Appendix 2-B - Aeroplane Simulator Validation Tests

4. Control Dynamics

The characteristics of an aeroplane flight control system have a major effect on the handling qualities. A significant consideration in pilot acceptability of an aeroplane is the "feel" provided through the cockpit controls. Considerable effort is expended on aeroplane feel system design in order to deliver a system with which pilots will be comfortable and consider the aeroplane desirable to fly. In order for a simulator to be representative, it too must present the pilot with the proper feel; that of the respective aeroplane. This fact is recognized in FAR121, AppendixH, PhaseII (LevelC), Simulator Requirement 10, which states: "Aircraft control feel dynamics shall duplicate the aeroplane simulated. This shall be determined by comparing a recording of the control feel dynamics of the simulator to aeroplane measurements in the takeoff, cruise and landing configuration".

Recordings such as free response to an impulse or step function are classically used to estimate the dynamic properties of electromechanical systems. In any case, it is only possible to estimate the dynamic properties as a result of only being able to estimate true inputs and responses; therefore, it is imperative that the best possible data be collected since close matching of the simulator control loading system to the aeroplane systems is essential. The required control feel dynamic tests are described in 2.B. of the Table of Validation Tests of this Appendix. For initial and upgrade evaluations, it is required that control dynamic characteristics be measured at and recorded directly from the cockpit controls. This procedure is usually accomplished by measuring the free response of the controls using a step or pulse input to excite the system. The procedure must be accomplished in takeoff, cruise and landing flight conditions and configurations.

For aeroplanes with irreversible control systems, measurements may be obtained on the ground if proper pitot-static inputs are provided to represent airspeeds typical of those encountered in flight. Likewise, it may be shown that for some aeroplanes, takeoff, cruise and landing configurations have like effects. Thus, one may suffice for another. If either or both considerations apply, engineering validation or aeroplane manufacturer rationale must be submitted as justification for ground tests or for eliminating a configuration. For simulators requiring static and dynamic tests at the controls, special test fixtures will not be required during initial and upgrade evaluations if the operator's QTG shows both test fixture results and the results of an alternate approach, such as computer plots which were produced concurrently and show satisfactory agreement. Repeat of the alternate method during the initial evaluation would then satisfy this test requirement.

5. Control Dynamics Evaluation

The dynamic properties of control systems are often stated in terms of frequency, damping and a number of other classical measurements which can be found in texts on control systems. In order to establish a consistent means of validating test results for simulator control loading, criteria are needed that will clearly define the interpretation of the measurements and the tolerances to be applied. Criteria are needed for underdamped, critically and overdamped systems. In case of an underdamped system with very light damping, the system may be quantified in terms of frequency and damping. In critically damped or overdamped systems, the frequency and damping is not readily measured from a response time history; therefore, some other measurement must be used.

For LevelsC and D simulators, tests to verify that control feel dynamics represent the aeroplane must show that the dynamic damping cycles (free response of the controls) match that of the aeroplane within specified tolerances of damping. The method of evaluating the response is described below for the underdamped and critically damped cases.

Underdamped Responses

Two measurements are required for the period, the time to first zero crossing (in case a rate limit is present) and the subsequent frequency of oscillation. It is necessary to measure cycles on an individual basis in case there are non-uniform periods in the response. Each period will be independently compared to the respective period of the aeroplane control system and, consequently, will enjoy the full tolerance specified for that period.

The damping tolerance should be applied to overshoots on an individual basis. Care should be taken when applying the tolerance to small overshoots since the significance of such overshoots becomes questionable. Only those overshoots larger than 5% of the total initial displacement should be considered significant. The simulator should show the same number of significant overshoots to within 1 when compared against the aeroplane data. This procedure for evaluating the response is illustrated in Figure1.

Critically Damped and Overdamped Response

Due to the nature of critically damped responses (no overshoots), the time to reach 90% of the steady state (neutral point) value should be the same as the aeroplane within ±10%. The simulator response should be critically damped also. Figure2 illustrates the procedure.

Tolerances

The following table summaries the tolerances, T. See Figures1 and 2 for an illustration of the referenced measurements.

T(P0) ±10% of P0
T(P1) ±20% of P1
T(Pn) ±10% of Pn
T(An) ±10% of A1, ±20% of Subsequent Peaks
T(Ad) ±5% of Ad
Overshoots ±1

Alternate Method for Control Dynamics

One aeroplane manufacturer has proposed, and TC accepts, an alternate means for dealing with control dynamics. The method applies to aeroplanes and artificial feel systems. Instead of free response measurements, the system would be validated by measurements of control force and rate of movement. For each axis of pitch, roll and yaw, the control shall be forced to its maximum extreme position for the following distinct rates. These tests shall be conducted at typical taxi, takeoff, cruise and landing configuration.

  1. Static Test - Slowly move the control such that approximately 100seconds are required to achieve a full sweep.
  2. Slow Dynamic Test - Achieve a full sweep in approximately 10seconds.
  3. Fast Dynamic Test - Achieve a full sweep in approximately 4seconds.

    Note: Dynamic sweeps may be limited to forces not exceeding 100lbs.

Tolerances

  1. Static Test - As per items in 2.B. of this Appendix.
  2. Dynamic Test - 2 lbs or 10% on dynamic increment above static test.

TC is open to alternative means such as the one described above. Such alternatives must, however, be justified and appropriate to the application. For example, the method described here may not apply to all manufacturers' systems and certainly not to aeroplanes with reversible control systems. Hence, each case must be considered on its own merit on an ad hoc basis. Should TC find that alternative methods do not result in satisfactory simulator performance, then more conventionally accepted methods must be used.

Figure 1 - Under-Damped Step Response

Figure 1 - Under-Damped Step Response (Displacement vs. Tme)


Figure 2 - Critically Damped Step Response

Figure 2 - Critically Damped Step Response (Displacement vs. Time)

6. Ground Effect

During landing and takeoff, aeroplanes operate for brief time intervals close to the ground. The presence of the ground significantly modifies the air flow past the aeroplane and, therefore, changes the aerodynamic characteristics. The close proximity of the ground imposes a barrier which inhibits the downward flow normally associated with the production of lift. The downwash is a function of height with the effects usually considered to be negligible above a height of approximately one wingspan. There are three main effects of the reduced downwash:

  1. a reduction in downwash angle at the tail for a conventional configuration;
  2. an increase in both wing and tail lift because of changes in the relationship of lift coefficient to angle of attack (increase in lift curve slope); and
  3. a reduction in the induced drag.

Relative to out-of-ground effect flight (at a given angle of attack), these effects result in higher lift in ground effect and less power required for level flight. Because of the associated effects on stability, they also cause significant changes in elevator (or stabilator) angle to trim and stick (column) forces required to maintain a given lift coefficient in level flight near the ground.

For a simulator to be used for takeoff and in particular landing credit, it must faithfully reproduce the aerodynamic changes which occur in ground effect. The parameters chosen for simulator validation must obviously be indicative of these changes. The primary validation parameters for longitudinal characteristics in ground effect are:

  1. elevator or stabilator angle to trim;
  2. power (thrust) required for level flight (PLF);
  3. angle of attack for a given lift coefficient;
  4. altitude/height; and
  5. airspeed.

This listing of parameters assumes that the ground effect data is acquired by tests during "fly-bys" at several altitudes in and out of ground effect. The test altitudes should, as a minimum, be at 10%, 30% and 70% of the aeroplane wingspan and one altitude out of ground effect, e.g. 150% of wingspan. Level fly-bys are required for LevelD, but not for LevelsC and LevelB. They are, however, acceptable for all levels.

If, in lieu of the level fly-by method for LevelsB and C, other methods such as shallow glidepath approaches to the ground maintaining a chosen parameter constant are proposed, then additional validation parameters are important. For example, if constant attitude shallow approaches are chosen as the test manoeuvre, pitch attitude and flight path angle are additional necessary validation parameters. The selection of the test method and procedures to validate ground effect is at the option of the organization performing the flight tests, however, rationale must be provided to conclude that the tests performed do indeed validate the ground-effect model.

The allowable longitudinal parameter tolerances for validation of ground effect characteristics are:

Elevator or Stabilator Angle ±1°
Power for Level Flight (PLF) ±5%
Angle of Attack ±1°
Altitude/Height ±10% or ±5_(1.5m)
Airspeed ±3 Knots
Pitch Attitude ±1°

The lateral-directional characteristics are also altered by ground effect. Because of the above-mentioned changes in lift curve slope, roll damping, as an example, is affected. The change in roll damping will affect other dynamics modes usually evaluated for simulator validation. In fact, Dutch-roll dynamics, spiral stability and roll rate for a given lateral control input are altered by ground effect. Steady heading sideslip will also be affected. These effects must be accounted for in the simulator modelling. Several tests such as "crosswind landing", "one engine inoperative landing" and "engine failure on takeoff" serve to validate lateral-directional ground effect since portions of them are accomplished while transiting altitudes at which ground effect is an important factor.

Date modified: