Autonomous PHM with blade-tip-sensors: algorithms and seeded fault experience

Abstract

Blade tip sensors embedded into the engine case have been used for decades to measure blade tip clearance and blade vibration. Many sensing technologies have been used; capacitive, inductive, optical, microwave, infra-red, eddy-current, pressure and acoustic. These sensors generate data streams far greater than have been historically used in engine diagnostic units. Data streams of about 10,000 samples per second per sensor are about the minimum achievable, with some sensor front-ends delivering data streams of greater than 1Megasamples per second per sensor. In a PHM application, this data cannot be stored for later human analysis, but must be analyzed and discarded. This paper outlines autonomous algorithms for the real-time analysis of this data stream for PHM purposes. The application of these algorithms to several seeded fault tests is described. The need for a series of additional seeded fault tests is highlighted, for the purpose of maturing these algorithms prior to introduction into service.