Genetic algorithm optimization of the hole-drilling method for non-uniform residual stress fields

摘要

The hole-drilling method, coupled with the influence functions method, is a versatile technique for measuring residual stresses. The latter determines a relationship between residual stress and measured relaxed strains for any experimental configuration and material. However, several parameters which describe the functional basis for the residual stress must be defined. This paper presents a procedure, based on the least squares method and genetic algorithms, optimally to choose these parameters to reduce the effects of noise and uncertainty in experimental measurements. The performance of the method was tested against finite element simulations for typical applications such as shot peening, showing that the method is flexible and accurate and allows the reconstruction of residual stress at deeper positions than other techniques.