Reliability of cold-formed steel screwed connections in tilt-and bearing

Van Wyk, Rudolf (2014-12)

Thesis (MEng)--Stellenbosch University, 2014.

Thesis

ENGLISH ABSTRACT: The South African National Standard for the structural use of cold-formed steel (SANS 10162-2) provides capacity prediction models for screwed connections. Screwed connections are designed against shear failure of the screw(s), section tear-out, net section failure and tilt-andbearing failure. Previous studies (Rogers & Hancock, 1997) showed that the capacity is typically determined by the tilt-and-bearing type failure mode. The aim of this document is to report on the reliability of single screwed connections in cold-formed steel against this critical failure mode. Predicted nominal capacities depend on the ultimate tensile strength of the steel, the thickness of the connected plates and the diameter of the screw. Design capacities are obtained by multiplying the nominal capacities by a capacity reduction factor of 0.5, according to SANS 10162-2. Reliability is assessed by means of FORM analyses, taking uncertainty in the prediction model and variability of input parameters into account. Laboratory testing of 222 single screwed connections allowed to statistically describe the model factor, i.e. the ratio of actual tested- over unbiased predicted capacity. For each connection, the steel strength, plate thickness and screw diameter were measured, with the measured values used to predict capacity. This implies that the model factor accounts for uncertainty in the prediction model and experimental setup, while the variability of input parameters is separately accounted for through appropriate statistical modelling. Variability in the input parameters was described using appropriate statistical distributions from expert literature (Holicky, 2009:199; JCSS, 2000). For steel strength, the mean value and standard deviation were obtained from tensile tests, while mean values and standard deviations of the plate thickness and screw diameter were obtained from the above mentioned measurements. The experimental work and numerical analysis resulted in a model factor with a mean just exceeding unity and a small standard deviation. This suggests that the design code under consideration is able to accurately predict the nominal capacity of screwed connections. The FORM analysis resulted in computed reliability indexes significantly higher than the corresponding target values which suggest conservative and reliable design formulations. Die eksperimentele werk en numeriese analise het gelei tot 'n model faktor met 'n gemiddeld hoër as een en 'n klein standaardafwyking. Dit dui daarop aan dat die ontwerp-kode onder oorweging in staat is om die nominale kapasitiet van skroef verbindings akkuraat te voorspel. Die betroubaarheid analise het gelei tot betroubaarheidsindekse aansienlik hoër as die ooreenstemmende teiken waardes wat daarop dui dat die ontwerp formulerings betroubaar en hoogs konserwatief is.

AFRIKAANSE OPSOMMING: Die Suid-Afrikaanse Nasionale Standaard vir die strukturele gebruik van koud gevormde staal (SANS 10162-2) bied voorspellingsmodelle vir die kapasitiet van skroef verbindings. Skroef verbindings word ontwerp teen skroef faling, staal profiel faling, die uitskeer van skroewe en ook faling weens skroef kanteling. Vorige studies (Rogers & Hancock, 1997) het getoon dat die kapasiteit gewoonlik bepaal word deur die skroef-kantel falingsmodus. Die doel van hierdie navorsing is om verslag te doen oor die betroubaarheid van tipiese enkel skroef verbindings in koud gevormde staal strukture teen hierdie kritiese falingsmodus. Voorspelde nominale kapasiteite hang af van die treksterkte van die staal, die dikte van die verbonde profiele en die diameter van die skroef. Volgens die SANS 10162-2 word die ontwerp kapasiteit verkry deur die nominale kapasiteit met 'n kapasiteitsverminderingsfaktor van 0.5 te vermenigvuldig. Betroubaarheid word ontleed deur middel van ʼn eerste orde betroubaarheidsmetode analise, met die in ag neming van onsekerheid in die voorspellingsmodel en wisselvalligheid van die parameters. Laboratoriumtoetse van 222 enkel skroef verbindings het ʼn statistiese beskrywing van die model faktor toegelaat. Die model faktor is bereken as die verhouding tussen die getoetste kapasitiet en die voorspelde kapasitiet. Die staal sterkte, profiel dikte en skroef diameter is gemeet vir elke verbinding met die gemete waardes wat gebruik is om die kapasiteit te voorspel. Dit beteken dat die model faktor slegs onsekerhede in die voorspellingsmodel en van die eksperimentele opstelling in ag neem, terwyl die wisselvalligheid van die parameters afsonderlik in ag geneem word deur toepaslike statistiese modellering. Variasie in die parameters is beskryf met gepaste statistiese verdelings voorgestel deur verskeie literatuur (Holicky, 2009:199; JCSS, 2000). Aangaande die staal sterkte, is die gemiddelde waardes en standaardafwykings verkry deur standaard trek toetse terwyl die gemiddelde waardes en standaardafwykings van die plaat dikte en skroef diameter verkry is deur die bogenoemde metings.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/95891
This item appears in the following collections: