Risk early warning and inducement identification method and system for event of aircraft rushing / deviating from runway

Abstract

The invention discloses a risk early warning and inducement identification method and system for an aircraft runway rushing / deviating event. The method comprises the steps that S0, a key parameter set and a risk factor set of the runway rushing / deviating event are obtained through FOQA standards, SOP and QRH; s1, constructing a corresponding relation between key parameters and risk factors based on a Delphi method; s2, constructing a runway running and deviating risk prediction model based on Transform, and realizing online risk identification; and S3, obtaining the most similar inducement based on a hypothesis testing method, and completing risk factor identification.

Description

Technical Field

[0001] The present invention belongs to the field of civil aviation safety technology and flight data application, and particularly relates to a risk warning and cause identification method and system for aircraft runway overrun / runoff events. Background Art

[0002] Runway overruns or excursions during the approach and landing phase are one of the leading causes of fatal accidents. Flight quality monitoring technology, based on post-flight data analysis, is a key means of causal analysis and risk assessment for runway overruns and excursions, and is of great significance to flight safety. However, current flight quality monitoring technology is still insufficient in terms of real-time early warning and tracing of the causes of runway overruns and excursions. This is mainly manifested in the following aspects:

[0003] First, the existing runway overrun / excursion detection method mainly combines flight data after the flight to determine whether a runway overrun / excursion has occurred. This makes it difficult to achieve real-time monitoring of aircraft status and online risk warnings, making it impossible for flight crews to detect potential risks in advance and further formulate corresponding strategies to prevent threats and errors from further evolving into runway overrun / excursion incidents.

[0004] Secondly, there is a lack of automated methods for identifying the causes of runway overruns and excursions. Currently, identifying the causes of runway overruns and excursions relies on manual analysis by flight safety experts combined with post-flight data to further determine the risk type. Automated identification of the causes of runway overruns and excursions could help flight safety experts better understand risk patterns and, to a certain extent, provide flight crews with information to support decision-making and take appropriate preventative measures. Summary of the Invention

[0005] To address the problems of the prior art, the present invention provides a method and system for risk warning and cause identification of aircraft overrun / runoff events. For aircraft overrun / runoff events, the present invention obtains key parameters related to aircraft takeoff and landing based on aircraft operation quality monitoring standards, flight standard operating procedures, and flight quick checklists, and establishes a risk factor fault tree to obtain the risk factors that cause aircraft overrun / runoff. Subsequently, the Delphi method is used to obtain the correspondence between key parameters and risk factors to achieve an objective analysis of risk factors. On this basis, an abnormality prediction model is established to achieve online risk monitoring and early warning, and combined with a hypothesis testing method to complete automated cause identification.

[0006] To achieve the above object, the present invention provides the following solutions:

[0007] A risk warning and cause identification method for an aircraft runway overrun / runoff event, the method comprising:

[0008] S0: Obtain a set of key parameters related to aircraft takeoff and landing through flight quality monitoring standards, flight standard operating procedures, and flight quick reference manuals; obtain a set of risk factors that cause runway overrun / excursion events by establishing a risk factor fault tree;

[0009] S1: Construct the corresponding relationship between key parameters and risk factors based on the modified Delphi method;

[0010] S2: Build a Transformer-based runway overrun / runoff risk prediction model. This model extracts runway overrun / runoff event features from flight data and uses historical flight data sequences to predict the aircraft's runway centerline deviation and remaining runway distance at a specific time in the future, enabling online risk identification.

[0011] S3: Based on the hypothesis testing method, the key parameters of the warning flights and normal flights that exceed the preset threshold are obtained. The most similar inducement is output based on the correspondence between the obtained key parameters and the risk factors to complete the risk factor identification.

[0012] Preferably, in said S0, the key parameter PV={v 1 ,v 2 ,...,v d} and risk factors PR={R1,R2,...,R m Specifically include:

[0013] Key parameter PV={v 1 ,v 2 ,...,v d}, d is the number of key parameters, x 1 The first key parameter; the key parameter PV includes two parts: PX and PW. PX is the key parameter set obtained through flight data. PX={x 1 ,x 2 ,...,x b}, b is the number of elements in PX, x 1 is the first element in PX; PW is a set of key parameters obtained through meteorological message data and airport pavement data, PW = {w 1 ,w 2 ,...,w c}, c is the number of elements in PW, and d=b+c, w 1 is the first element in PW;

[0014] Risk factor PR={R1,R2,...,R m}, m is the number of risk factors, and R1 represents the first risk factor.

[0015] Preferably, in said S1, the key parameter PV is established based on the modified Delphi method. 1 ,v 2 ,...,v d} and risk factors PR={R1,R2,...,R m}, specifically including:

[0016] S1.1: Organize an interdisciplinary expert team and design and distribute an anonymous scoring questionnaire to assess any risk factor R j , j∈[1,m] and any key parameter v i , the degree of association of i∈[1,d], and the subjective comments of experts;

[0017] S1.2: Based on the scoring questionnaire obtained from each expert, for each risk factor R j , j∈[1,m], construct the parameter association matrix Where n is the number of experts, is the key parameter v of the rth expert pair i and risk factors R j The correlation degree score is then introduced; the evaluation index mean μ i,j , consensus index ECI i,j , interquartile range IQR i,j , all the anonymous rating data of experts, mean μ i,j , consensus index ECI i,j , interquartile range IQR i,j And subjective comments are provided to each expert again, and each expert needs to focus on μ i,j <3.5, ECI i,j <0.7, IQR i,j The scores of the above two and the indicators mentioned in the subjective comments are improved and updated, and the process is repeated 2 to 3 times;

[0018] S1.3: According to the final scoring data, if μ i,j ≥3.5, the risk factor R j Affected by the key parameter v i The impact of is greater than the preset threshold, and the corresponding relationship between key parameters and risk factors is finally obtained, which is recorded as R j =f(v h ,...,v p ), 1≤h≤p≤m.

[0019] Preferably, in S2, a Transformer-based runway overrun / runoff risk prediction model is constructed, and runway overrun / runoff event features are extracted from flight data using the Transformer-based runway overrun / runoff risk prediction model. The aircraft's deviation from the runway centerline and the remaining runway distance at a certain moment in the future are obtained from a historical parameter sequence. The method for realizing online risk identification includes:

[0020] S2.1: For the flight data sequence of the aircraft [x1, x2, ..., x t ] T , using a 6-layer Encoder layer structure to transform the flight parameter sequence into a high-dimensional feature matrix Z with time-dependent and multi-parameter associations enc =[z1,z2,...,z t ] T , which is then input into the 4-layer Decoder layer for decoding. Based on the global features and historical information provided by the Encoder layer, the runway status at future moments is gradually deduced, and the final output is the distance d from the runway centerline at the future time t+k. t+k and the remaining runway distance L t+k ,in Represents the flight data sequence record of the flight at time t, Represents the recorded value of the first flight data at time t, Z enc is a high-dimensional time series feature matrix, is the time series feature vector of the flight at time t, b′ is the hidden layer dimension of the model;

[0021] S2.2: Based on historical flight data, set the 95% percentile of the runway centerline deviation distance of all flights at time t+k as the maximum value of the centerline deviation distance threshold range. The 5% quantile is set as the minimum value of the threshold interval of the distance from the center line The 5% quantile of the remaining runway distance of all flights at time k is set as the remaining runway distance threshold μ t+k ;when or When L t+k ≤μ t+k , it indicates that the aircraft is at risk of running off the runway at time t+k; when the aircraft is at risk of running off the runway, the key parameters of the current warning time t and time t are recorded. Then continue to monitor the risk status at the next moment.

[0022] Preferably, in S3, based on a hypothesis testing method, key parameters of the warning flight and the normal flight are obtained, and the most similar inducement is outputted in combination with the corresponding relationship between the obtained key parameters and the risk factors. The method for completing the risk factor identification includes:

[0023] When an early warning is issued at time t, the key parameters at time t Key parameters of other normal flights at time t Conduct hypothesis testing, where 1, 2, ..., N represents 1 to N different flights. (i∈[1,N]) represents the key parameters of the i-th flight at time t. The key parameters of the warning flight and the normal flight that exceed the preset threshold are recorded as the inducement parameter R′={v g ,...,v q}, 1≤g≤q≤m; then calculate the incentive parameter R′={v g ,...,v q} and each corresponding relationship R j =f(v h ,...,v p ), Jaccard similarity of j∈[1,m]; finally output the risk factor R with the largest Jaccard similarity j , complete risk factor identification.

[0024] The present invention also provides a risk warning and cause identification system for aircraft overrun / runway deviation events, the system being used to implement the aforementioned method, the system comprising: an acquisition module, a construction module, an online risk identification module, and a risk factor identification module;

[0025] The acquisition module is used to obtain a set of key parameters related to aircraft takeoff and landing through flight quality monitoring standards, flight standard operating procedures, and flight quick reference manuals; and to obtain a set of risk factors that cause runway overrun / runoff events by establishing a risk factor fault tree;

[0026] The construction module is used to construct the corresponding relationship between key parameters and risk factors based on the modified Delphi method;

[0027] The online risk identification module is used to build a Transformer-based runway overrun / runoff risk prediction model, extract runway overrun / runoff event features from flight data using the Transformer-based runway overrun / runoff risk prediction model, and use historical flight data sequences to predict the aircraft's deviation from the runway centerline and remaining runway distance at a certain time in the future, thereby achieving online risk identification;

[0028] The risk factor identification module is used to obtain key parameters that indicate that the difference between the warning flight and the normal flight exceeds a preset threshold based on a hypothesis testing method, and output the most similar inducement based on the correspondence between the obtained key parameters and the risk factors to complete the risk factor identification.

[0029] Preferably, in the acquisition module, the key parameter PV={v 1 ,v 2 ,...,v d} and risk factors PR={R1,R2,...,R m Specifically include:

[0030] Key parameter PV={v 1 ,v 2 ,...,v d}, d is the number of key parameters, v 1 The first key parameter; the key parameter PV includes two parts: PX and PW. PX is the key parameter set obtained through flight data. PX={x 1 ,x 2 ,...,x b}, b is the number of elements in PX, x 1 is the first element in PX; PW is a set of key parameters obtained through meteorological message data and airport pavement data, PW = {w 1 ,w 2 ,...,w c}, c is the number of elements in PW, and d=b+c, w 1 is the first element in PW;

[0031] Risk factor PR={R1,R2,...,R m}, m is the number of risk factors, and R1 represents the first risk factor.

[0032] Preferably, the construction module includes: a scoring questionnaire unit, a matrix construction unit, and a comparison unit;

[0033] The scoring questionnaire unit is used to form an interdisciplinary expert team and design and distribute anonymous scoring questionnaires to evaluate any risk factor R j , j∈[1,m] and any key parameter v i , the degree of association of i∈[1,d], and the subjective comments of experts;

[0034] The matrix construction unit is used to obtain the scoring questionnaire of each expert for each risk factor R j , j∈[1,m], construct the parameter association matrix Where n is the number of experts, is the key parameter v of the rth expert pair i and risk factors R j The correlation degree score is then introduced; the evaluation index mean μ i,j , consensus index ECI i,j , interquartile range IQR i,j , all the anonymous rating data of experts, mean μ i,j , consensus index ECI i,j , interquartile range IQR i,j And subjective comments are provided to each expert again, and each expert needs to focus on μ i,j <3.5, ECI i,j <0.7, IQR i,j The scores of the above two and the indicators mentioned in the subjective comments are improved and updated, and the process is repeated 2 to 3 times;

[0035] The comparison unit is used to calculate the final score data. i,j ≥.5, the risk factor R j Affected by the key parameter v i The impact of is greater than the preset threshold, and the corresponding relationship between key parameters and risk factors is finally obtained, which is recorded as R j =f(v h ,...,v p ), 1≤h≤p≤m.

[0036] Preferably, the online risk identification module includes: a network construction unit and a risk monitoring unit;

[0037] The network construction unit is used for the flight data sequence [x1, x2, ..., x t ] T , using a 6-layer Encoder layer structure to transform the flight parameter sequence into a high-dimensional feature matrix Z with time-dependent and multi-parameter associations enc =[z1,z2,...,z t ] T , which is then input into the 4-layer Decoder layer for decoding. Based on the global features and historical information provided by the Encoder layer, the runway status at future moments is gradually deduced, and the final output is the distance d from the runway centerline at the future time t+k. t+k and the remaining runway distance L t+k ,in Represents the flight data sequence record of the flight at time t, Represents the recorded value of the first flight data at time t, Z enc is a high-dimensional time series feature matrix, is the time series feature vector of the flight at time t, b′ is the hidden layer dimension of the model;

[0038] The risk monitoring unit is used to set the 95% percentile of the runway centerline deviation distance of all flights at time t+k as the maximum value of the centerline deviation distance threshold range based on historical flight data The 5% quantile is set as the minimum value of the threshold interval of the distance from the center line The 5% quantile of the remaining runway distance of all flights at time k is set as the remaining runway distance threshold μ t+k ;when or When L t+k ≤μ t+k , it indicates that the aircraft is at risk of running off the runway at time t+k; when the aircraft is at risk of running off the runway, the key parameters of the current warning time t and time t are recorded. Then continue to monitor the risk status at the next moment.

[0039] Preferably, in the risk factor identification module, based on a hypothesis testing method, key parameters of the warning flight and the normal flight that differ by a preset threshold are obtained, and the most similar inducement is output based on the corresponding relationship between the obtained key parameters and the risk factors. The process of completing the risk factor identification includes:

[0040] When an early warning is issued at time t, the key parameters at time t Key parameters of other normal flights at time t Conduct hypothesis testing, where 1, 2, ..., N represents 1 to N different flights. (i∈[1,N]) represents the key parameters of the i-th flight at time t. The key parameters of the warning flight and the normal flight that exceed the preset threshold are recorded as the inducement parameter R′={v g ,...,v q}, 1≤g≤q≤m; then calculate the incentive parameter R′={v g ,...,v q} and each corresponding relationship R j =f(v h ,...,v p ), Jaccard similarity of j∈[1,m]; finally output the risk factor R with the largest Jaccard similarity j , complete risk factor identification.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] (1) Traditional runway overrun / runoff detection methods determine whether a runway overrun / runoff event has occurred based on flight data after the flight, and are unable to provide online risk warnings. The present invention proposes a runway overrun / runoff risk prediction model, which builds a high-performance deep model based on Transformer, takes historical flight data sequences as input, and predicts the distance the aircraft will deviate from the runway centerline and the remaining runway distance at future moments. By providing risk warning information within a period of time before the runway overrun / runoff event occurs, real-time monitoring of the aircraft's flight status and runway overrun / runoff risk warnings can be achieved, and a certain amount of time is provided for the flight crew to further formulate corresponding strategies to avoid the runway overrun / runoff event.

[0043] (2) The traditional causal analysis method of runway overrun / runoff incidents requires flight safety experts to manually analyze the risk types that induce the runway overrun / runoff incidents in combination with post-flight data. This method consumes a lot of costs. The present invention proposes a method for constructing the correspondence between risk factors and key parameters. Based on the Delphi method, it integrates the experience of multiple experts to obtain the influence relationship and coupling mechanism between risk factors and key parameters, and further realizes automated cause identification by combining abnormal prediction models and hypothesis testing methods. This provides flight safety experts with risk cause information of runway overrun / runoff incidents, assists them in understanding risk patterns, and provides reference information in the post-flight causal analysis process. At the same time, it also provides relevant abnormal parameters to flight crew members, which can assist them in taking correct preventive measures to a certain extent to avoid the occurrence of runway overrun / runoff incidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in the embodiments. 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 any creative work.

[0045] Figure 1 This is a schematic diagram of the main steps S0-S3 of the embodiment of the present invention;

[0046] Figure 2 The present invention provides a flow chart of a risk warning and cause identification method for aircraft runway overrun / runoff events. DETAILED DESCRIPTION

[0047] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0048] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0049] Example 1

[0050] like Figure 1 、 Figure 2 As shown, the present invention provides a risk warning and cause identification method for aircraft runway overrun / runoff events, comprising the following steps:

[0051] S0: First, refer to the Flight Operation Quality Assurance (FOQA) standards, Standard Operating Procedures (SOPs) and Quick Reference Handbook (QRH) to comprehensively sort out the key parameters related to takeoff and landing. 1 ,v 2 ,...,v d}. Where d is the number of key parameters, v 1 The first key parameter is PV. The key parameter PV includes two parts: PX and PW. PX is the key parameter set obtained through flight data. PX={x 1 ,x 2 ,...,x b}, b is the number of elements in PX, x 1 is the first element in PX; PW is a set of key parameters obtained through meteorological message data, airport pavement data, etc., PW={w 1 ,w 2 ,...,w c}, c is the number of elements in PW, and d=b+c, w 1 It is the first element in PW. Then, for different stages (take-off and landing) and different types (deviation and overrun), corresponding risk factor fault trees T are established respectively, so as to further accurately and comprehensively obtain the risk factors that cause the aircraft to overrun / overrun the runway. Among them, the fault tree is constructed based on the fault event information related to the aircraft overrun / overrun, and includes multiple fault events related to the aircraft overrun / overrun and the causal relationship between multiple fault events. Specifically, the fault tree is constructed according to the top event T, the intermediate event PI corresponding to the top event = {I1, I2, ..., I j} and the sub-event PB corresponding to the intermediate event = {B1, B2, ..., B k}. PI is the set of intermediate events corresponding to the top event T, j is the number of elements in PI, and I1 is the first element in PI; PB is the set of sub-events corresponding to the intermediate event set, k is the number of elements in PB, and B1 is the first element in PI. There is a causal relationship between events at different levels, that is, the top event T is the result, and the intermediate event PI corresponding to the top event is the cause of the top event; the intermediate event PI is the result, and the sub-event PB corresponding to the intermediate event is the cause of the intermediate event. All basic events of the fault tree are defined as risk factors that cause the aircraft to overshoot / exit the runway: PR = {R1, R2, ..., R m}, where the basic event is the lowest level event in the fault tree, m is the number of risk factors, and R1 represents the first risk factor.

[0052] S1: Based on the modified Delphi method, the key parameter PV = {v 1 ,v 2 ,...,v d} and risk factors PR={R1,R2,...,R m}, is used to accurately identify and analyze the risk factors of overrunning / runway deviation. The specific steps are as follows:

[0053] (S1.1) Parameter-factor correlation assessment based on subject matter experts (SMEs)

[0054] A multidisciplinary expert team (15-20 people) was formed, including senior pilots (about 40%), aviation engineering experts (about 30%), flight data analysts (about 20%), and human factors experts (about 10%). An anonymous scoring questionnaire was designed and distributed to evaluate any risk factor R. j (j∈[1,m]) and any key parameter v i The degree of correlation between (i∈[1,d]) and the subjective comments of experts are collected. The rating levels include "1-irrelevant", "2-weakly irrelevant", "3-moderately relevant", "4-strongly relevant", and "5-strongly relevant".

[0055] (S1.2) Update of scoring data based on SME

[0056] Based on the scoring questionnaire of each expert obtained in (S1.1), for each risk factor R j (j∈[1,m]), construct the parameter association matrix Where n is the number of experts, is the key parameter v of the rth expert pair i and risk factors R j The correlation degree score is then introduced: mean μ i,j, consensus index ECI i,j , interquartile range IQR i,j ,in Key parameter v i and risk factors R j The mean score of the degree of association, if μ i,j ≥3.5, it indicates that the key parameter v i and risk factors R j Strong correlation; ECI i,j =1-σ i,j / μ i,j , is the score standard deviation, ECI i,j Used to describe the degree of concentration of expert opinions. If ECI i,j ≥0.7, indicating that the experts' opinions are highly consistent; IQR i,j for If IQR i,j ≥2, it indicates that there is a large disagreement in the ratings.

[0057] All experts’ anonymous rating data, mean μ i,j , consensus index ECI i,j , interquartile range IQR i,j And subjective comments are provided to each expert again, and each expert needs to focus on μ i,j <3.5, ECI i,j <0.7, IQR i,j The scores of ≥2 and the indicators mentioned in the subjective comments were combined, and the original scoring results were improved and updated. The process was repeated 2 to 3 times.

[0058] (S1.3) Correspondence confirmation based on scoring data

[0059] According to the final scoring data, if μ i,j ≥3.5, the risk factor R j Affected by the key parameter v i The influence of is greater, and finally the corresponding relationship between key parameters and risk factors is obtained, which is recorded as R j =f(v h ,...,v p ), 1≤h≤p≤m.

[0060] S2: Build a Transformer-based runway overrun / runoff risk prediction model. This model extracts runway overrun / runoff event characteristics from flight data and obtains the aircraft's runway centerline deviation and remaining runway distance at a certain point in the future from historical parameter sequences. The specific steps are as follows:

[0061] (S2.1) Building a Transformer Network

[0062] For the flight data sequence of the aircraft in the time range 0 to t [x1, x2, ..., x t ] T , using a 6-layer Encoder layer structure to transform the flight parameter sequence into a high-dimensional feature matrix Z with time-dependent and multi-parameter associations enc =[z1,z2,...,z t ] T It is then input into the 4-layer Decoder layer for decoding. Based on the global features and historical information provided by the Encoder layer, the runway status at future moments is gradually deduced, and the final output is the distance d from the runway centerline at the future time t+k. t+k and the remaining runway distance L t+k .in Represents the flight data sequence record of the flight at time t, Represents the recorded value of the first flight data at time t, Z enc is a high-dimensional time series feature matrix, is the time series feature vector of the flight at time t, b′ is the hidden layer dimension of the model, d t+k represents the distance from the runway centerline at the future time t+k predicted at time t, L t+k Represents the remaining runway distance at time t+k predicted at time t.

[0063] Specifically, the distance d from the runway centerline t+k The angle δ between the line connecting the center of the aircraft and the localizer and the centerline of the runway t+k and the distance between the aircraft and the localizer, d t+k =(L0+SL′ t+k )·tanδ t+k ; Remaining runway distance L t+k Calculated by subtracting the distance between the aircraft and the runway threshold from the total length of the runway, L t+k =SL′ t+k Among them, δ t+k is the angle between the center line of the aircraft and the localizer and the center line of the runway at time t+k, L0 is the distance between the localizer and the end of the runway, S is the runway length, L′ t+k is the vertical distance between the aircraft and the entrance at time t+k, d t+k >0 represents the maximum distance that the center of gravity of the aircraft deviates from the left side of the runway centerline (the direction of the aircraft's takeoff during takeoff and the direction of the aircraft's landing during landing), d t+k <0 represents the maximum distance the aircraft nose deviates to the right of the runway centerline (same as above).

[0064] A Transformer-based runway overrun / runaway risk prediction model is defined as follows:

[0065]

[0066] Among them, Represents the flight data sequence record of the flight at time t, is the time series feature vector of the flight at time t. dec (·) represents the processing operation of the Decoder layer, f enc (·) represents the processing operation of the Encoder layer, f tansformer (·) represents the operation process of the Transformer network. The input is the flight data sequence of the aircraft from 0 to t [x1, x2, ..., x t ] T , obtain the predicted distance d from the runway centerline at the future time t+k t+k and the remaining runway distance L t+k .

[0067] (S2.2) Runway Overrun / Excursion Risk Monitoring

[0068] Based on historical flight data, the 95% percentile of the runway centerline deviation distance of all flights at time t+k is set as the maximum value of the centerline deviation distance threshold interval. The 5% quantile is set as the minimum value of the threshold interval of the distance from the center line The 5% quantile of the remaining runway distance of all flights at time k is set as the remaining runway distance threshold μ t+k .

[0069] when or When L t+k ≤μ t+k , it indicates that the aircraft is at risk of running off the runway at time t+k. When the aircraft is at risk of running off the runway, record the key parameters of the current warning time t and time t. Then continue to monitor the risk status at the next moment.

[0070] S3: Based on the output of the runway overrun / excursion risk prediction model, the cause of the event is traced using hypothesis testing for online risk detection. The specific steps are:

[0071] When an early warning is issued at time t, the key parameters at time t Key parameters of other normal flights at time t Conduct hypothesis testing to obtain key parameters that are significantly different between the alert flights and the normal flights. Where 1, 2, ..., N represent 1 to N different flights. (i∈[1,N]) represents the key parameters of the i-th flight at time t.

[0072] QQ plot test parameters Whether it conforms to the normal distribution. If the parameter The variance is σ 2 , normal distribution N(μ,σ 2 ), then the Z test is used for significance test. The Z test is a parameter test method. First, the Z value is calculated according to the formula. in is the sample mean, μ is the population mean, σ is the population standard deviation, and n is the sample size. At the significance level α = 0.05, the corresponding critical value Z is obtained by looking up the standard normal distribution table. α , if |Z|>Z α , it indicates that there is a significant difference between the data and the normal data.

[0073] If the parameter If the data do not conform to the normal distribution, the box plot method is used for significance test. The box plot method is a non-parametric test method. First, the lower quartile Q1, upper quartile Q3 and median of the normal data are calculated, and a box plot is drawn. Then, the interquartile range IQR = Q3-Q1 is calculated, and the lower limit of the abnormal value is determined to be Q1-1.5IQR, and the upper limit is Q3+1.5IQR. Finally, if the data to be tested is or This indicates that there is a significant difference between the data and the normal data.

[0074] By parametric test (|Z|>Z α ) or nonparametric tests ( or ), output the key parameters of the warning flight and the normal flight that are significantly different, recorded as the inducement parameter R′={v g ,...,v q}, 1≤g≤q≤m. Finally, the key parameter PV obtained by S1 is combined with {v 1 ,v 2 ,...,v d} and risk factors PR={R1,R2,...,R m} to determine the most similar inducement for overrunning / running off the runway. The specific method is: calculate the inducement parameter R′={v g ,...,v q} and each corresponding relationship R j =f(v h ,...,vp The Jaccard similarity of )(j∈[1,m]) is as follows:

[0075]

[0076] Among them, J(R′,R j ) is larger, indicating that the inducement parameter R′={v g ,...,v q} and the risk factor R j The corresponding key parameters {v h ,...,v p The more similar the risk factor is, the more similar the risk factor is. The final output is the risk factor R with the largest Jaccard similarity. j , complete risk factor identification.

[0077] Example 2

[0078] The present invention also provides a risk warning and cause identification system for aircraft overrun / runway deviation events, the system being used to implement the method described in Example 1, the system comprising: an acquisition module, a construction module, an online risk identification module, and a risk factor identification module;

[0079] The acquisition module is used to obtain a set of key parameters related to aircraft takeoff and landing through flight quality monitoring standards, flight standard operating procedures, and flight quick reference manuals; and to obtain a set of risk factors that cause runway overrun / excursion events by establishing a risk factor fault tree;

[0080] A building module for constructing the correspondence between key parameters and risk factors based on the modified Delphi method;

[0081] The online risk identification module is used to build a Transformer-based runway overrun / runoff risk prediction model. This model extracts runway overrun / runoff event characteristics from flight data and uses historical flight data sequences to predict the aircraft's deviation from the runway centerline and remaining runway distance at a certain point in the future, thus achieving online risk identification.

[0082] The risk factor identification module is used to obtain the key parameters of the warning flights and normal flights that exceed the preset threshold based on the hypothesis testing method, and output the most similar inducement based on the correspondence between the obtained key parameters and the risk factors to complete the risk factor identification.

[0083] In this embodiment, in the acquisition module, the key parameter PV={v 1 ,v 2 ,...,v d} and risk factors PR={R1,R2,...,R m Specifically include:

[0084] Key parameter PV={v 1 ,v 2 ,...,v d}, d is the number of key parameters, v 1 The first key parameter; the key parameter PV includes two parts: PX and PW. PX is the key parameter set obtained through flight data. PX={x 1 ,x 2 ,...,x b}, b is the number of elements in PX, x 1 is the first element in PX; PW is a set of key parameters obtained through meteorological message data and airport pavement data, PW = {w 1 ,w 2 ,...,w c}, c is the number of elements in PW, and d=b+c, w 1 is the first element in PW;

[0085] Risk factor PR={R1,R2,...,R m}, m is the number of risk factors, and R1 represents the first risk factor.

[0086] In this embodiment, the construction module includes: a scoring questionnaire unit, a matrix construction unit, and a comparison unit;

[0087] Scoring questionnaire unit is used to form an interdisciplinary expert team and design and distribute anonymous scoring questionnaires to evaluate any risk factor R j , j∈[1,m] and any key parameter v i , the degree of association of i∈[1,d], and the subjective comments of experts;

[0088] Matrix construction unit for each risk factor R based on the scoring questionnaire obtained from each expert j , j∈[1,m], construct the parameter association matrix Where n is the number of experts, is the key parameter v of the rth expert pair i and risk factors R j The correlation degree score is then introduced; the evaluation index mean μ i,j , consensus index ECI i,j , interquartile range IQR i,j , all the anonymous rating data of experts, mean μ i,j , consensus index ECI i,j , interquartile range IQR i,j And subjective comments are provided to each expert again, and each expert needs to focus on μi,j <3.5, ECI i,j <0.7, IQR i,j The scores of the above two and the indicators mentioned in the subjective comments are improved and updated, and the process is repeated 2 to 3 times;

[0089] Comparison unit, used to calculate the final score data. i,j ≥3.5, the risk factor R j Affected by the key parameter v i The impact of is greater than the preset threshold, and the corresponding relationship between key parameters and risk factors is finally obtained, which is recorded as R j =f(v h ,...,v p ), 1≤h≤p≤m.

[0090] In this embodiment, the online risk identification module includes: a network construction unit, a risk monitoring unit;

[0091] The network construction unit is used to calculate the flight data sequence [x1, x2, ..., x t ] T , using a 6-layer Encoder layer structure to transform the flight parameter sequence into a high-dimensional feature matrix Z with time-dependent and multi-parameter associations enc =[z1,z2,...,z t ] T It is then input into the 4-layer Decoder layer for decoding. Based on the global features and historical information provided by the Encoder layer, the runway status at future moments is gradually deduced, and the final output is the distance d from the runway centerline at the future time t+k. t+k and the remaining runway distance L t+k .in Represents the flight data sequence record of the flight at time t, Represents the recorded value of the first flight data at time t, Z enc is a high-dimensional time series feature matrix, is the time series feature vector of the flight at time t, b′ is the hidden layer dimension of the model;

[0092] The risk monitoring unit is used to set the 95% percentile of the runway centerline deviation distance of all flights at time t+k as the maximum value of the centerline deviation distance threshold range based on historical flight data The 5% quantile is set as the minimum value of the threshold interval of the distance from the center line The 5% quantile of the remaining runway distance of all flights at time k is set as the remaining runway distance threshold μ t+k ;when or When L t+k ≤μ t+k , it indicates that the aircraft is at risk of running off the runway at time t+k; when the aircraft is at risk of running off the runway, the key parameters of the current warning time t and time t are recorded. Then continue to monitor the risk status at the next moment.

[0093] In this embodiment, the risk factor identification module obtains key parameters that indicate differences between alert flights and normal flights exceeding a preset threshold based on a hypothesis testing method, and outputs the most similar inducement based on the corresponding relationship between the obtained key parameters and risk factors. The process of completing risk factor identification includes:

[0094] When an early warning is issued at time t, the key parameters at time t Key parameters of other normal flights at time t Conduct hypothesis testing, where 1, 2, ..., N represents 1 to N different flights. (i∈[1,N]) represents the key parameters of the i-th flight at time t. The key parameters of the warning flight and the normal flight that exceed the preset threshold are recorded as the inducement parameter R′={v g ,...,v q}, 1≤g≤q≤m; then calculate the incentive parameter R′={v g ,...,v q} and each corresponding relationship R j =f(v h ,...,v p ), Jaccard similarity of j∈[1,m]; finally output the risk factor R with the largest Jaccard similarity j , complete risk factor identification.

[0095] The embodiments described above are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by persons skilled in the art should fall within the scope of protection defined by the claims of the present invention.

Claims

1. A risk warning and cause identification method for aircraft runway overrun / runoff events, characterized in that: The method comprises: S0: Obtain a set of key parameters related to aircraft takeoff and landing through flight quality monitoring standards, flight standard operating procedures, and flight quick reference manuals; obtain a set of risk factors that cause runway overrun / excursion events by establishing a risk factor fault tree; S1: Construct the corresponding relationship between key parameters and risk factors based on the modified Delphi method; S2: Build a Transformer-based runway overrun / runoff risk prediction model. This model extracts runway overrun / runoff event features from flight data and uses historical flight data sequences to predict the aircraft's runway centerline deviation and remaining runway distance at a specific time in the future, enabling online risk identification. S3: Based on the hypothesis testing method, the key parameters of the warning flights and normal flights that exceed the preset threshold are obtained. The most similar inducement is output based on the correspondence between the obtained key parameters and the risk factors to complete the risk factor identification.

2. The method according to claim 1, characterized in that In the S0, the key parameter PV={v 1 ,v 2 ,...,v d } and risk factors PR={R1,R2,...,R m Specifically include: Key parameter PV={v 1 ,v 2 ,...,v d }, d is the number of key parameters, v 1 The first key parameter; the key parameter PV includes two parts: PX and PW. PX is the key parameter set obtained through flight data. PX={x 1 ,x 2 ,...,x b }, b is the number of elements in PX, x 1 is the first element in PX; PW is a set of key parameters obtained through meteorological message data and airport pavement data, PW = {w 1 ,w 2 ,,w c }, c is the number of elements in PW, and d=b+c, w 1 is the first element in PW; Risk factor PR={R1,R2,...,R m }, m is the number of risk factors, and R1 represents the first risk factor.

3. The method according to claim 1, characterized in that In S1, the key parameter PV is established based on the modified Delphi method. 1 ,v 2 ,...,v d } and risk factors PR={R1,R2,...,R m }, specifically including: S1.1: Organize an interdisciplinary expert team and design and distribute an anonymous scoring questionnaire to assess any risk factor R j , j∈[1,m] and any key parameter v i , the degree of association of i∈[1,d], and the subjective comments of experts; S1.2: Based on the scoring questionnaire obtained from each expert, for each risk factor R j , j∈[1,m], construct the parameter association matrix Where n is the number of experts, is the key parameter v of the rth expert pair i and risk factors R j The correlation degree score is then introduced; the evaluation index mean μ i,j , consensus index ECI i,j , interquartile range IQR i,j , all the anonymous rating data of experts, mean μ i,j , consensus index ECI i,j , interquartile range IQR i,j And subjective comments are provided to each expert again, and each expert needs to focus on μ i,j <3.5, ECI i,j <0.7, IQR i,j The scores of the above two and the indicators mentioned in the subjective comments are improved and updated, and the process is repeated 2 to 3 times; S1.3: According to the final scoring data, if μ i,j > 3.5, then the risk factor R j Affected by the key parameter v i The impact of is greater than the preset threshold, and the corresponding relationship between key parameters and risk factors is finally obtained, which is recorded as R j =f(v h ,...,v p ), 1≤h≤p≤m.

4. The method according to claim 1, wherein In S2, a Transformer-based runway overrun / runoff risk prediction model is constructed. The Transformer-based runway overrun / runoff risk prediction model is used to extract runway overrun / runoff event features from flight data, and the aircraft's deviation from the runway centerline and remaining runway distance at a certain moment in the future are obtained from the historical parameter sequence. The method for realizing online risk identification includes: S2.1: For the flight data sequence of the aircraft [x1, x2, ..., x t ] T , using a 6-layer Encoder layer structure to transform the flight parameter sequence into a high-dimensional feature matrix Z with time-dependent and multi-parameter associations enc =[z1,z2,...,z t ] T , then input it into the 4-layer Decoder layer for decoding. Based on the global features and historical information provided by the Encoder layer, the runway status at future moments is gradually derived, and finally the deviation distance d from the runway centerline at the future time t+k is output. t+k and the remaining runway distance L t+k ,in Represents the flight data sequence record of the flight at time t, Represents the recorded value of the first flight data at time t, Z enc is the high-dimensional time series feature matrix, is the time series feature vector of the flight at time t, and b′ is the hidden layer dimension of the model; S2.2: Based on historical flight data, set the 95% percentile of the runway centerline deviation distance of all flights at time t+k as the maximum value of the centerline deviation distance threshold range. The 5% quantile is set as the minimum value of the threshold interval of the distance from the center line The 5% quantile of the remaining runway distance of all flights at time k is set as the remaining runway distance threshold μ t+k ;when or When L t+k < μ t+k , it indicates that the aircraft is at risk of running off the runway at time t+k; when the aircraft is at risk of running off the runway, the key parameters of the current warning time t and time t are recorded. Then continue to monitor the risk status at the next moment.

5. The method according to claim 1, wherein In S3, based on the hypothesis testing method, key parameters of the warning flight and the normal flight are obtained, and the most similar inducement is outputted based on the correspondence between the obtained key parameters and the risk factors. The method for completing the risk factor identification includes: When an early warning is issued at time t, the key parameters at time t Key parameters of other normal flights at time t Conduct hypothesis testing, where 1, 2, ..., N represents 1 to N different flights. represents the key parameters of the i-th flight at time t, and the key parameters of the warning flight and the normal flight that exceed the preset threshold are obtained, which are recorded as the inducement parameter R′={v g ,...,v q }, 1≤g≤q≤m; then calculate the incentive parameter R′={v g ,...,v q } and each corresponding relationship R j =f(v h ,...,v p ), Jaccard similarity of j∈[1,m]; finally output the risk factor R with the largest Jaccard similarity j , complete risk factor identification.

6. A risk warning and cause identification system for aircraft runway overrun / runoff events, the system being used to implement the method according to any one of claims 1 to 5, characterized in that: The system includes: an acquisition module, a construction module, an online risk identification module and a risk factor identification module; The acquisition module is used to obtain a set of key parameters related to aircraft takeoff and landing through flight quality monitoring standards, flight standard operating procedures, and flight quick reference manuals; and to obtain a set of risk factors that cause runway overrun / runoff events by establishing a risk factor fault tree; The construction module is used to construct the corresponding relationship between key parameters and risk factors based on the modified Delphi method; The online risk identification module is used to build a Transformer-based runway overrun / runoff risk prediction model, extract runway overrun / runoff event features from flight data using the Transformer-based runway overrun / runoff risk prediction model, and use historical flight data sequences to predict the aircraft's deviation from the runway centerline and remaining runway distance at a certain time in the future, thereby achieving online risk identification; The risk factor identification module is used to obtain key parameters that indicate that the difference between the warning flight and the normal flight exceeds a preset threshold based on a hypothesis testing method, and output the most similar inducement based on the correspondence between the obtained key parameters and the risk factors to complete the risk factor identification.

7. The system according to claim 6, characterized in that In the acquisition module, the key parameter PV={v 1 ,v 2 ,...,v d } and risk factors PR={R1,R2,...,R m Specifically include: Key parameter PV={v 1 ,v 2 ,...,v d }, d is the number of key parameters, v 1 The first key parameter; the key parameter PV includes two parts: PX and PW. PX is the key parameter set obtained through flight data. PX={x 1 ,x 2 ,...,x b }, b is the number of elements in PX, x 1 is the first element in PX; PW is a set of key parameters obtained through meteorological message data and airport pavement data, PW = {w 1 ,w 2 ,...,w c }, c is the number of elements in PW, and d=b+c, w 1 is the first element in PW; Risk factor PR={R1,R2,...,R m }, m is the number of risk factors, and R1 represents the first risk factor.

8. The system according to claim 6, wherein: The construction module includes: a scoring questionnaire unit, a matrix construction unit, and a comparison unit; The scoring questionnaire unit is used to form an interdisciplinary expert team and design and distribute anonymous scoring questionnaires to evaluate any risk factor R j , j∈[1,m] and any key parameter v i , the degree of association of i∈[1,d], and the subjective comments of experts; The matrix construction unit is used to obtain the scoring questionnaire of each expert for each risk factor R j , j∈[1,m], construct the parameter association matrix Where n is the number of experts, is the key parameter v of the rth expert pair i and risk factors R j The correlation degree score is then introduced; the evaluation index mean μ i,j , consensus index ECI i,j , interquartile range IQR i,j , all the anonymous rating data of experts, mean μ i,j , consensus index ECI i,j , interquartile range IQR i,j And subjective comments are provided to each expert again, and each expert needs to focus on μ i,j <3.5, ECI i,j <0.7, IQR i,j The scores of the above two and the indicators mentioned in the subjective comments are improved and updated, and the process is repeated 2 to 3 times; The comparison unit is used to calculate the final score data. i,j ≥3.5, the risk factor R j Affected by the key parameter v i The impact of is greater than the preset threshold, and the corresponding relationship between key parameters and risk factors is finally obtained, which is recorded as R j =f(v h ,...,v p ), 1≤h≤p≤m.

9. The system according to claim 6, wherein: The online risk identification module includes: a network construction unit and a risk monitoring unit; The network construction unit is used for the flight data sequence [x1, x2, ..., x t ] T , using a 6-layer Encoder layer structure to transform the flight parameter sequence into a high-dimensional feature matrix Z with time-dependent and multi-parameter associations enc =[z1,z2,...,z t ] T , which is then input into the 4-layer Decoder layer for decoding. Based on the global features and historical information provided by the Encoder layer, the runway status at future moments is gradually deduced, and the final output is the distance d from the runway centerline at the future time t+k. t+k and the remaining runway distance L t+k ,in Represents the flight data sequence record of the flight at time t, Represents the recorded value of the first flight data at time t, Z enc is a high-dimensional time series feature matrix, is the time series feature vector of the flight at time t, b′ is the hidden layer dimension of the model; The risk monitoring unit is used to set the 95% percentile of the runway centerline deviation distance of all flights at time t+k as the maximum value of the centerline deviation distance threshold range based on historical flight data The 5% quantile is set as the minimum value of the threshold interval of the distance from the center line The 5% quantile of the remaining runway distance of all flights at time k is set as the remaining runway distance threshold μ t+k ;when or When L t+k ≤μ t+k , it indicates that the aircraft is at risk of running off the runway at time t+k; when the aircraft is at risk of running off the runway, the key parameters of the current warning time t and time t are recorded. Then continue to monitor the risk status at the next moment.

10. The system according to claim 6, wherein: In the risk factor identification module, based on the hypothesis testing method, key parameters of the warning flight and the normal flight that exceed the preset threshold are obtained, and the most similar inducement is output based on the corresponding relationship between the obtained key parameters and the risk factors. The process of completing the risk factor identification includes: When an early warning is issued at time t, the key parameters at time t Key parameters of other normal flights at time t Conduct hypothesis testing, where 1, 2, ..., N represents 1 to N different flights. represents the key parameters of the i-th flight at time t, and the key parameters of the warning flight and the normal flight that exceed the preset threshold are obtained, which are recorded as the inducement parameter R′={v g ,...,v q }, 1≤g≤q≤m; then calculate the incentive parameter R′={v g ,...,v q } and each corresponding relationship R j =f(v h ,...,v p ), Jaccard similarity of j∈[1,m]; finally output the risk factor R with the largest Jaccard similarity j , complete risk factor identification.

Images