The Advent of Fixed-Base Simulation

Simulating flight has been an important part of aviation research since even before the Wright brothers. The wind tunnel, invented in the 1870s, represented one means of simulating flight conditions. The rudimentary Link trainer of the Second World War, although it did not attempt to represent any particular airplane, was used to train pilots on the proper navigation techniques to use while flying in the clouds. Toward the end of the Second World War, researchers within Government, the military services, academia, and private industry began experimenting with

analog computers to solve differential equations in real time. Electronic components, such as amplifiers, resistors, capacitors, and servos, were linked together to perform mathematical operations, such as arithme­tic and integration. By patching many of these components together, it was possible to continuously solve the equations of motion for a moving object. There are six differential equations that can be used to describe the motion of an object. Three rotational equations identify pitching, rolling, and yawing motions, and three translational equations identify linear motion in fore-and-aft, sideways, and up-and-down directions. Each of these equations requires two independent integration processes to solve for the vehicle velocities and positions. Prior to the advent of analog computers, the integration process was a very tedious, manual operation and not amenable to real-time solutions. Analog computers allowed the integration to be accomplished in real time, opening the door to pilot-in-the-loop simulation. The next step was the addition of controlling inputs from an operator (stick and rudder pedals) and output displays (dials and oscilloscopes) to permit continuous, real-time con­trol of a simulated moving object. Early simulations only solved three of the equations of motion, usually pitch rotation and the horizontal and vertical translational equations, neglecting some of the minor cou­pling terms that linked all six equations. As analog computers became more available and affordable, the simulation capabilities expanded to include five and eventually all six of the equations of motion (com­monly referred to as "six degrees of freedom” or 6DOF).

By the mid-1950s, the Air Force, on NACA advice, had acquired a Goodyear Electronic Differential Analyzer (GEDA) to predict aircraft handling qualities based on the extrapolation of data acquired from previous test flights. One of the first practical applications of simula­tion was the analysis of the F-100A roll-coupling accident that killed North American Aviation (NAA) test pilot George "Wheaties” Welch on October 12, 1954, one of six similar accidents that triggered an emer­gency grounding of the Super Sabre. By programming the pilot’s inputs into a set of equations of motion representing the F-100A, researchers duplicated the circumstances of the accident. The combination of sim­ulation and flight-testing on another F-100A at the NACA High-Speed Flight Station (now the Dryden Center) forced redesign of the aircraft. North American increased the size of the vertical fin by 10 percent and, when even this proved insufficient, increased it again by nearly 30 per­cent, modifying existing and new production Super Sabres with the

larger tail. Thus modified, the F-100 went on to a very successful career as a mainstay Air Force fighter-bomber.[719]

Another early application of computerized simulation analysis occurred during the Air Force-NACA X-2 research airplane program in 1956. NACA engineer Richard E. Day established a simulation of the X-2 on the Air Force’s GEDA analog computer. He used a B-17 bom­bardier’s stick as an input control and a simple oscilloscope with a line representing the horizon as a display along with some voltmeters for airspeed, angle of attack, etc. Although the controls and display were crude, the simulation did accurately duplicate the motions of the air­plane. Day learned that lateral control inputs near Mach 3 could result in a roll reversal and loss of control. He showed these characteristics to Capt. Iven Kincheloe on the simulator before his flight to 126,200 feet on September 7, 1956. When the rocket engine quit near Mach 3, the air­plane was climbing steeply but was in a 45-degree bank. Kincheloe remem­bered the simulation results and did not attempt to right the airplane with lateral controls until well into the entry at a lower Mach number, thus avoid­ing the potentially disastrous coupled motion observed on the simulator.[720]

Kincheloe’s successor as X-2 project pilot, Capt. Milburn Apt, also flew the simulator before his ill-fated high-speed flight in the X-2 on September 27, 1956. When the engine exhausted its propellants, Apt was at Mach 3.2 and over 65,000 feet, heading away from Edwards and apparently concerned that the speeding plane would be unable to turn and glide home to its planned landing on Rodgers Dry Lake. When he used the lateral controls to begin a gradual turn back toward the base,

the X-2 went out of control. Apt was badly battered in the violent motions that ensued, was unable to use his personal parachute, and was killed.[721]

The loss of the X-2 and Apt shocked the Edwards community. The accident could be duplicated on the simulator, solidifying the value of simulation in the field of aviation and particularly flight-testing.[722] The X-2 experience convinced the NACA (later NASA) that simulation must play a significant role in the forthcoming X-15 hypersonic research air­craft program. The industry responded to the need with larger and more capable analog computer equipment.[723]

The X-15 simulator constituted a significant step in both simulator design and flight-test practice. It consisted of several analog computers connected to a fixed-base cockpit replicating that of the aircraft, and an "iron bird” duplication of all control system hardware (hydraulic actua­tors, cable runs, control surface mass balances, etc.). Computer output parameters were displayed on the normal cockpit instruments, though there were no visual displays outside the cockpit. This simulator was first used at the North American plant in Inglewood, CA, during the design and manufacture of the airplane. It was later transferred to NASA DFRC at Edwards AFB and became the primary tool used by the X-15 test team for mission planning, pilot training, and emergency procedure definition.

The high g environment and the high pilot workload during the 10-min­ute X-15 flights required that the pilot and the operational support team in the control room be intimately familiar with each flight plan. There was no time to communicate emergency procedures if an emergency occurred— they had to be already imbedded in the memories of the pilot and team members. That necessity highlighted another issue underscored by the X-15’s simulator experience: the necessity of replicating with great fidelity the actual cockpit layout and instrumentation in the simulator. On at least two occasions, X-15 pilots nearly misread their instrumentation or reached for the wrong switch because of seemingly minor differences between the simulator and the instrumentation layout of the X-15 aircraft.[724]

Overall, test pilots and flight-test engineers uniformly agreed that the X-15 program could not have been accomplished safely or pro­ductively without the use of the simulator. Once the X-15 began flying, engineers updated the simulator using data extracted from actual flight experience, steadily refining and increasing its fidelity. An X-15 pilot "flew” the simulator an average of 15 hours for every flight, roughly 1 hour of simulation for every minute of flying time. The X-15 experience emphasized the profound value of simulation, and soon, nearly all new airplanes and spacecraft were accompanied by fixed-base simulators for engineering analysis and pilot/astronaut training.