A method and system for monitoring starter valve failures of an aircraft engine

Abstract

The application provides a method for monitoring start valve fault of an aircraft engine, which comprises reading QAR data of a flight section in a flight recorder, and decoding the QAR data to obtain corresponding engineering value data; determining a plurality of parameters related to the operating state of the start valve according to the parameter names contained in the engineering value data, and obtaining the parameter values corresponding to each parameter in the plurality of parameters at each time in the flight section; calculating the opening time and / or closing time of the start valve at each time based on the industry regulations according to the obtained parameter values at each time, and predicting the opening time and / or closing time of the start valve at future times through an exponential smoothing method; and determining that the start valve is faulty when it is judged that the predicted opening time and / or closing time of the start valve exceeds a set threshold.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of aviation, and particularly relates to a method and system for monitoring a starting valve fault of an aircraft engine. BACKGROUND

[0002] The safe and efficient operation of a domestic civil aircraft is the basis for market success, and the operation efficiency is the key to market success. However, the failure of engine starting caused by starting valve failure occurs from time to time in the current airline operation. If the aircraft engine cannot be started normally, it is easy to cause the aircraft to delay take-off and re-loop queue to wait for the taxi-out instruction, thereby causing flight delays, which has a great impact on airlines that pay attention to flight punctuality. The usual solution to this problem for airlines in the past is to use manual starting when the starting valve fails, and if it is still unsuccessful, temporary spare parts need to be replaced, and if there are no spare parts, the aircraft needs to be replaced. Such a processing method increases the workload of manual operation and cannot guarantee the punctual operation of the aircraft.

[0003] Research has found that the main causes of starting valve failure are internal rust, pollution, and internal mechanism performance degradation of the starting valve, and the main failure forms are that the starting valve cannot be opened, cannot be opened to the designated position, or cannot be closed. The existing flight data analysis software can analyze the working state and performance degradation of the starting valve by reading and quickly obtaining the massive QAR data stored in the quick access recorder (QAR), analyzing the historical flight data in the QAR data, and predicting the performance state of the starting valve at a future time, thereby monitoring the starting valve failure.

[0004] However, since the starting valve failure monitoring method used by the existing flight data analysis software does not perform noise reduction processing on the original signal, does not perform stationarity test on the flight data time series in the QAR data, and does not perform pattern recognition and related feature extraction on the starting valve related data, the monitoring result often cannot match the actual starting valve failure alarm and failure list, thereby causing the starting valve failure to be unable to be accurately monitored. SUMMARY

[0005] Therefore, the technical problem to be solved by the present application is to overcome the defects of the existing method for monitoring the starting valve failure, which has low accuracy and is difficult to accurately and efficiently monitor the starting valve failure of a domestic civil aircraft using massive QAR data, and a new method and system for monitoring the starting valve failure of an aircraft engine are proposed.

[0006] The present application solves the above technical problems by the following technical solutions.

[0007] In particular, the present application provides a method for monitoring a starting valve fault of an aircraft engine, characterized in that the method comprises the following steps:

[0008] reading QAR data of a flight section in a flight recorder, and decoding the QAR data to obtain corresponding engineering value data;

[0009] determining a plurality of parameters related to the operating state of the starting valve according to parameter names contained in the engineering value data, and obtaining parameter values of each of the plurality of parameters at each time point in the flight section;

[0010] calculating the opening time and / or closing time of the starting valve at each time point based on industry regulations according to the obtained parameter values at each time point, and predicting the opening time and / or closing time of the starting valve at a future time point through an exponential smoothing method; and

[0011] when it is judged that the predicted opening time and / or closing time of the starting valve exceeds a set threshold value, determining that the starting valve has a fault.

[0012] The method analyzes the causes and main forms of the starting valve fault, extracts and processes parameter data related to the starting valve, determines the working state and performance of the starting valve through the opening time and / or closing time of the starting valve, and based on known historical data, fits the opening time and / or closing time and predicts the opening time and / or closing time at a future time point, so as to generate a trend chart through the fitted and predicted data, and then determine the time when the fault is likely to occur according to the trend in the trend chart, and timely provide maintenance or replacement suggestions for the user.

[0013] According to an embodiment of the present application, the QAR data is a binary data packet based on the ARINC717 standard, and the step of decoding the QAR data comprises:

[0014] reading the QAR data according to subframes one by one and obtaining parameter information associated with each subframe, and then for each subframe, matching the parameter information with parameter configuration information in a parameter decoding table, and when the matching is successful, converting the binary value of each parameter in the parameter information into an engineering value according to the parameter decoding table.

[0015] According to an embodiment of the present application, the step of converting the binary value of each parameter in the parameter information into an engineering value comprises:

[0016] The numerical conversion of each parameter in the parameter information is performed according to the attribute of each parameter in the QAR data included in the parameter decoding table, and the attribute of the parameter includes parameter name, record slot number, subframe, record start bit, record end bit, parameter type, parameter sign information and record precision.

[0017] According to an embodiment of the present application, the parameter configuration information in the parameter decoding table is read by using Python language programming, and the QAR data is read and decoded into engineering value data by using C language programming.

[0018] According to an embodiment of the present application, the method further comprises judging whether the read QAR data is valid according to the data size of the read QAR data for each subframe; if the data size of the read QAR data is greater than a predetermined value, determining that the subframe and the corresponding QAR data are valid, and obtaining the start position of the valid QAR data.

[0019] According to an embodiment of the present application, the plurality of parameters related to the running state of the starting valve include automatic take-off running program parameters, starting valve non-closing parameters, engine high and low pressure rotor speed parameters, fuel flow parameters, rotation opening degree parameters and bleed air pressure parameters.

[0020] According to an embodiment of the present application, the method further comprises determining that the QAR data corresponding to the current subframe is invalid and eliminating the corresponding engineering value data when the engineering value data of the automatic take-off running program parameters and the engineering value data of the starting valve non-closing parameters do not correspond for each subframe.

[0021] According to an embodiment of the present application, the method further comprises converting the longitude and latitude data of the take-off and landing airport in the engineering value data into distance data by using the Euler distance algorithm, and matching the distance data with standard airport distance data, and obtaining the actual take-off and landing airport information and flight information when the matching is successful.

[0022] According to an embodiment of the present application, the method further comprises performing smoothness processing on the calculated opening time and / or closing time of the starting valve in the en-route segment by using the method of D-order difference to obtain a smooth time series, and then performing fitting and prediction of the opening time and / or closing time of the starting valve by using the exponential smoothing method.

[0023] According to an embodiment of the present application, the method further comprises adjusting the value of the index in the exponential smoothing method for multiple times, and determining the value of the index in the exponential smoothing method when the fitting curve of the opening time and / or closing time in the en-route segment is basically consistent with the distribution trend of the calculated opening time and / or closing time.

[0024] According to an embodiment of the present application, the set threshold value comprises a preliminary threshold value and an alarm threshold value, and the method further comprises determining to overhaul the starting valve when the opening time and / or closing time of the starting valve exceeds the preliminary threshold value, and determining to replace the starting valve when the opening time and / or closing time of the starting valve exceeds the alarm threshold value.

[0025] The present application also provides a computer readable medium having computer readable instructions stored thereon, which, when executed on a device, cause the device to perform the method described above.

[0026] The present application also provides a system for monitoring starting valve failure of an aircraft engine, characterized in that the system comprises a processor configured to obtain QAR data from a flight recorder and perform the method described above.

[0027] On the basis of common knowledge in the art, the above-mentioned preferred embodiments can be combined in any manner, i.e. to obtain each preferred embodiment of the present application.

[0028] The positive progress effect of the above-mentioned embodiments of the present application is that:

[0029] 1. The method can extract parameters related to the starting valve from a large amount of QAR data, avoid redundant calculation, and thus realize efficient and accurate monitoring of starting valve failure;

[0030] 2. The method eliminates inconsistent invalid QAR data by analyzing the consistency of the parameter values related to the starting valve, thereby ensuring the accuracy of starting valve failure monitoring;

[0031] 3. The latitude and longitude information is converted into distance information by the Euler distance method, so as to accurately and quickly locate the corresponding flight to verify the monitoring result, and perform corresponding starting valve maintenance or replacement to better serve the customer. At the same time, invalid QAR data can be further determined and eliminated through the converted distance information, so as to ensure the accuracy of the starting valve failure monitoring result;

[0032] 4. The method performs smoothness processing on the historical opening time and / or closing time by D-order difference, so as to ensure the accuracy of the prediction result;

[0033] 5. The method and system for monitoring starting valve failure of an aircraft engine according to the present application can accurately and efficiently monitor starting valve failure, provide maintenance or replacement suggestions for customers, and efficiently utilize relevant computing resources. BRIEF DESCRIPTION OF DRAWINGS

[0034] Figure 1 A flowchart of the method for monitoring starting valve failure of an aircraft engine according to the preferred embodiments of the present application.

[0035] Figure 2 A diagram showing the decoded engineering value data of the QAR data using the method according to the preferred embodiment of the present application.

[0036] Figure 3 A diagram showing the distribution of the start valve opening time at various times in the flight segment.

[0037] Figure 4 A diagram showing the distribution of the start valve closing time at various times in the flight segment.

[0038] Figure 5 A diagram showing the start valve opening time curve generated using the method in Figure 1 .

[0039] Figure 6 A diagram showing the start valve closing time curve generated using the method in Figure 1 .

[0040] Figure 7 A diagram showing the determination of the start valve malfunction in the start valve closing time curve generated using the method in Figure 1 .

[0041] Figure 8 A diagram showing the longitude data and the latitude data of the airport in the decoded QAR data.

[0042] Figure 9 A diagram showing the relevant information of the standard airport, including the longitude data and the latitude data of the standard airport.

[0043] Figure 10 A diagram showing the specific information of the flight determined by the Euclidean distance algorithm, in which the start valve malfunction occurs in the flight.

[0044] Figure 11 A diagram showing the opening time and the closing time of the start valve of a plurality of flights determined using the method for monitoring the start valve malfunction of the aircraft engine according to the preferred embodiment of the present application. DETAILED DESCRIPTION

[0045] In order to make the purpose, technical solutions and advantages of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings showing the embodiments according to the present application. It should be understood that, based on the embodiments described in the present application, all other embodiments obtained by those of ordinary skill in the art without any creative effort shall fall within the scope of protection of the present application.

[0046] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs; the terms "comprising," "comprises" and the like are to be construed as open-ended terms (i.e., meaning "including, but not limited to,") unless otherwise noted; singular terms shall be read as including plural terms (e.g., "a" will be understood to mean "one or more"); all numbers are approximate values; numerical parameters are only approximations of actual parameters unless otherwise indicated; and all databases known to the skilled artisan are incorporated by reference.

[0047] The QAR data is time series data recording the flight state of the aircraft, and has the characteristics of high bit, complexity and large data volume. The existing starting valve fault monitoring method cannot effectively process the massive QAR data, resulting in inaccurate starting valve fault monitoring results, so that the user cannot make correct maintenance or replacement decisions for the starting valve according to the fault monitoring results, thereby failing to ensure the normal starting of the engine, which causes the domestic civil aircraft to be unable to realize safe and efficient operation.

[0048] Based on at least the above problems, the present application proposes a new starting valve fault monitoring analysis method for domestic regional passenger aircraft engines. The method performs starting valve opening time and / or closing time fitting by analyzing known historical data in the QAR data, and obtains the predicted value of the starting valve opening time and / or closing time at future time by weighted calculation of the historical data through the exponential smoothing method, thereby establishing a starting valve fault monitoring analysis model for the engine.

[0049] Based on the starting valve fault monitoring analysis model, the user can timely monitor the trend and abnormal value of the starting valve opening time and / or closing time, assess the degree of performance decline of the starting valve in advance, and timely carry out maintenance work such as cleaning, lubrication, overhaul, replacement, etc., thereby improving the service life of the valve, avoiding the occurrence of associated damage to the engine and surrounding equipment, and effectively reducing spare parts inventory and on-site manual starting work, thereby not only improving the efficiency of route operation, but also achieving the effect of cost reduction and efficiency increase.

[0050] Specifically, as shown in Figure 1 The method for monitoring the starting valve fault of an aircraft engine according to some preferred embodiments of the present application can include the following main steps.

[0051] Step one, decoding the QAR data to obtain the corresponding engineering value data, which specifically includes reading the QAR data of a certain flight segment in the flight recorder, and decoding the read QAR data to obtain the corresponding engineering value data;

[0052] Step two, obtaining parameter values corresponding to each time in the flight segment, which specifically includes determining a plurality of parameters related to the running state of the starting valve according to the parameter names contained in the engineering value data, and then obtaining the parameter values corresponding to each time in the flight segment for each of the plurality of parameters;

[0053] Step three, predicting the opening time and / or closing time of the starting valve at the future time, which specifically includes calculating the opening time and / or closing time of the starting valve at each time based on industry regulations according to the obtained parameter values at each time, and predicting the opening time and / or closing time of the starting valve at the future time through the exponential smoothing method;

[0054] Step four, determining that the starting valve is malfunctioning when it is judged that the predicted value exceeds the set threshold, which specifically includes comparing the predicted opening time and / or closing time of the starting valve with the set threshold, and determining that the starting valve is malfunctioning when it is judged that the predicted value exceeds the set threshold.

[0055] Regarding step one, the QAR data can be a binary data packet based on the ARINC717 standard. At this time, the step of decoding the QAR data specifically includes reading the QAR data according to subframes one by one, and obtaining parameter information associated with each subframe. Subsequently, for each subframe, the parameter information is matched with the parameter configuration information in the parameter decoding table, and when the matching is successful, the binary value of each parameter in the parameter information is converted into the engineering value according to the parameter decoding table.

[0056] Further, the step of converting the binary value of each parameter in the parameter information into the engineering value can also include converting the numerical value of each parameter in the parameter information according to the attributes of each parameter in the QAR data included in the parameter decoding table, wherein the attributes of the parameter include the parameter name, the record word slot number, the subframe, the record start bit, the record end bit, the parameter type, the parameter sign information and the record precision.

[0057] Exemplarily, program codes can be written in C language and Python language to realize the decoding work of the original QAR data. Preferably, the program code written in Python language reads the decoding parameter table, and the program code written in C language reads the QAR data and decodes the QAR data into engineering value data.

[0058] At this time, the decoding process of the QAR data specifically includes obtaining a parameter decoding table from a main manufacturer, reading the parameter decoding table by using a program code written in Python language to obtain parameter configuration information and attributes of the parameters; calling a library file compiled by C language by using the program code written in Python language, and passing the parameter decoding table data including the parameter configuration information and the attribute information of the parameters to the program code written in C language in a parameter passing manner; reading the original QAR data by using the program code written in C language, and determining a starting position of valid data according to a starting word of a subframe in the ARINC717 specification; reading the QAR data subframe by subframe and obtaining parameter information of a plurality of parameters associated with each subframe, and sequentially converting binary data of each parameter into engineering value data according to the parameter decoding table, thereby realizing decoding and decoding calculation of the QAR data. The valid data refers to data with a data amount greater than a predetermined value, for example, 1 KB.

[0059] Preferably, the method for judging valid data can include, for each subframe, comparing a data amount of the read QAR data with a predetermined value to judge whether the data is valid. If the data amount of the read QAR data is greater than the predetermined value, it is determined that the subframe and the QAR data corresponding thereto are valid, and a starting position of the valid QAR data is obtained. Alternatively, if a file size of the read QAR data is greater than the predetermined value, it can also be determined that the subframe and the QAR data corresponding thereto are valid.

[0060] Alternatively, the step one further includes converting the decoded engineering value data into engineering value-relational structured data to facilitate subsequent analysis, and finally obtaining engineering value data corresponding to the decoded QAR data in the form of a table as shown in FIG. 2. Figure 2 Preferably, the engineering value data formed after decoding is displayed in the form of a table, each row representing engineering value data of all parameters associated with a subframe, and each column representing engineering value data of a parameter corresponding to all subframes.

[0061] It has been proved by practice that the efficiency and accuracy of the decoding method in the step one for decoding the QAR data can reach the performance of commercial software such as Airfase / AGS, so that the QAR data decoding work time is shortened from 2 to 3 days of manual operation per flight segment to minute-level automatic operation. Moreover, the decoding method can alternatively have data re-examination and verification functions, such as checking data consistency, filling missing data, correcting invalid data, removing abnormal data and repeated data, and the like.

[0062] As to step two, it is found that the parameters related to the running state of the starting valve include the automatic starting program parameter (YES & NO), the starting valve not closed parameter (YES & NO), the left and right engine high and low pressure rotor speed parameters, the fuel flow parameter, the rotation opening degree (FMV) parameter and the bleed air pressure (APU) parameter, and then the Chinese name of each parameter can be converted into the corresponding English name based on the parameter decoding table. Meanwhile, according to the engineering value data table generated after decoding, by reading the parameter name engineering value data in each column of the table header, and comparing the read engineering value data with the above-mentioned English name, the positions of the multiple parameters related to the running state of the starting valve in the engineering value data table can be determined. Subsequently, according to the determined position data of the multiple parameters and the engineering value data table generated after decoding, the parameter values corresponding to each of the multiple parameters related to the running state of the starting valve at each time in the flight segment can be obtained, forming the engineering value time sequence data corresponding to each parameter.

[0063] Furthermore, for the valid QAR data, the engineering value data of the automatic take-off running program parameter and the engineering value data of the starting valve not closed parameter should be corresponding. Therefore, when the engineering value data of the automatic take-off running program parameter and the engineering value data of the starting valve not closed parameter do not correspond, it can be determined that the QAR data corresponding to the current subframe is invalid. Exemplarily, when the engineering value data of the automatic starting program parameter is NO (or YES), and the engineering value of the starting valve not closed parameter is YES (or NO), it can be judged that the two parameter data do not correspond and the event is false, and then it can be determined that the corresponding QAR data is invalid. Further, for the engineering value data of the QAR data determined to be invalid, it can be removed by the way of supplementing 0, thereby ensuring to obtain accurate parameter values at each time.

[0064] As to step three, by analyzing the main failure forms of the starting valve, it is found that the opening time and / or closing time of the starting valve can more timely and accurately represent the working state and performance decline of the starting valve, so as to accurately monitor the failure of the starting valve. According to the ARINC767-2006 specification, MSG-3, FCOM3, aircraft health management manual and airline engineering documents and industry regulations, it can be determined that the opening time and / or closing time of the starting valve can be represented as a function of the above multiple parameters related to the running state of the starting valve, and then based on the confirmed valid parameter engineering value data at each time, the opening time and / or closing time of the starting valve at each time can be obtained by mathematical calculation through the business program code written. Preferably, the calculated opening time and / or closing time of the starting valve at each time is plotted into a graph, so as to obtain the opening time and / or closing time of the starting valve at each time in the flight segment as shown in FIG. 3. Figure 3 and Figure 4The start valve opening time distribution scatter diagram and the start valve closing time distribution scatter diagram shown.

[0065] Then, according to the calculated start valve opening time and / or closing time time series data at each time, the start valve opening time and / or closing time at future time is predicted by using the exponential smoothing method, and finally the start valve opening time trend chart (i.e. trend chart) shown in Figure 5 and the start valve closing time trend chart (i.e. trend chart) shown in Figure 6 According to the fitting results of the start valve opening time and / or closing time at known time by the exponential smoothing method, the index value in the exponential smoothing method can be determined.

[0066] Preferably, the single exponential smoothing method is used to predict the start valve opening time and / or closing time at future time. The single exponential smoothing method has only one smoothing parameter, which is suitable for fitting and predicting time series data with stationary characteristics. Moreover, unlike the moving average method which gives the same weight to each data, the exponential smoothing can give different weights to historical time data according to the actual situation. According to the general formula of the single exponential smoothing method, the smaller the weight of the earlier data, the smaller the role it plays in the prediction at future time. Therefore, the index value in the single exponential smoothing method can be adjusted multiple times by gradually reducing the step size to determine the reasonable weight of the data and to optimize the index data iteratively.

[0067] According to the communication with the technical experts of the airline, when the opening time and / or closing time curve fitted in the flight segment is basically consistent with the distribution trend of the calculated opening time and / or closing time, the index value in the single exponential smoothing method can be determined. Finally, the index data can be defined as the optimal value in a reasonable interval by the above method, and the engine valve fault monitoring chart shown in Figure 7 is output. As shown in the box in Figure 7 The fault reflected by the chart is highly consistent with the actual fault on the flight line, which shows that the method can accurately monitor the start valve fault information.

[0068] More preferably, step three further comprises using the D-order difference method to process the calculated start valve opening time and / or closing time in the flight segment to obtain a stationary time series by using the program code written in python language, and then using the single exponential smoothing method to fit and predict the start valve opening time and / or closing time.

[0069] With respect to step four, through the predicted start valve opening time and / or closing time, or through the graph or trend analysis generated by fitting and predicting the start valve opening time and / or closing time, the staff can estimate when the start valve fails, and thus make corresponding suggestions. Specifically, a settable threshold value data can be obtained from the airline, the predicted data is compared with the settable threshold value data, and when the predicted data exceeds the threshold value, the airline is informed that the start valve can fail at a certain time and needs to be repaired or replaced.

[0070] Preferably, the settable threshold value includes a preliminary threshold value and an alarm threshold value, when the opening time and / or closing time of the start valve exceeds the preliminary threshold value, it is determined to overhaul the start valve; when the opening time and / or closing time of the start valve exceeds the alarm threshold value, it is determined to replace the start valve.

[0071] Alternatively, the method for monitoring start valve failure according to the preferred embodiment of the present application includes converting the longitude and latitude data of the take-off and landing airport in the engineering value data into distance data by using the Euler distance algorithm, and matching the distance data with the standard airport distance data, and obtaining the actual take-off and landing airport information and flight information when the matching is successful.

[0072] Specifically, when it is found that the take-off and landing airport corresponding to each scatter point in the start valve failure monitoring graph is inaccurate, the longitude and latitude of the two airports of the take-off and landing airport in the QAR data decoded engineering value data can be converted into distance, and by matching with the converted standard airport distance, the actual take-off and landing airport information (i.e. the flight information of the aircraft) in the QAR data can be determined according to the nearest distance, so as to facilitate the verification of whether the start valve of the corresponding aircraft has a problem as soon as possible. At the same time, QAR data with incorrect airport can be screened out and removed to ensure the accuracy of the failure monitoring result.

[0073] For example, by comparing the airport information in the decoded QAR data as shown in Figure 8 with the standard airport information as shown in Figure 9 , it can be found that Figure 8 and Figure 9 have inconsistent longitude data and latitude data. At this time, the Euler distance algorithm can be used to compare the distance between the longitude data and latitude data in the QAR data and the longitude data and latitude data of each airport in the standard airport, and finally determine that the airport in the QAR data is Figure 9 Mudanjiang Hailang Airport as shown in

[0074] Preferably, according to the fitting and prediction data of the opening time and / or closing time of the start valve and the determined flight data of the start valve failure, a graph as shown in Figure 10 andFigure 11 The website shown is used to facilitate the airline to implement screening, view the opening time and / or closing time data of the start valve, and view the relevant information of the flight.

[0075] More preferably, the method for monitoring start valve failure according to the preferred embodiment of the present application is deployed on a domestic civil aircraft operation monitoring data analysis platform. The underlying data lake of the platform uses Lambda architecture and applies open source components to achieve batch data and streaming data processing through aviation data lake technology. Specifically, the aviation data lake technology includes a data acquisition layer, a message layer, a data ingestion layer, a batch processing layer, a fast processing layer, a service layer, and a data storage layer. The data acquisition layer includes connecting data sources, event interception and routing, and interfaces, which can be implemented through technologies such as Sqoop and Flume. The message layer uses Kafka components and can process TB to PB level data. Based on the batch and real-time characteristics of civil aircraft operation support data, the data ingestion layer can be implemented using Spark and Flink technologies. From the cost perspective, the batch processing layer can be implemented using HDFS, Pig, and Hive technologies. The fast processing layer can be implemented using Flume, Spark, and Flink technologies according to the scene. The service layer can quickly build a service side and integrate Sqoop, Pig, and Hive table export services through Spring Boot. The data storage layer can be implemented using common HDFS, HBASE, and Elasticsearch technologies.

[0076] In summary, the specific process of the method for monitoring start valve failure of an aircraft engine according to the preferred embodiment of the present application can be divided into an original QAR data decoding stage and an engineering value data processing stage after decoding. In the engineering value data processing stage, the corresponding field data is processed and analyzed by analyzing the causes of the failure and the main manifestations of the failure, and the trend of the generated start valve opening time and / or closing time trend chart is used to determine the time when the next failure may occur, and the airline is informed so that it can determine the approximate time for the start valve maintenance and replacement work of the corresponding aircraft.

[0077] Preferably, the exponential smoothing method and the Euler distance algorithm are both implemented based on Python language, and the QAR data decoded and stored in the aviation data lake are operated in the form of encapsulated functions in the program code, and finally output to the Postgresql database for business analysis. Compared with the traditional engine valve performance monitoring method, the start valve monitoring method through these algorithms can achieve higher precision fault monitoring of the start valve of the engine and more accurate short-term prediction of the performance of the engine.

[0078] According to some embodiments of the present application, a computer readable medium having stored thereon computer readable instructions which, when executed on a device, cause the device to perform the monitoring method as above can also be provided.

[0079] According to some embodiments of the present application, a system for monitoring start valve failure can also be provided, wherein the system comprises a processor configured to be capable of obtaining raw QAR data from a flight recorder and performing the monitoring method as above.

[0080] The method for monitoring start valve failure of an aircraft engine according to the preferred embodiments of the present application filters interference factors affecting time series of start valve failure monitoring from several aspects such as pattern recognition, data preprocessing, feature extraction, model tuning, etc., greatly improving the robustness of the model.

[0081] The model can be deployed on the "domestic civil aircraft operation monitoring data analysis platform" (FLYWIN.digital) to facilitate airlines to initiate preventive maintenance according to the monitoring results to improve the service life of the start valve, and greatly avoid the occurrence of associated damage to the engine and surrounding equipment, thereby reducing spare parts inventory and manual start work on site. Moreover, the most direct associated component of engine start valve failure is the engine, which has a value of millions of yuan, and by improving the service life of the start valve, damage to the engine can be avoided, thereby avoiding huge losses.

[0082] The start valve failure monitoring method and system not only can improve the efficiency of route operation, but also can achieve the effect of cost reduction and efficiency increase, and has obvious effect in maintenance engineering front-line work. The method and system for monitoring start valve failure of an aircraft engine are suitable for domestic ARJ21-700 aircraft, and are also suitable for C919 or CR929 mechanism failure. In addition, compared with A320, B737, CRJ and other aircraft models operated by domestic airlines, the corresponding parameters can be adjusted and monitored quickly after adjustment, so the method and system have broad application prospects.

[0083] Although the specific embodiments of the present application are described above, those skilled in the art should understand that these are only illustrative, and the protection scope of the present application is defined by the appended claims. Those skilled in the art can make various changes or modifications to these embodiments without departing from the principles and essence of the present application, and such changes and modifications fall within the protection scope of the present application.

Claims

1. A method for monitoring the starting valve malfunction of an aircraft engine, characterized in that, The method includes the following steps: Read the QAR data of a certain flight segment from the flight recorder and decode the QAR data to obtain the corresponding engineering value data; Based on the parameter names contained in the engineering value data, determine multiple parameters related to the operating status of the starting valve, and obtain the parameter value of each of the multiple parameters at each time in the flight segment; Based on the acquired parameter values ​​at various times, the opening and / or closing times of the starting valve at each time are calculated according to industry standards, and the opening and / or closing times of the starting valve at future times are predicted using an exponential smoothing method; and When the predicted opening time and / or closing time of the starting valve exceeds a set threshold, it is determined that the starting valve has malfunctioned. Specifically, the D-order difference method is used to perform stationarity processing on the calculated opening and / or closing times of the starting valve within the flight segment to obtain a stationary time series, and then the exponential smoothing method is used to fit and predict the opening and / or closing times of the starting valve.

2. The method according to claim 1, characterized in that, The QAR data is a binary data packet based on the ARINC717 standard. The steps for decoding the QAR data include: The QAR data is read frame by frame, and the parameter information associated with each subframe is obtained. Then, for each subframe, the parameter information is matched with the parameter configuration information in the parameter decoding table. When the match is successful, the binary value of each parameter in the parameter information is converted into the engineering value value in the parameter decoding table.

3. The method according to claim 2, characterized in that, The steps for converting the binary value of each parameter in the parameter information into an engineering value include: According to the attributes of each parameter in the QAR data included in the parameter decoding table, the numerical conversion is performed on each parameter in the parameter information. The attributes of the parameter include parameter name, number of record slots, subframe, record start bit, record end bit, parameter type, parameter sign information, and record precision.

4. The method according to claim 2, characterized in that, The parameter configuration information in the parameter decoding table is read using Python programming, and the QAR data is read and decoded into engineering value data using C programming.

5. The method according to claim 2, characterized in that, The method further includes: For each subframe, its validity is determined based on the amount of QAR data read. If the amount of QAR data read is greater than a predetermined value, the subframe and its corresponding QAR data are determined to be valid, and the starting position of the valid QAR data is obtained.

6. The method according to claim 1, characterized in that, The parameters related to the operating status of the starting valve include automatic takeoff operation procedure parameters, starting valve not closed parameters, engine high and low pressure rotor speed parameters, fuel flow parameters, rotation opening parameters, and bleed air pressure parameters.

7. The method according to claim 6, characterized in that, The method further includes: For each subframe, if the engineering value data of the automatic takeoff operation procedure parameters and the engineering value data of the start valve not closed parameters do not correspond, the QAR data corresponding to the current subframe is determined to be invalid, and the corresponding engineering value data is removed.

8. The method according to claim 1, characterized in that, The method further includes: The longitude and latitude data of the take-off and landing airports in the engineering value data are converted into distance data using the Eulerian distance algorithm. The distance data is then matched with standard airport distance data. When the match is successful, the actual take-off and landing airport information and flight information are obtained.

9. The method according to claim 1, characterized in that, The method further includes: The value of the exponent in the exponential smoothing method is adjusted multiple times, and the value of the exponent in the exponential smoothing method is determined when the fitted opening time and / or closing time curves within the flight segment are basically consistent with the distribution trend of the calculated opening time and / or closing time.

10. The method according to claim 1, characterized in that, The set thresholds include a preliminary threshold and an alarm threshold, and the method further includes: When the opening time and / or closing time of the starting valve exceed the preliminary threshold, it is determined that the starting valve needs maintenance. When the opening time and / or closing time of the starting valve exceeds the alarm threshold, it is determined that the starting valve should be replaced.

11. A computer-readable medium having stored thereon computer-readable instructions, wherein the computer-readable instructions, when executed on a device, cause the device to perform the method according to any one of claims 1 to 10.

12. A system for monitoring the failure of the starter valve of an aircraft engine, characterized in that, The system includes a processor configured to acquire QAR data for a flight segment from a flight recorder and perform the method according to any one of claims 1 to 10.

Images