FDI applied to affordable digital fly-by-wire flight control systems

Project members: R. Hallouzi, R. Babuška, V. Verdult, M. Verhaegen

Sponsored by: European Community GROWTH project ADFCS-II

Employing analytical redundancy based on mathematical models rather than hardware redundancy based on redundant hardware is a key issue in current research in the area of aircraft fault detection and identification (FDI). In this manner cost, weight and complexity can be reduced. The objective of this project is to apply model-based FDI methods to small commercial aircraft (see Figure 10). In previous work within the ADFCS-II project research has mainly focused on sensor faults. However, in order to achieve complete condition monitoring in aircraft, the development of FDI methods for actuators is an important step.

The FDI methods we propose are able to distinguish between total and partial actuator faults in an aircraft and can estimate how large these faults are. Partial faults are characterized by the fact that the control surface is still functioning, but only with reduced power (loss-of-effectiveness). Total faults are characterized by the fact the control surface is stuck at a certain position and does not react on the control input (e.g. stuck-in-place or hard-over fault).

The use of well-proven linear FDI methods is attractive in many aspects. However, aircraft are non-linear systems. In order to still be able to apply these linear FDI methods, multiple linear models have to be scheduled. These models are linearized at certain points in the flight envelope of the aircraft. The challenge is to use as few linearized models as possible to cover the whole flight envelope.

Figure 10: Galaxy business jet (left) and a schematic overview of a control system with model-based FDI (right).