Session: 13-12-01: Modeling of the Fracture, Failure, and Fatigue in Solids I

Paper Number: 164875

Application of ASME Code and Weldment Fatigue Design Guidelines to Life Prediction Under Multiaxial Stresses 

During service, weldments are often subjected to a combination of axial and shear loads. Even under single-channel loading conditions, the local geometry of the weld hotspots such as weld root, toe, and start/stop points creates a multiaxial state of stress which necessitates multiaxial fatigue analysis. To ensure safe design of welded joints against multiaxial cyclic damage, ASME offers guidelines for proportional and nonproportional loading analysis which can be used for damage computation and life prediction. Also, fatigue design codes such as recommendations by International Institute of Welding (IIW) and Eurocode 3 (EC3) provide fatigue classes in the form of nominal stress-life (S-N) curves. Based on the type of the welded joint geometry and the applied loading (axial or shear), the appropriate fatigue class can be selected and used for life prediction. Although the procedures for fatigue life prediction of welded joints under single-channel loading conditions have been regulated by weldment design codes, their multiaxial analyses have been limited to failure prediction based on interaction equations. In this work, multiaxial fatigue life prediction of welded tubular joints by ASME’s guideline is demonstrated and the results are compared to the predictions by IIW and EC3 fatigue classes.

The as-welded joints were fabricated by gas metal arc welding (GMAW) with 50% weld penetration using S45C steel and ER70S-3 as the base and weld metals, respectively. Experimental program included load-controlled constant amplitude fatigue tests under axial, torsional, in-phase and 90° out-of-phase combined loading conditions, all with and without the axial mean stress. Since failure under cyclic loading in welded joints often occurs at the weld, fatigue properties of the weld metal can be used to construct a reference S-N curve so that the influential factors on fatigue life can be directly taken into consideration. To capture the notch effect of the welded joint, axial and shear fatigue notch factors were determined by finite element analysis using the volume method of the theory of critical distance (TCD). The modified S-N line of the welded metal was then used for life prediction based on the equivalent stresses obtained by ASME’s guideline for multiaxial fatigue analysis. Depending on the proportionality of the applied loads, ASME provides procedures to calculate the comparable stress quantities to the obtained notched curve. The predictions by the ASME’s guideline were mostly conservative within an order of magnitude. For the case of torsional loading, using shear properties for life estimation resulted in a higher accuracy than for axial S-N curve of the welded joint, with predictions within a factor of 3 of the experimental fatigue lives.

For life prediction using the weld metal properties, the notch and surface finish effects can be explicitly incorporated into the S-N curve of the weld metal via fatigue notch and surface finish factors definable for any type of welded joint. However, the recommended S-N curves by the design codes are experimentally generated and applicable only to specific weldment configurations with inherent consideration of these effects. Therefore, fatigue classes for pure axial and pure shear loadings for the investigated joint type were selected. Using the interaction equations provided by the design codes, equivalent ranges of nominal stress were calculated and life predictions were made according to the axial fatigue class. In general, analysis based on IIW procedure resulted in more accurate life estimations as compared to EC3, since IIW distinguishes between the proportional and nonproportional loading cases and offers mean stress correction by enhancement factors. Using the uniaxial fatigue class for the case of torsional loading resulted in higher accuracy than the recommended shear fatigue curve.

Keywords: Fatigue life prediction; Welded joint; Multiaxial loading; Weld fatigue design codes

Presenting Author: Ahmad Razi The University of Memphis

Presenting Author Biography: Ahmad Razi is a PhD candidate with a background in solid mechanics and manufacturing. He obtained his bachelor's and master's degrees from Iran University of Science and Technology. His work mainly focuses on fatigue and fracture of welded joints and durability of welded components in service.

Authors:

Ahmad Razi The University of Memphis
Ali Fatemi The University of Memphis