Control Engineering

Link This | Email this | Blog This | Comments (0)

What is “control reversal?”
March 23, 2007

The term “control reversal” refers to a potentially disasterous phenomenon that occurs in high-speed fixed-wing aircraft, typically military jet fighters. It is associated with a more general aerodynamic phenomenon called “aeroelasticity,” in which the aircraft structure bends under aerodynamic loads.

Control reversal is generally important to control engineers not because it’s a common control problem—it’s not—but because its solution says a lot about what intelligent control systems are capable of. Intelligent control systems incorporating machine intelligence in the form of PLCs, PACs, PCs, etc. can follow sophisticated non-linear and even bizarre control algorithms.

The Wrights brothers’ great contribution to aviation was to realize that flight was not a lift or power problem, but a control problem. Specifically, a coordinated turn required controlled rotations about the aircraft’s vertical axis (yaw) and about its long axis in the forward-movement direction (roll).

Twisting the wing tips unbalances lift across the wing, creating a rolling moment. To see what happens, click on the image to visit the NASA Glenn Research Center.

The Wrights method of controlling roll is called “wing warping,” and requires deformable wing structures. To roll left (left wing moving downward), they applied a twist to their wings such that the leading edge on the port (left-hand looking forward) wing tip pointed downward, and the starboard (right hand looking forward) tip pointed up. I’m told that the idea popped into their heads when, during a discussion of the problem, one of them took a long, narrow box out of the wastebasket and twisted it!

It is not clear whether it was Glenn Curtis himself or his friend and fellow aviation enthusiast Alexander Graham Bell (of Bell Telephone fame) who suggested replacing the complex, flexible and previously patented wing-warping system with rigid wings and moveable ailerons. In either case, Curtis was the mechanical genius behind the project, so I like to give him credit. Since ailerons worked wonderfully, and had been used previously (though unsuccessfully) by New Zealand inventor Richard Pearse in 1902, and (with spectacular success) by Alberto Santos Dumont in France, the aileron idea was pretty much fair game for anyone wanting to get into the aircraft business. Hence, it became (and still is) the standard method for aircraft roll control. For it to work properly, however, the wing itself has to be as stiff as possible.

Fast forward to the late 20th Century and development of the F/A-18 for the U.S. Navy. By the time the F/A-18 was being designed, the demands of supersonic flight had pushed engineers to use extremely thin swept wings, which made for long, thin spars. Early F/A-18 versions had wings that proved to be just a little bit too thin.

During maneuvers at extreme speeds, pilots experienced control reversal. If they tried to roll the aircraft left, it rolled right instead! The harder they pushed their stick to the left, the harder the plane went right. Not only were the controls reversed, they were divergent, meaning that the harder they turned, the reverser they got. Pilots don’t like that.

The culprit was, of course, aeroelasticity. The ailerons positioned at the trailing edge tips applied torques twisting the wings in a sense opposite what the pilot intended. Raising the left aileron to push the left wing down twisted the wing tip so that the entire wing’s leading edge warped up. The whole phenomenon was speed dependent. At low speeds, the controls worked as advertised, but as the aircraft’s speed increased, the control effectiveness started to drop off. At a certain speed, the wing warp balanced the aileron input and control effectiveness was zero. At higher speeds, wing warping swamped ailerons, and the controls reversed.

The quick fix was to simply not go so fast. The Navy is not into going slowly, so they soon fitted thicker wings to better resist aerodynamically induced torques. Losing a little flight-performance was better than losing control of the plane. The fixed version could beat just about everyone else, so the pilots were happy, the Navy was happy, and (once the spin doctors were done) the taxpayers were happy with their new toy.

The story does not, however, end there. The engineers at NASA Dryden test facility got this crazy idea to actually use this aeroelastic effect to control the plane! They fitted a unit with the old-style thin wings, then went out to map the transfer function between control inputs and aircraft responses from barely moving to as fast as their test pilots dared to go, then wrote a computer program to convert pilot inputs into the appropriate control surface movements to make the craft do what the pilot wanted it to do. At low speeds, it acted like the standard F/A-18 fly-by-wire control system. At high speeds, when control reversal ruled the world, the program said: “Okay, the pilot wants to roll left, but at this speed controls are reversed, so I’ll just drop the left aileron and raise the right one. The wings will warp and bank the plane the other way. No problem!”

The moral of this story is that, while most control engineers don’t have to worry about control reversal, it’s important to realize that computerized control systems can be programmed to do just about anything you want them to. Since they have no intuition whatever, they can be programmed to act counterintuitively if that’s what the situation calls for. They can anticipate any bizarre, nonlinear behavior on the part of the mechanical system as long as you first test the system to determine how it will react throughout its operational envelope.

For more information about control-system programming, visit the Control Engineering website at http://www.controleng.com.

C.G. Masi, Control Engineering Senior Editor, charlie.masi@reedbusiness.com

For additional information, visit these Websites:

http://wright.nasa.gov/airplane/warp.html

http://en.wikipedia.org/wiki/Image:Aileron_roll.gif

Posted by Charlie Masi on March 23, 2007 | Comments (0)