Aviation has always been a powerful engine of innovation, and advanced control technologies have been at the forefront of enabling the industry to aggressively pursue performance targets in reliability, safety, efficiency, and environmental impacts. Recently, a new control-enabled advance in onboard fault detection has been developed and is now deployed on the Airbus A350 fleet.
In flight control systems, one of the anomalies to detect is termed the oscillatory failure case (OFC). This failure is an abnormal (small amplitude) oscillation of a control surface due to component malfunction in control surface servo-loops. This signal, of unknown amplitude and frequency, can propagate downstream from the control loop to the control surface (e.g., an aileron, elevator, or rudder), thereby causing vibrations in the airplane structure and producing structural loads that can reach levels incompatible with structural design objectives. There is no impact on the aircraft trajectory and control, but for structural design objectives this fault case must be properly taken into account. In other words, the ability to detect or not the targeted amplitudes can lead to add or to remove structural reinforcement, impacting the aircraft weight. Improving the OFC detection performances can thus lead to optimize the aircraft weight which in turns help decrease the aircraft environmental footprint (e.g., reduced fuel consumption).
The innovation story started earlier than the A350, with the superjumbo A380. Because of the use of new-generation actuators and more stringent load requirements, it was not possible to equip the A380 with legacy OFC detection strategies, which mainly relied on basic signal processing techniques. A basic model-based fault detection and isolation (FDI) approach was developed to cover OFC detection on all control surfaces. This “analytical redundancy” technique produces a fault indicator defined as the difference between the measured control surface position and an estimated position. A nonlinear hydraulic actuator model is used to estimate the position. In the A380 implementation, in order to reduce the computational burden, some model parameters were fixed to their most probable values (e.g., hydraulic pressure, actuator damping coefficient, etc.).
In order to improve this elementary model-based approach and to be compliant with more stringent load requirements (as well as a dedicated flight control system architecture), a joint parameter and state estimation technique has been developed for the A350, in partnership with University of Bordeaux, France.
The online physical parameter estimation of the actuator model allows for parameter variations during the aircraft flight and de facto improves the model accuracy. The estimation process is based on a nonlinear local filtering algorithm, based on robust control theory and shaped and adapted to the real-time constraints. The decision-making step is similar to the one used in the A380, meaning the use of an “incremental approach” suitable for critical embedded systems. The whole strategy allows for smaller fault amplitudes to be detected at an earlier stage compared with conventional systems. As aforementioned, the ultimate benefits include weight saving because of structural design optimization (not being compliant with load objectives would have led to add structural reinforcements).
This new FDI algorithm went through extensive V&V (Validation & Verification) activities, before certification and entering commercial service. The inaugural commercial flight of an A350 aircraft took place on January 15, 2015, between Doha and Frankfurt.
The main lessons learnt during this successful story we would like to emphasize here are the following:
- In the civil aviation industry, new technologies are only adopted when there is a clear need in terms of cost, performance or operational reliability benefit that cannot be adequately addressed through conventional techniques. In all cases, the level of safety of the aircraft must not be compromised. Introducing structural modifications to the in-service solutions entails risk and may require up to several years of V&V activities and maturation. This is also especially why an incremental approach is often suitable.
- To be assessed as worthwhile and to operate with the aircraft’s existing on-board systems, theoretical methods must be adequately shaped and adapted at an early design stage.
- Good average performance is of course necessary but it is by no means sufficient. The critical element is the achievable performance and robustness in extreme, unusual, non-standard and unexpected flight situations.
- A selected few, easy-to-tune, high-level parameters are decisive for the long-term use of an advanced solution during V&V activities.
- A major barrier is the certification of a new technique, particularly if it is structurally different to the in-service solutions. Here, the V&V activities play a crucial role, highlighting the importance of the above point 4.
- Control engineering can enable optimized solutions for critical and performance-sensitive industry sectors such as commercial aviation, provided that researchers expend the time and effort necessary to understand the application domain in depth. In particular, new theoretical paradigms must be adequately shaped and adapted early at the design stage.
This development is the result of a collaboration between the IMS lab, Bordeaux University, France, and the Airbus design office in Toulouse, France. It thus also serves as a case study for successful and impactful industry-academic collaboration.
Philippe Goupil, Airbus, Toulouse, France
Prof. Ali Zolghadri, Bordeaux Univ. – CNRS, France
More information on the development of the new FDI approach, including algorithmic details, the theory-to-practice process, and the industry/university collaboration, can be found in the following published sources:
- Goupil P. and Marcos A. (2014), The European ADDSAFE project: Industrial and academic efforts towards advanced fault diagnosis. Control Engineering Practice, vol. 31, October 2014, pp. 109–125.
- Zolghadri A., D. Henry, J. Cieslak, D. Efimov, P. Goupil (2014). Fault Diagnosis and Fault-Tolerant Control and Guidance for Aerospace Vehicles, from theory to application. Springer, Series: Advances in Industrial Control. 2014.
- Goupil, P. (2011). AIRBUS state of the art and practices on FDI and FTC in flight control system. Control Eng. Pract., vol. 19, no. 6, pp. 524–539.
- Goupil, P. (2010). Oscillatory failure case detection in the A380 electrical flight control system by analytical redundancy. Control Engineering Practice, 18(9).
- Zolghadri A.J.Cieslak, D.Efimov, D.Henry, P.Goupil, R.Dayre, A.Gheorghe, H.Leberre (2015). Signal and model-based fault detection for aircraft systems. IFAC SAFEPROCESS 2015, Paris, France.
- Zolghadri A. (2018). On flight operational issues management: past, present and future. Annual reviews in control. Volume 45, 2018, Pages 41-5.
- Zolghadri, A., Le Berre, H., Goupil, P. Gheorghe A., Cieslak, J., Dayre, R. (2016). Parametric approach to fault detection in aircraft control surfaces. AIAA Journal of Aircraft. Vol. 53, No. 3, pp. 846-85.