Development of a Plasma Spray Process Monitoring System through Aeroacoustic Signal Analysis

Abstract

Plasma spray coatings are vital to the capabilities of jet engines. They allow engines to operate at combustion temperatures that would otherwise melt the superalloy components. Coatings tighten clearance between rotating components, increasing engine compression. They prevent chemical attack and physical erosion. Plasma spray coatings are imperative to the durability and efficient operation of the modern jet engine. In this application coating material property variation has a significant cost. In addition to the variation inherent in the process, some of the biggest contributors to coating property variation have been traced to spray gun nozzle wear and powder feed variation[3, 4].

Presented here are multiple methods utilizing flow induced acoustic signals to quantify noise parameters, measure component wear, diagnose the plasma spray process and detect coating property deviation. Methods have been developed for offline and online analysis of components in addition to online process analysis. These include characterization of nozzle wear by throat roughness measurements and nozzle casting, offline detection of nozzle wear by attenuation of discrete tone generation and broadband signal variation, and offline measurement of powder port wear by jet screech frequency variation. Online methods include pre-ignition nozzle degree of wear measurement by discrete frequency changes; online parameter change detection, process deviation detection with potential source identification, as well as variation in coating property detection by broadband acoustic signal changes.

Offline methods allow for 100% accurate new nozzle manufacturer identification. By the same test nozzle wear state can be predicted with over 95% accuracy with the potential for a degree of wear determination. Internal diameter changes of less than 10 microns can similarly be detected. Analysis of online plasma spray acoustic signals as described here can distinguish nozzle state and powder feed variation with over 90% accuracy.

The capabilities developed here will aid in plasma spray process variation detection and contribute to identifying the source of this variation. This will improve coating quality and consistency, reduce failures, lower operational costs and ultimately make jet engines more economical, safer, and more fuel efficient with significant environmental and financial cost reduction.