Method for differentiating control failures in a system controlling an actuator, in
particular a gas-turbine engine stator
Field of the invention
5
The present invention relates to the field of gas-turbine engines, more particularly engines
equipping aircraft, and relates to the detection of faults in a system for controlling the
members modifying the operating state of the engine, such as the system regulating the pitch
angle of the variable-pitch fins of the compressor stator.
10
Prior art
In an engine equipping an aircraft, in particular a military aircraft, it is generally sought to
manage failures in the regulation system so as to afford better tolerance to faults and to
15 reduce the mission cancellation rate.
According to a known regulation method, control of the engine is based on a regulation
architecture with two interlinked levels. This is composed of engine regulation loops and local
regulation loops: the first control the state of the engine, for example its rotation speed or the
20 output speed of the LP compressor; the second control the position of the actuators acting on
the members modifying the state of the engine, which are the fuel metering devices, the input
guiding wheel, the variable pitch stator fins, the exhaust nozzle or others.
The local regulation groups are composed of three parts: a computing part, a control part and
25 a measuring part. The computing part produces the electric-current command to be applied
to the actuators. This command is produced so as to ensure the convergence of the actual
position of the actuators towards the desired position. The control part positions the output
member of the actuator according to the current command. Finally, the measuring part
assesses the position of the output member of the actuators.
30
The computing part is solely electrical and the control part also comprises mechanical
means.
The electrical part is duplicated. It is generally of the dual-channel type and generally
35 comprises two computers, each being able to send electric current to the actuators. The
WO 2014/049260 PCT/FR2013/052238
2
actuator is however controlled by only one computer at a time: the active computer. The
other computer - the passive computer - is ready to take over in the event of malfunctioning
of the first. There is then switching of the computer: the active computer becomes passive
and vice versa.
5
As for the mechanical part, this is not duplicated; there is only one mechanical control
channel.
Among the faults affecting the system controlling the focal regulation groups, the faults are
10 distinguished by the degree of importance of their magnitude. A simple fault corresponds to a
transparent or temporary failure of the chain of command, whereas a double fault
corresponds to a definitive failure.
Faults of electrical origin are also distinguished from those that are of mechanical origin since
15 the maintenance actions to be performed in the case of simple control faults of the electrical
type are not the same as those to be carried out in the case of simple control faults of the
mechanical type. However, with certain engines, the type of control fault is not diagnosed by
the regulation system. There are then risks that the maintenance actions are not suitable.
20 In the case of a dual-channel electrical control, when a simple control fault occurs, it causes
the computer to switch. The failed electrical control channel becomes passive. Thus the
single control fault is transparent in flight. However, for safety reasons, the maintenance
procedures require the dismantling of an item of equipment - the computer or control unit
concerned - on the ground.
25
The legitimacy of this dismantling varies according to the origin of the single control fault. It
results either from the failure of one of the two electrical control channels, or from a
temporary failure of the mechanical control channel. The latter may for example be transient
contamination of the fluid driving the actuator preventing functioning thereof.
30
It appears that, when the origin of the single control fault is mechanical, dismantling is not
justified.
Disclosure of the invention
35
WO 2014/049260 PCT/FR2013/052238
3
The present invention aims to improve the management of faults in engines that do not have
a diagnostic means of the single-control fault type. It thus relates to the implementation of a
diagnosis of the control fault in a local regulation loop, in particular the loop regulating the
pitch angle of variable-pitch stator fins. The purpose is to assist in determining the
5 maintenance action in the case of single control faults.
Thus the invention relates to a method for differentiating control faults in a system controlling
an actuator, the actuator having a movable member acting on the position of a member
modifying the state of a gas turbine engine, the system comprising an electrical control and
10 mechanical control, the electrical control comprising at least one computer for computing the
desired position of the movable member of the actuator and transmitting an instruction signal
to the single-channel mechanical control of the actuator, a fault being diagnosed when a
difference is detected between the desired position of the actuator and its measured position
during a given difference-confirmation period.
15
The method is characterised by the fact that the speed of movement of the actuator is
measured and that a fault of electrical origin is diagnosed when, during said differenceconfirmation
period, the speed of movement measured remains zero or above a
predetermined threshold speed, otherwise the fault is of temporary mechanical origin.
20
Through the method of the invention, a means is available for differentiating between an
electrical fault which, for safety reasons and in accordance with the maintenance procedures,
gives rise to the dismantling on the ground of an item of equipment and a temporary
mechanicai fault which objectively does not require dismantling. It is thus possible to decide
25 not to proceed with the dismantling of the equipment concerned if the type of single
mechanical fault is not repeated. If the fault of mechanical origin is repeated a certain
number of times, three for example, it may be opportune to proceed with dismantling. In
accordance with another, if the actuator is in the retracted position at the moment when the
difference is detected and remains in the position during the determined difference-
30 confirmation period, the system is declared to have a fault of an electrical origin.
This requirement is imposed by the fact that, since the electrical fault leads to an abrupt
movement towards the retracted position, it is possible for the actuator to be in its retracted
position before a difference is detected. In this case the speed remains zero during the
35 monitoring phase, which would indicate erroneously a temporary mechanical fault.
WO 2014/049260 PCT/FR2013/052238
4
In accordance with another feature, the value of the control signal being compared with a
given threshold, if the signal is higher than said threshold, and the speed of movement is
higher than the threshold speed, then the control fault is declared electrical.
5
The present invention concerns a control system using a single-channel or dual-channel
electrical control. It applies in particular to the local loop regulating the actuator controlling
the pitch angle of the variable-pitch fins of the engine compressor stator. It also applies to the
local loop controlling the inlet directing wheel (IDW).
10
Brief presentation of the figures
Fig. 1 illustrates the architecture of the regulation loops of a gas turbine engine;
15 Fig. 2 is a block diagram of the various steps performed during the method for differentiating
single control faults of the invention.
Detailed description of an embodiment of the invention
20 The diagram in Fig. 1 relates to the architecture for regulation of a gas turbine engine
propelling an aircraft. This engine is for example a double-slide engine with two concentric
rotors rotating at different speeds, a low-pressure rotor and a high-pressure rotor. Each rotor
comprises, on either side of the combustion chamber, a compressor compressing the air
towards the combustion chamber and a turbine supplied by the combustion gas from the
25 chamber and driving the compressor that is associated with it. The high-pressure
compressor receives the air from the low-pressure compressor and the low-pressure turbine
receives the gases partially expanded in the high-pressure turbine.
The control of this engine is based on a regulation architecture with two interlink levels, it is
30 composed of engine-regulation loops 2 that control the state of the engine 1, in particular its
rotor rotation speed and the output speed of the LP compressor.
The engine loops 2 are composed of three parts, a computing part 4, an engine part 1 and a
part 5 measuring the parameters of the engine.
WO 2014/049260 PCT/FR2013/052238
5
The computing part 4 fulfils the following functions:
definition of the desired state of the engine 1 according to the input parameters 3, which are
the aircraft flight conditions and the gas control lever position;
5
computing the actual state of the engine according to the measurement supplied by the
engine parameter measuring part 5;
computing the actuator position instruction 6. This instruction is produced so as to ensure the
10 convergence of the measured state of the engine towards the desired state of the engine.
The engine part 5 has its state modified by the position of the actuators.
The measuring part 5 assesses the operating parameters of the engine.
15
Within the engine loops, the system comprises local regulation loops 8 that provide the
position controi of the actuators. These actuators, jacks in particular, act on the members
modifying the state of the engine, namely the main fuel and post-combustion metering
devices (it is a flow meter rather than a jack for the post-combustion) and where applicable
20 the air inlet directing wheel, the variable-pitch stator fins, the exhaust nozzle and others. The
present invention relates in particular to controlling the pitch angle of the variable-pitch stator
fins. Varying the pitch of the stator fins makes it possible to optimise the functioning of the
HP compressor of the engine and to increase its pumping range.
25 The local loops 8 are composed of:
a computing part 10 that produces the current command 12 to be applied to the actuators.
The command 12 is produced so as to ensure the convergence of the actuator position
measured at 16 towards the desired position;
30
- a control part 14 that positions the actuator according to the current command;
- a measuring part 16 that assesses the position of the actuators.
35 The control chain 10, 14, 16 comprises two parts:
WO 2014/049260 PCT/FR2013/052238
6
• An electrical part 10. This part transmits the electrical energy to the actuators.
• A mechanical or hydromechanical part 14. This part transforms the electrical energy into
5 mechanical energy. For a hydromechanical actuator, this part comprises a servo-valve 141
that controls the supply to the actuator, such as a pressurised fluid jack. It should be noted
that the servo-valve itself comprises an electrical part with two duplicated control channels
and a mechanical part.
10 The electrical part 10 may be duplicated. In this case it comprises two computers 10 and 10'.
Each computer can send current 12 to the actuators, for example to the servo-valve 141 of a
hydromechanical actuator. However, the actuator 14 is controlled by only one computer at a
time: the active computer 10. The other computer 10' is passive and remains ready to take
over in the event of malfunctioning of the first. There is then switching of computer: the active
15 computer becomes passive and vice versa.
There are thus, in general, two electrical control channels. On the other hand, the
mechanical part is not duplicated. There is only one mechanical control channel.
20 Presentation of control faults
A control fault is due either to a failure at one of the two electrical control channels 10, 10' or
12 and also at the electrical part of the servo-valve 141 (the coils in particular), or to a failure
at the hydromechanical chain 141, 142. Faults of electrical origin are thus distinguished from
25 faults of mechanical origin.
The control channel associated with the active computer is said to be faulty when there is a
difference greater than a given threshold during a given confirmation period between the
control response and the theoretical response. This confirmation period is evaluated using a
30 theoretical model of the control chain.
Faults are also distinguished according to their degree of importance: single faults and
double faults, or even higher.
WO 2014/049260 PCT/FR2013/052238
7
The degree of importance of the fault (single or double) depends on the magnitude of the
failure.
For faults of electrical origin, the magnitude of the failure is characterised by the number of
5 electrical channels 10 or 10' concerned. In the case of a double-channel electrical control
system, the failure on a single electrical control channel causes a single control fault. A
failure on the two electrical-control channels 10 and 10' causes a double control fault. In the
case of a system with a single electrical control channel, there is no longer any concept of
single or double faults. A failure gives rise to the detection of a control fault.
0
For faults of mechanical origin, there is only one mechanical control channel. The magnitude
of the mechanical failure varies according to the duration of the failure: definitive or
temporary failure: the following table indicates the correspondence between the duration of
the mechanical failure and the control fault that it causes.
A period is said to be short if it is less than the confirmation period: confirmation of the fault
on the new active computer.
20 A period is said to be medium if it lies between the confirmation period and the sum of the
following periods: confirmation period + inhibition period. The latter is the period during which
the monitoring of the fault is inhibited following a switching of channels - this makes it
possible to leave time for the new active computer to correct the positioning.
25 A period is said to be high if it is greater than the sum of the following periods: confirmation
period + inhibition period.
Impact of single control faults on the availability of the engine in the case of a double-channel
electrical control.
WO 2014/049260 PCT/FR2013/052238
8
A single control fault gives rise to the switching of computers 10, 10'. The faulty control panel
remains passive. The single control fault is thus transparent in flight. However, in accordance
with the engine maintenance procedures according to the prior art and for safety reasons,
5 this fault gives rise to the routine dismantling of an item of equipment - computer 10, 10' or
control unit 14 of the actuator concerned - on the ground.
The legitimacy of this dismantling varies according to the origin of this single control fault.
10 The origins of a single control fault are:
• A failure of one of two electrical control channels
• A temporary failure of the mechanical control channel. This may for example be transient
15 contamination of the actuator.
In the case of a single control fault of mechanical origin, dismantling is not justified.
The following table presents the impacts, in terms of availability of the engine, of the single
20 control fault.
25
Necessity of differentiation of control faults
The above analysis shows that not dissociating the origins of single control faults leads to
dismantling of equipment that is not justified, it is however the recommended maintenance
procedure in the prior art as a safety measure.
WO 2014/049260
9
PCT/FR2013/052238
It has been noted that, during a period of observation of several months, a not insignificant
proportion of dismantling of the stator unit (comprising the variable-pitch fins) following
single control fault was unjustified. Dismantling is said to be unjustified if no failure has been
noted during investigations on the dismantled equipment.
5
It should also be noted that it would be possible to differentiate control faults by implementing
monitoring of the control current. The current sent to the servo-valve 141 is re-read by the
computer 10. The re-reading current is designated in the field as "wrap" current. This current
is compared with the one produced by the computer. A difference between these two
10 currents indicates a control fault of electrical origin. However, such a solution would result in
a requirement in respect of the hardware and the operating system of the computer and, for
an existing engine, this would assume a modification to the computer. To avoid such a
modification, the solution of the invention, simple to implement in this case, has been
developed.
15
In order to avoid inopportune dismantling in the case of a temporary fault on a mechanical
control channel, it is necessary to differentiate single control faults according to their origins.
Analysis of the effects of the faults to be distinguished
20
For a local group comprising a hydromechanical actuator, the control chain of a local
regulation loop consists of an active computer 10 or 10' that produces the control current 12
to be sent to the servo-valve 141. This current is calculated so as to make the difference
between the measured position of the actuator 142 tend towards the desired position
25 produced by the engine loops. The servo-valve 141 supplies the flow to the chambers of the
actuator. The supply pressures depend on the control current 12 received by the servo-valve
141. The speed of movement of the movable member of the actuator is determined by the
supply flow rate.
30 The movement of the jack is proportional to the variation in control current.
An abrupt movement, that is to say at the maximum speed associated with the given flight
conditions, of the movable member of the actuator (jack) towards the retracted position
corresponds to zero current. An abrupt movement, that is to say at the maximum speed
35 associated with the given flight conditions, of the movable member of the actuator (the jack)
WO 2014/049260
10
PCT/FR2013/052238
in the opposite position to the retracted position, corresponds to the maximum current. A
double-stage servo-valve is described in the patent US 3023782.
Effects of faults in an electric control channel
5
Faults in an electric control channel are generally due either to an open circuit or to a shortcircuit.
A short-circuit is detected by the computer circuit breaker, which then demands
zeroing of the current sent. These two types of failure cause the zeroing of the current read
by the servo-valve 141, which then demands the abrupt movement of the jack towards its
10 retracted position.
An electrical control fault may also be due (5% of cases) to a drift in the current sent by the
computer. The drift in the current sent by the computer causes a shift in the response of the
servo-valve. This shift is compensated for by the feedback loop but all the same causes an
15 increase in the response time of the loop.
Effects of temporary faults on the mechanical control channel
Temporary failures of the mechanical control channel result from a jamming of the slide of
20 the servo-valve in the case of transient contamination.
Analysis shows that there exists two types of slide jamming: wedging of particles between
the slide and the casing or shearing of particles between slide and supply outlet of the jack
chambers. The effect of the first type is to increase friction on the slides. The current/output
25 curve exhibits hysteresis. In other words, the response time of the control chain is increased.
A particle that becomes wedged between the slide and the supply output of a jack chamber
will be sheared. The slide is blocked during the shearing time (which may take a long enough
time to raise a control fault). The supply pressure of the jack chambers is kept constant
30 during blocking. This can happen only around the equilibrium position of the slide (closure of
the supply outlet of the jack chambers), and therefore at low speeds.
Temporary faults in the mechanical control channel are thus due to a momentary reduction in
the performances of the actuators. The response time of the loop is increased.
WO 2014/049260
11
PCT/FR2013/052238
The above analysis shows that the majority of faults in an electrical control channel can be
differentiated from temporary faults in the mechanical control channel through their
abruptness.
5 A small percentage (5%) of control faults of electrical origin have effects similar to the single
control faults of mechanical origin.
Method of differentiating single control faults of different origins
10 The method of the invention is based on the finding that an electrical fault in 95% of cases
causes movement of the jack at the maximum speed for given flight conditions towards the
retracted position, whereas as temporary hydromechanical fault restricts the jack movement
speed range.
15 Thus monitoring of the speed of the jack makes it possible to distinguish an electrical fault
from a temporary hydromechanical fault.
The speed of the jack is monitored at the time of the failure, that is to say from the detection
of a difference between the response of the control and the theoretical response, until the
20 control fault is confirmed, just before the switching of the computer. This corresponds to the
control fault confirmation phase.
A threshold speed is determined by calibration during engine tests. It is determined so as to
be less than the speed associated with an electrical control fault of the open circuit or short-
25 circuit type. It is assessed by zeroing the current sent to the servo-valve of the actuator.
If the jack moves in the direction of the retracted position and the speed of the jack is higher
than the threshold speed during the control fault confirmation phase, then the fault is said to
be of electrical origin. Otherwise the fault is said to be of mechanical origin.
30
Particular cases to be taken into account
Condition with respect to the position
WO 2014/049260 PCT/FR2013/052238
12
An electrical fault causes an abrupt movement towards the retracted position. It is possible
for the jack to reach the retracted position before a difference is detected between the
response of the feedback and the theoretical response. In this case, the speed remains zero
during the monitoring phase. It is therefore necessary to add the following condition: if the
5 jack is in the retracted position during the control fault confirmation phase, then the fault is
said to be of electrical origin.
Condition with respect to the current
In the case where the control current calculated is zero, a mechanical control fault may
cause a movement in the direction of the retracted position and that the speed of the jack is
higher than the threshold speed during the control fault confirmation phase. In order not to
misunderstand the origin of these mechanical control faults, it is necessary to add a condition
with respect to the current.
In the case of an electrical control fault, there is inconsistency between the calculated control
current and the control current actually received by the servo-valve. The control current
calculated is determined so as to slow down the jack or even reverse its direction of travel.
The calculated control current is not close to zero in the case of an electrical control fault.
The control fault is said to be electrical if the movement of the jack is abrupt while the control
current is not low.
Because of a differentiation of control faults, the dismantling logic following a single control
25 fault can be modified. The differences are indicated in the following table.
N.B.: A repetition of single control faults during a flight indicates an abnormal situation.
Dismantling is considered to be necessary in this case. The repetition threshold is fixed at 3.
10
15
20
WO 2014/049260
13
PCT/FR2013/052238
it should also be noted that this differentiation between electrical control faults and
mechanical control faults makes it possible to do away with tests on the ground of the
dismantled equipment, whether the system be single or double electrical control channel.
5
Fig. 2 is a logic diagram of the various operations. The references thus designate the
following operations:
100: the position X of the movable member of the actuator is measured;
10
101: the speed V of the movable member is worked out from the position;
102: the speed V is compared with the threshold speed;
15 103: the control current is compared with the low control current threshold;
104: the position of the movable member of the actuator is compared with a given stop
position;
20 106: a differentiation indication is worked out from the results of 102, 103 and 104 monitored
for a period corresponding to the phase confirming a control fault in the actuator;
107: verification of the absence of fault in the sensor sensing the position of the movable
member of the actuator;
25
108: verification of the control for detection;
109: verification that the two conditions are fulfilled;
30 110: the differentiation indication is taken into account only in the case of control fault
detection.

CLAIMS
1. Method for differentiating control faults in a system controlling an actuator (14) acting on
the position of a member modifying the state of a gas turbine engine, the system comprising
5 an electrical control and a mechanical control, the electrical control comprising at least one
computer (10, 10') for calculating the desired position of the movable member of the actuator
(14) and transmitting an instruction signal to the single-channel mechanical control of the
actuator, a fault being diagnosed when a difference is detected between the desired position
of the movable member of the actuator and its measured position for a given difference-
10 confirmation period, characterised by the fact that the speed of movement of the actuator is
measured and a fault of electrical origin is diagnosed when, during said differenceconfirmation
period, the speed of movement measured remains zero or above a
predetermined threshold speed, otherwise the fault is of temporary mechanical origin.
15 2. Method according to claim 1, according to which, if the actuator is in the retracted position
at the moment when the difference is detected and remains in this position for the given
difference-conformation period, the system is diagnosed as having a fault of electrical origin.
3. Method according to either claim 1 or claim 2, the value of the control signal being
20 compared with a given threshold, if the signal is above said threshold and the speed of
movement is higher than the threshold speed, then the control fault is diagnosed as being
electrical.
4. Method according to any of claims 1 to 3, the actuator controlling the position of the stator
25 fins of the engine compressor.
5. System for controlling an actuator (14) acting on the position of a member modifying the
state of a gas turbine engine, the system comprising an electrical control and a mechanical
control, the electrical control comprising at least one computer (10, 10'), only one being
30 active at a time, arranged to compute the desired position of the movable member of the
actuator (14) and to transmit an instruction signal to the single-channel mechanical control of
the actuator, a fault being diagnosed when a difference is detected between the desired
position of the movable member of the actuator and its measured position for a given
difference-confirmation period, characterised by the fact that it comprises an indicator
35 differentiating control faults of the actuator.
6. System according to the preceding claim, comprising a double-channel electrical control
with one computer (10, 10!) per channel, only one being active at a time.
5 7. System according to-either-claim-5 or claim 6, comprising a means for measuring the
speed of movement of the movable member of the actuator, the means for comparing the
speed with a speed and a fault differentiating indicating means such that the system
diagnoses a fault of electrica! origin when, during said difference-confirmation period, the
speed of movement measured remains zero or higher than' a predetermined threshoid
10 speed, otherwise the fault is said to be of temporary mechanical origin.