Fretting Fatigue Characterization of Steel Bolted Connections

Fretting fatigue is a multiaxial stress problem, which occurs on contacting components under normal pressure while they are subjected to cyclic loads and oscillatory relative movement of small amplitudes at the same time. Slip-critical bolted connections are commonly used in repetitive loading applications and are susceptible to fretting fatigue failure. Among the many factors that can affect the fretting fatigue life of the bolted connections, bolt preload and surface finish are considered primary. The main objective of this thesis was to characterize the fretting fatigue failure behaviour of slip-critical bolted connections made of high strength steel subjected to different axial stress ranges, considering the effects of bolt preload and surface finish type, including as-received mill scale steel plates (Class A surface finish) and sandblasted surfaces (Class B surface finish). Both experimental and numerical methods were employed. Experimentally, small-scale fretting fatigue tests were conducted on a total of 52 double lap bolted specimens under different combinations of bolt preload (90 kN and 145 kN), surface finish type (Class A and Class B) and axial stress range (327 MPa, 403 MPa, and 479 MPa). Digital Image Correlation (DIC) and a washer load cell were used to quantify relative displacement over the outer edge of the specimens and to measure the bolt preload variation throughout the tests, respectively. After completion of each experiment, the crack initiation location, the size of the contact area in terms of the stick/slip regions, the dominant surface damage mechanisms, and the crack propagation path were identified using optical and scanning electron microscopes (SEM). Forty-four specimens failed due to fretting fatigue in the middle plates; cracks initiated in the middle plates some distance away from bolt holes either at the stick-slip boundary or the contact edge, depending on the fretting fatigue condition. The fretting fatigue life measurements, presented in the form of S-N curves, showed that, regardless of the surface finish type, increasing the bolt preload reduced the fretting fatigue life in most cases, with the reduction being most pronounced for specimens with a Class B surface finish. The fretting fatigue lives exceeded the standard design curves for Detail Category A and B given in CSA-S16:19 (2019) by more than an order of magnitude in most cases, indicating that the design curve is conservative for the particular cases investigated here. The longest fretting fatigue lives were experienced at the lowest stress range by specimens with the Class B surface finish and lower bolt preload, and by those with the Class A surface finish and higher bolt preload. Mechanisms that prolonged fatigue life included erasure of initiated cracks due to abrasion in the former case, and the development of a protective debris layer in the latter case, as revealed by microscopic analysis. DIC was employed to characterize fretting fatigue frictional behaviour of specimens under the lowest stress range, where the widest range of results were observed. Regardless of the surface finish, increasing the bolt preload value reduced the steady state coefficient of friction (COF). Numerical analysis was used to better understand the fretting fatigue behaviour. To predict the crack initiation life and location, the finite element (FE) analysis software package ABAQUS was used in combination with the advanced fatigue durability analysis software FE-SAFE, within which the multi-axial Smith-Watson-Topper (SWT) criterion was applied. A parameter proposed by Ruiz, which incorporates slip displacement, was also used to estimate the crack initiation site by postprocessing of FE results in MATLAB, to evaluate the effect of different variables on the crack initiation location. Once the location of crack initiation was known, crack propagation analysis was implemented using ABAQUS via the virtual crack closure technique (VCCT) coupled with the extended finite element method (XFEM) in conjunction with the direct cyclic procedure to stimulate the crack growth life and path. Overall, the prediction of the crack initiation location using both the SWT criterion and the Ruiz parameter agreed well with experimental observations. A comparison between the predicted and experimental total fretting fatigue lives indicated that the numerical approach could estimate the fretting fatigue life with reasonable accuracy, which confirms the reliability of proposed approach. Increasing the bolt preload magnitude generally decreased the simulated crack initiation life, with the extent of the reduction being larger for models with the Class B surface finish. On the other hand, for models with the Class A surface, the crack propagation lives increased with an increase in the bolt preload, while it decreased for the Class B models. Furthermore, the higher COF associated with a Class B surface accelerated crack initiation and delayed crack propagation in comparison to equivalent models with the Class A surface finish. The numerical results showed that, for models with the Class A surface finish, the crack initiation process accounted for the vast majority of the total fatigue life, while crack propagation consumed a very small fraction. For the Class B models, on the other hand, raising the COF substantially increased the percentage of total life consumed by crack propagation.

Name thesis:degree_name

Doctor of Philosophy (Ph.D.)

Level thesis:degree_level

Doctoral

Discipline thesis:degree_discipline

Civil Engineering

Grantor

University of Saskatchewan

Year dc:date.issued

2023

Author dc:creator

• Zangouie, Alireza

Advisor dc:contributor.advisor

• Wegner, Leon

Committee members dc:contributor.committeemember

• Chang, WonJae
• Sparling, Bruce
• Feldman, Lisa
• Odeshi, Akindele
• Walbridge, Scott

• Fretting fatigue
• Steel bolted connections
• Bolt preload
• Surface finish
• Crack initiation
• Crack propagation

Language dc:language.iso

en

Handle dc:identifier.uri

https://hdl.handle.net/10388/14662

OAI identifier oai:identifier

oai:harvest.usask.ca:10388/14662