-T / T / +T | Comment(s)

Wednesday, November 1, 2006

R&D: Flight Testing at 0 Agl

With ongoing engineering work on the V-22, Boeing is pushing the limits on the use of simulation to cut the cost and time to certify new aircraft systems.

By John Croft

Cockpit simulation
Boeing uses the V-22 Flight Simulation Lab for rapid prototyping of control laws and crew systems, as well as evaluation of handling qualities.
The Flight-Control System Integration Rig
The Flight-Control System Rig is a room that contains V-22 control surface actuators, hydraulic and electrical systems and flight-control computers laid out in the configuration of the actual aircraft.
FOR 12 HR A DAY, THREE DAYS A WEEK, ENGINEERS, test pilots and technicians working on the V-22 Osprey program at Boeing's Integrated Defense Systems Rotorcraft plant here just south of Philadelphia put the military tilt-rotor through stress tests that both increase the safety and reduce the amount of flight time necessary to check crucial details of the aircraft's hardware and software. And they never leave the ground.

The trials, which feature pilots "flying" actual V-22 flight-control hardware and computer systems in three interconnected labs, are designed to check out man-machine performance during emergency procedures as well to investigate how the crew and the aircraft perform after system failures (known as "degraded modes" testing). The checks are being performed after each new flight-control system software revision for the V-22, upgrades that are being implemented as the aircraft is brought closer to initial operational capability, slated for next September.

Though Boeing's hardware-in-the-loop simulations have been going on for more than a decade at the plant, the effort continues to be on the leading edge of advanced simulation capabilities in the testing and evaluation environment. That and similar efforts in aerospace are designed to make flight testing safer and shorter. It's also part of a broader push-or perhaps ultimatum-in the industry to rely more heavily on simulations to reduce the risk and time needed to reach initial operational capability. Achieving that would reduce the number of costly flight tests for aircraft certification (particularly when it comes to proving functionality in a "net-centric" warfare environment). [Defined as a continuously evolving community of people and devices linked by a communications network, "net centric" is the Pentagon-mandated baseline requirement for all U.S. military information systems. Net centrism aims to turn an information advantage into a warfighting one.] "A lot of the more mundane (flight testing) requirements you could do in a simulator if you get your modeling right," said Phil Dunford, V-22 program manager for Boeing Rotorcraft Systems and the current president of the American Helicopter Society. "That will drive down the cycle time and costs." High-fidelity simulations could also play a role in accelerating the design cycle, cutting down on the number of wind-tunnel models needed as well as the number of wind-tunnel tests.

On the operational end of the testing spectrum, Dunford said, modeling and simulations will be key in coming up with cost-effective ways to prove that a new or modified vehicle operates as designed in a net-centric environment without investing in a live exercise. Dunford said Boeing and other companies are just now beginning to formulate ideas for how to accomplish the task. Boeing even created a vice president slot for analysis, modeling and simulations in the Integrated Defense Systems business unit. "There's no room to do business as usual," he added.

The V-22 work represents a baby step along the way, but a distinct starting point for using simulations and modeling to reduce risk and time to move a rotorcraft from design to initial operational capability. The most advanced of Boeing's efforts in Philadelphia is the hardware-in-the-loop simulator capability, also known as the Triple Lab Tie-in. As the name suggests, the Triple Lab, initially put together in 1995, is a combination of three labs-the Flight Simulation Lab, which includes a V-22 cockpit and computer equipment with a generic tilt-rotor mathematical model developed by Bell Helicopter; the Flight-Control System Integration Rig, a room with the V-22's control surface actuators, hydraulic and electrical systems and flight-control computers laid out in the configuration of the actual aircraft, and the Systems Integration Lab, an avionics- and software-qualification facility that generates the information that pilots see on the multi-function displays and other equipment in the V-22 cab. Boeing uses the Flight Simulation Lab as a stand-alone unit for rapid prototyping of control laws and crew systems, as well as evaluating handling qualities. The Flight-Control System Integration Rig's primary role as an individual lab is to test how the hardware and flight-control software will work together.

With the three labs joined via Mil Std 1553 connections, Ethernet and Scramnet (shared common random access memory network), however, helicopter designers now have an asset with the fidelity needed to run hardware-in-the-loop experiments with pilots, the actual control algorithms and real hardware, including the aircraft's control actuators.

As pilots in the cab move their controls during a Triple Lab simulation, digital signals representing the movement are sent to the flight-control integration rig, where they are converted to analog signals and sent to the flight-control computers. The flight-control computers execute the flight-control laws that generate commands to the hydraulic actuators for the flaperons, elevator, rudders, and swashplates, as well as the nacelles that set the tilt of the rotors. The actuators, pumps, reservoirs and switching/isolation valves in the rig are identical to the flight equipment. The lab also includes a "patch panel" that lets technicians introduce electrical failures. Simulated hydraulic failures are also possible with the rig.

When the actuators move, the position transducers measure the movement and send signals to the tilt-rotor mathematical model located in the Flight Simulation Lab, which then computes the V-22's dynamic response to the actuator movements and generates the cockpit instrument readings that will be sent to the flight deck. The math model also calculates the feedback forces on the control surfaces. It sends that data back to the flight-control integration rig, where add-on hydraulic actuators apply the forces directly to the control surface actuators, thereby providing feedback to the pilot as in the real aircraft in flight. The input from the math model, combined with the mission computer, also updates the cab's four multi-function displays (flight hardware, too) and the out-the-window visual display. The cab is surrounded by a 30-ft-dia dome and uses four projectors that shine a 180-deg horizontal and 90-deg vertical visual scene on the inside surface of the dome. The cab sits on a 10-ft base that can either be fixed or full-motion.

Boeing had planned to build a hardware-in-the-loop simulator for the Comanche helicopter program before the program was cut in February 2004. Boeing officials from the 787 program had also been interested in the pilot-in-the-loop capabilities of the V-22 simulator, said Curtis Walz, an engineer who until recently was the integrated product team lead for the V-22 autopilot system. Bell, Boeing's partner on the V-22 program, is using a hardware-in-the-loop simulator for testing its BA609 tilt-rotor, a smaller, commercial variant of the V-22. Along with giving Bell engineers the means to experiment with flight-control algorithms on the ground, the integrated system also allows BA609 test pilots to "try before they fly" to understand and experience a certain aircraft behavior and observe the system responses.

Most recently, Boeing has been using the Triple Lab to run a variety of on-ground tests each time engineers modify the flight-control system software for the aircraft. Software changes are sometimes caused by customer requested modifications or deficiencies discovered during flight tests. For instance, after an unintended takeoff and crash heavily damaged a V-22 at MCAS New River, N.C. March 27, Marine officials ordered a change to the software for the full-authority digital electronic controllers (FADECs) that govern the Osprey's two Rolls-Royce AE1107C Liberty engines. The FADEC software had been written on the assumption that the most critical situation in which the unit would fail would be in flight. Therefore, the engine defaults to a power-up condition to maintain altitude. It was not anticipated that a failure on the ground would cause the aircraft to fly before the pilots knew it. ("Fix Planned for V-22 FADECs In Wake of Uncommanded Liftoff," May 2006, page 10.)

Current production aircraft include the Block B V-22 for the U.S. Marine Corps and the Block 0/10 CV-22 for the Air Force. Testing is also under way for future versions of the aircraft, including studies of the Block C V-22 for the Marines and Block 20 V-22 for the Air Force.

One line of testing is for investigating the appropriateness of emergency procedures. In these simulations, test engineers introduce unannounced failures to see if pilots correctly decipher the failure from the Warnings, Cautions and Advisories subsystem and apply the appropriate emergency procedure in the flight manual.

Another type of testing is the degraded-modes program, in which engineers inject a failure into the simulator, tell the crew about it, then have the crew follow procedures while putting the aircraft through its flight envelope, finishing the simulation with a landing. Walz said the "stress testing" helps validate the requirement that the V-22 can be brought in for a safe landing after most double failures. Walz said the human-factors findings in the simulations are also helpful in trying to understand how pilots might react to a failure in flight.

As part of the degraded-modes testing, Boeing also evaluates aircraft handling qualities and controllability after failures. For this test, pilots are asked to fly an engineering maneuver, like precision hovering, after failures are introduced. Walz said tests can last 8-16 hr investigating a single failure.

"What we get out of it," he said, "is that we can be more specific in an operating document" regarding how to control the aircraft under those circumstances. Changes could include minimizing certain maneuvers or flight regimes.

According to Walz, Boeing performed 5,567 hr of simulated flight with the system (3,420 hr of engineering tests and 2,147 hr of piloted tests) between March 2001 and February 2006. For the autopilot alone, which has 11 modes for the Marine MV-22 and 12 modes for the Air Force's version, Walz estimated that the Triple Lab saved hours of flight testing. For example, he said, Boeing was able to reduce the number of gross-weight, center-of-gravity position and altitude combinations at which the autopilot had to be tested in flight by identifying the most critical cases in the Triple Lab.

"We eliminated test points by running a lot of conditions in the Triple Lab environment," Walz said.

Given his experience with the V-22 program, he said, hardware-in-the-loop simulations are here to stay. "You can look for anomalies, take precautions or make fixes with the system, zooming in on more critical type cases and leaning out the flight test program," he said. Future systems will be even more comprehensive.

"I don't believe we can eliminate flight test," he said, "but we can reduce it."

[X] Dismiss Ad