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

969

6

Figure 2 – Change of K  at cyclic elastic-plastic deformation of heterogeneous metal of a welded joint (base metal strength is higher than weld metal strength,  Y of base metal is referred to): 1 – weld metal, 2 – base metal.  = 0.80 (a),  = 0.70 (b),  = 0.55 (c).

2.2 Residual Welding Stress (RWS)

RWS is expedient to divide into components by the procedure of numerical evaluation and by the effect on fatigue strength (Figure 3), Karzov et al (1992):  Inherent RWS  res that may be expressed as a though-thickness (y-direction) epure self-balanced by force and by moment;  Local RWS  * res , for a multi-pass weld this fragment is a constant part of RWS and does not depend on joint type and size;  Reactive RWS  R in contrast with inherent RWS are not localized within double weld width, it spreads over large areas of a structure. In a certain “design” cross -section x-y it is not self-balanced. In terms of fatigue strength assessment it gives a full equivalent of the design stress.

Figure 3 – Scheme of RWS formation in a multi- pass weld in the area of acting reactive stress from the source “A” (welded insert) and source “B” (rectangular patch) (I), inherent RWS field (II) and its fragment, a local RWS field (III).

The initial physical information for RWS calculations described by Leonov and Manninen (2003) includes the temperature dependences of thermal and phase dilatation, yield stress of the material at its heating and cooling, specific heat and thermal diffusivity. Some related issues have been observed:  At a low-temperature  bainite or martensite-type phase transition the choice of a hypothetical dependence  Y ( T ) within the transition interval leads to more than 100% uncertainty in RWS estimates. To substantiate an adequate model of material behaviour the direct measurement of angular displacement of welded parts during cooling has been done. It was concluded that a hypothesis of the complete strength loss of material during its phase transition brings the best agreement with experiment.  Assessment of a short-time creep contribution into RWS level has shown that the creep process is can be neglected at RWS assessment in the “as - welded” state, but a post -weld heat treatment (PWHT) reduces RWS to 30% of its “as - welded” maximum value. The analysis of inherent RWS distribution gives that during multi pass welding a partial limitation of shrinkage displacement for upper beads from the cooled root beads forms a cross-section epure of stress acting transverse to