A method for predicting aircraft brake system faults based on big data analysis

Abstract

The invention discloses a method for predicting aircraft brake system faults based on big data analysis, and relates to the technical field of aircraft brake fault inspection. The method is based on the historical time series data of the fault alarm state parameters and multiple non-fault alarm state parameters of the aircraft brake system, on the one hand, calculates the correlation coefficients of each pair of fault alarm state parameters and non-fault alarm state parameters, on the other hand, extracts the time series data of each non-fault alarm state parameter before the fault alarm is triggered to build a corresponding vector library, and obtains the corresponding aircraft brake system fault prediction probability by performing ANN vector retrieval on the vector library, and then obtains the total prediction probability of the aircraft brake system fault according to all the correlation coefficients and the prediction probability, and finally triggers the execution of the aircraft brake system fault warning action when the total prediction probability reaches a preset probability threshold, so that based on the historical data of the operating parameters of the aircraft brake system, the ability of early fault prediction and warning can be provided through big data analysis.

Description

Technical Field

[0001] The present invention belongs to the technical field of aircraft brake fault inspection, and in particular relates to an aircraft brake system fault prediction method based on big data analysis. Background Art

[0002] The aircraft braking system refers to the device that slows down or stops the aircraft. It consists of three parts: (1) the wheel brake system, which consists of two parts: the brake and the control. The main components include various valves, decompression accelerators, brake control boxes and wheel speed sensors; (2) the reverse thrust brake device, which turns the exhaust gas of the jet engine at a certain angle and discharges it diagonally forward, generating a synthetic thrust in the opposite direction of the aircraft's movement, thereby slowing down the aircraft; (3) the pneumatic brake device, which is a speed brake or spoiler installed on the wing or fuselage. When in use, it opens outward to increase air resistance and reduce the speed of the aircraft.

[0003] At present, the fault diagnosis scheme of aircraft brake system mainly relies on experience and manually sets the threshold range of a single signal. When the single signal sensor value exceeds the corresponding threshold range, the corresponding signal sensor is judged to be faulty. However, since the aforementioned aircraft brake system fault diagnosis scheme is based on the real-time monitoring results of the signal sensor value for fault diagnosis, it lacks the ability to predict and warn faults in advance, which will delay the timing of fault response and thus affect flight safety. Therefore, how to predict and warn aircraft brake system faults based on the historical data of the operating parameters of the aircraft brake system through big data analysis to promote the purpose of timely response to faults is a topic that technicians in this field urgently need to study. Summary of the invention

[0004] The purpose of the present invention is to provide an aircraft brake system fault prediction method, computer equipment, computer-readable storage medium and computer program product based on big data analysis, so as to solve the problem that the existing aircraft brake system fault diagnosis scheme lacks the ability to predict and warn faults in advance, which easily leads to delays in fault response opportunities.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] In a first aspect, a method for predicting aircraft brake system failures based on big data analysis is provided, comprising:

[0007] Acquire historical time series data of all operating parameters of the aircraft brake system, wherein all the operating parameters include fault alarm state parameters and multiple non-fault alarm state parameters, and the historical time series data includes The unit time and the corresponding operating parameters are described in The historical parameter value of each unit moment in the unit moments;

[0008] For each pair of fault alarm state parameters and non-fault alarm state parameters among all the operating parameters, a corresponding correlation coefficient is calculated according to the historical time series data of the corresponding two parameters;

[0009] For each pair of fault alarm state parameters and non-fault alarm state parameters, first determine each fault alarm triggering moment according to the historical time series data of the corresponding fault alarm state parameters, then extract each previous time series data corresponding to each fault alarm triggering moment from the historical time series data of the corresponding non-fault alarm state parameters, and finally add each previous time series data as a first vector to the corresponding vector library, wherein the previous time series data refers to the time series data traced back from a fixed time before the corresponding fault alarm triggering moment to The historical time series data corresponding to the time window of the length of time, the fixed time refers to the time interval from the corresponding fault alarm triggering time. At a certain moment in time, Indicates the preset first duration. Indicates the preset second duration;

[0010] For each non-fault alarm state parameter in the plurality of non-fault alarm state parameters, first extract from the corresponding historical time series data the data traced back from the current moment to the The historical time series data corresponding to the time window of the time length is extracted, and the extraction result is used as the second vector, and then the ANN vector search is performed on the corresponding vector library based on the second vector to obtain a vector with the maximum similarity with the second vector, and finally the maximum similarity is used as the corresponding aircraft brake system failure prediction probability;

[0011] According to the correlation coefficients corresponding to the non-fault alarm state parameters and the predicted probability of the aircraft brake system failure, a total predicted probability of the aircraft brake system failure is obtained by superimposing and calculating;

[0012] If the predicted total probability of the aircraft brake system failure reaches a preset probability threshold, an aircraft brake system failure warning action is triggered.

[0013] Based on the above invention content, a new scheme for predicting and warning aircraft brake system faults based on historical data of aircraft brake system operating parameters is provided, that is, according to the historical time series data of the fault alarm state parameters and multiple non-fault alarm state parameters of the aircraft brake system, on the one hand, the correlation coefficients of each pair of fault alarm state parameters and non-fault alarm state parameters are calculated, and on the other hand, the time series data of each non-fault alarm state parameter before the fault alarm is triggered are extracted to construct a corresponding vector library, and the corresponding aircraft brake system fault occurrence prediction probability is obtained by performing ANN vector retrieval on the vector library, and then the total predicted probability of aircraft brake system fault occurrence is obtained by superposition calculation according to all the correlation coefficients and the predicted probabilities, and finally, when the predicted total probability reaches a preset probability threshold, the aircraft brake system fault warning action is triggered to execute, so that based on the historical data of the aircraft brake system operating parameters, the ability to predict and warn faults in advance can be provided through big data analysis, thereby promoting the purpose of timely responding to faults, ensuring flight safety, and facilitating practical application and promotion.

[0014] In a possible design, the fault alarm state parameter is a wheel speed sensor fault alarm state parameter, a brake pressure sensor fault alarm state parameter or a brake temperature sensor fault alarm state parameter.

[0015] In one possible design, the plurality of non-fault warning status parameters include aircraft speed, wheel speed, brake pressure, brake temperature and / or cumulative number of brakes.

[0016] In a possible design, for each pair of fault alarm state parameters and non-fault alarm state parameters, the corresponding correlation coefficient is calculated based on the historical time series data of the corresponding two parameters, including:

[0017] For a pair of fault alarm state parameters and non-fault alarm state parameters among all the operating parameters, a section of time series data is randomly selected from the historical time series data of the corresponding fault alarm state parameters to form a first sample. , and also extracting the first sample from the historical time series data corresponding to the non-fault alarm state parameter Another period of time series data during the same period constitutes the second sample ;

[0018] For the first sample Perform a normal distribution KS test to calculate the first test statistic value , and also for the second sample Perform the normal distribution KS check to calculate the second check statistic value ;

[0019] If the first check statistic value and the second check statistic value are greater than the preset threshold, the first sample is calculated according to the following formula With the sample The correlation coefficient :

[0020]

[0021] In the formula, represents a positive integer, In the first sample The Sample values, In the second sample The Sample values, Indicates the first sample The sample mean of The second sample The sample mean of ;

[0022] The correlation coefficient As the correlation coefficient corresponding to the pair of fault alarm state parameters and non-fault alarm state parameters.

[0023] In a possible design, after calculating corresponding correlation coefficients for each pair of fault alarm state parameters and non-fault alarm state parameters according to historical time series data of the corresponding two parameters, the method further includes:

[0024] At the beginning of each operation and maintenance task of the aircraft brake system, for each pair of fault alarm state parameters and non-fault alarm state parameters, the corresponding fault alarm state parameters and the parameters traced back from the current time to the The historical time series data corresponding to the time window of the length of time is used as the corresponding first verification sample, and the corresponding non-fault alarm state parameters and traced back from the current time to The historical time series data corresponding to the time window of the same length is used as the corresponding second verification sample, and then the corresponding correlation coefficient is updated and calculated according to the first verification sample and the second verification sample, wherein, Indicates the third preset duration.

[0025] In a possible design, after adding each of the previous time series data as a first vector to the corresponding vector library for each pair of the fault alarm state parameter and the non-fault alarm state parameter, the method further includes:

[0026] At the beginning of each operation and maintenance task of the aircraft brake system, for each pair of fault alarm state parameters and non-fault alarm state parameters, expired vectors are removed from the corresponding vector library, wherein the expired vector refers to the time from the corresponding fault alarm triggering time to the current time exceeding Any vector of duration, Indicates the fourth preset duration.

[0027] In a possible design, based on the correlation coefficients corresponding to the various non-fault alarm state parameters and the predicted probability of occurrence of the aircraft brake system fault, the total predicted probability of occurrence of the aircraft brake system fault is obtained by superimposing and calculating, including:

[0028] According to the correlation coefficients corresponding to the non-fault alarm state parameters and the predicted probability of aircraft brake system failure, the total predicted probability of aircraft brake system failure is calculated according to the following formula: :

[0029]

[0030] In the formula, represents the total number of parameters of the plurality of non-fault alarm state parameters, Indicates less than or equal to A positive integer, Indicates the first of the plurality of non-fault alarm state parameters The influence weight coefficient corresponding to the non-fault alarm state parameter is: Indicates that The correlation coefficient corresponding to the non-fault alarm state parameters, Indicates less than or equal to A positive integer, Indicates the first of the plurality of non-fault alarm state parameters The correlation coefficient corresponding to the non-fault alarm state parameters, Indicates that The predicted probability of aircraft brake system failure corresponding to the non-fault alarm state parameters.

[0031] In a possible design, when the fault alarm state parameter is a wheel speed sensor fault alarm state parameter, a brake pressure sensor fault alarm state parameter or a brake temperature sensor fault alarm state parameter, before triggering the execution of the aircraft brake system fault warning action, the method further includes:

[0032] For each fault alarm triggering moment, extract the historical parameter value at the corresponding fixed moment from the historical time series data of the sensor parameter corresponding to the fault alarm state parameter as a model output item, and take all non-sensing parameters and the corresponding first vector together as model input items, and then take the model input items and the model output items as corresponding sample data, wherein the non-sensing parameter refers to a parameter that is not a sensor parameter corresponding to the fault alarm state parameter among the multiple non-fault alarm state parameters;

[0033] Applying all the sample data to perform model training on the Attention-LSTM model to obtain a sensor parameter value estimation model corresponding to the fault alarm state parameter;

[0034] Triggering the execution of aircraft brake system fault warning actions, including:

[0035] The second vector of all the non-sensing parameters is introduced into the sensing parameter value estimation model, and the parameter estimation value of the sensing parameter corresponding to the fault alarm state parameter at the current moment is obtained as output.

[0036] In a possible design, the non-sensing parameter is also a parameter that has not currently triggered a fault warning event, wherein the currently not triggered fault warning event means that when the corresponding parameter corresponds to another fault alarm state parameter, the predicted total probability of the aircraft braking system failure corresponding to the other fault alarm state parameter and at the current moment does not reach the preset probability threshold.

[0037] In a possible design, the parameters of the Attention-LSTM model are searched using a grid search algorithm or a Bayesian search algorithm.

[0038] In a second aspect, the present invention provides a computer device comprising a memory, a processor and a transceiver which are communicatively connected in sequence, wherein the memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program to execute the aircraft braking system fault prediction method based on big data analysis as described in the first aspect or any design of the first aspect.

[0039] In a third aspect, the present invention provides a computer-readable storage medium having instructions stored thereon. When the instructions are executed on a computer, the aircraft braking system fault prediction method based on big data analysis as described in the first aspect or any design of the first aspect is executed.

[0040] In a fourth aspect, the present invention provides a computer program product, including a computer program or instructions, which, when executed by a computer, implements the aircraft brake system fault prediction method based on big data analysis as described in the first aspect or any design of the first aspect.

[0041] Beneficial effects of the above scheme:

[0042] (1) The present invention creatively provides a new solution for aircraft brake system fault prediction and early warning based on the historical data of aircraft brake system operating parameters, that is, based on the historical time series data of the aircraft brake system's fault alarm state parameters and multiple non-fault alarm state parameters, on the one hand, the correlation coefficients of each pair of fault alarm state parameters and non-fault alarm state parameters are calculated, and on the other hand, the time series data of each non-fault alarm state parameter before the fault alarm is triggered are extracted to construct a corresponding vector library, and the corresponding aircraft brake system fault occurrence prediction probability is obtained by performing ANN vector retrieval on the vector library, and then the total predicted probability of the aircraft brake system fault occurrence is obtained by superposition calculation based on all the correlation coefficients and predicted probabilities, and finally, when the predicted total probability reaches a preset probability threshold, the aircraft brake system fault early warning action is triggered. In this way, based on the historical data of the aircraft brake system's operating parameters, the ability to predict and warn faults in advance can be provided through big data analysis, thereby promoting the purpose of timely responding to faults, ensuring flight safety, and facilitating practical application and promotion;

[0043] (2) The accuracy of the comprehensive prediction results of aircraft brake system faults during the execution of the operation and maintenance tasks can also be ensured by dynamically updating the correlation coefficient at the beginning of each operation and maintenance task of the aircraft brake system;

[0044] (3) At the beginning of each operation and maintenance task of the aircraft brake system, the expired vectors in the vector library can be removed to ensure the accuracy of the comprehensive prediction results of the aircraft brake system faults during the operation and maintenance task execution;

[0045] (4) In case of sensor fault warning, the sensor value can be estimated based on the most recent time series data of non-sensing parameters, which is further beneficial for fault response and further ensures flight safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0047] Figure 1A schematic flow chart of a method for predicting aircraft brake system faults based on big data analysis provided in an embodiment of the present application.

[0048] Figure 2 A schematic diagram of the structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0049] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the present invention will be briefly introduced below in combination with the drawings and the description of the embodiments or the prior art. Obviously, the following description of the structures of the drawings is only some embodiments of the present invention. For ordinary technicians in this field, other embodiments can be obtained based on these embodiments without creative work. It should be noted that the description of these embodiments is used to help understand the present invention, but does not constitute a limitation of the present invention.

[0050] It should be understood that although the terms first and second, etc. may be used herein to describe various objects, these objects should not be limited by these terms. These terms are only used to distinguish one object from another object. For example, a first object can be referred to as a second object, and similarly, a second object can be referred to as a first object without departing from the scope of the exemplary embodiments of the present invention.

[0051] It should be understood that the term "and / or" that may appear in this document is merely a description of the association relationship between associated objects, indicating that there may be three relationships. For example, A and / or B can indicate three situations: A exists alone, B exists alone, or A and B exist at the same time. For another example, A, B and / or C can indicate the existence of any one of A, B and C or any combination of them. The term " / and" that may appear in this document describes another type of association object relationship, indicating that there may be two relationships. For example, A / and B can indicate two situations: A exists alone or A and B exist at the same time. In addition, the character " / " that may appear in this document generally indicates that the previous and next associated objects are in an "or" relationship.

[0052] Example

[0053] like Figure 1 As shown, the aircraft brake system fault prediction method provided in the first aspect of this embodiment and based on big data analysis can be, but is not limited to, executed by a computer device with certain computing resources, such as an airborne server or a controller of the aircraft brake system. Figure 1 As shown, the aircraft brake system fault prediction method may include but is not limited to the following steps S1 to S6.

[0054] S1. Obtain the historical time series data of all operating parameters of the aircraft brake system, wherein all operating parameters include but are not limited to fault alarm state parameters and multiple non-fault alarm state parameters, etc., and the historical time series data include but are not limited to the time series of The unit time and the corresponding operating parameters are described in The historical parameter values ​​of each unit moment in each unit moment, etc.

[0055] In the step S1, specifically, the fault alarm state parameters include, but are not limited to, wheel speed sensor fault alarm state parameters, brake pressure sensor fault alarm state parameters or brake temperature sensor fault alarm state parameters, etc., and their corresponding historical parameter values ​​can be represented by "1" to indicate that the aircraft brake system is in a fault alarm state at the corresponding unit time, and "0" to indicate that the aircraft brake system is in a non-fault alarm state at the corresponding unit time. Specifically, the multiple non-fault alarm state parameters include, but are not limited to, aircraft speed, wheel speed, brake pressure, brake temperature and / or cumulative number of brakes, etc. The unit time can be, but is not limited to, every 10 seconds, every 30 seconds or every 1 minute, etc., that is, the historical time series data includes, but is not limited to, the number of times the aircraft brake wheel was in the most recent The historical parameter values ​​at the minute, The historical parameter values ​​at the last minute, ..., the historical parameter values ​​at the last second minute, and the historical parameter values ​​at the last minute, etc. In addition, the historical time series data of all the operating parameters can be routinely collected and recorded by related equipment of the existing aircraft braking system.

[0056] S2. For each pair of fault alarm state parameters and non-fault alarm state parameters in all the operating parameters, a corresponding correlation coefficient is calculated based on the historical time series data of the corresponding two parameters.

[0057] In step S2, considering that the influence of different non-fault alarm state parameters on the fault alarm state parameters must be different, it is necessary to calculate the correlation coefficients of each pair of fault alarm state parameters and non-fault alarm state parameters, so as to be used for the subsequent comprehensive prediction of aircraft brake system faults and ensure the accuracy of the comprehensive prediction results. Specifically, for each pair of fault alarm state parameters and non-fault alarm state parameters, the corresponding correlation coefficients are calculated based on the historical time series data of the corresponding two parameters, including but not limited to the following steps S21 to S23.

[0058] S21. For a pair of fault alarm state parameters and non-fault alarm state parameters among all the operating parameters, randomly extract a section of time series data from the historical time series data of the corresponding fault alarm state parameters to form a first sample. , and also extracting the first sample from the historical time series data corresponding to the non-fault alarm state parameter Another period of time series data during the same period constitutes the second sample .

[0059] In the step S21, for example, the time series data from 00:00 on July 1 to 24:00 on August 31 can be extracted from the historical time series data of the fault alarm state parameter to form the first sample. , and also extracting the time series data from 00:00 on July 1 to 24:00 on August 31 from the historical time series data of the aircraft speed (i.e., one of the non-fault alarm state parameters) to form the second sample .

[0060] S22. For the first sample Perform a normal distribution KS test to calculate the first test statistic value , and also for the second sample Perform the normal distribution KS check to calculate the second check statistic value .

[0061] S23. If the first check statistic value and the second check statistic value are greater than the preset threshold, the first sample is calculated according to the following formula With the sample The correlation coefficient :

[0062]

[0063] In the formula, represents a positive integer, In the first sample The Sample values, In the second sample The Sample values, Indicates the first sample The sample mean of The second sample The sample mean of .

[0064] In the step S23, the preset threshold may be 0.05, for example.

[0065] S24. The correlation coefficient As the correlation coefficient corresponding to the pair of fault alarm state parameters and non-fault alarm state parameters.

[0066] S3. For each pair of fault alarm state parameters and non-fault alarm state parameters, first determine each fault alarm triggering moment according to the historical time series data of the corresponding fault alarm state parameters, then extract each previous time series data corresponding to each fault alarm triggering moment from the historical time series data of the corresponding non-fault alarm state parameters, and finally add each previous time series data as the first vector to the corresponding vector library, wherein the previous time series data refers to the time series data traced back from a fixed time before the corresponding fault alarm triggering moment to The historical time series data corresponding to the time window of the length of time, the fixed time refers to the time interval from the corresponding fault alarm triggering time. At a certain moment in time, Indicates the preset first duration. Indicates the preset second duration.

[0067] In step S3, the triggering moment of each fault alarm is the first moment of each fault alarm state, and thus can be conventionally determined based on the historical time series data of the fault alarm state parameters. The duration can be determined based on the duration required for fault response, and generally needs to be greater than the shortest duration required for fault response, such as 10 minutes. The duration can be set based on historical experience, such as 1 hour. For example, if a fault alarm is triggered at 12:10, the fixed time is 12:00, and the time series data from 11:01 to 12:00 can be extracted from the historical time series data of the non-fault alarm state parameters as the first vector and added to the vector library.

[0068] S4. For each non-fault alarm state parameter in the plurality of non-fault alarm state parameters, first extract from the corresponding historical time series data the state parameters traced back from the current moment to the The historical time series data corresponding to the time window of the time length is extracted, and the extraction result is used as the second vector. Then, based on the second vector, an ANN vector search is performed on the corresponding vector library to obtain a vector with the maximum similarity with the second vector. Finally, the maximum similarity is used as the corresponding aircraft brake system failure prediction probability.

[0069] In step S4, based on the example in step S3, if the current time is 13:12, the time series data from 12:13 to 13:12 can be extracted from the historical time series data of the non-fault alarm state parameters as the second vector. The ANN (Approximate Nearest Neighbor) vector retrieval is an existing technology for processing similarity search in large-scale data sets, which is mainly used to quickly find data items similar to the query object in large data sets; this technology has applications in many fields, including vector databases, retrieval enhancement generation, large-scale information retrieval, recommendation systems, drug discovery, and image search. The core of ANN vector retrieval is to use efficient indexing and search algorithms to approximately find the nearest neighbors in the data set (note that it is not accurate to find the nearest neighbors). This approximation method allows a certain error while ensuring search efficiency, thereby significantly improving the search speed when processing large-scale data sets. In general, ANN vector retrieval technology can significantly improve the efficiency of similarity search in large-scale data sets while ensuring that the retrieval results have a certain degree of accuracy by using efficient indexing and search algorithms, as well as specific graph methods such as HNSW (Hierarchical Navigable Small World Graph, which is a graph method for vector retrieval. It is based on the idea of ​​skip lists and achieves efficient approximate nearest neighbor search by constructing a navigable small world graph. It organizes data by using a navigable small world graph at each layer to form a hierarchical structure, so that the retrieval process can gradually approach the target vector space from top to bottom, thereby improving the search efficiency. In the process of graph composition, HNSW can ensure the connectivity of the graph by cutting the edges to ensure the smooth progress of the search process). Therefore, by performing ANN vector retrieval on the vector library based on the second vector, the retrieval result can be accurately and efficiently obtained: a vector with the maximum similarity to the second vector.

[0070] S5. Based on the correlation coefficients corresponding to the various non-fault alarm state parameters and the predicted probability of occurrence of aircraft brake system failure, a total predicted probability of occurrence of aircraft brake system failure is obtained by superimposing and calculating.

[0071] In step S5, since the correlation coefficient can be positively correlated with the influence weight coefficient of the non-fault alarm state parameter on the fault alarm state parameter (i.e., the larger the correlation system, the larger the influence weight coefficient), the purpose of comprehensive prediction of aircraft brake system failure can be achieved based on the correlation coefficients corresponding to the various non-fault alarm state parameters and the predicted probability of aircraft brake system failure. Specifically, the total predicted probability of aircraft brake system failure is obtained by superimposing and calculating based on the correlation coefficients corresponding to the various non-fault alarm state parameters and the predicted probability of aircraft brake system failure, including but not limited to the following steps:

[0072] According to the correlation coefficients corresponding to the non-fault alarm state parameters and the predicted probability of aircraft brake system failure, the total predicted probability of aircraft brake system failure is calculated according to the following formula: :

[0073]

[0074] In the formula, represents the total number of parameters of the plurality of non-fault alarm state parameters, Indicates less than or equal to A positive integer, Indicates the first of the plurality of non-fault alarm state parameters The influence weight coefficient corresponding to the non-fault alarm state parameter is: Indicates that The correlation coefficient corresponding to the non-fault alarm state parameters, Indicates less than or equal to A positive integer, Indicates the first of the plurality of non-fault alarm state parameters The correlation coefficient corresponding to the non-fault alarm state parameters, Indicates that The predicted probability of aircraft brake system failure corresponding to the non-fault alarm state parameters.

[0075] S6. If the predicted total probability of the aircraft brake system failure reaches a preset probability threshold, an aircraft brake system failure warning action is triggered.

[0076] In step S6, the preset probability threshold may be 85% for example. The aircraft brake system fault warning action is specifically but not limited to an audible and visual alarm action, and may be triggered and executed based on conventional techniques.

[0077] Therefore, based on the aircraft brake system fault prediction method described in the aforementioned steps S1 to S5, a new scheme for aircraft brake system fault prediction and early warning based on the historical data of the aircraft brake system's operating parameters is provided, that is, according to the historical time series data of the aircraft brake system's fault alarm state parameters and multiple non-fault alarm state parameters, on the one hand, the correlation coefficients of each pair of fault alarm state parameters and non-fault alarm state parameters are calculated, and on the other hand, the time series data of each non-fault alarm state parameter before the fault alarm is triggered are extracted to construct a corresponding vector library, and the corresponding aircraft brake system fault occurrence prediction probability is obtained by performing ANN vector retrieval on the vector library, and then the aircraft brake system fault occurrence prediction total probability is obtained based on all the correlation coefficients and prediction probabilities, and finally, when the predicted total probability reaches a preset probability threshold, the aircraft brake system fault early warning action is triggered. In this way, based on the historical data of the aircraft brake system's operating parameters, the ability to predict and warn faults in advance can be provided through big data analysis, thereby promoting the purpose of timely responding to faults, ensuring flight safety, and facilitating practical application and promotion.

[0078] Based on the technical solution of the first aspect, this embodiment further provides a possible design of how to dynamically update the correlation coefficient. That is, after calculating the corresponding correlation coefficient for each pair of fault alarm state parameters and non-fault alarm state parameters according to the historical time series data of the corresponding two parameters, the method also includes but is not limited to the following steps: at the beginning of each operation and maintenance task of the aircraft braking system, for each pair of fault alarm state parameters and non-fault alarm state parameters, first update the corresponding fault alarm state parameters and the correlation coefficients traced back from the current moment to the The historical time series data corresponding to the time window of the length of time is used as the corresponding first verification sample, and the corresponding non-fault alarm state parameters and traced back from the current time to The historical time series data corresponding to the time window of the same length is used as the corresponding second verification sample, and then the corresponding correlation coefficient is updated and calculated according to the first verification sample and the second verification sample, wherein, Indicates the preset third duration. The time length is exemplified but not limited to 1 month. The update calculation details of the aforementioned correlation coefficient can be referred to the aforementioned steps S22 to S24, which will not be described in detail here.

[0079] Therefore, based on the above-mentioned possible design one, the accuracy of the comprehensive prediction results of the aircraft brake system faults during the execution of the operation and maintenance tasks can be ensured by dynamically updating the correlation coefficient at the beginning of each operation and maintenance task of the aircraft brake system.

[0080] Based on the technical solution of the first aspect, the present embodiment further provides a possible design of how to remove expired vectors in the vector library, that is, after adding the respective previous time series data as first vectors to the corresponding vector library for each pair of fault alarm state parameters and non-fault alarm state parameters, the method further includes but is not limited to the following steps: at the beginning of each operation and maintenance task of the aircraft braking system, for each pair of fault alarm state parameters and non-fault alarm state parameters, remove expired vectors from the corresponding vector library, wherein the expired vector refers to a time period from the corresponding fault alarm triggering moment to the current moment that exceeds Any vector of duration, Indicates the fourth preset duration. The duration is exemplified but not limited to 2 months. For example, if the current time is 13:00 on September 7, and a vector in the vector library is for a fault alarm triggering time of 12:10 on July 7 and the time series data from 11:01 to 12:00 on July 7 is extracted from the historical time series data of non-fault alarm state parameters, then the vector is the expired vector and needs to be removed.

[0081] Therefore, based on the aforementioned possible design 2, at the beginning of each operation and maintenance task of the aircraft brake system, the expired vectors in the vector library can be eliminated to ensure the accuracy of the comprehensive prediction results of the aircraft brake system faults during the execution of the operation and maintenance task.

[0082] Based on the technical solution of the first aspect, this embodiment also provides a possible design three for how to estimate sensor values ​​during sensor fault warning, that is, when the fault alarm state parameter is a wheel speed sensor fault alarm state parameter, a brake pressure sensor fault alarm state parameter or a brake temperature sensor fault alarm state parameter, before triggering the execution of the aircraft brake system fault warning action, the method also includes but is not limited to the following steps S601 to S602.

[0083] S601. For each fault alarm triggering moment, extract the historical parameter value at the corresponding fixed moment from the historical time series data of the sensor parameter corresponding to the fault alarm state parameter as the model output item, and take all non-sensing parameters and the corresponding first vector together as the model input item, and then take the model input item and the model output item as a corresponding sample data, wherein the non-sensing parameter refers to a parameter that is not a sensor parameter corresponding to the fault alarm state parameter among the multiple non-fault alarm state parameters.

[0084] In step S601, for example, if the fault alarm state parameter is a wheel speed sensor fault alarm state parameter, the sensor parameter corresponding to the fault alarm state parameter is the wheel speed, and the non-sensor parameter is the aircraft speed, brake pressure, brake temperature or cumulative number of brakes, etc. In addition, considering that for sensor fault alarm state parameters corresponding to brake pressure or brake temperature, etc., it is also possible to trigger the execution of the corresponding aircraft brake system fault warning action based on the aforementioned steps S1 to S6, resulting in the corresponding sensor value (such as brake pressure value or brake temperature value) being unreliable at this time. Therefore, in order to ensure the accuracy of the subsequent sensor value estimation result, the non-sensor parameter needs to be a parameter that does not currently trigger a fault warning event, wherein the current non-triggered fault warning event means that when the corresponding parameter corresponds to another fault alarm state parameter, the predicted total probability of the aircraft brake system fault corresponding to the other fault alarm state parameter and at the current moment does not reach the preset probability threshold. For example, the other fault alarm state parameter is a brake pressure sensor fault alarm state parameter corresponding to the brake pressure. If, based on the above steps S1 to S6, it is found that the corresponding predicted total probability of the aircraft brake system failure at the current moment does not reach the preset probability threshold for the brake pressure sensor fault alarm state parameter, then the brake pressure can be used as the non-sensing parameter.

[0085] S602. Apply all the sample data to perform model training on the Attention-LSTM model to obtain a sensor parameter value estimation model corresponding to the fault alarm state parameters.

[0086] In step S602, the Attention-LSTM model is an improved model that adds a layer of attention mechanism to the existing LSTM (Long Short-Term Memory) model (which is often used for time series data prediction), wherein the added attention mechanism is used to allow the LSTM model to dynamically focus on part of the information during training so as to capture more important information and improve the performance of the model. Therefore, sensor value estimation can be performed based on the Attention-LSTM model. In addition, the specific training process of the Attention-LSTM model is a prior art means; specifically, the parameters of the Attention-LSTM model can be obtained by searching using, but not limited to, a grid search algorithm or a Bayesian search algorithm.

[0087] In addition, when triggering the execution of the aircraft brake system fault warning action, the following steps are specifically included: importing the second vector of all the non-sensing parameters into the sensor parameter value estimation model, and outputting the parameter estimation value of the sensor parameter corresponding to the fault alarm state parameter at the current moment.

[0088] Therefore, based on the aforementioned possible design three, the sensor value estimation can also be performed based on the latest time series data of the non-sensing parameters when a sensor fault warning is issued, which is further beneficial for fault response and further ensures flight safety.

[0089] like Figure 2 As shown, the second aspect of this embodiment provides a computer device for executing the aircraft brake system fault prediction method as described in the first aspect or any possible design in the first aspect, including a memory, a processor and a transceiver that are sequentially connected in communication, wherein the memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program to execute the aircraft brake system fault prediction method as described in the first aspect or any possible design in the first aspect. For example, the memory may include but is not limited to random access memory (Random-Access Memory, RAM), read-only memory (Read-Only Memory, ROM), flash memory (FlashMemory), first-in-first-out memory (First Input First Output, FIFO) and / or first-in-last-out memory (First Input Last Output, FILO), etc.; the processor may be but is not limited to a microprocessor of the STM32F105 series. In addition, the computer device may also include but is not limited to a power module, a display screen and other necessary components.

[0090] The working process, working details and technical effects of the aforementioned computer device provided in the second aspect of this embodiment can be referred to the aircraft brake system fault prediction method described in the first aspect or any possible design of the first aspect, and will not be described in detail here.

[0091] In a third aspect of the present embodiment, there is provided a computer-readable storage medium storing instructions for the aircraft brake system fault prediction method as described in the first aspect or any possible design of the first aspect, that is, the computer-readable storage medium stores instructions, and when the instructions are executed on a computer, the aircraft brake system fault prediction method as described in the first aspect or any possible design of the first aspect is executed. The computer-readable storage medium refers to a carrier for storing data, which may include but is not limited to computer-readable storage media such as floppy disks, optical disks, hard disks, flash memories, USB flash drives and / or memory sticks, and the computer may be a general-purpose computer, a special-purpose computer, a computer network or other programmable devices.

[0092] The working process, working details and technical effects of the aforementioned computer-readable storage medium provided in the third aspect of this embodiment can be referred to the aircraft brake system fault prediction method described in the first aspect or any possible design of the first aspect, and will not be described in detail here.

[0093] A fourth aspect of the present embodiment provides a computer program product, including a computer program or instructions, wherein when the computer program or the instructions are executed by a computer, the method for predicting a fault of an aircraft brake system as described in the first aspect or any possible design of the first aspect is implemented. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.

[0094] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the protection scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for predicting aircraft brake system faults based on big data analysis, characterized in that: include: Acquire historical time series data of all operating parameters of the aircraft brake system, wherein all the operating parameters include fault alarm state parameters and multiple non-fault alarm state parameters, and the historical time series data includes The unit time and the corresponding operating parameters are described in The historical parameter value of each unit moment in the unit moments; For each pair of fault alarm state parameters and non-fault alarm state parameters in all the operating parameters, a corresponding correlation coefficient is calculated according to the historical time series data of the corresponding two parameters, specifically including: for a pair of fault alarm state parameters and non-fault alarm state parameters in all the operating parameters, a section of time series data is randomly selected from the historical time series data of the corresponding fault alarm state parameters to form a first sample , and also extracting the first sample from the historical time series data corresponding to the non-fault alarm state parameter Another time series data of the same period constitutes the second sample ; For the first sample Perform a normal distribution KS test to calculate the first test statistic value , and also for the second sample Perform the normal distribution KS check to calculate the second check statistic value If the first check statistic value and the second check statistic value are greater than the first preset threshold, the first sample is calculated according to the following formula With the sample The correlation coefficient : In the formula, represents a positive integer, In the first sample The Sample values, In the second sample The Sample values, Indicates the first sample The sample mean of The second sample The sample mean of as a correlation coefficient corresponding to the pair of fault alarm state parameters and non-fault alarm state parameters; For each pair of fault alarm state parameters and non-fault alarm state parameters, first determine each fault alarm triggering moment according to the historical time series data of the corresponding fault alarm state parameters, then extract each previous time series data corresponding to each fault alarm triggering moment from the historical time series data of the corresponding non-fault alarm state parameters, and finally add each previous time series data as a first vector to the corresponding vector library, wherein the previous time series data refers to the time series data traced back from a fixed time before the corresponding fault alarm triggering moment to The historical time series data corresponding to the time window of the length of time, the fixed time refers to the time interval from the corresponding fault alarm triggering time. At a certain moment in time, Indicates the preset first duration. Indicates the preset second duration; For each non-fault alarm state parameter in the plurality of non-fault alarm state parameters, first extract from the corresponding historical time series data the data traced back from the current moment to the The historical time series data corresponding to the time window of the time length is extracted, and the extraction result is used as the second vector, and then the ANN vector search is performed on the corresponding vector library based on the second vector to obtain a vector with the maximum similarity with the second vector, and finally the maximum similarity is used as the corresponding aircraft brake system failure prediction probability; According to the correlation coefficients corresponding to the various non-fault alarm state parameters and the predicted probability of the aircraft brake system failure, the total predicted probability of the aircraft brake system failure is obtained by superimposing and calculating, specifically including: according to the correlation coefficients corresponding to the various non-fault alarm state parameters and the predicted probability of the aircraft brake system failure, the total predicted probability of the aircraft brake system failure is calculated according to the following formula : In the formula, represents the total number of parameters of the plurality of non-fault alarm state parameters, Indicates less than or equal to A positive integer, Indicates the first of the plurality of non-fault alarm state parameters The influence weight coefficient corresponding to the non-fault alarm state parameter is: Indicates that The correlation coefficient corresponding to the non-fault alarm state parameters, Indicates less than or equal to A positive integer, Indicates the first of the plurality of non-fault alarm state parameters The correlation coefficient corresponding to the non-fault alarm state parameters, Indicates that The predicted probability of aircraft brake system failure corresponding to the non-fault alarm state parameters; If the predicted total probability of the aircraft brake system failure reaches a preset probability threshold, an aircraft brake system failure warning action is triggered.

2. The aircraft brake system fault prediction method according to claim 1, characterized in that: The fault alarm state parameter is a wheel speed sensor fault alarm state parameter, a brake pressure sensor fault alarm state parameter or a brake temperature sensor fault alarm state parameter.

3. The aircraft brake system fault prediction method according to claim 1, characterized in that: The plurality of non-fault warning status parameters include aircraft speed, wheel speed, brake pressure, brake temperature and / or accumulated number of brakes.

4. The aircraft brake system fault prediction method according to claim 1, characterized in that: After calculating the corresponding correlation coefficient for each pair of fault alarm state parameters and non-fault alarm state parameters according to the historical time series data of the corresponding two parameters, the method further includes: At the beginning of each operation and maintenance task of the aircraft brake system, for each pair of fault alarm state parameters and non-fault alarm state parameters, the corresponding fault alarm state parameters and the parameters traced back from the current time to the The historical time series data corresponding to the time window of the length of time is used as the corresponding first verification sample, and the corresponding non-fault alarm state parameters and traced back from the current time to The historical time series data corresponding to the time window of the same length is used as the corresponding second verification sample, and then the corresponding correlation coefficient is updated and calculated according to the first verification sample and the second verification sample, wherein, Indicates the third preset duration.

5. The aircraft brake system fault prediction method according to claim 1, characterized in that: After adding each of the previous time series data as a first vector to the corresponding vector library for each pair of the fault alarm state parameter and the non-fault alarm state parameter, the method further includes: At the beginning of each operation and maintenance task of the aircraft brake system, for each pair of fault alarm state parameters and non-fault alarm state parameters, expired vectors are removed from the corresponding vector library, wherein the expired vector refers to the time from the corresponding fault alarm triggering time to the current time exceeding Any vector of duration, Indicates the fourth preset duration.

6. The aircraft brake system fault prediction method according to claim 1, characterized in that: When the fault alarm state parameter is a wheel speed sensor fault alarm state parameter, a brake pressure sensor fault alarm state parameter or a brake temperature sensor fault alarm state parameter, before triggering the execution of the aircraft brake system fault warning action, the method further includes: For each fault alarm triggering moment, extract the historical parameter value at the corresponding fixed moment from the historical time series data of the sensor parameter corresponding to the fault alarm state parameter as a model output item, and take all non-sensing parameters and the corresponding first vector together as model input items, and then take the model input items and the model output items as corresponding sample data, wherein the non-sensing parameter refers to a parameter that is not a sensor parameter corresponding to the fault alarm state parameter among the multiple non-fault alarm state parameters; Applying all the sample data to perform model training on the Attention-LSTM model to obtain a sensor parameter value estimation model corresponding to the fault alarm state parameter; Triggering the execution of aircraft brake system fault warning actions, including: The second vector of all the non-sensing parameters is introduced into the sensing parameter value estimation model, and the parameter estimation value of the sensing parameter corresponding to the fault alarm state parameter at the current moment is obtained as output.

7. The method for predicting aircraft brake system failure according to claim 6, characterized in that: The non-sensing parameter is also a parameter that has not currently triggered a fault warning event, wherein the currently not triggered fault warning event means that when the corresponding parameter corresponds to another fault alarm state parameter, the predicted total probability of the aircraft brake system failure corresponding to the other fault alarm state parameter and at the current moment does not reach the preset probability threshold.

8. The method for predicting aircraft brake system failure according to claim 6, characterized in that: The parameters of the Attention-LSTM model are searched using a grid search algorithm or a Bayesian search algorithm.

Images