PSI - Issue 14

Ilyin A.V. et al. / Procedia Structural Integrity 14 (2019) 964–977 Ilyin A.V., Filin V.Yu. / Structural Integrity Procedia 00 (2018) 000 – 000

965

2

1. Introduction

An assessment of fatigue strength should be a mandatory part of large-scale welded structures design. The presence of welded joints limits a cyclic strength as a whole as a result of the following: design stress/strain concentration in their geometry imperfections (first of all in the weld toe zones, WTZ), presence of residual welding stress (RWS) attaining the yield stress of a material. The third factor which may become especially significant for operation in corrosion media is a structural heterogeneity of metal within WTZ. There are several of available normative and methodical documents providing for such assessments. Among them it is necessary to mention the most expert DNV standards for marine engineering (DNVGL-RP-C203 Fatigue Design of Offshore Structures, 2016 and DNV Classification Notes No. 30.7 Fatigue Assessment of Ship Structures, 2014) and British standard BS 7910 Guide to methods for assessing acceptability of flaws in metallic structures, 2013+A1:2015 including an “ Assessment for Fatigue ” section. However the opportunities of their application have some limitations:  All standards include the fatigue curves of welded joints (Wöhler or S-N-curves) represented as straight lines in double logarithmic axes:     m N . (1) These dependences are defined in the area N > 10 4 ...10 5 cycles only and accepted as independent on the strength of material. However in the low-cycle area the allowable cyclic stress should depend on static strength of material.  RWS level in welded joints accepted for estimates is always close to the yield stress of material σ Y . As the consequence of this assumption, an effect of cyclic asymmetry is not taken into account because in any loading conditions a real local cycle in fracture origin will be extremely high. This supposition may lead to excessively conservative estimates when RWS is reduced by the pre-service pressure test, special treatment or at the expense of low thickness of welded joints. Moreover, for low-cycle fatigue under tension-compression cycle or for cycles with predominance of compression it is necessary to consider that a fatigue crack grows from WTZ up to the limiting size outside the high RWS area and even through the zones of compression RWS.  It is there accepted that a corrosion medium effect leads to the shift the numbers of cycles in S-N-curves in 2 to 3 times irrespective of a particular material. It may give a too conservative estimate when an optimum anticorrosion protection is present but bring in serious design errors if an applied high-strength material tends to a time-dependent crack growth. The following issues should also be noted: non-clear subdivision of the stress concentration factor into “ structural ” and “ technological ” components, no reference of S-N-curve approach to the fracture mechanics, no account of a high number of welded joint geometry parameters. It is likely that the estimates needed to solve the above issues have to include the quantitative representation of structural and technology factors having an effect on fatigue strength. The assessments should reflect as much as possible the physical nature of fatigue process consecutive stages. Nomenclature α S-N-curve coefficient accounting for the type of welded joint and thickness of its load-transmitting part  weld metal to base metal yield strength ratio  с cyclic yield strain,  с = S c / E  strain hardening exponent  p compliance of welded parts in relation to the tension load  angle of mating between the weld and base metal at the weld toe  weld toe radius  ef “ effective ” weld toe radius  design (service) stress  nominal stress range  a c conditional stress amplitude  res inherent RWS formed at free displacement of welded parts in x -direction