The invention relates to detecting damage to the rotor of an aircraft engine.
In civil aviation, regulations require aircraft engine vibration to be monitored. Such monitoring is generally performed by means of accelerometers.
The signal delivered by each accelerometer is initially processed in order to extract frequency components therefrom corresponding to the speeds of rotation of the rotors in the low pressure and high pressure stages. The amplitudes of these components are delivered in real time to the cockpit, and certain key values are stored for subsequent processing. In general, five to ten values are stored per flight, as measured under predefined stabilized conditions.
Unfortunately, it can happen that an aircraft engine rotor is damaged, for example by blades being torn off, without that event being detected because of the small amount of data gathered.
There thus exists at present a need for a system which enables damage to the rotor of an aircraft engine to be detected without error.
An object of the invention is to provide a method of detecting damage to the rotor of an aircraft engine in order to guarantee proper operation of the engine and safety of the aircraft.
Another object is to provide good diagnosis as soon as possible so as to have a maintenance team corresponding to the type of problem.
Yet another object is to provide preventative maintenance.
These objects are achieved by a method of detecting damage to the rotor of an aircraft engine including means for measuring vibration and speed in order to acquire data relating to rotor speed and to the amplitude and the phase of rotor vibrations during a determined flight, the method comprising the following steps:
The detection method of the invention may also comprise the following steps:
Similarly, the detection method of the invention may also comprise the following steps:
According to a first particular feature of the invention, the reference flight corresponds to the flight preceding said determined flight.
According to a second particular feature of the invention, the reference flight corresponds to a flight associated with a standard, reference engine.
Advantageously, the method of the invention comprises a step of updating the mean vectors of the reference flight from the data of said determined flight whenever the modulus or the maximum modulus of the vector difference does not exceed the predetermined threshold value.
The amplitude of said determined rotor speed range corresponds to a value of 1% to 10% of the nominal speed of the rotor.
The threshold value is determined depending on the location of the vibration detection means and correspond to a value lying in the range 2 thousandths of an inch (mils) to 5 mils.
Another object of the invention is to provide a system for detecting damage to the rotor of an aircraft engine that enables the above-defined method to be implemented.
This object is achieved by means of a system for detecting damage to the rotor of an aircraft engine provided with vibration measurement means and speed measurement means for acquiring data relating to the speed of the rotor and also to the amplitude and the phase of rotor vibration during a determined flight, the system comprising:
The detection system of the invention may also comprise:
Similarly, the detection system of the invention may also comprise:
The detection system of the invention includes at least one means for measuring vibration in a radial plane of the engine.
The invention also provides an aircraft engine comprising a compressor fitted with first rotary disks and a turbine fitted with second rotary disks, the engine implementing the above-defined detection system.
In a preferred embodiment, the detection system comprises first vibration measurement means at one of the first rotary disks and second vibration measurement means at the second rotary disks.
The invention also provides a computer program designed to implement the above-defined method when executed by a computer.
The invention will be better understood on reading the following description given by way of non-limiting indication, and with reference to the accompanying drawings, in which:
Each compressor comprises a rotary portion or “rotor” 5 and a stationary portion or “stator” 6 together with a casing 7. The rotor comprises a drum made up by assembling a plurality of disks together, the disks having moving blades 8 fixed thereto. The stator is constituted by a plurality of rows of stationary blades which can be fixed to the casing 7.
The turbine 4 comprises one or more stages in which each stage is made up of a set of stationary blades 9a and a set of moving blades 9b fixed on a disk.
In operation, the various blades are subjected to aerodynamic forces. In addition, the moving blades 8 and 9b are subjected to centrifugal force which is proportional to the square of the speed of rotation. Thus, the blades and their attachment points to the disks are dimensioned for the most severe operating conditions.
Since blades are somewhat strip-shaped, they can vibrate at a resonant frequency which depends on their shape, their dimensions, and the way in which they are fixed to the disk.
Such vibration can be excited by the wake that starts from the trailing edges of the blades, by mechanical forces stemming from the rotor in the event of the rotor being excessively out of balance, or by aerodynamic instabilities. Consequently, the resulting sustained vibrations can cause one or more blades to break or be lost.
Thus, certain kinds of damage to the rotary parts lead to a sudden change in the balance of the corresponding rotor, and can consequently lead to vibration.
In general, engine vibration is monitored by accelerometers or other vibration sensors. Each sensor may comprise two accelerometers at an angle to each other in a radial plane of the engine, and preferably at 90°. The sensors may be placed on the casing 7 over the turbine 4, over one of the compressors 1 and 2, or between the compressors. Naturally, it is preferable for a vibration sensor to be placed facing the moving disk(s) that are to be monitored.
In conventional manner, each vibration sensor or each measurement means delivers an electrical signal representative of the mechanical vibration of the compressor or of the turbine. After being amplified and filtered, the signal is converted into digital data by an analog-to-digital converter in order to be analyzed digitally.
In addition, a speed sensor 16 measures the speed of rotation of the rotor associated with the compressor, and another speed sensor (not shown) also exists which measures the speed of rotation of the rotor associated with the turbine.
As shown very diagrammatically in
In accordance with the invention,
The example of
Vibration v(t) is characterized by its amplitude A(t), i.e. by a maximum departure from an equilibrium position, and by its phase φ(t). Amplitude thus has the dimension of a length which in this case is expressed in micrometers (μm) or in mils, and phase can be expressed in radians or in degrees (°). In general, vibration is expressed as a complex variable having modulus A(t) and argument φ(t) in radians, and of the form v(t)=A(t)exp(iφ(t)).
Thus, at any given instant, a vibration is defined by an amplitude and by a phase or angle.
By way of example, the acquisition rate of data in
Thus, at the end of each flight, the central unit reads the data relating to the speed of rotation of the rotor (FIG. 2A), and to the amplitude and the phase of each vibration (
In step 20, the vibration of the rotor at a given instant t is expressed by a vibration vector V defined by the amplitude A(t) and the phase φ(t) at said instant t of said vibration.
In step 30 (see also FIG. 5), the vibration vectors coming from each vibration sensor are parameterized as a function of rotor speed. Thereafter, the spectrum of rotor speeds is subdivided into a plurality of speed “classes” or ranges, and consequently the vibration vectors are sorted in these ranges.
The subdivision need not necessarily be regular, and the amplitude of a speed range may correspond to a value of 1% to 10% of the nominal speed of the rotor. It is preferable to refine the subdivision at high speeds of the rotor since vibration, and consequently the risk of losing blades, is then higher. For example, for normalized rotor speeds of 80% to 110%, it may be appropriate to use subdivisions or ranges having an amplitude of 1%.
Furthermore, it is advantageous to enlarge the size of subdivision ranges at low speeds and even to eliminate the low end of the speed spectrum so as to avoid overloading data processing time and memory. For example, subdivisions can be made at 2% or more for normalized rotor speeds of less than 80%, and the portion below 20% can be eliminated.
In step 40, the coordinates of a mean vibration vector <V> are calculated for each speed range and for each sensor.
Similarly,
In step 50, the mean vector representing each speed range and each sensor is stored in memory.
Starting from step 50, the way these vectors vary is analyzed by steps 61 to 81 and/or by steps 62 to 82.
Thus, in step 61, for each speed range and for each sensor, the vector difference D is calculated between the mean vector <V1d> of a reference flight and the mean vector of the determined flight, naturally for the same speed range. It should be observed that the mean vectors <V1d> representing the various speed ranges of the reference flight have previously been stored in the memory 25 of the processor system 22 (see FIG. 1).
The reference flight may correspond to the flight preceding the determined flight. The reference flight may also correspond to a flight associated with a standard or reference engine, e.g. a test engine.
When the vector difference as calculated in step 61 departs from a critical zone around the coordinates of the reference mean vector, then it can be diagnosed that the rotor has suffered damage, for example a broken blade.
Thus, in step 71, the modulus d of each vector difference D is calculated, i.e. for each speed range. Thereafter, the values of these moduluses are stored in memory in step 81.
Subsequently, in step 90, the modulus d of the vector difference is compared with the predetermined threshold value. This threshold value may correspond to a value lying in the range 2 mils to 5 mils, for example.
The vibration sensor is more sensitive to the compressor being unbalanced than it is to the turbine being unbalanced. In addition, the sensitivity of the vibration sensors also depends on engine speed.
By way of example, the sensitivity of the sensor varies over the range about 200 centimeter grams per mil (cm.g/mil) to 300 cm.g/mil, i.e. an unbalance having a moment of about 200 cm.g corresponds to a vibration having an amplitude of 1 mil.
In general, a broken blade gives rise to an unbalance of about 2000 cm.g. Thus, given that there will normally be residual unbalance and given the differing sensitivities of the sensors, the rupture of a blade corresponds to a change in vibration amplitude of 2 mils to 5 mils. It should be observed that the range of variation for the threshold value may vary depending on the model of engine.
A process for calculating the statistical dispersion of vectors is described below with reference to steps 62 to 82. Thus, in step 62 (see also
It should be observed that it is also possible, in step 62, to calculate the vector difference between each vibration vector (as defined in step 30) of the determined flight and the mean vibration vector (stored in memory in step 50) of the determined flight for a given rotor speed range.
Then, in step 72, the modulus of the vector difference associated with each vibration vector is calculated and the maximum modulus dmax1 or dmax2 is determined. Thereafter, the maximum modulus, i.e. the largest modulus associated with the determined speed range is stored in the memory at step 82, or the larger modulus if there are only two of them.
Thereafter, in step 90, the maximum modulus for the determined speed range is compared with the predetermined threshold value.
When the maximum modulus (stored in memory in step 82) or a vector difference modulus (stored in memory in step 81) exceeds the predetermined threshold value, then a warning signal is issued on the cockpit screen or on a printer for the attention of maintenance personnel. The engine then needs to be examined appropriately before it is restarted.
In addition, when the modulus or the maximum modulus for the vector difference does not exceed the predetermined threshold value, the mean vectors of the reference flight are updated using the data from the determined flight so as to keep track of the normal aging of the engine.
It is also possible to envisage data relating to rotor speed and vibration being stored on a removable storage medium to enable the data to be processed by a computer on the ground after the aircraft has landed.
The method of detecting damage to a rotor in an aircraft engine comprising the above steps and implemented after a determined flight is particularly advantageous in that the pilot is not distracted by problems of this kind that are of little importance. Another advantage is the fact that the onboard processing system of the aircraft is not overloaded.
That said, it is entirely possible to detect rotor damage while in flight using a method similar to that of FIG. 3.
The steps of the flow chart shown in
Number | Date | Country | Kind |
---|---|---|---|
02 06530 | May 2002 | FR | national |
Number | Name | Date | Kind |
---|---|---|---|
4303882 | Wolfinger | Dec 1981 | A |
4435770 | Shiohata et al. | Mar 1984 | A |
4453407 | Sato et al. | Jun 1984 | A |
4685335 | Sato et al. | Aug 1987 | A |
4955269 | Kendig et al. | Sep 1990 | A |
5148711 | Twerdochlib et al. | Sep 1992 | A |
5258923 | Imam et al. | Nov 1993 | A |
5541857 | Walter et al. | Jul 1996 | A |
5544073 | Piety et al. | Aug 1996 | A |
5686669 | Hernandez et al. | Nov 1997 | A |
5744723 | Piety | Apr 1998 | A |
6098022 | Sonnichsen et al. | Aug 2000 | A |
6263738 | Hogle | Jul 2001 | B1 |
6321602 | Ben-Romdhane | Nov 2001 | B1 |
6445995 | Mollmann | Sep 2002 | B1 |
Number | Date | Country |
---|---|---|
1 118 920 | Jul 2001 | EP |
Number | Date | Country | |
---|---|---|---|
20040060347 A1 | Apr 2004 | US |