A QUESTION ABOUT ENGINES, ASKED THIRTY YEARS AGO

The Axisymmetric Vectoring Exhaust Nozzle, known as AVEN, was developed by GE Aerospace in the early 1990s to answer a question that had begun to preoccupy fighter engineers: whether the engine itself could contribute to manoeuvrability. Military aircraft had until then relied almost entirely on aerodynamic surfaces (ailerons, flaps, rudders) which work well across much of the flight envelope but whose effectiveness varies with speed and operating condition. As demands on fighter agility rose, engineers began looking for a second source of control.

Thrust vectoring was the answer, and the first implementations were two-dimensional. Rectangular nozzles using relatively flat flaps could redirect exhaust in a single axis – pitch or yaw, but not both. They worked, but at a cost: vectoring in only one plane, significant added weight on the engine, flow losses, and bulky designs that complicated maintenance and reduced efficiency.

GE Aerospace took a different approach. Rather than converting the engine’s naturally circular exhaust into a rectangular nozzle, AVEN preserved the axisymmetric flow while enabling thrust vectoring in any direction. It retained the efficiency of a conventional variable-area convergent-divergent nozzle and added multidirectional control, and (significantly for what followed three decades later) it was conceived as something that could be adapted to existing nozzle configurations rather than requiring a new propulsion architecture. The nozzle flew on a specially modified F-16 under the Multi-Axis Thrust Vectoring programme, tested at Edwards Air Force Base through the 1990s, accumulating 87 hours of ground testing and 135 flight hours across 95 sorties. The programme demonstrated that an axisymmetric vectoring nozzle could deliver reliable multidirectional vectoring without sacrificing performance across the envelope. Then it was put away.

A DIFFERENT PROBLEM, THREE DECADES LATER

What brought it back was a change in what makes a combat aircraft survivable. Shield AI’s argument, set out by its senior propulsion engineer for the X-BAT programme, is that in modern war gaming more aircraft are lost on the ground than in the air. Three decades of investment have gone into stealth to improve survivability in flight, while the same aircraft remain highly vulnerable parked on a runway that is easier to target than ever. Projecting combat power increasingly depends on dispersing aircraft across ships, islands, highways and expeditionary sites – and legacy fleets cannot do that.

Vertical take-off and landing is the enabler, and it turns the constraint inside out: an aircraft that does not need a runway treats any adequate surface as one. But building a tactical strike platform that launches and recovers vertically while retaining the speed, range, payload and survivability expected of a fighter makes propulsion the hardest problem on the programme. During vertical take-off, hover and landing the propulsion system becomes the aircraft’s primary means of holding attitude, and the demands on the nozzle are entirely different from those of forward flight – rapid, precise control movements sustained continuously, rather than occasional vectoring inputs to sharpen a turn.

Evaluated against that requirement, AVEN’s multidirectional vectoring provided the agility needed, and its flight history provided something equally valuable: confidence that the underlying hardware worked. Rather than starting a new propulsion system from first principles, Shield AI could build on the engineering already done and concentrate on adapting it.

WHAT HAD TO CHANGE

The distinction matters, because the nozzle that flew in the 1990s was designed to improve manoeuvrability in forward flight. X-BAT asks it to serve as the aircraft’s primary flight control system during vertical operations, continuously redirecting thrust to hold attitude at a speed and precision the original programme had no reason to test for. Converting it meant reconsidering how the nozzle actuates, how it interfaces with the engine and how it communicates with the aircraft’s flight controls.

The aircraft’s propulsion arrangement is otherwise conventional: a single jet engine mounted down the centreline, in the manner of an F-16. At take-off the engine lights its afterburner to generate the thrust needed for vertical launch while AVEN vectors it for control. Many modern fighters already produce more thrust than their own weight but expend it on a runway take-off roll; X-BAT uses the same margin to lift off vertically. Landing inverts the problem, requiring thrust to be modulated closely against the aircraft’s weight while the nozzle continuously adjusts direction to hold a stable descent.

Working with GE Aerospace’s Edison Works team, Shield AI took an AVEN directly from a storage warehouse, refurbished the original hardware, integrated it with the F110-GE-129E engine and returned it to operation. Integration, actuation and engine light-off testing were completed at GE Aerospace’s Peebles Test Operation, validating that nozzle, engine, actuators and control systems functioned as a single propulsion system. It was the first fully integrated AVEN test campaign since the original flight programme more than thirty years earlier.

WHY STARTING WITH OLD HARDWARE WAS THE FAST ROUTE

The programme’s own assessment is that beginning with flight-proven hardware compressed the development timeline substantially. Rather than spending years developing and qualifying a new nozzle architecture, the engineering effort could be directed at the problems specific to vertical flight, moving from integration to ground testing in a fraction of the time and reducing technical risk along the way. GE Aerospace engineers had already spent years designing, testing and flying the nozzle, and that body of work did not have to be recreated.

The AVEN now under test is not the final configuration. Shield AI and GE Aerospace are reengineering the nozzle around modern materials, guidance and control systems and production techniques, while retaining the original flight-proven design as the baseline – with future iterations intended to be faster, lighter, more responsive and optimised for autonomous vertical flight. The next step is X-BAT’s vertical flight testing, at which point the nozzle transitions from an experimental thrust-vectoring device into the system that enables the aircraft to lift off, hover, transition and land.

The wider point the programme draws is that a technology conceived to explore the limits of fighter manoeuvrability has become the foundation of a different generation of aircraft entirely. Its first chapter established that multidirectional thrust vectoring could work; its second is establishing what that capability enables.