Rock-structure interaction : improving the consistency in the finite element modelling of rock foundations for bridges

Fourie, Dylan Andrew (2020-03)

Thesis (MEng)--Stellenbosch University, 2020.

Thesis

ENGLISH ABSTRACT: The design and modelling of foundations crosses two civil engineering disciplines, namely structural- and geotechnical engineering. The structural engineer goes into great detail when sizing foundations to ensure effective load transfer from the superstructure to the underlying geomaterials. This is usually accomplished by deriving the load and moment taken down from the superstructure onto the foundation. This load takedown is normally established as a first estimate based on either a fixed-base or an assumed springs stiffness model in structural finite element (FE). The loads transferred from the superstructure to the various piers and foundations will vary depending on the fixity assumed by the structural engineer and could result in large discrepancies when modelled with the same stiffness when certain foundations are stiffer than others. This becomes more critical in large bridge structures with tall piers where even the slightest differentials in displacement at the base of subsequent piers of the structure could lead to significant differential tilt and settlement at the top of the piers – resulting in significant load re-distribution between softer and stiffer foundations. It is therefore proposed that the analysis process is and should be an iterative process between the structural- and geotechnical engineer as settlement and distortion is best estimated by the geotechnical engineer whilst load take-down, due to these varying foundation stiffnesses, are best estimated by the structural engineer. This iteration should continue until convergence is reached between the two models. This study aims to compile a guideline to optimize the iteration process between the geotechnical- and structural engineer, and to assist the geotechnical engineer in improving the consistency in the finite element modelling (FEM) of the interaction between the structure and the rock. This was achieved by modelling a bridge footing on rock using a 3D geotechnical FE software package; obtaining the footing’s settlement and rotation; deriving structural springs and inserting these revised springs back into a structural FE software package to determine the revised load takedown. This allows for more realistic modelling by the bridge engineer. A simplified method was proposed of applying an eccentric loading, which provided accurate results when the footing was assumed to be fully rigid. The settlement values from the geotechnical model differed with less than 10% from the structural model. The derived springs, thus, model the rock-structure interaction more accurately and can be used for rigid and flexible foundations.

AFRIKAANSE OPSOMMING: Die ontwerp en modellering van fondasies oorskry twee siviele ingenieurswese dissiplines, naamlik strukturele- en geotegniese ingenieurswese. Die strukturele ingenieur gaan in diepte om die grootte van die fondasies te bepaal om te verseker dat die vrag van die bobou korrek en effektief oorgeplaas word na die onderliggende geomateriale. Dit word gewoonlik gedoen deur die vrag en moment te bepaal wat af geneem word van die bobou en geplaas word op die fondasie. Hierdie vrag wat afgeneem word, word gebruik as ‘n eerste skatting gebasseer op ‘n vasgestelde of aanvaarde veer-styfheid model in strukturele eindige element - ”finite element” (FE). Die vragte wat oorgeplaas word van die bobou na die verskeie steunpilare en fondasies sal verskil afhangende van die vastheid aangeneem deur die strukturele ingenieur en kan lei tot groot teenstrydighede wanneer die modellering dieselfde styfheid waarde gebruik terwyl sekere fondasies stywer is as ander. Dit is nog meer noodsaaklik by groot brug strukture met hoë steunpilare waar die kleinste verskil in verplasing by die basis van die struktuur se steunpilare kan lei tot ‘n beduidende kantel en besinking bo die steunpilare, wat lei tot betekenisvolle vrag her-verspreiding tussen die sagter en stywer fondasies. Dit word dus voorgestel dat die analise ‘n iteratiewe proses is en behoort te wees tussen die strukturele- en geotegniese ingenieur aangesien die geotegniese ingenieur die besinking en verdraaing meer akkuraat kan bepaal, terwyl die strukturele ingenieur beter is met die bepaling van die vrag afname, weens verskeie fondasie styfhede. Hierdie iteratiewe proses behoort aan te hou totdat konvergensie tussen die twee modelle bereik is. Hierdie studie beoog om ‘n riglyn bymekaar te stel om die iteratiewe proses tussen die geotegniese- en strukturele ingenieur te optimaliseer, asook om die geotegniese ingenieur te help om die konsekwentheid in die eindige element modellering – “finite element modelling” (FEM) – van die interaksie tussen die struktuur en die rots te verbeter. Dit was bereik deur die modellering van ‘n brug voetstuk op rots (deur gebruik te maak van ‘n 3D geotegniese FE sagteware paket); die verkryging van die voetstuk se besinking en rotasie; die afleiding van strukturele vere en deur die aangepaste waardes by die strukturele FE sagteware paket in te voeg om die nuwe vrag afname te bepaal. Dit maak dit moontlik vir die brug ingenieur om meer akkurate modellering te doen. ‘n Vereenvoudigde metode is aanbeveel wat ‘n eksentrieke vrag uitoefen en akkurate resultate verskaf het toe die voetstuk aanvaar was as heeltemal rigied. Die waardes van besinking wat deur die geotegniese ingenieur bepaal is, verskil met minder as 10% van die strukturele model. Die afgeleide vere modelleer dus die rots struktuur interaksie meer akkuraat en kan gebruik word vir rigiede asook buigsame fondasies.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/107825
This item appears in the following collections: