PSI - Issue 33

6

Brito et al. / Structural Integrity Procedia 00 (2019) 000–000

R.F.N. Brito et al. / Procedia Structural Integrity 33 (2021) 665–672

670

From the experimental analysis with aluminum substrates, the models with 12.5 mm ≤ L O ≤37.5 mm presented a P m higher than their CFRP counterpart by 30.7, 24.1, 8.7%, respectively; the case of L O = 50 mm, the aluminum joint had a slightly lower P m by 4.4%. The improvement in the results with the metallic substrate are since the CFRP could present interlaminar damage while for L O = 50 mm, the aluminum substrate fails. The interlaminar damage occurs due to the presence of normal stresses (  y ), present at the butt-ends of the steps, as previously demonstrated by Valente et al., though the analysis of stepped-lap joints with steel, aluminum, and CFRP substrates (Valente et al., 2019), being this effect accentuated with strong adhesives.

Figure 3. Experimental comparison between CFRP and aluminum substrates and its effect on P m .

3.2. Stress distributions As previously mentioned, stress distributions were obtained and normalized. The normalized peel stress (  y /  avg ) is shown in Figure 4a. The  y /  avg present an overall trend as a SLJ, except for the disruptions at each step, which are caused by the stress concentrations at the butt-ends of each step (Chowdhury et al., 2016). Besides the stress concentrations in such points, it can be observed that the stress distribution is approximately equal in all the steps, agreeing with the findings of a previous elastic-plastic study (Ichikawa et al., 2008). The peak stresses at x / L O =0 and x / L O =1 increase with L O , presenting values of 0.76, 1.09, 1.49, and 1.91, respectively for the four L O analyzed here; this is caused by the rotation effects inside the joint, increasing with L O . Normalized shear stresses (  xy /  avg ) for the four L O cases are reported in Figure 4b. Contrary to the  y /  avg which was approximately uniform along L O , the  xy /  avg presented large peaks at the butt-ends of each step; nevertheless, the behavior per step is similar to that from a SLJ (Mortensen and Thomsen, 1997), such effect is known as shear-lag and it is caused by the difference on stiffness between the substrates (Nunes and Moreira, 2013). The difference on stiffness is more accentuated at the butt-ends. The stress concentration at those points, increasing both  y /  avg and  xy /  avg , indicating that joint failure starts at those points; this effect is less critical in ductile adhesives because of adhesive plasticization (Nunes and Moreira, 2013).