CN111721480A

CN111721480A - Civil aircraft unit oxygen system leakage early warning method based on flight data

Info

Publication number: CN111721480A
Application number: CN202010606697.2A
Authority: CN
Inventors: 孙有朝; 彭冲; 苏思雨
Original assignee: Nanjing University of Aeronautics and Astronautics
Current assignee: Nanjing University of Aeronautics and Astronautics
Priority date: 2020-06-29
Filing date: 2020-06-29
Publication date: 2020-09-29
Anticipated expiration: 2040-06-29
Also published as: CN111721480B

Abstract

The embodiment of the invention discloses a civil aircraft oxygen system leakage early warning method based on flight data, which is applied to real-time monitoring and risk early warning of the civil aircraft oxygen system; the method comprises the steps of collecting flight data related to a unit oxygen system of a monitored flight, processing the flight data and extracting characteristics, introducing the extracted characteristic values into a flight unit oxygen system leakage model for calculation, and performing grading early warning processing on the calculation results. The invention can automatically collect the flight data related to the oxygen system of the aircraft unit, analyze the flight data, judge the performance of the oxygen system of the aircraft unit, and finally inform the maintenance personnel of the early warning result, thereby realizing the closed-loop control and the health management of the oxygen system of the aircraft unit, and the maintenance personnel can carry out the maintenance according to the situation and timely eliminate the faults, thereby reducing the maintenance cost of the aircraft.

Description

Civil aircraft unit oxygen system leakage early warning method based on flight data

Technical Field

The invention relates to the technical field of civil aircraft system safety management and risk early warning.

Background

When the civil aircraft loses pressure at high altitude, the oxygen system plays a very important role in ensuring the life safety of passengers on the aircraft, and has very important influence on the safety of the operation of the aircraft. The oxygen system of the unit is a fixed high-pressure oxygen system which is stored by a fixed oxygen cylinder and provides oxygen for breathing for the primary driver, the secondary driver and the observer. The crew oxygen system can provide emergency oxygen for crew members in the cockpit in case of decompression, smoke and fire, and can provide supplementary oxygen for improving night flight eyesight or eliminating fatigue.

At present, an airline company is required to check whether the pressure of an oxygen cylinder of a unit meets a release standard every day for daily inspection and maintenance of the oxygen system of the unit, and if the pressure is lower than the standard, the oxygen cylinder of the unit is replaced; if the pressure meets the requirement, the pressure of the oxygen cylinder of the unit is not recorded and continuously monitored and managed. Therefore, if the abnormal slow leakage of the oxygen system of the aircraft crew is difficult to find in time, the aircraft crew oxygen bottle is frequently replaced by the crew, the workload and the maintenance cost of the maintenance of the air route are increased, but the leakage fault is not really eliminated, and the potential safety hazard always exists.

In addition, the following difficulties often exist in monitoring whether the aircraft crew oxygen system is abnormal or not by analyzing flight data and establishing a crew oxygen system leakage model: the pressure of the oxygen cylinder of the aircraft is influenced by a plurality of factors such as air pressure, temperature and the like, and the temperature of the oxygen cylinder of the aircraft is determined by the temperature of the cockpit and the atmospheric temperature; a temperature sensor is not arranged at the position of the oxygen cylinder of the unit, so that the temperature cannot be directly measured; at the same time, these influencing factors also change from moment to moment during the flight of an aircraft. Therefore, currently, the monitoring threshold of the crew oxygen system is usually set directly according to engineering experience, and when the crew oxygen pressure change of a flight exceeds the set threshold, an alarm is given. The monitoring threshold value set according to engineering experience is often too large, so that the leakage phenomenon of the oxygen system of the unit is frequently missed and failure can not be timely eliminated, and the operation safety of the airplane cannot be guaranteed.

Disclosure of Invention

The purpose of the invention is as follows: the invention provides a civil aircraft unit oxygen system leakage early warning method based on flight data, which aims to solve the problems that in the prior art, the maintenance cost is high, faults cannot be cleared in time and the like.

The technical scheme is as follows: the invention provides a civil aircraft unit oxygen system leakage early warning method based on flight data, which specifically comprises the following steps:

step 1: collecting flight data of monitored flights of civil aircraft A;

step 2: preprocessing the acquired flight data, wherein the preprocessing comprises missing data supplement;

and step 3: extracting characteristic values of the preprocessed data, wherein the specifically extracted characteristic values comprise: crew oxygen pressure P when aircraft is at takeoff airport_dAtmospheric temperature T of take-off airport_dOxygen pressure P of unit in approach stage_aAnd the lowest atmospheric temperature T in the course of the route_min；

And 4, step 4: obtaining historical flight data of n flights without leakage faults, wherein the model of an airplane executing the n flights is the same as that of a civil airplane A; preprocessing the historical flight data, extracting characteristic values from the processed data, establishing a unit oxygen system leakage model according to the extracted characteristic values, and extracting P from the step 3_d、T_d、P_aAnd T_minLeading the oxygen pressure into the model to obtain the predicted value P of the oxygen pressure of the unit at the approach stage of the monitored flight_pre-aThereby obtaining P_pre-aAnd P_aResidual error between_a；

And 5: will be provided with_aAnd standard deviation S_dFor comparison, if_a≤2.5S_dIf the flight is normal, the oxygen system of the unit is normal, and the next flight is continuously monitored; if 2.5S_d＜_a＜3.5S_dTurning to step 6; if the residual error_a≥3.5S_dThen, an alarm is given to the maintenance personnel of the engineering; said S_dIntroducing a sample consisting of historical flight data of flights without leakage faults into an oxygen system leakage model to obtain a standard deviation of sample residual errors;

step 6: collecting p pieces of historical flight data of the civil aircraft A adjacent to the monitored flight, and preprocessing the data by using the preprocessing method in the step 2;

and 7: and 6, extracting characteristic values of the data preprocessed in the step 6, wherein the specifically extracted characteristic values comprise: departure time D of each flight_tThe crew oxygen pressure P at the takeoff airport of the aircraft on each flight_d1Temperature of the cockpit of each flight_cocAnd start in each flightAtmospheric temperature T of an airport_d1Based on T_cocAnd T_d1Oxygen pressure P of the aircraft units at the take-off airport_d1Conversion to T at standard temperature_stdOxygen pressure value P of the unit_std；

And 8: according to D_tAnd P_stdEstablishing a model of the variation of the oxygen pressure of the civil aircraft A unit along with the time by using least square normal regression fitting, thereby obtaining the unit oxygen pressure leakage rate β of the aircraft₅；

And step 9: selecting continuous N pieces of historical flight data of abnormal-free airplanes of an h-frame unit oxygen system, wherein the h-frame airplanes are the same as the civil airplane A in type; performing characteristic extraction on flight data of each flight of each airplane to obtain the unit oxygen pressure P when the airplane is positioned at a take-off airport in each flight of the s-th airplane_d-sAnd departure time of each flight, will P_d-sConverting into oxygen pressure value P of unit at standard temperature_stdsAccording to P of each flight of the s-th aircraft_stdsAnd taking off time, and establishing a time-dependent change equation of the oxygen pressure of the set of the s-th aircraft by using a least square method so as to obtain the oxygen pressure leakage rate β of the set of the s-th aircraft_5·sAnd(s) is 1,2 and … h, the crew oxygen pressure leakage rate of the h aircrafts is averaged, and thus the crew oxygen pressure leakage rate β of the crew oxygen system abnormal-free aircraft is obtained_std；

Step 10, according to β₅And β_stdDetermining a judgment parameter α, if α is less than or equal to 5 degrees, ensuring that the crew oxygen system of the airplane is normal without slow leakage risk, and if α is more than 5 degrees, ensuring that the crew oxygen slow leakage risk exists in the airplane and giving an alarm to crew maintenance personnel.

Further, in the step 1, the flight data of the monitored flight of civil aircraft a is collected in the quick access recorder.

Further, the flight data of the monitored flight obtained in the step 1 is complete flight cycle time sequence data, and the sequence of the flight data is 5 columns, namely time, flight phase, unit oxygen pressure, atmospheric temperature and cockpit temperature parameters;each row has m rows, and the rows correspond to the flight time of the flight; in the step 2, a near supplementation method is adopted for missing data supplementation, and the preprocessed time sequence data is X ═ X_ijWhere i is 1, 2.., m; j ═ 1, 2., 5, x_ijThe data of the ith row and the jth column in X.

Further, the step 3 specifically includes:

step 3.1: crew oxygen pressure P when positioning an aircraft at a takeoff airport_dAnd atmospheric temperature T of take-off airport_dThe data extraction time of (a) is set as the time t when the flight enters the engine non-starting stage_dLet the length of the sampled data be t, then

Wherein x is_i3Is the data of the ith row and the 3 rd column in the flight cycle time sequence data X, X_i4The data of the 4 th row in the flight cycle time sequence data X;

step 3.2: oxygen pressure P of unit in approaching stage_aThe data extraction time of (2) is set as the time t when the flight enters the approach stage_aIf the length of the sampled data is t, then

Step 3.3: determining a minimum atmospheric temperature T in an airway process_minIs the time t of the first occurrence of the minimum value of the atmospheric temperature_minIf the length of the sampled data is t, then

Further, the establishing of the unit oxygen system leakage model in the step 4 specifically includes:

step 4.1: combining the eigenvalues extracted in the step 4 into an eigenvalue matrix E₁：

Wherein P is_dn·nCrew oxygen pressure, T, at the takeoff airport of an aircraft on the nth flight in historical flight data representing no leakage faults_dn·nAtmospheric temperature, T, at takeoff airport of nth flight in historical flight data representing no leakage fault_{min n·n}Lowest atmospheric temperature, P, during the course of the nth flight in the historical flight data representing no leakage faults_an·nThe crew oxygen pressure of the nth flight in the historical flight data representing no leakage fault in the approach phase;

step 4.2: to E₁And analyzing characteristic value data: with P_anAs a dependent variable, with P_dn、T_dn、T_{min n}Is an independent variable, P_an＝[P_an-1，P_an-2，…P_an-n]，P_dn＝[P_dn-1，P_dn-2，…P_dn-n]，T_dn＝[T_dn-1，T_dn-2，…T_dn-n]，T_{min n}＝[T_{min n-1}，T_{min n-2}，…T_{min n-n}]And establishing a regression fitting equation by adopting a multiple linear regression method so as to determine the leakage model of the oxygen system of the flight unit: p_pre-a＝β₀+β₁·P_dn+β₂·T_dn+β₃·T_{min n}Wherein β₀To fit constant terms, β₁、β₂、β₃Are coefficients.

Further, each historical flight data in the step 6 is a complete flight cycle time sequence data, 5 columns are provided in total, and time, flight phase, unit oxygen pressure, atmospheric temperature and cockpit temperature parameters are sequentially provided; each column has m1 rows, and the preprocessed time sequence data is X_e＝{x_IJ1,2, m 1; 1,2, 5, X_eIs the e-th time sequence data, e is 1,2, … p, x_IJIs X_eRow I and column J;

p in said step 7_stdComprises the following steps:

where t1 is the sample length, t_d1For the moment x when the flight enters the engine non-starting stage in the historical data collected in the step 6_I4For the data of row I, column 4 in each time series data.

Further, step 8 specifically adopts a least square method to perform group D on p in step 7_tAnd P_stdPerforming linear regression fitting to establish a time-varying equation of the oxygen pressure of the civil aircraft A: p_std＝β₄+β₅·D_tWherein β₄Fitting equation constant terms.

Further, the method for determining the threshold α in step 9 is as follows:

α＝arctan(β₅)-arctan(β_std)。

has the advantages that: the invention designs a civil aircraft unit oxygen system leakage early warning method based on flight data, which comprises the steps of automatically acquiring relevant flight data of an aircraft unit oxygen system, analyzing the data, carrying out grading processing on the calculation result of a leakage model of the flight unit oxygen system, judging the performance of the aircraft unit oxygen system, finally informing a maintenance staff of the aircraft service with the early warning result, guiding the development of maintenance work and timely troubleshooting; by establishing a time-varying equation of the oxygen pressure of the aircraft unit of the monitored aircraft, the oxygen leakage rate of the aircraft unit is obtained, the remaining service time of the aircraft unit oxygen system is predicted, and the crew can carry out maintenance according to the situation, so that the maintenance cost of the aircraft is reduced. The method is automatically operated and executed by platform software, no crew member is required to participate in the operation, the unsupervised monitoring and closed-loop control of the oxygen system of the aircraft unit are realized, the crew member only needs to make a decision according to an output result, and the operation efficiency of an airline company is improved.

Drawings

FIG. 1 is a flow chart of the method of the present invention.

FIG. 2 shows the flight of 100 monitored flights according to the present inventionThe predicted value P of the oxygen pressure of the flight unit at the approach stage is obtained after the data is calculated by the leakage model of the oxygen system of the flight unit_pre-aAnd the actual value P_aResidual error between_aAnd (5) distribution diagram.

FIG. 3 is a plot of crew oxygen pressure versus time plotted from least squares linear regression fit of historical flight data for an aircraft.

Detailed Description

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate an embodiment of the invention and, together with the description, serve to explain the invention and not to limit the invention.

As shown in fig. 1, the present embodiment provides a leakage early warning method for a civil aircraft oxygen system based on flight data, which specifically includes the following steps:

step 1: a civil aircraft unit oxygen system leakage early warning method based on flight data can monitor a plurality of overhead flights simultaneously. FLIGHT data (M is 100 in this embodiment) related to the CREW oxygen system recorded by M FLIGHTs of the same model, including but not limited to TIME (TIME), FLIGHT PHASE (FLIGHT _ PHASE), CREW oxygen pressure (CREW _ OXY _ PRS), atmospheric temperature (SAT), and COCKPIT temperature (TEMP _ clock).

Step 2: and performing preprocessing work on the acquired flight data, including missing data supplement and the like. The missing data supplement method is near supplement and is suitable for flight parameter data without mutation. The acquired flight data of the monitored flight is complete flight cycle time sequence data, and the sequence of the flight data is 5 columns, namely time, flight stages, unit oxygen pressure, atmospheric temperature and cockpit temperature parameters; each row has m rows, and the rows correspond to the flight time of the flight; the preprocessed time sequence data is X ═ { X ═ X_ijWhere i is 1, 2.., m; j ═ 1, 2., 5, x_ijIs ith row and jth column data.

And step 3: and (3) carrying out characteristic value extraction work on the preprocessed data, wherein the specific characteristic values comprise: crew oxygen pressure P when aircraft is at takeoff airport_dAtmospheric temperature T of take-off airport_dOxygen pressure P of unit in approach stage_aAnd the lowest atmospheric temperature T in the course of the route_min. The characteristic value extraction method comprises the following steps:

(1) determining the crew oxygen pressure P at the takeoff airport of the aircraft according to the FLIGHT PHASE (FLIGHT _ PHASE)_dAtmospheric temperature T of take-off airport_dData extraction point t of_dCorresponds to the moment when the flight has just entered the engine off (eng.stop) phase, i.e. when double-start has just been detected 2N2 < 55%. If the length of the sampled data is t, then

x_i3For the data in row i, column 3, x in the time series data of the flight cycle_i4Is the data of the 4 th column of the ith row in the flight cycle timing data.

(2) Determining the crew oxygen pressure P in the approach PHASE according to the FLIGHT PHASE (FLIGHT _ PHASE)_aData extraction point t of_aSetting the length of the sampling data as t corresponding to the moment when the flight just enters the APPROACH (APPROACH) stage

(3) Determining a minimum atmospheric temperature T in an airway process_minData extraction point t of_minSetting the length of the sampling data as t corresponding to the time of the initial occurrence of the minimum value of the atmospheric temperature

(4) In this embodiment, the sample data length t is set to 19s, that is, the characteristic value is an average value within 20 s.

And 4, step 4: importing the characteristic value extracted in the step 3 into an oxygen system leakage model of the flight unit for calculation to obtain a predicted value P of the oxygen pressure of the flight unit at the approach stage of the monitored flight_pre-aAnd the actual value P_aResidual error between_a. The model of the leakage of the oxygen system of the flight set is established as follows:

(1) and (4) selecting historical flight data without leakage faults of 200 flights, wherein the historical flight data is from the same model as the model in the step 1. Data processing and characteristic value extraction are carried out through the steps 1-3, and a characteristic value matrix formed by the extracted characteristic values is as follows:

(2) to E₁The characteristic value data in the interior is analyzed by P_dn、T_dn、T_{min n}Is an independent variable, P_an＝[P_an-1，P_an-2，…P_an-n]，P_dn＝[P_dn-1，P_dn-2，…P_dn-n]，T_dn＝[T_dn-1，T_dn-2，…T_dn-n]，T_{min n}＝[T_{min n-1}，T_{min n-2}，…T_{min n-n}]Establishing a regression fitting equation by adopting a multiple linear regression method, and determining the leakage model of the oxygen system of the flight unit as follows:

P_pre-a＝β₀+β₁·P_d+β₂·T_d+β₃·T_min

formula (III) β₀≈107.97，β₁≈0.97，β₂≈-1.44，β₃≈1.88。

(3) P of oxygen pressure of unit in approach phase_a＝P_pre-a+_a，_a～N(μ，σ²) Unbiased estimation of (normal distribution of the path), sigma

In the formula

Is the residual mean of the sample, where k is 1, 2.

(4) Leading the extracted characteristic value of the monitored flight into a model for calculation to obtain a predicted value P of the oxygen pressure of the unit at the approach stage of the monitored flight_pre-aResidual error from actual value_a。

And 5: residual standard deviation S calculated by the residual error calculated by the model in the step 4 and sample data consisting of abnormal flights of the unit oxygen system through the leakage model of the flight unit oxygen system_dAnd comparing and carrying out grading early warning treatment. If the residual error of the flight is monitored_a≤2.5S_dJudging that the oxygen system of the unit is normal without leakage risk, and monitoring the next flight; if the residual is 2.5S_d＜_a＜3.5S_dIf the possibility of slow leakage of the oxygen system of the unit exists, turning to step 6; if the residual error_a≥3.5S_dAnd then the alarm is given to the maintenance personnel of the engine and the leakage of the oxygen system of the unit is checked.

FIG. 2 is the residual error of flight data of 100 monitored flights calculated by the flight crew oxygen system leakage model in the embodiment_aAnd (5) distribution diagram. It can be seen that the calculated residual error of most flight data_aAre all less than 2.5S_dWithin the normal range; residual 2.5S after small part (circled) flight data calculation_d＜_a＜3.5S_dIf the unit oxygen system has a slow leakage tendency, turning to step 6; does not exceed 3.5S_dOf flights.

Step 6: residual 2.5S_d＜_a＜3.5S_dWhen the airplane is monitored within the range, the airplane has the tendency of slow leakage of a crew oxygen system, and needs to be further judged. 30 flights of historical flight record data are collected that the aircraft was next to before performing the monitoring flight (the flight is circled in fig. 2). And (4) carrying out pretreatment work according to the method of the step 2.

And 7: extracting characteristic values of the historical flight record data preprocessed in the step 6,in step 6, each piece of historical flight data is a piece of complete flight cycle time sequence data, the number of columns of the complete flight cycle time sequence data is 5, each column has m1 rows, and the specific characteristic values include: departure time D of each flight_tThe crew oxygen pressure P at the takeoff airport of the aircraft on each flight_d1Temperature of the cockpit of each flight_cocAnd the atmospheric temperature T of the takeoff airport in each flight_d1Based on T_cocAnd T_d1Oxygen pressure P of the aircraft units at the take-off airport_d1Conversion to T at standard temperature_stdOxygen pressure value P of the unit_std；

(1)P_d1，P_d1In a manner consistent with that described in step 3.

(2) Recorded departure time D for each flight_tThe aircraft departure time for the corresponding flight; aircraft cockpit temperature T at takeoff airport_cocCorresponding to the time t when the flight just enters the engine off (ENG_d1If the sample data length is t1, the sample data length is set to

x_I4For the data of row I, column 4 in each time series data.

(3) The sample data length t is set to 19s, i.e., the characteristic value is the average value within 20 s.

(4) The oxygen pressure P of the unit when the airplane is positioned at the take-off airport_d1Conversion to standard temperature T_stdOxygen pressure value P of lower unit_stdIs calculated by the formula

In this example T_std＝20℃，D_t，P_stdThe eigenvalue matrix of the composition is as follows:

wherein D_t·pFor the departure time of the aircraft of the P flight, P_std·pP for the P-th flight point_stdIn (1).

And 8: for eigenvalue matrix E₂Establishing an equation P of the change of the oxygen pressure of the aircraft unit along with the time by using least square normal linear regression fitting_std＝β₄+β₅·D_tObtaining the oxygen pressure leakage rate β of the aircraft crew₅In formula β₄≈1856.75，β₅≈-2.42。

Step 9, the oxygen pressure leakage rate β of the aircraft crew₅Crew oxygen pressure leakage rate β of healthy state aircraft (aircraft without abnormal state of crew oxygen system)_stdComparing, calculating and judging parameters α, if α is less than or equal to 5 degrees, judging that the crew oxygen system is normal without slow leakage risk, if α is more than 5 degrees, judging that the crew oxygen system is in slow leakage risk, giving an alarm to crew maintenance personnel, and checking the leakage of the crew oxygen system, wherein the crew oxygen pressure leakage rate β of the aircraft in a healthy state is calculated_stdThe method for calculating the judgment parameter α includes:

(1) crew oxygen pressure leak rate β of healthy aircraft_stdThe unit oxygen pressure leakage rate β is obtained by selecting continuous M (M is 50 in the embodiment) flight data of 30 airplanes in the same type of health state, and sequentially establishing an equation of change of the aircraft unit oxygen pressure with time according to the steps 7 and 8_5·s(s ═ 1, 2.., h), according to the formula

Finding β_stdThe unit oxygen pressure leakage rate β obtained in the embodiment has universality_std＝-0.4543。

(2) The formula of the judgment parameter α is α ═ arctan (β)_std)-arctan(β₅) The judgment parameter α obtained in this embodiment is 43.15 degrees, which is much larger than 5 degrees, so that the risk of slow leakage of oxygen of the unit exists, and an alarm should be given to the crew maintenance personnel to check the leakage of the oxygen system of the unit.

FIG. 3 shows an embodiment of performing residual 2.5S_d＜_a＜3.5S_dThe time-varying graph of the crew oxygen pressure plotted by least squares linear regression fitting of 30 adjacent historical flight record data of the airplane before the flight, the time-varying graph of the crew oxygen pressure comprising the airplane and the standard crew oxygen pressure leakage rate β of the airplane according to the health state_stdAnd (5) drawing a time-varying curve of the oxygen pressure of the unit. It can be clearly seen that the crew oxygen decay rate of the aircraft is much greater than that of a healthy aircraft, and it can be determined that the aircraft has a crew oxygen system leakage fault.

The embodiments of the present invention have been described in detail with reference to the drawings, but the present invention is not limited to the above embodiments, and various changes can be made within the knowledge of those skilled in the art without departing from the gist of the present invention.

Claims

1. A civil aircraft unit oxygen system leakage early warning method based on flight data is characterized by specifically comprising the following steps:

step 1: collecting flight data of monitored flights of civil aircraft A;

And 4, step 4: obtaining historical flight data of n flights without leakage faults, wherein the model of an airplane executing the n flights is the same as that of a civil airplane A; preprocessing the historical flight data, extracting characteristic values from the processed data, establishing a unit oxygen system leakage model according to the extracted characteristic values, and extracting P from the step 3_d、T_d、P_aAnd T_minLeading the model into the model to obtain the approach stage of the monitored flightOxygen pressure prediction value P of unit_pre-aThereby obtaining P_pre-aAnd P_aResidual error between_a；

and 7: and 6, extracting characteristic values of the data preprocessed in the step 6, wherein the specifically extracted characteristic values comprise: departure time D of each flight_tThe crew oxygen pressure P at the takeoff airport of the aircraft on each flight_d1Temperature of the cockpit of each flight_cocAnd the atmospheric temperature T of the takeoff airport in each flight_d1Based on T_cocAnd T_d1Oxygen pressure P of the aircraft units at the take-off airport_d1Conversion to T at standard temperature_stdOxygen pressure value P of the unit_std；

And step 9: selecting continuous N pieces of historical flight data of abnormal-free airplanes of an h-frame unit oxygen system, wherein the h-frame airplanes are the same as the civil airplane A in type; performing characteristic extraction on flight data of each flight of each airplane to obtain the unit oxygen pressure P when the airplane is positioned at a take-off airport in each flight of the s-th airplane_d-sAnd departure time of each flight, will P_d-sConversion to standardOxygen pressure value P of unit at temperature_{s ds}According to P of each flight of the s-th aircraft_{s ds}And taking off time, and establishing a time-dependent change equation of the oxygen pressure of the set of the s-th aircraft by using a least square method so as to obtain the oxygen pressure leakage rate β of the set of the s-th aircraft_5·sAnd(s) is 1,2 and … h, the crew oxygen pressure leakage rate of the h aircrafts is averaged, and thus the crew oxygen pressure leakage rate β of the crew oxygen system abnormal-free aircraft is obtained_std；

2. The aircraft crew oxygen system leakage early warning method based on flight data as claimed in claim 1, wherein in step 1, the flight data of the monitored flight of civil aircraft a is collected in a fast access recorder.

3. The civil aircraft crew oxygen system leakage early warning method based on flight data as claimed in claim 1, wherein the flight data of the monitored flight obtained in step 1 is a complete flight cycle time sequence data, which has 5 columns in total, and sequentially comprises time, flight phase, crew oxygen pressure, atmospheric temperature and cockpit temperature parameters; each row has m rows, and the rows correspond to the flight time of the flight; in the step 2, a near supplementation method is adopted for missing data supplementation, and the preprocessed time sequence data is X ═ X_ijWhere i is 1, 2.., m; j ═ 1, 2., 5, x_ijThe data of the ith row and the jth column in X.

4. The civil aircraft crew oxygen system leakage early warning method based on flight data as claimed in claim 3, wherein the step 3 is specifically:

step 3.2: oxygen pressure P of unit in approaching stage_aThe data extraction time of (2) is set as the time t when the flight enters the approach stage_aThen, then

Step 3.3: determining a minimum atmospheric temperature T in an airway process_minIs the time t of the first occurrence of the minimum value of the atmospheric temperature_minThen, then

5. The civil aircraft crew oxygen system leakage early warning method based on flight data according to claim 1, wherein the establishing of the crew oxygen system leakage model in the step 4 specifically comprises:

Wherein P is_dn·nCrew oxygen pressure, T, at the takeoff airport of an aircraft on the nth flight in historical flight data representing no leakage faults_dn·nAtmospheric temperature, T, at takeoff airport of nth flight in historical flight data representing no leakage fault_minn·nRoute of nth flight in historical flight data representing no leakage faultMinimum atmospheric temperature in the process, P_an·nThe crew oxygen pressure of the nth flight in the historical flight data representing no leakage fault in the approach phase;

step 4.2: to E₁And analyzing characteristic value data: with P_anAs a dependent variable, with P_dn、T_dn、T_minnIs an independent variable, P_an＝[P_an-1,P_an-2,…P_an-n]，P_dn＝[P_dn-1,P_dn-2,…P_dn-n]，T_dn＝[T_dn-1,T_dn-2,…T_dn-n]，T_minn＝[T_minn-1,T_minn-2,…T_minn-n]And establishing a regression fitting equation by adopting a multiple linear regression method so as to determine the leakage model of the oxygen system of the flight unit: p_pre-a＝β₀+β₁·P_dn+β₂·T_dn+β₃·T_minnWherein β₀To fit constant terms, β₁、β₂、β₃Are coefficients.

6. The civil aircraft crew oxygen system leakage early warning method based on flight data as claimed in claim 1, wherein each historical flight data in step 6 is a complete flight cycle time sequence data, there are 5 columns in total, and time, flight phase, crew oxygen pressure, atmospheric temperature and cockpit temperature parameters are sequentially set; each column has m1 rows, and the preprocessed time sequence data is X_e＝{x_IJ1,2, m 1; 1,2, 5, X_eIs the e-th time sequence data, e is 1,2, … p, x_IJIs X_eRow I and column J;

p in said step 7_stdComprises the following steps:

7. The aircraft crew oxygen system leakage early warning method based on flight data as claimed in claim 1, wherein step 8 specifically adopts a least square method to p groups D in step 7_tAnd P_stdLinear regression fitting is carried out to establish a time-varying equation P of the oxygen pressure of the civil aircraft A_std＝β₄+β₅·D_tWherein β₄Fitting equation constant terms.

8. The aircraft crew oxygen system leakage early warning method based on flight data according to claim 1, wherein the threshold value α determination method in the step 9 is as follows:

α＝arctan(β₅)-arctan(β_std)。