Influence of analysis parameters in the evaluation of residual stresses with incremental hole-drilling method

Titanium grade 5 alloy is wide used in aerospace applications due to the good mechanical properties at high temperatures, its good corrosion resistance and especially to the very promising strength to weight ratio. Consequently, in the last years its use is increasing mainly for the manufacturing of welded joints. In order to better characterize this material many mechanical tests are carried out, including residual stress measurements that provide important information about the state of material subjected to either thermal or mechanical treatments. These kind of measurements could be carried out both by using X-Ray diffraction and Hole Drilling Method, which are the only two standardized methods commonly applied. The possibility to successful use X-Ray method for in depth residual stress measurements is strongly reduced by two reasons: the size of the area subjected to chemical etching that is large enough to put away the analysis point and to the nature of material itself particularly subjected to oxidation during the chemical treatment. For this reason, Hole Drilling Method is commonly preferred: it is capable to measure the strain relaxation of components by drilling a hole millimeter along the material thickness and measuring the strain in proximity of it. This technique provide results according to different algorithms, known in scientific literature, that could deeply differ one from each other due to the mathematics they are based on. Accuracy of the results is strongly dependent on the measurement procedure and on the algorithms adopted to calculate stresses starting from strain measurements. Small measurement errors on the strain can imply big errors in the calculated stress. In this study, the influence of strain measurement interpolation is evaluated on titanium grade 5 butt laser welded plates, by critically comparing two different measurement procedures, namely integral and HDM, both implementing in one commercial system. Moreover, the effects on the accuracy of the proper choice of the number of depth calculus increments, the step distribution as well as the influence of the plate thicknesses are investigated according to ANOVA analysis.