Turbomachine monitoring method and associated system

EP4594601A1Pending Publication Date: 2025-08-06SAFRAN SA
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
EP2023793418
Authority / Receiving Office
EP · EP
Patent Type
Applications
Current Assignee / Owner
Priority Date
2022-09-28
Filing Date
2023-09-13
Publication Date
2025-08-06

AI Technical Summary

Technical Problem

Current turbomachine monitoring methods are passive, requiring complete signal acquisition before analysis, which prevents real-time detection and corrective actions during operation, leading to inefficiencies and wear due to clearance fluctuations between rotor and stator blades.

Method used

A method involving a strain gauge and position sensor that captures deformation and blade position signals, using a low-pass filter to calculate touch indicators in real-time or near real-time, allowing for immediate corrective actions by modifying blade tip clearance.

Benefits of technology

Enables active monitoring and rapid correction of blade tip clearance, improving turbomachine efficiency and reducing wear by detecting touches in real-time, unlike previous passive methods that required offline processing.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 1.1
    Figure 1.1
Patent Text Reader

Abstract

One aspect of the invention relates to a method (2) for monitoring a turbomachine (1) comprising a stator (11), a rotor (12) comprising blades (13), a strain gauge (15) attached to the stator (11) and a position sensor (14) for detecting the position of the blades (13), the gauge (15) being configured to capture a first signal comprising two components, the position sensor (14) being configured to capture a second signal representing the position of each blade (13), the method (2) comprising, on each revolution of the rotor (12), with one revolution being detected on the basis of the second signal: acquiring (21) the first and second signals; angularly resampling (22) the first signal; extracting (23) the first component from the resampled first signal by sliding-filtering the resampled first signal; and, on the basis of the filtered resampled first signal, calculating (24) at least one contact indicator for each blade (13) for the revolution.
Need to check novelty before this filing date? Find Prior Art

Description

DESCRIPTION TITLE: Turbomachine monitoring method and associated system TECHNICAL FIELD OF THE INVENTION

[0001] The technical field of the invention is that of turbomachinery monitoring.

[0002] The present invention relates to a method of active monitoring of a turbomachine and in particular a method of active monitoring of touches between a stator and a rotor blade of a turbomachine, making it possible to obtain a touch indicator between a stator and a rotor blade at each rotation of the rotor. TECHNOLOGICAL BACKGROUND OF THE INVENTION

[0003] A turbomachine is a machine that transforms the kinetic energy of a fluid into mechanical energy, and vice versa, via a rotating assembly called a rotor. The stationary part of the turbomachine is called the stator.

[0004] These turbomachines have different designs depending on their function: turbines, pumps, compressors, turbochargers, turboshaft engines, etc. Nevertheless, most of them share a common architecture consisting of a rotor, which is a rotating part mounted on a shaft, and a stator, which is a stationary part connected to a structure via a bearing. The structure is, for example, that of an aircraft engine. The rotor comprises a plurality of blades designed to accelerate the airflow passing through the turbomachine. To establish relative rotational motion between the rotor and the stator, the blades must be spaced apart from the stator. This spacing is commonly called the "blade tip clearance" or "JSA." In an aircraft engine, the blade tip clearance is an important design and operating parameter of a turbomachine.Minimizing this clearance prevents excessive amounts of airflow from bypassing the rotating blade array, thus improving the turbomachine's energy efficiency. However, this clearance is subject to fluctuations due to thermal expansion and mechanical phenomena dependent on the turbomachine's operating cycle.

[0005] In some cases, these clearance fluctuations occur asymmetrically, and the clearance between certain rotor blades and the stator becomes zero. Zero clearance This results in friction, also known as contact, between the rotating and stationary parts of the turbomachine. This can lead to excessive wear on the stator's abradable surface and a loss of efficiency caused by increased leakage at the blade tips.

[0006] For these reasons, it is desirable to detect and monitor keystrokes. Identifying these keystrokes allows for monitoring and controlling these phenomena within a preventive maintenance framework. The amplitude of the proposed indicators can be used indirectly to assess wear on contact interfaces following keystrokes, within a preventive maintenance context.

[0007] To this end, French patent application FR3104255A1, entitled "METHOD FOR MONITORING A TURBOMACHINE, DEVICE, SYSTEM, AIRCRAFT AND COMPUTER PROGRAM PRODUCT," concerns a monitoring method for detecting scuffing by analyzing signals from strain gauges attached to the rotor or stator. This monitoring method requires complete acquisition of the vibration signal over several rotor revolutions, followed by a signal processing method enabling the separation of blade contributions and the calculation of indicators "offline," i.e., after a period of turbomachine operation, as the calculations require significant computing power. The aim is therefore to characterize these scuffs in order to estimate the stresses and fatigue of the blades. This invention is of interest for developing turbomachine degradation models and for passive turbomachine monitoring.

[0008] One limitation of FR3104255A1 is that the monitoring performed is passive, in that it requires acquisition of the complete signal before analysis, not allowing for the implementation of corrective actions during the operation of the turbomachine after the detection of a touch.

[0009] There is a need to be able to carry out active monitoring, in real time or near real time, of the operation of a turbomachine and in particular of the play at the top of the blade, allowing corrective actions to be implemented for detected touches. SUMMARY OF THE INVENTION

[0010] The invention offers a solution to the problems mentioned above, by enabling active monitoring of the blade tip clearance of a turbomachine, by providing, at the end of each rotor revolution, a touch indicator between a blade and the stator for the rotor revolution for each rotor blade.

[0011] One aspect of the invention relates to a method for monitoring a turbomachine comprising a stator, a rotor comprising blades, a strain gauge fixed to the stator, and a rotor blade position sensor. The strain gauge is configured to capture a first signal comprising a first component representing stator deformations caused by rotation of the rotor blades relative to the stator, and a second component representing stator deformations caused by distinct elements of the rotor blades. The rotor blade position sensor is configured to capture a second signal representing the position of each rotor blade. The method comprises, at each current rotor revolution, a rotor revolution being detected from the second signal: acquisition of the first and second signals.Angular resampling of the first signal from the second signal to obtain a first resampled signal with a predefined angular resolution; extraction of the first component of the first resampled signal by filtering the first resampled signal; the filtering comprising the calculation, for each sample of the first resampled signal, of a sum weighted by a low-pass filter coefficient; the weighted sum comprising the sum of the value of the sample of the first resampled signal of the current rotor revolution and the value of each sample of a plurality of samples of the same angular position of the first component of the first resampled signal extracted during a predefined integer number of previous rotor revolutions; calculation, from the filtered first resampled signal, of at least one touch indicator for each rotor blade for the current rotor revolution.

[0012] Thanks to the invention, a touch can be detected, when it occurs, at each rotor revolution, therefore in real time or near real time. This is made possible by the use of a linear, time-invariant, causal, impulse-response filter. finite, as well as a blade position signal and a stator deformation signal. The blade position signal allows for angular resampling of the time-domain deformation signal, i.e., at a constant angular step, in order to avoid problems related to changes in the turbomachine's rotational speed.

[0013] The filter is applied in a sliding fashion, meaning that with each new rotation of the turbomachine's rotor, it is applied to the deformation signal captured during that rotation and to the deformation signal captured during a plurality of previous rotations. This allows for a real-time or near-real-time touch indicator, as the computation is resource-efficient, enabling it to be integrated into the turbomachine's control computer and implemented during turbomachine operation.

[0014] It is then possible to transmit this touch information to the turbomachine's microcontroller. The microcontroller can then modify the blade tip clearance based on the received touch information, unlike prior art solutions, which required a large dataset and therefore did not allow for adjustment of the blade tip clearance during turbomachine operation. Thanks to the invention, the turbomachine has greater efficiency and less wear than in the prior art, because the correction of the blade tip clearance is rapid after a touch is detected.

[0015] In addition to the features mentioned in the preceding paragraph, the monitoring method according to one aspect of the invention may have one or more of the following complementary features, considered individually or in all technically possible combinations: the method further comprises a step of detecting, in at least one of the calculated touch indicators, a touch between the stator and the blade corresponding to the indicator; a rotor revolution is detected as two passages of the same rotor blade; the passages of the same rotor blade are extracted from the second signal based on predefined information regarding the number of turbine rotor blades; the extraction of the first component of the first resampled signal by filtering the first resampled signal includes the application of the filter L defined according to the number of turns of rotors on which the weighted sum is calculated, P the number of rotor blades, L the angular resolution, and denotes the coefficients of the low-pass filter. The weighted sum is calculated using the following formula: m k (n) = m k the first component extracted from the first resampled signal, I the number of rotor revolutions over which the weighted sum is calculated, P the number of rotor blades, L the angular resolution, x k the first resampled signal acquired during the current rotor rotation, x^™ 9 a concatenation of the first mechanical component of each first signal of each previous rotor revolution, ht the filter coefficients. The low-pass filter coefficient is an average filter coefficient calculated according to the formula h m = with / the number of rotor revolutions over which the weighted sum is calculated, the calculation, from the first filtered resampled signal, of at least one touch indicator for each rotor blade per rotor revolution is performed according to the formula: With p a blade number among the rotor blades, k the current rotor revolution, A^ the indicator of the p-th rotor blade at the k-th rotor revolution, L the angular resolution, x k the first resampled signal acquired during the current rotor revolution, m k the first component extracted from the first resampled signal, a concatenation step of the first signal acquired at the current rotor revolution with a set of the first signals acquired at previous rotor revolutions to form with x k Lona the concatenated signal, x^ 1 the concatenation of the first mechanical component of each first signal acquired at the previous rotor revolutions, xk the first resampled signal acquired during the current rotor revolution, I the number of rotor revolutions on which the weighted sum is calculated, P the number of rotor blades, L the angular resolution, a touch is detected when a value of the touch indicator exceeds the value of the sum of an average of the touch indicators and three times a standard deviation of the touch indicators calculated during a so-called "reference" (no touch) operation of the turbomachine.

[0016] Another aspect of the invention relates to a monitoring system for a turbomachine comprising a stator and a rotor comprising blades, the monitoring system comprising a turbomachine monitoring device, a strain gauge fixed to the turbomachine stator, and a position sensor for the turbine blades of the turbine rotor. The strain gauge is configured to capture a first signal comprising a first component representative of stator deformations caused by rotation of the rotor relative to the stator, and a second component representative of stator deformations caused by distinct elements of the rotor blades. The position sensor for the rotor blades is configured to capture a second signal representative of the position of each rotor blade. The device comprises: an input for acquiring the first signal, an input for acquiring the second signal, anda data processing unit configured to implement the monitoring method according to the invention.

[0017] Another aspect of the invention relates to an aircraft comprising a turbomachine including a stator and a rotor configured to be driven in rotation relative to the stator, and a monitoring system (30) according to one aspect of the invention, in which the strain gauge is fixed to the stator.

[0018] Another aspect of the invention relates to a computer program product comprising program code instructions for executing the steps of the method for monitoring a turbomachine according to the invention, when this program product is executed by at least one data processing unit.

[0019] Another aspect of the invention relates to a computer-readable recording medium comprising instructions which, when executed by a computer, cause the computer to implement the monitoring method according to the invention.

[0020] The invention and its various applications will be better understood by reading the following description and examining the accompanying figures. BRIEF DESCRIPTION OF THE FIGURES

[0021] The figures are presented for illustrative purposes only and are in no way limiting to the invention. Figure 1 shows a schematic representation of a turbomachine monitored by a monitoring method according to the invention, Figure 2 shows a schematic representation of a blade position signal used in a turbomachine monitoring method according to the invention, Figure 3 shows a schematic representation of a turbomachine monitoring method according to the invention, Figure 4 shows a temporal and schematic representation of blade position and deformation signals used in a turbomachine monitoring method according to the invention, Figure 5 shows a temporal and schematic representation of touch indicators obtained by the turbomachine monitoring method according to the invention. DETAILED DESCRIPTION

[0022] Unless otherwise specified, the same element appearing on different figures has a unique reference.

[0023] The invention relates to a method for monitoring a turbomachine.

[0024] A turbomachine monitored by the method according to the invention is schematically represented in Figure 1.

[0025] The turbomachine 1 comprises (not shown) a nacelle, a fan, a low-pressure casing, a high-pressure casing, and a combustion chamber. The low-pressure casing includes a low-pressure compressor, a low-pressure turbine, and a low-pressure shaft connecting the low-pressure compressor to the low-pressure turbine. The high-pressure casing includes a high-pressure compressor, a high-pressure turbine, and a high-pressure shaft connecting the high-pressure compressor to the high-pressure turbine.

[0026] When the turbomachine is running, the high-pressure turbine drives the high-pressure compressor via the high-pressure shaft. The low-pressure turbine drives the low-pressure compressor and the blower via the low-pressure shaft. The blower generates a primary airflow and a secondary airflow (or bypass flow). The primary airflow passes successively through the low-pressure compressor, the high-pressure compressor, the combustion chamber, the high-pressure turbine, and the low-pressure turbine.

[0027] The high-pressure turbine comprises a stator 11 and a rotor 12 configured to be driven in rotation relative to the stator 11. The stator 11 is fixedly mounted on a structure, for example, on an aircraft structure, or for example, on the high-pressure turbine housing. The rotor 12 comprises a plurality of blades 13, for example, three blades as shown in Figure 1.

[0028] The turbomachine 1 is monitored by a monitoring system. The monitoring system includes a blade position sensor 14, a strain gauge 15 and a monitoring device 16. In Figure 1, the sensor 14 and the gauge 15 are fixed to the stator 11 of the turbomachine 1, and the monitoring device 16 receives signals from the sensor 14 and the gauge 15 via a connection, preferably a wired connection.

[0029] The blade position sensor 14 is a sensor that provides a signal representing the position of at least one blade 13 of the turbomachine 1. Preferably, such a sensor 14 is a blade passage sensor 14, that is, a sensor 14 that produces a signal when a blade 13 passes near the sensor 14, preferably in front of the sensor 14. This sensor 14, which allows the times of blade passage to be known, is also known as the " "Top-blade sensor." For example, the position sensor 14 is a proximity sensor, such as an optical probe, measuring the arrival time of the rotor blades 13 of the rotor 12. Such an optical probe, or optoelectronic sensor, detects a voltage signal in volts. Thus, when a blade 13 crosses the field of view of the sensor 14, the sensor 14 acquires a voltage signal in volts from the passage, with a bell-shaped waveform. As the blade approaches the sensor 14, the voltage signal increases to a maximum Vmax, corresponding to the instant when the blade 13 is aligned with the sensor 14. Beyond this point, the blade 13 moves away from the sensor, and the voltage signal gradually decreases. The waveform obtained during the passage of a wave is illustrated in Figure 2. The position signal, also called the dawn-passing signal, acquired is transmitted to device 16 via a first input 16-a of device 16.

[0030] The strain gauge 15 is configured to obtain a signal representative of a deformation of the stator 11. Such a strain gauge is also called a "strain gauge" and is preferably a resistive wire extensometer. The acquired strain signal is transmitted to the device 16 via a second input 16-b of the device 16. The strain signal comprises a first component representative of deformations of the stator 11 caused by a rotation of the rotor 12 relative to the stator 11 creating contact between the rotor blades 13 and the stator 11, and a second component representative of deformations of the stator 11 caused by separate elements of the rotor blades 13. This second component is also called "noise".

[0031] Device 16 is a monitoring device for the turbomachine 1 and includes an acquisition chain that, in particular, digitizes the received deformation and position signals. The monitoring device 16 also includes a processor and memory (not shown). The processor of device 16 is configured to execute the steps of a monitoring process that detects contact between the stator 11 and the rotor 12 of the turbomachine 1. To this end, the processor of device 16 implements a computer program. The computer program consists of program code instructions which, when executed by the processor of device 16, cause device 16 to implement the monitoring process according to the invention. The computer program is stored in the memory of device 16, and the processor accesses the program. by making a request to the memory. The memory can also be used to store other data, and in particular at least one of the signals received from sensors 14 and 15. An aircraft comprising a turbomachine 1 then includes the device 16 for monitoring the turbomachine 1 if it is desired that the turbomachine 1 be monitored by the method according to the invention.

[0032] The monitoring method 2 for turbomachine 1, implemented by the monitoring device 16, is schematically represented in Figure 3. The monitoring method 2 comprises two initialization steps 11 and 12, and six steps 21 to 26, implemented at each rotor revolution 12 of the turbomachine 1. The monitoring method 2 makes it possible to detect, at each rotor revolution 12 of the turbomachine 1, whether one or more touches have occurred and which blade(s) 13 are involved in this or these touches. In this respect, the monitoring of the turbomachine 1 is said to be "active" or "real-time or near-real-time," meaning that a touch is detected at the latest at the end of the rotor revolution 12 during which it occurred, unlike prior art solutions which require a complete acquisition of a large number of turbomachine revolutions to detect, a posteriori, the touches that occurred during this acquisition.

[0033] Method 2 according to the invention includes an initialization step, since in the invention, to detect a touch during a current rotor revolution 12, it is necessary to have acquired the deformation and blade position signals 13 during a predefined integer / - 1 of previous revolutions. The method therefore functions as a sliding window taking into account / revolutions, to detect a touch in the last rotor revolution 12 among the / revolutions.

[0034] The initialization (Init) of process 2 will now be described. The initialization comprises two steps, I1 and I2, allowing the accumulation of / revolutions of the rotor 12 of the turbomachine 1. "Accumulation" refers to the acquisition and storage of signals. The deformation signals acquired during this initialization period are denoted by x0(t):

[0036] In a first step 11, the position signals of at least one blade 13 and of the stator deformation 11 are acquired by the monitoring device 16, that is to say, they are received respectively from the position sensor 14 and the gauge of deformation 15, respectively via the first input 16-a and via the second input 16-b of the monitoring device 16.

[0037] The sampling frequency of each input signal is configurable. The sampling frequency must be set sufficiently high, for example, greater than 100 kHz, and the signals are acquired under steady-state operating conditions, i.e., conditions where the speed variation is less than 5%. The two signals obtained are called v(t) for the blade position signal and x(t) for the deformation signal. These signals are shown in Figure 4, with respect to time, in normal operation of the turbomachine 1 comprising three blades 13, as shown in Figure 3. The blade position signal v(t) consists of a series of bell curves representing the passage of the different blades 13 of the turbomachine 1. This signal therefore makes it possible to calculate the times of passage of each of the blades 13 at a precise point, for example, near the sensor 14 or when the blade 13 and the sensor 14 are aligned, as well as the rotation period of the rotor 12.

[0038] Either the moment of passage of the p ième dawn in the i ième Turbomachine 1 revolution. These transition times are calculated from the detection of a rising edge of the position signal above a predefined threshold v0 < V max , i.e.

[0040] This is illustrated in Figure 3. Regarding the strain signal x(t), it generally includes a transient wave in the presence of contact between a rotor blade 13 and the stator 11. This transient wave is due to the vibrational response to the impact caused by this contact. In the absence of this contact, the strain signal is theoretically zero. However, the presence of strong environmental noise, also called the second component of the strain signal, renders this distinction ineffective, hence the need for appropriate processing to eliminate these interferences.

[0041] To compensate for velocity fluctuations that cause variations in the time length of the periods, the deformation signal is transformed into an angular signal at a step I2 using cubic interpolation, preferably cubic spline interpolation; that is, it is resampled to an angular signal from the position signal. The position signal is used to determine the start and end of a rotor revolution. A uniform angular grid 0 = s. A0 is therefore fixed, where s is a A positive integer denoting the index relative to the angular position and A0 is the angular increment corresponding to the difference between two samples and calculated as follows:

[0042] A0 = — P*L

[0043] With P representing the number of blades 13 of the rotor 12 of the turbomachine 1 and L an integer representing the angular resolution. In order to avoid any loss of information, L must satisfy the following condition:

[0045] With f rotor denoting an average operating speed of the turbomachine 1 and f e The sampling frequency, 1.05, designates a maximum velocity fluctuation of 5%. The value 1.05 can be modified without departing from the invention, according to the fluctuation requirements.

[0046] The angular signal obtained after initialization is therefore a vector 1 ... I

[0048] The length of the vector x l ° ng is therefore equal to L * P * I. The points L are obtained by interpolation. It should be noted that the angular signal has a length of L * P * I.

[0049] After the initialization which enabled obtaining the deformation and position signals for I rotor revolutions 12 of the turbomachine 1, the following steps of the process 21 to 26 are reproduced at each current rotor revolution 12 of the turbomachine 1.

[0050] In a first step 21, the deformation and blade position signals 13 are acquired, for the current rotor revolution 12, in the same way as in step 11, i.e., by receiving the signals through the monitoring device 16. Considering the current rotor revolution 12 as the K-th iteration, a rotor revolution is defined as being between two successive passes of the first blade 13, therefore between the instants Indeed, if a first blade 13 of the rotor 12 passes twice in succession in front of the position sensor 14, the rotor 12 has completed one revolution. "Successive" refers to two successive passes of the same blade 13. Passages of the same blade 13 are separated by passes of the other blades 13 of the rotor 12. To identify when the same blade 13 passes a second time in front of the position sensor 14, the monitoring device 16 has knowledge, that is, it stores or has access to information, of the number of blades 13 of the rotor 12 of the monitored turbomachine 1. For example, in the example shown in Figure 1, the monitoring device 16 knows the number of blades 13 of the rotor 12 of the turbomachine 1, and knows that a first blade 13 passes a second time after the detected passage of two other blades 13 in front of the position sensor 14, because the turbomachine 1 in Figure 1 comprises three blades.

[0051] The signal obtained x fc (t) is therefore defined on one revolution of the rotor 12:

[0053] In a second step 22, the deformation signal is resampled in the angular domain for the current rotor revolution 12, in the same way as in step 12, i.e., with the same angular resolution L. The time-domain deformation signal associated with the current revolution / + k of rotor 12, i.e., captured during the kth revolution of rotor 12, is resampled in the angular domain to obtain the vector x k (s) for s = 1 ... P * L.

[0054] In a third step 23, the mechanical component related to the rotation of the rotor blades 13 of the rotor 12 relative to the stator 11 is extracted. This allows for the later detection of contact between a blade 13 and the stator 11 by analyzing this mechanical component. This extraction is performed by filtering, as described below.

[0055] The invention proposes the use of a linear, causal time-invariant filter with a finite impulse response. At each given angular position in the current rotation of the rotor 12, the sample is denoised, thus removing its second component due to elements other than the blades 13 of the rotor 12, by applying a weighted sum to all the samples in the current rotation of the rotor 12.

[0056] For this purpose, the h filter is proposed. T = [h ±1 ... , h(i +1 P * L ] of length (7 + 1) * P * L such that

[0058] the coefficients of a low-pass filter.

[0059] In a preferred embodiment, the coefficients of the low-pass filter are the coefficients of a mean filter with II e number of rotor revolutions 12 taken into account to calculate the average, including the current revolution and the previous / - 1 revolutions.

[0060] The mechanical component of rotor 12 is therefore estimated by summing the samples whose values ​​are weighted by the filter coefficients and taking into account the mechanical component values ​​acquired in the previous / - 1 revolutions of rotor 12: l ...P * L

[0062] The filtering process includes calculating, for each sample of the resampled distortion signal x k , of a sum weighted by a low-pass filter coefficient, the weighted sum comprising, at each instant, therefore at each angular step, of the current rotation of the rotor 12, the sum of the value of the sample of the resampled deformation signal x kand the value of each sample from a plurality of samples of the same angular position of the mechanical component related to the rotating blades 13 relative to the stator 11 of the resampled strain signal acquired during a predefined integer number of previous revolutions of the rotor 12, each sample of the same angular position being acquired at a different revolution of the rotor 12. A "sample of the same angular position as a sample obtained at the current revolution" is understood to mean a sample acquired during a revolution of the rotor 12 preceding the current revolution of the rotor 12, and having a number of angular steps relative to the first sample of the revolution considered equal to the number of angular steps of the sample obtained at the current revolution relative to the first sample of the current revolution. For example, a sample of x k at the kth turn, called the current turn, obtained at angular step number 7, is of "same angular position" as a sample of x^™ 9preceding the current turn obtained at angular step number 7. This is made possible by the fact that the filter h T includes nuisance coefficients, i.e. with a value of 0 at positions other than angular position number 7, and a low-pass filter coefficient at angular position number 7.

[0063] Each sample in the current iteration thus has a filter coefficient h applied to it. T and is summed with the other samples of the same angular position obtained in the previous 7 - 1 rounds, to which the same filter coefficient h T is also applied. This allows for the creation of a moving average that acts between rotor revolutions 12, enabling the sliding and therefore real-time or near-real-time nature of the invention. Thus, the noise component is attenuated or eliminated at each revolution of the rotor 12, to highlight the mechanical component related to the rotation of the rotor blades 13 relative to the stator 11.

[0064] The signal m k (n) obtained includes the first mechanical component of the deformation signal, that is to say the mechanical component related to the rotation of the blades 13 of the rotor 12 relative to the stator 11, for the current revolution of rotor 12, that is to say for the kth revolution in the present example.

[0065] The monitoring method according to the invention then comprises a calculation step 24 based on the first filtered resampled signal m k (n), of at least one touch indicator for each blade 13 of the rotor 12 for the current rotor revolution.

[0066] The first component m k (n) allows the extraction of indicators for the P auges 13 of the rotor 12 of the turbomachine 1. The invention uses an indicator comprising the ratio of the energy of the first mechanical component to the k-th revolution of the rotor 12 m k (n) on the first raw signal x k(n) of the same revolution, calculated at different ranges of the revolution, reflecting the position of the blades 13 of the rotor 12. The indicator A is therefore calculated according to the following formula for the indicator of blade p at iteration k:

[0068] This indicator returns a value close to zero when the signal from the p-th blade 13 is absent, and rises with the energy of this first mechanical component, which is related to the intensity of the friction.

[0069] At step 25, the signal (s) is updated by concatenating the vector x k (_s~) for s = 1 ... P * L with the signal x^°_ n 1 9 s') including the mechanical component of the deformation signal for the previous 7 - 1 rotor revolutions 12 to form:

[0070]

[0071] With x k Lon9 the concatenated signal, x^ 9the concatenation of the first mechanical component of each first signal acquired at the previous rotor revolutions 12, x k the first resampled signal acquired during the current rotor 12 revolution, / the number of rotor 12 revolutions on which the weighted sum is calculated, P the number of rotor 12 blades, L the angular resolution.

[0072] At step 26 of process 2 according to the invention, a touch is detected by comparing the indicator of each blade 13 of the rotor 12 with a threshold. The threshold is, for example, a threshold equal to or substantially equal to the mean of the indicator plus three times the standard deviation of the indicator, i.e., p + 30, calculated during so-called "reference" operation (without touch) of the turbomachine. This makes it possible to determine, as shown in Figure 5, a touch start time D and a touch end time F, corresponding to the times at which the indicator has a value greater than p + 3 and a value less than p + 3 respectively.

[0073] The touch and touch time information obtained by the method according to the invention can then be used to modify the blade tip clearance on the next rotor turn directly, making blade tip clearance monitoring active, unlike the prior art, which could not modify the blade tip clearance because the touch information was not calculated during operation and was not calculated in real time or near real time.

[0074] Aircraft high-pressure turbines feature an active blade tip clearance control system to improve fuel efficiency by dynamically controlling the blade tip clearance. A common strategy in such a system involves a valve that mixes hot and cold air from the turbomachine 1 compressor outlet and the turbomachine 1 bypass duct, respectively, at a desired temperature. This air is routed to flow through tubes surrounding the stator 11 at each turbine stage. This air expands or contracts the turbine casing to achieve the desired blade tip clearance. It should be noted that the aforementioned valve is automatically adjusted by a computer called a FADEC (Full Automatic Digital Engine Control) according to the position of the thrust lever of the aircraft carrying the turbomachine using the estimated blade tip clearance. The "FADEC" computer can implement method 2 according to the invention, and is then the monitoring device 16 described previously.

Claims

CLAIMS

1. Method (2) for monitoring a turbomachine (1) comprising a stator (11), a rotor (12) comprising blades (13), a strain gauge (15) fixed to the stator (11) and a position sensor (14) of the blades (13) of the rotor (12), the strain gauge (15) being configured to capture a first signal comprising a first component representative of deformations of the stator (11) caused by a rotation of the blades (13) of the rotor (12) relative to the stator (11), and a second component representative of deformations of the stator (11) caused by elements distinct from the blades (13) of the rotor (12), the position sensor (14) of the blades (13) of the rotor (12) being configured to capture a second signal representative of the position of each blade (13) of the rotor (12), the method (2) comprising, at each current revolution of the rotor (12), a rotation of the rotor (12) being detected from the second signal: - acquisition (21) of the first signal and the second signal, - angular resampling (22) of the first signal from the second signal, to obtain a first resampled signal according to a predefined angular resolution, - extraction (23) of the first component of the first resampled signal by filtering the first resampled signal, the filtering comprising the calculation, for each sample of the first resampled signal, of a sum weighted by a low-pass filter coefficient, the weighted sum comprising the sum of the value of the sample of the first resampled signal of the current revolution of the rotor (12) and the value of each sample of a plurality of samples of the same angular position of the first component of the first resampled signal extracted during a predefined integer number of previous revolutions of the rotor (12), - calculation (24), from the first filtered resampled signal, of at least one touch indicator for each blade (13) of the rotor (12) for the current rotor revolution (12).

2. Method (2) according to claim 1 further comprising a step of detecting (26), in at least one of the calculated touch indicators (24), a touch between the stator (11) and the blade (13) corresponding to the indicator.

3. Method (2) according to any one of the preceding claims, according to which a revolution of the rotor (12) is detected as two successive passages of the same blade (13) of the rotor (12), the passages of the same blade (13) of the rotor (12) being extracted from the second signal based on predefined information on the number of blades (13) of the rotor (12) of the turbomachine (1).

4. A method (2) according to any preceding claim wherein extracting (23) the first component of the first resampled signal by filtering the first resampled signal comprises applying the filter h T = [h ±1 ..., h I+1 P ^ L ] of length (Z + 1) * P * L defined according to with I a number of rotor revolutions (12) over which the weighted sum is calculated, P the number of blades (13) of the rotor (12), L the angular resolution, and the coefficients of the low-pass filter.

5. Method (2) according to the preceding claim according to which the calculation of the weighted sum is carried out according to the formula: With m kthe first extracted component (23) of the first resampled signal, I the number of rotor revolutions (12) over which the weighted sum is calculated, P the number of blades (13) of the rotor (12), L the angular resolution, x k the first resampled signal acquired during the current rotor revolution (12), a concatenation of the first mechanical component of each first signal of each previous revolution of the rotor (12) and h t the filter coefficients.

6. A method (2) according to any preceding claim wherein the low-pass filter coefficient is a filter coefficient average calculated according to the formula h m = with / the number of rotor revolutions (12) over which the weighted sum is calculated.

7. Method (2) according to any one of the preceding claims according to which the calculation (203), from the first resampled signal filtered, of at least one touch indicator for each blade (13) of the rotor (12) for the revolution of the rotor (12) is made according to the formula: With p a blade number among the blades (13) of the rotor (12), k the current rotor revolution (12), the indicator of the p-th blade (13) of the rotor (12) at the k-th rotor revolution (12), L the angular resolution, x k the first resampled signal acquired during the current rotor revolution (12), m k the first extracted component (23) of the first resampled signal.

8. Method (2) according to any one of the preceding claims further comprising a step of concatenating (26) the first signal acquired at the current rotor revolution (12) with a set of the first components of the first signals acquired at the previous rotor revolutions (12) to form: With x k Lon9 the concatenated signal, x^™ 9 the concatenation of the first mechanical component of each first signal acquired at the previous rotor revolutions (12), x k the first resampled signal acquired during the current rotor revolution (12), I the number of rotor revolutions (12) over which the weighted sum is calculated, P the number of blades (13) of the rotor (12), L the angular resolution.

9. Method (2) according to any one of the preceding claims according to which a touch is detected when a value of the touch indicator exceeds the value of p + 3o, with p an average of the touch indicators and o a standard deviation of the touch indicators calculated during the reference operation of the turbomachine.

10. A system for monitoring a turbomachine (1) comprising a stator (11) and a rotor (12) comprising blades (13), the monitoring system comprising a turbomachine monitoring device (16), a strain gauge (15) attached to the stator (11) of the turbomachine (1) and a position sensor (14) of the blades (13) of the rotor (12) of the turbomachine (1), the strain gauge (15) being configured to capture a first signal comprising a first component representative of deformations of the stator (11) caused by a rotation of the rotor (12) relative to the stator (11), and a second component representative of deformations of the stator (11) caused by distinct elements of the blades (13) of the rotor (12), the position sensor (14) of the blades (13) of the rotor (12) being configured to capture a second signal representative of the position of each blade (13) of the rotor (12), the device (16) comprising: - an acquisition input of the first signal (16-a), - an acquisition input of the second signal (16-b), and, - a processor configured to implement the monitoring method (2) according to any one of the preceding claims.

11. Aircraft comprising: - a turbomachine (1) comprising a stator (11) and a rotor (12) configured to be driven in rotation relative to the stator (11), and - a monitoring system (30) according to claim 10, wherein the strain gauge (15) is attached to the stator (11).

12. Computer program product comprising program code instructions for executing the steps of the method (2) for monitoring a turbomachine (1) according to one of claims 1 to 9, when this program product is executed by at least one data processing unit.