Method, apparatus, system, aircraft, and computer program product for monitoring a turbine

Abstract

Method, device, system, aircraft and computer program product for monitoring a turbine, the turbine comprising a stator and a rotor. The method comprises the step of obtaining an input signal, the input signal representing a deformation of the stator or the rotor of the turbine. The input signal is obtained by a strain gauge attached to the stator or the rotor. The input signal comprises: a first part, the first part representing a deformation of the stator or a deformation of the rotor caused by the rotation of the rotor blades relative to the stator; and a second part, the second part representing a deformation caused by different elements of the rotor blades. The method comprises the step of resampling the input signal to obtain a resampled input signal, the resampled input signal comprising a predetermined integer number of samples in each revolution of the turbine rotor. The method further comprises the step of processing the resampled input signal, the step comprising: filtering the resampled input signal such that a filtered input signal with the second part attenuated can be obtained; and separating the filtered input signal into a plurality of third parts, each third part representing the contribution to the deformation caused by a respective associated blade. The method further comprises the step of detecting a contact between the blade associated with the third part and the stator in the third part.

Description

Technical Field

[0001] The present disclosure relates to the general field of monitoring of turbines. More particularly, the present disclosure relates to a monitoring method capable of detecting friction between a stator and a rotor of a turbine. The present disclosure also relates to a device and a system for implementing such a method. These frictions between the stator and the rotor of the turbine are considered abnormal phenomena. Background Art

[0002] A turbine is a machine that enables the conversion of the kinetic energy of a fluid into mechanical energy (and conversely, the conversion of mechanical energy into the kinetic energy of a fluid) by a rotating component called a rotor. The stationary part of the turbine is called a stator.

[0003] These turbines have different designs according to their functions: turbines, pumps, compressors, turbocompressors, turbojets, etc. However, most of these turbines have a common architecture, including a rotor (a rotating component mounted on a shaft) and a stator (a fixed component connected to the structure of an aircraft) through bearings. The rotor includes a plurality of blades, and the plurality of blades are designed to accelerate the flow of air through the turbine. In order to build a relative rotational movement between the rotor and the stator, the blades must be spaced apart from the stator, and this spacing is generally referred to as the "blade tip clearance" or JSA. In an aircraft, the blade tip clearance is an important design and functional parameter of the turbine. The minimization of this clearance avoids excessive air flow bypassing the rotating blade row, thereby increasing the energy output of the turbine. However, due to thermal expansion and mechanical phenomena depending on the operating cycle of the turbine, this clearance undergoes fluctuations.

[0004] In some cases, these fluctuations of the clearance occur asymmetrically, and the clearance between some blades of the stator and the rotor becomes zero. The zero clearance results in friction between the rotating component and the fixed component of the turbine. This may lead to excessive wear of the wear-resistant surface and a loss of efficiency due to an increase in the leakage flow at the blade tip level.

[0005] For these reasons, it is desirable to be able to detect and monitor the instant at which the friction occurs. The determination of the time and amplitude of the friction enables the monitoring and control of these phenomena in the case of preventive monitoring. The amplitude of the proposed indicator can be indirectly used to evaluate the wear of the contact interface after friction in the case of preventive maintenance. Summary of the Invention

[0006] In this context, the present invention particularly solves the technical problem of monitoring of turbines by providing a method, a device, a system, an aircraft, and a computer program product capable of performing such monitoring.

[0007] Accordingly, according to the first embodiment, there is provided a method for monitoring a turbine, the turbine including a stator and a rotor. The method includes: a step of acquiring an input signal, the input signal representing a deformation of the stator or a deformation of the rotor of the turbine. The input signal has been acquired by a strain gauge attached to the stator or the rotor. The input signal includes: a first part, the first part representing a deformation of the stator or a deformation of the rotor caused by the rotation of the blades of the rotor relative to the stator; and a second part, the second part representing a deformation caused by an element separated from the blades of the rotor. The method includes a step of resampling the input signal to obtain a resampled input signal, the resampled input signal including a predetermined integer number of samples in each revolution of the rotor of the turbine. The method further includes a step of processing the resampled input signal, the step including: filtering the resampled input signal to obtain a filtered input signal, wherein the second part is attenuated, and separating the filtered input signal into a plurality of third parts, each third part representing a contribution to the deformation caused by a respective associated blade. The method further includes a step of detecting friction between the blade associated with the third part and the stator in the third part.

[0008] The method provides the following advantages:

[0009] - The use of the measurement of deformation and the processing of the associated signals enables good friction detection accuracy to be obtained. This is because the method particularly enables the effects of aerodynamic noise and the effects of interference that may come from different mechanical sources and / or power sources to be limited.

[0010] - In addition, the strain gauge is not very sensitive to the physical stresses borne by the turbine. In particular, the strain gauge has excellent temperature tolerance and is more tolerant than other types of sensors used to determine physical quantities related to the turbine. For example, in particular, the strain gauge is more tolerant than a capacitance sensor.

[0011] The step of resampling the input signal provides the advantage of enabling the signal to be synchronized with the rotation angle of the shaft of the turbine. In particular, the concept of periodicity in mechanical signals is inherently related to the rotation angle, and thus analysis in the angular domain enables the cancellation of speed fluctuations (which may cause slight variations in the duration of the period). This step also provides the advantage of obtaining a resampled input signal that has an integer and fixed number of samples in each revolution of the turbine.

[0012] During the detection of friction, the filtering step performed in the step of processing the signal provides the advantage of limiting the effects of deformation phenomena that are not caused by the rotation of the blades of the rotor relative to the stator.

[0013] The step of detecting the friction between the blades associated with the third part and the stator in the third part enables the detection of which blade causes the friction.

[0014] The monitoring method can be implemented in the following manner.

[0015] In an embodiment, the filtering step in the step of processing the signal includes the following steps: the step of determining the values of a plurality of parameters of a model representing the resampled input signal, and the step of using the model and the values of the plurality of parameters to determine the modeled input signal. Further, the separating step in the step of processing the signal separates the modeled input signal into a plurality of third parts associated with the blades.

[0016] In an embodiment, the step of determining the values of the plurality of parameters determines the values of the plurality of parameters as those values that minimize the deviation between the resampled input signal and the input signal modeled using the representative model and the values.

[0017] In an embodiment, the step of determining the values of the plurality of parameters of the model uses the following equation:

[0018]

[0019] where:

[0020]

[0021] is a column vector of size

[0022] P

[0023] that includes the values of the

[0024] P

[0025] parameters of the representative model,

[0026] Φ is a matrix of size

[0027] Q*P

[0028] where the element in row

[0029] q

[0030] and column

[0031] p

[0032] has the value

[0033] p q-1 ,

[0034] x

[0035] is of size

[0036] Q

[0037] The column vector, which includes the number of revolutions of the rotor

[0038] Q

[0039] Samples of the input signal resampled during the period of

[0040] This embodiment enables: a step of determining the values of the parameters of the model with lower computational complexity compared to using other methods.

[0041] In an embodiment, the step of determining the modeling input signal uses the following equation:

[0042]

[0043] Where:

[0044]

[0045] Is a column vector of size

[0046] P

[0047] The column vector, which includes the corresponding values of the

[0048] P

[0049] P parameters of the representative model,

[0050] Φ

[0051] Is a matrix of size

[0052] Q*P

[0053] Wherein, in the row

[0054] q

[0055] And column

[0056] p

[0057] The element in has the value

[0058] p q-1 ,

[0059]

[0060] Is a column vector of size

[0061] Q

[0062] The column vector, which includes the samples of the modeling input signal for the number of revolutions of the rotor

[0063] Q

[0064] Of the period.

[0065] This embodiment enables having a step of determining a modeling input signal with a lower computational complexity compared to using other methods.

[0066] In an embodiment, the separating step includes windowing the modeling input signal using different time-domain windows, for example, to obtain a plurality of third parts.

[0067] In an embodiment, the separating step uses the following equation:

[0068]

[0069] where:

[0070]

[0071] is a sample of the third part associated with the blade

[0072] r

[0073] associated third part

[0074] n,

[0075]

[0076] is a sample of the modeling signal

[0077] n,

[0078]

[0079] is the windowed signal associated with the blade

[0080] r

[0081] associated, where

[0082]

[0083] Q

[0084] is the number of revolutions of the rotor during the total duration of the modeling input signal,

[0085] R

[0086] is the total number of blades of the rotor.

[0087] In an embodiment, the step of detecting friction between the blade associated with the third part and the stator in the third part includes: determining the energy of the third part in at least one first revolution of the rotor. The detecting step further includes: comparing the determined energy with a given threshold and detecting friction or no friction based on the comparison.

[0088] In an embodiment, the step of determining the energy of the third part uses the following equation:

[0089]

[0090] wherein:

[0091] I r [q]

[0092] is the value of the energy in the revolution

[0093] r

[0094] associated with the third part in the rotor,

[0095] q

[0096] and

[0097] N

[0098] is the number of samples of the modeling input signal in the revolution

[0099] q

[0100] and

[0101]

[0102] is the sample

[0103] r

[0104] associated with the third part

[0105] n.

[0106] In an embodiment, the given threshold is calculated according to the following equation:

[0107]

[0108]

[0109]

[0110] wherein:

[0111] λ r

[0112] is the given threshold associated with the blade

[0113] r

[0114] and

[0115] Q ref

[0116] is the given number of revolutions of the rotor

[0117] I r [q]

[0118] is the value of the energy of the third part associated with the blade

[0119] r

[0120] during revolution.

[0121] q

[0122] of the revolution.

[0123] This embodiment enables an easy setting of a threshold enabling the detection of friction. In particular, this is based on the energy of the samples of the third part obtained by the monitoring device during a reference period. Thus, the method implemented in the monitoring device can be used to determine the detection threshold without the need for a special method to determine the detection threshold. Furthermore, this embodiment enables the direct determination of the detection threshold based on the turbine.

[0124] According to an embodiment, there is also provided a device for monitoring a turbine, the turbine including a stator and a rotor. The device includes an input section for receiving an input signal representing a deformation of the stator or a deformation of the rotor. The input signal has been obtained by a deformation gauge attached to the stator or the rotor. The input signal includes: a first part representing a deformation of the stator or a deformation of the rotor caused by the rotation of the blades of the rotor relative to the stator; and a second part representing a deformation of the stator or a deformation of the rotor caused by an element separated from the blades of the rotor. The device includes a data processing unit configured to perform the following steps: a step of obtaining the input signal obtained by the deformation gauge attached to the stator or the rotor; and a step of resampling the input signal to obtain a resampled input signal, the resampled input signal including a predetermined integer number of samples in each revolution of the rotor of the turbine; and a step of processing the resampled input signal, the step including: filtering the resampled input signal to obtain a filtered input signal, wherein the second part is attenuated, and separating the filtered input signal into a plurality of third parts, each third part representing the contribution to the deformation caused by the respective associated blade. The method further includes a step of detecting friction between the blade associated with the third part and the stator in the third part.

[0125] The device provides the following advantages:

[0126] - The use of deformation measurement and the processing of the associated signals enables good friction detection accuracy. This is because the method enables, in particular, the limitation of the influence of aerodynamic noise and the influence of possible interferences from different mechanical sources and / or power sources.

[0127] - Additionally, the strain gauge is not very sensitive to the physical stresses to which the turbine is subjected. In particular, the strain gauge has excellent temperature tolerance and is more tolerant than other types of sensors used to determine physical quantities related to the turbine. For example, in particular, the strain gauge is more tolerant than a capacitance sensor.

[0128] According to an embodiment, there is also provided a monitoring system including the monitoring device as described previously and a strain gauge adapted to transmit an input signal representing a deformation of the stator or a deformation of the rotor of the turbine. As part of the system, the strain gauge is connected to the input part of the monitoring device.

[0129] According to an embodiment, there is also provided an aircraft including a turbine. The turbine includes: a stator and a rotor, the rotor being adapted to be rotationally driven relative to the stator; and the monitoring system as described above, wherein the strain gauge is attached to the stator or the rotor.

[0130] According to an embodiment, there is also provided a computer program product including program code instructions for performing the steps of the method for monitoring a turbine as shown previously when the program product is executed by at least one data processing unit. Description of the Drawings

[0131] Other features and advantages of the present invention will become more apparent from the following description, which is illustrative only and non - limiting and must be read with reference to the drawings, in which:

[0132] - Figure 1 An aircraft is shown, which includes a turbine and a device for monitoring the turbine.

[0133] - Figure 2 Steps 201 and 202 of the method for monitoring a turbine are shown.

[0134] - Figure 3 The result of the step of filtering the signal received from the strain gauge is shown.

[0135] - Figure 4 The result of the step of separating the filtered signal is shown.

[0136] - Figure 5 The result of the step of detecting friction using the separated signal is shown.

[0137] - Figure 6 The step of separating the filtered input signal is shown. Detailed Description

[0138] Figure 1Aircraft 101 is schematically shown. Aircraft 101 includes a turbine 102.

[0139] In Figure 1 the example shown, turbine 102 includes a nacelle, a fan, a low-pressure section, a high-pressure section, and a combustion chamber. The low-pressure section includes a low-pressure compressor, a low-pressure turbine, and a low-pressure shaft that connects the low-pressure compressor to the low-pressure turbine. The high-pressure section includes a high-pressure compressor, a high-pressure turbine, and a high-pressure shaft that connects the high-pressure compressor to the high-pressure turbine.

[0140] When the turbine is operating, the high-pressure turbine rotationally drives the high-pressure compressor via the high-pressure shaft. The low-pressure turbine rotationally drives the low-pressure compressor and the fan via the low-pressure shaft. The fan generates a primary air flow and a secondary air flow (or bypass flow). The primary air flow successively passes through the low-pressure compressor, the high-pressure compressor, the combustion chamber, the high-pressure turbine, and the low-pressure turbine.

[0141] The high-pressure turbine includes a stator 103 and a rotor 104, and the rotor is adapted to be rotationally driven relative to the stator 103. The stator 103 is fixedly mounted on a structural part of the aircraft, for example, on the casing of the high-pressure turbine. The rotor 104 includes a plurality of blades 105. The turbojet engine 102 further includes a strain gauge 106. In Figure 1 it, the gauge is attached to the stator of the high-pressure turbine, or the gauge can also be attached to the rotor of the high-pressure turbine.

[0142] The aircraft further includes a monitoring device 107, and the detection device enables the detection of friction between the stator 103 and the rotor 104. The device 107 includes an input part 107-a that is used to receive an input signal representing the deformation of the stator 103 from the strain gauge 106. The input part 107-a includes an acquisition line that, among other things, also provides digitization of the input signal. The sampling frequency of the input signal is configurable. Advantageously, this frequency is about several tens of kHz, and this frequency depends on the turbine. The device also includes a data processing unit 107-b. The data processing unit 107-b is configured to execute the steps of a monitoring method, and the steps of the monitoring method enable the detection of friction between the stator 103 and the rotor 104 of the turbine 102. Generally, the data processing unit 107-b includes at least one processor for executing a computer program. The computer program includes program code instructions that are configured to implement a method for monitoring the turbine 102 when these instructions are executed by the processor of the data processing unit 107-b. In addition, the device also includes a memory 107-c that is used to store data, particularly for storing the input signal received from the strain gauge 106.

[0143] In an embodiment, the sensor 108 is arranged on the turbine 102, and the sensor gives an item of information representing the blade passage time. This sensor 108 is typically arranged on the stator. The sensor 108 is used to indicate the time of blade passage, and this sensor is also referred to as a blade tip sensor. The sensor 108 can be, for example, an optical probe that optically measures the passage of the blade.

[0144] The monitoring device 107 enables the following to be obtained at the output:

[0145] - an indication of the friction of the blade,

[0146] - a detection threshold associated with these frictions, and

[0147] - the time of these frictions.

[0148] These items of information can be displayed on a screen or stored in the memory 107-c to send these items of information to another device, such as a fixed monitoring station, if applicable.

[0149] Referring to Figure 2 , in Figure 1 the monitoring device 107 shown is capable of detecting the friction between the stator 103 and the rotor 104 of the turbine 102, and this monitoring device operates as follows.

[0150] In a first step 201, the input section 107-a of the device 107 acquires an input signal. This input signal represents the deformation of the stator 103 or the deformation of the rotor 104 of the turbine 102. The input signal has been generated by a strain gauge 106 attached to the stator 103 or the rotor 104. The input signal includes: a first part and a second part, the first part representing the deformation of the stator 103 or the deformation of the rotor 104 caused by the rotation of the blades 105 of the rotor 104 relative to the stator 103, and the second part representing the deformation caused by an element separated from the blades 105 of the rotor 104.

[0151] In an embodiment, the signal acquired by the strain gauge 106 is sampled at a sampling frequency

[0152] F s

[0153] (The sampling frequency is, for example, approximately 100 kHz).

[0154] Then, the processing unit 107-b performs a step 202 of resampling the input signal to obtain a resampled input signal, and the resampled input signal includes a predetermined integer number of samples in each revolution of the rotor of the turbine. This resampling can be described as angular sampling.

[0155] Resampling of the signal received from the deformable gauge 106 uses the blade position signal generated by the sensor 108. The sensor 108 generates a position signal that represents the time at which the blade 105 passes in front of the sensor 108. An indication of the number of blades 105 of the rotor 104 can also be used. The signal obtained after resampling is an "angle" signal (i.e., a signal sampled at fixed angular intervals). This resampled signal is obtained by interpolating the signal received from the deformable gauge 106. The resampled signal includes samples spaced apart by an angle

[0156] Δθ

[0157] which is the angular sampling period. In addition, N

[0158] samples of the resampled signal are precisely associated with each revolution of the rotor.

[0159] Then, the device performs step 203 of processing the resampled input signal to detect friction.

[0160] Step 203 of processing the signal includes filtering 204 the resampled input signal. This filtering enables the acquisition of a filtered signal in which a second part is attenuated.

[0161] This second part can be considered noise. Thus, the input signal has the following form:

[0162]

[0163] where:

[0164] L

[0165] represents the total number of samples,

[0166] n

[0167] represents a given sample,

[0168] x[n]

[0169] represents the sample of the input signal

[0170] n,

[0171] d[n]

[0172] represents the first part of the sample of the input signal

[0173] n

[0174] of,

[0175] w[n]

[0176] represents the sample of the input signal

[0177] n

[0178] The second part of

[0179] One way to perform the filtering 204 is, for example, to perform step 205 and then step 206. Step 205 is to determine the values of a plurality of parameters of a model that represents the resampled input signal, and step 206 is to determine the input signal modeled using the model and the values of the plurality of parameters. This way of performing the filtering 204 is illustrative and non - limiting, and the filtering 204 can be performed in other ways. The step 203 of processing the signal may also include separating 207 the resampled input signal into a plurality of third parts. Each third part represents the contribution to the deformation caused by the respective associated blade. This separation 207 is performed after the filtering 204 of the signal. If the filtering is performed by steps 205 and 206, the separation 207 is performed on the modeled input signal.

[0180] Next, the data processing unit 107 - b performs step 208, which is to detect the friction between the blade 105 associated with the third part and the stator 103 in the third part.

[0181] One way to perform the detection 208 is, for example, to perform step 209 and then step 210. Step 209 is to determine the energy of the third part during at least one first revolution of the rotor, and step 210 is to compare the determined energy with a given threshold. Friction is detected or not detected based on the comparison. This way of performing the detection 208 is illustrative and non - limiting, and the detection 208 can be performed in other ways.

[0182] In the present disclosure, two symbols will be used to represent samples:

[0183] n

[0184] And

[0185]

[0186] n

[0187] Is used to represent a sample in an absolute sense,

[0188] n

[0189] Is the i - th sample, and the start of the resampled input signal is taken as the initial time.

[0190]

[0191] Is used to represent a sample in a relative sense related to one revolution among a plurality of revolutions. Thus,

[0192]

[0193] Denote the i-th sample

[0194]

[0195] And take the start of the rotor turnover when the sample is recorded as the initial time.

[0196] Using this symbol, there is only a single sample

[0197] n

[0198] But there are multiple samples

[0199]

[0200] Sample

[0201]

[0202] Each sample in is respectively associated with one of the

[0203] Q

[0204] turnovers of the rotor of the turbine.

[0205] Using the equation

[0206]

[0207] To perform the change from

[0208] n

[0209] To

[0210]

[0211] Where

[0212] N

[0213] Represents the number of samples per turnover, where

[0214]

[0215] Represents

[0216] a

[0217] Divided by

[0218] b

[0219] The remainder of.

[0220] In the same way, turnover

[0221] q

[0222] sample

[0223]

[0224] absolute position

[0225] n

[0226] given by the equation

[0227] n = N*(q - 1)+n

[0228] given

[0229] The data processing unit 107 - b can perform the determination step 205 in different ways. By way of illustrative and non - limiting example, this step can be performed by considering the parameter values to be those that minimize the deviation between the input signal generated by the strain gauge 106 and the modeled input signal using the representative model and the said values.

[0230] The input signal generated by the strain gauge 106 can be modeled as follows:

[0231]

[0232] where:

[0233] q

[0234] represents one revolution in the revolution of the turbine 102,

[0235] Q

[0236] represents the number of revolutions in which the input signal is located,

[0237]

[0238] represents one sample out of a plurality of samples representing one revolution,

[0239] N

[0240] represents the number of samples in each revolution,

[0241]

[0242] represents the i - th sample of the modeled signal of the turbojet engine associated with the revolution

[0243] q

[0244] associated

[0245]

[0246] P

[0247] is the number of parameters,

[0248]

[0249] representing the samples of the input signal of the representative model

[0250]

[0251] associated parameters

[0252] p.

[0253] The value of the order of the polynomial (and thus the value of the number of parameters)

[0254] P

[0255] must be high enough to account for variations in the amplitude of the input signal. Advantageously,

[0256] P

[0257] will have a value between 5 and 20.

[0258] Using this model of the input signal, the values of the parameters will be obtained by the following equation:

[0259]

[0260] where,

[0261]

[0262] is the sample of the input signal generated by the strain gauge 106 at the time

[0263] q

[0264] of revolution

[0265]

[0266]

[0267]

[0268] is a column vector of size

[0269] P

[0270] which includes the values of the

[0271]

[0272] parameters associated with the samples of the representative model

[0273] P

[0274] values of the

[0275] By way of non-limiting and purely illustrative examples, the values of the plurality of parameters of the model can be determined using the following equations:

[0276]

[0277] wherein,

[0278]

[0279] is a column vector of size

[0280] P

[0281] that includes the values of the P parameters of the representative model associated with the samples

[0282]

[0283] associated with

[0284] P

[0285] parameters of the model,

[0286] Φ

[0287] is a matrix of size

[0288] Q*P

[0289] wherein the element in row

[0290] q

[0291] and column

[0292] p

[0293] has the value

[0294] p q-1 ,

[0295]

[0296] is a column vector of size

[0297] Q

[0298] that includes samples of the input signals generated by the strain gauges during the number of revolutions

[0299] Q

[0300] of the rotor and, where applicable, samples of the resampled input signals, all the samples having the same relative position during the revolution of the rotor

[0301]

[0302] and each sample being respectively associated with the

[0303] Q

[0304] associated with one of the plurality of turns in rotation.

[0305] After determining the plurality of parameters, the monitoring device 107 determines the modeling input signal. The step 206 of determining the modeling input signal can be performed in different ways. By way of example but not limitation, the modeling input signal

[0306]

[0307] has the following form:

[0308]

[0309] where:

[0310]

[0311]

[0312] is a column vector of size

[0313] P

[0314] that includes the values of the P

[0315]

[0316] parameters associated with the samples of the representative model

[0317] P

[0318] parameters,

[0319] Φ

[0320] is a matrix of size

[0321] Q * P

[0322] wherein the element in row

[0323] q

[0324] and column

[0325] p

[0326] has the value

[0327] p q-1 ,

[0328]

[0329] is of size

[0330] Q

[0331] The column vectors, which include the number of revolutions of the modeling input signal with respect to the rotor

[0332] Q

[0333] Samples, all having the same relative position during the revolution of the rotor

[0334]

[0335] And each sample is respectively related to the

[0336] Q

[0337] associated with one of the Q revolutions of the rotor.

[0338] In this embodiment, the modeling input signal

[0339]

[0340] is the input signal (resampled and filtered if applicable) generated by the strain gauge 106, where the second part has been attenuated.

[0341] The processing unit 107-b is configured to separate the contributions of different blades 105. This separation step 207 is performed based on the resampled input signal, which is filtered to attenuate the second part. In the embodiment where the filtering step 204 is based on steps 205 and 206, the signal on which the separation is based is the modeling input signal

[0342]

[0343] This separation 207 can be performed by windowing the modeling input signal using different time-domain windows. This windowing enables obtaining a plurality of third parts.

[0344] This separation can be performed in different ways. By way of illustrative and non-limiting example, the samples of the third part

[0345]

[0346] will be obtained by multiplying the samples of the modeling input signal

[0347]

[0348] by the samples of a signal

[0349] F(n)

[0350] including a time window. Different types of windows can be used, such as Gaussian windows, Hanning windows, or rectangular windows. In the case of a rectangular window, this gives:

[0351]

[0352] wherein:

[0353]

[0354] is a sample of a third part associated with the blade

[0355] r

[0356] n,

[0357] which is a sample of the modeling signal

[0358]

[0359] n,

[0360] which is a windowed signal associated with the blade

[0361]

[0362] r

[0363] r

[0364] wherein,

[0365] Q

[0366]

[0367] is the number of revolutions completed by the rotor during the total duration of the modeling input signal,

[0368] and

[0369] R

[0370] is the total number of blades of the rotor.

[0371] To calibrate the start of the time window, the position signal from sensor 108 can be used, which gives an item of information representing the end of the time when the blade passes through. Using this sensor 108, the position of each blade can be known, and the windowed signal is applied to facilitate the time associated with the blade of interest.

[0372]

[0373] is obtained from the vector

[0374]

[0375] in the following manner. In the first step, the revolution of the rotor associated with the absolute sample

[0376] n

[0377] is determined

[0378] q

[0379] Know the number of samples in one revolution of the rotor

[0380] N

[0381] This revolution of the rotor

[0382] q

[0383] is

[0384] q = floor(n / N)+1

[0385] or

[0386] floor(x)

[0387] is the largest integer less than

[0388] x

[0389] of the maximum integer

[0390] In the second step, the relative position of the sample

[0391]

[0392] is determined using the following formula

[0393]

[0394] Therefore, knowing the relative position

[0395]

[0396] will be able to select the correct vector associated with the absolute sample

[0397] n

[0398] associated

[0399]

[0400] and in the vector

[0401]

[0402] by selecting the vector

[0403]

[0404] of the sample

[0405] q

[0406] to obtain the value

[0407]

[0408] In an embodiment, the step 501-a of determining the energy of the third part is performed by summing the squares of the samples of the third part over one revolution of the rotor. For each blade, this summation is performed independently. Thus, the energy is obtained by the following equation:

[0409]

[0410] where:

[0411] I r [q]

[0412] is the value of the energy of the third part associated with the blade

[0413] r

[0414] during the revolution

[0415] q

[0416] of the rotor,

[0417] N

[0418] is the number of samples of the resampled input signal associated with the revolution

[0419] q

[0420] of the rotor,

[0421]

[0422] is the sample

[0423] r

[0424] of the third part associated with the blade

[0425] n.

[0426] To calculate the detection threshold, the energy of the third part in a plurality of reference revolutions is used. The detection threshold is based on the average value of the energy of the third part in a plurality of reference revolutions (

[0427]

[0428] ) and the standard deviation of this energy (

[0429]

[0430] ). Thus, the threshold has the following form:

[0431]

[0432] In these different equations:

[0433] -λ r

[0434] is a given threshold associated with the blade

[0435] r

[0436] associated with the blade

[0437] -Q ref

[0438] is a given number of revolutions of the rotor

[0439] -I r [q]

[0440] is the value of the energy of the third part associated with the blade

[0441] r

[0442] associated with the blade during the revolution

[0443] q

[0444] in the revolution

[0445] It is possible to define a confidence interval associated with the detection threshold, and this confidence interval can be determined in the following manner:

[0446] By making the assumption that there is no friction and using a reference signal in which there is no friction to determine the empirical probability density of the indicator

[0447] Based on this empirical probability density, determine the confidence interval

[0448] Therefore, this embodiment enables the calculation of the threshold based on the average value and standard deviation of the indicator over a plurality of

[0449] Q ref

[0450] cycles that are reference cycles. These reference cycles are cycles during which there is no friction between the blades of the rotor and the stator

[0451] In the embodiment, the monitoring device 107 or the monitoring method is configured such that the user can specify:

[0452] - parameters of the turbine (such as the number of blades), and

[0453] - parameters for detecting friction (such as the order of the polynomial used to model the input signal, and thus the number of parameters of this model of the input signal, the type of separation window, or the confidence interval associated with the detection threshold).

[0454] Figures 3 to 5Shows the influence of different steps of a monitoring method capable of detecting friction on the signal from the deformable gauge 106. The signals used in these figures are real signals measured on a turbine including three blades.

[0455] Figure 3 The amplitude of the signal received from the deformable gauge 106 is shown on the top curve. This signal includes: a first part, which represents the deformation of the stator 103 or the rotor 104 caused by the rotation of the blade 105 of the rotor 104 relative to the stator 103; and a second part, which represents the deformation caused by an element separated from the blade of the rotor. The middle curve shows the signal filtered by the filtering step 204. In this filtered signal, the second part is greatly attenuated. Therefore, the middle curve mainly shows the first part of the signal received from the deformable gauge. This filtered signal corresponds to the contribution of all the blades to the first part. By using this filtered signal, the time of contact of the blades can be easily detected (based on revolution 4500 and up to revolution 5200). By comparison, these same times cannot be detected in the signal of the top curve because the first part is hidden by the noise generated by the second part. Finally, Figure 3 The bottom curve shows the second part of the signal received from the deformable gauge 106.

[0456] Figure 4 Shows the effect of separating the filtered signal into a plurality of third parts, each third part representing the contribution to the deformation caused by one of the three blades. The filtered signal is the signal obtained on the Figure 3 middle curve. Figure 4 The signal shows these three parts for the revolution from revolution 5000 to revolution 5010.

[0457] Figure 5 Shows the energy of the third parts associated with the three blades of the turbine. These three curves are shown respectively with three detection thresholds 501-a to 501-c. Therefore, between revolution 4400 and revolution 5200, the energy of the third parts is greater than the detection threshold, which corresponds to the detection of friction.

[0458] In the resampling step 202, based on the blade tip sensor 108 that measures the passage of the blades of the turbine, the signal received from the deformable gauge 106 is converted into an angular signal x(θ). Θ represents the rotational angle of the shaft of the turbine with P blades (full rotation of the shaft of the turbine).

[0459] The first blade will be centered at an angle θ = 0 + kΘ (k is an integer), the second blade will be centered at an angle θ = Θ / P + kΘ, the third blade will be centered at an angle θ = 2*Θ / P + kΘ, and the i-th blade will be centered at an angle θ = i*Θ / P + kΘ.

[0460] If the center of the window is aligned with the angular position of the blade, multiplying the filtered input signal (by polynomial synchronous averaging) by a periodic window (the periodic window repeats at each full rotation) enables obtaining the contribution of blade i. The general equation of the window is:

[0461]

[0462] For the first blade, the equation of the window used is as follows:

[0463]

[0464] For the second blade, the window used for the first blade is offset by an angle that separates the two blades (this angle is equal to the angle of rotation of the turbine divided by the number of blades, Θ / P). The equation of the window used is as follows:

[0465]

[0466] For the third blade, the window used for the first blade is offset by an angle that is twice the angle that separates the two blades. The equation of the window used is as follows:

[0467]

[0468] For the i-th blade, the window used for the first blade is offset by an angle that is i times the angle that separates the two blades. The equation of the window used is as follows:

[0469]

[0470] Figure 6 Shows separation step 207 for a turbine including three blades. In this figure, time is shown from left to right. The first revolution of the turbine is at 601, the second revolution is at 602, and the third revolution is at 603. At 604, the contributions of all the blades are shown. At 605, the contribution of the first blade is shown. This contribution is represented by multiplying the contributions of all the blades by the window F0(θ). At 606, the contribution of the second blade is shown. This contribution is represented by multiplying the contributions of all the blades by the window F1(θ). At 607, the contribution of the third blade is shown. This contribution is represented by multiplying the contributions of all the blades by the window F2(θ).

Claims

1. A method for monitoring a turbine, the turbine comprising a stator and a rotor and a strain gauge attached to the stator or the rotor, the method comprising: - obtaining an input signal by means of the strain gauge, wherein the input signal comprises: a first part representing the deformation of the stator or the rotor caused by the rotation of the blades of the rotor relative to the stator; and a second part representing the deformation of the stator or the rotor caused by an element separated from the blades of the rotor; - resampling the input signal to obtain a resampled input signal, the resampled input signal comprising a predetermined integer number of samples in each revolution of the rotor of the turbine; - processing the resampled input signal, the step comprising: filtering the resampled input signal to obtain a filtered input signal in which the second part is attenuated, and separating the filtered input signal into a plurality of third parts, each third part representing the contribution to the deformation caused by a respective associated blade; and - detecting friction between the blade associated with the third part and the stator in the third part.

2. The method according to claim 1, wherein, Filtering the resampled input signal comprises: determining values of a plurality of parameters of a model representing the resampled input signal; and using the model representing the resampled input signal and the determined values of the plurality of parameters to determine a modeled input signal.

3. The method according to claim 2, wherein, The determined values of the plurality of parameters are those that minimize the deviation between the resampled input signal and the input signal modeled using the model representing the resampled input signal and the determined values.

4. The method according to claim 2, wherein The values of the plurality of parameters of the model representing the resampled input signal are determined using the following equation where: is a column vector of size P, which includes the values of the multiple parameters of the model representing the resampled input signal, Φ is a matrix of size Q*P, where the element in row q and column p has the value p q-1 , x is a column vector of size Q, the column vector comprising samples of the resampled input signal during the number of revolutions Q of the rotor.

5. The method according to claim 2, wherein The modeled input signal is determined using the following equation where: is a column vector of size P that includes the values of the multiple parameters of the model representing the resampled input signal, Φ is a matrix of size Q*P, where the element in row q and column p has the value p q-1 , x is a column vector of size Q, the column vector comprising samples of the resampled input signal during the number of revolutions Q of the rotor.

6. The method according to claim 2, wherein Separating the filtered input signal comprises windowing the modeled input signal using different time domain windows.

7. The method according to claim 6, wherein, The filtered input signal is separated using the following equation where: is a sample n of a third part associated with blade r, is the sample n of the modeling input signal, is a windowed signal associated with blade r, where, Q is the number of revolutions of the rotor during the total duration of the modeled input signal, R is the total number of blades of the rotor.

8. The method according to claim 1, wherein Detecting friction between the blade associated with the third part and the stator in the third part comprises: determining the energy of the third part in at least one first revolution of the rotor; and comparing the determined energy with a given threshold.

9. The method according to claim 8, wherein The energy of the third part is determined using the following equation where: I r [q] is the value of the energy of the third part associated with the blade r in the revolution q of the rotor, N is the number of samples of the resampled input signal in revolution q, Sample n of the third part associated with blade r.

10. The method according to claim 8, wherein, The method further comprises calculating the given threshold using the following equation λ r = μ r ( ref ) + 3σ r ( ref ) where: λ r is the given threshold value, Q ref is the given number of revolutions of the rotor I r [q] is the value of the energy of said third part associated with blade r in revolution q.

11. A device for monitoring a turbine, the turbine comprising a stator and a rotor and a strain gauge attached to the stator or the rotor, the device comprising: - an input section for obtaining an input signal acquired by the strain gauge, the input signal comprising: a first part representing the deformation of the stator or the rotor caused by the rotation of the blades of the rotor relative to the stator; and a second part representing the deformation of the stator or the rotor caused by an element separated from the blades of the rotor; and - a data processing unit configured to perform the following steps: - a step of acquiring the input signal; and - a step of resampling the input signal to obtain a resampled input signal, the resampled input signal comprising a predetermined integer number of samples in each revolution of the rotor of the turbine; - a step of processing the resampled input signal, the step comprising: filtering the resampled input signal to obtain a filtered input signal in which the second part is attenuated, and separating the filtered input signal into a plurality of third parts, each third part representing the contribution to the deformation caused by a respective associated blade; and - a step of detecting friction between the blade associated with the third part and the stator in the third part.

12. A monitoring system, the monitoring system comprising: - the device according to claim 11, - a strain gauge adapted to transmit an input signal representing the deformation of the stator or the rotor of the turbine; wherein the strain gauge is connected to the input section of the device.

13. An aircraft, the aircraft comprising: - a turbine, the turbine comprising a stator and a rotor, the rotor being adapted to be rotationally driven relative to the stator, and - the monitoring system according to claim 12, wherein the strain gauge is attached to the stator or the rotor.

14. A non-transitory computer-readable medium having program code instructions stored thereon, which when executed by a processor cause the processor to implement the method according to claim 1.

Images