Fly-by-wire and interactions

Aircraft Engineering and Aerospace Technology

ISSN: 0002-2667

Article publication date: 1 October 1999

428

Keywords

Citation

Ford, T. (1999), "Fly-by-wire and interactions", Aircraft Engineering and Aerospace Technology, Vol. 71 No. 5. https://doi.org/10.1108/aeat.1999.12771eaf.001

Publisher

:

Emerald Group Publishing Limited

Copyright © 1999, MCB UP Limited


Fly-by-wire and interactions

Fly-by-wire and interactions

Keywords: Flight controls, Aircraft

Over the years that fly-by-wire (FBW) systems have become established in an increasing range of civil transport aircraft, attention in reports and at conferences has addressed the number of control law functions that significantly interact with the structure. Some of the flight control law modes affect aeroelasticity and have to be modelled when analyses are carried out.

In Airbus aircraft control law modes are either autopilot or flight normal laws. The former automatically maintain parameters such as altitude, descent rate, etc., over relatively long periods of time. The low frequency response of the aircraft is modified when the autopilot mode is active and a different response to turbulence is to be expected other than that when the aircraft is under another control law or without an electronic flight control system (Figure 1).

Figure 1 The influence of the autopilot and normal control laws on the low frequency pitch response of the Airbus A320.

The normal control laws reduce pilot workload by providing autotrim and protection against exceeding operational limits. In a FBW aircraft when the pilot acts on the stick, an aircraft response is directly ordered (e.g. roll rate). The pilot order is sent to the computer and the aircraft response to this order is fed back to the computer. The computer then gets the difference between the pilot order and the aircraft response, and processes the control surface deflection required to get the aircraft response as ordered by the pilot, on a short-term basis. The computer also carries a high speed integrator which achieves a long-term stabilization function by processing the deflection of the control surface required to maintain the aircraft's flight path, stick free. The computer also ensures that the operational limits are not breached and for this to be achieved, additional information is required which is obtained via feedback parameters, such as those concerning acceleration rates.

In this way the feedback and high speed integrator achieve stability, stick free, and manoeuvrability, on pilot demand. The control surfaces are constantly active, even when the stick is free. In the event of computer or sensor failures, the flight normal laws may be reconfigured to either an alternate law (ALTN), or direct law and these other laws have to be considered when undertaking aeroelastic analyses. As far as the normal control laws are concerned aeroelastic stability is not of major concern since the gains and actuation rates are modest. The presence of the various protection functions, however, has considerable influence on loads prediction.

These operational limits concern protection for pitch attitude, load factor, bank angle, stall, and overspeed. The purpose of these is to:

  • give full authority to the pilot in order to consistently achieve the best possible aircraft performance in extreme conditions (such as windshear, high turbulence, etc);

  • reduce the risks of overcontrolling or overstressing the aircraft; and

  • provide the pilot with an easy, instinctive and immediate procedure to obtain these aims.

Concerning pitch attitude, the pitch normal law will keep the aircraft roughly in level flight, autotrimmed, while it decelerates. Then when VLS (minimum normal speed = 1.23Vs1g) is reached, the pilot should take action. If there is no action, the deceleration continues until the autotrim stops, which occurs at a predetermined angle of attack (Alpha Prot). At this point if there is still no action on the stick (or on thrust), the aircraft will gently descend keeping constant the angle of attack. If the pilot pulls stick aft, a higher angle of attack is ordered and if full aft stick is pulled, Alpha Max is ordered. This protection is an aerodynamic feature, but thrust is required to maintain the desired flight path. Other features are also present to help this aspect of protection.

For bank angle, on a commercial aircraft 30ú is not normally exceeded. If the aircraft exceeds the normal bank flight envelope, the pilot is immediately aware and the protection prevents any major upset or pilot mishandling to bring the aircraft into an uncontrollable state. It provides full authority to the pilot to achieve most efficiently any required roll manoeuvre.

Load factor protection has been influenced as a result of incidents and demonstrations which have shown that an immediate recovery at 2.5g provides a larger obstacle clearance than a manoeuvre recovery achieved hesitantly, with a higher G load reached later (typically more than 2 seconds delay). As a consequence, a G load limiter has been set in the normal (and alternate) law, which protects the aircraft against overstresses by maintaining it within its structural limitations. In short, the load factor protection associated to the high angle of attack protection gives full authority to the pilot to react immediately and instinctively for an evasive manoeuvre while minimising the risks of overstress or loss of control.

The high speed protection is based on the following considerations. The maximum speed/Mach of the flight envelope is defined by VMO/MMO. It may happen, in exceptional circumstances, that this speed is overshot without major impact on flight safety until design speed limits VD/MD are reached. Beyond these speeds (1.2 VD/MD) aircraft control problems and structural consequences due to high loads may be encountered. In order to protect the aircraft against such phenomena a positive nose-up G demand is added to the pilot demand on the stick; this being proportional to the amount of speed overshoot beyond VMO/MMO but limited to 0.75G. Thus, if a dive is achieved, stick free, the aircraft will slightly overshoot VMO/MMO and fly back into the flight envelope. Stick full-forward, the aircraft will significantly overshoot VMO/MMO without reaching VD/MD, and will stabilise at VMO + 15/MMO + 0.04. Thus, if a steep dive is required, the pilot has full authority to achieve the manoeuvre safely, and if a major upset is experienced, the high speed protection enhanced by the pitch protection minimizes the risk of potential structural damage.

Other structural implications

Two other control law functions that may be of structural significance are load control and stability augmentation. Load control has primarily been used on Airbus aircraft to reduce inner wing bending moments and hence to save structural weight. The A320 employs a gust load alleviation system to reduce the wing bending moments arising from turbulence to the level of those arising from manoeuvres. The A330/A340, on the other hand, employs a manoeuvre load alleviation system to reduce the bending moment arising from manoeuvres to the level of those arising from turbulence (Figure 2).

Figure 2 The effect of the vertical velocity error signal limit represented by an equivalent gain, on the low frequency pitch response of the Airbus A340 with autopilot active.

The redistribution of aerodynamic load is achieved through the deployment of ailerons and outboard spoilers. For gust alleviation where the gust is sensed via accelerometers on the fuselage, the surface deployment rates must be high. The combination of inetrial feedback, high gains and high actuation rates means that strong coupling is required between the structure and control system. For aircraft the size of the A340, manoeuvres are fairly slow events in comparison to turbulence encounters, and thus the demands on control surface actuation rates are modest. Consequently, aeroelastic stability is not an issue when dealing with manoeuvre load alleviation systems. However, all load control systems have to be considered when predicting aircraft loads. There is little doubt that these systems will reduce the targeted loads, but the effects on loads elsewhere are less clear.

Stability augmentation functions increase damping in certain aeroelastic modes, and such functions have been considered for the A340 in two forms. The first is a lateral damping augmentation function which was developed as a provision for increasing the damping in certain aeroelastic modes, although it became apparent later on in the design stage that it was not required. The other is a comfort in turbulence function incorporated subsequent to first flight. Flight tests suggested that the fuselage vibration modes were sensitive to turbulence (lowly damped). The purpose of the system is to increase damping in the relevant aeroelastic modes.

Aeroelastic analyses are required to define these systems and to assess the aeroelastic stability implications, since they directly aim to modify aeroelastic damping. A major challenge is achieving a good standard of aeroelastic model early in the design so that the need for such systems can be identified.

Various considerations

The Airbus electronic flight control system (EFCS) control laws are inherently non-linear, falling into several categories. It is not always necessary to model the non-linear properties of the control laws and for certain analyses it is not desirable because the most appropriate analysis technique does not permit it. A major concern over recent years has come from the airworthiness requirement for loads prediction due to continuous turbulance. Another non-linear dimension results from the gain scheduling required to achieve a common feel/handling quality system thoughout the operational envelope. Although this does not directly affect modelling techniques, it does reduce the possibility for interpolating and extrapolating to determine design conditions which, in turn, necessitates a greater amount of analysis.

As mentioned, non-linearity and the prediction of loads due to continuous turbulence are a problem since the properties of the control system have significant impact on the loads predicted. Further problems are due to the assumption of linearity in the definition of design load stated within the airworthiness requirements. Over recent years this has had major impact on the loads prediction effort.

Particular tasks include the effectiveness of the A320 gust load alleviation function to continuous turbulence, which had to be demonstrated. This necessitated the development of non-linear analysis techniques. Another concern is that the A340 with autopilot active shows a very large low frequency response when the autopilot is modelled ignoring its non-linear elements. Atmospheric turbulence has a high low-frequency content, so the interaction of the closed-loop response of the aircraft/autopilot system with the turbulence yields high loads. This, however, is unrealistic because a limit applied to the vertical error signal significantly reduces the response at the low frequency. It is therefore necessary to account for system non-linearity in the turbulence analysis. Because of these and other problems British Aerospace Airbus has developed very full nonlinear, continuous turbulence analyses which have been very successfully applied.

In summary, there are few problems in modelling control systems interaction with structures to a level of accuracy sufficient for structural design, but electronic control systems have had a major impact. This is because the non-linear aspects of these systems have provided many challenges in complying with the airworthiness regulations covering loads arising from continuous turbulence. Also, electronic control systems, with the object of minimising the variation in handling quality throughout the operational envelope, demand more extensive coverage of design conditions for aeroelastic analyses than a conventionally controlled aircraft. This is further compounded when the control system is non-linear. Efforts have been made to resolve these and other problems, such as definition and standards, and work continues on accuracy of predictions and development of the representation of unsteady aerodynamics in the aeroelastic model, particularly in the transonic region.

Information gratefully received from British Aerospace Airbus and Airbus Industrie.

Terry Ford

Related articles