Strain-based fatigue life estimation techniques are commonly used in the automotive industry to model the likely service behaviour of safety-critical components contained within, especially, chassis and suspension systems. Mostly, these analytical models assume that the material properties, loads and part geometry are invariant. However, the ‘real world’ is probabilistic and not deterministic in nature. The work presented in this paper surveys several different analytical techniques plus two empirically based methods that may be used to define lower bounds to both strain-life fatigue damage curves and cyclic stress-strain parameters. In addition, an empirical method proposed by Conle (SAE Technical Paper
2002-01-1279) which applies a pseudo notch factor, K
s, to account for material data scatter in fatigue life estimations was assessed as part of a series of LCF models using both constant- and variable-amplitude loads. Work has also been undertaken to correlate the various methods examined with real-world data from component fatigue tests. The results indicate that the minus 3SD from standard error method provides the best correlation between analytical models and the statistical lower bound, B
10, life obtained from laboratory tests.