Section Based Profile Tolerance Assessment of 3D Scanned Compressor Airfoils

This paper introduces a turbomachinery specific method for profile tolerance assessment of 3D scanned compressor airfoils.

A 3D scanner is used to digitize a blade or a blisk to obtain an STL mesh of the entire surface. The alignment of the STL mesh to nominal geometry is done via the component datum structure in either disc datum features (in case of a blisk) or contact faces in blade root (in case of a blade). The aligned SLT mesh is cut through the airfoil to extract profile sections. This is typically done in Cartesian coordinates to obtain few planar sections at the desired distance to the platform. In order to allow for a complete airfoil surface coverage, sections should better be extracted in the cylindrical coordinate system for a constant relative blade height or span (interpolated between platform and blade tip). The cylindrical section can be conformal mapped to 2D space by conserving either length or angle to then decompose the 2D profile section into thickness and camber distribution. Among other results also the intersection points of camber line with profile line, the inversely calculated leading edge (LE) and trailing edge (TE) point are obtained for each section. This inverse profile definition ultimately provides a non-dimensional airfoil description in relative height (or span) and relative chord (x/c).

Commonly the allowed profile tolerance reduces when getting closer to the edges (i.e. the leading edge) to account for the increased sensitivity of flow to geometric deviation in that area. The developed profile tolerance assessment method utilizes the discussed non-dimensional description to elegantly inspect for flexible profile tolerance criteria. Therefore sections are extracted from both the actual and the nominal geometry for the same span coordinates vector. Each actual section is best-fitted to its nominal counterpart while recording shift and rotation. For edges processing the LE & TE points are overlaid which may reduce the best-fit procedure to an angle adjustment. The best-fitted sections are assessed regarding the deviation of the actual coordinates to the interpolated counterpart of the nominal section. The local deviation is compared to the local tolerance by using the non-dimensional airfoil coordinate being the connecting element. Ultimately these local ratios or differences of deviation and tolerance may be plotted elegantly in a developed view showing the profile tolerance assessment of the whole airfoil in only one 2D figure. This plot allows the manufacturing engineer to quickly judge the overall airfoil deviation characteristic and suggests where to investigate the dataset in more depth.