Fatigue life prediction of type 304 stainless steel under sequential biaxial loading

Xu Chen, Di Jin, Kyeong‐Hwa KimInternational Journal of Fatigue 2005 · DOI

TL;DR

This paper investigates the fatigue life of type 304 stainless steel under sequential biaxial loading, a common scenario in engineering components. Existing fatigue models struggle to accurately predict life under these complex, non-proportional loading conditions. The authors propose a modified damage model that accounts for additional hardening observed in non-proportional loading, showing improved predictions within a factor-2 scatter band.

What Problem Does This Paper Address?

Engineering components often experience complex, multi-axial loading that changes over time, known as sequential biaxial loading. Predicting how long such components will last under these conditions is crucial for ensuring safety and reliability. While simple models like Miner's rule exist, they often fail to accurately predict fatigue life when the loading path changes or when the loading is non-proportional (e.g., axial and torsional loads applied out of phase). This non-proportionality can lead to "additional hardening," where the material becomes stronger than expected, but also potentially alters the fatigue damage accumulation. Previous research has shown that type 304 stainless steel exhibits this additional hardening, but existing models don't adequately capture this effect, leading to non-conservative (i.e., overly optimistic) life predictions. This paper aims to address this gap by developing and validating a new fatigue damage model that specifically accounts for the effects of sequential, non-proportional loading.

Methods

The researchers conducted low-cycle fatigue tests on type 304 stainless steel specimens using a specialized tension-torsion machine. They subjected the specimens to various sequential loading paths: axial followed by torsional, torsional followed by axial, in-phase axial-torsional followed by 90°-out-of-phase, and vice versa. Strain control was used, and failure was defined by a 10% drop in load.

To evaluate existing models, they applied Miner's linear damage rule, the double linear damage rule, the damage curve approach, and Morrow's plastic work interaction rule to their experimental data. They then proposed a modified damage model that incorporates a term to quantify the non-proportionality of the loading path, drawing on a definition from previous work. This new model aims to better capture the additional hardening observed under non-proportional loading.

Datasets

The study utilized experimental data generated from fatigue tests on type 304 stainless steel specimens. No pre-existing datasets were used for training or evaluation. The material properties (chemical composition, elastic modulus, yield strength, etc.) are provided in Table 1. Fatigue life data for various sequential loading paths are presented in Table 2.

Key Insights and Ideas

  • Non-proportional loading causes significant additional hardening: The tests revealed that when axial and torsional loads are applied out of phase, the material exhibits substantial additional hardening compared to in-phase loading or purely axial/torsional loading.
  • Loading sequence matters, but not always as expected: While cross-hardening was observed in axial-then-torsional loading, it was absent in torsional-then-axial. The effect of loading sequence on fatigue damage was not always clear-cut, with some data points falling below the linear damage line.
  • Existing models are insufficient: Standard fatigue models like Miner's rule, the double linear damage rule, and the damage curve approach consistently provided non-conservative predictions, meaning they overestimated the fatigue life.
  • A new model for non-proportionality is needed: The authors developed a modified damage model that explicitly accounts for the non-proportionality of the loading path, showing improved predictive capability.

Key Results

  • The experimental data showed that sequential loading, particularly 90°-out-of-phase loading, led to significant additional hardening in type 304 stainless steel.
  • Existing fatigue models (Miner's rule, double linear damage rule, damage curve approach, Morrow's plastic work rule) were found to be non-conservative, predicting longer fatigue lives than observed experimentally.
  • The proposed modified damage model, which incorporates a non-proportionality parameter (J), showed improved predictions. When validated against the experimental data, this model achieved correlations within a factor-2 scatter band, a significant improvement over previous models.
  • For the specific loading conditions tested, the non-proportionality parameter J was found to be between 0.31 and 0.35. The material constant β in the modified model was determined to be 3 for best correlation with the data.

Limitations and Future Work

The authors acknowledge that the number of tests performed was relatively small, and the scatter in the data made it challenging to draw definitive conclusions about the effect of loading sequence. They suggest that further testing is needed to verify the proposed dependence of damage on the non-proportionality between loading paths. Additionally, the proposed model's dependence on loading conditions needs further evaluation. The paper also notes that a deeper understanding of the failure mechanisms, particularly at the microscopic level, is required to fully explain phenomena like the significant damage accumulation observed in some cases. Future work should focus on more extensive testing and potentially incorporating microstructural analysis to refine fatigue life prediction models for complex loading scenarios.

Summary updated

BibTeX
@article{chen2005fatigue,
  author = {Xu Chen and Di Jin and Kyeong‐Hwa Kim},
  title = {{Fatigue life prediction of type 304 stainless steel under sequential biaxial loading}},
  journal = {International Journal of Fatigue},
  year = {2005},
  doi = {10.1016/j.ijfatigue.2005.05.003},
}

Loading related content...

Figure