Device and method for monitoring the operation of an aircraft engine oil circuit

Abstract

The invention relates to a method for monitoring an oil pressure in an oil circuit of an aircraft engine on the basis of a previously generated normalisation model, the monitoring method including the steps of: - obtaining at least one sample of current values, each sample including a current oil temperature value associated with a current oil pressure value and with a current high-pressure engine speed value, which values are acquired during a flight of the aircraft; - for each sample, providing the current oil temperature, oil pressure and high-pressure engine speed values to the normalisation model so as to obtain a normalised oil pressure value for the sample; and - comparing the normalised oil pressure value(s) with at least one threshold beyond or below which the oil pressure in the oil circuit of the engine is considered to be abnormal.

Description

[0001] Description

[0002] Title of the invention: Device and method for monitoring the operation of an aircraft engine oil circuit

[0003] Technical Field

[0004] The invention relates to the general field of aeronautics and in particular to the monitoring of the operating status of aircraft. More specifically, the invention targets the field of monitoring the oil circuits of aircraft engines, for example turbomachinery.

[0005] Previous technique

[0006] Before being put into service, aircraft engines are subjected to a series of tests designed to validate their entry into service.

[0007] Furthermore, during their operation, aircraft engines are regularly subjected to tests in order to assess the need to carry out certain maintenance operations.

[0008] Among the tests regularly performed, a corrected oil pressure (COP) measurement test must be validated to allow the engine to be put into or put back into service. Corrected oil pressure is an indicator used to assess the health of an engine's oil circuit. This indicator is calculated from the following measurements: oil supply pressure (EOP), oil temperature (EOT), and the high-pressure core speed (N2). This COP indicator corresponds to the oil supply pressure made independent of the test conditions, via the following formula:

[0009] [Math. 1]

[0010] HAIR corrigée = EOP + aN2 + f>EOT + y

[0011] with EOP in psid (pounds per square inch differential), N2 in revolutions per minute (rpm), and EOT in °C, a, 6, and y are numerical coefficients. The pound per square inch (PSI) is a unit of measurement corresponding to the result of one pound-force applied to a surface area of ​​one square inch. A suffix is ​​often placed after PSI to indicate the zero reference measurement: PSIA for absolute pressure, PSIG for barometric pressure, and PSID for differential pressure. Since the feed pump is a positive displacement type, its output flow rate is, to a first approximation, proportional to its rotational speed, i.e., N2. Furthermore, this POILC indicator allows for consideration of the influence of ambient temperature by taking into account the oil temperature measurement EOT in the calculation of the corrected oil pressure. This indicator is sensitive to the environment in which the EOT, EOP, and N2 measurements are acquired.

[0012] If a discrepancy is found between the POILC indicator and the applicable thresholds, a study is conducted to identify the cause of this failure.

[0013] When an engine is to be subjected to a chain of tests and / or maintenance operations, in which the tests and operations are carried out in a particular predetermined order and a test is not validated, the engine may have to be subjected again to all or part of the tests and operations carried out previously after corrective maintenance operations have been implemented.

[0014] The test and operations chain typically includes:

[0015] - an inspection to confirm the scope of the work,

[0016] - a partial or total engine dismantling, depending on the need.

[0017] - Cleaning of parts for inspection

[0018] - an inspection of the parts to allow comparison of their characteristics with operating limits,

[0019] - logistical management of components for engine reassembly,

[0020] - reassembly of the engine,

[0021] - a verification of engine performance by means of a test bench test.

[0022] Verification that the corrected oil pressure meets the applicable thresholds is performed at the end of the line, during testing on the test bench. When the corrected oil pressure does not meet the applicable thresholds, an airflow check is performed. This airflow check aims to verify the impermeability of the oil supply and recovery circuits to identify the presence or absence of a fault.

[0023] However, any fault detected via the corrected oil pressure test directly leads to increased costs and reduced availability, not only for the engines but also for the test benches. Indeed, detecting a fault during the POILC test results in increased engine downtime, potentially lasting several days, known as "Turnaround Time" (TAT), and increased resource costs, particularly for identifying and correcting the cause of the fault. Furthermore, the increased downtime generally leads to delays in the delivery of the engine, as well as other engines scheduled for testing, due to the test bench being occupied by the faulty engine.

[0024] Furthermore, a major drawback is the potential for a "workscope escalation" procedure if corrected oil pressure and permeability tests reveal an abnormally high oil level. Such an escalation procedure significantly increases the duration of maintenance actions, requiring a new engine test after the detected fault has been corrected. This considerably lengthens not only the overall maintenance time but, more importantly, the test bench downtime.

[0025] To limit unnecessary repetition of maintenance actions, and especially to minimize test bench occupancy, permeability tests can be systematically performed to anticipate the result of a high corrected oil pressure during the test bench test. However, these tests, carried out systematically before the corrected oil pressure test, unnecessarily prolong maintenance for the vast majority of engines that are already well within the corrected oil pressure limits.

[0026] There is therefore a real need to optimize downtime. There is also a need for better visibility to limit the impact of potential delivery delays. More generally, there is a need to improve the oil circuit monitoring procedure.

[0027] The invention aims to meet all or part of these needs.

[0028] Description of the invention

[0029] The invention relates to a method for generating a normalization model of an oil pressure value contained in an oil circuit of an aircraft engine, the method comprising the following steps:

[0030] - obtaining at least one data sample, preferably a set of data samples, each data sample comprising an EOT oil temperature value associated with an EOP oil pressure value and a high pressure N2 regime value acquired during a flight;

[0031] - training at least one prediction model from at least a part of the data sample(s), so that once trained the trained prediction model is able to determine a predicted value of oil pressure from oil temperature and high pressure speed values ​​supplied as input to said prediction model;

[0032] - parameterization of the normalization model from the prediction model, the normalization model being parameterized so as to provide at least one normalized oil pressure value as output from at least the current values ​​of oil pressure, oil temperature and high-pressure speed provided as input to said normalization module,

[0033] the normalized oil pressure value being determined from at least one difference between the current oil pressure value and a predicted oil pressure value calculated from the prediction model and current oil temperature and high pressure speed values.

[0034] The method according to the invention allows the determination of a standardized oil pressure value, from current values ​​conventionally measured by means of sensors on board during aircraft flights. The method according to the invention makes it possible in particular to estimate an oil pressure from flight data, in other words from data acquired when the engine is operating in flight, unlike the corrected oil pressure which proves to be unusable when calculated from flight data because it is too noisy.

[0035] Furthermore, the invention allows for the determination of a standardized indicator that is comparable not only from one flight to another, but also from one engine to another. The invention relates to the determination of a new oil pressure indicator that does not require the engine to be immobilized, and in particular does not require testing on a test bench.

[0036] By "oil temperature EOT associated with an oil pressure value EOP and a high pressure regime value N2", we mean values ​​acquired at the same instant during a flight, relating to the same aircraft engine oil circuit.

[0037] Preferably, the data sample values ​​are acquired during a flight phase in which engine operation is stable, for example, during a cruise phase or a taxi phase (also called a ground roll phase). Preferably, the data samples are acquired during a cruise phase.

[0038] The flight phases classically include: taxiing phase, takeoff phase, climb phase, cruise phase, descent phase, approach and landing phase and post-landing phase.

[0039] Preferably, several data samples are obtained, the samples being able to relate to a plurality of flights carried out with different engines or types of engines.

[0040] The training step for at least one prediction model may include, for each engine or engine type, training a prediction model specific to said engine or engine type from the data sample(s) relating to flights carried out with said engine or engine type. A prediction model specific to each engine or engine type can thus be obtained, potentially improving the prediction accuracy of said prediction model(s).

[0041] The training of the prediction model aims to enable the determination of a prediction function linking oil pressure values ​​to oil temperature values ​​and high-pressure engine speed.

[0042] The prediction function can be an affine function.

[0043] The prediction function can take the following form:

[0044] [Math. 2]

[0045] EOP pred = f P red^EOT, 7V2) = kl * N2 + k2 * EOT + k3

[0046] with kl, k2, k3 being three numerical coefficients, EOT and N2 the input values ​​respectively relating to an oil temperature and an associated high-pressure regime, and EOPpred the output corresponding to a predicted oil pressure value. The coefficients kl, k2, k3 can in particular be obtained by optimizing the equation:

[0047] [Math. 3]

[0048] EOP hist = f P red E0T hist ,N2 hist ) = kl * N2 hist + k2 * EOT hist + k3

[0049] with EOPhist, EOThist and N2hist vectors grouping the values ​​of the data samples.

[0050] The normalization model can be configured to allow, once configured, the calculation of the normalized oil pressure value, notably through the application of the following formula:

[0051] [Math. 4]

[0052] EOPnorm EOP cour fpred(J^ Tcour> 2 cour

[0053] Preferably, the generation process also includes, before the parameterization step of said normalization model, the following steps:

[0054] - determination of a reference value for oil temperature and a reference value for high pressure operating speed by statistical analysis of at least a part of the oil temperature values ​​and the high pressure operating speed values ​​of the data sample set, respectively, - after the training step of at least one prediction model, deduction of at least one reference value for oil pressure by providing the reference values ​​for oil temperature and high pressure operating speed to the prediction model;

[0055] the parameterization of said normalization model being further carried out from the reference value of oil pressure and the normalized value of oil pressure being determined from this said reference value of oil pressure.

[0056] In particular, the normalization model can be configured to allow, once configured, the calculation of the normalized oil pressure value from the application of the following formula:

[0057] [Math. 5]

[0058] EOP nO rm ~ EOP re f + (EOP cour fpred EOT cour , N2 cour y)

[0059] In particular embodiments of the invention, the oil pressure reference value is specific to an engine or engine type, the oil pressure reference value being calculated from a prediction model specific to said engine or engine type, respectively.

[0060] The invention also relates to a method for monitoring oil pressure contained in an oil circuit of an aircraft engine from a normalization model generated by means of a generation method according to the invention, the monitoring method comprising the following steps:

[0061] - obtaining at least one sample of current values, each sample including a current EOT oil temperature value COU r associated with a current oil pressure value EOPcour and a current high-pressure regime value N2 CO ur acquired during a flight of said aircraft;

[0062] - for each sample, providing current values ​​of oil temperature, oil pressure, and high pressure regime to the normalization model so as to obtain a normalized value of oil pressure for said sample;

[0063] - comparison of the normalised oil pressure value(s) to at least a threshold above or below which the oil pressure contained in the engine oil circuit is considered to be abnormal.

[0064] The monitoring method according to the invention makes it possible to obtain an estimate of a normalized oil pressure, thus allowing the result of a corrected oil pressure test to be anticipated on an engine test bench. Therefore, the method according to the invention provides an indicator that can be implemented systematically, without requiring the engine to be immobilized and without lengthening maintenance procedures. The method according to the invention can advantageously replace the systematic implementation of the permeability test, which can then be carried out only when absolutely necessary.

[0065] In particular embodiments of the invention, where the normalization model includes the determination of a prediction model specific to an engine or engine type, the monitoring method includes providing the normalization model of the engine or engine type, so as to allow the use of a prediction model specific to said engine or engine type in the calculation of the normalized pressure value.

[0066] In particular embodiments of the invention, several thresholds are predetermined, defining varying degrees of failure. Preferably, a plurality of samples is obtained, with comparison to at least one threshold being performed by comparing an average of the standardized oil pressure values ​​to at least one threshold, preferably the average being a moving average calculated over a predefined number of standardized values.

[0067] This step allows for a smoothing of values ​​over time.

[0068] The plurality of samples may correspond to all or part of the flights carried out since the engine was put into service and / or since the last maintenance operation of the oil circuit.

[0069] The engine of the aircraft in question can be identified as being at risk if, after a predetermined number of flights, the oil pressure in the engine's oil circuit is considered abnormal. In other words, an alert can be triggered if, after a predetermined number of flights, the oil pressure in the engine's oil circuit is considered abnormal. This can advantageously limit the detection of false positives. The alert may inform an operator that the results of the corrected oil pressure and permeability tests applied to the engine are likely to be negative, thus revealing an abnormally high oil level. This alert allows for the anticipation of a possible extension of maintenance operations.

[0070] The methods according to the invention may also include a step of verifying the samples obtained, said verification step comprising, for each sample in the set of data samples

[0071] - comparing the oil temperature value to an initial range of reference values ​​to identify whether the oil temperature value is abnormal,

[0072] - comparing the oil pressure value to a second range of reference values ​​to identify if the oil pressure value is abnormal,

[0073] - comparing the high oil pressure speed value to a third range of reference values ​​in order to identify whether the high oil pressure speed value is abnormal,

[0074] - the removal of the sample if at least one of the values ​​of said sample is identified as abnormal.

[0075] For example, the first interval is [20°C; 250°C], the second interval is [20psid; 272psid], and the third interval is [60%; 117%]. Preferably, the first interval is [40°C; 180°C], the second interval is [40psid;

[0076] 130psid], and the third interval is [90%; 110%].

[0077] In general, the methods according to the invention may include verifying that the set of samples does not contain duplicates, missing, aberrant or corrupted values, or that they were acquired during a particular flight phase. All or part of the steps, preferably all the steps, of the methods according to the invention may be implemented by computer.

[0078] Thus, the invention also relates to a computer program comprising code instructions which, when implemented, allow the execution of the steps of a process according to the invention.

[0079] The invention also relates to a recording medium readable by means of a computer comprising a computer program.

[0080] The aforementioned features and advantages, as well as others, will become apparent upon reading the detailed description that follows. This detailed description refers to the attached drawings.

[0081] Brief description of the drawings

[0082] The attached drawings are schematic and are primarily intended to illustrate the principles of the presentation.

[0083] In these drawings, from one figure to another, identical elements (or parts of elements) are identified by the same reference symbols.

[0084] [Fig. 1] Figure 1 illustrates an example of the implementation of a generation process according to the invention,

[0085] [Fig. 2] Figure 2 represents an example of the implementation of a monitoring method according to the invention,

[0086] [Fig. 3] Figure 3 represents another example of the implementation of a monitoring method according to the invention,

[0087] [Fig. 4] Figure 4 represents an example of thresholds,

[0088] [Fig. 5] Figure 5 is an example of results obtained by implementing a monitoring method according to the invention, [Fig. 6] Figure 6 schematically illustrates a maintenance procedure conventionally implemented (prior art),

[0089] [Fig. 7] Figure 7 schematically illustrates the use of a monitoring method according to the invention in a maintenance procedure,

[0090] [Fig. 8] Figure 8 illustrates a generation device according to the invention,

[0091] [Fig. 9] Figure 9 illustrates a monitoring device according to the invention,

[0092] [Fig. 10] Figure 10 represents a physical architecture of a device according to the invention,

[0093] [Fig. 11] Figure 11 schematically represents an oil circuit of a turbomachine.

[0094] Description of the implementation methods

[0095] To make the explanation more concrete, examples of the implementation of the generation and monitoring processes are described in detail below, with reference to the attached drawings. It should be noted that the invention is not limited to this example.

[0096] Figure 1 represents a process for generating 100 of a MODnorm normalization model of an oil pressure value contained in an oil circuit of an aircraft engine.

[0097] Figure 11 illustrates an example of such an oil circuit 5 of a turbomachine 4. The oil circuit 4 includes a reservoir 10, a lubrication unit 12 and a supply circuit 14. A supply 16 feeds the oil circuit.

[0098] The 100 generation process includes:

[0099] - a step El 10 of obtaining one or more samples {xi, ..., Xh, ..., XM} of data;

[0100] - a step E120 of training one or more MODpred prediction models, from all or part of the data samples obtained in step El 10;

[0101] - a step E150 of parameterizing a MODnorm normalization model from the prediction model(s). Each sample Xh includes at least one oil temperature value EOTh associated with an oil pressure value EOPh and a high pressure regime value N2h acquired at the same time during a flight.

[0102] Oil temperature, oil pressure, and high-pressure operating speed values ​​for the sample set can be acquired during one or more flights, for one or more engines or engine types or aircraft types. These values ​​are acquired using onboard sensors.

[0103] In particular embodiments, each sample may include information relating to the engine or engine type or to a particular type of aircraft.

[0104] Thus, the set of samples obtained may be specific to a particular engine, engine type, or aircraft type. Alternatively, the set of samples obtained may include data relating to different engines, engine types, or aircraft types.

[0105] Preferably, the samples include, for each engine, at least one sample per flight carried out since the last maintenance operation, or since the engine was put into service.

[0106] In some embodiments, the samples can be obtained from flight reports, called "Post Flight Summary (PFS)," which are typically generated and recorded after each flight. At least one sample, and in particular plus or minus eight samples, can be obtained per flight, preferably acquired during the cruise phase. These flight reports can be recorded, for example, in a database from which the data are retrieved in step El 10. In some embodiments, step El 10 involves acquiring the sample values ​​using onboard sensors 18, 20. A sensor 20 can be connected to the oil circuit, and in particular to the supply circuit 14, and configured to allow the acquisition of oil temperature and oil pressure values. A sensor 18 can be connected to a gearbox 22 and configured to allow the acquisition of high-pressure speed values.Preferably, the samples contain values ​​acquired during a relatively stable engine flight phase, from a thermal perspective and at high engine speed and high pressure, for example, during cruise or taxiing. Thus, all samples were acquired under substantially similar conditions, notably those applied on the test bench for calculating the corrected oil pressure. Furthermore, this allows for a standardized oil pressure indicator that is comparable from one flight to another, and potentially from one engine to another.

[0107] Data samples can be preprocessed in such a way as:

[0108] - to remove duplicates, outliers, corrupted samples, etc. - to convert samples into a more suitable format, particularly for statistical and computational processing.

[0109] - to be sorted, to facilitate the subsequent use of the samples.

[0110] Any commonly used preprocessing method can be employed. In particular, when the samples contain values ​​acquired during a cruise phase, a preprocessing step can be performed to verify that the values ​​in each data sample are consistent with such a cruise phase.

[0111] Similarly, when the samples contain values ​​acquired for a taxi phase, a preprocessing step can be performed to verify that the values ​​in each data sample are consistent with such a taxi phase. This can be applied to any flight phase.

[0112] To verify that the values ​​are compatible with a flight phase, the following steps can be implemented for each sample:

[0113] - comparison of the oil temperature value to an initial range of reference values ​​in order to identify whether the oil temperature value is abnormal for said flight phase,

[0114] - comparison of the oil pressure value to a second range of reference values ​​in order to identify whether the oil pressure value is abnormal for said flight phase,

[0115] - comparison of the high oil pressure speed value to a third range of reference values ​​in order to identify if the high oil pressure speed value is abnormal for said flight phase.

[0116] When at least one value in a sample is outside the range, the sample is preferably discarded.

[0117] For a cruising phase, the first, second and third intervals can correspond respectively to the following intervals: [20°C;250°C],

[0118] [20psid;272psid], and [60%; 117%], or even at the following intervals: [40°C;180°C], [40psid;130psid], and [90%; 110%].

[0119] The data samples preferably consist only of samples containing values ​​acquired in flight during operation.

[0120] In specific embodiments, the MODnorm normalization model can be parameterized to be applied specifically according to the engine, engine type, and / or aircraft type. Thus, the normalized oil pressure can be determined differently depending on the engine, engine type, and / or aircraft type.

[0121] In particular embodiments, the standardization model is identical regardless of the engine, engine type or aircraft type.

[0122] In the following sections, the generation process is described in order to generate a parameterized normalization model that allows for the determination of a standardized oil pressure specific to each engine. A predictive model can then be determined for each engine.

[0123] The EOP oil pressure differs from one engine to another and from one flight to another, being strongly correlated with the operating temperature of the EOT oil and the rotational speed of the high-pressure shaft N2, the drive(s) of the prediction model(s) make it possible to link these values ​​together, from historical data, based on experience.

[0124] The prediction model(s) can be regularly updated from the continuous acquisition and recording of samples over time. Preferably, each prediction model is obtained by statistical analysis and in particular by implementing linear regression.

[0125] An example of the implementation of an E120 step is detailed below.

[0126] In this example, for each engine, a prediction model is determined from the sample(s) relating to said engine.

[0127] Each prediction model can be determined by optimization, by estimating the three coefficients k1, k2 and k3 satisfying the relation:

[0128] [Math. 6]

[0129] EOP hist = f P red .E0T hist ,N2 hist ) = kl * N2 hist + k2 * EOT hist + k3

[0130] with EOPhist = (EOPi, ..., EOPh, ..., EOPM) a vector grouping the pressure values ​​of the sample(s) relating to said engine; N2hist = (N2i, ..., N2h, ..., N2M) a vector grouping the high pressure engine speed values ​​of the sample(s) relating to said engine; EOThist = (EOTi, ..., EOTh, ..., EOTM) a vector grouping the temperature values ​​of the sample(s) relating to said engine.

[0131] The generation process according to the invention may include the optional but preferred steps E130 and E140:

[0132] - determination of an EOT oil temperature reference value re f and a high-pressure regime reference value N2 re f by statistical analysis, respectively, of at least a portion of the EOThist oil temperature values ​​and the N2hist high-pressure regime values ​​from the data sample set, and

[0133] - deduction of at least one reference value of oil pressure EOPref by providing the reference values ​​of oil temperature EOTref and high-pressure operating temperature N2 re f to the MOD prediction model pre d.

[0134] The reference values ​​for temperature and high-pressure regime determined in step E130 may correspond to average or median values.

[0135] The reference values ​​for temperature and high-pressure regime can be determined by:

[0136] - defining, for each engine, a representative value of the depression values ​​acquired during flights for said engine, the representative value being a median or average value;

[0137] - selecting, for each engine, the data sample having the EOP oil pressure value closest to the representative value defined for said engine;

[0138] - calculating the average or median of the temperature and high-pressure regime values ​​from said selected samples.

[0139] For example, a reference temperature value might be 91°C and a reference high-pressure regime value might be 100%. This example is provided for illustrative purposes only.

[0140] In step E140, an oil pressure reference value EOPref can be deduced for each engine, from the engine-specific prediction model, in other words, determined using samples relating to that engine, to which the reference temperature values ​​EOT are provided. re f and high-pressure N2 regime re f:

[0141] [Math. 7]

[0142] EOP re f = klN2 re f + k2EOT re f + k3

[0143] At step E150, the normalization model is parameterized.

[0144] The MODnorm normalization model is parameterized based on the MOD prediction model(s). pre d, the normalization model being parameterized so as to provide as output a normalized value of oil pressure EOPnorm from current values ​​of oil pressure EOPcour, oil temperature EOTcour and high pressure N2 CO ur provided as input to said normalization module.

[0145] The normalized oil pressure value EOPnorm is determined from at least one difference between the current oil pressure value EOPcour and a predicted oil pressure value EOPpred calculated from at least one prediction model MOD pre d and current EOT oil temperature values COU r and high-pressure N2 regime C0Ur. The normalization model can be parameterized to provide a normalized oil pressure value that is engine-dependent. In this case, the prediction model used to calculate the predicted oil pressure value is engine-dependent, with the engine serving as an input value for the normalization model. Alternatively, multiple normalization models can be parameterized, each specific to an engine, and the process could then involve selecting a specific normalization model based on the engine.

[0146] The normalization model can be configured to allow, once configured, the calculation of the normalized oil pressure value by applying the following formula:

[0147] [Math. 8]

[0148] EOP norm — EOP re + (EOP cour — fpred( OT cour , W2 cour ))

[0149] The parameterization of the normalization model may also include a step of determining one or more thresholds defining limits beyond or below which the normalized oil pressure values ​​are considered abnormal.

[0150] Several thresholds can be determined, defining degrees of failure of varying degrees.

[0151] The thresholds can be specific to an engine.

[0152] The threshold(s) can be predetermined based on reference values, particularly the EOPref oil pressure reference value(s). Specifically, a threshold can be a range of values ​​beyond or below which the normalized oil pressure value is considered abnormal.

[0153] The threshold(s) may include all or part of the following intervals as illustrated in Figure 4: [EOP re f-3o; EOP r ef+3o], [EOP ref-4o; EOP re f-3o], [EOPref+3o ; EOPref+4o], where o is a standard deviation value calculated from oil pressure values ​​of the samples, specifically only from the oil pressure values ​​of the samples selected and used to determine the reference values ​​for temperature and high-pressure operating speed. In the example shown in Figure 4, a first threshold corresponds to the interval Io=[EOPref-3o ; EOPref+3o], shown in solid gray. A normalized value within this interval is indicative of an engine considered healthy. A second threshold corresponds to the union of the intervals h and i; [EOP re f-4o ; EOPref- 3o]U[EOP re f+3o; EOP re f+4o], shown in hatching. A normalized value within this interval IiUI'i indicates an engine considered to be at risk. A normalized value within none of the intervals lo, Ii, and l'i indicates an engine considered to be faulty.

[0154] The generation process thus provides a normalization model enabling the calculation of an oil pressure monitoring indicator in an aircraft engine, without the need to immobilize the engine and without the need for measurements on a test bench.

[0155] The parameterized normalization model can be stored in a monitoring device 2, comprising for example a computer program including code instructions which, when implemented, allow the execution of the steps of a monitoring process 200 according to the invention, as described below.

[0156] Correspondingly, the invention relates to a device 1 for generating a MODnorm normalization model of an oil pressure value contained in an oil circuit of an aircraft engine, illustrated in figure 8, the device 1 comprising: - a module M10 for obtaining at least one data sample configured to implement a step El 10 of a generation process according to the invention;

[0157] - an M12 training module of at least one prediction model configured to implement an E120 step of a generation process according to the invention;

[0158] - Optionally, modules M14 and M16 for determining reference values ​​configured to implement steps E130 and E140, respectively, of a generation process according to the invention; - a module M18 for parameterizing the normalization model configured to implement step E150 of a generation process according to the invention. The generation device may be a computer program or a storage medium for such a program.

[0159] The invention also relates to a monitoring method 200 based on the use of a normalization model generated by means of a generation method according to the invention. Examples of the implementation of such a method are shown in Figures 2 and 3.

[0160] The said monitoring method is applied to an aircraft engine.

[0161] Method 200 for monitoring oil pressure in an aircraft engine oil circuit from a MODnorm standardization model includes:

[0162] - a step E210 of obtaining a sample Xi of current values ​​acquired during a flight of said aircraft;

[0163] - a step E220 of supplying current values ​​to the MODnorm normalization model so as to obtain a normalised oil pressure value EOPnorm; - a step E240 of comparing the normalised oil pressure value to at least a threshold beyond or below which the oil pressure contained in the engine oil circuit is considered to be abnormal.

[0164] The sample Xi comprises at least one current oil temperature value EOTcour associated with a current oil pressure value EOP C our,i and at a current value of a high-pressure N2 regime C our,i.

[0165] Preferably, several samples xi, Xi, x n are obtained in step E210, a normalized oil pressure value being determined for each sample in step E220, by providing, for each sample, the current EOT oil temperature values CO ur,i, EOTcour , EOT CO EOP oil pressure ur,n CO ur,i , EOPcour , EOPcour,n, and high-pressure regime N2 CO ur,i, N2 C0U r,i, N2 C0U r,n to the MODnorm normalization model so as to obtain a normalized oil pressure value EOPnorm,i, EOPnorm, EOPnorm,n for said sample xi, Xi, x n A sample can be obtained for each flight of said engine.

[0166] Figure 5 illustrates normalized oil pressure values ​​EOPnorm,i, EOPnorm, EOPnorm,n obtained for a plurality of flights of a given engine.

[0167] Preferably, for each flight since the last maintenance action or since the engine was put into service, a single sample is obtained.

[0168] Alternatively, multiple samples can be obtained for each flight. These samples can be taken from a portion of the available flight reports for the engine, meaning those accessible, for example, through a database. For instance, the samples could correspond to values ​​from all available flight reports since the last maintenance operation, or since the engine entered service. This provides an indicator whose evolution over time can be analyzed, particularly useful for monitoring the evolution of the same oil circuit that has not been replaced or serviced during maintenance operations.

[0169] For each flight, the current oil pressure value may be a representative value of the oil pressure values ​​of the sample plurality relating to said flight, such as a mean or a median, or the oil pressure value of the sample plurality closest to said representative value.

[0170] When the current value of oil pressure is a representative value of the oil pressure values ​​of the relative sample plurality of said vol such as a mean or a median, the current value of oil temperature is also a representative value of the oil temperature values ​​of the relative sample plurality of said vol, such as a mean or a median, and the current value of high pressure regime is also a representative value of the high pressure regime values ​​of the relative sample plurality of said vol, such as a mean or a median.

[0171] When the current oil pressure value is the oil pressure value of the sample plurality closest to said representative value, the current oil temperature value is the temperature value of the sample corresponding to said current oil pressure value, and the current high pressure speed value is the high pressure speed value of the sample corresponding to said current oil pressure value.

[0172] Preferably, the monitoring procedure includes a step E230 for determining moving average values ​​from the standardized oil pressure values:

[0173] [Math. 9]

[0174]

[0175] Thus, the standardized pressure values ​​are smoothed over time and flights.

[0176] The moving average window can be between 5 and 50, or even between 10 and 30. The moving average window can be around 20. In other words, averages are calculated by grouping 20 successive normalized pressure values ​​over time.

[0177] This smoothing of normalized oil pressure values ​​reduces noise in the data while minimizing the time it takes to detect an anomaly. In step E240, the normalized pressure value(s), possibly smoothed, are compared to at least one threshold.

[0178] When the normalized pressure value is above or below which the oil pressure contained in the engine oil circuit is considered to be abnormal.

[0179] Preferably, several thresholds can be predetermined, defining varying degrees of failure. Depending on the position of the normalized value relative to the different thresholds, different actions can be taken or at least suggested.

[0180] A first threshold can define a healthy state when the normalized value, possibly smoothed, does not exceed said first threshold, and an abnormal state when it exceeds said first threshold. A second threshold can define a risky state when the normalized value, possibly smoothed, does not exceed said second threshold, and a failing state when it exceeds said second threshold. The second threshold is complementary to the first threshold and serves to more precisely assess states considered abnormal.

[0181] For example, in the example in Figure 4, the first threshold can be the interval lo and the second threshold the union of the intervals h and l'i: if the normalized value, possibly smoothed, is within lo, the engine is considered to be healthy, if the normalized value, possibly smoothed, is not within lo, but is within IiU l'i, the engine is considered to be at risk, if the normalized value, possibly smoothed, is not within lo, and is not within IiU l'i, the engine is considered to be faulty.

[0182] Whenever the oil pressure in the engine's oil circuit is considered abnormal, an alert may be generated, or even transmitted to an operator; in particular, an indication may be given to an operator, for example, to plan maintenance actions.

[0183] In certain embodiments, the alert is generated when the oil pressure in the engine oil circuit is considered abnormal for a predetermined number of flights, making it possible to reduce the number of false positives and false negatives.

[0184] A false positive is an engine incorrectly identified as defective or at risk. Conversely, a false negative is an engine incorrectly identified as healthy.

[0185] In specific embodiments, the alert is generated when a predefined number of normalized oil pressure values, possibly smoothed, considered abnormal relative to at least one threshold is reached. Preferably, this predefined number of normalized values, possibly smoothed, is relative to a specific threshold and to a total number of normalized values, possibly smoothed, considered.

[0186] For example, an alert indicating an abnormal-risk state is generated when K normalized values, possibly smoothed, are not within the first interval but are within the second interval, preferably considering a total number T of normalized values. K could be around 70 and T around 100. In another example, K could be around 350 and T around 500. Preferably, K and T are chosen so that the K / T ratio is around 70%. However, the K / T ratio is not limited to 70% and can be between 60% and 90%.

[0187] Similarly, an alert indicating an abnormal or faulty state can be generated when L normalized values, possibly smoothed, are neither within the first interval nor within the second interval, preferably considering a total number T of normalized values. Preferably, L and T are chosen such that the L / T ratio is approximately 70%. However, the L / T ratio is not limited to 70% and can be between 60% and 90%. The K / T and L / T ratios can be different.

[0188] In some embodiments, the threshold(s) define values ​​beyond which a standardized oil pressure value is considered abnormal. In other words, in some embodiments, the threshold(s) define only maximum values ​​that must not be exceeded.

[0189] Preferably, the threshold(s) define minimum and maximum values, as mentioned previously.

[0190] The monitoring method may include the selection of a specific normalization model, or the provision of a value relative to the engine indicating to the normalization model which parameterization to use, in other words which specific prediction model to use, from among those parameterized during an implementation of a generation method according to the invention.

[0191] The monitoring method according to the invention can advantageously be implemented in a conventionally implemented engine maintenance procedure.

[0192] Typically, maintenance procedure 300 involves the following successive steps:

[0193] - Optionally, performing a permeability test of the engine oil circuit during an optional test step T310,

[0194] - the implementation of a set of maintenance operations carried out on the aircraft, and more specifically on the engine in the workshop, during an E320 stage,

[0195] - a T330 test stage involving the execution of a set of engine function and performance tests,

[0196] - a T340 step of checking the oil level in an oil circuit of said engine via the determination of the POILC.

[0197] - when the POILC indicates an abnormal oil level (Y), a step T310 of carrying out a permeability test of the engine oil circuit, otherwise (N), the engine is considered operational and the engine can be put back into service during a step E350.

[0198] - When the T310 permeability test indicates the presence of a fault (Y), then the engine can be subjected to an escalation process during an E360 step; otherwise (N), the engine is considered operational and can be returned to service during an E350 step. The escalation process includes potentially corrective actions aimed at either carrying out further investigations and / or corrections, or validating a possible return to service of the engine.

[0199] Such a maintenance procedure, which is known, is illustrated in Figure 6.

[0200] Thus, the invention also aims at the use of a monitoring method 200 prior to an aircraft engine maintenance procedure, the procedure 300 being adapted according to said monitoring method 200.

[0201] In particular, when said monitoring procedure 200 indicates an oil pressure in the engine oil circuit considered to be abnormal (Y), for example when an alert is generated, then the maintenance procedure may include a step T310 of implementing a permeability test of the engine oil circuit, after the implementation of step E320 of implementing a set of maintenance operations carried out on the aircraft and more particularly on the engine in the workshop, and possibly prior to the implementation of this step E320 of implementing a set of maintenance operations carried out on the aircraft and more particularly on the engine in the workshop.Furthermore, when the permeability test, implemented before or after step E320, indicates the presence of a failure (Y), an escalation process involving potentially corrective actions aimed at either conducting further investigations and / or additional corrections may be implemented during step E362. After such step E362, the procedure may include the implementation of step E320 or step T330.

[0202] When these permeability tests do not detect any failure (N), then the maintenance procedure proceeds in the usual way.

[0203] In figures 6 and 7, the references "Y" mean that at the end of the test, the motor is considered abnormal / defective, while "N" means that at the end of the test, the motor is considered healthy / normal.

[0204] On the contrary, when said monitoring method 200 indicates an oil pressure contained in the engine oil circuit considered to be normal (N), for example when no alert is generated, then the maintenance procedure is unchanged, in particular the maintenance procedure does not include an optional step of carrying out a permeability test of the engine oil circuit T310 (dotted line in figure 6), carried out in the prior art in a predictive manner.

[0205] One such use is illustrated in Figure 7.

[0206] Correspondingly, the invention relates to a device 2 for monitoring oil pressure contained in an oil circuit of an aircraft engine from a MODnorm standardization model, illustrated in figure 9, the device comprising: - a sample acquisition module M20 configured to implement a step E210 of a monitoring process according to the invention,

[0207] - an M22 module for determining one or more standardized oil pressure values ​​configured to implement step E220 of a monitoring process according to the invention,

[0208] - optionally a smoothing module M24 configured to implement a step E230 of a monitoring process according to the invention,- an evaluation module M26 configured to implement a step E240 of a monitoring process according to the invention.

[0209] In particular embodiments, the generation 1 and / or monitoring 2 devices have the hardware architecture of a computer, as shown in Figure 10. It should be noted that some elements of this architecture may be confused with corresponding elements of the aircraft.

[0210] Preferably, devices 1 and 2 are not carried on board the aircraft.

[0211] More specifically, generation 1 and / or surveillance 2 devices may include a PC processor, ROM read-only memory, RAM random-access memory, and communication means.

[0212] The read-only memory of the devices constitutes a recording medium readable by the processor and on which is recorded a computer program according to the invention, comprising instructions for the execution of the steps of processes 100, 200, respectively according to the invention detailed above.

[0213] This computer program defines equivalently functional modules (software) of the devices illustrated in figures 8 and 9.

[0214] Although the present invention has been described with reference to specific embodiments, it is evident that modifications and changes can be made to these examples without departing from the general scope of the invention as defined by the claims. In particular, individual features of the various embodiments illustrated / mentioned can be combined in additional embodiments. Therefore, the description and drawings should be considered in an illustrative rather than restrictive sense.

[0215] It is also evident that all the characteristics described with reference to a process are transposable, alone or in combination, to a device, and conversely, all the characteristics described with reference to a device are transposable, alone or in combination, to a process.

[0216] All values, models, samples said to be "specific to an engine" may alternatively be, in particular embodiments of the invention, specific to a type of engine or a type of aircraft.

[0217] Alternatively, a single prediction model is determined from all samples, regardless of the engine, engine type, or aircraft type to which they relate; the prediction model is not specific to any engine, engine type, or aircraft type. In this case, a single oil pressure reference value is determined, if applicable.

Claims

Demands 1. A method for generating (100) a normalization model (MODnorm) of an oil pressure value contained in an oil circuit of an aircraft engine, the method comprising: - obtaining at least one data sample, each data sample comprising an oil temperature value (EOT) associated with an oil pressure value (EOP) and a high pressure regime value (N2) acquired during a flight; - training of at least one prediction model (MOD) pre d) from at least a portion of the data sample(s), such that once trained, the prediction model is capable of determining a predicted oil pressure value from oil temperature and high-pressure operating speed values ​​provided as input to said prediction model; - parameterization of the normalization model (MODnorm) from the prediction model (MOD pred) the normalization model being parameterized so as to provide at least one normalized oil pressure value (EOPnorm) as output from at least the current oil pressure (EOPcour) and oil temperature (EOT) values COU r) and high-pressure regime (N2 CO ur) provided as input to said normalization module, the normalized oil pressure value (EOPnorm) being determined from at least one difference between the current oil pressure value (EOPcour) and a predicted oil pressure value (EOP pre d) calculated from the prediction model (MOD pre d) and current values ​​of oil temperature (EOTcour) and high-pressure operating speed (N2) C0U r).

2. A generation method according to claim 1, further comprising, before the parameterization step of said normalization model, the following steps: - determination of a reference oil temperature value (EOT) ref) and a high-pressure regime reference value (N2 re (f) by statistical analysis, respectively, of at least a portion of the oil temperature values ​​and high-pressure operating speed values ​​from the data sample set, after training at least one prediction model, deduction of at least one oil pressure reference value (EOPref) by providing the oil temperature reference values ​​(EOT) re f) and high-pressure regime (N2 re f) to the prediction model (MOD pre d); the parameterization of said normalization model being further carried out from the oil pressure reference value (EOPref) and the normalized oil pressure value being determined from this said oil pressure reference value (EOPref).

3. A generation method (100) according to any one of the preceding claims, wherein several data samples are obtained, the samples relating to a plurality of flights carried out with different engines or types of engines, the training stage of at least one prediction model comprising, for each engine or engine type, the training of a prediction model specific to said engine or engine type from the data sample(s) relating to flights carried out with said engine or engine type.

4. Generation method (100) according to claims 2 and 3, wherein the pressure reference value (EOPref) is specific to an engine or engine type.

5. Generation method (100) according to any one of the preceding claims wherein at least one prediction model is generated by linear regression.

6. A method (200) for monitoring oil pressure in an aircraft engine oil circuit from a standardization model (MODnorm) generated by means of a method according to any one of the preceding claims, the monitoring method comprising: - obtaining at least one sample (xi) of current values, each sample including a current oil temperature value (EOT) COU r) associated with a current oil pressure value (EOPcour) and a current high-pressure regime value (N2 CO ur) acquired during a flight of said aircraft; - for each sample (xi), provision of current oil temperature values ​​(EOT) COU r), oil pressure (EOPcour), and high-pressure regime (N2 CO ur) to the normalization model (MODnorm) so as to obtain a normalized oil pressure value (EOPnorm) for said sample; - comparison of the normalised oil pressure value(s) to at least one threshold above or below which the oil pressure contained in the engine oil circuit is considered to be abnormal.

7. A monitoring method (200) according to claim 6 wherein a plurality of samples (xi, x n ) is obtained, the comparison at at least one threshold being carried out by comparing an average of the normalised oil pressure values ​​at at least one threshold, preferably the average being a moving average carried out over a predefined number of normalised values.

8. Monitoring method according to any one of claims 6 and 7, an alert is triggered when, for a number of flights beyond a predetermined number of flights, the oil pressure contained in the engine tank is considered to be abnormal.

9. Computer program comprising code instructions which, when implemented, allow the execution of steps of a process according to any one of the preceding claims.

10. Recording medium readable by means of a computer comprising a computer program according to claim 9.

Images