Disk crack detection and diagnosis for gas turbine engines

Prognostics and health management (PHM) is believed to become a built-in capability for the next generation military aircraft, e.g., Joint Strike Fighter (JSF). Health monitoring for gas turbine engines is certainly an essential component for such PHM systems. This paper presents a technique of detecting and diagnosing disk cracks of gas turbine engines under spin testing by analyzing the shaft displacement data. Using synchronous averaging and spectral analysis, an approach to acquiring the unbalance index (UI) is firstly established. For crack detection, a complex UI is obtained for each spin test cycle near the top speed during the speed-up process, and its amplitude and phase are trended against cycle numbers. Results have shown that the trending plots can be effective in detecting changes of unbalance status due to cracks in a rotating disk. For crack diagnosis, detail analysis is then carried out on the UI for every speed of a whole spin cycle using the Nyquist plot, i.e., the real part versus the imaginary part for the full speed spectrum in the spin cycle. Results from an engine disk spin test have shown that the Nyquist plots can be used to extract valuable diagnostic information on the relationship between initial mass unbalance and crack induced unbalance, and probably even on the propagation of cracks.