PSI - Issue 17

Hayder Al-Salih et al. / Procedia Structural Integrity 17 (2019) 682–689

685

Al-Salih/ Structural Integrity Procedia 00 (2019) 000 – 000

4

Fig. 1. a) Relative displacement along length of crack path and b) Convergence along length of crack path

Theoretically, convergence values should approach 100% near the crack tip as crack openings go to zero. However, known crack tip locations on the C(T) specimens typically corresponded with convergence values between 90% and 95%. At higher load cases convergence values typically decreased, indicating relative displacement continued beyond the actual crack tip location. This is likely due to larger amounts of crack tip plasticity occurring at higher applied stress intensities. 2.3. Application to Distortion-Induced Fatigue Loading The crack characterization methodology based on in-plane C(T) specimens was applied to data obtained from a half-scale girder to-cross-frame subassembly. This bridge component specimen allowed for loading with a realistic distortion-induced fatigue mechanism. Loading out-of-plane for 21,000 cycles produced a vertical fatigue crack at the stiffener to web weld. The 44.5 mm (1.75 in.) crack was composed of three linear sections, with two vertical segments connected by small diagonal segment. Similar to the in plane C(T) tests, multiple load levels were defined based on realistic fatigue loading. A stereo-camera setup was used to obtain three dimensional displacement data, and convergence was calculated based on displacement in the direction perpendicular to the girder web. Using convergence values of 90% and 95%, crack lengths were calculated using the DIC crack characterization methodology. For all load cases, the average total crack length corresponding to 90% convergence was 40.6 mm (1.59 in.), while 95% convergence predicted an average crack length of 45.0 mm (1.77 in.). These predictions under-predict and over-predict the crack length by 9% and 1%, respectively, indicating the 90% to 95% convergence range is appropriate for distortion-induced fatigue cracks subjected to out of-plane loading. However, it should be noted that the although more geometrically-complex than the linear crack on the in-plane C(T) specimen, the out-of-plane fatigue crack was fairly simple for a distortion-induced fatigue crack. Often cracks on steel highway bridges can be extremely complex, with multiple cracks located in the same region and bifurcation creating two or more crack tips on individual cracks.

3. Objective

The objective of the study discussed herein was to evaluate the ability of a DIC-based fatigue crack characterization methodology to quantify a complex, bifurcated out-of-plane crack subjected to distortion-induced fatigue. A branching crack was tested under out of-plane loading, and DIC data was collected at various load levels. The efficacy of the DIC crack characterization methodology was examined for two bifurcated crack tips emanating from a single crack.

4. Experimental Approach

4.1. Girder Test Setup and Loading

A girder-to-cross-frame subassembly, shown in Fig. 2a, was used for distortion-induced fatigue testing. The half-scale girder, fabricated from A36 steel, had a length of 2845 mm (112 in.), a depth of 917 mm (36.1 in.), and a web thickness of 10 mm (0.375 in.). The bottom flange of the girder subassembly was connected to the reaction floor of the laboratory, restraining i t from out-of-plane