This paper presents two turbomachinery-specific methods for profile tolerance assessment of compressor airfoils that process 3D scan data. This optical inspection technology digitizes the entire surface of the part into a triangulated mesh, which is aligned to nominal geometry and then processed to extract densely arranged profile sections.

For the reverse engineering method, the profile sections are decomposed into thickness and camber distributions. These distributions and the camber line are used to identify the profile parameter vector of the reverse engineering model. The deviation of the actual geometry is obtained by subtracting its parameters from those of the nominal geometry. Parameter-span graphs reveal airfoil shape deviation and allow for quantification of blisk scatter. The design-like parameters are meaningful and enable an intuitive engineering judgement of the actual geometry deviation.

The profile tolerance assessment method utilizes the camber line from the reverse engineering method to elegantly check against variable profile tolerance limits. The actual section is best-fitted to its nominal counterpart and assessed regarding its deviation relative to the allowed local tolerance. This ratio is plotted in a developed view summarizing the result of the profile tolerance assessment for the whole airfoil in a single graph. Thus, the condensed results allow for effective utilization of the high-resolution in airfoil sections.

Ultimately, the paper widens the view from one airfoil towards the assessment of the entire blisk. The blisk data is presented by statistical processing of deviation fields of all airfoils, in terms of mean and standard deviation. These statistical quantities are plotted onto the airfoil contour to e.g. represent the average airfoil thickness of the blisk. The standard deviation plot points to airfoil sections of larger geometric scatter and reveals areas of a non-robust manufacturing process.

This content is only available via PDF.
You do not currently have access to this content.