Date of Award
Fall 2016
Document Type
Thesis
Department
Mechanical Engineering
First Advisor
Fournelle, Raymond
Second Advisor
Rice, James
Third Advisor
Schaefer, Matthew
Abstract
This thesis presents the study of fatigue crack propagation in a low carbon steel (ASTM A36) and two different weld metals (AWS A5.18 and AWS A5.28). Fatigue crack propagation data for each weld wire is of interest because of its use for predicting and analyzing service failures. Fatigue crack growth test specimens were developed and fabricated for the low carbon steel base metal and for each weld wire. Weld specimens were stress relieved prior to fatigue testing. Specimens were tested on a closed-loop servo hydraulic test machine at two different load ratios. Fatigue test data was collected to characterize both Region I and Region II crack propagation for each material. Test materials were characterized and fracture surfaces were analyzed. Experimental test results were compared to fatigue striation measurements taken using a scanning electron microscope (SEM). Region II fatigue crack propagation data for ASTM A36 was found to be in agreement with existing R=0.05 and R=0.6 data for ferritic-pearlitic steels. Region II fatigue crack propagation data for weld metal was generally the same as ASTM A36 and within the limits of other weld metals. Scanning electron microscopy of the Region II fracture surfaces showed that they all exhibited similar fracture features (striations), indicating that the crack propagation mechanism was the same in all cases. Region I fatigue crack propagation data resulted in higher ∆K_thvalues for AWS A5.18 as compared to AWS A5.28. ∆K_thvalues for ASTM A36 were in agreement with published values for mild steel. ∆K_thvalues were greater for load ratios R=0.05 as compared to R=0.6. The greater ∆K_th values for R=0.05 are thought to be caused by crack closure. ∆K_th values for ASTM A36 and AWS A5.18 were greater than those of AWS A5.28. The grain structure of AWS A5.28 was found to be finer than those of ASTM A36 and AWS A5.18 and is thought to be the cause of the lower ∆K_th values.