Air University Review, July-August 1968

Development of Aerospace Escape Systems

James W. Brinkley

Two decades have passed since the first operational ejection seat was installed in a Unites States Air Force aircraft within these two decades the escape systems used by the Air Force have evolved from a relatively simple, ballistically catapulted propelled cockpit escape system of the F-111. This evolution has been marked by a progression of subsystem improvements, including the automatically opened lap belt, automatically actuated parachute, and seat/man separation devices. Each has significantly enhanced the efficacy of escape system. The overall performance envelope of the escape system has been extended by the development of seat stabilization techniques, forcefully deployed parachutes, and methods of protection from windblast and altitude, such as the encapsulated seat.

Aerospace medical research has been an inherent part of the escape-system development effort. One need only consider the factors that limit the performance capability of a given escape system to recognize this relationship. Figure 1 illustrates a typical performance envelope for an open ejection seat assuming level flight and no sink rate. The performance limit from point A to point B is established by the requirement to provide altitude protection for the ejectee. The limit between points B and C is determined by the power capability of the aircraft. The ejectee’s ability to withstand high dynamic pressures and rapid deceleration defines the limit from point C to D.  The human tolerance to parachute-opening shock determines the limit of the envelope between D and E. Point E represents the highest velocity at ground-level air density at which the personnel parachute may be opened without injury to the parachutist. (In some escape systems the parachute opening is automatically retarded until the velocity decays below this limit.) The limit between D and E is caused by the time delay and resultant ejection trajectory loss until the ejectee has decelerated to a safe parachute-opening velocity. The trajectory height remains a critical parameter from point E through points F and G. In this range a combination of trajectory height and horizontal velocity determines the actual limits. Point F represents the airspeed that, for a given system, is considered sufficient to obtain an open parachute with the inherent trajectory height available. Point G represents the limit of the ejection trajectory height solely and is directly or indirectly influenced by the ejectee’s ability to tolerate ejection acceleration without spinal injury. The line between points F and G represents the trade-off between the trajectory height and horizontal velocity parameters. None of these limits represent an insurmountable barrier. The altitude limit has been overcome by use of the pressure suit and the enclosed escape system. The dynamic pressure limit has been extended by enclosing the ejectee to eliminate windblast injury and by improving the drag characteristics of the escape system to reduce the magnitude of the rapid deceleration. Furthermore, current research and development efforts are disclosing methods to reduce the low-altitude escape barriers. Nevertheless, advances in aerospace-system performance and the concentration of fatalities outside of current escape-system performance envelopes continue to spotlight vividly the requirement for expansion of the escape envelope.

Fig 1.  Typical escape-system performance

Figure 1. Typical escape-system performance  

Determining the ability of the human to withstand a given environmental stress is basically the same problem that confronts the aerospace-vehicle designer when he must determine whether a vehicle subsystem can function within the environment. The vehicle designer may test the subsystem to failure numerous times, enabling him to describe with a relatively high degree of accuracy such factors as the performance decrement with each increase in stress, as well as the variability of the level of stress causing system failure. The point of human failure, however, may be the level of stress that will cause physical or psychological injury or even fatality, so of course destructive tests with a living human are impossible. Unfortunately, the human factors that limit the performance envelope of escape systems are indeed ones involving injury and fatality. Thus the aeromedical problem is a paradox in that the level of stress that will cause injury must be defined without deliberately causing injury in the process. To circumvent this situation, injury limits must be investigated by use of accident data, cadaver tests, operational experience, and laboratory tests to stress levels which provoke pain rather than injury. All these techniques are laden with problems that eliminate them as an individual solution. The purpose of this article is to review the development of aerospace escape systems from the aeromedical standpoint and describe recent advances that have made it possible to define more completely the human factors limiting the escape-system performance envelope. Performed within the current envelope limits, ejections have reached a success rate of 95 percent; however, the overall success rate is 88 percent.1

historical background

Early in World War II the German Force recognized the need for a method of assisted escape from disabled high-performance aircraft. By the end of the war they had developed and flight-tested the ejection seat, and it had been used approximately sixty times.

In 1943 the Allied forces began to consider seriously the requirement for an escape-system. A study of accident reports conducted at that time by the United States Army indicated that 12.5 percent of all emergency bailouts accomplished in the preceding 12-month period had resulted in fatal injuries. Nonfatal injuries were experienced by 45.5 percent the personnel involved in the bailouts. Of aviators abandoning single-engine aircraft, approximately 24 percent had been fatally injured. The alarming nature of these statistics was emphasized when they were compared to parachute training records. These records revealed that less than 1 percent of personnel involved in parachute training exercises were killed, and 1.5 percent injured. Analysis of emergency bailout reports indicated that escape was complicated by the centrifugal forces developed by the disabled aircraft, collision with the aircraft structure after egress, and lack of sufficient time to accomplish escape.

The British began investigating methods of assisted escape in 1944, and by 24 July 1945 they had completed their first in-flight ejection test using a human subject.

In August 1945 the United States Army Air Forces initiated an effort to design an ejection seat to provide an escape capability for the F -80. The basic design concept was patterned after the German Heinkel 162 ejection seat. Study of the German seat and its catapult indicated that they were inadequate for use in the F-80. The ejection velocity of the seat did not provide sufficient clearance between the seat and the vertical stabilizer at the maximum airspeed of the F-80. Hence primary emphasis was given to the development of a new ballistic catapult.

The most fundamental problem in the design of an ejection catapult is in assuring that the ejection velocity is adequate to provide safe clearance between the seat trajectory and the aircraft tail surfaces. In a ballistic catapult, the velocity is limited by the length of the ejection stroke and the magnitude of the seat acceleration. Since the ejection stroke of a ballistically propelled seat is determined by the depth of the cockpit, the required seat velocity must be obtained by maximizing the catapult acceleration.  This must be accomplished within the restrictions imposed by the capability of the human anatomy to withstand the acceleration without injury. The injury most commonly experienced during the catapult phase of the ejection sequence is compression fracture of the vertebrae of the spinal column, with associated damage to the surrounding muscle and ligamentous structures.

Until 1945 little effective research had been accomplished to determine man’s tolerance to short-duration acceleration. German scientists had conducted numerous ejection experiments with human subjects,2 but their instrumentation was considered inadequate, so that many of their conclusions were viewed with considerable skepticism by scientists in the United States.3 The conclusions that were more seriously considered were based on data from tests of cadaver vertebrae under static load conditions.  The data from these tests indicated that the vertebral segments would fail at 23 to 25 g. Since it is not possible to extrapolate directly to the dynamic loading situation, these data did not represent the answer to the catapult design problem, but they were a basic piece of the puzzle and a reasonable starting point.

The response of elastic systems to static versus dynamic loads had been studies by Frankland,4 and so the rate of application of the acceleration was recognized as a critical parameter in optimizing the catapult acceleration. More specifically, from a theoretical standpoint it is possible to use the static tolerance limits only if the acceleration is applied so slowly that the loads within the elastic systems, in this case the human spine, do not dynamically overshoot the input loads, i.e., the seat acceleration. The experimental effort was oriented toward finding a workable trade-off between the parameters of acceleration magnitude and acceleration onset rate. The experimental effort was conducted by the Army Air Forces Aero Medical Laboratory at Wright Field. Using catapults developed by the Frankford Arsenal, the laboratory exposed volunteer human subjects to progressively higher catapult accelerations until the required seat velocity had been reached and proved tolerable. Two of the subjects participating in the aeromedical experiments were concurrently being prepared to flight-test the ejection seat, an effort being managed by the Aircraft Laboratory and the Flight Test Division at Wright Field. The ejection-seat research and development effort reached a major milestone on 17 August 1946 when the first human ejection within the United States was completed from a P-61B aircraft.

The first operational ejection did not occur until three years later. On 29 August 1949 a successful ejection was accomplished from an F-86A at an airspeed of approximately 500 knots and an estimated altitude of 1000 to 2000 feet.

From the very beginning of the development of escape systems, human tolerance to the acceleration environment has been a major stumbling block. As soon as the acceleration criteria for the ejection catapult had been provided, it was recognized that the major human tolerance problem remaining to limit the useful range of the ejection seat was rapid deceleration following ejection at high airspeed and air density. Once again the limitations of the human body were found to be inadequately defined, and therefore the required limit of the escape-system performance envelope could not be established. The first approximation of the envelope was accomplished by ejecting human subjects from a flight-test aircraft. Ejections were accomplished at progressively higher airspeeds-up to 482 knots true airspeed at 9000 feet altitude—without injury to the subjects. In view of these tests the investigators concluded that ejection could be safely accomplished to approximately 600 knots equivalent airspeed (EAS). This capability appeared to be adequate for the high-performance aircraft of that time period.

In order to provide better estimates human tolerance to the ejection-seat deceleration and windblast under more controlled conditions, the Aero Medical Laboratory conducted a series of human experiments using rocket-sled techniques.5 This extensive series of tests provided data that remain the basis for human tolerance criteria for rapid deceleration when the acceleration is applied chest to back or back to chest.

acceleration tolerance parameters

From the experimental evidence available and an understanding of the basic mechanics involved, it is now possible to identify the parameters that directly or indirectly determine if the ejectee will be injured by the ejection acceleration. These parameters are acceleration environment, acceleration modification, and imposed acceleration versus the ejectee’s load-bearing ability. The most fundamental parameters in determining the potential for injury are the characteristics of the acceleration imposed at the interface between the subject and his environment and his ability to tolerate the acceleration. The characteristics of the acceleration are defined in terms of the acceleration-time function, i.e., they are defined primarily by the time to maximum acceleration, the duration of maximum acceleration, and the duration of the entire acceleration. From the earliest human testing, the rate of application of acceleration (the rate of onset) or the time period to the maximum acceleration level has been known, at least empirically, to have a direct effect on the tolerability of the acceleration environment. More specifically, higher acceleration levels have been observed to be more easily tolerated if the rate of onset of acceleration is reduced. In fact, as mentioned earlier, if the acceleration is increased gradually, it is feasible to reach an acceleration level equal to the static load that would cause injury. Conversely, if the acceleration is applied too rapidly, the acceleration within a mechanical system such as the human body may theoretically exceed the input acceleration by as much as a factor of two if the system is undamped.6  

The ability of the human body to tolerate the acceleration is quite complex and is dependent upon factors such as position of the body with respect to the acceleration vector, maintaining the body position during acceleration, and the subject’s age, musculoskeletal development, etc. As an example of the magnitude of effect of these factors, the tolerance to acceleration applied parallel to the spinal column (+ Gz) is approximately one-half of that which can be tolerated when the acceleration is applied perpendicular to the spinal column (+ Gx) as shown in Figure 2. With the exception of the body position, these factors are fixed in each subject, but they vary widely in the total population and therefore must be handled on a statistical basis.

Fig 2.  Acceleration terminology, arrows indicating direction of intertial responses

Figure 2. Acceleration terminology, arrows indicating direction of inertial responses  

The imposed acceleration is controlled by two rather broad factors: the characteristics of the acceleration environment and the modification of the acceleration by structure and equipment interposed between the human body and the environment. Each of the factors contains many variables within it. The acceleration produced by the ejection catapult varies as a function of the weight of the seat/man combination, the preignition temperature of the catapult propellant, and variations in the catapult propellant. The degree of acceleration modification is dependent on the deformation characteristics of the structure and equipment through which the acceleration environment is transmitted to the subject. The characteristics of structure and equipment may either amplify or attenuate the acceleration.

The ejection-seat cushion was the first piece of equipment known to modify acceleration adversely.7 The deformation the seat cushion during ejection isolates the ejectee from his environment for a brief period, allowing the seat velocity initially to exceed significantly that of the ejectee. Once the seat cushion bottoms, the ejectee is rapidly accelerated until he and the seat reach the same velocity. In the process the acceleration of the subject usually exceeds the acceleration he would have experienced had he bee rigidly coupled to the seat. Other similar factors that might adversely modify the acceleration transmitted to the subject include elasticity of his restraint system, slackness of the restraint, and flexion of the seat structure.

The acceleration environment may be modified beneficially by interposing acceleration attenuating components, such as crushable foam, hydraulic shock absorbers, crushable honeycomb, air bag decelerators, etc. Such components have been used in escape systems of the B-58 and B-70 capsules and the F-111 crew escape module as well as the Project Mercury and Project Apollo vehicles.

human analog

Two difficult problems in the development of escape systems have been, first, the lack of data on the capability of the human being in the hostile environments sometimes encountered in emergency escape; and, second, the ambiguity and inconsistency of the data that have been made available to the escape-system designer. The first problem stems from the inherent difficulties in conducting human tolerance research and from the discontinuity of research emphasis on the escape-system development problem. The second problem developed initially out of the inconsistencies in the methodology used by the first investigators to describe the acceleration environment and the associated human response. This situation has been perpetuated by the slow pace the research and by lack of a unifying methodology.

Within the last five to ten years there has been a gradual improvement in the methodology. The most significant advance has resulted from the use of analogs to describe the human response to mechanical environments, such as the short-duration acceleration associated with ejection. Since the response of the human body that results in injury is mechanical in nature (i.e, deformation of body tissue to the point of failure), the analog is a mechanical system composed of elements analogous to the mass, elasticity, and damping properties for the body. The response of the analog is describable by differential equations containing terms representing the positions of the mechanical elements with respect to time. Thus, by computation of the response of the analog elements for a given acceleration input, the response of the human body can be predicted. The accuracy of the analog may be determined by comparing the computed response of the analog with the measured response of the human body to noninjurious acceleration environments created in the laboratory and also by analysis of the environments known to have caused accidental injury.

The human response to short-duration accelerations applied parallel to the vertebral column (+ Gz) is currently studied by use of a mechanical analog, since the critical mode of injury in this axis, compression fracture of the vertebrae, is a structural phenomenon. The simplest analog for this axis is a mechanical model composed of a mass, a spring, and a damper. The system elements are lumped-parameter elements, e.g., all the mass of the human body that acts upon the vertebrae to cause deformation is represented by the mass element. A diagram of the analogous model is shown in Figure 3. This model is used to predict the maximum deflection and associated force within the vertebral column for any given short-duration acceleration environment. The properties of the model elements have been derived from existing data. For instance, the spring stiffness has been determined from tests of cadaver vertebral segments, and the damping ratio has been calculated from measurements of mechanical impedance during vibration and impact of human subjects.8

Fig 3.  Spinal-injury model

Figure 3. Spinal-injury model  

Study of the spinal-injury model provides a great deal of insight into the findings of earlier experimental evidence. It reveals rather simply that there is a physical basis for the observation that the acceleration level that may be safely tolerated is a function of the time required to reach the acceleration level. In this instance the experimental data can be used to verify the adequacy of the model.  This verification provides additional confidence in the model’s ability to predict the human response in environments that have not been as thoroughly investigated.

One of the greatest limitations of the method previously used to describe human tolerance to spinal injury was the oversimplification to fit the acceleration-time history of the environment in terms of a trapezoidal wave form. The trapezoidal wave form was selected because it fit the accelerations produced within the laboratory and, at least initially, the operationally encountered environment. Unfortunately, as escape systems became more complex, the acceleration environment became complex. With the advent of the escape capsule, the complexity of the acceleration environment reached a level that could not be handled by the method.

The acceleration environment of the escape capsule represents a major departure from the open ejection seat, since the ejectee remains within the capsule during parachute deployment, descent, and landing phases of the escape sequence. Because of variability of the descent velocity, oscillation of the capsule beneath the parachute, horizontal wind drift, and nature of the landing surface, the acceleration wave forms during landing impact are highly irregular and the direction of the acceleration vector unpredictable. It was this almost insurmountable problem that provided the greatest stimulus to the analog development. The human analog has not evolved to the point of sophistication that the problem of changing direction of the acceleration vector can be handled adequately, but the irregularity of the wave form is no longer a difficult problem. The limitation thus far is the lack of biological data to incorporate into the human analog.

operational verification

The human analog provides more than a basis for handling the complex acceleration profiles of the operational world; it provides a statistical method of predicting the injury potential of any given environment. Since the biologically important response of the analog can be measured in terms of a single parameter, such as the peak force within the system or the corresponding deflection of the system, the analog avoids the problem of relating the multiple descriptors of the environment, i.e., acceleration magnitude, pulse duration, time to peak acceleration, etc., which may vary individually or in combination. Using the analog, the analysis of environments known to cause injury can provide a correlation between the single human response parameter and the risk of injury associated with it.

Some insight into the probability-of-injury relationship can be seen from the cadaver test data that have been used to develop the stiffness characteristics of the spinal-injury model. As one might expect, the breaking strengths of the tested vertebral segments varied. The breaking strengths of these specimens can be statistically described. These values appear to be normally distributed about the mean value of the breaking force or deflection. Thus, for a given force or deflection there is a specific probability of failure. However, since these breaking strengths are obtained from research with cadaver material, the direct application of these data to the operational circumstance must be limited.

To provide correlation between the predictions of the model and the injuries being experienced operationally, an analytical effort has been in progress to collect data describing the accelerations produced by aircraft escape systems and to determine the spinal compression fracture rates associated with the use of each system. This proved to be no simple matter. Either instrumented ejections had not been accomplished or there had been so few instrumented ejections and the measurements so varying that they could not be described with any statistical confidence. However, a large body of data was found to exist from catapult qualification firings. These data were collected under controlled conditions and were sufficient in number to allow statistical analysis. Use of catapult qualification data has the potential failing of not being representative of the environment that causes the reported injury. The inherent variability of catapult performance due to preignition temperature of the propellant, the ejected weight, and the variability of the propellant make it virtually impossible to estimate the actual environment that caused the injury. Nevertheless, from the standpoint of controlling the design of catapults by use of the analog approach, use of the catapult qualification data provides an important advantage. Briefly, since the catapult performance data are usually the only comprehensive and statistically valid data collected during system development, it appears practical from a system design standpoint to relate the statistical distribution of qualification test results to the rate of operational injury. Conversely, although very interesting from a research standpoint, it did not appear practical to determine an absolute injury level and then attempt to relate this level to the probability of its occurrence operationally.

The relationships between the acceleration environment and the risk of injury that have been derived to date are given in Figure 4. The response parameter of the model has been nondimensionalized and expressed in terms of dynamic response index (DRI) values. The probability of injury determined from the cadaver data is indicated by the broken line. The operational escape-system data points have been obtained by calculating the DRI from an acceleration-time history representative of the escape system’s mean catapult performance characteristics at nominal conditions of weight and temperature and then determining the actual spinal-injury rate associated with the system. Only compression fractures of the vertebral column that have been attributed to ejection acceleration, rather than landing impact, are used. Each data point represents the injury rate associated with at least 25 successful nonfatal ejections.

As might be expected, the operational system data points in Figure 4 are in most cases somewhat higher than would be predicted by the cadaver data. The more obvious reason for this difference would be that intact living vertebral column embedded in the torso would almost certainly be stroll than cadaver segments. Since there are relatively few data points, it is not possible to describe a complete distribution of the DRI versus probability of versus probability of injury. Nevertheless the distribution should be similar to the distribution of the cadaver breaking strength.  Such a predicted distribution is represented by the solid line in Figure 4. The points that represent the Navy F-4B and the Air Force F-4C ejection seats deviate from this prediction, the latter considerably. This can be partially explained by the fact that the seat of F-4B and F-4C is so designed that the axis of the ejectee’s spinal column is not aligned with the acceleration vector of the catapult. The spinal column may be misaligned as much as 10 to 16 degrees, so that during ejection the spinal column flexes forward, creating a greater load concentration on the anterior surfaces of the vertebrae, which results in a greater compression fracture potential.9 The more deviant position of the F-4C point is caused by the poorer degree of restraint, a different survival kit-seat cushion configuration, and the operational use of the between-the-legs D-ring as the primary ejection actuation control. In the Navy F-4B the face curtain is used primarily rather than the D-ring. The face curtain provides head restraint that retards flexion of the torso and spinal column. Use of the D-ring exaggerates the tendency for flexion.

Fig 4.  Probability of spinal injury predicted from cadaver data compared to operational experience

Figure 4. Probability of spinal injury predicted from cadaver data to operational experience  

An additional source of variation in the data plotted in Figure 4 is the difference in the cushions used with the various ejection seats. Use of the modeling approach also permits depiction of the mechanical response characteristics of components between the ejectee and the acceleration source which might modify the imposed acceleration.10 Recent ejection tests using the F-4 ejection seat have illustrated this point. In these tests the inertial response of the human subject was measured by placing force transducers between the seat and the subject. Figure 5 compares the measured force with the computed force which would have been measured had the subject been a rigid mass. Figure 6 compares the measured force and computed force when a seat cushion was placed between the seat pan and the subject.

Fig 5.  Comparison of computed and measured force with subject sitting on a rigid seat pan

Figure 5. Comparison of computed and measured force with subject sitting on a rigid seat pan  


Fig 6.  Comparison of computed and measured force with subject sitting on a seat cushion

Figure 6. Comparison of computed and measured force with subject sitting on a seat cushion  

Using the analog as a tool, the equipment designer finds it feasible to design the escape-system components quantitatively rather than use the former “cut and try” methods. Furthermore, in designing the seat cushion one might conceivably be able to determine analytically a reasonable trade-off between acceleration protection and crew comfort. These two requirements are currently considered to be antagonistic.

application of the analog

The F-111 crew escape module is a radical departure from previous escape systems. The crew module is an integral part of the forward fuselage of the aircraft, encompassing the pressurized cabin and forward portion of the wing glove. During ejection the crew module is severed from the aircraft fuselage by a shaped charge train and then propelled away from the aircraft by a solid-propellant rocket motor. After separation the module stabilization glove provides stability and aerodynamic lift. A parachute provides additional stabilization and deceleration of the module. The recovery parachute is forcibly deployed, and opening shock is minimized by reefing the parachute for 2.5 seconds. Landing impact on ground or water is attenuated by controlled gas expulsion from an impact-attenuation bag with blowout plugs. During development testing of the crew module it became apparent that the acceleration environment produced during the catapult phase of the escape sequence would be unlike that of previous egress systems. The acceleration data obtained in ejections from a high-speed rocket sled indicated that the acceleration experienced in the axis parallel to the vertebral column (+Gz) is significantly influenced by the airspeed and air density. The acceleration profiles shown in Figure 7 illustrate the effect. The first acceleration pulse, occurring in the initial 0.05-second period, was not apparent at all in either the development computer studies or wind-tunnel tests. The second, longer-duration pulse had been predicted; however, the magnitude appeared to be greater than anticipated. The first pulse, referred to as the “popgun” effect, represented a challenge to the traditional acceleration limit parameter, the rate of onset of acceleration. Specifications governing the procurement of previous escape systems limited the maximum allowable rate of onset in this direction to 300g/sec. The rate of onset of the popgun pulse is approximately 1000g/sec. Even though the popgun pulse exceeded established design limits, data from Air Force and contractor experimentation with human subjects during the development of the B-58 escape capsule indicated that such a pulse shape might be tolerable. Furthermore, analytical work with the spinal-injury analog provided a theoretical basis for the belief. At this point in the crew-module development program it became apparent that application of the analog might prove fruitful. The alternative was redesign of the escape system. With the acceleration-time history recorded from the track test ejection at 250 knots (equivalent airspeed) as the input function, the DRI-time history was computed. The computation results are shown in Figure 8. Assuming that use of the spinal analog provides a reasonable prediction of the response of the vertebral column, these results show that the inertial responses within the human do not correspond proportionately to the acceleration input function. They reveal that the time duration of the popgun pulse is so brief in terms of the dynamic response characteristics of the analog that the response is significantly reduced. The spinal-injury analog has since been more thoroughly explored and has been incorporated into the design guidance for the F-Ill crew escape module as well as all future escape-system development efforts.

Fig 7.  Comparison of acceleration of the F-111 crew escape module during ejection from a rocket sled at two equivalent airspeeds: 250 knots (above) and 450 knots (below)

Figure 7. Comparison of acceleration of the F-111 crew escape module during ejection from a rocket sled at two equivalent airspeeds: 250 knots (above) and 450 knots (below)  


Fig 8.  Acceleration recorded from F-111 crew escape module and computed analog response

Figure 8. Acceleration recorded from F-111 crew escape module and computed analog response  

The modeling approach is also being applied to the problem of describing the influence of the motion of the ejectee’s center of gravity upon escape-system trajectories. When the rocket catapult was introduced, its purpose was to extend the length of the ejection acceleration stroke beyond the limits of the cockpit depth. This enabled the escape system to clear the aircraft vertical stabilizer at extremely high speeds without increasing the acceleration magnitude. Unfortunately, without the constraint of the rails, the rocket catapult thrust vector must be directed through the center of gravity of the seat/man combination, else adverse pitch or roll moments will result. With high-impulse rockets, the condition results in rotation of the seat during rocket burning. This unstable condition increases the possibility of seat/man/parachute interference and limits the trajectory height at low airspeeds.

It is difficult to design for the center of gravity of the seat/man combination in the operational circumstance, since its position is not fixed. First, the center of gravity varies as a function of the anthropometric characteristics of the ejectee. Second, the weight and position of equipment worn by the crewman will add to the overall variability. Finally, during the catapult firing the human body will slump and the equipment will be displaced, so that the target for the thrust vector, the seat/man center of gravity, becomes a moving target. Laboratory study of the motion of the center of gravity of the human body during short-period acceleration shows that the motion can be predicted by a simple mechanical analog. Several systems have been developed to counteract the effect of the center-of-gravity eccentricity. The simplest concept employs a bridle and drag line that introduce opposing moments. A more complex system involves the use of a gyro-controlled vernier rock assembly.11

For two reasons this discussion of aeromedical efforts to support the development of aerospace escape systems has concentrated largely on defining spinal-injury limits. First, spinal injury has traditionally concerned potential users of escape systems because that risk has been associated with the use of escape system ever since their inception. Second, the technical approach used to investigate and define the causative factors and to quantify the probability of this injury mode is illustrative of the technical approach currently being use to investigate human tolerance to short-duration acceleration applied in other directions. Furthermore, this approach will be used to define more completely man’s response to other environmental hazards associated with emergency escape, such as windblast, parachute-opening shock, angular motion, and the combination of such stresses as are still inadequately understood. The increased performance capability of advanced aircraft demands that these definitions be provided if injury and fatality rates are to be maintained at or below current levels.

In the past the development of each new escape system has disclosed the inadequacies of our knowledge of the human being. As a result much of our research has been committed to solving these problems as they have appeared. Hopefully the lesson has been learned. Additional emphasis is being placed on the problems of advanced systems. The need to develop operational escape systems for space vehicles adds urgency to this requirement. A more comprehensive evaluation of the probability of injury or fatality must be provided and related to the likelihood of occurrence of a given stress, so that the basic feasibility of space escape systems may be determined. The analog provides a powerful tool in this task, but the fundamental biological properties that limit man’s capability must be more precisely defined by an increased research effort if the analog is to be maximally useful.

Aerospace Medical Research Laboratories


1. R.H. Shannon and C.H. Sawyer, “man/Seat/Chute Interference in USAF Ejections,” SAFE Engineering, Vol. 1, No. 4, October/November 1967.

2. S. Ruff, “Bried Acceleration: Less Than One Second,” German Aviation Medicine in World War II, Chap. IV-C, Vol. I, U.S. Government Printing Office, Washington, D.C., 1950.

3. D.T. Watts, A.B. Mendelson, H.N. Hunter, A.T. Kornfield, and J.R. Poppen, “Tolerance to Vertical Acceleration Required for Seat Ejection,” Aviation Medicine, December 1947.

4. J.M. Frankland, Effects of Impact on Simple Elastic Structure, Navy Department, David Taylor Model Basin, Washington, D.C., Report 481, April 1942.

5. J.P. Stapp, Human Exposures to Linear Deceleration, AF Technical Report No. 5915, Part 1, June 1949, and Part 2, December 1951.

6. Frankland, op. cit.

7. F. Latham, “A Study in Body Ballistics: Seat Ejection,” Proceedings of the Royal Society, B, Vol. 147, 1957.

8. R. R. Coermann, The Mechanical Impedance of the Human Body in Sitting and Standing Positions at Low Frequencies, ASD Technical Report 61-492, Aerospace Medical Research Laboratories, September 1961. The sources of data and modeling techniques have been more thoroughly described by E. L. Stech and P. R. Payne in Dynamic Models of the Human Body, AMRL-TR-66-157, in press.

9. G. C. Mohr, J. W. Brinkley, L. E. Kazarian, and W. W. Millard, “Variations of Spinal Alignment in Egress Systems and Their Effect,” to be published in Aerospace Medicine.

10. P. R. Payne, Personnel Restraint and Support System Dynamics, AMRL-TR-65-127, October 1965.

11. H. R. Moy, B. Nichols, Jr., and R. G. McIntyre, “A Rotation and Trajectory Control System for Rocket Catapult Ejection Seats,” SAFE Engineering, Vol. 1, No.1, March/April 1967.


James W. Brinkley (B.S., Ohio State University) is a research scientist, Aerospace Medical Research Laboratories, Air Force Systems Command, Wright-Patterson AFB, Ohio. Since his graduation in 1958 he has been involved in the development of escape systems for the B-58, B-70, and F-l11 and in research in support of Projects Mercury, Gemini, Apollo, and the Manned Orbiting Laboratory. Mr. Brinkley has written numerous technical papers and reports dealing with acceleration tolerance, restraint systems, and impact attenuation. He was named AMRL Scientist of the Year in 1966.


The conclusions and opinions expressed in this document are those of the author cultivated in the freedom of expression, academic environment of Air University. They do not reflect the official position of the U.S. Government, Department of Defense, the United States Air Force or the Air University.

Home Page | Feedback? Email the Editor