Method for monitoring an aircraft landing gear with at least one accelerometer on the landing gear wheel and aircraft using this method

Description

[0001] The invention relates to the field of aeronautics and more particularly to the monitoring of an aircraft landing gear. BACKGROUND OF THE INVENTION

[0002] Aircraft landing gear is constantly monitored. Landing gear is rarely equipped with a hard landing detector, which can signal a particularly rough landing, which could potentially lead to the deformation of certain parts of the landing gear, in particular the axle(s) carrying the landing gear wheels. This detector constitutes additional on-board equipment, a source of additional weight and maintenance. Examples of detectors are known, described in documents EP3118030 A1, EP3336485 A1 and US8543322 B1.

[0003] Furthermore, there is an unmet need to monitor the deflections of the landing gear axles in real time, to check whether they are in line with the expected values ​​in service or whether, on the contrary, they reach values ​​suggesting plastic deformation of a portion of the landing gear. These plastic deflections are in practice only detectable during a heavy maintenance operation during which the landing gear is removed from the aircraft and returned to the maintenance workshop. SUBJECT OF THE INVENTION

[0004] The invention aims to propose a means of real-time monitoring of an aircraft landing gear which does not require equipping the landing gear with ad hoc devices or removing the landing gear. PRESENTATION OF THE INVENTION

[0005] In order to achieve this goal, a method for monitoring an aircraft landing gear is proposed, comprising the step of using information from at least one accelerometer secured to a wheel of the landing gear mounted for rotation on an axle of the landing gear, to estimate a deflection angle of the axle at least when the wheel is in contact with the ground.

[0006] Thus, we take advantage of the presence of an accelerometer on the lander wheels to exploit the signal in order to generate information which is not directly linked to the wheel, but useful for monitoring the lander itself and its deformations in service.

[0007] In particular, the wheels of the landing gear are equipped with tire pressure monitoring systems (TPMS). The latest devices of this type under development are energy-autonomous, transmit their information remotely, and are equipped with at least one three-axis accelerometer. Some are even equipped with two accelerometers, one to identify the flight phase (in particular by detecting the acceleration caused by the impact of the wheels on the ground during landing), and the other to identify the angular position of the device when the aircraft is stationary.

[0008] Thus, and according to an advantageous embodiment of the invention, the signals from the accelerometer(s) integrated into a tire pressure monitoring device attached to the wheel are used. Thus, the presence of the accelerometers in the tire pressure monitoring device is taken advantage of to exploit the signals. The signals from these accelerometers are available at any time, and can therefore be used to generate the information sought about the landing gear.

[0009] The invention also relates to an aircraft equipped for implementing this method. DESCRIPTION OF FIGURES

[0010] The invention will be better understood in light of the following description of a particular embodiment of the invention, with reference to the figures of the appended drawings, among which: [ Fig. 1 ] there figure 1is a schematic view of an aircraft landing gear, the wheels being shown in dotted lines to show the axle which carries them; [ Fig. 2 ] there figure 2 is a perspective view of half a rim of one of the wheels of the landing gear of the figure 1 , equipped with a tire pressure monitoring device; [ Fig. 3 ] there figure 3 is a diagram illustrating the various reference points used for the exploitation of the signals from the accelerometers, the wheel being seen here from the side; [ Fig. 4 ] there figure 4 is a partial schematic view of the lander of the figure 1 seen from the front, illustrating in an exaggerated manner the deflection of the axle which one seeks to estimate by the method of the invention. DETAILED DESCRIPTION OF THE INVENTION

[0011] The invention relates to the monitoring of an aircraft landing gear 100 as illustrated in the figure 1. The landing gear 100 comprises a leg 101 secured to the structure of the aircraft, carrying an axle 102 on which wheels 103 are rotatably mounted. Each wheel 103 comprises a rim 104 carrying a tire 105. In a manner known per se, each rim comprises two half-rims bolted together. On the figure 2 an outer half-rim 106 is illustrated, equipped with a tire pressure monitoring device 107. Said device incorporates at least one accelerometer 108, here a triaxial accelerometer, the signals of which are used according to the invention. For the description of the use of the accelerometer signals, the following direct references are used: Ground-bound G marker Reference A linked to the axle Marker M linked to the wheel Repère C lié à l'accéléromètre Premier axe Xg Xa Xm Xc Second axis Yg Yes Ym Yc Third axis Zg Za Zm Zc The landmarks, which are illustrated in Figures 3 and 4 , are constructed as follows: Reference G: direct orthonormal reference frame linked to the wheel considered non-inclined and mounted on a perfect, non-flexed axle, having as its origin the point O, in which the Z axis g is vertical, the X axis g is horizontal and follows the forward rolling direction, and the Y axis g is horizontal and completes the direct orthonormal trihedron; Reference A: direct orthonormal reference frame linked to the axle, and whose origin is on the geometric axis of the latter at a point O, substantially at the center of the wheel 103. As is more particularly visible at figure 4 , we move from the G frame to the A frame by a rotation of angle α around the X axis g which is the deflection angle of the axle to be estimated. The Y axis a extends along the axis of rotation of the wheel. The origin of this frame moves along the X axis g with a speed V 0. The transition matrix from frame G to frame A is the matrix: Rot G → A = 1 0 0 0 cos α sin α 0 − sin α cos α M frame: direct orthonormal frame linked to the wheel, and whose origin is a point M located at a distance R of the Y axis a. The point M corresponds substantially to the center of the accelerometer 108 carried by the pressure monitoring device 107. The reference point M is deduced from the reference point A by a rotation around the Y axis a by an angle β , which is the angle of rotation of the wheel when the aircraft is moving. The rotational speed of the wheel is therefore oh = ḃ . The Y axis has the same direction as the Y axis a . The Z axis m extends in a radial direction, while the X axis m extends in an orthoradial direction. The transition matrix from the A frame to the M frame is the matrix: Rot A → M = cos β 0 − sinβ 0 1 0 sinβ 0 cosβ Reference frame C: direct reference frame (not necessarily orthonormal) linked to the accelerometer 108 carried by the pressure monitoring device 107. The three axes of the reference frame C correspond to the three input axes of the three elementary accelerometers forming the accelerometer 108. It should be noted that, in theory, the reference frames C and M are the same. In practice, assembly defects mean that they are slightly offset from each other. We move from the reference frame M to the reference frame C by 6 small rotations defining the alignment errors (or misalignments) θ xy , θ xz , θ yx , θ yz , θ zx , θ zy accelerometers, which define the passage matrix making it possible to obtain the accelerations felt by the accelerometer 108 in the C frame from the measurements made in the M frame: Proj M → C = 1 θ xy θ xz θ yx 1 θ yz θ zx θ zy 1

[0012] All calculations made, and retaining only the first order terms, the accelerations c x,c (substantially ortho-radial), c y,c(in a direction substantially parallel to a central axis of the axle), and c z,c (substantially radial) measured by the elementary accelerometers of the triaxial accelerometer 108 along the three axes of the C reference frame are: γ x , c γ y , c γ z , c ≈ − g . sinβ + g . θ xz . cosβ + τ + cosβ . V ˙ o − θ xz . Rω 2 + b x g . sinα − g . θ yx . sinβ − θ yz . cosβ − θ yz . Rω 2 + b y g . cosβ − g . θ zx . sinβ + sinβ . V ˙ o − Rω 2 + b z In this expression, g is the acceleration of Earth's gravity, t is the ratio between the distance R and the rolling radius d of the wheel, and the quantities bx , by , bz are offsets (commonly called biases) affecting the acceleration measurements on each of the axes of the accelerometer 108. It appears that the transverse acceleration c y,c includes different terms, including: 1- a term in g.sina, substantially constant during a rolling phase on the ground, and which therefore depends directly on the deflection angle α of axle 102; 2- the offset term by , substantially constant; 3- periodic variable terms with the rotation of the wheel in sinβ, cosβ ; 4- finally a variable term with the rotation speed of the wheel in oh 2< . By exploiting the transverse acceleration according to the invention γ y,c , it is therefore possible to estimate the deflection angle alpha of axle 102.

[0013] According to a first particular type of implementation of the method of the invention, the angle is estimated β of rotation of wheel 103 from orthoradial acceleration c x,c and / or radial acceleration c z,c , and from this estimate, we estimate the rotation speed oh = ḃ . These two estimates allow us to estimate the variable terms 3 / and 4 / to eliminate them from the transverse acceleration γ y,c , in order to keep only the quantity in by + g.sina, which will be processed to extract the deflection angle α .

[0014] A second particular type of implementation of the invention comprises the filtering of the periodic variable terms with the angle β rotation of the wheel 103 using a low-pass filter whose cut-off frequency is lower than the rotation frequency of the wheel (equal to oh / 2 π ) . We thus obtain the filtered accelerations: γ x , c , f γ y , c , f γ z , c , f ≈ τ . V ˙ o − θ xz . Rω 2 + b x g . sinα − θ yz . Rω 2 + b y − Rω 2 + b z

[0015] At sufficiently high speed, the offset term bz is negligible, so the filtered acceleration c z,c,f constitutes a good estimator of the term - Rω 2< . Substitute in the expression of γ y,c,f , we obtain: γ y , c , f − θ yz γ z , c , f ≈ g . sinα + b y So it is again possible to estimate the same quantity by + g.sina, and therefore to monitor the deflection angle α of the axle. This last relationship can still be modeled in the form: y = h . X + v Or v is a measurement noise, with: y = γ y , c , f h = 1 γ z , c , f X = g . sinα + b y θ yz We can then estimate the two components of the state vector X by accumulating the measures y And h over the period during which the filtering is effective, precisely after the initial transient at the time of impact of the wheel 103 on the ground and before the rotation speed of the wheel is too slow (i.e. when the wheel frequency is lower than the cutoff frequency of the filter), then estimating the state vector X by any suitable method, such as a pseudo-inverse method.

[0016] A first algorithm allowing the implementation of this second particular type of implementation of the method of the invention is now described. It comprises the following five steps: A- We begin the acceleration measurements c y,c And c z,cat a fixed frequency (100 Hz for example) then they are filtered using a low-pass filter with a cut-off frequency of a few Hz. The start of sampling takes place at the start of landing, detectable when | c z,c | exceeds a certain threshold or is in saturation. Filtering can start on saturated values, allowing the transitional period to pass more quickly and filtering to be established more quickly; B- Once the saturation phase has been exceeded, the following data is taken, at the same frequency or at a lower frequency: y i = γ y , c , f t i h t i = 1 γ z , c , f t i . where the you are the sampling times; C- Sampling is stopped when the wheel speed is lower than a given threshold, which can be detected with a threshold on | c z,c,f |. D- By putting the collected data in the following matrix form: Y = y 1 … y n H = h 1 … h n we obtain the estimate of the state vector X by calculating X̂ = MY with the pseudo-inverse M = ( H t< .H)-1< . H t< E-Finally, the estimate of the deflection angle α is obtained by a = arcsin( X̂ (1) / g ) (neglecting the offset by ), and the estimate of the misalignment yz is obtained by θ̂ yz = X̂ (2) .

[0017] This first algorithm can be activated as soon as the lander's wheels touch the ground. It is possible to detect this instant t 0 by monitoring tire pressure by detecting a sudden increase in tire pressure on the rim in question, and thus triggering the acquisition of measurements. The start of processing may be delayed if the measurements become saturated. In particular, acceleration c z,cintegrates centripetal acceleration, which can reach a few tens of g at the start of rotation. In this case, low-pass filtering can nevertheless be activated, allowing steady state to be reached more quickly (elimination of the transient). After a predefined delay, the measurements are taken to develop the quantities Y And H. The end of the sampling occurs either after a predefined time, or when the acceleration c z,c,f falls below a predetermined threshold. Below this threshold, low-pass filtering is no longer effective in eliminating sinusoidal variable terms. The calculation of α̂ can then be carried out.

[0018] The calculations can be carried out in real time in one of the aircraft's computers (for example, calculation means on board the wheel or the landing gear), or remotely by transmitting the data to a data collection device, on each flight or during a visit of the aircraft to a maintenance center.

[0019] The calculation of the pseudo-inverse can be replaced by a recursive least-squares method. This can smooth out the computational load over acquisitions. When a large number of points are used, for example, for the least-squares estimation of the state vector, the prior low-pass filtering step A / can be omitted, with the least-squares estimate then fulfilling the filtering role.

[0020] A second simplified algorithm is now proposed. Indeed, the previous algorithm is sensitive to misalignment θ yz . If the misalignment yzis known at least approximately before the flight (obtained with the first algorithm for example, and assuming that it remains sufficiently constant over time, which is generally the case), and assuming that, at sufficiently low speed, the term - θ yz .Rω 2< is small in front g.sina, we correct the filtered transverse acceleration c y,c,f in the following manner: γ y , c , cor ≈ γ y , c , f − θ yz γ z , c , f The estimate of the deflection angle α is then obtained directly by: α ^ = arcsin γ y , c , cor / g .

[0021] An alternative implementation of the method of the invention applies to a wheel whose pressure monitoring device 107 comprises a second accelerometer 109, also triaxial, arranged diametrically opposite relative to the first accelerometer 108. By calling c x,c, 1 , c y,c, 1 , c z,c ,1 the accelerations measured by the first accelerometer 108, and c x,c, 2, c y,c, 2, c z,c, 2 the accelerations measured by the second accelerometer 109, the radial accelerations y z,c, 1 and c z,c, 2 are worth respectively: γ z , c , 1 = g . cosβ − g . θ zx , 1 . sinβ + sinβ . V ˙ o − Rω 2 + b z , 1 γ z , c , 2 = − g . cosβ + g . θ zx , 2 . sinβ − sinβ . V ˙ o − Rω 2 + b z , 2 So the average of these two accelerations is: γ ¯ z , c = γ z , c , 1 + γ z , c , 2 2 = g θ zx , 2 − θ zx , 1 2 . sinβ − Rω 2 + b z , 1 + b z , 2 2 In this average, the variable term in g.cosβ has disappeared. A sinusoidal variable term remains, but of low amplitude, much easier to filter. This averaged acceleration is then used to implement the method of the invention.

[0022] Thanks to the method of the invention, it is therefore possible to estimate the deflection angle of the axle using only the signals from one or more accelerometers carried by the associated wheel. This estimation can be used to monitor the deformations of the axle in different forms such as for example: monitoring during aircraft operation of axle deflection by checking at each landing that the axle deflection remains within values ​​compatible with normal elastic deformations; recording of the deflections thus estimated to establish a trend curve and detect any possible drift in this deflection over time; detecting significant deflection following a landing, which could potentially indicate plastic deformation of the axle, this detection being the result of one of the accelerations measured by the accelerometer(s) exceeding a predetermined threshold.

[0023] The invention is not limited to what has just been described, but on the contrary encompasses any variant falling within the scope defined by the claims.

[0024] In particular, although the signals from the triaxial accelerometers integrated into the tire pressure monitoring device are used here, it is more generally possible to use an accelerometer that is not necessarily triaxial, since the deflection angle can even be estimated using only the filtered transverse acceleration. γ y , c , f ≈ g . sinα − θ yz . Rω 2 + b y if we also know how to estimate the rotation speed of the wheel (for example using the speed information provided by a tachometer) to eliminate the term in Rω 2< .

[0025] The G marker can be linked to the track.

[0026] The invention can be applied to an aircraft carrying an electronic control unit (comprising for example a processor and a memory) which is connected to the accelerometers and which executes a computer program comprising instructions arranged to implement the method. The electronic control unit is for example that of the aircraft monitoring system ACMS (Aircraft Condition Monitoring Systems). Alternatively, the electronic control unit is housed in the landing gear hold and can be interrogated by a radio or wired link using a maintenance terminal.

Claims

1. A method of monitoring an aircraft landing gear (100) provided with a device (107) for monitoring the pressure of a tyre (105), the tyre (105) being carried by a wheel (103) of the landing gear, the wheel (103) being mounted to rotate on a landing gear axle (102), the method comprising the step of estimating an angle of deflection α of the axle at least when the wheel is in contact with the ground, characterised in that said estimation uses information from at least one accelerometer (108) integrated into the pressure monitoring device (107).

2. The method according to Claim 1, wherein a transverse acceleration γy,c measured by the accelerometer (108) in a direction substantially parallel to a central axis of the axle (102) and dependent on the angle of deflection α of the axle is used to estimate the latter.

3. The method according to Claim 2, wherein an angle β of rotation of the wheel (103) is estimated from an orthoradial acceleration γx,c and / or a radial acceleration γz,c measured by the accelerometer (108), and a speed of rotation ω = β̇˙ is then estimated, these two estimates being utilised to eliminate from the transverse acceleration γy,c terms varying with the angle of rotation and the speed of rotation, and then the angle of deflection of the axle is estimated from the transverse acceleration γy,c thus reprocessed.

4. The method according to Claim 2, wherein the transverse acceleration γy,c and a radial acceleration γz,c measured by the accelerometer (108) are filtered to obtain a filtered transverse acceleration γy,c,f and a filtered radial acceleration γz,c,f by eliminating the terms varying periodically as a function of an angle of rotation β of the wheel.

5. The method according to Claim 4, wherein filtering is started from a landing onset that is detected when the quantity |γz,c| exceeds a certain threshold or is in saturation.

6. The method according to Claim 4, wherein, after the filtering has been established: - the data is collected y i = γ y , c , f t i h t i = 1 γ y , c , f t i . where the ti are the sampling instants, γy,c,f is the filtered transverse acceleration, and γz,c,f is the filtered radial acceleration; - the sampled data are put into the following matrix form: Y = y 1 … y n H = h 1 … h n - a state vector X is estimated by calculating X̂ = M.Y with the pseudo-inverse M = (Ht.H)-1.Ht ; - the angle of deflection of the axle is estimated by calculating α̂ = arcsin(X̂(1) / g) where g is the acceleration of gravity.

7. The method according to Claim 6, wherein the pseudo-inverse calculation is replaced by a recursive least-squares method.

8. The method according to Claim 4, wherein a corrected acceleration (γy,c,cor) is calculated from the filtered accelerations γy,c,f and γz,c,f : γ y , c , cor ≈ γ y , c , f − θ yz γ z , c , f Where θyz is given or estimated beforehand, and the angle of deflection of the axle is estimated by the calculation α ^ = arcsin γ y , c , cor / g where g is the acceleration of gravity.

9. The method according to Claim 4, wherein two accelerometers (108, 109) are used that are arranged on the wheel (103) diametrically opposite each other and measuring respective radial accelerations γz,c,1 and γz,c,2, and wherein the mean radial acceleration γ ¯ z , c = γ z , c , 1 + γ z , c , 2 2 is formed and used as the radial acceleration for estimating the deflection angle of the axle.

10. An aircraft comprising at least one landing gear provided with a wheel to which is attached at least one accelerometer connected to an electronic control unit of a device (107) for monitoring the pressure of a tyre (105) performing a computer program comprising instructions arranged to implement the method according to any one of the preceding claims.