Come fly with me

Discover fly-by wire, the technical wizardry keeping modern aircraft in the air.

When you are sitting comfortably at 30,000ft, all that's stopping you from spiralling at several hundred miles an hour into the land or water below is a technology called fly-by-wire (FBW), a suite of hardware and software control systems that has revolutionised powered flight over the past 20 years.

Although its introduction into civil airliners such as the Airbus A320 and Boeing 777 has been widely touted, FBW was first exploited by the military, which is where much of the running in its research and development is still being made.

It's widely known that controlling an aircraft in flight is a matter of adjusting the control surfaces - the ailerons on the wings and the rudder and elevators on the tail. The wings also have flaps, set next to the ailerons, for extra lift on take-off and extra drag when landing.

Traditionally, these were operated mechanically from the control column in the cockpit, via cables routed through the aircraft and using pulleys, cranks, wires and, for hydraulically-assisted controls, hydraulic pipes. They're still in widespread use, although usually now on small aircraft.

But all those mechanical components make them heavy and, like other safety-critical technology, they usually need back-up systems in case of failure. This increases their weight further - not ideal in a vehicle you're trying to get off the ground, and not great for fuel consumption. They're not that great either at compensating for changes in aerodynamic conditions, and they won't necessarily stop an aircraft going into a stall or a spin. And pilot-induced oscillation (PIO), where the pilot inadvertently overcorrects commands in opposite directions to try to counteract the aircraft's reaction to the previous command - particularly dangerous during landing - can still occur.

More modern systems

Modern FBW systems do away with all this, and by means of replacement control-surface actuators are operated by onboard computers. Instead of a physical connection via cables and pulleys, the cockpit controls now operate signal transducers that generate the appropriate commands, which in turn are processed by the computer software to determine how best to move the actuators at each control surface to provide the desired response.

But while these systems eliminate the complexity and weight of a mechanical circuit, the main concern still has to be reliability. Mechanical or hydraulic control systems tend to fail gradually, whereas if the computers crashed so could the aircraft. So virtually all FBW systems incorporate triply or quadruply redundant computers and signal wiring to reduce the possibility of a general failure to vanishingly small levels, and the software has to be developed according to various standards (see Standards box).

As well as in-flight control, the software allows an aircraft's handling characteristics to be designed precisely, within the overall limits of what is possible with the aerodynamics and structure of the aircraft. For example, it can prevent stalls, spins and PIOs by preventing pilots from exceeding preset limits known as the aircraft's flight envelope.

These internal systems are only part of the picture, however - to do their job they need to know what's happening in the atmosphere outside, so they need information on engine thrust, adjusting that too when necessary, and data from external sensors on pressure and airflow to calculate the true velocity vector of the aircraft. In essence, an FBW system is a set of feedback loops that keep the aircraft stable and operating within its flight envelope.

Range of benefits

The technology also brings a range of benefits for the aircraft designer and pilot. For the designer, it allows the natural aerodynamic stability of the aircraft to be relaxed, slightly for a transport aircraft and more for a military plane, which means the control surfaces can be made smaller. In fact, for military aircraft it's not unusual any more for a fighter to be deliberately designed to be inherently unstable to give it a high degree of manoeuvrability. For the pilot it reduces workload, freeing them up for other tasks, and in civil airliners it improves economy because of the reduced weight of the aircraft.

One of the leading developers of the technology is BAE Systems, and Chris Fielding, flight control systems technol-ogist at the company, explains how these systems are designed.

He says: "We start by gathering aerodynamic data from the wind-tunnel testing. From that, we develop a mathematical model for the aircraft simulation.

"We then build in the feedback loops and control laws to cover the flight envelope for that particular aircraft, as well as the mathematical models for the sensors and actuators. The feedback gains and command-path filters then have to be tuned to get the best handling characteristics for the aircraft.

Going into further detail, Fielding says that in order to design the control laws, a grid of 'operating points' has to be selected across the envelope. This results in a set of localised controllers for those points, which then have to be integrated to cover the envelope.

This integration is usually carried out using gain scheduling which, explains Fielding, is where an inverse schedule is applied to compensate for dynamic pressure changes in order to maintain a 'constant loop gain'; without it, the feedback loop would become unstable, eventually leading to an uncontrollable aircraft. The information for this scheduling usually comes from the external pressure and airflow sensors.

Unstable aircraft

Again, usually, the control system is designed for a baseline configuration, using a nominal set of aerodynamic data and parametric tolerances based on past project experience and uncertainties in the available wind tunnel data. Other significant factors that need to be taken into account in the design include fuel state, airbrakes, undercarriage operation and ground handling - all of which can have a significant effect on the design in terms of stability, handling and airframe loading.

These are general principles and yet, says Fielding, each type of aircraft will have its own flight control system architecture. "From our point of view the Tornado, for example, is an aerodynamically stable aircraft, and has fixed control laws for whatever external stores it is carrying, such as under-wing or under-fuselage tanks," he says. "But the Typhoon, say, which is unstable, has automatic gain selection within its control laws, depending on the 'store group' being carried, to give it design robustness and optimise its handling."

And there are broad differences between these systems when used for military or civil applications. "Military craft carry military personnel, obviously, sometimes at supersonic speeds, so the prime design driver of these aircraft is mission effectiveness," Fielding says. "Civil aircraft, however, carry passengers at subsonic speeds, and their prime design driver is economy."

But Fielding points out that, whatever the application, reliability and safety must always be the fundamental requirements. "Our systems use what's called 'similar redundancy' - with multiple parallel computing lanes using the same hardware and software. Civil aircraft, by contrast, use 'dissimilar redundancy' - different hardware and software in each computing lane.

"Also, the actual level of redundancy depends on the application," he says. "We use triplex as well as quadruplex systems - three or four multiple computing lanes. So, for controlling an aircraft's angle of attack, for example, we might use a triplex arrangement and for gathering air-speed information, say, it might be quadruplex."

System delays

Inevitably, all this hardware - the sensors, actuation systems, and digital computers and their interfaces - introduces lags and delays into the system, which tend to reduce the aircraft's stability margins and impose limits on its performance. And since, with current technology, the control laws are usually implemented through the computer software, digital processing effects have to be taken into account as well.

So, Fielding says: "We minimise the inevitable lags and delays within the closed-loop system by using a high clock speed - typically of the order of 200Hz for processing control algorithms. We also use anti-aliasing filters to remove the high-frequency element of the feedback signals, to prevent aliasing."

This, in general, is the current 'state of the art', although development and research of course continues apace.

"In terms of development," says Fielding, "we're seeing increasing miniaturisation in the sensors, while flight systems computers are delivering higher processing throughput and greater data storage. And on the actuation side, there is growing use of electrical rather than electro-mechanical devices for the control surfaces, which is helping us to reduce the weight and maintenance costs."

Further down the line he also outlines developments such as total vehicle management systems that might integrate the functionality of hitherto separate airframe systems; the use of laser-based air data systems, particularly on stealth aircraft, and the use of thrust vectoring, and novel control methods such as nose suction or blowing.

More complex systems

Dr Mark Lowenberg, head of the Department of Aerospace Engineering at the University of Bristol, says: "We're now seeing increasingly complex FBW systems on modern manned aircraft - both civil as well as military - and this has led to a difficulty in certifying such systems. The clearance of the flight-control laws for a modern fighter aircraft, for example, can be incredibly time-consuming and hence expensive; as a result there is research focused on methods that can perform the necessary analysis to clear the system as effectively as existing methods but far more quickly.

"Another thread of research being tackled in our department and elsewhere deals with enhancing the aircraft design process such that all aspects of the control system - physical as well as control law logic - are accounted for much earlier in the design process," he adds. Socioeconomic factors are creating the need for industry to procure aircraft that are optimised in a truly holistic sense. "The idea is to integrate the consequences of a range of onboard systems in terms of weight, power consumption and other influences to arrive at a better optimised outcome."

Although debate still surrounds the date of the world's first powered, heavier-than-air flight, it's beyond dispute that aeronautics has come along way in the past 100 years or so.

1956 saw the earliest configuration of an FBW system fitted to an RAF Avro Vulcan bomber. The first FBW airliner wasn't seen until 1969 during the maiden flight of Concorde.

1974 saw the maiden flight of the F-16 Fighting Falcon, the world's first aircraft designed to be aerodynamically unstable and hence controlled by FBW and in 1977 digital FBW was used on Space Shuttle in free-flight approach and landing.

The first airliner with digital FBW controls, the Airbus A320, had its maiden flight in 1987.

While we may marvel at the technological wizardry keeping stealth bombers and the Airbus aloft, advances such as hypersonic aircraft and private commercial space flight coming over the horizon suggest this is only the beginning for FBW. 

Recent articles

Info Message

Our sites use cookies to support some functionality, and to collect anonymous user data.

Learn more about IET cookies and how to control them