Statistical analysis of residual stress determinations using neutron diffraction

Within Task Group 1 (TG1) of the European Network on Neutron Techniques standardization for Structural Integrity (NeT), residual stress determinations by neutron diffraction were performed on a single weld bead on a plate (BoP) specimen by several laboratories. By following a pre-defined protocol, residual stress distributions have been estimated in the plate. Each residual strain/stress instrument at each institute have their own characteristics and their own methods by which the data is analyzed; however, the final estimate of the stress distributions are expected to be ‘the same’ to within the quoted uncertainty. This implies that this quoted uncertainty for each stress determination should be realistic.For specific measurement locations in the BoP specimen a modified Bayesian approach was used to obtain an average stress. This special approach is less affected by outliers, although care must be taken as it has been observed that in the case where there are few data points in the average, a strong bias can occur towards data points with a relatively small quoted uncertainty. In the data analyses used here, a ‘fairer average’ less influenced by such bias is implemented.Presented here is an overview of the main features that influence the stress and associated uncertainty determination, such as reference values, choice of moduli values, grain size effects and positioning etc.