Ocena potresnega tveganja zidanih stavb : doktorska disertacija

Snoj, Jure (author), Dolšek, Matjaž (mentor)

In this thesis a new methodology for the seismic risk assessment of masonry buildings is presented. It is based on a probabilistic approach, which makes it possible to determine seismic risk using various performance measures such as the probability of exceeding a certain damage state, the probability of exceeding a certain economic loss, the expected annual loss, and the loss dependent on a given seismic intensity. Using these performance measures engineers and project stakeholders can make informed decisions as to whether particular existing buildings should be repaired, strengthened, or replaced, or simply choose the best variant in the design of a new building. The methodology consists of four main steps: hazard analysis, structural analysis, damage analysis and loss analysis. The results from these four separate analyses are combined together in a probabilistic manner, using the total probability theorem, thus making it possible to explicitly take into account numerous uncertainties. However, in this study the main focus was on the modelling uncertainties, and on the aleatoric uncertainties which are due to record-to-record variability. A database of experimental data obtained in the testing of numerous masonry specimens was established, and analysed in order to define fragility functions for masonry elements failing in shear or flexure, while also evaluating the uncertainties in the limit-state deformations of the masonry walls. The possibility of reducing the impact of uncertain input parameters by means of non-destructive tests was also explored. By measuring its ambient and forced vibrations, the vibration periods of an old masonry building in Ljubljana were determined, and used to calibrate corresponding structural models. It was shown that such non-destructive experimental investigations can contribute to more reliable seismic performance assessments of buildings, even when the performance assessment procedure is based on a nonlinear method of analysis. In addition to the explicit consideration of uncertainties, damage to the components of buildings was also explicitly simulated on the basis of the relationship between the seismic intensity and the engineering demand parameters which were obtained by simplified nonlinear analysis. The latter consisted of pushover analysis, together with incremental dynamic analysis of a SDOF model. The applicability of the proposed methodology was demonstrated by means of examples of a residential and a commercial building, which were constructed from hollow clay bricks or solid bricks. Under the presumptions of this study, the expected annual loss for buildings constructed with hollow clay bricks was 60 €/100m2 of the gross floor area, and 125 €/100m2 in the case of buildings which were made from older material consisting of solid bricks. Disaggregation of the results showed that weaker earthquakes with lower intensities contribute significantly towards the expected annual losses. It was also found that nonstructural elements, which are key components in the loss assessment model, can contribute more than 50 % of the total loss. The modelling uncertainties have a large impact on the results of the pushover analysis and of those obtained by incremental dynamic analysis, as well as on the probability of exceeding damage states. However their impact on loss estimates is more pronounced in the case of the fragility and loss functions, seismic hazard, and the costs of replacing the building.

• Language: Slovenian
• Keywords: gradbeništvo, disertacije, potresno tveganje, zidane stavbe, analiza negotovosti, občutljivostna analiza, analiza izgub, verjetnost porušitve, ambientne in vsiljene vibracije, inkrementalna dinamična analiza, potisna analiza, ekonomska izguba, pričakovana letna izguba
• Work type: Doctoral dissertation
• Typology: 2.08 - Doctoral Dissertation
• Organization: FGG - Faculty of Civil and Geodetic Engineering
• Place of publishing: Ljubljana
• Publisher: [J. Snoj]
• Year: 2014
• Number of pages: XXXIV, 241 str.
• PID: 20.500.12556/RUL-32347
• UDC: 624.042.7(043.3)
• COBISS.SI-ID: 6790753
• Publication date in RUL: 10.07.2015

Secondary language

• Language: English
• Title: Seismic risk assessment of masonry buildings

Abstract

V doktorski disertaciji predstavimo metodologijo za oceno potresnega tveganja zidanih stavb. Metodologija temelji na verjetnostnem pristopu, kar omogoča vrednotenje potresnega tveganja z različnimi merami kot so verjetnost prekoračitve izbranega mejnega stanja, verjetnost prekoračitve določene ekonomske izgube, pričakovana letna izguba ter pogojne izgube pri izbranih vrednostih intenzitete potresa. Vse omenjene mere potresnega tveganja omogočajo investitorjem informirano odločitev o sanaciji/utrditvi/zamenjavi obstoječih objektov oziroma izbiro najboljše variante pri projektiranju novih. Metodologija za oceno potresnega tveganja je sestavljena iz štirih analiz: analize potresne nevarnosti, analize obnašanja konstrukcije, analize poškodovanosti in analize izgub. Rezultati vsake od ločenih analiz so obravnavani probabilistično in med seboj povezani z izrekom o popolni verjetnosti. Takšen okvir omogoča eksplicitno upoštevanje številnih negotovosti, v okviru disertacije pa smo se osredotočili na upoštevanje modelnih negotovosti in negotovosti v potresni obtežbi. Vzpostavili smo bazo podatkov iz laboratorijskih eksperimentov na številnih zidanih preizkušancih in izpeljali funkcije ranljivosti za zidove, ki se porušijo bodisi v strigu bodisi v upogibu, s čimer smo vrednotili negotovosti v mejnih pomikih zidov. Raziskali smo tudi možnost zmanjšanja vpliva negotovosti vhodnih parametrov z uporabo nedestruktivnih preiskav. Z merjenjem ambientnih in vsiljenih vibracij smo določili nihajne čase stare zidane stavbe v Ljubljani, nato pa smo s parametrično analizo dokazali, da je s kalibracijo matematičnega modela stavbe glede na izmerjen elastični nihajni čas mogoče zmanjšati tudi vpliv negotovosti na rezultate nelinearne analize obnašanja konstrukcije. Poleg vpliva negotovosti smo v oceni potresnega tveganja eksplicitno simulirali poškodovanost komponent. Zvezo med potresnimi zahtevami in intenziteto potresa smo določili s poenostavljeno nelinearno metodo na osnovi potisne analize in inkrementalne dinamične analize na modelu z eno prostostno stopnjo. Predlagano metodologijo smo demonstrirali na primerih stanovanjske in poslovne zidane stavbe, ki sta bili zgrajeni bodisi iz opečnih votlakov bodisi iz polnih zidakov. Pri stavbah iz boljšega materiala je bila ob predpostavkah v tej študiji in izbiri modela pričakovana letna izguba 60 €/100 m2 bruto tlorisne površine, pri stavbah iz starejšega materiala pa 125 €/100 m2. Rezultate smo razčlenili po intenzitetah potresa, pri čemer se je izkazalo, da k pričakovanim letnim izgubam zelo veliko prinesejo šibki potresi, ki ne povzročijo velikih izgub, vendar se glede na uporabljen model potresa pojavijo precej pogosto. Ugotovili smo tudi, da nekonstrukcijske komponente prispevajo več kot 50 % k celotnim izgubam obravnavanih objektov, zaradi česar so ključnega pomena za oceno izgub. Izkazalo se je tudi, da imajo modelne negotovosti velik vpliv na rezultate potisnih analiz, inkrementalnih dinamičnih analiz in na verjetnost prekoračitve mejnih stanj, pri oceni izgub pa prevlada močan vpliv funkcij ranljivosti in funkcij izgub ter vpliv potresne nevarnosti in nadomestitvene vrednosti objekta.

• Keywords: civil engineering, thesis, seismic risk, masonry buildings, uncertainty analysis, sensitivity analysis, loss analysis, probability of failure, measurement of ambient and forced vibrations, incremental dynamic analysis, pushover analysis, economic loss, expected annual loss