Device and method for monitoring damage to an engine caused by the ground surface overflown

Abstract

The invention relates to a method for evaluating degradation suffered by a helicopter engine as a result of its environment in flight, the method comprising the following steps: a) obtaining flight data for at least one flight characterising at least one flight trajectory comprising at least latitudes, longitudes and heights from the ground during the at least one flight; b) obtaining characteristic data of the flight environment, referred to as terrain data, defining one or more types of terrain overflown during the at least one flight; c) determining at least one degradation index from the analysis of the terrain data to identify areas likely to cause engine damage, crossed with the flight data; d) determining the presence of degradation according to the at least one degradation index.

Description

[0001] Description

[0002] Title of the invention: Device and method for monitoring engine damage caused by the ground surface overflown

[0003] Technical Field

[0004] This presentation concerns maintenance operations in aeronautics, in particular the monitoring of the operating status of aircraft, commonly carried out using devices called "Health Monitoring System", and more specifically the monitoring of the operating status of helicopter engines.

[0005] Previous technique

[0006] Unlike airliners designed to fly at very high altitudes, helicopters generally fly at relatively low altitudes (e.g., below 2500 meters), and are therefore more likely to be subjected to polluted, corrosive, abrasive and / or erosive environments.

[0007] Furthermore, takeoff and landing phases can be carried out on any type of ground which could then damage the helicopter.

[0008] The helicopter engine is also particularly exposed to the external environment due to its positioning under a rotating wing which can facilitate the introduction of external elements into the air intake of said engine.

[0009] Therefore, due to its use and structure, it is necessary to regularly perform specific maintenance operations to verify that the helicopter, and more particularly its engine, is operational. These maintenance operations typically follow a predetermined schedule provided by the manufacturers, possibly adapted according to the helicopter's operating area. Some operations may be mandatory at regular intervals, while others may only be recommended.

[0010] The operations performed are therefore not necessarily required, nor are they carried out at the most opportune time. It is well known to implement preliminary checks to assess the relevance of performing any maintenance operation. These preliminary checks consist of performing engine performance tests and determining whether or not the engine requires maintenance. However, these tests do not allow us to confirm that maintenance operations are necessary or to correct the fact that the engine is not meeting expected performance levels, as the performance shortfall could, for example, stem from a component unrelated to the maintenance operations performed. These checks only offer an indirect view of the engine's condition and therefore do not allow us to determine if the problem originates from the engine itself and, if so, whether it is clogged, eroded, oxidized, or has suffered impacts.These checks do not allow for the identification of appropriate maintenance operations.

[0011] Therefore, there is a need to enable maintenance operations to be better adapted to the actual condition of the engine. There is also a need to optimize maintenance schedules to avoid any unnecessary helicopter downtime.

[0012] Furthermore, even though helicopters typically have filters designed to limit the introduction of particles into the engine, their effectiveness remains limited and the introduction, even limited, of more or less fine particles into the engine can cause fouling and / or erosion.

[0013] In particular, erosion, which corresponds to the removal of material through abrasion by aerosol particles, can significantly reduce engine lifespan, notably due to the degradation of compressor performance. Furthermore, pollution, carried for example by dust, can cause fouling of the air intake or auxiliary systems, such as a blow-off valve controlled by compressed air. Pollution can also lead to a reduction in engine lifespan (fouling of cooling circuits, loss of combustion chamber efficiency, etc.). Generally speaking, there is a real need for methods to identify damage sustained by a helicopter engine or to prevent future damage, and to optimize maintenance operations.

[0014] The invention aims to meet all or part of these needs.

[0015] Description of the invention

[0016] This presentation concerns a method for evaluating the degradation caused to a helicopter engine by its in-flight environment. The method is implemented by computer and comprises the following steps:

[0017] a) obtaining flight data for at least one flight, said data characterizing at least one flight trajectory including at least latitudes, longitudes and ground heights during said at least one flight;

[0018] b) obtaining data characteristic of the flight environment, called terrain data, and more particularly characteristics of the environment overflown during said at least one flight such as characteristics of a terrain overflown;

[0019] c) determination of at least one degradation index from the analysis of ground data cross-referenced with flight data, so as to identify areas likely to cause engine damage;

[0020] d) determination of the presence of engine degradation based on at least one degradation index.

[0021] The characteristic data of the flight environment makes it possible in particular to define one or more types of terrain overflown.

[0022] The method according to the invention can advantageously allow consideration of the impact of the terrain overflown on an engine in order to plan appropriate maintenance operations.

[0023] The method according to the invention may further include a preliminary step of cross-referencing flight data and ground data, filtering the flight data, with data relating to a ground height above a predefined threshold value not being considered for the remainder of the method. The predefined threshold value may be on the order of 1 to 10 times the rotor diameter of the helicopter.

[0024] The method according to the invention makes it possible to consider ground heights during a helicopter flight that can cause the creation of particle clouds depending on the composition of the ground being flown over and the helicopter's proximity to the ground. Thus, particularly during very low-altitude flights, typically below 60 meters, the helicopter in flight can create a cloud of particles that can enter the engine and damage it.

[0025] Field data preferably includes multispectral images, hyperspectral images, and / or radar images.

[0026] Field data is advantageously obtained from satellite image databases and / or geological databases.

[0027] Field data can be obtained via application programming interfaces, commonly known as APIs, and open source databases.

[0028] The flight data preferably includes at least ground heights during said at least one flight and airspeeds during said at least one flight.

[0029] In particular, ground height helps determine if the helicopter is close enough to influence the surface of the terrain being overflown. Depending on the type of terrain, and especially the type of surface being overflown, the ground height at which a helicopter in flight exerts an influence on that terrain can vary.

[0030] Air speed is advantageously taken into account, so as to refine the estimation of the possible influence of the helicopter on the surface of the terrain flown over; the greater the relative speed with respect to the air, the greater the cloud of particles created can be, in height, width and / or density.

[0031] In addition, the flight data may include the blade pitch so as to allow the lift of the helicopter to be taken into account. Other flight data, conventionally acquired by means of sensors on board the helicopter, may also be obtained.

[0032] Preferably, flight data is acquired during flight using sensors onboard the helicopter. In other words, the invention utilizes readily available flight data without the need for additional sensors. Flight data can be obtained after at least one flight. That is, flight data can be retrieved and used while the helicopter is on the ground, once at least one flight has been completed.

[0033] Flight data can be acquired at high frequency, typically at a frequency on the order of Hertz, for example 2 Hz.

[0034] The flight path can take the form of a set of data samples, each sample of which is associated with at least one geographical indication, including a ground height, latitude, and longitude.

[0035] The geographical indication can correspond to a point in space, a segment, an area, and take for example the form of a vector (latitude, longitude, height above ground) or ([min latitude, max latitude], [min longitude, max longitude], [min height above ground, max height above ground]).

[0036] Each sample can also be associated with other types of data such as air speed, blade pitch, engine speed, etc.

[0037] Determining the presence of degradation may involve determining a state of degradation from at least one degradation index.

[0038] Determining the presence of damage, and in particular the extent of damage, can be done by taking into account engine characteristics such as engine type, brand, technologies used, and materials. This list is not exhaustive.

[0039] The determination of the presence of degradation, in particular the state of degradation, can be carried out by comparison to a reference value, for example a value beyond which a maintenance operation is necessary. The method according to the invention may further include a step of adapting a maintenance plan associated with said engine according to the presence of degradation and / or at least one degradation index and / or the state of degradation where applicable.

[0040] In particular, the adaptation may include: imposing, delaying, removing or bringing forward a maintenance operation, and / or changing the periodicity of a maintenance operation.

[0041] The method according to the invention can advantageously make it possible to identify the type of particles likely to be introduced into the engine and thus plan maintenance operations that are adapted accordingly.

[0042] Said at least one flight may include all flights carried out since the last maintenance performed or since the engine was put into service.

[0043] The analysis of field data may involve the application of a classification method followed by the analysis of geochemical characteristics by image processing, the field data including images, in particular multispectral images.

[0044] Thus, based on field data, a map of the geographical area flown over can advantageously be determined according to the content of the terrain and its composition, particularly on the surface.

[0045] Preferably, the classification method is configured to identify land types from a set of predefined land types, particularly based on the land's composition. For example, the classification method can be configured to identify all or some of the following classes: "built-up land," "forest," "bare soil," "sand," "clay soil," etc. This list is not exhaustive.

[0046] In general, the classification method can be configured to identify, on the one hand, terrain types that may be laden with particles likely to be stirred up by the helicopter's approach, such as sandy, dusty, or industrial terrain, and on the other hand, terrain types that present little or no risk of potential engine damage. The classification can be binary. Alternatively, the classification can include more than two predefined classes. This can improve the accuracy of the analysis, allowing, for example, the distinction between high-impact, medium-impact, and low-impact zones, and / or zones likely to cause engine fouling, zones likely to cause engine erosion, zones likely to cause both fouling and erosion, and zones unlikely to cause engine damage.

[0047] Any classification method can be used.

[0048] Geochemical analysis, on the other hand, allows for the characterization of the ground surface composition. This analysis is advantageously applied only to terrains whose classification is considered potentially damaging to a helicopter.

[0049] The analysis of geochemical characteristics can be parameterized to identify surface types from a set of predefined surface types, particularly based on the composition of the terrain.

[0050] Preferably, this analysis of geochemical characteristics is carried out using a machine learning algorithm, previously trained, for example by means of a neural network, a random forest, or a support vector machine (SVM).

[0051] The analysis of geochemical characteristics aims to distinguish the types of soil on the surface of the land flown over, in other words the types of surfaces, in particular so as to make it possible to identify the particles that can cause fouling of the particles that can cause erosion.

[0052] The analysis of geochemical characteristics can provide as output a categorization of terrain types among predefined categories of terrain types, including a categorization of surface types among predefined categories of surface types.

[0053] The classification and analysis of geochemical characteristics advantageously allows for a reduction in size. This makes it possible to group soils by severity category, for example: none, low, medium, high, and to provide a more reliable analysis of engine damage.

[0054] The process may include, for each type of terrain overflown, in particular for each type of surface overflown, the determination of a flight time at a ground height below a predefined threshold value for the type of terrain considered, respectively of surface, the degradation index being defined according to one or more flight times.

[0055] In one particular embodiment, the analysis also takes into account flight data, such as ground height, air speed, blade pitch, etc., with the cross-referencing of flight data and ground data being carried out simultaneously with the analysis of geochemical characteristics.

[0056] The degradation index can be an estimated rate of fouling, an estimated level of erosion, a duration of exposure to a polluted and / or erosive environment, or a category, for example, heavily fouled / lightly fouled / not fouled. These examples are provided for illustrative purposes only and are not exhaustive.

[0057] Each type of terrain, or each type of surface, can be previously associated with a weighting coefficient, representing the impact of said type of terrain or said type of surface on the engine.

[0058] The degradation index can be a sum of the exposure times to each type of terrain, or to each type of surface, weighted by a weighting coefficient depending on the types of terrain, or the types of surface respectively.

[0059] The weighting coefficients may have been determined through a statistical analysis of historical data.

[0060] Weighting coefficients can be determined using a machine learning algorithm, previously trained on historical data, for example using a neural network, a random forest, or a support vector machine (SVM). The historical data typically takes the form of a set of samples, each sample relating to at least one flight for a helicopter engine and comprising:

[0061] - flight data relating to at least one flight, the flight data being as defined above,

[0062] - relevant field data as defined above, and

[0063] - an actual state of engine damage caused by said at least one flight. The actual state of engine damage may be determined after said at least one flight, preferably the actual state of damage including a degradation value caused by said at least one flight. The degradation value may be determined by comparing an actual state of damage before said at least one flight with an actual state of damage after said at least one flight.

[0064] The degradation value and / or the actual damage states can be determined by operators, for example during maintenance operations. Alternatively and / or additionally, the degradation value and / or the actual damage states can be determined by implementing an evaluation method according to the invention.

[0065] Thus, historical data is advantageously updated and allows for better estimation of weighting coefficients over time.

[0066] The invention also relates to a method for determining weighting coefficients, said method comprising the following steps:

[0067] - obtaining historical data, and preferably analyzing field data from said historical data in order to identify the types of terrain, or even the types of surfaces;

[0068] - statistical analysis of said data, in particular using a machine learning algorithm,

[0069] - deduction of weighting coefficients based on said analysis.

[0070] All or part of the steps in the evaluation process may be implemented by computer; preferably all steps of the process are implemented by computer. All or part of the steps in the determination process may be implemented by computer; preferably all steps of the process are implemented by computer.

[0071] 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.

[0072] The invention further relates to a device for implementing a method according to the invention comprising at least one processor configured to implement the computer program.

[0073] The device may also include a memory for storing flight data and field data, and where appropriate actual damage statuses, degradation indices, degradation statuses, and / or maintenance plans.

[0074] The device may also include means of communication enabling the acquisition of flight data and terrain data, including means of communication via Wifi, 4G or other cellular networks.

[0075] 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.

[0076] Brief description of the drawings

[0077] The attached drawings are schematic and are primarily intended to illustrate the principles of the presentation.

[0078] In these drawings, from one figure to another, identical elements (or parts of elements) are identified by the same reference symbols.

[0079] [Fig. 1] Figure 1 represents the steps of an example of the implementation of an evaluation method according to an embodiment of the invention,

[0080] [Fig. 2] Figure 2 illustrates the operation of a geohash method applied to flight data. [Fig. 3] Figure 3 schematically represents a cross-referencing of filtered flight data with terrain types.

[0081] [Fig. 4] Figure 4 schematically represents a device according to an example of an embodiment of the invention,

[0082] [Fig. 5] Figure 5 represents a material architecture of a device according to an example of an embodiment of the invention.

[0083] Description of the implementation methods

[0084] To make the explanation more concrete, an example of the implementation of an evaluation method 300 is described in detail below, with reference to the attached drawings. It should be noted that the invention is not limited to this example.

[0085] Correspondingly, a device 1 configured to implement process 300 is also described in Figure 5.

[0086] The evaluation process aims to enable a quantification of the level of exposure of an engine to an erosive / polluted environment, in particular due to proximity to soil composed of particles likely to cause erosion and / or fouling of a helicopter engine due to their absorption by the latter.

[0087] Indeed, the geological conditions and soil types encountered by aircraft can significantly influence the rate of engine erosion and fouling. Flights in dusty, sandy, or industrial environments therefore require protective measures and increased monitoring to maintain aircraft engine performance and minimize potential damage. The evaluation process 300 includes a flight data acquisition step E310.

[0088] Flight data can be acquired at high frequencies, particularly at a frequency of around 2 Hz, in flight using onboard sensors. These onboard sensors typically allow the acquisition of the following data: - speed, including the helicopter's forward speed and / or engine rotation speed,

[0089] - altitude,

[0090] - pressure,

[0091] - height above ground, in other words altitude, the distance from the ground

[0092] - temperature, internal and / or external,

[0093] - geolocation coordinates.

[0094] Flight data includes at least one flight path defined by ground heights, latitudes and longitudes acquired during at least one flight.

[0095] Flight data can take the form of vectors, each containing at least one piece of spatial information: latitude(s), longitude(s), ground height(s), and possibly additional information such as speeds, altitudes, pressures, temperatures, etc.

[0096] Flight data is advantageously synchronized, for example via internal clocks in the helicopter, or deduced from a data acquisition frequency.

[0097] Flight data can be obtained via Wi-Fi, 4G or other cellular data networks.

[0098] The evaluation process 300 includes a step E320 of obtaining field data 32, which may take the form of one or more maps, including in particular multispectral satellite images.

[0099] The field data 32 can at least partly come from geological databases.

[0100] The field data includes minimal information allowing the determination of the typology of the land flown over, in particular the surface area of ​​the land flown over.

[0101] Ideally, field data should include data from different sources and databases. In particular, field data may include:

[0102] - multispectral images,

[0103] - hyperspectral images,

[0104] - radar images.

[0105] The resolution of these images can be from 5 to 30 meters.

[0106] Field data can come from acquisitions by satellites such as Sentinel 1, Sentinel 2, Landsat 8, Hyperion, AVIRIS, and RADARSAT 2.

[0107] The ground data may include a plurality of images representing the same geographical area including the terrains overflown during said at least one flight.

[0108] Multispectral images can be used, for example, to identify vegetation types, soil types, and for land mapping. Hyperspectral images can be used to determine the properties of soils, minerals, and other geological features in greater detail. Radar images can provide information at night and through cloud cover. Among other things, radar images can provide information about ground surface structure and soil moisture.

[0109] Flight data and / or terrain data can be pre-processed so as to be associated with a geographic cell, for example by applying a geohash method and possibly a timehash method.

[0110] A simplified example of the application of a geohash method is shown in Figure 2.

[0111] A flight trajectory Tv representing flight data acquired during a flight is illustrated in a graph where longitudes are on the x-axis and latitudes are on the y-axis. Each point pi corresponds to a vector acquired at time t, for a particular latitude, a particular longitude, and a particular ground height.

[0112] A geographic cell Cj can be defined by a range of longitudes, a range of latitudes, and optionally a range of altitudes / ground heights. These ranges can be on the order of kilometers; in particular, the cells can form roughly squares with sides of 1 km. This example is not exhaustive, however, as the size of the ranges considered is advantageously determined according to the desired precision.

[0113] A geographic cell can also be defined by a time interval, for example, on the order of a few minutes, such as ten minutes. This example is not exhaustive, however, as the size of the interval can be advantageously determined according to the desired precision.

[0114] In general, geohashing methods involve a spatial division that partitions the Earth's surface into cells of varying sizes, hierarchically and with a specific encoding, typically in the form of character strings. These methods are widely used in geoinformatics, particularly for spatial research, geographic indexing, and mapping. A geohashing method is also based on:

[0115] - a hierarchical structure: the Earth's surface is divided into a grid forming cells of roughly square or rectangular shapes delimited according to geographical coordinates (latitude and longitude). The cells can be subdivided into sub-cells, and so on, thus forming an increasingly fine mesh with increasingly smaller cells.

[0116] - Encoding of cells, sub-cells, etc. The encoding is conventional, based on characters composed of a base-32 alphabet, which allows for efficient compression of geographic coordinates. The longer the encoding associated with a cell, the more precise the location, as a character is added for each subdivision.

[0117] These traditional methods allow for efficient storage and fast searches. Furthermore, the hierarchical structure allows for adjusting the level of precision according to specific needs.

[0118] Timehash methods rely on using sliding time windows placed side-by-side to aggregate data and thus reduce the amount of data. The window size can be adjusted according to the desired accuracy.

[0119] The example in Figure 2 is applied to flight data. Of course, such methods can also be applied to ground data.

[0120] The 300 evaluation process includes a step E330 for analyzing ground data to identify areas likely to cause damage. A degradation index is derived from the combination of flight data and the ground areas thus identified.

[0121] Flight data is combined with ground data, and more specifically with areas identified as potentially damaging. Graphically, this data combination can be represented as an overlay of flight and ground data within a common spatial reference frame. Such an example is illustrated in Figure 3.

[0122] Prior to this E330 step, and in particular prior to the cross-referencing of flight and ground data, the evaluation process may include an E315 step of filtering flight data.

[0123] In particular, this E315 filtering step involves comparing flight altitudes with a threshold value; only flight data relating to altitudes less than or equal to said threshold value are subsequently analyzed. The threshold value is advantageously predetermined to correspond to the helicopter's ground distance from which the effects of its vertical downdraft raise dust, typically between 1 and 10 times the helicopter's rotor diameter.

[0124] Alternatively or additionally, the E315 filtering step can be performed by comparing a forward speed, a collective pitch position of the main rotor or a power delivered by the motor, against a threshold value or threshold values.

[0125] Threshold values ​​can potentially be determined or adjusted by implementing machine learning algorithms using historical data. This allows for refining threshold values ​​as data is acquired, leading to a more accurate analysis of flight data. For example, a machine learning algorithm can be trained to determine a maximum altitude at which engine damage is observed for a majority of historical data points. The historical data may include flight data for multiple flights, with each flight data point associated with a level of engine damage caused by that flight, which can be determined by a maintenance operator.

[0126] A ground impact zone corresponding to an area impacted by the effects of the vertical blast can be approximated by a square footprint with sides 1 to 10 times the helicopter's rotor diameter. Any other shape can be considered. Data can be overlaid to superimpose ground impact zones with areas identified as likely to cause damage. In other words, the overlay involves determining the intersection of ground impact zones with areas identified as likely to cause damage.

[0127] The resolution of flight data and ground data can advantageously be defined from the size of this ground impact zone.

[0128] Step E330 may include an analysis of the physical characteristics of the terrains overflown, during a sub-step E332, in particular for the terrains overflown corresponding to the flight data remaining after filtering, possibly taking into account a radius of influence of the helicopter.

[0129] This analysis of the physical characteristics of the terrains flown over can be carried out by means of a classification identifying one or more types of terrain flown over from a set of predefined terrain types.

[0130] The process may involve selecting, from field data, the relative geographical areas of typical terrains likely to damage the helicopter, in particular by causing fouling or erosion.

[0131] The process may include a second analysis, during substep E334, of the composition of the surfaces of the terrain overflown for the selected geographical areas, based on at least a portion of the terrain data. This second, more detailed analysis can categorize terrain types according to their influence on the engine and its performance, in particular by analyzing the surfaces of the terrain types to more precisely distinguish between specific surfaces leading to erosion and those leading to engine fouling.

[0132] This second analysis can be implemented using a machine learning algorithm.

[0133] Step E330 may include a substep E336 of determining at least one degradation index from the cross-referencing of flight data Tv, preferably flight data previously filtered Tvtiit by the implementation of a step E315, and ground data previously classified and categorized.

[0134] Such a cross-referencing of data is shown schematically in figure 3.

[0135] The filtered Tvfin flight data is cross-referenced with previously analyzed ground data to identify terrain and surface types. The shades of gray represent different categories 32'a, 32'b, 32'c of surface types associated with geographical areas of the ground data 32.

[0136] All or part of step E330 can be performed directly by a machine learning algorithm. In particular, the machine learning algorithm can be pre-trained to provide as output a degradation index based on input data including: flight data and ground data, possibly classified ground data, i.e. ground data after implementation of an E332 classification step, or even categorized ground data, i.e. after implementation of an E334 classification step.

[0137] Preferably, the machine learning algorithm is trained to take into account the time required for the dust cloud to form and dissipate. This can be done during the algorithm's training. The degradation index can be an exposure time for each type of terrain, preferably from among the selected terrain types likely to damage the engine. The degradation index can be an exposure time for each type of surface. The degradation index can be a fouling rate and / or an erosion rate.

[0138] The degradation index can be a sum of the exposure times to each type of terrain, or to each type of surface, weighted by a weighting coefficient depending on the types of terrain, or the types of surface respectively.

[0139] These coefficients can be determined beforehand by statistical analysis or by means of a machine learning algorithm.

[0140] The degradation index can be cumulative. In particular, the degradation index can accumulate over time, with each flight. It can be reset after maintenance is performed.

[0141] Process 300 includes a step E340 of deduction of the presence of degradation based on at least one degradation index.

[0142] The presence of degradation may involve determining a state of degradation.

[0143] The state of degradation can be determined by comparing the degradation index to a reference threshold.

[0144] Depending on the state of degradation and / or the degradation indicator(s), a maintenance plan can be defined, adjusted, modified.

[0145] In particular, a maintenance operation can be planned based on the state of degradation or the degradation index before a failure occurs or before the engine can no longer deliver sufficient power or its operability (pumping margin) is impacted.

[0146] Triggering or, more generally, scheduling maintenance operations based on degradation indicators and / or the aircraft's condition minimizes helicopter downtime while ensuring optimal operation. Maintenance is thus triggered only when necessary, significantly increasing helicopter operational availability, particularly in erosive and polluted environments, compared to currently used generic pre-programmed maintenance plans. For example, some helicopters currently undergo endoscopic compressor inspections every 50 flight hours. However, this lengthy and complex procedure is not always necessary at the time it is performed. The process described above allows for validating a satisfactory compressor condition without resorting to endoscopic inspection.This operation can then be implemented only when the state of degradation indicates damage. Thanks to the invention, the number of maintenance operations can thus be significantly reduced.

[0147] Ground conditions (geological or otherwise) can significantly impact aircraft engine maintenance costs, including increased maintenance frequencies, component replacements, regular cleaning, additional inspections, and costs associated with reduced engine life, corrosion, and unscheduled downtime. This necessitates proactive and well-planned management to minimize these costs and ensure engine reliability and performance. As previously mentioned, weighting factors can be determined through statistical analysis or by using a machine learning algorithm based on historical data.

[0148] In general, the determination of the degradation index(es) and / or the state of degradation and / or the planning of maintenance operations can be carried out through prior statistical analysis or by means of a machine learning algorithm from historical data in order to identify existing correlations between field data and flight data.

[0149] Thus, the invention also relates to a method for determining (not shown) correlation coefficients between types of terrain overflown and the impact of these overflights on the engines of the helicopters that performed these flights, each correlation coefficient being associated with a type of terrain and being representative of the damage caused by the terrain type to a helicopter engine. The determination method comprises:

[0150] - obtaining historical data including flight data, terrain data and actual engine damage status,

[0151] - preferably, the analysis of the field data of said historical data in order to identify the types of terrain, or even the types of surfaces;

[0152] - the statistical analysis of said historical data, in particular by means of a machine learning algorithm,

[0153] - deduction of correlation coefficients based on said analysis.

[0154] Historical data takes the form of a set of samples, each sample relating to at least one flight for a helicopter engine and comprising:

[0155] - flight data relating to at least one flight, the flight data being as defined above,

[0156] - relevant field data as defined above, and

[0157] - an actual state of engine damage determined after said at least one flight, preferably the actual state of damage including a degradation value caused by said at least one flight. The degradation value can be determined by comparing an actual state of damage before said at least one flight with an actual state of damage after said at least one flight.

[0158] The degradation value and / or the actual damage states can be determined by operators, for example during maintenance operations. Alternatively and / or additionally, the degradation value and / or the actual damage states can be determined by implementing an evaluation method according to the invention.

[0159] These steps make it possible to determine the types of terrain and in particular surfaces that have an impact on engine performance, but also to determine which types of terrain and in particular specific surfaces lead to erosion and / or engine fouling. The more significant the impact is analyzed as being, the higher the correlation coefficient can be.

[0160] Correspondingly, the invention relates to a device 1 configured to implement an evaluation method. Such a device is illustrated in Figure 4.

[0161] Device 1 comprises:

[0162] - an M30 data acquisition module configured to implement a step E310 of a process according to the invention;

[0163] - an M32 data acquisition module configured to implement a step E320 of a process according to the invention;

[0164] -an M34 analysis module configured to implement an E330 step of a process according to the invention;

[0165] - an M36 determination module configured to implement a step E340 of a process according to the invention.

[0166] Flight data can come from 3 onboard sensors.

[0167] Field data can come from one or more remote databases, including satellite image databases, particularly accessible via APIs.

[0168] In particular embodiments, device 1 has the hardware architecture of a computer, as shown in Figure 5. It should be noted that some elements of this architecture may be confused with corresponding elements of the helicopter.

[0169] Preferably, device 1 is not carried on board the helicopter.

[0170] More specifically, device 1 may include a PC processor, a read-only memory (ROM), a random-access memory (RAM), and communication means.

[0171] The read-only memory of device 1 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 the process 300 according to the invention detailed above and in particular illustrated in figures 1 to 3. This computer program defines in an equivalent way functional modules (software) of device 1, such as in particular the data acquisition modules M30, M32, the analysis module M34, and the determination module M36.

[0172] 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.

[0173] 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.

[0174] In general, the invention is based on combining conventionally acquired flight data with data accessible via APIs relating to the flight environment, such as meteorological data, ground type, and / or ground or air composition. In certain embodiments, the ground data can be stored locally, for example, in a database providing a program that can be run on a standard computer at the helicopter's base, without requiring an internet connection. By leveraging a readily available dataset, and in particular without the need for additional sensors, the invention provides a maintenance support tool.

[0175] The invention is also easily implementable.

Claims

Demands 1. Method for evaluating (300) degradation caused to a helicopter engine by its in-flight environment, the method comprising the following computer-implemented steps: a) obtaining (E310) flight data (30) for at least one flight, said data characterizing at least one flight trajectory (Tv) comprising at least latitudes, longitudes and ground heights during said at least one flight; b) obtaining (E320) data (32) characteristics of the flight environment, called terrain data, defining one or more types of terrain overflown during said at least one flight; c) determination (E330) of at least one degradation index from the analysis of ground data cross-referenced with flight data, so as to identify areas likely to cause engine damage; d) determination (E340) of the presence of engine degradation based on at least one degradation index.

2. Evaluation method according to claim 1 further comprising a step (E315), prior to the cross-referencing of flight data and ground data, of filtering the flight data, the data relating to a ground height greater than a predefined threshold value not being considered for the remainder of the method.

3. A method according to any one of the preceding claims, wherein the determination of the presence of degradation involves the determination of a state of degradation from at least one degradation index.

4. A method according to any one of the preceding claims comprising a step of adapting a maintenance plan associated with said engine according to the presence of degradation and / or at least one degradation index and / or the state of degradation where applicable.

5. A method according to the preceding claim in which the analysis of field data comprises the application (E332) of a classification method and then the analysis (E334) of geochemical characteristics by image processing, the field data comprising images, in particular multispectral images.

6. A method according to claim 5, the classification method being parameterized to identify types of terrain from among a set of predefined terrain types, the analysis of geochemical characteristics being parameterized to identify types of surfaces from among a set of predefined surface types, the process further comprising for each type of terrain overflown, in particular for each type of surface overflown, the determination of a flight time at a ground height below a predefined threshold value for the type of terrain considered, respectively of surface, the degradation index being defined according to one or more of said flight times.

7. Method according to claim 6, each type of terrain, or each type of surface, being previously associated with a weighting coefficient, representative of the impact of said type of terrain or said type of surface on the engine, the degradation index being a sum of the exposure times to each type of terrain, or to each type of surface, weighted by a weighting coefficient depending on the types of terrain, or the types of surface respectively.

8. A method according to claim 7, wherein the weighting coefficients are determined by means of statistical analysis or a machine learning algorithm, from historical data, the historical data taking the form of a set of samples, each sample relating to at least one flight for a helicopter engine and comprising: - flight data relating to said at least one flight, characterizing at least one flight trajectory including at least one latitude, longitude and ground height during said at least one flight, - terrain data defining one or more types of terrain overflown during said at least one flight, and - an actual state of engine damage caused by said at least one flight.

9. Computer program comprising code instructions which, when implemented, enable the execution of the steps of an evaluation 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