"Current ice-protection systems attempt to prevent or remove an ice accretion and may provide limited sensing of the presence of ice," said Michael Bragg, professor and head of aeronautical and astronautical engineering at the UI. "But these systems use little, if any, information about the present state of the aircraft. Our approach is to provide the pilot with a near real-time characterization of the effect that ice is actually having upon his aircraft."
When ice accumulates on flight surfaces, it can change an aircraft's performance, stability and controllability. Accidents can occur not only from degraded aerodynamic performance, but also from well-intentioned pilots making bad decisions in the absence of adequate information.
"Pilots expect an aircraft to respond in a certain way to their commands, and when it doesn't, they might assume the wrong reason and take improper measures that can result in a dangerously unstable aircraft," said Tamer Basar, the Fredric G. and Elizabeth H. Nearing Professor of Electrical and Computer Engineering at the UI. "We have to provide more relevant information to the pilot so that he can make informed decisions and safely fly an aircraft under severe icing conditions."
Using systems identification techniques, the researchers first modeled the effects that ice can have on an aircraft's flight dynamics. Then they developed methods to detect and characterize those effects. "Instead of relying only upon an ice-thickness sensor, for example, we're measuring the changes in aircraft performance and control during an icing encounter," said James Melody, a graduate student in the university's Coordinated Science Laboratory. "We use a neural network to extract information from the flight dynamics and various other sensors to better inform the pilot of the current state of his aircraft."
Ultimately, the researchers want their ice-management system to automatically adapt the flight control system to make an aircraft easier - and safer - to fly when iced. For larger, newer aircraft, the system could operate autonomously, while still keeping the pilot properly informed.
"We are putting our smart icing system between the ice protection system and the pilot," Bragg said. "After the ice-protection system has done all it can do, we want our system to analyze the situation and then help make adjustments to improve safety. As the aviation industry moves toward fully autonomous flight control systems, we're ensuring that icing and safety are part of that process."
A flight test conducted earlier this year is being used to validate the researchers' models and algorithms. A prototype of the smart ice-management system will be flight-tested next year.
The National Aeronautics and Space Administration is supporting the work.