Quantitative estimation of nonmonotonic residual stress depth-profiles using an extended Kypris-Jiles model of the magnetic Barkhausen noise spectrum

Abstract

Using nondestructive techniques to quantitatively estimate residual stresses along the depth is necessary to improve the ability to predict the real fatigue life of pieces for many applications. Magnetic Barkhausen noise has been proven to successfully estimate the residual stress at the surface produced by machining, plastic deformation, phase transformation or surface treatments such as shot peening, also allowing one to obtain information of the residual stress depth-profile in shot peened pieces which presented similar depth-profile shapes. However, residual stress depth-profiles with nonmonotonic or different shapes have not been successfully estimated. In the present study, an extended approach is developed in order to estimate these stresses independent of the shape of the residual stress depth-profile. The approach proposed here improves an existing model of the Barkhausen noise spectrum (Kypris-Jiles model) by adding the effect of the attenuation of the applied magnetic field on the Barkhausen noise. This extended approach is used to estimate the residual stress depth-profiles of samples with different depth-profiles using a calibration process. The approach is validated by estimating the residual stress depth-profiles, with errors smaller than 70 MPa in a depth of 130 mu m, in all the samples studied