METHOD FOR MONITORING THE TEMPERATURE OF BRAKES ON AN AIRCRAFT LANDING GEAR.

Abstract

A device for monitoring a maximum temperature reached during a landing by a brake of a landing gear of an aircraft is implemented. The monitoring device uses a prediction model to estimate (102) a maximum temperature reached by the brake during the landing. An error between the estimated maximum temperature and the measured maximum temperature is then determined (103). If the error between the estimated maximum temperature and the measured maximum temperature is greater than a first predefined threshold S1, when the measured maximum temperature is greater than the estimated maximum temperature, or a second predefined threshold S2, when the measured maximum temperature is lower than the estimated maximum temperature, and if a total number of errors is greater than a third predefined threshold S3, then an alert message is generated (106) for the attention of the crew and / or ground personnel.It is thus possible to anticipate maintenance actions on the elements of a landing gear. Figure to be published with the abstract: Fig. 2.

Description

Title of the invention: METHOD FOR MONITORING THE TEMPERATURE OF BRAKES ON AN AIRCRAFT LANDING GEAR. Technical field

[0001] The present invention relates to a method and a device for monitoring the temperature of brakes of a landing gear, such as for example a main landing gear or "Main Landing Gear" (MLG), of an aircraft. More particularly, the invention relates to a detection for each of the brakes of the same landing gear, of a difference between an estimated maximum temperature of a brake and an actual maximum temperature measured of said brake. STATE OF PRIOR ART

[0002] The main functions of a landing gear are to allow an aircraft to move on the ground. In particular, a landing gear allows the following to be controlled: taxiing maneuvers between the different locations of an aerodrome (i.e., towing, taxi, etc.), the takeoff run, the cushioning of the landing impact, and, thanks to an associated braking system, the stopping of the aircraft over an acceptable distance.

[0003] Generally, each wheel of a landing gear is equipped with a braking system including brakes, as well as temperature sensors located on or near these brakes. When the brakes of a landing gear are activated, their temperature increases. The temperature sensors therefore measure the temperature of each brake individually and then transmit these measurements to respective brake temperature monitoring units ("Brake Temperature Monitoring Unit"), located on the landing gear. In one example, a landing gear includes four wheels and therefore four brakes and therefore four associated monitoring units. Each monitoring unit then transmits the brake temperature data to a Braking and Steering Control Unit, integrated into the aircraft's avionics systems.

[0004] Before takeoff, to ensure the safe operation of the aircraft and to avoid degradation of its performance, the brakes must not exceed a limit temperature (for example, greater than 400°C). This may involve, for example, preventing the brakes from heating to temperatures above their safe operating range.

[0005] The braking and piloting control unit monitors the evolution of the brake temperature during a given landing, but also over a predefined period during which several landings took place.

[0006] Thus, at the end of landing, when the temperature of at least one of the brakes of a landing gear is higher than the predefined threshold, the braking and piloting control unit transmits via a human-machine interface of the ECAM (Electronic Centralized Aircraft Monitor) an alert message to the crew in the event of abnormal behavior of the brakes. For example, if the temperature of a brake is higher than 100°C, a message informing them of the possibility of taking off the aircraft appears on a human-machine interface of the ECAM. If the temperature exceeds 300°C, a message indicating that takeoff must be delayed to allow the brakes to cool appears on the human-machine interface of the ECAM.

[0007] In another example, the braking and piloting control unit makes it possible to detect a possible asymmetry in the temperature of the brakes of the same landing gear. An asymmetry may be the result of abnormal braking conditions such as: oxidation of the brake lining, residual braking or released brakes. Thus, if the temperature asymmetry between the brakes reaches a predefined maintenance threshold, noted S (for example 150°C), then an alert message, generated by the braking and piloting control unit and then transmitted to the ECAM for example, indicates that a maintenance action or check is required (for example: brake repair, brake replacement, etc.).

[0008] However, according to this technique, it is difficult to identify precisely which brake, or which other landing gear equipment (e.g.: wheel, braking system equipment, etc.) of the landing gear generates the temperature asymmetry between the brakes and requires maintenance action or verification (e.g.: repair, replacement, etc.).

[0009] Indeed, an abnormally high maximum temperature for a brake can influence the maximum temperature value of another brake of the landing gear.

[0010] It is then desirable to overcome these drawbacks of the state of the art.

[0011] It is particularly desirable to provide a solution that makes it possible to monitor the evolution over time of the maximum temperature of each brake of the same landing gear individually and thus anticipate the maintenance actions to be carried out on brakes or any other equipment of the landing gear (for example: wheels, sensors, etc.). Furthermore, it is desirable to provide a solution that makes it possible to identify precisely which brake, or which other equipment of the landing gear (for example: wheels, sensors, etc.) requires a maintenance action or check. Statement of the invention

[0012] A method is proposed here for monitoring a maximum temperature reached during a landing by a brake of a landing gear of an aircraft. The method is implemented by a monitoring device. The method comprises:

[0013] a phase of using a model for predicting a maximum temperature reached by the brake during landing comprising the following steps

[0014] - obtaining a current set of values ​​of a plurality of braking parameters, for a running landing;

[0015] - estimate, using the prediction model, a maximum temperature during the landing which should be reached by said brake, from the values ​​of said current set.

[0016] The method further comprises a comparison phase comprising:

[0017] - obtain a maximum measured temperature reached by the brake during the current landing;

[0018] - determine whether an error between the estimated maximum temperature and the temperature maximum measured temperature is greater than a first predefined threshold SI, when the maximum measured temperature is greater than the estimated maximum temperature, or a second predefined threshold S2, when the maximum measured temperature is lower than the estimated maximum temperature;

[0019] - when the error is greater than the first predefined threshold SI or the second threshold predefined S2, determine whether, on a set of landings comprising a number N of landings from a sliding window and the current landing, a total number of errors is greater than a third predefined threshold S3, otherwise repeat the phase of using the prediction model and the comparison phase for a following landing;

[0020] - when the total number of errors is greater than the third predefined threshold S3, generate an alert message, otherwise repeat the prediction model usage phase and the comparison phase for a subsequent landing.

[0021] Thus, it is possible to monitor the temperature of each brake of a landing gear upon landing of each wheel independently. An alert is then triggered when a large difference is observed between the measured temperature and the estimated temperature of the analyzed brake, which makes it possible to anticipate maintenance operations on the different elements of the landing gear.

[0022] According to a particular embodiment, prior to the use phase of the prediction model, the method includes a training phase of a machine learning model comprising: associating a reference set of braking parameter values that has been obtained for each landing of a set of reference landings with a reference value of the maximum temperature that has been reached by the brake of the landing gear during the landing in question.

[0023] In a particular embodiment, estimating the maximum temperature reached by the brake for the current landing, from the values of the current set of braking parameters, further comprises: calculating a weighted sum between the estimated maximum temperature for the brake and an estimated maximum temperature for another brake of the landing gear.

[0024] In a particular embodiment, the method further comprises determining a synthetic estimated temperature as the weighted sum of a sliding average, over the sliding window of N landings, of the estimated maximum temperature of the brake and of a measured temperature of the other brake of the landing gear.

[0025] In a particular embodiment, the error between the estimated maximum temperature and the measured maximum temperature is calculated as a weighted sum between an average absolute error between a sliding average, over the sliding window of N landings, of the maximum temperature reached by the brake and the synthetic estimated temperature, and a symmetrical average absolute error in percentage between the sliding average, over the sliding window of N landings, of the maximum temperature reached by the brake and the synthetic estimated temperature.

[0026] In a particular embodiment, the braking parameters are among: a braking energy, a maximum braking power, a duration of activation of the brake fans, a duration of reversal of the thrust of the engines, a duration of the landing, an initial temperature of the brake, a duration between a moment when the maximum temperature of the brake is reached and a moment when a rotation speed of the wheel reaches the value of the 95% percentile, a static air temperature, an average of the ground speed, a maximum value of the ground speed, a sum of the currents applied by a servovalve of the braking system, an average altitude, a manufacturer serial number of the aircraft, an aircraft engine type identifier, an aircraft model identifier.

[0027] In a particular embodiment, the alert message includes an indication that a maintenance action or check must be performed on the brake, and an indication of a type of maintenance action or check to be performed.

[0028] In a particular embodiment, the method further comprises adapting the alert message by indicating:

[0029] - information on a first category of problem and information on a first category of action or maintenance check to be carried out which is adapted to the first category of problem, when the maximum measured temperature is higher than the estimated maximum temperature and the error between the estimated maximum temperature and the maximum measured temperature is higher than the first predefined threshold SI, or

[0030] - information on a second category of problem and information on a second category of action or maintenance check to be carried out which is adapted to the second category of problem, when the maximum measured temperature is lower than the estimated maximum temperature and the error between the estimated maximum temperature and the maximum measured temperature is greater than the second predefined threshold S2.

[0031] Also proposed here is a device for monitoring a maximum temperature reached during a landing by a brake of an aircraft landing gear. The monitoring device comprising electronic circuitry configured to implement:

[0032] a phase of using a model for predicting a maximum temperature reached by the brake during landing comprising the following steps:

[0033] - obtaining a current set of values ​​of a plurality of braking parameters, for a running landing;

[0034] - estimate, using the prediction model, a maximum temperature during the landing which should be reached by said brake, from the values ​​of said current set;

[0035] said method further comprising a comparison phase comprising:

[0036] - obtain a maximum measured temperature reached by the brake during the current landing;

[0037] - determine whether an error between the estimated maximum temperature and the temperature maximum measured temperature is greater than a first predefined threshold SI, when the maximum measured temperature is greater than the estimated maximum temperature, or a second predefined threshold S2, when the maximum measured temperature is lower than the estimated maximum temperature;

[0038] - when the error is greater than the first predefined threshold SI or the second threshold predefined S2, determine whether, on a set of landings comprising a number N of landings from a sliding window and the current landing, a total number of errors is greater than a third predefined threshold S3, otherwise repeat the phase of using the prediction model and the comparison phase for a following landing;

[0039] - when the total number of errors is greater than the third predefined threshold S3, generate an alert message, otherwise repeat the prediction model usage phase and the comparison phase for a subsequent landing.

[0040] Also proposed here is an aircraft comprising a monitoring device as described above according to one embodiment. Brief description of the drawings

[0041] The above-mentioned and other features of the invention appear will become more clearly apparent from reading the following description of at least one exemplary embodiment, said description being made in relation to the attached drawings, among which:

[0042] [Fig.l] illustrates in diagrammatic form the steps of the method for monitoring the maximum temperature of a brake of a landing gear of an aircraft during a landing, according to one embodiment;

[0043] [Fig.2] illustrates in diagrammatic form a preliminary step of characterization of braking parameters, according to one embodiment;

[0044] [Fig.3] illustrates in graphic form an example of the result obtained after implementation description of the method for monitoring the maximum temperature of a brake of a landing gear of an aircraft during a landing, according to one embodiment;

[0045] [Fig.4] illustrates in graphic form another example of the result obtained after implementation of the method for monitoring the maximum temperature of a brake of a landing gear of an aircraft during a landing, according to one embodiment;

[0046] [Fig.5] schematically illustrates an example of hardware architecture of a device monitoring according to one embodiment; and

[0047] [Fig.6] schematically illustrates, in side view, an aircraft equipped with a device for monitoring, according to one embodiment.

[0048] DETAILED DESCRIPTION OF EMBODIMENTS

[0049] The general principle of the invention relates to the monitoring of the maximum temperature reached by each of the brakes of the same landing gear of an aircraft (for example: main landing gear), independently of each other, during their activation. More specifically, it is the difference, or error, between a so-called "estimated" maximum temperature and a so-called "real" or "measured" maximum temperature of a brake which is monitored for each brake of a landing gear during the activation of the latter.

[0050] Subsequently, in order to illustrate the method described below, it is considered that the activation of the brakes takes place during a landing of the aircraft. It should be noted that the activation of the brakes of the landing gear of the aircraft can also take place during taxiing, or during a takeoff phase of the aircraft during emergency situations.

[0051] Hereinafter, "landing" or "landing phase" means the period extending from the moment the aircraft touches the ground (i.e., flight phase No. 8) to the moment the aircraft's engines are cut (i.e., flight phase No. 10). In one embodiment, a margin is taken before flight phase No. 8 (for example: 1 minute) and / or after flight phase No. 10 (for example: 10 minutes), to ensure that the entire landing is covered.

[0052] The estimated maximum temperature is obtained through the use of a device of monitoring, noted DISP, of a model for predicting the maximum temperature of a brake that is reached during a landing (also simply called prediction model hereinafter). More particularly, this prediction model is implemented in a machine learning module of the DISP monitoring device (eg, artificial intelligence algorithm). In order to obtain this prediction model, a machine learning model is trained (training phase) to associate parameter values, called "braking parameters", with the maximum temperature reached by each brake of the landing gear as measured during a set of reference landings.At the end of this training phase, the prediction model is used (use phase) to estimate the maximum temperature of each landing gear brake from new values ​​of the braking parameters obtained for a new landing, which makes it possible to compare the maximum temperature thus estimated with a maximum temperature value actually measured during this new landing.

[0053] Hereinafter, the term "braking parameter" means parameters capable of exerting an influence on the temperature of the brakes during the landing of the aircraft (i.e., increasing the temperature of the brakes, or decreasing the temperature of the brakes during landing). In one example, these braking parameters are: braking pressure, static air temperature, wind speed and direction, vertical acceleration, duration of application of reverse thrust of the engines, alternative braking pressure, duration of activation of an anti-skid (or "Anti-skid" in English), duration of activation of a brake fan, duration of landing, gross weight of the aircraft, braking energy, rotation speed of the wheels of the landing gear on landing, etc.

[0054] Thanks to this prediction model, it is possible to estimate, for each landing, the maximum temperature reached by each of the landing gear brakes individually under so-called "nominal" landing conditions for a given type of aircraft (i.e., conventional landing conditions for an aircraft assuming that the brakes do not require any maintenance action or check).

[0055] Thus, it is possible for a landing considered, to compare the estimated maximum temperature of each of the brakes with the actual maximum temperature measured for these brakes during this landing. Depending on the difference, or error, between these estimated and measured maximum temperatures, an alert message is generated by the DISP monitoring device in order to notify the crew (for example via the ECAM human-machine interface) and / or the ground personnel by means of air-ground communication, of a need for maintenance to be checked and / or carried out on one of the brakes and / or other equipment of the landing gear (for example: wheel, sensor, etc.).

[0056] [Fig.l] illustrates in diagrammatic form the steps of the method for monitoring the maximum temperature reached during a landing of the brakes of a landing gear according to one embodiment. This method for monitoring the maximum temperature of the brakes is implemented in the monitoring device DISP, as described below in connection with [Fig.5].

[0057] The DISP monitoring device comprises electronic circuitry configured in particular to collect in real time from one or more sensors (for example temperature sensor) and / or information systems of the aircraft (for example: braking and steering control unit or BSCU or aerodynamic data computer or ADC for "Air Data Computer" in English), information on: the temperature of each brake of the landing gear of the aircraft, including the initial temperature of the brakes at the start of the landing phase, braking parameters (for example: duration of activation of the brake fans, duration of reversal of the thrust of the engines, etc.).

[0058] During a preliminary step (not shown), these braking parameters are characterized. These braking parameters will then be used for the training phase of the machine learning model and the phase of using the prediction model.

[0059] It should be noted that, in a particular embodiment, it is possible to classify these braking parameters according to their influence on the temperature of the brakes during landing. Indeed, certain braking parameters influence the temperature of the brakes more significantly than others. Thus, in order to limit the number of data, among the braking parameters identified previously, only those having a greater influence on the temperature of the brakes during landing than others will be used for the training phase of the machine learning model and the phase of use of the prediction model.

[0060] [Fig.2] thus illustrates in diagrammatic form a preliminary step of characterizing the braking parameters according to one embodiment.

[0061] During a sub-step 201 COL_ST, data concerning the braking parameters are obtained from one or more sensors and / or an information system of the aircraft, such as the ADC. This data collection is carried out over a predefined period corresponding to a flight window characteristic of the landing phase of the aircraft as defined previously. Furthermore, this data collection is carried out according to a predefined sampling rate for each braking parameter. Indeed, the sampling rate of the braking parameters depends on the recording frequency of the sensor or system in question, for example 2Hz.

[0062] Thus, for a period corresponding to the landing phase of the aircraft, a time series of values ​​for each braking parameter is obtained.

[0063] In order to simplify the prediction model, during a sub-step 202 EXT_V, these time series of values ​​are processed to extract characteristic values ​​of the braking parameters. In other words, a single characteristic value of the braking parameter is extracted instead of having the entire evolution of the values ​​of the time series for this braking parameter. For example, it is possible to extract, from a time series of values ​​corresponding to the evolution of the temperature of the brakes during the entire landing phase, a characteristic value which is the maximum temperature of the brakes on landing. Overall, from the time series, characteristic values ​​are extracted to be used as input data for the prediction model.

[0064] It is thus possible to characterize, during a sub-step 203 ID_PC, braking parameters for estimating the maximum temperature of the brakes.

[0065] Consequently, at the end of the preliminary step, the braking parameters are characterized so that they can be used to train the machine learning model. Thus, during the phase of use of the prediction model, the estimation of the maximum temperature reached by the brakes during the landing phase takes into account these different braking parameters which influence the temperature of the brakes during the landing phase. Thus, the training, then the use of the prediction model for the estimation of the maximum temperature of the brakes during the landing phase, are carried out from input data (or values) characteristic of these braking parameters.

[0066] Thus, during a step 101 PHS_E corresponding to the training phase of the automatic learning model for obtaining the prediction model, the latter is trained to associate characteristic values ​​of braking parameters with a maximum measured temperature reached for each landing of a set of landings which serves as a reference. This automatic learning model is for example implemented by an artificial intelligence module MOD 506 of the monitoring device DISP.

[0067] For this, the DISP monitoring device obtains from one or more sensors and / or information systems of the aircraft (for example: ADC system), for each landing of a set of landings called "reference", characteristic values ​​called "reference" of braking parameters and maximum temperature of a brake reached during each landing of the reference landing set. Reference landing is understood to mean a landing for which the values ​​of the braking parameters and maximum temperature of the brakes reached during the landings are standard values ​​serving as reference, accepted and in conformity with a large number of measurements.

[0068] It should be noted that the machine learning model is trained for each landing gear brake independently of the others. This is because a brake may tend to behave differently depending on its side, etc. This is why the machine learning model is trained for each brake independently, and for each aircraft.

[0069] Furthermore, all braking parameters identified during the preliminary step must be available for the flight to be considered valid for consideration by the machine learning model. Otherwise, the flight is considered invalid and the landing data is not taken into account for training the machine learning model and then for using the prediction model.

[0070] Thus, the DISP monitoring device obtains, for each landing gear brake and for each landing of the reference set of landings, the reference characteristic values ​​of: i. Braking parameters such as: braking energy and maximum braking power, brake fan activation time, engine thrust reversal time; landing time corresponding to the time between the moment when the maximum wheel rotation speed associated with the brake is reached and the moment when the maximum brake temperature is reached, initial brake temperature corresponding to the average of the brake temperature values ​​obtained during a predefined time interval corresponding to the start of the landing phase; time between the moment when the maximum brake temperature is reached and the moment when the wheel rotation speed reaches the 95% percentile value, static air temperature; average ground speed; maximum ground speed value; sum of currents applied by a servovalve acting as a hydraulic control device of the braking system;average altitude (i.e., the average altitude during the landing phase relative to sea level); aircraft manufacturer's serial number (MSN); an aircraft engine type identifier (each engine having different thrust reverser powers, for example); an aircraft model identifier (each aircraft model having different aerodynamic characteristics, for example); ii. the maximum brake temperature reached during landing. Note that if the maximum temperature is lower than a predefined minimum temperature, or if the maximum temperature is higher than a predefined maximum temperature, then the flight is considered invalid and the data is not taken into account for training the learning model automatic.

[0071] The reference characteristic values ​​of these braking parameters, as well as the maximum temperature of the brakes are then used as input data to train the machine learning model and thus obtain the prediction model.

[0072] In one embodiment, these input data are used to train a stacking regression machine learning algorithm. Indeed, this type of machine learning algorithm gives excellent performance with an average absolute error of about 15°C (about 7%). This is an ensemble method that combines several models, and which consists of stacking the results of each estimator and using a regressor to calculate the final prediction.

[0073] The machine learning model is therefore trained on the basis of the reference characteristic values ​​of the braking parameters and maximum temperature of the brakes of the reference set of landings in order to obtain the prediction model. This prediction model is then used in a use phase to estimate the maximum temperature of each brake, for a new landing i (z an integer greater than 0).

[0074] Thus, during the use phase (denoted PHS_UT) of the prediction model by the monitoring device DISP, via its artificial intelligence module MOD 506, the monitoring device DISP obtains (in the same way as for the training phase described previously) the values ​​of the braking parameters, as well as the maximum temperature reached by each of the brakes of the landing gear for a new landing i.

[0075] The monitoring device DISP, from the braking parameters described above, via the prediction model, estimates, during a step 102 EST_TEMP, a maximum temperature reached by each brake of a landing gear for this new landing i.

[0076] However, in order to strengthen the estimation of the maximum temperature reached by a brake and to reduce the risk of overfitting, the estimated maximum braking temperature for a brake is calculated as a weighted sum of two prediction models of a brake pair for the same landing gear. For a given brake pair of the landing gear, the weighting factor then defines the weight of each brake in the estimation of the maximum braking temperature. In other words, for a brake X analyzed, the estimated maximum temperature takes into account the maximum temperature estimated, via the prediction model, of another brake Y of the same landing gear.

[0077] Once the maximum temperature is estimated using the prediction model, the DISP monitoring device implements a temperature comparison phase. estimated and measured temperatures comprising steps 103 to 106 described below.

[0078] During a step 103 COMP_TEMP, for each brake, the monitoring device DISP compares the maximum temperature estimated via the use of the prediction model and the maximum temperature measured in real time during landing i.

[0079] More particularly, the monitoring device DISP determines a difference, or an error between the estimated maximum temperature and the maximum temperature measured for landing i. For this, the monitoring device DISP determines: the sliding average over a sliding window of N (N an integer greater than 0) landings of the estimated maximum temperature of an analyzed brake (for example brake X), the sliding average over the sliding window of N landings of the maximum temperature reached by this analyzed brake, the maximum braking temperature measured of another brake (of the brake pair comprising the analyzed brake and another brake, for example brake Y) of the same landing gear to limit the difference between the estimated maximum braking temperature and that measured on the other brake of the same landing gear.

[0080] It should be noted that the maximum measured braking temperature of the other brake is obtained from the sliding average of the estimated maximum temperature of the analyzed brake, the sliding average over the sliding window of N landings of the maximum temperature reached by this other brake and a parameter making it possible to limit the impact of the actual measured temperature for the other brake on the calculation of the error. Without this parameter for limiting the actual measured temperature for the other brake the error would be very high, but only because of the difference between the estimated temperature of the analyzed brake and the measured temperature of the other brake.

[0081] It should be noted that the sliding window of N landings corresponds to a fixed number of N landings preceding the current landing i, for the same aircraft and the same brake. In other words, for each new flight, the last N flights are taken into account in a sliding manner, that is to say that for a flight i+1, the flight i is added to the sliding window of N landings and the oldest flight is removed from this sliding window.

[0082] Then, according to a particular embodiment, the DISP monitoring device determines the synthetic estimated temperature as the weighted sum of the running average of the estimated maximum temperature of brake X and the measured maximum braking temperature of brake Y of the same landing gear. The weighting factor corresponds to the impact of the proportion of brake Y on the estimated maximum braking temperature. From the running average of the estimated maximum temperature of the analyzed brake and the synthetic estimated temperature, the DISP monitoring device determines the mean absolute error (MAE) of the differences between the actual values ​​and the synthetic estimated values, and the symmetrical mean absolute percentage error (SMAPE).

[0083] Finally, the DISP monitoring device determines the error between the estimated temperature and the measured temperature as the weighted sum of the mean absolute error (MAE) and the symmetrical mean absolute percentage error (SMAPE), where the weighting factor is the proportion of absolute and relative errors. Absolute errors facilitate the detection of abnormally high temperatures, while relative errors facilitate the detection of abnormally low temperatures.

[0084] A brake anomaly detection strategy is then applied by the DISP monitoring device, based on the error between the measured maximum temperatures and the estimated maximum temperatures of the analyzed brake.

[0085] Thus, during a step 104 DET_ER, from the calculation of this error, or difference, between the estimated and measured maximum temperatures of the brake analyzed for the landing of a given current flight i, the monitoring device DISP determines whether this error is greater than: - a first predefined threshold SI, when the maximum measured temperature is higher than the estimated maximum temperature; - a second predefined threshold S2, when the maximum measured temperature is lower than the estimated maximum temperature.

[0086] It is thus possible to detect different categories of problems at the level of the brakes (or other equipment of the landing gear) because a maximum temperature measured too high (i.e., the maximum temperature measured is higher than the estimated maximum temperature) or a maximum temperature measured too low (i.e., the maximum temperature measured is lower than the estimated maximum temperature) are representative of different problems. To remedy these problems, maintenance actions or maintenance verifications specific to each problem are typically applied.

[0087] In one example, when there is significant brake wear or piston friction, the brake in question will tend to heat up more than expected (i.e., the measured maximum temperature is higher than the estimated maximum temperature). In another example, when a servo valve or pressure sensor fails, the maximum temperature of the brake can be lower than expected (i.e., the measured maximum temperature is lower than the estimated maximum temperature).

[0088] In another example, a measured maximum temperature lower or higher than the estimated maximum temperature can be representative of a brake temperature sensor that indicates an abnormally very low or conversely very high temperature (“drift” for example).

[0089] If the measured temperature is greater than the predefined threshold SI or the predefined threshold S2, then, during step 105 DET_NER, the monitoring device DISP determines whether during a set of landings comprising the N previous flights of the sliding window and the current flight i, the number, noted NT, total of times where the error is greater than the predefined threshold SI or S2 (as the case may be) is greater than a third predefined threshold S3.

[0090] Thus, when the number NT is greater than the third predefined threshold S3, then the monitoring device DISP, during step 106 G_MSG, generates an alert message for the crew and / or ground personnel. In other words, - an alert message for the brake in question is generated if the maximum measured temperature of the brake is abnormally high, i.e. if the number NT of times the measured temperature of the brake is greater than the estimated maximum temperature plus the predefined threshold SI (for example: 60°C), exceeds the predefined threshold S 3; - an alert message for the brake in question is generated if the maximum measured brake temperature is abnormally low, i.e. if the number NT of times the measured brake temperature is lower than the estimated maximum temperature minus a predefined threshold S2 (for example: 60°C), exceeds the predefined threshold S3.

[0091] This alert message is for example transmitted to the ECAM for display on its human-machine interface for the attention of the crew. Alternatively or additionally, the alert message is transmitted to a ground control system for display on a human-machine interface for the attention of the ground personnel.

[0092] In one embodiment, this alert message is a text message indicating that a brake identified, for example using an identifier, in this alert message has an abnormally high or low temperature and requires maintenance action or verification. In one embodiment, this alert message further comprises an indication of a type of maintenance or verification to be performed.

[0093] In one embodiment, this alert message is adapted according to the different situations described below (i.e., different message depending on the situation encountered). In particular, such an adapted alert message comprises:

[0094] - information on a first category of problem and information on a first category of action or maintenance check to be carried out which is adapted to the first category of problem, when the maximum measured temperature is higher than the estimated maximum temperature and the error between the estimated maximum temperature and the maximum measured temperature is higher than the first predefined threshold SI, or

[0095] - information on a second category of problem and information on a second category of action or maintenance check to be carried out which is adapted to the second category of problem, when the maximum measured temperature is lower than the maximum estimated temperature and the error between the tem estimated maximum temperature and the measured maximum temperature is higher than the second predefined threshold S2.

[0096] In one example, when the error between the estimated maximum temperature and the measured maximum temperature is greater than the first predefined threshold SI, when the measured maximum temperature is greater than the estimated maximum temperature, then the alert message indicates that there is significant wear of the brakes, or friction of the piston, or a problem with a temperature sensor, and that an action of replacement or verification of this equipment must be carried out.

[0097] In another example, when the error between the estimated maximum temperature and the measured maximum temperature is greater than the second predefined threshold S2, when the measured maximum temperature is lower than the estimated maximum temperature, then the alert message indicates that there is a problem with a servovalve or a temperature sensor of the braking system and that an action of repair or verification of this servovalve or the temperature sensor must be carried out.

[0098] Only one alert message for one brake and per landing gear is triggered at a time to avoid alerts for the same problem. An anomaly in one brake (especially in the case of a brake whose temperature is lower than the estimated temperature) can have an impact on the normal braking behavior of another brake of the same landing gear.

[0099] [Fig.3] represents in graphic form an example of a result obtained after implementation of the monitoring method, according to one embodiment.

[0100] In this example, the estimated average maximum values ​​of the temperature of brakes numbers 3 and 4 (denoted respectively Temp_est_3 and Temp_est_4), as well as the actual, measured values ​​of the temperature of brakes 3 and 4 (denoted respectively Temp_mes_3 and Temp_mes_4) are represented for the same aircraft and the same landing gear (see top graph of [Fig.3]).

[0101] The average error between the estimated maximum temperature and the maximum temperature measured for brakes 3 and 4 (denoted respectively Err_moy_3 and Err_moy_4) is also represented (see bottom graph of [Fig.3]).

[0102] In addition to the temperature asymmetry between brakes 3 and 4, the prediction model accurately estimates the maximum temperatures of brakes 3 and 4 (COMP_N).

[0103] By using the prediction model described above, the DISP monitoring device is able to anticipate a temperature asymmetry between the estimated maximum temperature and the measured maximum temperature for each of the landing gear brakes independently and therefore to identify precisely which brake requires maintenance action or verification.

[0104] The temperature error, or deviation, between the estimated and measured maximum temperatures of the brake 4 continues to increase (ASY) until an alert message is generated. In other words, when the number of times the measured maximum temperature is higher than the estimated maximum temperature of the predefined threshold SI (e.g. 60°C) over a plurality of landings is higher than the predefined threshold S3, then the monitoring device DISP generates an alert message.

[0105] On the contrary, the error, or difference, in temperature between the maximum temperatures estimated and measured for the brake 3 is low, i.e. lower than the predefined threshold SI.

[0106] Thus, the DISP monitoring device is capable of accurately detecting which brake requires maintenance action or check. In this example, brake 4 is the brake that requires the maintenance action or check.

[0107] Once the maintenance action or check has been carried out, the error, or deviation, between the estimated and measured maximum temperatures tends towards low values ​​(COMP_N).

[0108] [Fig.4] represents in graphic form another example of the result obtained after implementation of the monitoring method, according to one embodiment.

[0109] Before the maintenance action (ASY), the measured brake 1 temperature tends to be higher than the estimated maximum temperature for this brake. This is probably an oxidation effect during braking, which causes brake 1 to heat up more than expected. After the brake replacement (COMP_N), the prediction model accurately predicts the maximum braking temperature at each landing corresponding to the maximum temperature under nominal conditions.

[0110] Thus, the DISP monitoring device can, through the use of the prediction model, detect most of the temperature asymmetry events between an estimated maximum temperature and a measured maximum temperature with a confidence rate of 100%.

[0111] The DISP monitoring device, via the prediction model, is able to also detect faults requiring maintenance such as: brake wear, wheel friction and sensor problems. For example, it is possible to detect:

[0112] - brake wear or wheel friction;

[0113] - sensor problems;

[0114] - cases of abnormally cold brakes, for example linked to a servovalve or a sensor faulty pressure.

[0115] [Fig. 5] schematically illustrates an example of hardware architecture of the DISP monitoring device, which then comprises, connected by a communication bus 510: a processor or CPU (“Central Processing Unit” in English) 501; a RAM (“Random Access Memory” in English) 502; a ROM (“Read Only Memory” in English) 503, for example a Flash memory; a data storage device, such as a hard disk drive (HDD) or a storage media reader, such as an SD card reader (Secure Digital) 504; at least one communication interface 505 allowing the DISP monitoring device to interact with different sensors and / or avionics systems.

[0116] The processor 501 is capable of executing instructions loaded into the RAM 502 from the ROM 503, an external memory (not shown), a storage medium, such as an SD card, or a communication network (not shown). When the DISP monitoring device is powered on, the processor 501 is capable of reading instructions from the RAM 502 and executing them. These instructions form a computer program causing the processor 501 to implement the behaviors, steps and algorithm described herein.

[0117] In one embodiment, the monitoring device DISP further comprises an artificial intelligence module MOD 506 configured to implement a machine learning model during the training phase PHS_E, then for the use of the prediction model during the use phase PHS_UT as described herein.

[0118] In a variant, the DISP monitoring device comprises the artificial intelligence module MOD 506 configured to use the prediction model during the PHS_UT use phase as described here. This artificial intelligence module MOD 506 being previously trained during the PHS_E training phase in a device different from the DISP monitoring device.

[0119] All or part of the behaviors, steps and the algorithm described herein may thus be implemented in software form by executing a set of instructions by a programmable machine, such as a DSP (Digital Signal Processor) or a microcontroller, or be implemented in hardware form by a machine or a dedicated component (chip) or a set of components (chipset), such as an FPGA (Field-Programmable Gate Array) or an ASIC (Application-Specific Integrated Circuit). Generally speaking, the DISP monitoring device comprises electronic circuitry arranged and configured to implement the behaviors, steps and the algorithm described herein.

[0120] In an exemplary embodiment, the DISP monitoring device may be implemented in parallel with a BTSM, to provide enhanced functionality and / or redundancy.

[0121] In an exemplary embodiment, the DISP monitoring device may be integrated into an avionics system of the aircraft 600 described in connection with [Fig. 6], or may be connected to an avionics system in any suitable manner, such that the DISP monitoring device may communicate estimated values ​​of tem brake temperature to the avionics system of the aircraft 600. For example, the DISP monitoring device may be integrated or connected to a controller of a BTMS of the aircraft 600.

[0122] [Fig.6] schematically illustrates, in side view, an aircraft 600 equipped with a DISP monitoring device 601, according to one embodiment. According to the embodiment of [Fig.6], the DISP monitoring device 601 belongs to the avionics system of the aircraft 600.

[0123] In other examples, the DISP monitoring device may be completely independent of any onboard system of the aircraft 600. In these examples, the DISP monitoring device may be part of an offboard system, such as a portable maintenance device, which may or may not be capable of communicating with the onboard systems of the aircraft 600, or it may comprise a separate onboard system. In these examples, the DISP monitoring device is provided with suitable means for receiving control commands and / or for providing estimated temperature values, such as a display or user interface.

Claims

Claims

1. Method for monitoring a maximum temperature reached during a landing by a brake of a landing gear of an aircraft, the method being implemented by a monitoring device (DISP), said method comprising: a usage phase (PHS_UT) of a model for predicting a maximum temperature reached by the brake during landing comprising the following steps - obtaining a current set of values ​​of a plurality of braking parameters, for a current landing; - estimate (102), using the prediction model, a maximum temperature during landing which should be reached by said brake, from the values ​​of said current set; said method further comprising a comparison phase comprising: - obtain a maximum measured temperature reached by the brake during the current landing; - determining (104) whether an error between the estimated maximum temperature and the measured maximum temperature is greater than a first predefined threshold SI, when the measured maximum temperature is greater than the estimated maximum temperature, or a second predefined threshold S2, when the measured maximum temperature is lower than the estimated maximum temperature; - when the error is greater than the first predefined threshold SI or the second predefined threshold S2, determining (105) whether, on a set of landings comprising a number N of landings of a sliding window and the current landing, a total number of errors (NT) is greater than a third predefined threshold S3, otherwise repeating the phase of use of the prediction model and the comparison phase for a following landing; - when the total number of errors (NT) is greater than the third predefined threshold S3, generate (106) an alert message, otherwise repeat the use phase (PHS_UT) of the prediction model and the comparison phase for a following landing.

2. Method according to claim 1, further comprising, prior to the use phase (PHS_UT) of the prediction model: a training phase (PHS_E) of a machine learning model comprising: associating a set of reference values ​​of braking parameters which was obtained for each landing of a set of reference landings with a reference value of maximum temperature which was reached by the landing gear brake during the landing in question.

3. A monitoring method according to claims 1 or 2, wherein estimating the maximum temperature reached by the brake for the current landing, from the values ​​of the current set of braking parameters, further comprises: calculating a weighted sum between the estimated maximum temperature for the brake and an estimated maximum temperature for another brake of the landing gear.

4. A monitoring method according to claim 3, further comprising determining a synthetic estimated temperature as the weighted sum of a running average, over the running window of N landings, of the estimated maximum brake temperature and a measured temperature of the other landing gear brake.

5. Monitoring method according to claim 4, in which the error between the estimated maximum temperature and the measured maximum temperature is calculated as a weighted sum between a mean absolute error (MAE) between a sliding average, over the sliding window of N landings, of the maximum temperature reached by the brake and the synthetic estimated temperature, and a symmetrical mean absolute percentage error (SMAPE) between the sliding average, over the sliding window of N landings, of the maximum temperature reached by the brake and the synthetic estimated temperature.

6. A monitoring method according to claims 1 to 5, wherein the braking parameters are among: a braking energy, a maximum braking power, a duration of activation of the brake fans, a duration of reversal of the thrust of the engines, a duration of the landing, an initial temperature of the brake, a duration between a moment when the maximum temperature of the brake is reached and a moment when a rotational speed of the wheel reaches the value of the 95% percentile, a static air temperature, an average of the ground speed, a maximum value of the ground speed, a sum of the currents applied by a servovalve of the braking system, an average altitude, a manufacturer serial number of the aircraft, an engine type identifier of the aircraft, an aircraft model identifier.

7. A monitoring method according to claims 1 to 6, wherein the alert message includes an indication that a maintenance action or check needs to be performed on the brake, and an indication of a type of maintenance action or check to be performed.

8. Monitoring method according to claim 7, further comprising adapting the alert message by indicating: - information on a first category of problem and information on a first category of action or maintenance check to be carried out which is adapted to the first category of problem, when the maximum temperature measured is higher than the estimated maximum temperature and the error between the estimated maximum temperature and the maximum temperature measured is higher than the first predefined threshold SI, or - information on a second category of problem and information on a second category of action or maintenance check to be carried out which is adapted to the second category of problem,when the measured maximum temperature is lower than the estimated maximum temperature and the error between the estimated maximum temperature and the measured maximum temperature is greater than the second predefined threshold S2.,

9. A device (DISP) for monitoring a maximum temperature reached during a landing by a brake of a landing gear of an aircraft, the monitoring device (DISP) comprising electronic circuitry configured to implement: a phase of use (PHS_UT) of a model for predicting a maximum temperature reached by the brake during the landing comprising the following steps - obtaining a current set of values ​​of a plurality of braking parameters, for a current landing; - estimating (102), using the prediction model, a maximum temperature during the landing which should be reached by said brake, from the values ​​of said current set; said method further comprising a comparison phase comprising: - obtaining a measured maximum temperature reached by the brake during the current landing;- determining (104) whether an error between the estimated maximum temperature and the measured maximum temperature is greater than a first predefined threshold SI, when the measured maximum temperature is greater than; the estimated maximum temperature, or at a second predefined threshold S2, when the measured maximum temperature is lower than the estimated maximum temperature; - when the error is greater than the first predefined threshold SI or the second predefined threshold S2, determining (105) whether, on a set of landings comprising a number N of landings of a sliding window and the current landing, a total number of errors (NT) is greater than a third predefined threshold S3, otherwise repeating the phase of use of the prediction model and the comparison phase for a following landing; - when the total number of errors (NT) is greater than the third predefined threshold S3, generate (106) an alert message, otherwise repeat the use phase (PHS_UT) of the prediction model and the comparison phase for a following landing.

10. Aircraft (600) comprising a monitoring device (DISP) according to claim 9.

Images