A probabilistic model for LCF

Abstract

Fatigue life of components or test specimens often exhibit a significant scatter. Furthermore, size effects have a non-negligible influence on fatigue life of parts with different geometries. We present a new probabilistic model for low-cycle fatigue (LCF) in the context of polycrystalline metal. The model takes size effects and inhomogeneous strain fields into account by means of the Poisson point process (PPP). This approach is based on the assumption of independently occurring LCF cracks and the Coffin-Manson-Basquin (CMB) equation. Within the probabilistic model, we give a new and more physical interpretation of the CMB parameters which are in the original approach no material parameters in a strict sense, as they depend on the specimen geometry. Calibration and validation of the proposed model is performed using results of strain controlled LCF tests of specimens with different surface areas. The test specimens are made of the nickel base superalloy RENE 80.

A log-normal format for failure probability under LCF: Concept, validation and definition of design curve

“…Socie [14,15] provided a complete description of the probabilistic model for LCF and how to estimate the prospective life distribution for a notched member with a Monte Carlo simulation. More recent papers develop a probabilistic model based on Weibull distribution for modelling the dependence of LCF life on size [16,17].…”

Evaluation of component‐similar rotor steel specimens with a local probabilistic approach for LCF

“…Hence, a thorough size effect consideration is an essential task for LCF life prediction. Additionally, steels are very well fulfilling the requirement of the locally independent crack initiation sites for the probabilistic model by Schmitz et al 11,12 which is used for the examination in this paper. They are also well described by FEA models using isotropic material laws.…”

“…From this elastic‐plastic stress field, all local strain values ε a at the surface are computed with the Ramberg‐Osgood equation. The strain and temperature fields ε a ( x ), T ( x ) are the input for the surface integral that calculates the Weibull scale parameter of the LCF crack initiation distribution for the specimen under a given load . Two predictions of the crack initiation life distributions were calculated.…”

“…6,7 That is why research activities have taken probabilistic LCF modelling approaches for rotor steels and forged Ni-base superalloys into their focus. [8][9][10] Applying the local probabilistic model described in Schmitz et al 11,12 to a typical rotor material and geometry was hence a logical step for validation and size effect studies. On the one hand, real rotor disc geometries often have a large surface, which can imply a negative statistical size effect impact on the risk for crack initiation.…”

“…This paper focuses on modelling the statistical size effect for crack initiation in rotor disc cooling borehole features with the local probabilistic model. 11,12 Component-similar specimens made of X12CrMoWVNbN10-1-1 are used for validating different prediction scenarios. Using single batch material parameters results in an approximately correct size effect shift prediction.…”