Method for predicting a current wear state of an identified tire mounted on an identified aircraft

Abstract

The present invention relates to a computer-implemented prediction method (200) for predicting a remaining landing before threshold (remaining LPT) corresponding to reaching a removal threshold of an identified tire, output by a prediction model. In said prediction method, a value of a removal threshold of an identified tire installed on an identified aircraft is compared to a remaining landing before threshold (remaining LPT) before reaching the removal threshold of the identified tire, such that a system (100) executing said prediction method creates a maintenance plan for the identified tire.

Description

Technical Field

[0001] This invention relates to a computer-based method for predicting the current wear condition of aircraft tires based on parameters that affect their service life. Background Technology

[0002] Aircraft tires (or "aircraft tires" or simply "tire"), like those on other vehicles, are wear parts that are replaced when they reach a removal threshold (the tread depth falls below the tire's service limit) or become too structurally degraded to continue use. The term "service life," as used below, refers to the lifespan of an aircraft tire, that is, the time until it reaches its wear removal threshold and is no longer in use; or in other words, the time an aircraft tire can provide its intended service without encountering durability issues (requiring early removal). In the aviation industry, it is well known that service life is measured by counting the number of landings a tire makes between its initial installation on an aircraft and its removal from the aircraft (this is also known as "landings per carcass," "landings per tread," or LPT).

[0003] Existing technologies exist for predicting when maintenance is needed based on a tire's expected lifespan. For example, US Patent 10,295,333 discloses a method for determining the tread condition of a tire based on an image of the tread. Furthermore, US Patent 10,179,487 discloses a method capable of generating a display representing the potential wear level of a tire or its tread condition. This display is partly based on measurements of tread depth obtained beforehand. However, neither of these solutions is applicable to aircraft tires. Therefore, neither of these solutions falls within the scope of predicting aircraft tire lifespan based on the number of landings.

[0004] WO2020 / 169833 discloses a method for detecting the deterioration of aircraft tires. In this method, determining the position of the center point of the actual tire includes a first step: defining a normal vector through the surface of a first three-dimensional object passing through each captured point. The method includes a second step: estimating the position of the center point of the actual tire based on the normal vector (the estimation of the center point position is an iterative process). The method also includes a third step: matching the first three-dimensional object with a theoretical tire of known size and orientation to obtain a second three-dimensional object forming the matched tire. The second three-dimensional object is then transformed to obtain one or more two-dimensional objects, which are analyzed to detect the deterioration of the actual tire.

[0005] WO2019116782A1 discloses a device that measures the remaining groove height of at least a first tire mounted on an aircraft and predicts the wear level of other tires mounted on the aircraft (including the first tire). Wear calculations are performed based on a number of parameters, including the wear energy and internal pressure of the aircraft tires.

[0006] However, for a given tire model, wear exhibits significant performance deviations (measured by the number of landings), easily exceeding 100% between the observed minimum and maximum values. Currently, these deviations are a maintenance and operational problem because they prevent airlines from applying predictive maintenance strategies that would allow them to optimize operations and costs. Currently, whether an aircraft tire must be removed due to structural damage is unpredictable, as it relates to probabilistic events (e.g., damage to the tire caused by a foreign object present on the runway or taxiway). This type of damage is commonly referred to as “foreign object damage” or “FOD” and occurs when an aircraft passes over a runway or taxiway strewn with rigid objects (the terms “aircraft” and “aircraft” are used interchangeably). Objects causing FOD include any type of object capable of damaging tires (including, but not limited to, loose metal pieces, sidewalk debris, catering supplies, building materials, rocks, sand, baggage, and wildlife). These objects are located at terminal entrances, runways and taxiways, and other surfaces that aircraft traverse as they move across the ground.

[0007] The inability to predict when tires will be removed at the end of their wear life (when the remaining tread height becomes less than or equal to 1.2 mm ± 0.2 mm) imposes significant limitations and associated economic consequences on customers in terms of inspection, maintenance team management, and inventory management of tires, brakes, and wheels. Furthermore, because aircraft tires have the shortest lifespan among the tire-brake-wheel-landing gear components, the entire maintenance chain of an aircraft's tire-brake-wheel-landing gear system (e.g., systems such as the ATA 32) relies on aircraft tires.

[0008] Theoretically, based on physical models, it's possible to estimate the remaining tread depth of an aircraft tire and then infer its remaining lifespan (by the number of landings or the remaining tread depth). However, for sufficient accuracy, this model requires numerous physical parameters (including but not limited to tread temperature, tire load, tire pressure, and tire speed). Without placing many sensors within the tire, obtaining these physical parameters is difficult.

[0009] Recent advancements in machine learning and data analytics, combined with platforms for computing and storing data, have paved the way for developing new predictive methods to manage aircraft tires, overcoming the shortcomings of existing techniques. Machine learning, a well-known technique in artificial intelligence, essentially involves "training" on a large number of scenarios. By adjusting weighting coefficients during the training phase, machine learning can predict outcomes for new situations presented. It's important to note that there are various machine learning approaches, including supervised learning (where the algorithm is trained and learns on a set of labeled data until it achieves the desired results), unsupervised or semi-supervised learning (where data is not labeled, allowing the network to learn and improve the algorithm's accuracy), reinforcement learning (where the algorithm is rewarded for positive outcomes and penalized for negative outcomes), and active learning (where, during learning, the algorithm requests examples and labels to optimize its predictions) (see https: / / www.lebigdata.fr / reseau-de-neurones-artificiels-definition).

[0010] Therefore, it would be beneficial to explore the relationship between aircraft tire usage (i.e., parameters affecting aircraft tire lifespan) and tire lifespan performance. This performance knowledge includes understanding the wear profiles observed due to parameters of aircraft usage.

[0011] Therefore, the present invention relates to a method for predicting the remaining service life of aircraft tires, wherein the remaining service life is indirectly estimated through the tire's wear condition. The remaining service life corresponds to potential remaining wear, which serves as a predictive maintenance tool, enabling better scheduling of all activities related to aircraft tire maintenance planning. Summary of the Invention

[0012] This invention relates to a computer-implemented prediction method for predicting the number of remaining landings corresponding to reaching a threshold for the removal of identified tires installed on an identified aircraft, the prediction method comprising the following steps:

[0013] - The step of introducing parameters affecting the identified tires into a system performing the prediction method, the system including a communication network for managing data input into the system, the communication network having one or more communication servers, the one or more communication servers managing data corresponding to the parameters affecting the identified tires, and having at least one communication device for capturing this data and sending this data to the servers, the step including the following steps:

[0014] - A step of acquiring parameters affecting the identified tires, said step being performed by the system's communication device, wherein the acquired parameters include data corresponding to historical and general information of the identified tires; and

[0015] - The step of creating a wear state training database, which is incorporated into a model for predicting the remaining number of landings corresponding to the removal threshold of identified tires;

[0016] - The step of training a prediction model to predict the remaining number of landings corresponding to the removal threshold of the identified tires, wherein the machine learning method receives the obtained influence parameters and data from the training database as input, so that the processor can obtain the known wear state corresponding to the number of landings performed by the identified tires.

[0017] - A step of predicting the remaining number of landings before reaching the removal threshold of identified tires, wherein the remaining number of landings of identified tires is calculated based on data corresponding to influencing parameters; and

[0018] - Comparison step, wherein the remaining number of landings before the removal threshold of the identified tire is reached, output by the prediction model, is compared with the value of the removal threshold of the identified tire, so that the system creates a maintenance plan for the identified tire.

[0019] In some embodiments of the method, the influencing parameters include:

[0020] -Historical information, which includes data corresponding to the historical flights of identified aircraft with the identified tires installed; and

[0021] - General information, which includes data corresponding to the identified tires, including the installation location of the identified tires on the identified aircraft.

[0022] In some embodiments of the method, the maintenance plan created by the system in the comparison step includes:

[0023] - A plan to repair identified tires when the remaining landing count output by the predictive model exceeds the removal threshold defined for identified tires; and

[0024] - Check the plan for the identified tires when the remaining number of landings output by the prediction model is equal to or less than the removal threshold defined for the identified tires.

[0025] In some embodiments of the method, the method further includes a supervised learning method that receives acquired influence parameters and data from a training database as input, enabling the processor to acquire known wear states corresponding to the number of landings performed by the identified tires in order to construct a predictive model.

[0026] In some implementations of the method, the supervised learning method includes a supervised learning method of the Gradient Boosting Regressor (GBR) type.

[0027] In some embodiments of the method, the training database includes images of wear curves corresponding to known wear states and the number of landings performed by the corresponding identified tires.

[0028] In some embodiments of the method, in the prediction step, the remaining number of landings for the identified tires is calculated based on data corresponding to the impact parameters of future landings.

[0029] In some embodiments of the method, the method further includes the step of simulating the destination airport of the identified aircraft based on historical flight data.

[0030] In some embodiments of the method, the simulation steps are performed via Markov chains, in which each state represents an airport, and connections between airports represent the probability of departing from one airport and landing at another.

[0031] In some embodiments of the method, the simulation steps are repeated multiple times using a Monte Carlo loop to predict the end-of-life of identified tires.

[0032] Other aspects of the invention will become apparent from the following detailed description. Attached Figure Description

[0033] The features and various advantages of the invention will become more apparent from the following detailed description taken in conjunction with the accompanying drawings, in which the same reference numerals refer to the same parts, wherein:

[0034] Figure 1 and Figure 2 This is a diagram of an airplane tire.

[0035] Figure 3 A schematic diagram of an embodiment of the prediction method of the present invention is shown.

[0036] Figure 4 A schematic diagram of a system for implementing the prediction method of the present invention is shown.

[0037] Figure 5 A flowchart of one embodiment of the prediction method of the present invention is shown. Detailed Implementation

[0038] The solution proposed in this invention aims to estimate the lifespan of aircraft tires without adding new sensors, utilizing only historical information about the tire's flight history (e.g., data corresponding to departure date, arrival date, airports visited, and climate data) and general information about the tire (e.g., data corresponding to its retreading level, location on the aircraft, theoretical average number of landings at that location, and performance data excluding similar tires removed due to FOD). This data, describing most of the tire's usage, will be used in a machine learning model to predict the tire's current wear condition based on parameters affecting its lifespan.

[0039] Regarding the characteristics of the aircraft tires involved in this invention, their geometry must be taken into account. Figure 1 and Figure 2 This diagram includes a tire P, which conventionally comprises two circumferential beads designed to secure the tire to a rim. Each bead includes annular reinforcing bead wires. The tire's composition is typically described by representations of its constituent components in the radial plane (i.e., the plane containing the tire's axis of rotation). Radial, axial, and circumferential directions represent directions perpendicular to the tire's axis of rotation, parallel to the tire's axis of rotation, and perpendicular to any radial plane, respectively. The expressions "radially," "axially," and "circumferentially" refer to the tire's "radial direction," "axial direction," and "circumferential direction," respectively. The expressions "radially inner" and "correspondingly radially outer" refer to directions "closer to and correspondingly further away from the tire's axis of rotation in the radial direction," respectively.

[0040] refer to Figure 1 The tire P includes an internal limit F that collectively defines the limit of the tire P's sidewall F. I and external limit F E Internal limit F I Separate the tire sidewall F from the rim (not shown) to which the tire will be mounted. The tire P also has a rim radius R. J The rim radius R J Defined as the internal limit F between the center point C of the tire and the sidewall F separating the rim and the tire. I The distance between them. Tire P also has a radius defined as the rim radius R. J The tire P has twice the inner diameter of the sidewall. The tire P also has a tire radius R. P The tire radius R P Defined as the external limit F between the center point C and the sidewall F representing the tread surface of the tire. E The distance between them. Tire P also has a radius R defined as the tire radius. P Twice the diameter of the tire.

[0041] refer to Figure 2 An inflated and unloaded tire P has several parameters related to its geometry, including the nominal section width L. P and height H P (height H) P It is usually represented as width L P (Percentage). Tire P also has a measurement D indicating the diameter of the rim on which the tire will be mounted. J (This measurement result is basically equal to the inner diameter of the tire sidewall F) I It should be understood that each of these parameters can be expressed in an equivalent known unit of length measurement (e.g., millimeters (mm) or inches (in)).

[0042] Now for reference Figure 3 and Figure 4 In this context, the same reference numerals denote the same elements. Figure 3 A schematic diagram of an embodiment of a method (or “prediction method” or “method”) for predicting the number of remaining landings corresponding to reaching a removal threshold for an identified tire is shown. The prediction method of the present invention is defined based on the wear state of an identified tire, by predicting a wear profile, by the remaining groove height (including grooves capable of removing water to allow contact and adhesion with the tread pattern elements of the runway or taxiway), or by another means for predicting potentially remaining usable tread. It should be noted that the wear state is represented by a value or a set of values ​​indicating the difference between a new state (represented by a tire never installed on an aircraft) and a wear state (represented by a tire that has ceased use due to reaching a removal threshold through wear). The wear state can take the form of a single value such as a minimum groove height, a set of values ​​defining the height of the tread in the radial plane of the tire (e.g., a two-dimensional curve), or a set of values ​​defining the height of the tread on all or part of the outer surface of the tire (e.g., a three-dimensional curve).

[0043] As used herein, the "remaining useful life" of an identified tire refers to the potential remaining wear based on the values ​​of parameters affecting the tire's service life. The remaining useful life can be determined continuously by collecting data corresponding to the parameters affecting the useful life of the identified tire, or at regular, predefined, or occasional intervals. Based on the collected data, the method of the present invention can define the current wear state of the tire to determine the remaining useful life of the identified tire, which is calculated in terms of available landings.

[0044] Figure 4A system 100 for implementing the prediction method of the present invention is shown. The system includes a communication network 102 that manages data input into the system 100 from various sources. The communication network 102 includes one or more communication servers (or “servers”) 102a that manage data corresponding to historical information and general information about identified tires. The term “identified tire” (singular or plural) is used herein to refer to a tire present in the physical environment of the system 100 and mounted on an identified aircraft (i.e., a tire still in service on the identified aircraft). Identified tires may include one or more known sensors for generating or capturing data (e.g., data corresponding to the operating environment of the identified aircraft or a portion thereof). Sensors may include a set of sensors for transmitting data about the operating characteristics of the identified tire. For example, sensors may include one or more speed sensors, one or more acceleration sensors, one or more sensors related to adhesion friction, one or more sensors related to braking, and / or a combination of sensors for collecting data about one or more aspects of the dynamics of the identified tire. The sensor can also transmit stored data about the identification of the identified tire (including but not limited to its traceability regarding production, distribution and / or storage, production date, retreading history (if applicable), and its location and installation history).

[0045] The term "historical information" (singular or plural) is used herein to refer to data corresponding to the historical flights of the identified aircraft to which the identified tires are mounted. This data may include, but is not limited to: data corresponding to the departure and / or arrival dates of the identified aircraft, data on airports visited by the identified aircraft (this data can be collected from various sources including historical flight data attributed to the airline 104 to which the identified aircraft belongs), data on the aircraft (including but not limited to manufacturer, aircraft model version, identification code, etc.), and weather and / or climate conditions 106 during the historical flights.

[0046] The term "general information" (singular or plural) is used herein to refer to data corresponding to an identified tire. This data may include, but is not limited to: the size of the identified tire (which may be indicated by the tire type and / or its nomenclature), construction code (e.g., "-" for bias ply, "R" for radial), traceability information regarding production (e.g., the name and / or trademark of the identified tire's manufacturer, date and place of manufacture, distribution and / or storage), serial number, rated load, rated speed, expected tread landings or LPT and / or unique identifier. For example, for a 52x21.0R22 tire, the number "52" indicates the tire diameter in inches, the number "21.0" indicates the cross-sectional width at the widest point of the inflated new tire, the letter "R" indicates a radial tire, and the number "22" indicates the rim diameter in inches.

[0047] General information may also include the identified tire mounting location 110 (including installation history). It should be noted that... Figure 4 The installation locations shown are given as examples.

[0048] General information may also include the retreading level of the identified tire (if applicable). The term "retreading" is used here to refer to the method of restoring a worn tire to a suitable operating condition by replacing the tread rubber and one or more layers. In the retreading process, the old tread product is removed and replaced with new material. Retreading is a known and prescribed method in the aerospace industry. Data corresponding to the retreading level of the identified tire is managed by the airline 104 and / or the manufacturer (indicated by reference numeral 108).

[0049] In one embodiment of the invention, historical information and / or general information may be generated and / or managed at least in part by one or more airports (or by a network of one or more airports including specific airports). In this embodiment, system 100 is designed to simulate the destination airport of the identified aircraft based on historical information. It should be noted that airline 104 manages its aircraft in different ways, and therefore its management model can be modified. Thus, in this embodiment, by personalizing the simulation of the destination airport, general information can also be used to clarify the relationship between historical information and the attributes of the identified aircraft.

[0050] The communication network 102 of system 100 includes one or more communication devices (not shown) that capture and collect data and transmit it to server 102a. One or more communication devices include one or more portable devices, such as mobile network devices (e.g., mobile phones, laptops, one or more portable devices connected to the network including “augmented reality” and / or “virtual reality” devices, and / or any combination and / or any equivalent). In all cases, the communication devices may include clothing and / or wearable devices connected to the network and worn by one or more operators (where each operator is a human or a known device such as a robot). For example, a monitoring device worn by an airline pilot can monitor flight conditions via video and transmit the corresponding data (i.e., historical information of identified tires) to server 102a of communication network 102.

[0051] One or more communication devices may also include one or more remote computers capable of transmitting data via communication network 102. For example, a portable device of system 100 may send historical and / or general information about the identified tires to a remote computer of system 100. Based on the transmitted data, the remote computer may send a report on the maintenance of the identified tires to the portable device, indicating plans for anticipated commitments (e.g., recommending discontinuation of the identified tires because the predictive model of the present invention indicates that their tread depth is too low).

[0052] The communication network 102 may include wired or wireless connections and may employ any data transmission protocols known to those skilled in the art. Examples of wireless connections may include, but are not limited to: radio frequency (RF) connections, satellite connections, (analog or digital) cellular or mobile phone connections, Connections include Wi-Fi connections, infrared connections, ZigBee connections, local area network (LAN) connections, wireless local area network (WLAN) connections, wide area network (WAN) connections, near field communication (NFC) connections, connections according to other wireless communication standards and configurations, their equivalents, and combinations of these elements.

[0053] It should be understood that communication network 102 implies the use of one or more processors as understood by anyone skilled in the art. The term "processor" (or, alternatively, the term "programmable logic circuit") (singular or plural) refers to one or more means capable of processing and analyzing data and including one or more software packages for processing said data (e.g., including one or more integrated circuits, one or more controllers, one or more microcontrollers, one or more microcomputers, one or more programmable logic controllers (PLCs), one or more application-specific integrated circuits, one or more machine learning algorithms, and / or one or more other known equivalent programmable circuits known to those skilled in the art).

[0054] Therefore, this invention utilizes artificial intelligence (or "AI")-based methods and tools to complete the transmission of some information (based on received historical and general information). Machine learning analysis is performed using a machine learning model, such as an artificial neural network comprising multiple layers. The machine learning analysis receives processed data (i.e., historical and general information of identified tires) from the communication network 102 as input. In this embodiment, the data is received by inputting it into the various layers of the machine learning model. The algorithm is capable of continuous improvement for all aircraft tires, thereby ensuring that the system 100 improves itself from its acquired experience, particularly in selecting between maintenance plans and plans to inspect identified tires.

[0055] Refer again Figure 3 and Figure 4 And further reference Figure 5 Schemes of the method for predicting wear condition of the present invention will now be disclosed. In each embodiment, the prediction method is implemented by a computer, such that system 100 can construct a model (or “prediction model”) for predicting the wear condition of an identified tire (in terms of the number of landings or the wear condition of an aircraft tire).

[0056] refer to Figure 5 The figure illustrates one embodiment of the prediction method 200 of the present invention, in which a prediction model can predict the remaining number of landings of an identified tire to determine its remaining service life. The remaining service life of the identified tire includes the number of landings (or tread landings or LPT) of the identified aircraft corresponding to the predicted wear state of the identified tire. Therefore, in this embodiment, the method involves constructing a model for predicting the remaining number of landings of an identified tire to determine a maintenance schedule. The maintenance schedule for the identified tire includes a choice between a repair schedule (where the identified tire is still installed on the identified aircraft) and an inspection schedule (where the identified tire is inspected to ensure it is removed at the correct time for refurbishment or introduced into a scrap line including recycling).

[0057] At the beginning Figure 5 In the prediction method 200 shown, a prediction model is constructed based on one or more parameters affecting the wear of identified tires. To construct the prediction model, the method includes a step 202 of introducing the influence parameters of the identified tires, which is performed by system 100. In this step, the influence parameters of the identified tires are acquired by a communication device of system 100. The acquired influence parameters include data corresponding to historical and general information about the identified tires. This data is stored (e.g., in one or more databases 120, 122 of system 100) (see...). Figure 3 Furthermore, these data are updated continuously or intermittently throughout the prediction method.

[0058] The introductory step includes creating a training database (or “database”) for incorporating wear states into the predictive model. The created training database may include a reference database of aircraft tires to be installed on the identified aircraft. This database may include previously created references (e.g., a table of remaining service life of aircraft tires at multiple wear thresholds). The database may include parameters corresponding to multiple commercially available tires (including parameters that form part of the general information described above). The specific source of the wear states at the multiple wear levels is not critical to the method described herein, which is equally effective using data obtained solely from tires that have reached the wear threshold. For example, the system may implement the method of the invention using data obtained from measurements performed on tires received at the factory after removal and installation of the assembly.

[0059] This system implements the method of the invention to help airlines send information about the wear progress of identified tire treads to staff at the airports served by the airline, perhaps during aircraft inspections. The system utilizes data from measurements of wear thresholds of “stocked” tire treads to prepare personalized alerts to be sent to the aircraft operator (i.e., the airline or the airport served by the airline) to predict the maintenance needs of identified tires.

[0060] The created database can include images of wear profiles corresponding to known wear states and the corresponding number of landings. The wear profiles of identified tires are defined by the contours of the outer surface of the tread during use. Therefore, unused new tread is considered the limit data (envelope) worn to 0%, and tread meeting the manufacturer's recommended removal conditions is considered the limit data worn to 100%. The curves described herein can include one or more (3D) surface measurements (performed beforehand) reproducing all or part of the outer surface of the tread. Furthermore, the curves described herein can include one or more (2D) linear measurements (performed beforehand) in one or more planes containing the axis of symmetry of the tread. Therefore, the training database includes anticipated images (and corresponding data) corresponding to the (3D, 2D, 1D) curves of worn tires and aircraft landings.

[0061] The prediction method 200 of the present invention (the method is as follows) Figure 5 The method (shown) further includes step 204, which involves training a predictive model to predict the remaining number of landings corresponding to reaching a removal threshold for the identified tire. As used herein, the "removal threshold" for the identified tire refers to the minimum regulatory depth at which the tire must be replaced (or another threshold defined by the user). In this step, the machine learning method receives data from acquired influence parameters (i.e., historical and general information about the identified tire) and a created training database as input. After system 100 acquires historical and general information corresponding to the identified tire, the processor can collect known wear states corresponding to the number of landings performed on the identified tire to construct a predictive model.

[0062] In one implementation, the machine learning method employed in training step 204 includes a supervised learning method. Supervised learning methods may include one or more neural networks (e.g., autoencoders, ANNs, CNNs, RNNs, perceptrons, long short-term memory (LSTM), Hopfield networks, Boltzmann machines, deep belief networks, deconvolutional neural networks, generative adversarial networks (GANs), etc.) and their complements and equivalents. One or more CNNs can be trained using ground-truth data generated from data representing the influencing parameters (e.g., data incorporated into the training database described above).

[0063] In one implementation, training step 204 includes a supervised learning method, specifically a gradient boosting regressor (GBR) type supervised learning method. The GBR learning method receives data corresponding to historical and general information about the identified tires (including, but not limited to, the location of the identified tires on the identified aircraft, the retreading level of the identified tires, and the proportion of landings at each airport visited by the identified tires) as input data. The GBR learning method aims to predict the remaining number of landings (remaining LPT) corresponding to reaching a removal threshold for the identified tires. Therefore, the predictive output of the predictive model will be a prediction of the wear state of the identified tires (e.g., characterized by the date range of reaching the removal threshold for the identified tires). For example, this enables the preventative removal of tires with excessively high wear (i.e., wear exceeding a predetermined wear threshold to ensure proper tire operation). For instance, in some usage scenarios, the wear rate on the shoulder of an identified tire may cause the working layer to appear before reaching the slip limit. Providing wear profiles of the worn aircraft tires can prevent immediate and unpredictable removal.

[0064] The prediction method 200 of the present invention further includes a step 206 of predicting the remaining number of landings (or “remaining LPT”) before reaching the removal threshold of the identified tires. The remaining number of landings for the identified tires can be calculated based on data corresponding to impact parameters for future landings. It should be noted that the future landing data can be hypothetical or real, depending on the type of management employed by the airline for its fleet and its ability to determine the routes to which aircraft will be assigned within a given timeframe. The offset between the actual wear condition and the predicted number of landings is represented by a calculation error, which indicates changes in the tread of the identified tires. This calculation error can be fed into a training database (as described above) to improve the predictive capability of the prediction model.

[0065] The prediction method 200 further includes a comparison step 208, in which the remaining LPT (output by the prediction model) before reaching the removal threshold of the identified tire is compared with the value of the removal threshold. This removal threshold is defined by the user (e.g., an airline) and / or the manufacturer of the identified tire to organize maintenance operations.

[0066] In the comparison step, if the remaining number of landings output by the prediction model is higher than the removal threshold defined for the identified tire, system 100 instructs plan 210 for repairing the identified tire. Similarly, if the remaining number of landings output by the prediction model is equal to or lower than the removal threshold, system 100 instructs plan 212 for inspecting the identified tire. When the tread of an identified tire reaches or is close to reaching the wear removal threshold, system 100 considers the wear condition of the worn tire obtained from wear condition measurements performed on aircraft tires received after use at the factory. In cases where an identified tire must be replaced with another tire of the same type, the prediction model is updated (see [link to relevant documentation]) during the process of replacing the worn tire with an identified tire that is considered to have a new tread. Figure 5 (See reference numeral 214 in the attached figure). It should be noted that replacing an identified tire may include retreading or recycling (or other end-of-life treatment) the identified tire.

[0067] Refer again Figure 3 In all embodiments of the prediction method of the present invention, the method may further include an optional simulation step 400, which is designed to simulate the destination airport of the identified aircraft. In these embodiments, the simulation is designed to simulate the destination airport of a given aircraft based on historical flight data of the airline to which the given aircraft belongs. The simulation is performed via a Markov chain, where each state represents an airport, and connections between airports represent the probability of departing from one airport and landing at another. These probabilities are estimated based on historical information to allow for stochastic scheduling when the airline is unaware of the future usage of its aircraft. The use of Markov chains enables the learning of key features of the data to develop and evaluate corresponding uncertainty models.

[0068] In an embodiment of the method of the present invention aimed at simulating the destination airports of an identified aircraft, simulation of multiple destinations makes it possible to calculate the proportion of landings of the identified tires at each of the multiple airports identified in the simulation (thus utilizing both "real" and simulated data). This is achieved through a Monte Carlo loop (see...). Figure 3 (Referring to the attached figure 402) Step 400 of simulating the destination is repeated multiple times to complete the lifespan of the identified tires. Subsequently, each simulation is a subject predicted by the above model to finally obtain the predicted landing number distribution of the identified tires.

[0069] Example:

[0070] *The identified tire installed at location P of the identified aircraft has Z LPTs at time t. The knowledge of the LPT distribution at location P is obtained through historical information.

[0071] *Considering the possible random plotting of trajectories, for the theoretical LPT (i.e., LPT_th) t Each random plot will include the simulated trajectory LPT_th. t -Z destination.

[0072] *By calculating the percentage of frequent occurrences on the complete trajectory of the identified tire's lifespan, a predictive model is applied to each complete trajectory to obtain the distribution of the total number of LPTs that can be performed by the identified tire.

[0073] Assuming airlines employ more planning management and know the aircraft's destination weeks in advance, they can replace simulations with the introduction of real destinations (or even opt for intermediate semi-deterministic management if the airline doesn't fully know how its aircraft will be used but is able to define relatively probable destinations). Before implementing Markov chains to identify long-term time-dependent models of the prediction process (which reveal the behavior of relevant characteristics required for uncertainty models), the nature of the datasets (historical and general information) used to evaluate the models must be studied and understood.

[0074] In all embodiments of the prediction method of the present invention, one or more steps may be performed iteratively.

[0075] Although the implementation of the prediction method of the present invention has been described herein using a neural network through a machine learning model, other types of machine learning models can be used. Other types of machine learning models include, but are not limited to, those utilizing linear regression, logistic regression, decision trees, support vector machines, Naive Bayes methods, k-nearest neighbor (k-NN) algorithms, k indicating a group, random forests, dimensionality reduction algorithms, and gradient descent.

[0076] Therefore, this invention uses readily available data to predict the lifespan of aircraft tires (and thus maintenance schedules), thereby creating a reliable predictive model. While human experts demonstrate flexibility in making plans for maintenance or aircraft tire inspections, they lack the analytical capabilities to determine the vast amounts of data required to make real-time decisions about whether an identified aircraft must be repaired. Employees also lack the analytical capabilities to make such decisions. Therefore, it is necessary to employ methods that utilize data on the various rules of experience employed by human experts to modify airline operations in order to predict maintenance reports.

[0077] System 100 may include pre-programmed management information. For example, prediction method adjustments may be associated with parameters of the typical physical environment in which system 100 operates (e.g., parameters of the airport being visited). In some embodiments, for example, system 100 (and / or devices associated with system 100) may receive voice commands or other audio data indicating the start or stop of capturing data corresponding to historical and / or general information of identified tires, or the start or stop of movement of communication devices. Requests made to system 100 may include requests for the current state of an automatic prediction method cycle. The generated response may be represented in an auditory, visual, tactile manner (e.g., through a tactile interface) and / or in a virtual and / or augmented manner. This response, along with the corresponding data, may be recorded in a neural network.

[0078] The terms "at least one" and "one or more" are used interchangeably. The given range "between a and b" includes the values ​​"a" and "b".

[0079] Although specific embodiments of the disclosed apparatus have been described and illustrated, it will be understood that various changes, additions, and modifications can be made without departing from the spirit or scope of the invention. Therefore, no limitation should be imposed on the scope of the invention other than that set forth in the appended claims.

Claims

1. A computer-implemented prediction method (200) for predicting the number of remaining landings (remaining LPT) corresponding to reaching a threshold for the removal of identified tires installed on an identified aircraft, the prediction method comprising the following steps: - A step (202) of introducing parameters affecting the identified tires into a system (100) performing the prediction method, the system (100) including a communication network (102) for managing data input into the system, the communication network having one or more communication servers (102a), the one or more communication servers (102a) managing data corresponding to the parameters affecting the identified tires, and having at least one communication device for capturing this data and sending it to the server, the step including the following steps: - A step of acquiring parameters affecting the identified tires, said step being performed by a communication device of the system (100), wherein the acquired parameters include data corresponding to historical and general information of the identified tires; and - The step of creating a wear state training database, which is incorporated into a model for predicting the remaining number of landings corresponding to the removal threshold of identified tires; - The step (204) of training a prediction model to predict the remaining number of landings corresponding to the removal threshold of the identified tires, wherein the machine learning method receives the obtained influence parameters and data from the training database as input, enabling the processor to obtain the known wear state corresponding to the number of landings performed by the identified tires; - Step (206) of predicting the remaining number of landings (remaining LPT) before reaching the removal threshold of identified tires, wherein the remaining number of landings of identified tires is calculated based on data corresponding to the influencing parameters; and - Comparison step (208), wherein the remaining number of landings (remaining LPT) before reaching the removal threshold of the identified tires, output by the prediction model, is compared with the value of the removal threshold of the identified tires, so that the system (100) creates a maintenance plan for the identified tires. The prediction method (200) further includes a step (400) of simulating the destination airport of the identified aircraft based on historical flight data. The simulation step (400) is performed via a Markov chain, in which each state represents an airport, and the connection between each airport represents the probability of departing from one airport and landing at another airport.

2. The prediction method (200) according to claim 1, wherein, Influencing parameters include: - Historical information, which includes data corresponding to the historical flights of the identified aircraft with the identified tires installed; and - General information, which includes data corresponding to the identified tires, including the installation location of the identified tires on the identified aircraft.

3. The prediction method (200) according to claim 1 or 2, wherein, The maintenance plan created by the system (100) in comparison step (208) includes: - A plan to repair identified tires when the remaining landings (remaining LPT) output by the predictive model are higher than the removal threshold defined for identified tires; and - Check the plan for identified tires when the remaining number of landings (remaining LPT) output by the prediction model is equal to or less than the removal threshold defined for identified tires.

4. The prediction method (200) according to claim 1, wherein, Supervised learning methods take the acquired influence parameters and data from the training database as input, enabling the processor to acquire the known wear state corresponding to the number of landings performed by the identified tires in order to build a predictive model.

5. The prediction method (200) according to claim 4, wherein, Supervised learning methods include those of the Gradient Boosting Regressor (GBR) type.

6. The prediction method (200) according to claim 4, wherein, The training database includes images of wear curves corresponding to known wear states and the number of landings performed by the corresponding identified tires.

7. The prediction method (200) according to claim 1, wherein, In the prediction step (206), the remaining number of landings (remaining LPT) of the identified tires is calculated based on the data corresponding to the impact parameters of future landings.

8. The prediction method (200) according to claim 1, wherein, The simulation steps (400) are repeated multiple times using a Monte Carlo cycle (402) to predict the scrapping of identified tires.

Images