Improved aviation safety and reliability can be achieved through enhanced in-flight diagnostic capabilities for aircraft gas turbine engines. Since faults in sensors, actuators, or components can lead the aircraft engine into undesirable operating conditions, it is critical to detect faults as early as possible so that the necessary corrective actions can be taken. A challenge in developing an in-flight fault-detection system is to make it adaptive to engine health degradation. Engine health degradation is a normal, expected process that all aircraft engines will undergo, whereas a fault is an abnormal, unexpected event. However, both health degradation and faults influence the engine output, from which the existence of faults is detected. If an in-flight fault-detection system cannot adapt to health degradation, it will eventually lose its diagnostic effectiveness.
Structure of the in-flight fault-detection system: y, measured engine output; ucmd, control commands; WSSR (weighted sum of squared residuals), fault indicator signal.
Long description of figure 1.
To address this issue, researchers at the NASA Glenn Research Center developed a hybrid Kalman filter (HKF). This new type of estimation technique combines the advantages of the linear and nonlinear Kalman filter approaches. The uniqueness of the HKF is its structure: a hybrid of a nonlinear onboard engine model (OBEM) and piecewise linear models. Because of this hybrid structure, the HKF possesses significant advantages that make this technique well suited for in-flight, real-time diagnostic applications.
The first advantage is that the reference health baseline of the HKF can be updated to the health condition of the degraded engine in a relatively simple manner. In-flight diagnostics is accomplished by processing the measured information relative to an established reference health baseline. As the aircraft engine degrades over time because of usage, its operating condition deviates from this baseline, causing health-degradation-induced shifts in the measurements. Such health degradation influences are compensated for by periodic updates to the reference health baseline. Although this baseline update is a necessary step, some diagnostic techniques have a structure that makes this update highly complex and impractical. Therefore, the structure of the HKF, which lends itself to a simple health baseline update, is a significant advantage.
The second advantage is the HKF capability to capture nonlinear, off-design engine operations. Aircraft engines undergo such operations as the engine control system responds to faults or non-fault-related factors, such as bleed air or horsepower extractions for aircraft services. This HKF capability is due to the utilization of the OBEM in which the nonlinear relationships between the engine parameters are embedded.
The third advantage is the numerical robustness of the HKF. An OBEM can be a concern for numerical stability depending on how it is implemented and used in the diagnostics process. In the HKF approach, the OBEM runs as a standalone engine model; it generates expected engine output for given control command input. On the basis of the information provided by the OBEM, the HKF estimates engine variables. Therefore, the numerical stability of the OBEM is not influenced by the estimation performance of the HKF.
The HKF-based fault-detection system was developed, and its diagnostic capability was investigated in a simulation environment. Extensive study indicates that HKF is a promising technology for aircraft engine in-flight diagnostics.
Kobayashi, Takahisa; and Simon, Donald L.: Hybrid Kalman Filter Approach for Aircraft Engine In-Flight Diagnostics: Sensor Fault Detection Case. ASME Paper GT2006-90870 (NASA/TM--2006-214418), 2006, pp. 745-755. http://gltrs.grc.nasa.gov/Citations.aspx?id=167
ASRC Aerospace Corp. contact: Takahisa Kobayashi, 216-433-3739, Takahisa.Kobayashi-1@nasa.govLast updated: October 15, 2007
Responsible NASA Official:
Gynelle.C.Steele@nasa.gov
216-433-8258
Point of contact for NASA Glenn's Research & Technology reports:
Cynthia.L.Dreibelbis@nasa.gov
216-433-2912
SGT, Inc.
Web page curator:
Nancy.L.Obryan@nasa.gov
216-433-5793
Wyle Information Systems, LLC
