Advanced

Reliability-based assessment procedures for existing concrete structures

Jeppsson, Joakim LU (2003)
Abstract (Swedish)
Popular Abstract in Swedish

Möjligheten att använda tillförlitlighetsteori som ett verktyg för utvärdering av säkerhet och återstående livslängd av skadade konstruktioner har undersökts. Det framkommer att tillgängliga riktlinjer och stödjande dokument är öppna för tolkningar och detta leder till resultat med oacceptabla skillnader. Tolkningar förkommer även vid deterministiska beräkningar men tillförlitlighetsteorin är känsligare för gjorda antaganden och skillnaderna i resultaten blir därför större.



En litteraturstudie visar att livslängdsprediktioner inte är pålitliga mer än 10 till 15 år fram i tiden beroende på de stora osäkerheter som är kopplade till analysen. Baserat på dessa förutsättningar så... (More)
Popular Abstract in Swedish

Möjligheten att använda tillförlitlighetsteori som ett verktyg för utvärdering av säkerhet och återstående livslängd av skadade konstruktioner har undersökts. Det framkommer att tillgängliga riktlinjer och stödjande dokument är öppna för tolkningar och detta leder till resultat med oacceptabla skillnader. Tolkningar förkommer även vid deterministiska beräkningar men tillförlitlighetsteorin är känsligare för gjorda antaganden och skillnaderna i resultaten blir därför större.



En litteraturstudie visar att livslängdsprediktioner inte är pålitliga mer än 10 till 15 år fram i tiden beroende på de stora osäkerheter som är kopplade till analysen. Baserat på dessa förutsättningar så föreslås prediktiva modeller som är öppna för införandet av ny information samt metoder för att göra detta. Inom medicinsk statistik och reglerteknik används modeller baserad på linjär regression för detta ändamål och för tillfällen då övervakning av en funktion behövs.



Två testfall har använts, en betongdamm och en järnvägsbro. Från fallet med betongdammen kan man fastslå att de deterministiska riktlinjer som används bör utvecklas vidare. Statistiska beskrivningar av viktiga parametrar såsom, islaster och friktionskoefficienter behövs om tillförlitlighetsteori skall användas för utvärdering dammsäkerhet.



Järnvägsbron visar att modellosäkerheterna för brottmoder som används inom betongkonstruktionen saknas. Används Baysiska metoder för att inkludera testdata i analyserna så bör det fastslås vad som kan användas som a-priori information. I det här sammanhanget framkom det att en stokastisk modell för betongens hållfasthetstillväxt behövs för att etablera a-priori informationen. En sådan här modell kan också användas för kvalitativa bedömningar av betongen, om där är stora skillnader mellan förväntade värde och uppmätta värde bör betongen undersökas med avseende på nedbrytningsmekanismer av typen inre frostskador eller alkali kisel syra reaktioner.



Tillförlitlighetsteori är väl lämpat för utvärdering av existerande konstruktioner då man bland annat får tillgång till känslighetsanalyser som ger bra beslutsunderlag för både säkerhetsutvärderingar och livslängdsuppskattningar. (Less)
Abstract
A feasibility study of reliability theory as a tool for the assessment of present safety and residual service life of damaged concrete structures has been performed in order to find a transparent methodology for the assessment procedure.



It is concluded that the current guidelines are open to interpretation and that the variation in the results obtained regarding the structural safety is too great to be acceptable. Interpretations by the engineer are also included when deterministic methods are used, but probabilistic methods are more sensitive to the assumptions made and the differences in the results will therefore be greater.



In a literature survey it is concluded that residual service life... (More)
A feasibility study of reliability theory as a tool for the assessment of present safety and residual service life of damaged concrete structures has been performed in order to find a transparent methodology for the assessment procedure.



It is concluded that the current guidelines are open to interpretation and that the variation in the results obtained regarding the structural safety is too great to be acceptable. Interpretations by the engineer are also included when deterministic methods are used, but probabilistic methods are more sensitive to the assumptions made and the differences in the results will therefore be greater.



In a literature survey it is concluded that residual service life predictions should not be expected to be valid for more than 10 to 15 years, due to the large variability of the variables involved in the analysis. Based on these conclusions predictive models that are suitable for the inclusion of new data, and methods for the incorporation of new data are proposed. Information from the field of medical statistics and robotics suggests that linear regression models are well suited for this type of updated monitoring. Two test cases were studied, a concrete dam and a railway bridge. From the dam case, it was concluded that the safety philosophy in the deterministic dam specific assessment guidelines further development. Probabilistic descriptions of important variables, such as ice loads and friction coefficients, are needed if reliability theory is to be used for assessment purposes.



During the study of the railway bridge it became clear that model uncertainties for different failure mechanisms used in concrete design are lacking. If Bayesian updating is to be used as a tool for incorporation of test data regarding concrete strength info the reliability analysis, a priori information must be established. A need for a probabilistic description of the hardening process of concrete was identified for the purpose of establishing a priori information. This description can also be used as qualitative assessment of the concrete. If there is a large discrepancy between the predicted value and the measured value, the concrete should be investigated regarding deterioration due to, for example internal frost or alkali silica reactions.



Reliability theory is well suited for the assessment process since features of the reliability theory such as sensitivity analysis give good decision support for matters concerning both safety and service life predictions. (Less)
Please use this url to cite or link to this publication:
author
opponent
  • Professor Elfgren, Lennart, Luleå University of Technology
organization
publishing date
type
Thesis
publication status
published
subject
keywords
concrete, reliability theory, service life, assessment, monitoring, hydraulic engineering, Civil engineering, offshore technology, soil mechanics, Väg- och vattenbyggnadsteknik
pages
198 pages
publisher
Structural Engineering, Lund University
defense location
Lund Institute of Technology, V-building, room V:C, John Ericssons väg 1, LUND
defense date
2003-04-28 13:15
external identifiers
  • other:ISRN: LUTVDG/TVBK-03/1026-SE (198)
ISSN
0349-4969
language
English
LU publication?
yes
id
f05dab3f-6350-4be7-a432-2a7c52382243 (old id 21165)
date added to LUP
2007-05-28 14:22:43
date last changed
2016-09-19 08:44:58
@phdthesis{f05dab3f-6350-4be7-a432-2a7c52382243,
  abstract     = {A feasibility study of reliability theory as a tool for the assessment of present safety and residual service life of damaged concrete structures has been performed in order to find a transparent methodology for the assessment procedure.<br/><br>
<br/><br>
It is concluded that the current guidelines are open to interpretation and that the variation in the results obtained regarding the structural safety is too great to be acceptable. Interpretations by the engineer are also included when deterministic methods are used, but probabilistic methods are more sensitive to the assumptions made and the differences in the results will therefore be greater.<br/><br>
<br/><br>
In a literature survey it is concluded that residual service life predictions should not be expected to be valid for more than 10 to 15 years, due to the large variability of the variables involved in the analysis. Based on these conclusions predictive models that are suitable for the inclusion of new data, and methods for the incorporation of new data are proposed. Information from the field of medical statistics and robotics suggests that linear regression models are well suited for this type of updated monitoring. Two test cases were studied, a concrete dam and a railway bridge. From the dam case, it was concluded that the safety philosophy in the deterministic dam specific assessment guidelines further development. Probabilistic descriptions of important variables, such as ice loads and friction coefficients, are needed if reliability theory is to be used for assessment purposes.<br/><br>
<br/><br>
During the study of the railway bridge it became clear that model uncertainties for different failure mechanisms used in concrete design are lacking. If Bayesian updating is to be used as a tool for incorporation of test data regarding concrete strength info the reliability analysis, a priori information must be established. A need for a probabilistic description of the hardening process of concrete was identified for the purpose of establishing a priori information. This description can also be used as qualitative assessment of the concrete. If there is a large discrepancy between the predicted value and the measured value, the concrete should be investigated regarding deterioration due to, for example internal frost or alkali silica reactions.<br/><br>
<br/><br>
Reliability theory is well suited for the assessment process since features of the reliability theory such as sensitivity analysis give good decision support for matters concerning both safety and service life predictions.},
  author       = {Jeppsson, Joakim},
  issn         = {0349-4969},
  keyword      = {concrete,reliability theory,service life,assessment,monitoring,hydraulic engineering,Civil engineering,offshore technology,soil mechanics,Väg- och vattenbyggnadsteknik},
  language     = {eng},
  pages        = {198},
  publisher    = {Structural Engineering, Lund University},
  school       = {Lund University},
  title        = {Reliability-based assessment procedures for existing concrete structures},
  year         = {2003},
}