Numerical Analysis of Threaded Rods

Abstract

Denne masteroppgaven er en numerisk analyse av gjengede stag innskrudd i treverk. To modeller ble forsøkt kalibrert basert på eksperimentelle resultater. En modell hadde et gjenget stag utsatt for aksialkraft, i både trykk og strekk, mens en annen modell ble utsatt for en sideveis kraft. For begge modeller ble det gjort forsøk hvor kraften virket på parallelt og normalt på fiberretningen til treverket. Stivheten til modellen hvor staget er utsatt for aksialkraft i strekk viste god overensstemmelse med eksperimentelle resultater, og ikke fullt så god overensstemmelse i trykk. Utdragskraft ble også undersøkt i modellen hvor staget ble utsatt for aksialkraft, og kraften som skulle til for å rive ut staget viste god overensstemmelse med eksperimenter. Basert på resultater fra denne oppgaven, er det sannsynlig at det er mulig å kalibrere en modell som kan gi realistiske resultater, uten for mye kompleksitet.

This thesis is a numerical analysis of threaded rods embedded in wood. Two models were calibrated based on experimental results. One had a threaded rod subjected to tension and compression, while the other model was concerned with lateral loading. In the axially loaded model, the stiffness in tension and compression was studied both parallel and perpendicular to the grain. This was also done for the laterally loaded model. The stiffness of the axially loaded model in tension was in general in good agreement with experimental results, while less so for the same model in compression. The author was not able to calibrate the stiffness of the laterally loaded model. Withdrawal failure was also modelled for the axially loaded model, and the withdrawal strength was in general in good agreement with experimental results. A parametric study was conducted parallel to the calibration, and showed good and predictable results. Based on the results of this thesis, it can be possible to successfully calibrate and create a realistic model, that yields accurate results without using too much computational power.