Reliability-based calibration of reinforced concrete design formulas : master thesis no.: 171/II. GR

Cerar, Benjamin (author), Turk, Goran (mentor), Yang, Yuguang (comentor), Rózsás, Árpád (comentor)

Structural design codes are one of the most important components in the process of designing structures. In this thesis we deal with probabilistic calibration of existing design codes, specifically the Eurocodes. In the process of developing this thesis, we pose the following research question: What are the partial safety factors for concrete (%C) for the selected Eurocode 2 (EN 1992-1:2018 Draft 4) design expressions obtained by performing a reliability-based code calibration? To answer the question we use an implementation of a probabilistic (reliability-based) code calibration procedure, developed at TU Delft and TNO, which allows parameters of the resistance and the load effects to be considered as random variables. In this thesis, we consider the resistance model uncertainty as one of the random variables and estimate its parameters with the method of maximum likelihood using an experimental database. Reliability analysis is performed for a large number of design scenarios. The parameters of probabilistic models for the random variables are based either on literature or expert judgment. The partial factors are free parameters of a weighted objective function, which measures the absolute deviation of the calculated reliability levels from the target reliability for all design scenarios. Here we use a symmetric objective function, the target reliability %target = 4.7, which corresponds to a 1-year reference period and consequence class 2 as defined in EN 1990, and a weight function that is based on our judgment of prevalent design scenarios. The calculated sensitivity factors % 2 indicate that only the model uncertainty is a significant parameter among the resistance variables with values between 0.5 and 0.7. For other random variables on the resistance side we mostly obtain % 2 values lower than 0.2. The reliability levels mostly follow the same behavior: the values range between 3.5 and 5.0 and are the lowest when the variable load represents most of the applied load (high load ratio values). The traffic load combination shows different behavior in terms of reliability levels, with few design scenarios achieving the target reliability. To answer the research question: the calibrated material partial factors are %C = 1.54 and %C = 1.46 for the two considered subsets of the one-way shear resistance formula, %C = 1.61 and %C = 1.46 for the subsets of the punching shear resistance formula. It appears that the partial factors are close to the present value %C = 1.50, however, more research should be done to investigate the reliability level when designing according to the traffic load combination. Additionally, we recommend exploring the calibration for different levels of target reliability, as well as performing a sensitivity analysis in order to determine whether it makes sense to include the random variables that show low % 2 values in the calibration.

• Language: English
• Keywords: civil engineering, master thesis, reliability analysis, structural design codes, uncertainty
• Work type: Master's thesis/paper
• Typology: 2.09 - Master's Thesis
• Organization: FGG - Faculty of Civil and Geodetic Engineering
• Place of publishing: Ljubljana
• Publisher: [B. Cerar]
• Year: 2020
• Number of pages: XX, 55 str., [25] str. pril.
• PID: 20.500.12556/RUL-121870
• UDC: 624.012.45(043.3)
• COBISS.SI-ID: 37806083
• Publication date in RUL: 04.11.2020

Secondary language

• Language: Slovenian
• Title: Kalibracija enačb mejnega stanja za armirano betonske konstrukcije na osnovi zanesljivosti konstrukcij : magistrsko delo št.: 171/II. GR

Abstract

Standardi za projektiranje gradbenih konstrukcij predstavljajo enega izmed najpomembnejsih vidikov % v procesu gradnje. V magistrski nalogi obravnavamo probabilisticno kalibracijo enačb mejnih stanj v % omenjenih standardih, natančneje v Evrokodih. Tekom te magistrske naloge postavimo raziskovalno % vprašanje: kolik % sna je vrednost delnega varnostnega faktorja za beton ( % %C) za izbrane enačbe mejnega % stanja v EN 1992-1:2018 D4 ob kalibraciji na podlagi analize zanesljivosti? Na vprašanje odgovorimo % z uporabo procedure za probabilisticno kalibracijo mejnih stanj, ki so jo razvili raziskovalci iz organizacije TNO in Tehniske univerze v Delftu. V nalogi obravnavamo modelno negotovost kot slučajno % spremenljivko. Parametre njene porazdelitve izračunamo z metodo največjega verjetja na podlagi ek-sperimentalnih podatkov. Analiza zanesljivosti je opravljena za veliko število generiranih projektnih % scenarijev. V sklopu probabilisticne kalibracije obravnavamo delne varnostne faktorje kot spremenljivke % utežene ciljne funkcije, ki meri povprečno odstopanje izračunanih indeksov zanesljivosti od ciljne za- % nesljivosti. V tej nalogi uporabimo simetrično ciljno funkcijo, ter ciljno zanesljivost % %target = 4.7, ki po definiciji v EN 1990:2002 ustreza enoletnem referencnem obdobju. Uporabimo tudi utežno funkcijo, ki je določena na podlagi lastne presoje o merodajnosti generiranih projektnih scenarijev. Faktorji % občutljivosti % % 2 nakazujejo na to, da je le modelska negotovost (vrednosti med 0.5 in 0.7) relevantna slučajna spremenljivka za odpornostni model, saj so preostale vrednosti manj % se kot 0.2. Izračunani indeksi zanesljivosti so večinoma podobni za vse uporabljene obtežne kombinacije: vrednosti nihajo med % 3.5 in 5.0 in so najmanjše, ko je spremenljiva obtežba prevladujoča. Izjema je obte % zna kombinacija % za promet, kjer večina projektnih scenarijev ne doseže ciljne zanesljivosti. Izračunani delni varnostni % faktorji za beton so: %C = 1.54 in %C = 1.46 za pogoja projektne strižne odpornosti armiranobetonskih % elementov brez strižne armature, ter % %C = 1.61 in %C = 1.46 za pogoja projektne odpornosti proti preboju armiranobetonskih elementov brez strižne armature. Izračunani delni varnostni faktorji so podobni % trenutni vrednosti %C = 1.50. Kljub temu bi bila potrebna nadaljna analiza stopnje zanesljivosti obeh projektnih enačb, % se zlasti, ko je uporabljena obte % zna kombinacija za promet. Priporočena je tudi kalibracija % z drugimi vrednostmi indeksa ciljne zanesljivosti, kot tudi obšutljivostna analiza, s katero bi bilo mogoče% določiti vpliv slučajnih spremenljivk z nizkimi % % 2 vrednosti na izračunani delni varnostni faktor.

• Keywords: gradbeništvo, magistrska dela, GR, analiza zanesljivosti, enačbe mejnih stanj, negotovost