PSI - Issue 38

Isabel Huther et al. / Procedia Structural Integrity 38 (2022) 466–476 Author name / Structural Integrity Procedia 00 (2021) 000 – 000

467

2

1. Introduction Cranes industries, pressure vessel companies, turbomachinery are subjected by their codes, standards to proof load, over-speed, or internal over-pressure respectively. The aim is to qualify the structural integrity of the mechanical systems As said, these initial overload techniques are often imposed by Standards like ASTM [1] or DNVGL [2]. They specified the conditions of initial overload to apply based on professional knowledge. For example, ANSI/ASTM A391-86 standard specifies the required nominal and proof load for different chain sizes and grades. For a 10 mm grade 80 chain, the design load is 25% of its breaking strength and the proof load is approximately 57% of the breaking strength. However, no beneficial or negative effect on in-service behaviour is generally given. No justification is pointed out. The purpose of static pre-loading is therefore to mechanically relax the residual stresses inherent in shrinkage during post-weld cooling. During pre-loading, the most stressed areas become locally plasticised in the areas where the effect of the stress concentration coefficient occurs. The localized areas become the site of residual compressive stresses during the overall springback of the structure when the preload is cancelled. The hydraulic test performed on these pressure vessels is the best known example of preloading. In the following example, a 4-point bending preload with an applied nominal stress equal to 95% of the yield strength of the base metal has been implemented resulting in plasticization at the weld toe since the stress concentration corresponds to a Kt ≈ 2 [1]. It has been seen in the literature that the effect of initial overload on the fatigue life of steels is the result of a competition between (a) work hardening, (b) residual stresses and (c) damage, introduced by the plastic deformation of the material. In a structure, the geometry of the part has a very important role, especially on the areas plasticized by the pre-load and the distribution of residual stresses. There is an important effect of the loading mode (pre-load, then fatigue). It is important that the areas plasticized by the pre-load correspond to the critical areas in fatigue. And so, there is an important effect of the material, especially concerning its response to plastic deformation at the damage and work hardening levels. For steels, it may be seen that, in general, pre-strain increases fatigue strength at large numbers of cycles, with two exceptions: 1) if the plasticity introduced by the pre-strain causes damage to the material (micro-cracking or a degradation of the surface condition), the fatigue limit can be decreased, 2) If the monotonic strain-stress curve is characterised by a Lüders step, a study has shown a slight decrease in fatigue strength [8]. The aim of this study is to evaluate the effect of one overload on the fatigue behaviour for various levels of plastic strains on longitudinal non-load carrying welded specimen in S355 steel grade. The hypothesis made is that the initial overload will introduce compressive residual stresses at the weld toe, the area where the fatigue crack initiates and propagates. These stresses would normally have a beneficial effect on fatigue strength as after hammer-peening, HFMI or shot-peening [7]. Before fatigue tests, static tests are performed to understand the effect of an overload on local plasticity and residual stresses. Fatigue tests were performed in tension (R = 0.1) on longitudinal welded gusset specimens without overload to establish the reference and specimens after one overload for three levels (corresponding to a local stress near the weld toe equal to yield stress (YS), 1.25·YS and 1.56·YS).