      |
|||
|
|
The original motivation of this
action group was the adverse interactions between the
human pilotand the aircraft dynamics, called PIO (
pilot-in-the-loop oscillations) or APC ( aircraft-pilot
coupling), which occurred not only on highly manoeuvrable
fighter aircraft, but also on modern civil airliners with
fly-by-wire flight control systems (accidents of US
American YF-22, the Swedish JAS39, high altitude upset of
MD-11). PIO can be considered as a closed loop
destabilisation of the aircraft-pilot loop, and often
occurs under situations when the pilot proves to be
unable to adapt himself to a sudden change of the vehicle
dynamics during a high demanding flying task. Three categories of PIO were
recently distinguished [1]: -category I PIO concerns mainly
linear characteristics of pilot and vehicle
oscillations -category II PIO is characterised
by quasi-linear pilot and vehicle models, but with
non-linear effect on rate and /or position limiting
effects, -category III PIO is
characterised by highly non-linear pilot-vehicle
interactions including multiple dominant non-linearities
and transitions in the pilot and aircraft behaviour, such
as mode changes, modification in cues (e.g. from attitude
to load factor). The existing handling qualities criteria
were established mainly for category I PIO and can be
considered as effective, which means that it is possible
to avoid PIO due to linear effects (lags, time delays,
etc.) by design. Currently, no criterion really exist or
validated for category PIO II, and it is too difficult to
find general category III PIO prevention techniques due
to the rich variety of highly diverse phenomena causing
it. Following the recommendations of
the GARTEUR Exploratory Group EG-18, the GARTEUR Action
Group FM-AG-12 began the work in April 1999, and gathered
several participants from 12 European organisations (
industrials , Universities and five research
establishments) working during a two and a half year
research program, with an emphasis on cat. II PIO
problems. The objectives of the action
group were : - to develop procedures combining
experiments and effective analysis methods which prove
that a highly augmented aircraft is sufficiently free
from PIO proneness. - to initiate the Concept of a
European Handbook for PIO testing of highly augmented
aircraft. The outcome of this work would
aim at fulfilling industrial needs for the design of
PIO-free flight control systems, and the safety of flight
tests of high augmented aircraft . The work programme covered three
aspects with the following objectives : 1) Experimentation on
ground-based simulators in order to find the manoeuvres
that best identify PIO tendencies. Even if all the PIO
aspects could not be covered by the pilot-in-the-loop
groundbased simulation, it was thought that specific and
selective simulation tasks could play an effective role
at a stage of discovery search of PIO events, before
proceeding to flight test. Existing PIO data-bases were
completed by new simulator experiments in order to
correlate the analysis results with pilot
comments. 2) Development of analysis
techniques and criteria that best predict PIO tendencies.
Several new methods were developed and their capability
to predict whether or not a given pilotvehicle system
(PVS) is PIO prone were compared with piloted simulations
: the Time Domain Neal-Smith Criterion (TDNS), the
Describing Function/Vector margin(ROBAN),the Describing
Function/Mu-setting, and the Bifurcation approach. With
the exception of the first method (TDNS) which was
already known, the others methods were specifically
developed within the action group. 3) Development of on line
algorithms that best predict the PIO phenomenon and best
compensate the effects of rate limiting. The adverse effects of rate
saturation can be alleviated by new rate limiter
concepts, which limit the maximum rate, but introduce
less phase lag. These concepts are called phase
compensation rate limiters or phase compensation filters.
Several approaches have been developed by different
organisations, such as by NASA, DASA, DLR, SAAB. SAAB has
a lot of operational experience with phase compensation,
since their approach is installed in the production
software of the JAS39 aircraft. The filter has been
validated by extensive simulator and in-flight testing.
However, there was still a need for further research work
on this subject. A new technique, based on H-infinity
approach, was developed for this purpose.
All these studies were applied to
a realistic aircraft model ADMIRE which was developed by
FOI (Sweden), based on the Generic Aerodata Model (GAM),
developed by SAAB AB. ADMIRE is implemented in
Matlab/simulink environment and represents a single seat
fighter aircraft with a delta-canard configuration,
covers a large domain of flight envelop (high angle of
attack, up to Mach 2.5), with several inputs such as left
and right canard, leading edge flap, four elevons, rudder
and thrust setting. Five ground-based flight
simulators were used in the study, both fixed and moving
base : they are located respectively at FOI in Sweden
(FOSIM -moving base- and FENIX -fixed base), at
NLR’s National Simulation Facility in the
Netherlands (NSF moving base), and at CEV/Istres in
France ( SEM and M-2005 fixed bases ). Experimental results The simulators campaign results
were : - a collection of data and PIO
ratings to validate off line theoretical analysis results
from the new PIO prediction methods, - an assessment of the
effectiveness of the tracking tasks developed to detect
PIO. Several pilots, from several organisations (CEV,
FMV, EADS, SAAB, NLR, DLR), with various background
(research organisations, industrial, tests pilots)
participated to the simulation trials. On the overall, the HUD tracking
tasks, pitch attitude capture and hold and bank angle
capture and hold proved to be similar in effectiveness.
The HUD tracking task for the combined axes seemed to be
the most promising . It was suggested that further
development of experimental methods should be focussed on
multiaxes acquisition manoeuvres to predict PIO
proneness. In general, analytical
predictions were in good agreement with the pilot ratings
for all the cases with at least one of the four methods.
The best results seemed to be obtained with the following
methods, in the decrease order of quality: the
bifurcation approach, the musetting, the robust analysis,
the Time domain Neal Smith criterion. Nevertheless, these
results have to be considered with caution, because
either the shortage of available data has not allowed to
fix comprehensive limits for the criteria, or the use of
original limits has led to contradictory results in some
cases. By processing further the newly generated data
bases, analysis work has to be continue to refine the
limits defined in the criteria and to further the
understanding of the relation with the pilot’s
rating. The new H-infinity method of
rate-limit compensation was found to be useful in terms
of both effectiveness and ease of tuning, when compared
to the two other approaches (DLR and SAAB). More research remained to be
conducted on experimental verification of these results
and suitable methods for implementation. A concept for a European Handbook
(similar to the US MIL Standard [2]) for PIO
testing of highly augmented aircraft was delivered and
set in the GARTEUR web site. It should be updated
regularly by the users. This document is supported by a
toolbox, also available in the same web site, which
should help an aircraft manufacturer to develop PIO free
flight control systems or to test a given aircraft
against PIO. .
