Open Access

ARTICLE

Probability Methods for Estimation of Cleavage Fracture Toughness from Small Data Sets

by R. Moskovic¹, P. E. J. Flewitt¹

1 British Nuclear Group (BNFL), Engineering and Technology Services, Berkeley Centre, Berkeley Glos. GL13 9PB, UK.
2 Also Interface Analysis Centre, University of Bristol, Bristol. BS2 8BS, UK.

Structural Durability & Health Monitoring 2005, 1(1), 83-94. https://doi.org/10.3970/sdhm.2005.001.083

Abstract

Consideration of the structural integrity is one of the inputs when evaluating potential solutions to plant problems. Structural integrity assessments of components forming the pressure boundaries of nuclear plant evaluate safety margins against cleavage fracture. These assessments consider the reserve factors between the applied stress and fracture toughness of the material as well as temperature margins between the operating temperature and the temperature at which the steel is ductile as defined by upper shelf behaviour. To carry out these structural integrity assessments, estimates of cleavage fracture toughness are required. The approach presented in this paper allows for differences between cleavage fracture toughness properties of different materials associated with material to material variability. Mean cleavage fracture toughness properties are described as a function of temperature, section thickness and ductile crack growth. The random scatter is described by standard statistical probability distributions with variance a constant percentage of the mean cleavage fracture toughness. Micromechanisms of cleavage fracture and numerical modelling show that fracture toughness is determined by the microstructure of the steel which also defines the work hardening characteristics, the yield stress and its temperature dependence. On this basis, temperature and thickness dependence, and scatter of cleavage fracture toughness are specific to a given type of material. This paper presents methods for estimation of cleavage fracture toughness from small data sets using simple statistical tools and cleavage fracture toughness curves fitted to large databases for materials with the specification similar to the small data set. Applications of this methodology are presented. The methodology is validated by comparing the predicted values with data that were not used in the analyses. One of the applications illustrated involves predictions of cleavage fracture toughness for the neutron irradiated condition. This is validated by comparing predictions with data measured on specimens removed from ex-service reactors.

---