Mini Shear

For the axial crush rails hot stamped and quenched in 400 and 700 °C dies, since the area from which specimens could be machined from each rail was limited, as was the number of parts available, an alternative to the butterfly specimen was used to obtain fracture strains for the simple shear stress state. This geometry is smaller than the butterfly and was developed by Peirs et al. [53]. Previous work at the University of Waterloo has found it to be an extremely clever

and effective method of characterizing material in simple shear [54]. The geometry of this specimen is shown in Figure 28. In the gauge section, a state of simple shear is induced.

However, due to the resulting rotation of the gauge section, the geometry features a slight eccentricity between the cut-outs in order to ensure a relatively consistent stress state up to larger strains.

41 Figure 28: Mini shear specimen geometry

Since these parts were to be produced from axial crush rails, the tear-drop shaped cut-out was machined in the rail, followed by machining the outer dimensions of the specimen. Figure 13 shows the location of the rail from which this specimen was machined. Work done by Omer [37]

showed that the microstructure of the rail in this region is uniform, with average micro hardnesses of 245 and 195HV for the 400 and 700 °C rails, respectively.

The mini shear specimens were tested using the hydraulic Instron Model 1331 tensile apparatus. A crosshead velocity of 0.03 [mm/s] was used, as this was found to induce a nominal strain rate of 0.01 s^-1. Although this specimen doesn’t typically exhibit any thinning [54], as would be expected for simple shear, stereoscopic DIC was utilized in spite of the fact that 2-D DIC could have been considered sufficient. For capturing images of this test, two Point Grey Research GRAS-50S5M-C 5.0MP cameras fitted with Tamron SP 180mm f/3.5 Di macro lenses were used. Frame rates were selected to yield between 300-400 images for each test. For computation of true logarithmic strains at the onset of fracture, a circle with a diameter of 1.2 mm was located in the centre of the gage section. A subset size of 29-31 pixels was used for DIC analysis, with a step size of 5 and strain filter size of 7.

42

3 UNIAXIAL AND NOTCHED TENSILE EXPERIMENTS AND SIMULATIONS

This chapter presents the data obtained from the uniaxial and notched tensile tests that were conducted. While these tests were conducted for the fully bainitic, fully martensitic, intermediate forced air quench, and intermediate Gleeble material conditions, only the results for the fully bainitic material quench condition are presented here, while the results from the remainder of the material conditions are included in Appendix A-Appendix D. Experimental data presented in this chapter includes plots of load-displacement response as well as a summary of selected mechanical properties. Additionally, for each test, local area strains were obtained by measuring the cross sections of the fracture specimen.

In addition to the experimental results from the uniaxial and notched tensile tests, results from finite element simulations of these tests are also provided in this chapter. A brief description of the modeling approach is provided, as well as pictures of the mesh used for each specimen geometry.

It should be noted, over the course of testing the notched tensile specimens and developing corresponding finite element models for each experiment, the use of these specimens for fracture characterization came under scrutiny, and have ultimately not been used for developing the fracture loci presented in Chapter 5. Although results from these experiments and simulations have been published [55], it was not possible to definitively identify the location of fracture initiation in the experiments. Of additional concern with these experiments was the fact that the stress state continually evolves as the specimen is tested, prompting question over the use of such tests for calibration of stress state-dependent fracture loci.Uniaxial Tensile – Fully bainitic material quench condition

43

For each of the material conditions, the first test done was the uniaxial tensile test at quasi-static strain rates (0.003 s^-1). The specimen geometry is shown in Figure 14. Load-displacement response was obtained for each test using load measurements from the load cell mounted on the Instron and displacements obtained from 2-D (fully bainitic, fully martensitic, and intermediate forced air) and 3-D DIC (intermediate Gleeble, 400 and 700 °C tailored). To derive engineering stress-engineering strain data, the dimensions of each of the specimens were measured prior to testing. Rather than mounting a physical extensometer, virtual extensometers were placed along the gauge length of the specimens using Vic-2D or Vic-3D. Ideally, local strain paths and failure strains could be determined by computing the strains locally in the region where necking occurs.

Unfortunately, due to issues of paint adhesion, seemingly a result of the aluminum silica coating on hot stamped USIBOR^® 1500-AS [4], analysis of the necked region proved extremely difficult.

Instead, failure strains were determined by measuring area reduction in extended depth of field (EDOF) images of the fracture surfaces. Figure 29 shows EDOF images of one specimen as an example. The measured area reduction takes into consideration the effect of the fracture surface angle. All of the EDOF images are included in Appendix A.

Figure 29: Extended depth of field images for measuring area reduction at failure (left: fully bainitic condition, right: fully martensitic condition)

44

The fully bainitic uniaxial specimens exhibited reasonable repeatability in terms of engineering stress-engineering strain response, as Figure 30 shows. For the fully bainitic material condition, the ductility of the material resulted in some of the paint applied for DIC flaking off in the area of the neck. While this would influence the fidelity of local strain measurements in that area, the virtual extensometer is still capable of capturing the macroscopic response of the material, as the paint remained intact in the other areas at the ends of the gauge length. A typical contour plot of equivalent strain is also shown in Figure 30.

Figure 30: Engineering stress-strain curve for fully bainitic uniaxial tensile tests and typical

contour plot of equivalent strain one frame before fracture. Note paint separation in necked region of specimen

Table 5 details the local strains at failure from the area reduction of the fully bainitic uniaxial tensile specimens.

Table 5: Properties of fully bainitic uniaxial tensile specimens

Sample B1 B2 B3 B4 Average Standard

For the material quench conditions included in Appendix B:

0

45

 Three fully martensitic uniaxial tensile specimens were tested and exhibited

reasonably consistent engineering stress-engineering strain response. When the EDOF images were being produced in order to measure area, specimen M2 could not be located. Thus, only the area strains of specimens M1 and M3 area included in Table 27.

 While the forced air quench apparatus was being calibrated to produce an

intermediate microstructure, repeated microhardness measurements of the material produced indicated greater variability than the material quenched either in still air or still oil. However, the five intermediate forced-air quench uniaxial tensile specimens tested demonstrated acceptable repeatability. Total elongation is similar to the fully martensitic material quench condition, but greater area strains were measured.

 Five intermediate Gleeble quench uniaxial tensile specimens were tested. Good

repeatability was observed in terms of UTS, total elongation, and area strain at fracture. In comparison with the intermediate forced air quench uniaxial tensile specimens, the intermediate Gleeble specimens exhibited greater elongation but lower area strains at fracture.

 The five uniaxial tensile specimens tested from each of the tailored hot stamped axial

crush rails were tested using an electromechanical MTS Criterion 45 tensile apparatus fitted with a 100 kN loadcell, rather than the servohydraulic Instron 1331.

 While the 400 °C tailored hot stamped parts had measured microhardness similar to that of the fully bainitic material quench condition [37], the measured area strains were considerably greater.

46