An investigation into the reliability performance of bridges designed according to TMH-7

Basson, Suzanne Esther (2020-04)

Thesis (MEng)--Stellenbosch University, 2020.

Thesis

ENGLISH ABSTRACT: South Africa’s bridge design code, called Technical Methods for Highways 7 (TMH-7), was last revised in 1988 and is partially based on an outdated code. Since then, the road freight industry has expanded significantly, further encouraged by changes to the traffic regulations. The expected reliability performance of TMH-7 in catering for the increased traffic demand is not clear. The purpose of this study is to investigate the structural performance of a new highway bridge designed according to TMH-7 and loaded with actual traffic loads. The study focus on the performance of TMH-7 for normal traffic conditions i.e. NA loading. A reliability analysis can provide a reliability index that measures the safety level of the structure. For the investigation, reliability analyses were performed for two case studies, based on actual traffic load effects derived from site-specific Weigh-in-Motion (WIM) data. The load effect is represented by the bending moment at midspan of a simply supported bridge. For both case studies the extreme traffic load effects were extracted and described in a probabilistic manner. By means of statistical projection the maximum load effect distribution for a 50-year reference period was obtained. The limit state function was then formulated to define the failure mode at the ultimate limit state (ULS). The First Order Reliability Method (FORM) was used to obtain an overall reliability index and the results were interpreted by comparing the reliability index to target values from existing standards. For the reliability analysis in the first case study, failure occurs when the actual traffic load effects exceed the design load effect for NA loading. Spans ranging from 5 to 50 metres were investigated. It was found that NA loading generally performs well for a typical highway bridge, especially for longer spans ranging from 30 to 50 metres. However, a poor reliability performance is seen for short narrow span bridges (especially for 5 m and 10 m spans), which agrees with previous studies. In the second case study, a critical element reliability analysis was conducted on a spine beam of a 20 m reinforced concrete twin spine deck. Failure occurs when the load effects exceed the resistance of the critical spine beam. High reliability indices were obtained, which indicate that NA loading is performing well for the bending moment capacity of a 20 m span bridge. Furthermore, a sensitivity analysis revealed that the traffic load and model uncertainty for resistance, have the most significant effect on the obtained reliability indices. With regard to bending moments at midspan, deficiencies in the traffic load model for NA loading were identified. Spans were also identified where design could be optimised to be more cost-effective. The study contributes to an improved understanding of the performance of TMH-7 for normal traffic conditions and can help direct future revisions of the traffic load model. It is recommended to extend the study to include other load effect types and to provide separate probabilistic descriptions for different loading event types.

AFRIKAANSE OPSOMMING: Suid-Afrika se brugontwerpkode, Tegniese Metodes vir Snelweë 7 (TMH-7), was laas opgedateer in 1988 en is gedeeltelik gebaseer op ‘n uitgedateerde kode. Sedertdien het die padvrag-industrie aansienklik uitgebrei, aangemoedig deur veranderinge in die verkeersregulasies. Dit is onduidelik watter betroubaarheidsprestasie van TMH-7 verwag kan word, gegewe die toenemende verkeersbehoefte. Die doel van die studie is om die strukturele werkverrigting van ‘n nuwe brug op ‘n snelweg, wat ontwerp is volgens TMH-7, met werklike verkeersbelasting te ondersoek. Die studie fokus op die doeltreffendheid van TMH-7 vir normale verkeerstoestande, m.a.w. NA-belasting. ‘n Betroubaarheidsanalise kan gebruik word om ‘n betroubaarheidsindeks te verkry wat die struktuur se veiligheidsvlak meet. Die ondersoek behels die uitvoering van ‘n betroubaarheidsanalise vir twee gevallestudies wat gebaseer is op die las-effek van werklike verkeersbelasting, wat afgelei is van terrein-spesifieke Weeg-in-Beweging data. Die buigmoment by midspan van ‘n enkel-span brug verteenwoordig die las-effek. Die maksimum las-effekte van die verkeersbelasting was bepaal en beskryf op ‘n waarksynlikheidswyse. Statistiese projeksie was gebruik om die maksimum las-effekte vir ‘n 50-jaar verwysingsperiode te bepaal. Daarna was die grenstoestand van swigting geformuleer by faling. Die Eerste Orde Betroubaarheidsmetode was gebruik om ‘n algehele betroubaarheidsindeks te vind, en die resultate was geïnterpreteer deur die betroubaarheidsindeks te vergelyk met teikenwaardes van bestaande standaarde. Faling in die eerste gevallestudie vind plaas wanneer die werklike las-effekte van die verkeer die ontwerp las-effek van NA-belasting oorskrei. Spanlengtes van 5 tot 50 meter was ondersoek. Daar is bevind dat NA-belasting goed presteer vir ‘n tipiese snelwegbrug, veral vir langer spanne tussen 30 en 50 meter. ‘n Swak betroubaarheidsvlak word egter gesien vir kort en smal spanlengtes (veral tussen 5 en 10 meter), wat ooreenstem met vorige studies. ‘n Kritiese-element betroubaarheidsanalise was uitgevoer op ‘n flensbalk van ‘n 20 m gewapende beton brugdek met twee geflensde balke. Faling vind plaas wanneer die las-effekte die weerstand van die balk oorskry. Hoë betroubaarheidsindekse was verkry, wat aandui dat NA-belasting goed presteer vir die buigmomentkapasitiet van ‘n 20 m span brug. Verder het ‘n sensitiwiteitsanalise die verkeersbelasting en model-onsekerheid vir weerstand geïdentifiseer as die veranderlikes wat die grootste invloed het op die betroubaarheidsindekse Sover dit buigmomente by midspan betref, was tekortkominge in die verkeersbelastingsmodel vir NA-belasting geïdentifiseer, sowel as spanlengtes waar ontwerp geoptimiseer kan word om meer koste-effektief te wees. Die studie dra by tot 'n beter begrip van die doeltreffendheid van TMH-7 vir normale verkeerstoestande en kan dien as riglyn vir toekomstige hersienings van die verkeersbelastingsmodel. Dit word aanbeveel om die studie uit te brei om ook ander laseffek tipes in te sluit en om aparte aarskynlikheidsbeskrywings te gee vir verskillende tipes las-gebeurtenisse.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/107753
This item appears in the following collections: