 The Lewis Flight Propulsion Laboratory of the National Advisory Committee for Aeronautics has investigated the problems of survival in light airplane accidents because this type of accident is responsible for many of the casualties of civil aviation. It was the purpose of this work to determine how people are injured during such crashes. In order to obtain the desired information on crash survival, several small tandem two-place light airplanes were crashed. A stall spin accident was chosen for study. Such an accident occurs when an airplane stalls and enters into a spin during a landing approach. The airplane is too close to the ground to recover from the spin. A survey made by Cornell Crash Injury Research indicated that this type of accident is a frequent source of fatalities in light airplane crashes. To simulate such a crash, the axis of the crash was rotated. The ground was simulated by a mound of earth and the airplane ran along the ground in a level attitude to strike the mound of earth. The mound was located so that the airplane's left wingtip, left landing wheel, and the cell would strike at the same time in the same manner as in the stall spin accidents that were being simulated. Two dummies were installed in the airplane. The dummy in the front seat was a standard Air Force dummy used in the testing of parachutes. In the construction of this dummy, no attempt was made to simulate the resilience of the human body, although the mass distribution of the component parts was similar to that of a human being. An Air Force anthropomorphic dummy was installed in the rear seat. This dummy is a reasonable replica of the human body in both mass distribution and resilience of human tissues. Motion pictures were taken from various vantage points around the crash site. These motion pictures show how the dummy passengers moved during the crash and how the airplane's structure deforms. Meanwhile, the telemeter transmitting station carried in the airplane radiated the crash loads measured on the passengers and in the cockpit. The transmitted data were recorded by a receiving station at the crash site. Three crashes were made in this investigation. The airplanes were propelled by their own power along a runway toward the crash barrier. A slipper mounted on a monorail guided each airplane into the barrier. The speed of the airplane at impact in the three crashes was 60, 47 and 42 miles per hour. The motion pictures obtained verify that most of the injuries that result from light airplane accidents are caused by the occupant striking part of the airplane inside the cockpit. These injuries occur in two ways. When the airplane's structure collapses and strikes the occupant, this happens to the occupant in the front seat of this model. The occupant can also be injured when he is improperly restrained and moves from his normal position to strike objects in the cockpit as illustrated here by the occupant in the rear seat. The next motion picture sequence of the 60 mile per hour crash shows how the front of the airplane collapses and crushes the dummy in the front seat. Now watch the dummy in the front seat. The red liquid that obscures the airplane in the latter part of the pictures was used to study the fuel spillage during a crash. There is little chance that the occupant in the front seat would survive such a crash. The lower part of the dummy was trapped in the collapsing airplane structure. The instrument panel moved rearward and the dummy's head hit it with sufficient force to leave a dent. Even if the instrument panel is not pushed rearward, an improperly restrained occupant may hit it with his head. This is illustrated in the next motion picture sequence of a crash which takes place at 42 miles per hour. Even though the fuel spillage force structure was not pushed back into the front dummy's lap, the dummy struck the instrument panel because he was restrained only by a seat belt and his torso was free to move. Again, watch the dummy in the front seat. As before, the dummy in the front seat struck the instrument panel. The neck of the dummy in the rear seat broke because of an imperfection in its construction. Injuries of this type are not likely in an actual crash. Here is the dent in the instrument panel. These results explain why the survey conducted by Cornell Crash Injury Research found so many head injuries in actual crashes. This survey found that head injuries were inflicted in 88 percent of the accidents. If an occupant is not to be injured when wearing only a seat belt for restraint, sufficient space must be clear ahead of the occupant to allow him to flex over the seat belt. The occupant in this airplane was restrained by only a seat belt. The front seat and rear control stick were removed, so that there was sufficient room for complete flexure of the dummy's torso. The torso of the dummy moved forward and downward until the chest contacted the thighs. It is apparent that if injuries resulting from contact with solid structure are to be avoided when using only a seat belt for restraint, a distance of about 45 inches ahead of the seat must be free of any solid objects. This much of a clear space is seldom available in an airplane. This dangerous movement is reduced to safe values when the occupant is properly restrained by a seat belt and a shoulder harness. The dummy in the rear seat in the next crash wore a seat belt and a shoulder harness. This dummy moved forward and out of his seat about 8 to 10 inches. The forward movement was limited by the harness. In the most forward position, the torso was about vertical. Notice this action in these slow motion pictures of the crash. Watch the dummy in the rear seat. Here again the dummy lost his head due to a weakness in the neck structure of the dummy. This would not have happened to a human occupant. If the harness is made of material that stretches excessively, the occupant may still strike objects inside the cockpit. The harness restraining the rear dummy in the 60 mile per hour crash stretched sufficiently to allow the dummy's head to strike the back of the dummy in the front seat. This situation was aggravated by the collapse of the structure between the front and rear seats. This moved the front dummy rearward. Again, watch the dummy in the rear seat. Unfortunately, the fuel mist obscured the actual impact of the dummy's head with the front dummy. However, a post crash inspection revealed a dent in the dummy's helmet. Here is the position of the dummy's head when it struck the front dummy. It has been shown that some of the benefits of a shoulder harness are lost if the harness stretches excessively in the crash. If the restraining harness is to serve its purpose, it must not break. In order to tell the designer how strong to make this harness, the forces in these straps were measured in the crashes. To obtain this information, tensiometers were located at each end of the seat belt and on the common juncture of the shoulder harness. The sum of the forces on each end of the seat belt and the force measured in the shoulder harness in the 60 mile per hour crash for the dummy in the rear seat are plotted against time after the impact with the barrier. The seat belt restraining force has two peaks, one of 4,000 pounds and a second of 3,200 pounds. The major seat belt forces endured for about one-tenth of a second. The combined pull on both shoulder harness straps is approximately equal to half that of the seat belt, showing that the seat belt supported most of the load. The restraining forces in the seat belt were higher when the passenger was restrained by a seat belt only, even when the impact speed was reduced from 60 to 47 miles per hour. During the 47 mile per hour crash in which the dummy was restrained only by a seat belt, peak seat belt forces were 4,400 pounds and 3,000 pounds, as compared to a peak of 4,000 pounds for the seat belt for the 60 mile per hour crash. Seat belts and harness capable of withstanding these dynamic loads can be comfortable and light in weight. To summarize, the information presented so far has shown that injuries and crashes result, one, when the cockpit collapses and the occupant is crushed by the airplane structure, two, when the seat belt and shoulder harness stretch excessively under the crash loads, and three, when the occupant is restrained by only a seat belt. Even if the structure remains intact, both seat belt and shoulder harness are necessary in small cockpits if serious body blows are to be avoided. The information presented here has also indicated the loads produced in the restraining harness during a crash. Even if the occupant is properly restrained during a crash, he may still be injured by the deceleration he encounters. The severity of the injury received from a deceleration depends on the magnitude of the deceleration, the rate at which it increases, commonly called the rate of onset, and the duration of the deceleration. These conclusions were reached from the work of Lieutenant Colonel Stapp in the report entitled Human Exposures to Linear Deceleration. In this study by Stapp, tests were made with human beings carefully supported by specially designed seat belt, shoulder harness, and leg straps. Under such conditions, the experimenter subjected himself to a deceleration that had a maximum of 47 Gs. The deceleration endured for 0.228 seconds and had a rate of onset of 735 Gs per second. The resulting injuries were minor. When the rate of onset increased from 735 to 1550 Gs per second, signs of shock were observed. To obtain an indication of the injuries that may result from the decelerations encountered in these experimental crashes, the decelerations of the rear dummy's chest were measured. These decelerations had about the same duration, and the magnitudes were very little higher than those reported by Stapp. A maximum deceleration of 50 Gs was measured in the rear dummy's chest in the 60 mile per hour crash. The maximum for the 47 mile per hour crash was 46 Gs. And the maximum for the 42 mile per hour crash was 32 Gs. From the standpoint of magnitude and duration, these decelerations are within the tolerable limits established by Stapp's work. However, the rate at which the deceleration increases in these light airplane crashes varied from 2200 Gs per second for the 60 mile per hour crash to 950 for the 42 mile per hour crash. It's expected that these high rates of onset would cause momentary unconsciousness. It may be concluded therefore that the decelerations measured in these crashes are not large enough to fatally injure the occupant in the rear seat. Oddly enough, the deceleration of the occupant during a crash is often higher than that of the fuselage floor. This effect is shown in the following motion picture sequence, in which the action is slowed to about one sixtieth of normal. Graphs of the deceleration of the chest and floor are superimposed over the motion picture. These graphs develop in phase with the airplane's action. Notice that during the period that the seat belt and harness stretch, the dummy decelerates less than the fuselage floor under him. During this period, the dummy acquires velocity relative to the local airplane structure. When the seat belt and harness stretch is complete, the passenger is decelerated rapidly to the speed of the airplane. This causes the dummy to have larger peak decelerations than the fuselage floor. The deceleration of the fuselage floor had a peak of 35 Gs, whereas the peak for the chest was 50 Gs. This increase in the chest deceleration over that of the fuselage floor was also found in the 47 mile per hour crash when the dummy was restrained only by a seat belt. This amplification of peak decelerations may be even greater if the seat belt and shoulder harness are slack when the crash occurs. The slack and stretch in these members may cause failure of the seat belt and harness as a result of the high crash deceleration loads produced. The deceleration imposed on the occupant and his restraints depends on the deceleration of the fuselage floor. The deceleration of the fuselage floor in turn depends on the strength of the airplane structure and the distribution of the airplane weight. In the 47 mile per hour crash shown at about 140th of normal speed, notice how the deceleration of the fuselage floor reaches excessive peaks. Every time a main structural element supports the crash load, a peak deceleration occurs. When the loaded structure breaks, the deceleration drops. The maximum deceleration of the cockpit, therefore, depends directly on the strength of the structure as long as the cockpit has not collapsed completely. This is illustrated by a comparison of the maximum decelerations measured in the 60 and 47 mile per hour crashes with those of the 42 mile per hour crash. The maximum deceleration in all cases is about the same, ranging between 26 and 33 Gs. These decelerations represent the maximum load that the airplane structure can support. The higher kinetic energy of the higher speed crash is dissipated by a more extensive crushing of this fuselage floor structure. This increase in the crumpling of the fuselage floor structure increases the time during which the cockpit decelerations exist. The extent to which the crumpling of the fuselage floor structure reduces the deceleration applied to the cockpit floor can be seen by comparing the deceleration of the engine with that of the fuselage floor in the 60 mile per hour crash. Notice in the next motion picture sequence that the engine deceleration rises immediately upon impact, whereas the floor deceleration rises gradually until extensive crumpling of the fuselage floor structure has taken place. This crumpling of the fuselage absorbs a considerable portion of the crash energy. When the floor structure has crumpled and the load is applied directly to the fuselage floor, the engine and floor decelerate at about the same rate. The peak deceleration of the engine was 62 Gs while the floor had a peak of only 35 Gs. For this reason, it's desirable to place as much of the airplane as possible forward of the cockpit. This arrangement has two principal advantages. First, it places much of the airplane structure in front of the cockpit to crumple and cushion the crash flow. And second, it places much of the airplane mass in front of the cockpit. By reducing the mass behind the cockpit, the load on the cockpit structure is reduced and failure of this structure is less likely in a crash. It may be concluded therefore that the chances of surviving a crash are better if both the seatbelt and the shoulder harness of proper design are used. The chances of impact survival are also better if the airplane has much of its structure and mass ahead of the cockpit. The information presented in this motion picture is reported in NACA Technical Note 2991 entitled Accelerations and Passenger Harness Loads measured in full-scale light airplane crashes.