An assessment of the effect of cutting welded samples on residual stress measurements by chill modelling

Abstract

In making residual stress measurements in welds, a common problem is that the samples are often cut before measurement, which may alter the stresses. To estimate the effects of cutting, a simplified method of modelling residual stresses in welds was developed, known as chill modelling, which requires limited material data and no welding process data. The method can be performed with elastic finite element analysis. The method does not predict the value of the stresses, just the relative reduction that occurs after cutting. The method was validated against published synchrotron measurements of welded plates, which were cut and remeasured a number of times. For welded plates only, an empirical equation was developed that predicts the change in stress after cutting.