Comparative Dimensional Accuracy of Dovetail and Airfoil EDM Defects on Gas Turbine Blades

Gas turbine blades operate under severe thermomechanical loads therefore the geometric examination of damage features must be precise and consistent. This work evaluates a non contact digital image-processing method for measuring artificial defects produced by die sinking electrical discharge machining (EDM) on two blade surfaces: a serrated dovetail joint and a smooth-contoured airfoil. Six defects were manufactured (three per region) with depths of 0.5, 1.0, and 2.0 mm. Twenty repeated measurements per defect (120 total) were used to quantify accuracy (relative error) and precision (standard deviation), followed by non-parametric and variance tests, and validation against a 0.001 mm-resolution digital micrometer. Length and width accuracies were typically 97.5–99.9%, and overall accuracy exceeded 95% for planar features; the median accuracy did not differ between regions (Mann–Whitney test, P-value 0.577). Precision was significantly lower on the dovetail (Levene test, P-value <0.001), attributed to serration-induced shadows and specular reflections that destabilize edge detection. Micrometer validation showed minimal deviation for straight-line features, while on the airfoil the optical method captured a more representative projected arc length than the micrometer’s chord measurement. These results support digital image processing as an efficient, reliable alternative for non-destructive inspection of complex turbine geometries.