Method for monitoring a thrust fault of an aircraft turbofan

Abstract

A method for monitoring a thrust fault of a turbofan during a modification of the thrust setting of the turbofan, the method including a step of processing the thrust setting via a filtering function and a transient-phase model such as to obtain a modelled thrust, a step of comparing the modelled thrust to the actual thrust such as to determine a thrust difference, a step of comparing the thrust difference to an alarm threshold; and a step of emitting an alarm in the event of exceeding the alarm threshold, wherein at a given iteration, in which the prior modelled thrust is known, the transient-phase model provides a time constant in accordance with the prior modelled thrust, and the filtering function provides a modelled thrust in accordance with the time constant obtained, the prior modelled thrust and the thrust setting.

Claims

1. A method for monitoring a thrust fault in an aircraft turbojet engine via an on-board computer of an aircraft when a thrust setting of said turbojet engine is modified, an effective thrust of the turbojet engine being modified during a transient phase to attain a desired thrust setting, said method comprising: processing the thrust setting via a filtering function and a transient-phase model, obtained using experience feedback from measurements taken on a same type of turbojet engine in order to obtain a modelled thrust; measuring the effective thrust; comparing said modelled thrust with said effective thrust in order to determine a thrust difference; comparing said thrust difference with an alarm threshold; and emitting an alarm in an event of said alarm threshold being exceeded; wherein, at a given iteration, in which a prior modelled thrust is known, the transient-phase model supplies a time constant from the prior modelled thrust, and the filtering function supplies a modelled thrust from the time constant obtained, the prior modelled thrust and the thrust setting.

2. The method according to claim 1, wherein the filtering function is initialized depending on a sign of the thrust difference when the turbojet engine is started, the prior modelled thrust being equal to the effective thrust of the turbojet engine during the initialization.

3. The method according to claim 1, wherein the filtering function is a low-pass function.

4. The method according to claim 1, wherein, as the turbojet engine comprises an idle-speed regulating device suitable for pre-empting a setting, defined via a throttle control lever, by an idle speed value that depends on determined ambient conditions of the turbojet engine, the method further comprises, for monitoring excess thrust, measuring at least one ambient parameter of the turbojet engine in order to define whether the throttle control lever setting is pre-empted by the idle speed value.

5. The method according to claim 4, further comprising determining a modelled idle speed value modelled via an idling model which associates a modelled idle speed value with one or more values of ambient parameters of the turbojet engine.

6. The method according to claim 5, wherein the idling model associates the modelled idle speed value with at least one of an ambient pressure of the turbojet engine and an ambient temperature of the turbojet engine.

7. The method according to claim 1, wherein the alarm threshold is a threshold that can be parameterised, and the method further comprising monitoring the state of at least one auxiliary condition of the turbojet engine, and applying a penalty parameter to said alarm threshold when an abnormal state of said at least one auxiliary condition is detected.

8. The method according to claim 7, wherein the penalty parameter is applied to said alarm threshold which depends on a criticality of the at least one auxiliary condition being monitored.

Description

DESCRIPTION OF THE DRAWINGS

(1) The invention will be better understood upon reading the following description, given solely by way of example, and with reference to the accompanying drawings in which:

(2) FIG. 1 is a schematic diagram of the method for monitoring the inadequate thrust of a turbojet engine according to the invention;

(3) FIG. 2 is a schematic diagram of the step of processing the thrust setting when implementing the method from FIG. 1;

(4) FIG. 3 is a first diagram of monitoring excess thrust and a second diagram of monitoring inadequate thrust;

(5) FIG. 4 is a schematic diagram of a step of managing idling when monitoring excess thrust;

(6) FIG. 5 is a general schematic diagram of the parameterisation of the alarm threshold;

(7) FIG. 6 is a schematic diagram of the parameterisation of the alarm threshold by monitoring the speed of change of the thrust difference;

(8) FIG. 7 is a schematic diagram of the parameterisation of the alarm threshold by monitoring the speed of rotation of a high-pressure spool of the turbojet engine; and

(9) FIG. 8 is a schematic diagram of the parameterisation of the alarm threshold by detecting a pumping event and/or a cut-off event of the turbojet engine.

(10) It should be noted that the drawings disclose the invention in a detailed manner for implementing the invention, but said drawings could, of course, be used to better define the invention if need be.

DESCRIPTION OF ONE OR MORE EMBODIMENTS

(11) FIG. 1 is a schematic illustration of a method according to the invention for monitoring a thrust fault of an aircraft turbojet engine by means of an on-board computer of said aircraft when modifying the thrust setting of said turbojet engine.

(12) The invention will be described for a dual-spool turbojet engine comprising a low-pressure spool and a high-pressure spool. For the sake of clarity, the speed of rotation of the low-pressure spool, also referred to as speed N1, will be used to determine the thrust of the turbojet engine. It is self-evident that other parameters of the turbojet engine which depend on the thrust could also be used, in particular, the parameter EPR (engine pressure ratio).

(13) The thrust setting N1.sub.CONS corresponds in this example to the set speed of the low-pressure spool. Preferably, the thrust setting N1.sub.CONS is defined by the pilot of the aircraft, on which the turbojet engine is mounted, using a throttle control lever.

(14) Advantageously, the thrust setting N1.sub.CONS is known precisely and reliably. The thrust setting N1.sub.CONS is redundant, thus ensuring that the monitoring method uses sound, in other words uncorrupted, input. The reliability of the monitoring method according to the invention is therefore improved.

(15) As indicated in the preamble, the effective thrust of the turbojet engine N1.sub.EFF is modified during a transient thrust phase until it attains the thrust setting N1.sub.CONS. The effective thrust N1.sub.EFF corresponds, in this example, to the effective speed of the low-pressure spool. Preferably, the effective thrust N1.sub.EFF is measured redundantly by sensors of the turbojet engine and is considered to be a sound input.

(16) According to the invention, with reference to FIG. 1, the method comprises: a step of processing 1 the thrust setting N1.sub.CONS during the transient thrust phase so as to determine a modelled thrust N1.sub.MOD; a step of measuring the effective thrust N1.sub.EFF, a step of comparing said modelled thrust N1.sub.MOD with said effective thrust N1.sub.EFF so as to determine a thrust difference ; a step of comparing said thrust difference with an alarm threshold S; and a step of emitting an alarm ALARM should said threshold S be exceeded.

(17) The method is noteworthy in that, during the processing step 1, with reference to FIG. 2, the thrust setting N1.sub.CONS is processed using a filtering function F and a transient-phase model M so as to obtain a modelled thrust N1.sub.MOD.

(18) As shown in FIG. 2, the processing step is iterative. At a given iteration, since a prior modelled thrust N1.sub.MOD(OLD) is known, the transient-phase model M supplies a time constant T from the prior modelled thrust N1.sub.MOD(OLD), the filtering function F supplies a modelled thrust N1.sub.MOD from the time constant obtained T, the prior modelled thrust N1.sub.MOD(OLD) and the thrust setting N1.sub.CONS.

(19) In other words, the invention proposes transforming the thrust setting N1.sub.CONS into a modelled thrust N1.sub.MOD which takes account of the transient thrust phase of said turbojet engine. Therefore, the modelled thrust N1.sub.MOD and the actual thrust N1.sub.EFF are correlated during the transient thrust phase, thus allowing them to be compared in a relevant way at all times in the transient phase. This allows any thrust fault (inadequate thrust or excess thrust) to be detected reactively. Therefore it is no longer necessary to wait for the thrust to stabilise in order to detect a thrust fault as in the prior art.

(20) As illustrated in FIG. 2, the thrust setting N1.sub.CONS is filtered by a low-pass filtering function F of approximately 2 during the processing step 1. The filtering function F is a transfer function, which is known per se to a person skilled in the art, which allows a modelled thrust N1.sub.MOD to be supplied depending on the time constant T in order to allow a comparison with the effective thrust N1.sub.EFF. Preferably, the filtering function F has a unity gain such that at the end of the transient phase, the modelled thrust N1.sub.MOD equals the thrust setting N1.sub.CONS.

(21) Still with reference to FIG. 2, the transient-phase model M associates a modelled thrust value N1.sub.MOD with a time constant value T. The transient-phase model M is conventionally obtained using experience feedback from measurements taken on the same type of turbojet engine. As illustrated in FIG. 2, the modelled thrust N1.sub.MOD of a particular iteration is defined from the modelled thrust N1.sub.MOD(OLD) obtained at the preceding iteration and the time constant T.

(22) Preferably, the transient-phase model M is different for an increase or reduction of the thrust. Provision is also made for a transient-phase model M for each type of modification of the thrust.

(23) As illustrated in FIG. 3, the excess thrust and the inadequate thrust of the turbojet engine are monitored by two distinct methods which comprise different processing steps 1.sub.EXCESS, 1.sub.INAD given that they make use of different transient-phase models M.sub.EXCESS, M.sub.INAD and different filtering functions F.sub.EXCESS, F.sub.INAD. Similarly, the alarm thresholds are different S.sub.EXCESS, S.sub.INAD depending on the thrust fault being monitored as illustrated in FIG. 3.

(24) For the sake of clarity, the invention is described below for the case of inadequate thrust with reference to the general schematic illustration in FIG. 1.

(25) Preferably, as illustrated in FIG. 2, an initialisation INIT of the filtering function F is carried out, firstly, when starting the turbojet engine and, secondly, depending on the sign of the thrust difference .

(26) Given that the filtering function F is approximately 2, an initialisation allows said filtering function to begin for predetermined modelled thrust values N1.sub.MOD(OLD). An initialisation of the filtering function F depending on the sign of the thrust difference makes it possible to prevent a divergence of monitoring methods depending on the type of fault monitored. By way of example, if the effective thrust N1.sub.EFF is greater than the modelled thrust N1.sub.MOD (the hypothesis of excess thrust), the filtering function F.sub.INAD of the method for monitoring the inadequate thrust (see FIG. 3) can be initialised in order to prevent inconsistent results from being supplied, in other words, false alarms. In particular, such an initialisation is advantageous when accelerations are linked to decelerations.

(27) Preferably, during the initialisation, the prior modelled thrust N1.sub.MOD(OLD) is equal to the effective thrust N1.sub.EFF of the turbojet engine.

(28) Taking Account of the Idle-Speed Regulating Device

(29) According to a preferred aspect of the invention, the turbojet engine comprises an idle-speed regulating device suitable for pre-empting a thrust setting, defined by means of a throttle control lever, by an idle speed value depending on determined ambient conditions of the turbojet engine.

(30) For example, with reference to FIG. 4, when the pilot of the aircraft wishes to limit the thrust of the turbojet engine as much as possible, the idle-speed regulating device pre-empts the thrust setting N1.sub.MAN defined by the throttle control lever by an idle speed value N1.sub.RAL so that the reduction in thrust is not too great for the turbojet engine compared with limitations other than the speed N1 (speed N2, minimal fuel flow rate to limit the risk of cut-off, etc.). The idle speed value N1.sub.RAL is not calculated, but results from various limitations of the idle-speed regulating device so as to provide an optimal thermodynamic cycle for the turbojet engine when it decelerates.

(31) However, with regard to the method for monitoring an excess thrust fault according to the invention, the fact that the idle speed value N1.sub.RAL corresponds to the thrust setting N1.sub.CONS during idling has to be taken into account.

(32) For this purpose, with reference to FIG. 4, in order to monitor excess thrust, the method comprises a step of comparing the lever thrust setting N1.sub.MAN with the idle speed value N1.sub.RAL, the maximum thrust value being used as a thrust setting N1.sub.CONS in the above-mentioned processing step 1. In other words, if the thrust setting N1.sub.MAN defined by the throttle control lever is greater than the idle speed value N1.sub.RAL induced by the idle-speed regulating device, the monitoring method remains in conformity with the embodiment shown in FIG. 1.

(33) However, if the thrust setting N1.sub.MAN defined by the throttle control lever is less than the idle speed value N1.sub.RAL induced by the idle-speed regulating device, the monitoring method uses the idle speed value N1.sub.RAL as a thrust setting N1.sub.CONS in the above-mentioned processing step 1.

(34) For the sake of reliability of the monitoring method, it is important to know the idle speed value N1.sub.RAL independently of the idle-speed regulating device. Accordingly, the idle speed value is modelled as described below.

(35) In the example in FIG. 4, the idle speed value N1.sub.RAL-MOD is obtained from an idling model M.sub.RAL that associates a modelled idle speed value N1.sub.RAL-MOD with one or more values of ambient parameters of the turbojet engine. However, it is self-evident that the modelled idle speed value N1.sub.RAL-MOD could be obtained by other means. By measuring ambient parameters of the turbojet engine which are characteristic of the actuation of the idle-speed regulating device, the idle speed value N1.sub.RAL-MOD can be determined indirectly.

(36) In this example, with reference to FIG. 4, the idling model M.sub.RAL associates a modelled idle speed value N1.sub.RAL-MOD with the ambient pressure P.sub.amb of the turbojet engine and the ambient temperature T.sub.amb of the turbojet engine. In practice, the ambient parameters P.sub.amb and T.sub.amb of the turbojet engine are redundant parameters which are considered to be reliable. In particular, these parameters do not have common modes with the thrust faults to be monitored. It is self-evident that other ambient parameters could also be suitable. Similarly to the transient-phase model M, the idling model M.sub.RAL is obtained using experience feedback from measurements taken on the same type of turbojet engine.

(37) With reference to FIG. 4, the monitoring method comprises a step of measuring the ambient parameters P.sub.amb and T.sub.amb of the turbojet engine in order to define the modelled idle speed value N1.sub.RAL-MOD with the help of the idling model M.sub.RAL. Once the modelled idle speed value N1.sub.RAL-MOD has been obtained, said value is compared with the thrust setting N1.sub.MAN of the control lever, the maximum thrust being used as a thrust setting N1.sub.CONS for the processing step 1.

(38) Because the idling is taken into account, the precision with which the excess thrust is monitored is improved. In particular, it makes it possible to avoid excess thrust faults from being declared when the turbojet engine is regulated normally on a high idle speed value.

(39) Variable Alarm Threshold

(40) In order to improve the robustness of the monitoring method compared with the dispersions between different turbojet engines of the same type, the alarm threshold S is, preferably, a variable threshold that takes account of auxiliary conditions in order, for example, to reduce the value of the alarm threshold S and thus increase the sensitivity of the detection. In other words, with reference to FIG. 5, the method uses monitoring of at least one auxiliary condition of the turbojet engine COND1, COND2. If a malfunction relating to said auxiliary condition COND1, COND2 is detected, the alarm threshold S is penalised by a penalty parameter PEN1, PEN2 in order to favour rapid detection of a thrust fault.

(41) Preferably, the method implements monitoring of a plurality of auxiliary conditions COND1, COND2 of the turbojet engine. The auxiliary conditions COND1, COND2 are ranked according to their criticality. The more critical the auxiliary condition COND1, COND2, the greater the penalty parameter PEN1, PEN2 of the alarm threshold S. Thus, the threshold S is adaptive to allow a rapid reaction while limiting the risk of false alarms, which is advantageous.

(42) Depending on the type of thrust fault monitored (inadequate thrust or excess thrust), the penalty parameter PEN1, PEN2 can be subtracted/added to the threshold S. It is self-evident that the penalty parameter PEN1, PEN2 could also be in the form of a multiplier coefficient, a mathematical function, etc. Preferably, the penalty parameter PEN1, PEN2 can also depend on the auxiliary condition monitored COND1, COND2 in order to modulate the alarm threshold S progressively.

(43) A plurality of auxiliary conditions will be described below in association with their penalty parameters for monitoring inadequate thrust of a turbojet engine.

(44) a) Measurement of the Speed of Change of the Thrust Difference

(45) By way of example, with reference to FIG. 6, the method comprises a step of measuring the speed of change of the thrust difference, that is, the parameter

(46) $\frac{d}{\mathrm{dt}},$
and a step or comparing the speed of change of the thrust difference

(47) $\frac{d}{\mathrm{dt}}$
with a predetermined threshold S.sub.d.

(48) If the threshold S.sub.d, is not exceeded, the alarm threshold is not modified. However, if the threshold S.sub.d is exceeded, the alarm threshold S is reduced by a penalty parameter PEN.sub.d, thus having the consequence of making the detection of a thrust fault more sensitive, as illustrated in FIG. 6.

(49) In other words, if the change in the thrust difference is very rapid, it can be concluded that the effective thrust N1.sub.EFF of the turbojet engine corresponds less and less to the modelled thrust N1.sub.MOD, and this is characteristic of a thrust fault.

(50) Preferably, the penalty parameter PEN.sub.d depends on the speed of change of the thrust difference , thus allowing the alarm threshold S to be modulated progressively. By way of example, the penalty parameter PEN.sub.d can be in the form of a coefficient which is equal to 0.5 for a rapid change in the thrust difference and in the form of a coefficient which is equal to 0.2 for a very rapid change in the thrust difference .

(51) b) Measurement of the Speed of the High-Pressure Spool

(52) By way of example, with reference to FIG. 7, the method comprises a step of measuring the speed of the high-pressure spool, that is, the speed N2 of the turbojet engine.

(53) Similarly to the low-pressure spool, the effective speed N2.sub.EFF of the high-pressure spool is compared with a modelled speed N2.sub.MOD which is determined directly depending on the position of the throttle control lever or indirectly by means of a transient-phase model suitable for the change in speed N2 of the high-pressure spool which is similar to the transient-phase model M described previously.

(54) In other words, a step of measuring the difference .sub.N2 between the effective speed N2.sub.EFF and the modelled speed N2.sub.MOD is carried out and a step of comparing 7 this difference of high pressure speed .sub.N2 with a high pressure threshold S.sub.N2 of predetermined value is carried out as shown in FIG. 7.

(55) If the threshold S.sub.N2 is not exceeded, the alarm threshold S is not modified. However, if the threshold S.sub.N2 is exceeded, the alarm threshold S is reduced by a penalty parameter PEN.sub.N2, thus having the consequence of making the detection of a thrust fault more sensitive.

(56) It can be concluded that the effective speed N2.sub.EFF of the turbojet engine corresponds less and less to its modelled speed N2.sub.MOD, and this is characteristic of a thrust fault.

(57) This type of auxiliary condition makes it possible to increase the reliability of the monitoring. The relationship of the speed N1 over the speed N2 is not linear over time. In particular, the response time of the high-pressure spool is lower than that of the low-pressure spool. Thus, the high-pressure spool accelerates faster at low speed than the low-pressure spool. For this reason, monitoring, firstly, the speed of the low-pressure spool N1 and, secondly, the speed of the high-pressure spool N2, allows optimum monitoring of a thrust fault.

(58) Some turbojet engines are regulated during the transient phases in accordance with a speed setting N2.sub.CONS or in accordance with the acceleration in

(59) $\frac{\mathrm{dN}2}{\mathrm{dt}}.$
In addition, according to this first hypothesis, a step of measuring the difference between the effective speed N2.sub.EFF and the speed setting N2.sub.CONS is carried out. In other words, the speed setting N2.sub.CONS replaces the modelled speed N2.sub.MOD. If the turbojet engine is regulated in accordance with the acceleration, the acceleration

(60) $\frac{\mathrm{dN}2}{\mathrm{dt}}$
replaces the modelled speed N2.sub.MOD.
c) Detection of Auxiliary Events

(61) As well as measuring the speed of the low-pressure spool N1 and of the high-pressure spool N2, the turbojet engine may comprise means of detecting auxiliary conditions that could affect the thrust of the turbojet engine.

(62) By way of example, with reference to FIG. 8, the turbojet engine may comprise means of detecting pumping of the turbojet engine in order to apply a pumping penalty parameter PEN.sub.PUMP to the alarm threshold S if such pumping is detected.

(63) Similarly, still with reference to FIG. 8, the turbojet engine may comprise means of detecting a cut-off of the turbojet engine in order to apply a cut-off penalty parameter PEN.sub.CUT to the alarm threshold S if such a cut-off is detected.

(64) In the example in FIG. 8, the penalty parameters PEN.sub.PUMP and PEN.sub.CUT are used in combination but it is self-evident that they could be applied independently to the alarm threshold S. Thus, if an auxiliary event that could affect the thrust is detected (pumping or cut-off), the value of the alarm threshold S is reduced in order to improve the sensitivity of the detection.

(65) Furthermore, in a preferred embodiment of the invention, said penalty parameters PEN.sub.PUMP and PEN.sub.CUT may cancel the alarm threshold S in order to trigger an alarm directly.

(66) d) Measuring the Change in Speed of the Low-Pressure Spool or the High-Pressure Spool

(67) Preferably, the method may comprise a step of measuring the speed of change of the difference .sub.N1, .sub.N2 during the transient phase, that is, the parameters

(68) $\frac{dN1}{\mathrm{dt}}\mathrm{or}\frac{dN2}{\mathrm{dt}},$
and a step of comparison to a predetermined threshold S.sub.N1 or S.sub.N2. Similarly, the penalty parameter may depend on the speed of change of the difference .sub.N1, .sub.N2, thus allowing the alarm threshold S to be modulated progressively.

(69) Advantageously, this type of monitoring allows any thrust fault to be detected very reactively. A thrust fault causes an abnormal change of the difference .sub.N1 or .sub.N2 which can be rapidly detected.

(70) e) Use of a Priority Alarm

(71) In a preferred embodiment, the penalty parameters have a negative value for a method for monitoring an inadequate thrust. Thus, even if the thrust difference is nil, an alarm can be emitted directly. In other words, a penalty parameter allows priority alarms to be triggered which are independent of the thrust difference . Advantageously, this allows a highly reactive monitoring method to be made available.

(72) The auxiliary conditions monitored by the turbojet engine have been presented independently in FIGS. 6 to 8 but it is self-evident that they can be monitored in combination.

(73) If a thrust fault is detected on a turbojet engine, the fault can be corrected with the aid of the other turbojet engine which is also in its transient phase by modifying its own thrust setting. Thus, the thrust is monitored and corrected in real time, thus increasing safety and reliability.