharp irregular motions which resus in large high frequency lateral accelerations (Kasal etal. 1990). The ‘windows facing the atrium fell down, and the computer ‘equipments shined and/or tured over. Heavy building equipments in the penthouse shifted significantly (Fig, 3) Fig. 3: Type-3 Pounding Damage Example. Survey Findings and Comments, ~ The following. are some of the general survey findings and comments: (The majority of reported cases are in urban areas Including San Francisco (e.g., Fig.4), Oakland (Fig. 5), Sana Cruz and Watsonville (@) Pounding ypically involved muli-story buildings constructed prior to about 1930. They are typically of ‘masonry construction ‘with or without stel skeletal verical load resisting systems. Very litle consideration was given for separation between’ such buildings to Preciude pounding. In many cases, hey are in contact With eachother. (G)_ Fewer modern buildings suffered pounding. In such buildings, relatively larger separations exist. However, it 's noted that many modern buildings having expansion Joins suffered pounding due t9 small separations (4). There is evidence of correlation between occurrences ‘ot pounding and soft foundation sll conditions. ‘This may belatributed to the more intense shaking typically ese Fig. Francisco Downtown. Centr. reported for such sll conditions and/or from the possible setlement and rocking of the structures located on soft soils. (©). Special pounding cases were also observed. They Include; severe pounding at unsupported part (¢4., ridheigh of columns or walls; pounding promoted by torsional behavior of building; and pounding berween the buildings sharing a common wall (© Olde buildings that suffered Type-I damage typically also had Type-2 damage (Le., falling bricks). Modern buildings that pounded usually had Type-4 damage, and Several of them also suffered Type-3 damage. The sirvey ‘has relative distributions for damage Types 1 and 4 of 21% and 79%, respectively, Many ofthe present Type ‘damage cases will become damage Types 1, 2, and/or 3 ‘when a future more severe earthquake affects the region. ‘The Type-4 damage cases may be thought of as precursors forthe major pounding damage yet to occur. 3. POUNDING TIME HISTORY ANALYSIS elation, = The writs ave developed ize computer pounding analysis programs SLAM and SLAM rhc fre made pubis avalable (Maison and Kasi 198, 199. The programs ideale bulngs a dee dimensional” GD)" maidegree-ofeedom (DOF) sysems. The SLAM program assumes that a buling ier colies wth & rigid ace elding and he SLAMC2 (Fig. 6)prostam considers that bth aiings fee exible. Pounding i sumed to occ a sighs fice level having a Tigi dapragm, ‘Tre pound problem sale as having wo tear sas nS Ste bullings vibrate witout comac, and in State 2" ‘he buldings af n coma.” A nolinear problem ess Ss the feaponseossies fom one ner sate the ther. Thee esltions were made ava saring point inthe pounding investigation inorder to make the problem” manageable, wile reting ipa 3D: founding nunc characerisics. The programs employ Seeretaly exact soon shame Qvason and Rash 1:90), and hey a alo computational efter Fig, 6: The Pounding Problem and SLAM-2 Idealization, Sample Analyses Rens. - An exning 15-0 ste ‘mom resistant fame buiing Buling A, petiad = 113 see is assumed to calide with an adjacent Dexbie xory balding Bullding B, period = 0.8 sec). The flor mats of Bling Bis eonsigered tres times that of Bulding A” SLAM dynam analyses were conducted Using 0-4 areal earthquakes. (Kas etal. 1990) Danping tat i act 5% and 1.5 inches separation Is conslered for pounding eae. Local cont sifess (Gi, 6) ato SOODD Win sonsdering he past suis {httison and Rasat 1990). Pounding fs a severe load Condivon, Impacts a he pounding lve eli age Sd quick acreraion Pulses, nd’ the peak” foe San cn rte dan 0 et oie rom he torpoundng care (compare Figs. 1) ard Troan buildings, pounding produces peak dis, shears, OTM's, tongues and acelraon aus Sor) ievels that ae gretr than ose from the no pounding aut (ep Fag. 8). Bullding.midhelght_ pounding {Bulding’Ayincreates shears above pounding Tevel as Veta eceleratns inde vil of de nga F.9 Bling top level pounding Guling B) decreases peak shart over fhe ene bling eight win. he Exception ofthe stories in the viity of the impact. Furter analyses were conducted by varying the Moor rae, Arh ference in ie relatv mas ieeases, the Sere effects of pounding increase ne bulding faving the lewoe mais, These Toetions of pounding tuple responses correspond to the observed damage isations ine resent earhausk (8, i. Fig. also shows the response quan when the builengs have 208 damping. Tis clear fom the gure Sha his damping reverse respons signa. to ul (see Tbe hes. ACCELERATION (6) ABS. ACEELERATIN () rat (See Fig. 7: Absolute Floor Accelerations, Pounding Case ((S-story against 8-tory Building) vs. No-Pounding Case ee ties Bonen) we actiwal °F aPacreunel @ Fig. 8: Response Envelops: Pounding Case (Damping = 58 of 208) vs. No-Pounding Case (Damping = 5%). ‘Tis issue is furer discussed in Sec. 8 "The flexible adjacent building cases studied have many eends that are similar to those ffom a rigid adjacent bulling case. ‘The righ adjacent building ease, therefore, was the fst subject of the writers” study, results of which sxe discussed below. 3498 4, SPECTRAL RULE FOR POUNDING RESPONSE ‘Through numerous SLAM analytical sues, the writers found that the non-linear pounding peak response of SDOF as well a5 MDOF systems is not sensitive to the details of the particular earthquake history as long a5 the farhquakes have a common spectrum characteristics (Kasal etal. 1990). Based on this, the following method to predit the peak pounding responses were developed. ‘No Pounding and Fixed Spring. Systems. The technique is based on response spectrum analyses of two basic linear systems (Fig. 9): (1) no pounding system (the building vibrating without contac), and (2) fixed spring sjstem (Oe Building vibrating in continuous contact with the adjacent strucute). The peak response of the pounding system Is predicted’ by considering the Aistibution of earthquake energy in both systems inthe form of Kinetic energy and stain energy in each linear stue, “The peak pounding responses of MDOF system are Calculated a5 follows (Kasai etal. 1990) fw} = afuy) and (u} = 2luy) + rlue) in which {u'} and {u*} = the peak negative and positive Ghplacement vectors, reapecvey: (ue) and ox) = the pedi dspiacementvecior obtained fron common used frltinode response spec analysis of he no pounding Syxem and te Ted sping system, respective. The Siparaton rato Ps dened ate rao of the atcos Separation diane divided by the pak dplacoment of SE nopounding sytem a the conesponding story Evel Theat and are baited fom simple equations fonsiing of the kinetic enerpes_ a2 appronmatly Somputed vsing te ist mocal partpaon factor and tariguake parudo-elocky apeca (kaa eta 199) Esimaions of he oer peak pounding responses sich a5 Ais, shers, and OTM can be aie in tsar wy Fig. 10: Theory vs. Average of SLAM-Analysis Results (Gix Amificial Earthquakes, 0.48) ‘Accuracy. » The theory was verified by more than 700, cases (Kasal and Patel 1990) comparing the theorsicd results to those from SLAM analyses for SDOF systems as well as MDOF systems. Fig. 10 illustrates the good accuracy of the theory for predicting MDOF pounding System peak response for various separations in a mi height pounding case and a top pounding "case, respectively. Note that vertical location of pounding significantly. influences the distribution of story peat sponses through the height of the bullding, and thatthe ‘hears remaln almost the same with the separation rai from 010 2/3, ‘5, BUILDING APPURTENANCES RESPONSE ‘The writers observed damage to bulldng appurtenances such’ electrical and mechanical equipments, buldng parapets and curtain walls which was taused by poundiag Sf buildings during. Loma Prieta earthquake’ (Types Gamage, See. 2). As discussed eatier Fig. 7) te peak flor accelerations can be more than 10 times thse from the no-pounding case. It was also found tat aid adjacent building case gives the results similar to tote {rom a celavely heavy exible adjacent building cave ‘The following stdies consider the righ adjacent Dung ae (Kata etal. 1980). ‘loot Acceleration Response Spectra. - The foot acceleration response spectra (FARS) atthe top pounding level of tie 15-Rory bullding sre showm in Fig 1. They Indica that pounding is especially harmful for equipment ‘o¢ secondary systems having short periods (=O se) This effect is not covered by existing indueal design spec. For example, sce the Network Equipment ‘Building System (NEBS) design spectrum given by Bel Comimnicaton Research (BELLCORE 1968), which is ‘ery close to the FARS of no-pounding case. The FARS inthe pounding case can be as much as 30 times higher than tose in no-pounding case. Based on these, ie commonly considered metiod of designing the secondary systems to have shore periods to reduce the system Fesponse may be effective Only when no pounding eeu, but would be significantly unconsevative in a pounding condition. Neglecting the eect of damping, the aceleration dof ly pounding level ducing pounding (Stato 2)" approximately expressed trom equilibrium as Via ViekO-d1/m > ‘where Vj, Uy and m, = story shear, displacement, and mass of the’ th floor level, respectively, = Yoel ‘contact stiffess (Fig. 6), and $= av-cest Separation distance. The writers have found tat the peak, at tae 2 is approximately obtained by substituting into Eg. 2 the peak ¥,, peak Vie, and peak u, chat are estimated using the simplified method explained earlier (Eq, 1) ‘The walters have also found that the ratlo between pounding FARS and no-pounding FARS, hereby defined 5s a spectrum. amplification, remains very stable egardless of different separation ratios (0 0 about 23) and earthquakes types (Fig. 12) (Kasal et al. 1990). Because ofthis effort and considering Eq. 2, a simplified method of obtaining pounding FARS seems possible. a 3608 a reli (stcoues) Fig 1: FARS, Flor Acceleration Response Specra (3Story Building), Fig. 12: FARS (due to pounding) Normalized to FARS (de to no-pounding), 18¢Story Bullding 6. ANALYSIS OF DAMAGED BUILDINGS ‘The writers are conducting corelaive pounding analyses ‘of actual buildings damaged during Loma Prieta earthquake, “The following describes one of the sample ‘analytical studies: ‘Pounding between the 10 story bullding and $ story building explained in Sec. 2 war analyzed (es Fig. 1. ‘The building pounded near the comet of the buildings (Fig, 13). A 3D-dynamic analyses were performed using SLAM-2. ‘Fig. 14 shows an analysis result using a 2 digecona earthquake motion (0.163) recorded near the study buildings during the Loma Prieta event, Note the large shear above pounding level. and large torsion

    You might also like