Effect of post-weld heat treatment on the fatigue strength of HFMI-treated mild steel joints

Martin Leitner, W. Mössler, A. Putz, M. Stoschka

Research output: Contribution to journalArticlepeer-review


Welding as a thermo-mechanical joining process generally induces residual stresses and distortion in welded components or structures. Mechanical post-treatment methods as the high-frequency mechanical impact treatment (HFMI) reduces the geometrical notch and introduces compressive stresses in the locally treated weld toe area, whereas post-weld heat treatment (PWHT) globally affects the whole structure. In this paper, the fatigue strength of HFMI-treated transverse non-load-carrying attachments and cruciform joints made of structural mild steel S355 before and after PWHT is investigated. Comprehensive tumescent fatigue tests and evaluation of notch topography, residual stress and distortion show the influence of the investigated post-treatment methods. To analyse the effect of distortion on the resulting stress condition during the fatigue tests, simulations and strain gauge measurements are carried out for different load cases. Finally, a local fatigue assessment based on the effective notch stress approach shows that an additional PWHT is not beneficial for fatigue strength. As an increase in distortion of the samples, and an influence on the base material properties, caused by the heat-treatment is not observable, the decrease in fatigue is mainly caused by the entire relieve of manufacturing induced (as-welded/HFMI-treated) prior compressive residual stresses to an almost zero stress value.

Original languageEnglish
Pages (from-to)861-873
Number of pages13
JournalWelding in the World
Issue number6
Publication statusPublished - 1 Nov 2015
Externally publishedYes

ASJC Scopus subject areas

  • Mechanics of Materials
  • Mechanical Engineering
  • Metals and Alloys


Dive into the research topics of 'Effect of post-weld heat treatment on the fatigue strength of HFMI-treated mild steel joints'. Together they form a unique fingerprint.

Cite this