An image processing technique for erosion analysis on compressor blade geometry

Abstract

Particle ingestion into aircraft engines is a major concern due to its physical- and performance-related implications; largely due to compressor blade material erosion. Current methods for monitoring this erosion (such as visual tools) lack adequate reliability and consistency, while only providing qualitative insight into changes in blade geometry. This study presents a novel application for image processing to assess compressor blade erosion using two methods: 1) quantitative derivation of total chord, blade thickness, leading and trailing edge radii, and area at the blade tip using image binarization and a pixel-to-inch scaling factor and 2) qualitative visualization of material loss on the blade surface by aligning pre- and post-ingestion images of a rotor blade using feature-based detection. These methods are tested on a Rolls-Royce M250-C20B turboshaft engine, in which full-scale engine testing is performed to study particle ingestion effects on performance. The results showed a <1% uncertainty for blade geometry measurements, with the total chord also having a 1% difference compared to the value collected from a 10 µm resolution caliper. Furthermore, image registration showed substantial span-wise leading edge erosion at the 1st-stage rotor and the start of trailing edge erosion at the tip - which amounted to a ∼13% decrease of total chord - after a particle exposure of 1.3 hours and a dose of 2.6 kg of 100 µm sieved quartz. These erosion patterns showed results comparable to those of other studies on compressor blade particle erosion in the literature. The findings found in this study lay the foundation for image processing to be used as an accurate alternative to traditional aircraft engine inspection methods.