Site specific traffic load factor approach for the assessment of existing bridges

Perez, Sergi (2020-12)

Thesis (PhD)--Stellenbosch University, 2020.

Thesis

ENGLISH ABSTRACT: Maintenance of existing bridges is a matter of on-going concern for engineers and local authorities. In most developed countries, a large percentage of bridges in service nowadays was constructed decades ago. The design of these structures was executed using codes and guidelines that could be now outdated. In addition, owing to increased traffic loads, traffic intensity and deterioration of materials due to aggressive environments over time, it is necessary to evaluate bridge performance to assure the continuance of high levels of safety and comfort for users. Even though important advances have been accomplished in this matter, especially regarding the resistance side of this issue, there are still opportunities for enhancements of the traffic load models used for the assessment of existing bridges. Recent technological advances as Weigh-in-Motion recordings and probabilistic approaches such as Monte Carlo simulations of artificial traffic allow for more accurate estimation of the traffic load effects on bridges. This is, however, not an option for the majority of practicing engineers and it is not desirable from the bridge-owners’ perspective due to the potentially high associated cost of investigations. In this contribution, an innovative, general and simple procedure for the estimation of the site-specific load effects on existing bridges is presented. Traffic design load model effects are modified through the application of various site-load factors in order to adapt the load effects to the specific conditions of the traffic circulating over the assessed bridge. In contrast to previous studies, the here developed method is based on simple observation of basic traffic descriptors such as average daily truck traffic, or percentage of long vehicles. These can be easily obtained, which presents a novel and simplified approach in comparison to WIM data based reduction factors. The calibration of the site-load factors is performed using a new Monte Carlo routine based on bivariate copula functions. The simulation of artificial traffic not only allows for the generation of extreme events that are not usually recorded by WIM stations but also allows for the modification of certain variables in the traffic generated to evaluate its influence on the load effects. It therefore also allows for a simulation of traffic not encountered on major highways. Shorter remaining service life is usually at play when assessing existing structures. This is usually taken into account by reduced reference periods and consequently lower characteristic values of the variable loads. Reduced target reliability levels and therefore, lower safety margins, are also considered in the assessment due to the increased cost of strengthening and achieving an optimal cost of safety. Both concepts are applied here for the calibration of the partial safety factors. Moreover, an additional partial safety factor is introduced in this thesis. Traffic flow is not constant and fluctuates with the time, therefore this affects the estimation of the traffic descriptors necessary for the calculation of the site-specific load effects. This new partial factor accounts for this uncertainty and mitigates its undesirable consequences. The proposed procedure is validated using reliability analysis techniques to ascertain that the intended reliability levels are achieved in any scenario. Lastly, a general framework for the calibration of the design traffic load models for the assessment of existing bridges is established here. The most important aspects of this work are discussed and summarised for its applicability in any other regions. The presented results show the important modifications on the design load model load effects that can be achieved by taking into account site specific traffic conditions from simple traffic counts. Especially, the estimation of the traffic load effects in minor roads with very light traffic can be significantly improved. In addition, the new safety partial factors calibrated with reduced reference periods for shorter remaining service life and reduced target reliability indices allow for a more accurate assessment of existing bridges. The service life of the assessed bridges could be, therefore, extended or unnecessary and costly repairs avoided.

AFRIKAANSE OPSOMMING: Die instandhouding van bestaande brûe is 'n saak wat die ingenieurs en die plaaslike owerhede deurlopend bekommer. In die meeste ontwikkelde lande is 'n groot persentasie brûe wat tans in diens is, dekades gelede gebou. Die ontwerp van hierdie strukture is uitgevoer met behulp van kodes en riglyne wat nou verouderd kon wees. Vanweë verhoogde verkeersbelasting en verkeersintensiteit en agteruitgang van materiale as gevolg van aggressiewe omgewings met verloop van tyd, is dit ook nodig om die prestasie van die brug te evalueer om die kontinuïteit van hoë veiligheid en gemak vir gebruikers te verseker. Alhoewel belangrike vooruitgang hiertoe bewerkstellig is, veral wat betref die weerstandkant van hierdie kwessie, maak die meeste lande steeds gebruik van ontwerpstandaarde vir die beoordeling van bestaande brûe. Onlangse tegnologiese vooruitgang soos Weeg-in-Mosie stasie en waarskynlike benaderings soos Monte Carlo-simulasies van kunsmatige verkeer maak voorsiening vir 'n akkurate beraming van die effek op verkeersbelasting op brûe. Dit is egter nie 'n opsie vir die meerderheid praktiserende ingenieurs nie, en dit is ook nie wenslik uit die perspektief van die brug-eienaars nie as gevolg van die koste verbonde aan ondersoeke. In hierdie bydrae word 'n innoverende en eenvoudige prosedure vir die beraming van die terrein-spesifieke laseffekte op bestaande brûe aangebied. Effekte van die vragmodel van die verkeersontwerp word gewysig deur die gebruik van verskillende terreinbelastingfaktore om die laseffekte voldoende toe te pas op die spesifieke omstandighede van verkeer wat oor die beoordeelde brug sirkuleer. Hierdie terreinbelastingfaktore is 'n funksie van basiese verkeersbeskrywers en kan maklik verkry word deur verkeerstellings. Die kalibrasie van die terreinbelastingfaktor word uitgevoer met behulp van 'n nuwe Monte Carlo-roetine gebaseer op tweevoudige kopula-funksies. Die simulasie van kunsmatige verkeer maak nie net voorsiening vir ekstreme gebeure wat gewoonlik nie deur WIM-stasies opgeneem word nie, maar ook om die veranderlikes in die verkeer wat gegenereer word, te verander om die invloed daarvan op die laseffekte te evalueer. Korter oorblywende lewensduur word gewoonlik ondervind by die beoordeling van bestaande strukture. Dit word gewoonlik in ag geneem deur verminderde verwysingsperiodes en gevolglik die laer waardes van die veranderlike vragte. Verlaagde teikenbetroubaarheidsvlakke en laer veiligheidsperke word ook oorweeg as gevolg van die hoër koste om bestaande strukture te versterk as dit met die ontwerpstadium vergelyk word. Beide konsepte word hier toegepas vir die kalibrering van die gedeeltelike veiligheidsfaktore. Verder word 'n addisionele gedeeltelike veiligheidsfaktor in hierdie tesis bekendgestel. Verkeersvloei is nie konstant nie en wissel met tyd, daarom beïnvloed dit die beraming van die verkeersbeskrywers wat nodig is vir die berekening van die terrein-spesifieke laseffekte. Hierdie nuwe gedeeltelike faktor is verantwoordelik vir hierdie onsekerheid en verminder die ongewenste gevolge. Die voorgestelde prosedure word gevalideer met behulp van tegnieke vir betroubaarheidsanalise om te bepaal of die beoogde betroubaarheidsvlakke in enige scenario bereik word. Laastens word hier 'n algemene raamwerk vir die kalibrering van die ontwerp-verkeersbelastingmodelle vir die beoordeling van bestaande brûe ingestel. Die belangrikste aspekte van hierdie werk word bespreek en saamgevat vir die toepaslikheid daarvan in enige ander streek. Die aangebied resultate toon die belangrike modifikasies op die lasbelastingeffekte van die ontwerpbelastingmodel wat bereik kan word deur die spesifieke verkeerstoestande in ag te neem uit eenvoudige verkeerstellings. Die beraming van die effek van die verkeersbelasting op geringe paaie met baie ligte verkeer kan veral aansienlik verbeter word. Daarbenewens het die nuwe veiligheidsgedeeltes wat gekalibreer is met 'n verminderde verwysingsperiode vir korter oorblywende lewensduur en verminderde teikenbetroubaarheidsindekse, 'n akkurater beoordeling van die bestaande brûe moontlik gemaak. Die gebruiksduur van die brûe se gebruiksduur kan dus verleng of onnodig en duur herstelwerk vermy word.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/109086
This item appears in the following collections: