EP3970035A1

EP3970035A1 - Method for configuring a digital filter for attenuating a frequency associated with a torsion mode of a power transmission line of a turbine engine

Info

Publication number: EP3970035A1
Application number: EP20715900.5A
Authority: EP
Inventors: Typhaine RASLAIN
Original assignee: Safran Aircraft Engines SAS
Current assignee: Safran Aircraft Engines SAS
Priority date: 2019-05-13
Filing date: 2020-04-06
Publication date: 2022-03-23
Also published as: US20220220979A1; FR3096159B1; FR3096159A1; WO2020229058A1; CN114127716A

Abstract

The invention concerns a method for configuring a digital filter for attenuating a torsion mode of a power transmission line of an aircraft turbine engine (1), the mode being associated with a frequency F_T falling within a confidence interval le, the digital filter being a low-pass filter and: - described by a causal transfer function that is stable and equal to the quotient N(z)/D(z), - intended to be integrated into a pre-existing control loop of the turbine engine (1), so as to filter signals sampled at a frequency F_E, the loop being closed and having a gain increased, in absolute value, by a value V in its bandwidth. Moreover, the method includes: - a step (100) of calculating zeros of N(z), such that the filter attenuates the frequency F_T, - a step (200) of updating the zeros of N(z), such that the gain of the filter satisfies, in the interval le, a first gain margin, - a step (300) of determining poles of D(z), such that, in the bandwidth of the loop: - the phase of the filter satisfies a phase margin, - the gain of the filter satisfies a second gain margin.

Description

Title of the invention: Method of parameterizing a digital filter for the attenuation of a frequency associated with a torsion mode of a power transmission line of a turboshaft engine

Prior art

The present invention belongs to the field of aircraft turbine engines, such as airplanes or helicopters. It relates more particularly to a method of parameterizing a digital filter model for the attenuation of a torsion mode of a power transmission line of an aircraft turbine engine. The invention finds a particularly advantageous application, although in no way limiting, in the case of a turbine engine comprising non-faired propulsion means.

[0002] The use of turbine engines in the design of aircraft, both civil and military, such as for example airplanes and helicopters, is very widespread today. Turbine engines make it possible to develop the power necessary for aircraft flights, the mass of which most often reaches several tens of tonnes.

Turboshaft engines are available in different versions (gas turbines, turbojets, turboprop engines, etc.) all governed by the same operating principle, namely the conversion of kinetic and thermal energy, resulting from gas production (typically by combustion of a hydrocarbon), into mechanical energy intended to set at least one shaft coupled to propulsion means, such as a rotor provided with propellers, for example.

[0004] Conventionally, a turbine engine comprises a gas generator and a turbine, of the linked or free type, positioned downstream of the gas generator, with reference to the direction of flow of the gases in the turbine engine. This turbine is driven in rotation by the flow of gas generated, so as to set a transmission line (also called a “power transmission line”) in rotation. The transmission line comprises, in a known manner, at least one shaft directly coupled to the turbine, also called a “turbine shaft”, as well as at least one output shaft coupled to the propulsion means. Optionally, a Planetary type speed reducer connects the turbine shaft to the output shaft, in order to reduce the speed of rotation of the propulsion means.

[0005] The operation of the turbine engine is conventionally controlled by a set of logics which form what is called engine control. Among these logics, some rely on feedbacks to constitute closed-loop control logics. Such a loop aims to control a parameter of

operation of the turbine engine, such as a rotational speed of a rotor of the propulsion means, for example, to meet a pre-established control strategy. To this end, the loop takes measurements of said parameter, and compares these measurements with a setpoint. Any deviation between the measurements and the setpoint is transmitted to a control device capable of generating a control signal which is transmitted to the turbine engine in order to compensate for said deviation, the control process then being iterated along said loop.

[0006] The control signal generated by the control device thus impacts the operation of the turbomachine, which notably includes the assembly formed by the turbine, the transmission line and the propulsion means. However, this assembly is characterized in particular, like any mechanical system, by a certain stiffness - here in rotation - which for sizing constraints may not be sufficient with regard to the highly inertial elements which are found at its ends. This problem is further aggravated by increasing the length of the transmission line or by increasing the number of its components, via the introduction of a reduction gear for example. The transmission line then presents a torsion mode, the frequency of which is typically positioned outside, but nevertheless relatively close, of the bandwidth of use of the transmission line. Consequently, there is a risk that the command generated by the control device will excite the torsion mode of the transmission line, or amplify the resonance following an excitation external to the closed loop. Such a configuration is problematic insofar as the torsional oscillations can have an amplitude capable of greatly degrading the fatigue strength of the transmission line, and therefore cause premature wear of the equipment and cause plasticization or rupture of the shaft. . At least conceptually, and in order to limit the excitation of the transmission line according to its torsion mode, it could be envisaged, according to a first alternative, to orient the mechanical design and the dimensioning of the turbine engine so that said mode of torsion is sufficiently far from the bandwidth of the control loop. In this way, the power of the control signals would be sufficiently attenuated at the frequencies of said torsion mode.

[0008] According to a second alternative, it could be envisaged not to modify the mechanical architecture of the turbine engine, but rather to take into account the knowledge of the torsion mode (frequency, amplitude) in the design of the control loop of the turbine engine. . In other words, once the design of the turbine engine has stopped, and the torsion mode has been identified, the control loop is designed so that it does not activate said torsion mode.

[0009] These two alternatives nevertheless go against the classic cycle of

design of a turbine engine. In fact, the sizing of a turbine engine aims to define global production constraints, that is to say which are imposed on the turbine engine when it is considered in its entirety (or even ideally when the environment in which it is intended for integration is taken into account). Such constraints concern, for example, mass, cost, size, modes of operation and use, etc. Nevertheless, a turbine engine is a complex architectural system, in the sense that it is manufactured by means of a large quantity of parts, so that it is at the same time difficult to decline all these high level constraints for each of said different parts. and to anticipate, before carrying out them, the behavior of the final machine.

[0010] Consequently, the complex dynamic behaviors linked to integration effects are often discovered late in the design cycle and only validated during engine tests, once all of said parts have been produced and assembled.

[0011] It is therefore understood that if the initial sizing does not make it possible to avoid a mode of torsion of the line, the mechanical design and / or the control logic associated with the control loop must be reviewed, which is particularly long and expensive, and therefore to be avoided.

Disclosure of the invention

The present invention aims to remedy all or part of the

drawbacks of the prior art, in particular those set out above, by proposing a solution which makes it possible to attenuate a mode of torsion of a line of

transmission of power from a turbine engine, so as to avoid any

material resizing of said turbine engine as well as any alteration of the pre-existing regulation logic of operation of said turbine engine.

To this end, and according to a first aspect, the invention relates to a method of setting a digital filter for the attenuation of a torsion mode of a power transmission line of a turboshaft engine. aircraft, said mode being associated with a frequency F_T included in a confidence interval le, the digital filter being of the low-pass type and:

- described by a causal transfer function in z, stable and equal to the quotient N (z) / D (z), where N and D are polynomial functions, N being of degree strictly greater than 1,

- intended to be integrated into a pre-existing control loop of the turbine engine, so as to filter control signals generated by a control device of said loop and sampled at a frequency F_E, said loop being closed and associated with a pass band in which the gain of the loop is increased, in absolute value, by a value V.

In addition, said method is implemented by a setting device and comprises:

- a calculation step, as a function of the frequencies F_T and F_E, of complex numbers forming zeros of N (z), so that the filter attenuates the frequency F_T,

- a step of updating the zeros of N (z), so that the gain of the filter satisfies, in the confidence interval le, a first gain template

predetermined according to the amplitude of the torsion mode,

- a step of determining real numbers forming poles of D (z), so that, in the bandwidth of the loop:

- the phase of the filter satisfies a predetermined phase mask as a function of the frequency F_E,

the gain of the filter satisfies a second predetermined gain template as a function of the value V.

The step of calculating the zeros of N (z) makes it possible to initiate the parameterization of the filter by precisely targeting the frequency F_T to be attenuated. At this stage, the frequency behavior of the filter phase is not taken into account.

[0015] Following the calculation step, the update step makes it possible to release

the attenuation hitherto targeted only on the frequency F_T, so as to take into account the confidence interval le in which said frequency F_T is included. In other words, this step makes it possible to take into account the uncertainty associated with the value of the frequency F_T, and therefore to make the final attenuation sought for the digital filter more robust with respect to this uncertainty.

[0016] It should be noted that at this stage, again, the

frequency behavior of the filter phase. On the other hand, the behavior of the gain of the filter is itself substantially stopped, and is likely to vary only marginally during the subsequent step of determining the poles of D (z). More precisely, the fact that the modulus of the gain decreases, in absolute value and in the interval le, during the updating step is countered by the fact that the latter increases again during the updating step. determination of the poles of D (z).

Finally, the step of determining the poles of D (z) aims to place the poles of D (z) so as to control the evolution of the phase of the filter on the passband of the loop, which allows in fine, to control the phase of the filter over the entire frequency spectrum envisaged (zones A, B and C). In addition, also constraining the gain of the filter in the passband of the loop helps to ensure that the useful information contained in the control signals can continue to flow to the actuators of the turbine engine.

[0018] Thus, the invention makes it possible to integrate at the output of the closed loop control device a digital filter configured to attenuate the torsion mode associated with the power line, without physically resizing the turbine engine, as well as without modifying the logic of pre-existing regulation (i.e. ie the operation of a control system operating according to said pre-existing closed loop). The digital filter obtained by the parameter setting process only supplements the pre-existing regulation logic.

[0019] It should be noted that the setting of the digital filter is

advantageously done by decoupling the placement of the zeros from the numerator N (z), to ensure sufficient attenuation in the interval le, from the placement of the poles of the denominator D (z), to essentially control the phase of the digital filter.

[0020] Such a decoupling is advantageous because it makes it possible to obtain a linear filter

discreet, but also a good compromise between the intended behavior of the filter and a low filter degree. The filter obtained in this way is also very easily implemented in an ECU software from the

libraries of elementary functions known to those skilled in the art and commonly used for the production of certified aeronautical software.

[0021] In particular embodiments, the parameterization method may further include one or more of the following characteristics, taken in isolation or in any technically possible combination.

In particular embodiments, the step of updating the zeros of N (z) comprises a sub-step of reducing the respective moduli of the zeros according to a predetermined step, the sub-step of reduction being executed iteratively until the first amplitude template is satisfied.

Thus reducing the respective moduli of the zeros determined during the calculation step makes it possible to move the latter away from the unit circle, and therefore thus to widen the attenuation zone of the filter so as to take into account the interval trust the. This also makes it possible to obtain a good compromise between the efficiency and the complexity of the configuration.

In particular embodiments, the poles of D (z) are

all considered equal, the pole determination step comprising: - a sub-step of selecting a pole strictly between -1 and 1,

- a sub-step of translating the selected pole along the real axis and at a predetermined pitch, so as to obtain a translated pole,

said translation sub-step being executed iteratively as long as the phase template and the second gain template are not satisfied, the pole selected during an iteration corresponding to the translated pole obtained during the previous iteration.

[0025] Thus determining the poles of D (z) advantageously makes it possible to control the evolution of the phase of the filter over the passband of the loop, which ultimately makes it possible to control the phase of the filter over the entire frequency spectrum . In addition, the fact of also constraining the gain of the filter and thus normalizing it in the passband of the loop ensures that the useful information contained in the control signals can continue to be obtained.

routed to the actuators of the turbine engine.

[0026] In particular embodiments, the first gain template corresponds to an increase, in the confidence interval le, of the value of the gain by the opposite of the amplitude of the torsion mode.

In particular modes of implementation, the phase template

corresponds to an increase in the phase shift introduced by the filter in the passband of the closed loop.

In particular embodiments, the degree of N (z) is equal to 2, so as to obtain, during the calculation step, zeros z_1 and z_2 according to the following formulation:

z_1 = exp ((2xixnxF_T) / F_E) and z_2 = exp ((- 2xixnxF_T) / F_E).

The fact that N (z) is of degree equal to 2 makes it possible to limit the complexity of the filter, as well as to precisely target the frequency F_T to be attenuated.

In particular modes of implementation, the degree of D (z) is equal to 3.

[0031] The fact that D (z) is of a degree equal to 3 advantageously makes it possible to limit the complexity of the filter, while making it possible to have a strictly clean filter.

In particular modes of implementation, the frequency F_T and

the confidence interval are determined beforehand during a test bench test campaign of the oscillatory behavior of the line of

power transmission.

In particular modes of implementation, said method comprises, following the step of determining real numbers forming poles of D (z), a step of validating the temporal behavior of the digital filter, said validation step consisting in verifying that the temporal response of the filter to a step signal is increasing monotonously.

[0034] According to a second aspect, the invention relates to a control system

intended to be on board an aircraft comprising a turbine engine, said turbine engine comprising a power transmission line having a torsion mode associated with a frequency F_T included in a confidence interval le, said system comprising means for receiving a relative instruction at a predetermined parameter, a control device configured to generate control signals sampled at a frequency F_E and means for measuring said parameter, said control system forming a closed control loop associated with a passband in which the gain is increased , in absolute value, by a value V. In addition, the control loop comprises a digital filter parameterized by means of a method according to the invention, said digital filter being integrated into said loop so as to filter the control signals.

[0035] According to a third aspect, the invention relates to a computer program comprising a set of program code instructions which, when they are executed by a processor, configure said processor to implement a parameterization method according to invention.

[0036] According to a fourth aspect, the invention relates to a recording medium readable by a computer on which is recorded a computer program according to the invention.

According to a fifth aspect, the invention relates to a device for parameterizing a digital filter, said filter being intended to attenuate a torsion mode of a power transmission line of an aircraft turbine engine, said mode being associated with a frequency F_T included in a confidence interval le, the digital filter being of low-pass type and:

- intended to be integrated into a pre-existing control loop of the turbine engine, so as to filter control signals generated by a control device of said loop and sampled at a frequency F_E, said loop being closed and associated with a passband in which the gain of the loop is increased, in absolute value, by a V value,

said device comprising:

- a calculation module, configured to calculate, as a function of the frequencies F_T and F_E, complex numbers forming zeros of N (z), so that the filter attenuates the frequency F_T,

- an update module, configured to update the zeros of N (z), so that the gain of the filter satisfies, in the confidence interval le, a first predetermined gain mask as a function of the amplitude of the torsion mode,

- a determination module, configured to determine real numbers forming poles of D (z), so that, in the bandwidth of the loop:

[0038] According to a sixth aspect, the invention relates to an aircraft comprising a

turbine engine, said turbine engine comprising a power transmission line exhibiting a torsion mode associated with a frequency F_T within a confidence interval le. In addition, said aircraft comprises a control system according to the invention.

Brief description of the drawings

[0039] Other characteristics and advantages of the present invention will emerge from the description given below, with reference to the accompanying drawings which illustrate an exemplary embodiment thereof without any limiting nature. In the figures:

[Fig. 1] FIG. 1 schematically represents an embodiment of a system, called a power transmission system, of an aircraft turbine engine;

[Fig. 2] FIG. 2 schematically represents an example, known to those skilled in the art, of operation, in nominal mode, of a control system of a turbine engine; [Fig. 3] FIG. 3 schematically represents the gain of the transfer function associated with the closed control loop according to which the control system of FIG. 2 is configured;

[Fig. 4] FIG. 4 represents a flowchart of an embodiment of a method for parameterizing a digital filter according to the invention, said method making it possible to attenuate a mode of torsion of a power line of a turbine engine

[Fig. 5] FIG. 5 schematically represents a preferred mode of implementation, according to the invention, of the parameterization method of FIG. 4, in which said method comprises a step of validating the temporal behavior of the digital filter;

[Fig. 6A] FIG. 6A represents the change in the gain of the digital filter obtained, during the parameterization process according to the invention, after a step of calculating the zeros of said filter;

[Fig. 6B] FIG. 6B represents the evolution of the phase of the digital filter obtained, during the parameterization process according to the invention, after a step of calculating the zeros of said filter;

[Fig. 7A] FIG. 7A represents the evolution of the gain of the digital filter obtained, during the parameterization process according to the invention, after a step of updating the zeros of said filter;

[Fig. 7B] FIG. 7B represents the evolution of the phase of the digital filter obtained, during the parameterization process according to the invention, after a step of updating the zeros of said filter;

[Fig. 8A] FIG. 8A represents the evolution of the gain of the digital filter obtained, during the parameterization process according to the invention, after a step of

determining the poles of said filter;

[Fig. 8B] FIG. 8B represents the evolution of the phase of the digital filter obtained, during the parameterization method according to the invention, after a step of determining the poles of said filter; Description of embodiments

[0040] The present invention finds its place in the field of turbine engines for aircraft, and more particularly in the field of the damping of one or more mechanical elements forming part of a turbine engine.

By "damping", we refer here to the control of the amplitude

oscillations, insofar as these are likely to be

associated with a frequency equal to, or at least close to, the frequency of an eigenmode characteristic of the assembly formed by said elements

mechanical, this mode being likely to lead to premature wear of said assembly when it is maintained over time or when it is repeatedly energized. In other words, the concept of damping corresponds to the fact of seeking a damping of oscillations evolving at a frequency

predetermined and likely to damage the mechanical components considered.

[0042] Figure 1 shows schematically an embodiment of a

system 10, called the power transmission system, of an aircraft turbine engine 1.

[0043] In practice, the power transmission system 10 may also include other elements than those shown in Figure 1, but which nevertheless fall outside the scope of the present invention.

As illustrated in Figure 1 by no means limiting, the system 10 of

power transmission comprises a gas generator 11. This generator 11 typically corresponds to a combustion chamber in which hydrocarbons are ignited to generate gases at high temperature as well as at high speed. The gases generated are then conveyed to a so-called power turbine 12, which is thus set in motion. Such a power turbine 12 may be designated by the expression "free turbine" by those skilled in the art in certain engine architectures.

A turbine shaft is coupled, at its respective ends, to the power turbine 12, as well as to a speed reducer 13 of the epicyclic type. Another shaft, called the output shaft, and opposite the turbine shaft by relative to the speed reducer 13, for its part extends between the speed reducer 13 and propulsion means 14 of the aircraft.

The turbine shaft is therefore rotated by the power turbine 12. The speed reducer 13 allows for its part to rotate the output shaft at a reduced speed relative to that of the shaft. turbine. Finally, the propulsion means 14 are in turn driven by the output shaft.

[0047] The assembly formed by the turbine and output shafts, as well as by the

speed reducer 13, is commonly referred to as a "power transmission line". It is indeed understood that this assembly is responsible for ensuring the transmission of the rotational energy of the power turbine 12 to the propulsion means 14.

The rest of the description is aimed more specifically, but in a manner

in no way limiting, a turbine engine 1 of the turboprop type for an airplane. This is therefore a turbine engine 1, the main thrust of which is obtained by the rotation of at least one propeller comprising a plurality of blades. For example, and preferably, the propulsion means 14 comprise two non-ducted contra-rotating propellers, which makes it possible in particular to improve the

propulsive efficiency.

[0049] However, following other examples not detailed here, nothing excludes

consider other types of turbine engines, such as a

turbojet. The invention is in fact applicable to any type of turbine engine whose operation is to be controlled so that the transmission line is not excited according to a torsion mode that is specific to it. There is also nothing to exclude considering another type of turbine downstream of the gas generator, such as for example a linked turbine of a type known per se, as well as an aircraft of another type, such as for example a helicopter.

It is noted that the mode of torsion of the transmission line results not

only the torsional flexibility thereof, this flexibility being a function of the materials used in its manufacture as well as of its length and its diameter, but also of the fact that it is rotated between the power turbine 12 and the propulsion means 14 which are, for their part, elements having an inertia much greater than that of the transmission line (thus than speed reducer 13 in this example). In other words, during nominal operation of the turbine engine 1, the transmission line is liable to be subjected to a torque capable of exciting its torsion mode according to a natural frequency F_T associated with this mode.

The operation of the turbine engine 1 is conventionally controlled by a control system 20 on board the aircraft.

FIG. 2 schematically represents an example, known to those skilled in the art, of operation, in nominal mode, of the control system 20 of the turbine engine 1. Such a figure is also designated by the expression "block diagram d 'enslavement'.

By "nominal mode", we refer here to the mode in which the control system 20 acts on the turbine engine 1 when the latter is subjected to constraints which may affect its operation, but which have nevertheless been taken into account in the design of turbine engine 1 before dynamic response tests of the transmission line are performed.

[0054] In a conventional manner, and as shown in FIG. 2, the

control system 20 comprises at the input means 21 for receiving a setpoint of a type known per se, such as for example a computer. According to this exemplary embodiment, the setpoint is representative of a desired speed of rotation of a rotor of the propulsion means. It should however be noted that other physical parameters can be considered to define the instruction, such as for example a predetermined orientation of the aircraft. The choice of a parameter depends in particular on the control strategy that is chosen to ensure the thrust of the aircraft.

Such a control system 20 comprises a control device 22

configured to generate control signals intended to be transmitted to actuators (not shown in Figure 2) of the aircraft. Such actuators are, for example means configured to deliver a quantity

of hydrocarbons determined at the gas generator 11, such as for example a fuel metering device.

The control device 22 comprises for example one or more

processors and storage means (magnetic hard disk, memory electronics, optical disc, etc.) in which data and a computer program are stored, in the form of a set of program code instructions to be executed to implement all or part of the control of the operation of the turbine engine 1. Alternatively or in addition, the control device 22 also comprises one or more programmable logic circuits, of the FPGA, PLD, etc. type, and / or specialized integrated circuits (ASIC), and / or a set of discrete electronic components, etc. adapted to implement the control of the operation of the turbine engine 1.

In other words, the control device 22 comprises a set of means configured in software (specific computer program) and / or hardware (FPGA, PLD, ASIC, etc.) to implement the control. operation of the turbine engine 1.

In a particular exemplary embodiment, the control device 22 is configured according to a "PID" type model (acronym of the expression "

Proportional, Integrator, Differentiator ") that a person skilled in the art knows how to implement. Such an exemplary embodiment nevertheless constitutes only an alternative embodiment, and nothing excludes having a control device configured according to a model of a different type not detailed here.

The control system 20 also comprises, at the output, a measurement of the speed of rotation of the rotor of the propulsion means 14, typically by means of dedicated sensors 23, such as for example phonic wheels. This measured speed is redirected to the input of the control system 20 so that the latter operates according to a closed loop control logic.

The control device 22 is an integral part of the closed control loop, and is therefore configured to generate a control signal on the basis of a difference between the speed setpoint and the speed measurement. The speed setpoint is therefore controlled, and the control device acts as a corrector to compensate for said difference. The control signal is then transmitted to the actuators, which has an effect on the turbine engine 1, and therefore ultimately also on the transmission line (change of speed of rotation of the turbine and output shafts, and therefore of the means of propulsion 14). It should be noted that a control signal generated in response to a deviation from the setpoint corresponds to a digital signal. The sampling frequency of the control signals generated during operation of the turbine engine 1 is denoted F_E in the remainder of the description, and is for example equal to 50 Hz. Nothing excludes however, according to other examples not detailed here , to consider a sampling frequency F_E other than 50 Hz.

The control loop as known until now, and illustrated in Figure 2, is associated with a transfer function representative of the frequency response, to a control signal, of the assembly formed by the actuators and the turbine engine 1. Insofar as this assembly corresponds to a physical assembly of the real world, the transfer function here is of the low-pass type.

It will be clear to those skilled in the art that the expressions "gain of the control loop" and "gain of the transfer function associated with the control loop" have the same meaning have the same meaning in the description which follows.

[0064] FIG. 3 schematically represents the gain of the transfer function associated with the closed control loop, and corresponds to a graph on a semi-logarithmic scale (Bode diagram). This graph has an x-axis representing the frequency f in Hertz (Hz), as well as an y-axis representing the gain of the GdB filter in decibels (dB).

As illustrated in Figure 3, the change in gain as a function of frequency comprises three areas, namely:

- a zone A corresponding to the pass band of the transfer function (in this example, it extends between 0 Hz and approximately 1 Hz), and in which the gain of the loop is increased, in absolute value, by a value V, for example equal to 0 dB,

- a zone C corresponding to the attenuated band of the transfer function (in this example, it extends beyond substantially 10 Hz),

- a zone B corresponding to the transition band of the transfer function, and located between zones A and C.

The frequency F_T of the torsion mode is located in zone B. Also, the frequency F_T is sufficiently close to zone A for it to be necessary to consider attenuating it, in order to avoid the excitation of the associated torsion mode, and therefore to avoid any risk of premature material wear. The present invention proposes a solution to this problem, a solution which does not require modifying the mechanical architecture of the turbine engine, nor modifying the pre-existing regulation logic of the operation of the turbine engine.

[0067] For example, the frequency F_T is equal to 7 Hz as well as associated with a

confidence interval le, the respective limits of which are 6.5 Hz and 7.5 Hz. Nothing excludes, according to other examples not detailed here, from considering other values for the frequency F_T as well as for the interval of trust him.

It should be noted that the frequency F_T is associated with an interval of

trust him. The existence of such a confidence interval is justified by the fact that the frequency F_T cannot be known with absolute precision, insofar as dispersions may exist from one motor to another during a mass production.

[0069] Thus, according to a preferred example, the frequency F_T is determined during a campaign of bench tests of the dynamic behavior of the power transmission line. It should be noted that such a test campaign is carried out once the mechanical sizing of the turbine engine as well as the design of the regulation logic have been completed. The confidence interval therefore depends on the precision of the measurements carried out during the tests, but also on the number of tests carried out according to statistical methods known to those skilled in the art, and not detailed here because outside the scope of the invention. .

However, nothing excludes considering other methods of obtaining the frequency F_T as well as its confidence interval le. For example, they can be obtained by digital simulation, which therefore requires fine modeling of the various mechanical parts forming the turbine engine as well as a mathematical simulation model representative of the dynamic behavior of these parts.

Nothing excludes either, for the choice of the, to take into account any other

parameter that would be identified as the source of variation of the frequency F_T, such as for example the dispersions linked to the manufacturing processes of the parts which constitute the power transmission line or even the evolution of these parameters during the life of these said parts.

[0072] FIG. 4 represents a flowchart of an embodiment of a method for parameterizing a digital filter for the attenuation of the torsion mode associated with the frequency F_T.

Said parameterization method is implemented by a control device.

parameter setting (not shown in the figures) which comprises for example one or more processors and storage means (magnetic hard disk, electronic memory, optical disc, etc.) in which data and a computer program are stored, under the form of a set

program code instructions to be executed in order to implement all or part of the steps of the parameterization method. Alternatively or in

additionally, the parameter-setting device also comprises one or more programmable logic circuits, of the FPGA, PLD, etc. type, and / or specialized integrated circuits (ASIC), and / or a set of discrete electronic components, etc. suitable for implementing all or part of the steps of the parameterization process.

In other words, the parameter setting device comprises a set of means configured in software (specific computer program) and / or hardware (FPGA, PLD, ASIC, etc.) to implement the various steps. of the parameterization process.

The digital filter parameterized by means of the parameterization method is

intended to be integrated into the pre-existing control loop so as to filter the control signals generated by the control device 22. In other words, the digital filter is intended to be integrated, for example in software, at the output of the control device 22.

The digital filter of the present invention is sought in the form of a low pass filter, in particular so as not to disturb the behavior of the transfer function associated with the control loop.

The digital filter is further sought so that the transfer function associated with it is causal, stable and equal to the quotient N (z) / D (z), where N and D are polynomial functions, N being of degree strictly greater than 1. The transfer function is therefore a rational fraction. As it is causal, this implies that the degree of the denominator is strictly greater than that of the numerator. The stability criterion implies that the poles of D (z) are all included in the unit circle of the complex plane. It will also appear clearly to those skilled in the art that the argument z of the functions N (z) and D (z) corresponds to the notation of a complex variable conventionally used for the manipulation of discrete signals, the link of which with the continuous representation is done through the z transform. We then have the following formula:

z = exp (2m * p / F_E)

where p is the Laplace variable.

The roots of the numerator N (z) and of the denominator D (z) are called

zeros and poles respectively.

The parameterization method comprises several steps. In general principle, the process consists first of placing the zeros of the numerator in order to target the attenuation of the frequency F_T. These zeros are then updated to take into account the uncertainty of the value of the frequency F_T. It is only after the numerator setting is completed that the denominator in turn is set by placing its poles, primarily to adjust the phase of the digital filter.

The parameterization method firstly comprises a step 100 of calculating, as a function of the frequencies F_T and F_E, of complex numbers forming zeros of N (z), so that the filter attenuates the frequency F_T.

[0081] The objective of step 100 is to target a first placement of the zeros of N (z) in order to ensure an attenuation of the frequency F_T.

In a preferred embodiment, the degree of N (z) is equal to 2. This implies that N (z) has two zeros denoted z_1 and z_2 respectively. These zeros z_1 and z_2 are calculated according to the following formulation during calculation step 100:

z_1 = exp ((2xixnxF_T) / F_E) and z_2 = exp ((- 2xixnxF_T) / F_E).

Calculating the zeros z_1 and z_2 in this way amounts to determining a digital filter precisely targeting the frequency F_T as being the frequency to be attenuated. The fact that such zeros are first determined on the unit circle is representative of the poor damping of the frequency F_T by the power transmission line.

However, nothing excludes calculating, during step 100, zeros z_1 and z_2 in a different manner, since the latter make it possible to exclude a frequency zone substantially centered around the frequency F_T associated with the mode. twist. Preferably, the zeros are determined during step 100 close to the unit circle (therefore of modulus substantially equal to 1), ideally on the unit circle, in order to easily initialize the parameterization process.

The fact of choosing a degree of N (z) equal to 2 makes it possible to limit the complexity of the filter. However, it should be noted that this choice is only an alternative implementation of the invention. For example, N can be parameterized so as to be of degree strictly greater than 2, for example equal to 4, as long as the constraint according to which the filter is causal is respected.

The parameterization method then comprises a step 200 of updating the zeros of N (z), so that the gain of the filter satisfies, in the confidence interval le, a first amplitude mask predetermined as a function of of the amplitude of the torsion mode.

This step 200 of updating the zeros makes it possible to take into account

the uncertainty associated with the value of the frequency F_T, and therefore to make the final attenuation sought for the digital filter more robust with respect to this uncertainty.

In a particular embodiment, the step 200 for updating the zeros of N (z) comprises a sub-step of reducing the respective moduli of the zeros according to a predetermined step. The fact of reducing the respective moduli of the zeros determined during the calculation step 100 makes it possible to move them away from the unit circle (towards the interior of the unit circle), and therefore to thus widen the attenuation zone of the filter so as to take into account the confidence interval le.

The sub-step of reducing the modules is then iterated as long as the

first gain template is not satisfied.

For example, the modulus reduction step is set equal to 0.01. In this way, and during a first iteration of the reduction sub-step, the Updated zeros have modulus equal to 0.99. It is further understood that if the reduction sub-step is iterated for example five times, the zeros obtained at the end of the update step 200 will have a modulus equal to 0.95.

However, following other examples not detailed here, nothing excludes

consider a step greater or less than 0.01. Nothing excludes either considering a reduction in moduli which is not of the additive type, but for example of the multiplicative type.

The fact of gradually reducing the modules by iteration makes it possible to obtain a good compromise between the calculation time and the complexity of the configuration.

[0093] However, it should be noted that this way of proceeding only constitutes a variant implementation of the invention. For example, the update of the zeros of step 100 can be carried out by means of an optimization algorithm, for example a shape optimization algorithm, aiming to optimize a predetermined cost function as a function of the amplitude of the torsion mode. Such an optimization algorithm nevertheless increases the complexity of the step 200 of updating the zeros of N (z).

By way of non-limiting example, the first gain template corresponds, in the confidence interval le, to an increase, preferably strict, of the value of the gain by the opposite of the amplitude of the resonance of the twist mode. For example, if the torsion mode results in a 3 dB peak on the Bode diagram of the system at the frequency F_T, the filter is designed so that the gain minimally compensates for this amplification. In terms of template, this results in a maximum constraint of -3dB at the frequency F_T.

In fact, the updating of the zeros according to the invention reduces, in absolute value, the amplitude of the gain of the filter relative to the gain obtained at the end of the single calculation step 100. Therefore, fixing such a first gain mask allows to provide a constraint to stop updating zeros, so that the filter will completely attenuate the frequency F_T in the confidence interval le.

The choice of such a first gain template is only a variant

implementation of the invention. Other variants are therefore conceivable, such as having a first gain template corresponding to a gain value, in the interval le, greater than or equal to the opposite of the amplitude of the torsion mode. For example, the first gain template can correspond to a gain value of between 90% and 95% of the amplitude of the torsion mode. In fact, the step subsequent to the step 200 of updating the zeros, and which is described later, has the effect of further reducing the gain of the digital filter at the frequency F_T, so that it is possible to achieve a compromise between the amplitude of the torsion mode, the gain of the filter at the end of step 200 and the length of the interval Ic.

It is therefore understood that steps 100 and 200 are essentially aimed at

parameterize the digital filter to calibrate its gain around the frequency F_T, the phase and gain calibration on the rest of the frequency spectrum, and in particular on the passband of the closed loop, being carried out

later.

For this purpose, the parameterization method comprises a step 300 of

determination of real numbers forming poles of D (z). The fact of finding the poles of D (z) in the form of real numbers makes it possible to guarantee the

damped behavior of the filter.

Said step 300 is performed under constraints, namely that the poles of D (z) are determined so that, in the bandwidth of the loop:

[0100] This step 300 therefore aims to place the poles of D (z) so as to control the evolution of the phase of the filter over the passband of the control loop, which ultimately makes it possible to control the phase of the filter. over the entire frequency spectrum considered (zones A, B and C). Furthermore, the fact of also constraining the gain of the filter in this way in the passband of the loop makes it possible to guarantee that the useful information contained in the control signals can continue to be conveyed to the actuators of the turbine engine 1 without attenuation. In a preferred embodiment, the poles of D (z) are all considered to be equal. Considering the poles all equal to each other makes it possible to obtain an advantageous compromise between the complexity of the parameterization (and therefore the calculation time and the necessary calculation means) and the precision of the behavior of the filter. It should nevertheless be noted that the choice of poles which are all equal to one another only constitutes an implementation variant of the invention. For example, the real poles can be determined so as to be all distinct from each other, or even so that only certain poles are equal to each other.

[0102] In a particular embodiment, the step 300 of determining the poles of D (z) comprises:

- a sub-step of selecting a pole strictly between -1 and 1,

- a sub-step of translating the selected pole along the real axis and at a predetermined pitch, so as to obtain a translated pole.

The translation sub-step is then iterated as long as the phase template and the second gain template are not satisfied. In order to carry out such an iteration, the pole selected during an iteration corresponds to the translated pole obtained during the previous iteration.

[0104] For example, said translation step along the real axis is set equal to 0.01 in the direction of decreasing reals. In this way, and during a first iteration of the translation sub-step, the pole is equal to 0.89 if the pole initially selected is equal to 0.9. It is further understood that if the translation substep is iterated for example five times, the pole obtained at the end of step 300 will be equal to 0.85 if the very first pole selected is equal to 0.9.

[0105] Nothing, however, excludes, following other examples not detailed here, from

consider a step greater or less than 0.01, as well as a translation in the direction of increasing real numbers. Typically, the direction of translation depends on the position of the pole initially selected relative to the terminals -1 and 1. Nothing excludes either having a direction of translation along the real axis which changes between at least two iterations , for example in the case where the direction of translation is determined by means of an optimization algorithm aimed at optimizing a predetermined cost function as a function of the frequency F_E and of the value V. The fact of determining the pole of D (z) by iteration of translations makes it possible to obtain a good compromise between the calculation time and the complexity of the parameterization.

[0107] However, it should be noted that this way of proceeding constitutes only a variant implementation of the invention. Thus, according to considerations similar to those described above in the case of step 200, the determination of the pole of D (z) can be carried out by means of an optimization algorithm, for example an optimization algorithm of form. Such an optimization algorithm, however, increases the complexity of step 300.

[0108] By way of non-limiting example, the phase mask corresponds to an increase in the phase shift introduced by the filter on zone A, which corresponds to the passband of the closed loop. For example, this increase corresponds to a predetermined multiple of the product of period 1 / F_E by the maximum pulse delimiting zone A.

[0109] The fact of increasing the phase shift in the passband of the loop makes it possible to avoid an excessive distortion of the phase during the control of the operation of the turbine engine 1. In other words, proceeding in this way advantageously makes it possible not to destabilize the existing control loop, by limiting the phase shift effects introduced, it being understood that any digital processing necessarily generates effects on the phase.

[0110] In addition, the fact of normalizing the gain of the filter makes it possible not to impact the gain of the closed loop at low frequency.

[01 1 1] In a preferred embodiment, the numerator D (z) is set so as to be of a degree equal to 3. This choice advantageously makes it possible to meet the need while limiting the complexity of the filter. However, it should be noted that this choice is only an alternative implementation of the invention. For example, D (z) can be parameterized so as to be of degree strictly greater than 3, for example equal to 5, as long as the constraint according to which the filter is causal is respected.

[01 12] FIG. 5 diagrammatically represents a preferred embodiment of the parameterization method, in which said method comprises, following step 300 of determining real numbers forming poles of D (z), a step 400 for validating the temporal behavior of the digital filter. By “validation of the temporal behavior”, one refers here to the fact of checking that the output of the digital filter follows in time an expected behavior in response to a known input signal.

[0113] In other words, the validation step 400 makes it possible to ensure that the filter

digital parameterized according to the invention does not exhibit unsuitable behavior.

In said preferred embodiment, said validation step 400 consists of verifying that the temporal response of the filter to a step signal

(Heaviside function) is monotonically increasing. Such a verification step corresponds to the study of the step response of the digital filter. It is therefore not detailed here further. It is only specified that the temporal behavior of the filter is effectively validated when said temporal response is increasing monotonously.

[0115] Once the validation step 400 is complete, and when the behavior

time of the parameterized filter is not ultimately satisfactory, step 300 of determining the poles of D (z) and step 400 of validation are iterated until the behavior of the digital filter is validated. In other words, the poles of D (z) are readjusted. To readjust the poles of D (z), step 300 can for example be executed by choosing real poles which are all identical, but for which the first pole selected before any translation differs from that selected during the first implementation of the method leading to the filter whose behavior is not satisfactory.

FIGS. 6A, 6B, 7A, 7B, 8A, 8B represent the respective evolutions of the gain (FIGS. 6A, 7A, 8A) and of the phase (FIGS. 6B, 7B, 8B) of the digital filter obtained step by step during of an example of implementation of said parameterization method.

In this example of implementation, the frequencies F_E and F_T are

respectively equal to 50 Hz and 7 Hz. The interval corresponds here to [6.5 Hz,

7.5 Hz], and the value V is taken to be 0 dB. Further, the filter transfer function is sought in the form of an N (z) / D (z) quotient, where N is of degree 2 and D is of degree 3 with the poles all equal to each other. [0118] FIG. 6A schematically represents the frequency evolution of the gain of the digital filter at the end of the step 100 of calculating the parameterization method. As illustrated in FIG. 6A, the gain increases considerably, in absolute value, at the level of the frequency F_T, which corresponds well to the

expected behavior for targeted mitigation.

[0119] Figure 6B shows schematically, for its part, the evolution

frequency of the phase of the digital filter at the end of step 100 of calculating the parameterization process. As illustrated in Figure 6B, the phase of the filter is not yet under control since it increases beyond 180 ° for frequencies above the frequency F_T.

[0120] Note that the zeros of N (z) determined at the end of step 100 are respectively equal to 0.637 + 1 * 0.771 and 0.637 - i * 0.771.

[0121] It is also noted that FIGS. 6A and 6B have been obtained by simulation of the behavior of the digital filter, thanks to the parameter setting device.

[0122] Figures 7A and 7B correspond to the respective updates of Figures 6A and 6B, once the step 200 of updating the zeros has been performed.

As can be seen in FIG. 7A, the gain of the digital filter has decreased, in absolute value, in the vicinity of the frequency F_T, more

specifically in the meantime. The fact remains that this gain remains, in absolute value, greater than the amplitude of the torsion mode.

[0124] The phase, for its part, and as illustrated in FIG. 7B, has hardly changed.

[0125] Moreover, the updated zeros associated with the case of FIGS. 7A and 7B

are respectively 0.606 + i * 0.732 and 0.6064 * 0.732. We therefore see a decrease in modulus relative to the zeros obtained at the end of step 100 and associated with Figures 6A and 6B.

[0126] FIGS. 8A and 8B correspond to the respective updates of FIGS. 7A and 7B, once step 300 of determining the poles has been carried out. Note that at the end of step 300, the poles are all determined to be 0.5.

[0127] As can be seen in FIG. 8A, the gain of the digital filter has increased, in absolute value, in the vicinity of the frequency F_T, more specifically in the meantime. There is therefore a very strong attenuation targeted on the mode of torsion of the power transmission line. In addition, the gain remains below 0 dB in the passband, which means that the digital filter modifies the low frequency control signal very slightly.

[0128] The phase, for its part, and as illustrated in FIG. 8B, remains between 0 ° and 180 ° in absolute value over the whole of the frequency spectrum envisaged (zones A, B and C), which makes it possible to avoid any side effect during the execution of the control loop, such as for example an excessive phase shift which could lead to an inversion of the control signal.

[0129] In general, the invention remains of course applicable for a torsion mode positioned not in zone B but also in zone A or else zone C.

[0130] The invention therefore advantageously makes it possible to configure a digital filter to effectively attenuate the torsion mode of the power transmission line around the frequency of said torsion mode, without degrading the gain of the pre-existing control loop on the rest. of the frequency spectrum, by limiting the phase shift effects introduced, and without any undesirable temporal behavior being introduced into the pre-existing regulation logic. In addition, the parameterization method makes it possible to obtain a very efficient digital filter of reasonable order, typically less than 5, for example equal to 3, which is compatible with a real-time implementation in the pre-existing control logic.

[0131] Finally, it should be noted that a transfer function associated with a filter

digital parameterized according to the invention can be easily implemented in pre-existing control software. To this end, those skilled in the art have access to function libraries allowing them to generate such a digital filter at the output of the control device 22 of the control loop.

Claims

[Claim 1] Method of parameterizing a digital filter for

the attenuation of a torsion mode of a power transmission line of an aircraft turbine engine (1), said mode being associated with a frequency F_T within a confidence interval le,

the digital filter being of low-pass type and:

- comprising a causal transfer function in z, stable and equal to the quotient N (z) / D (z), where N and D are polynomial functions, N being of degree strictly greater than 1,

- intended to be integrated into a pre-existing control loop of the turbine engine (1), so as to filter control signals generated by a control device (22) of said loop and sampled at a frequency F_E, said loop being closed and associated at a passband in which the gain of the loop is increased, in absolute value, by a value V,

said method being implemented by a parameter setting device and comprising:

- a step (100) of calculating, as a function of the frequencies F_T and F_E, of complex numbers forming zeros of N (z), so that the filter attenuates the frequency F_T,

- a step (200) of updating the zeros of N (z), so that the gain of the filter satisfies, in the confidence interval le, a first predetermined gain template as a function of the amplitude of the mode of torsion,

- a step (300) of determining real numbers forming poles of D (z), so that, in the bandwidth of the loop:

[Claim 2] The method of claim 1, wherein the step (200) of updating the zeros of N (z) comprises a substep of decreasing the respective moduli of the zeros by a predetermined step, the substep the decrease being performed iteratively until the first amplitude template is satisfied.

[Claim 3] A method according to any one of claims 1 to 2, in which the poles of D (z) are all considered equal to each other, the step (300) of determining the poles comprising:

- a sub-step of selection of a pole strictly between -1 and 1,

a sub-step of translating the selected pole along the real axis and according to a predetermined pitch, so as to obtain a translated pole,

[Claim 4] A method according to any one of claims 1 to 3, wherein the first gain mask corresponds to an increase, in the confidence interval le, of the value of the gain by the opposite of the amplitude of the gain. twist mode.

[Claim 5] A method according to any one of claims 1 to 4, in which the phase mask corresponds to an increase in the phase shift introduced by the filter in the passband of the closed loop.

[Claim 6] A method according to any one of claims 1 to 5, in which the degree of N (z) is equal to 2, so as to obtain, during the step (100) of calculation, zeros z_l and z_2 according to the following formulation:

z_l = exp ((2xixnxF_T) / F_E) and z_2 = exp ((- 2xixnxF_T) / F_E).

[Claim 7] A method according to any one of claims 1 to 6, wherein the degree of D (z) is equal to 3.

[Claim 8] A method according to any one of claims 1 to 7, in which the frequency F_T and the confidence interval are determined beforehand during a campaign of tests on a test bench of the dynamic behavior of the transmission line. power transmission.

[Claim 9] A method according to any one of claims 1 to 8, said method comprising, following the step (300) of determining real numbers forming poles of D (z), a step (400) of validation of the temporal behavior of the digital filter, said validation step (400) consisting in verifying that the temporal response of the filter to a step signal is increasing monotonously,

the step (300) of determining the poles of D (z) and the step (400) of validation being iterated as long as the behavior of the digital filter is not validated.

[Claim 10] Control system (20) intended to be carried on board an aircraft comprising a turbine engine (1), said turbine engine (1) comprising a power transmission line exhibiting a torsion mode associated with a frequency F_T included in a range of confidence le, said system (20) comprising means for receiving (21) a setpoint relating to a predetermined parameter, a control device (22) configured to generate control signals sampled at a frequency FE and means for measurement (23) of said parameter, said control system (20) forming a closed control loop associated with a pass band in which the gain is increased, in absolute value, by a value V, said system being characterized in that the loop control comprises a digital filter parameterized by means of a method according to any one of claims 1 to 9, said digital filter being integrated in said loop so as to filter the signals of ordered.

[Claim 11] Computer program comprising an assembly

program code instructions which, when executed by a processor, configure said processor to implement a parameterization method according to any one of claims 1 to 9.

[Claim 12] A computer readable recording medium on which is recorded a computer program according to claim 11.

[Claim 13] Device for parameterizing a digital filter, said filter being intended to attenuate a torsion mode of a power transmission line of an aircraft turbine engine (1), said mode being associated with a frequency F_T within a confidence interval on, the digital filter being of low-pass type and:

said device comprising:

- a determination module, configured to determine real numbers forming poles of D (z), so that, in the bandwidth of the loop

[Claim 14] An aircraft comprising a turbine engine, said turbine engine comprising a power transmission line exhibiting a torsion mode associated with a frequency F_T included in a confidence interval le, said aircraft further comprising a control system according to claim 10.