Towards a rapid assessment of turbine blade performance on wing

Abstract

Currently decisions to replace gas turbine components tend to be based on visual inspection and/or an appropriate component age derived from an analysis of the nominal design. However, inspection processes of gas turbine engines on wing are continuing to improve in fidelity with the development of, for example, borescopes capable of capturing images, videos and laser point measurements. As the wealth of on wing inspection data increases there is a gap in how this data are utilized to make service decisions. Similarly, basing service life on the nominal design ignores geometry variations that can occur due to both manufacture and through-life degradation and the growing amount of geometric data from high resolution 3D scans of components post manufacture and at the end of life. The following article proposes a framework for the on-wing prediction of turbine blade performance and demonstrates the progress made towards this. High resolution 3D scans of turbine blades post manufacture and at the end of their life are used to capture both blade manufacturing variation and typical through life degradation. Parametric representations of this damage, in terms of location and scale, are mapped on to a surface mesh of either the nominal or as-manufactured blade and fed into a CFD analysis. The result is a large database of damaged turbine blades with associated performance data which is then used to construct a neural network model which could be used to predict the performance of a blade on wing. While currently focusing on turbine blade efficiency the framework could be expanded in the future to eventually predict overall blade life.

More information

Identifiers

ORCID for Jennifer Forrester: orcid.org/0000-0002-6257-4603

ORCID for David Toal: orcid.org/0000-0002-2203-0302

Catalogue record

Export record