CN116010889A - Intelligent recognition method for abnormal flight state of aviation aircraft - Google Patents
Intelligent recognition method for abnormal flight state of aviation aircraft Download PDFInfo
- Publication number
- CN116010889A CN116010889A CN202211725975.1A CN202211725975A CN116010889A CN 116010889 A CN116010889 A CN 116010889A CN 202211725975 A CN202211725975 A CN 202211725975A CN 116010889 A CN116010889 A CN 116010889A
- Authority
- CN
- China
- Prior art keywords
- flight
- abnormal
- data
- delta
- ref
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 230000002159 abnormal effect Effects 0.000 title claims abstract description 78
- 238000000034 method Methods 0.000 title claims abstract description 38
- 238000003066 decision tree Methods 0.000 claims abstract description 37
- 238000013145 classification model Methods 0.000 claims abstract description 11
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 10
- 238000012549 training Methods 0.000 claims abstract description 10
- 238000005457 optimization Methods 0.000 claims abstract description 4
- 238000006073 displacement reaction Methods 0.000 claims description 30
- 230000005856 abnormality Effects 0.000 claims description 27
- 230000008859 change Effects 0.000 claims description 15
- 238000004364 calculation method Methods 0.000 claims description 9
- 238000013138 pruning Methods 0.000 claims description 9
- 239000013598 vector Substances 0.000 claims description 9
- 238000005259 measurement Methods 0.000 claims description 8
- 238000010276 construction Methods 0.000 claims description 6
- RZVHIXYEVGDQDX-UHFFFAOYSA-N 9,10-anthraquinone Chemical compound C1=CC=C2C(=O)C3=CC=CC=C3C(=O)C2=C1 RZVHIXYEVGDQDX-UHFFFAOYSA-N 0.000 claims description 4
- 230000007246 mechanism Effects 0.000 claims description 4
- 230000009471 action Effects 0.000 claims description 3
- 230000009194 climbing Effects 0.000 claims description 3
- 238000002790 cross-validation Methods 0.000 claims description 3
- 238000005192 partition Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 3
- 238000010845 search algorithm Methods 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 238000007405 data analysis Methods 0.000 abstract description 2
- 238000001514 detection method Methods 0.000 abstract description 2
- 230000036541 health Effects 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Traffic Control Systems (AREA)
Abstract
The invention relates to the technical field of aviation flight data analysis, in particular to an aviation aircraft abnormal state detection method. The method comprises the following steps: 1) Selecting historical standard flight data of the aviation aircraft by aviation aircraft flight experts, and marking the flight phases of the historical standard flight data; 2) Designing an optimization classification regression decision tree CART algorithm by using a marked data set D marked with the flight phases to construct a flight phase classification model M; 3) Inputting flight data of an abnormal flight state to be identified; 4) Performing flight phase division on the flight data set X by using the constructed classification model M; 5) Identifying abnormal flight states in the forest by using an optimized isolated forest algorithm IF; 6) And outputting the abnormal flight state time series data. According to the method, the historical flight data of the aircraft are utilized, a model of a normal flight state is established by a training algorithm, and the established model is utilized to identify an abnormal flight state.
Description
Technical Field
The invention relates to the technical field of aviation flight data analysis, in particular to an aviation aircraft abnormal state detection method.
Background
The flight state of an aircraft refers to the movement of the aircraft at a certain instant. The normal flight state of the aircraft refers to a state in which the aircraft flies according to a pre-booked course or a desired state under the control of the flight control system. The abnormal flight state of the aircraft refers to that the aircraft is greatly deviated from an expected flight state due to the fact that a certain component part of the aircraft breaks down or is influenced by the interference of external environment. For example, the deviation between the actual flight route and the set route is large, the deviation between the given attitude angle, speed and heading and the corresponding measured values is large in a long time, and the abnormal flight state of the aviation aircraft can bring flight risks, and if the abnormal flight state cannot be recognized and processed in time, serious flight safety accidents can be caused. Therefore, the identification of the abnormal flight state of the aviation aircraft has important significance for guaranteeing the flight safety of the aviation aircraft. The method has the advantages that the abnormal flight state of the aviation aircraft is identified, decision support is provided for real-time control decision and health management of the aviation aircraft, the aviation aircraft is convenient for ensuring personnel to search for the reasons of abnormal events in advance, the occurrence of flight accidents is avoided, and the flight safety is ensured.
Aiming at the problem of automatic recognition of the flight state, xie Chuan et al construct a flight state recognition method based on an expert knowledge base and a knowledge inference engine, and the recognition method is limited by the expert knowledge base. Meng Guanglei et al in the literature constructed a dynamic bayesian network model for maneuver identification with maneuver correspondence to the flight parameters of the flight simulation training as the subject of investigation, and identified the flight status by means of intelligent methods based on recursive reasoning of the network model. Zhou Chaodeng A flight state identification method based on an improved dynamic time warping algorithm is provided mainly aiming at the characteristics of strong randomness and different lengths of tactical maneuver data, different contribution degrees are set by using different characteristic parameters of the flight state, the frame matching distance between flight parameter data and standard template data is calculated, and the flight state is identified according to the distance. All documents realize automatic identification of certain flight states under a specific application scene, and are not suitable for scenes of identifying abnormal flight states of the aviation aircraft.
Therefore, there is a need for an intelligent recognition method for abnormal flight states of an aircraft, which can solve the above technical problems.
Disclosure of Invention
The invention aims to overcome the defects of the prior art, and provides an intelligent identification method for abnormal flight states of an aircraft, which defines multiple characteristic parameters of abnormal gestures, abnormal speeds, abnormal tracks and abnormal control of an actuating mechanism of the aircraft in different flight stages, introduces a grid search method to optimize parameters of a classification decision tree, realizes accurate classification of the flight stages, builds an isolated forest by adding abnormal influence of the characteristic parameters, improves the accuracy rate of identifying the abnormal flight states, and has important application value in the aspect of aircraft flight monitoring and has important significance in improving the intelligent monitoring capability of flight safety.
The invention relates to an intelligent recognition method for abnormal flight states of an aviation aircraft, which is characterized by comprising the following steps of:
1) Selecting historical standard flight data of the aviation aircraft by aviation aircraft flight experts, and marking the flight phases of the historical standard flight data;
the method comprises the following steps: the flight expert selects historical standard flight data, marks 6 basic flight stages of take-off, climb, fly-flat, turn, descent and landing on the flight data, and marks the marked flight data set as D= { D 1 ,D 2 ,......,D 6 }；
2) Designing an optimized classification regression decision tree (Classification And Regression Tree, CART) algorithm by using the marked data set D marked with the flight phases to construct a flight phase classification model M;
the method comprises the following steps:
step 1: for standard action template data d= { D 1 ,D 2 ,......,D 6 -calculating a standard flight state D r N-element characteristic parameter vector of the u-th time point
The altitude change rate delta h, the pitch angle change rate delta alpha, the roll angle change rate delta beta, the yaw angle change rate delta gamma and the course angle change rate delta between the two, and calculating a coefficient formula (1) of the data set D
In p k The proportion of the kth category in the data set D in the D is represented in a simpler way than the way of calculating the logarithm by adopting the information gain ratio. For data set D, calculate the coefficient of Kennel formula (2) of characteristic parameter c
Where |D| represents the number of all time-series vectors of the data set D, E represents the total number of the valued categories of the characteristic parameter c on the data set D, |D e I represents the total number of vectors with e-value for all features c in the data set D, gini (D e ) The coefficient of kunning is calculated according to formula (1).
Step 2: the data set D is divided into two parts according to the characteristic parameter C with the minimum coefficient value of the foundation to obtain left and right nodes, and the left and right nodes are marked as D left And D right ；
Step 3: two conditions for stopping constructing the decision tree are provided, namely, the relation between the threshold value and the current coefficient of the foundation is judged, if the current coefficient of the foundation is smaller than the threshold value, the current node stops constructing the decision tree subtree, otherwise, the construction is continued; checking whether classification features capable of continuously decomposing the feature parameter set D exist or not, and if not, stopping constructing a decision tree;
step 4: left and right child node D of D node left And D right Recursively executing steps 1 to 4 until exiting from meeting the condition of step 3, returning to decision makingTraining a model M by a tree;
step 5: to avoid the over fitting problem of the decision tree model, pruning the decision tree. Pruning, deleting non-leaf nodes { T ] 1 ,T 2 ,T 3 ,......,T n Alpha with minimum surface error rate gain value i Corresponding to non-leaf node T i Left and right child nodes of a), alpha i The calculation formula (3) of (2) is
Wherein R (i) represents the error generated after the leaf node replaces the ith non-leaf node, and the calculation formula is that
R(i)＝r(i)p(i), (4)
Where R (i) represents the error rate of node i, p (i) represents the percentage of the number of samples on node i to the number of samples in the entire training set, R (T) i ) Representing the subtree T when node i is not clipped i The sum of the errors of all leaf nodes, i.e
Replacement of non-leaf node T with leaf node of subtree i Repeating the process until no non-leaf nodes can be replaced, pruning is completed, and returning to the decision Tree;
step 6: setting target loss function (accuracy, precision, recall or F 1 Score), placing all the super parameters of the decision tree in a network, adjusting the parameters of the decision tree by using a grid search algorithm to ensure that the classification of the decision tree model M is optimal, and simultaneously, adopting a data cross-validation mode to improve the classification performance of the decision tree model M in order to avoid the influence caused by the division of a training set and a testing set;
3) Inputting flight data of an abnormal flight state to be identified;
the method comprises the following steps: the aircraft flight data set for the input t time points is noted as x= { X 1 ,X 2 ,……,X t }
4) Performing flight phase division on the flight data set X by using the constructed classification model M;
the method comprises the following steps: flight data set x= { X for t time points using classification model M 1 ,X 2 ,…,X t Automatically dividing the flight phase and marking the flight phase of each time period;
5) Identifying abnormal flight states in the system by utilizing an optimization Isolated Forest (IF) algorithm;
the method comprises the following steps:
step 1: calculating abnormal characteristic parameter values of all flight phases according to the table 1 for the X marked with the flight phases, wherein the characteristic parameter set of each flight phase is recorded as S;
taking 6 most basic flight phases of take-off, climbing, flat flight, turning, descending and landing as an example, each flight phase can be represented by characteristic parameters such as flight speed, height, attitude angle, heading angle and the like, and simultaneously, each flight phase corresponds to different flight control laws, and main characteristic parameters of the flight phase are established;
wherein, abnormal characteristic parameters are classified as: the calculation methods of the attitude abnormality, the speed abnormality, the track abnormality and the actuating mechanism control abnormality are as follows:
(1) Abnormal characteristic parameters of posture
The abnormal flying attitude of the aviation aircraft refers to that the measured values of attitude angles (pitch angle theta, yaw angle phi and roll angle phi) deviate from given values (given pitch angle theta_ref, given yaw angle phi_ref and given roll angle phi_ref) for a long time, and the deviation mean values (E (delta theta), E (delta phi) and E (delta phi)) of the attitude angles at each N moment points are calculated as the abnormal attitude characteristic parameters, namely
Wherein θ i 、ψ i 、φ i The measurement values of the pitch angle, yaw angle, and roll angle, θ_ref, at the i (i=1, 2,., N) th time point are respectively represented i 、ψ_ref i 、φ_ref i The given values of the i (i=1, 2,.., N) th moment point pitch angle, yaw angle, and roll angle are respectively represented.
(2) Speed anomaly characteristic parameter
The abnormal speed of the aviation aircraft refers to that the measured value V of the flying speed deviates from a given speed value V_ref for a long time, and the deviation average E (delta V) of the speeds at the N time points is calculated as the characteristic parameter of the abnormal speed, namely
Wherein V is i 、V_ref i The i (i=1, 2,., N) th point in time aircraft flight speed measurement and setpoint are respectively indicated.
(3) Characteristic parameters of track anomalies
The aircraft track abnormality refers to the deviation of an actual flight path from a set path, the path is composed of a plurality of waypoints, and the average value E (delta d) of the distance deviation of the waypoints at N time points is calculated as the characteristic parameter of the track abnormality, namely
Wherein the measured values and given values of the waypoint location longitude, latitude, and altitude at the i (i=1, 2., N) time are denoted as P, respectively i (x i ,y i ,h i ) And P_ref i (x_ref i ,y_ref i ,h_ref i )。
Calculating the course angle deviation mean value E (delta PSI) of N moment points as the characteristic parameters of the track abnormality, namely
Wherein PSI i 、PSI_ref i The measured value and the given value of the heading angle at the i (i=1, 2,., N) time are respectively represented.
(4) Abnormal control of an actuator
The aircraft actuator control abnormality refers to the deviation of actuator displacement (aileron displacement dx, rudder displacement dty, rudder lifting displacement dtz, flap aileron displacement dtjy, front wheel displacement dtw) from actuator instructions (aileron instruction dtxc, rudder instruction dtyc, elevator instruction dtzc, flap aileron instruction dtjyc, front wheel instruction dtwc), and the deviation mean value (E (delta dtx), E (delta dty), E (delta dtz), E (delta dtjy) and E (delta dtw)) of each actuator instruction at each N time points is calculated as the actuator control abnormality characteristic parameters, namely
Wherein, dtx i 、dty i 、dtz i 、dtjy i 、dtw i The i (i=1, 2,., N) th point in time aileron displacement, rudder lifting displacement, flap aileron displacement, and front wheel displacement measurements, dtxc, are shown, respectively i 、dtyc i 、dtzc i 、dtjyc i 、dtwc i Respectively representing the given values of the i (i=1, 2,., N) th point-in-time aileron command, rudder command, elevator command, flap aileron command, and front wheel command;
step 2: and selecting the characteristic parameter with the largest abnormal influence from the abnormal characteristic parameter set S, and selecting the data point closest to the mean value as the root node of the tree.
Step 3: and setting the abnormal characteristic parameter value smaller than the current partition point as a Left node Left of the binary tree and larger than or equal to a Right node Right of the binary tree.
Step 4: nodes Left and Right recursively construct the tree by repeating steps 9 and 10 until only one data cannot continue to be constructed, or the tree has reached a defined height, stopping construction.
Step 5: calculating an anomaly score s (x, n) of the sample x:
wherein: h (x) is the path length of the tree where sample x is located, c (n) is the path average length of each tree [197] ：
Wherein: h (n-1) =ln (n-1) +ζ (ζ is euler constant, and generally takes the value of 0.58). Data samples x with anomaly scores s (x, n) close to 1 are outlier data points.
6) And outputting the abnormal flight state time series data.
The method comprises the following steps: outputting an aviation aircraft flight data set X= { X 1 ,X 2 ,……,X t Abnormal time series of flight conditions in }.
The intelligent recognition method for the abnormal flight state of the aviation aircraft can recognize the abnormal flight state of the aviation aircraft, provides decision support for real-time control decision and health management of the aviation aircraft, facilitates the guarantee personnel of the aviation aircraft to find the reasons of abnormal events, avoids the occurrence of flight accidents and ensures the flight safety.
Drawings
FIG. 1 is a flow chart of a method for intelligently identifying abnormal flight conditions of an aircraft.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Example 1
Referring to fig. 1, an intelligent recognition method for abnormal flight states of an aircraft in this embodiment includes the following steps:
1) Selecting historical standard flight data of the aviation aircraft by aviation aircraft flight experts, and marking the flight phases of the historical standard flight data;
the method comprises the following steps: the flight expert selects historical standard flight data, marks 6 basic flight stages of take-off, climb, fly-flat, turn, descent and landing on the flight data, and marks the marked flight data set as D= { D 1 ,D 2 ,......,D 6 }；
2) Designing an optimization classification regression decision tree CART algorithm by using a marked data set D marked with the flight phases to construct a flight phase classification model M;
the method comprises the following steps:
step 1: for standard action template data d= { D 1 ,D 2 ,......,D 6 -calculating a standard flight state D r N-element characteristic parameter vector of the u-th time point
The altitude change rate delta h, the pitch angle change rate delta alpha, the roll angle change rate delta beta, the yaw angle change rate delta gamma and the course angle change rate delta between the two, and calculating a coefficient formula (1) of the data set D
In p k The proportion of the kth category in the data set D in the D is represented in a simpler way than the way of calculating the logarithm by adopting the information gain ratio. For data set D, calculate the coefficient of Kennel formula (2) of characteristic parameter c
Where |D| represents the number of all time-series vectors of the data set D, E represents the total number of the valued categories of the characteristic parameter c on the data set D, |D e I represents the total number of vectors with e-value for all features c in the data set D, gini (D e ) The coefficient of kunning is calculated according to formula (1).
Step 2: the data set D is divided into two parts according to the characteristic parameter C with the minimum coefficient value of the foundation to obtain left and right sections of the data set DPoint, designated as D left And D right ；
Step 3: two conditions for stopping constructing the decision tree are provided, namely, the relation between the threshold value and the current coefficient of the foundation is judged, if the current coefficient of the foundation is smaller than the threshold value, the current node stops constructing the decision tree subtree, otherwise, the construction is continued; checking whether classification features capable of continuously decomposing the feature parameter set D exist or not, and if not, stopping constructing a decision tree;
step 4: left and right child node D of D node left And D right Recursively executing the steps 1 to 4 until exiting from the condition meeting the step 3, and returning to the decision Tree training model Tree;
step 5: to avoid the over fitting problem of the decision tree model, pruning the decision tree. Pruning, deleting non-leaf nodes { T ] 1 ,T 2 ,T 3 ,......,T n Alpha with minimum surface error rate gain value i Corresponding to non-leaf node T i Left and right child nodes of a), alpha i The calculation formula (3) of (2) is
Wherein R (i) represents the error generated after the leaf node replaces the ith non-leaf node, and the calculation formula is that
R(i)＝r(i)p(i), (4)
Where R (i) represents the error rate of node i, p (i) represents the percentage of the number of samples on node i to the number of samples in the entire training set, R (T) i ) Representing the subtree T when node i is not clipped i The sum of the errors of all leaf nodes, i.e
Replacement of non-leaf node T with leaf node of subtree i Repeating the process until no non-leaf nodes can be replaced, pruning is completed, and returning to the decision Tree;
step 6: setting target loss function (accuracy, precision, recall or F 1 Score), placing all the super parameters of the decision tree in a network, adjusting the parameters of the decision tree by using a grid search algorithm to ensure that the classification of the decision tree model M is optimal, and simultaneously, adopting a data cross-validation mode to improve the classification performance of the decision tree model M in order to avoid the influence caused by the division of a training set and a testing set;
3) Inputting flight data of an abnormal flight state to be identified;
the method comprises the following steps: the aircraft flight data set for the input t time points is noted as x= { X 1 ,X 2 ,……,X t }
4) Performing flight phase division on the flight data set X by using the constructed classification model M;
the method comprises the following steps: flight data set x= { X for t time points using classification model M 1 ,X 2 ,…,X t Automatically dividing the flight phase and marking the flight phase of each time period;
5) Identifying abnormal flight states in the forest by using an optimized isolated forest algorithm IF;
the method comprises the following steps:
step 1: calculating abnormal characteristic parameter values of all flight phases according to the table 1 for the X marked with the flight phases, wherein the characteristic parameter set of each flight phase is recorded as S;
taking 6 most basic flight phases of take-off, climbing, flat flight, turning, descending and landing as an example, each flight phase can be represented by characteristic parameters such as flight speed, altitude, attitude angle, heading angle and the like, and meanwhile, each flight phase corresponds to different flight control laws, and main characteristic parameters of the flight phase are established, as shown in table 1.
Table 1 table of flight phase characteristics parameters of an aircraft
The abnormal characteristic parameters in the table are classified as follows: the calculation methods of the attitude abnormality, the speed abnormality, the track abnormality and the actuating mechanism control abnormality are as follows:
(1) Abnormal characteristic parameters of posture
The abnormal flying attitude of the aviation aircraft refers to that the measured values of attitude angles (pitch angle theta, yaw angle phi and roll angle phi) deviate from given values (given pitch angle theta_ref, given yaw angle phi_ref and given roll angle phi_ref) for a long time, and the deviation mean values (E (delta theta), E (delta phi) and E (delta phi)) of the attitude angles at each N moment points are calculated as the abnormal attitude characteristic parameters, namely
Wherein θ i 、ψ i 、φ i The measurement values of the pitch angle, yaw angle, and roll angle, θ_ref, at the i (i=1, 2,., N) th time point are respectively represented i 、ψ_ref i 、φ_ref i The given values of the i (i=1, 2,.., N) th moment point pitch angle, yaw angle, and roll angle are respectively represented.
(2) Speed anomaly characteristic parameter
The abnormal speed of the aviation aircraft refers to that the measured value V of the flying speed deviates from a given speed value V_ref for a long time, and the deviation average E (delta V) of the speeds at the N time points is calculated as the characteristic parameter of the abnormal speed, namely
Wherein V is i 、V_ref i The i (i=1, 2,., N) th point in time aircraft flight speed measurement and setpoint are respectively indicated.
(3) Characteristic parameters of track anomalies
The aircraft track abnormality refers to the deviation of an actual flight path from a set path, the path is composed of a plurality of waypoints, and the average value E (delta d) of the distance deviation of the waypoints at N time points is calculated as the characteristic parameter of the track abnormality, namely
Wherein the measured values and given values of the waypoint location longitude, latitude, and altitude at the i (i=1, 2., N) time are denoted as P, respectively i (x i ,y i ,h i ) And P_ref i (x_ref i ,y_ref i ,h_ref i )。
Calculating the course angle deviation mean value E (delta PSI) of N moment points as the characteristic parameters of the track abnormality, namely
Wherein PSI i 、PSI_ref i The measured value and the given value of the heading angle at the i (i=1, 2,., N) time are respectively represented.
(4) Abnormal control of an actuator
The aircraft actuator control abnormality refers to the deviation of actuator displacement (aileron displacement dx, rudder displacement dty, rudder lifting displacement dtz, flap aileron displacement dtjy, front wheel displacement dtw) from actuator instructions (aileron instruction dtxc, rudder instruction dtyc, elevator instruction dtzc, flap aileron instruction dtjyc, front wheel instruction dtwc), and the deviation mean value (E (delta dtx), E (delta dty), E (delta dtz), E (delta dtjy) and E (delta dtw)) of each actuator instruction at each N time points is calculated as the actuator control abnormality characteristic parameters, namely
Wherein, dtx i 、dty i 、dtz i 、dtjy i 、dtw i The i (i=1, 2,., N) th point in time is the measured values of aileron displacement, rudder lifting displacement, flap aileron displacement, and front wheel displacement, respectively，dtxc i 、dtyc i 、dtzc i 、dtjyc i 、dtwc i Respectively representing the given values of the i (i=1, 2,., N) th point-in-time aileron command, rudder command, elevator command, flap aileron command, and front wheel command;
step 2: and selecting the characteristic parameter with the largest abnormal influence from the abnormal characteristic parameter set S, and selecting the data point closest to the mean value as the root node of the tree.
Step 3: and setting the abnormal characteristic parameter value smaller than the current partition point as a Left node Left of the binary tree and larger than or equal to a Right node Right of the binary tree.
Step 4: the nodes Left and Right repeat the steps 2 and 3 recursively constructing the tree until only one data cannot continue to be constructed, or the tree has reached a defined height, stopping construction.
Step 5: calculating an anomaly score s (x, n) of the sample x:
wherein: h (x) is the path length of the tree where sample x is located, c (n) is the path average length of each tree:
wherein: h (n-1) =ln (n-1) +ζ (ζ is euler constant, and generally takes the value of 0.58). Data samples x with anomaly scores s (x, n) close to 1 are outlier data points.
6) And outputting the abnormal flight state time series data.
The method comprises the following steps: outputting an aviation aircraft flight data set X= { X 1 ,X 2 ,……,X t Abnormal time series of flight conditions in }.
Although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (3)
1. An intelligent recognition method for abnormal flight state of an aviation aircraft is characterized by comprising the following steps:
1) Selecting historical standard flight data of the aviation aircraft by aviation aircraft flight experts, and marking the flight phases of the historical standard flight data;
the method comprises the following steps: the flight expert selects historical standard flight data, marks 6 basic flight stages of take-off, climb, fly-flat, turn, descent and landing on the flight data, and marks the marked flight data set as D= { D 1 ,D 2 ,......,D 6 }；
2) Designing an optimization classification regression decision tree algorithm by using the marked data set D marked with the flight phases to construct a flight phase classification model M;
3) Inputting flight data of an abnormal flight state to be identified;
the method comprises the following steps: the aircraft flight data set for the input t time points is noted as x= { X 1 ,X 2 ,……,X t }；
4) Performing flight phase division on the flight data set X by using the constructed classification model M;
the method comprises the following steps: flight data set x= { X for t time points using classification model M 1 ,X 2 ,…,X t Automatically dividing the flight phase and marking the flight phase of each time period;
5) Identifying abnormal flight states in the forest by utilizing an optimized isolated forest algorithm;
6) Outputting abnormal flight state time sequence data;
the method comprises the following steps: outputting an aviation aircraft flight data set X= { X 1 ,X 2 ,……,X t Abnormal time series of flight conditions in }.
2. The intelligent recognition method of abnormal flight state of aviation aircraft according to claim 1, wherein the step 2) specifically comprises the following steps:
step 1: for standard action template data d= { D 1 ,D 2 ,......,D 6 -calculating a standard flight state D r N-element characteristic parameter vector of the u-th time point
The altitude change rate delta h, the pitch angle change rate delta alpha, the roll angle change rate delta beta, the yaw angle change rate delta gamma and the course angle change rate delta between the two, and calculating a coefficient formula (1) of the data set D
In p k The proportion of the kth category in the data set D in the D is represented in a simpler way than the way of calculating the logarithm by adopting the information gain ratio. For data set D, calculate the coefficient of Kennel formula (2) of characteristic parameter c
Where |D| represents the number of all time-series vectors of the data set D, E represents the total number of the valued categories of the characteristic parameter c on the data set D, |D e I represents the total number of vectors with e-value for all features c in the data set D, gini (D e ) Calculating a coefficient of Kernine according to formula (1);
step 2: the data set D is divided into two parts according to the characteristic parameter C with the minimum coefficient value of the foundation to obtain left and right nodes, and the left and right nodes are marked as D left And D right ；
Step 3: two conditions for stopping constructing the decision tree are provided, namely, the relation between the threshold value and the current coefficient of the foundation is judged, if the current coefficient of the foundation is smaller than the threshold value, the current node stops constructing the decision tree subtree, otherwise, the construction is continued; checking whether classification features capable of continuously decomposing the feature parameter set D exist or not, and if not, stopping constructing a decision tree;
step (a)4: left and right child node D of D node left And D right Recursively executing the steps 1 to 4 until exiting from the condition meeting the step 3, and returning to the decision tree training model M;
step 5: pruning the decision tree, so-called pruning, to remove non-leaf nodes { T }, in order to avoid overfitting problems of the decision tree model 1 ,T 2 ,T 3 ,......,T n Alpha with minimum surface error rate gain value i Corresponding to non-leaf node T i Left and right child nodes of a), alpha i The calculation formula (3) of (2) is
Wherein R (i) represents the error generated after the leaf node replaces the ith non-leaf node, and the calculation formula is that
R(i)＝r(i)p(i), (4)
Where R (i) represents the error rate of node i, p (i) represents the percentage of the number of samples on node i to the number of samples in the entire training set, R (T) i ) Representing the subtree T when node i is not clipped i The sum of the errors of all leaf nodes, i.e
Replacement of non-leaf node T with leaf node of subtree i Repeating the process until no non-leaf nodes can be replaced, pruning is completed, and returning to the decision tree M;
step 6: setting a target loss function, placing all the super parameters of the decision tree in a network, and adjusting the parameters of the decision tree by using a grid search algorithm to ensure that the classification of the decision tree model M is optimal. Meanwhile, in order to avoid the influence caused by the division of the training set and the testing set, a data cross-validation mode is adopted to improve the classification performance of the decision tree model M.
3. The intelligent recognition method of abnormal flight state of aviation aircraft according to claim 1, wherein the step 5) specifically comprises the following steps:
step 1: calculating abnormal characteristic parameter values of all flight phases according to the table 1 for the X marked with the flight phases, wherein the characteristic parameter set of each flight phase is recorded as S;
taking 6 most basic flight phases of take-off, climbing, flat flight, turning, descending and landing as an example, each flight phase can be represented by characteristic parameters such as flight speed, height, attitude angle, heading angle and the like, and each flight phase corresponds to different flight control laws to establish abnormal characteristic parameters of the flight phase;
wherein, abnormal characteristic parameters are classified as: the calculation methods of the attitude abnormality, the speed abnormality, the track abnormality and the actuating mechanism control abnormality are as follows:
(1) Abnormal characteristic parameters of posture
The abnormal flying attitude of the aviation aircraft refers to that the measured values of attitude angles (pitch angle theta, yaw angle phi and roll angle phi) deviate from given values (given pitch angle theta_ref, given yaw angle phi_ref and given roll angle phi_ref) for a long time, and the deviation mean values (E (delta theta), E (delta phi) and E (delta phi)) of the attitude angles at each N moment points are calculated as the abnormal attitude characteristic parameters, namely
Wherein θ i 、ψ i 、φ i The measurement values of the pitch angle, yaw angle, and roll angle, θ_ref, at the i (i=1, 2,., N) th time point are respectively represented i 、ψ_ref i 、φ_ref i Respectively representing the given values of the i (i=1, 2,., N) time point pitch angle, yaw angle, and roll angle;
(2) Speed anomaly characteristic parameter
The abnormal speed of the aviation aircraft refers to that the measured value V of the flying speed deviates from a given speed value V_ref for a long time, and the deviation average E (delta V) of the speeds at the N time points is calculated as the characteristic parameter of the abnormal speed, namely
Wherein V is i 、V_ref i Respectively representing i (i=1, 2,., N) th point in time aircraft speed measurements and given values;
(3) Characteristic parameters of track anomalies
The aircraft track abnormality refers to the deviation of an actual flight path from a set path, the path is composed of a plurality of waypoints, and the average value E (delta d) of the distance deviation of the waypoints at N time points is calculated as the characteristic parameter of the track abnormality, namely
Wherein the measured values and given values of the waypoint location longitude, latitude, and altitude at the i (i=1, 2., N) time are denoted as P, respectively i (x i ,y i ,h i ) And P_ref i (x_ref i ,y_ref i ,h_ref i )；
Calculating the course angle deviation mean value E (delta PSI) of N moment points as the characteristic parameters of the track abnormality, namely
Wherein PSI i 、PSI_ref i Representing a measured value and a given value of the heading angle at the i (i=1, 2,., n.), respectively;
(4) Abnormal control of an actuator
The aircraft actuator control abnormality refers to the deviation of actuator displacement (aileron displacement dx, rudder displacement dty, rudder lifting displacement dtz, flap aileron displacement dtjy, front wheel displacement dtw) from actuator instructions (aileron instruction dtxc, rudder instruction dtyc, elevator instruction dtzc, flap aileron instruction dtjyc, front wheel instruction dtwc), and the deviation mean value (E (delta dtx), E (delta dty), E (delta dtz), E (delta dtjy) and E (delta dtw)) of each actuator instruction at each N time points is calculated as the actuator control abnormality characteristic parameters, namely
Wherein, dtx i 、dty i 、dtz i 、dtjy i 、dtw i The i (i=1, 2,., N) th point in time aileron displacement, rudder lifting displacement, flap aileron displacement, and front wheel displacement measurements, dtxc, are shown, respectively i 、dtyc i 、dtzc i 、dtjyc i 、dtwc i Respectively representing the given values of the i (i=1, 2,., N) th point-in-time aileron command, rudder command, elevator command, flap aileron command, and front wheel command;
step 2: selecting the characteristic parameter with the largest abnormal influence from the abnormal characteristic parameter set S according to the experience of a flight expert, and selecting the data point closest to the mean value as the root node of the tree;
step 3: setting the abnormal characteristic parameter value smaller than the Left node Left of the current partition point as a binary tree and larger than or equal to the Right node Right of the binary tree;
step 4: the nodes Left and Right repeat the steps 2 and 3 to recursively construct the tree until only one data cannot be continuously constructed or the tree reaches a limited height, and the construction is stopped;
step 5: calculating an anomaly score s (x, n) of the sample x:
wherein: h (x) is the path length of the tree where sample x is located, c (n) is the path average length of each tree:
wherein: h (n-1) =ln (n-1) +ζ (ζ is euler constant, and takes a value of 0.58), and data sample x having an anomaly score s (x, n) close to 1 is an anomaly data point.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211725975.1A CN116010889A (en) | 2022-12-30 | 2022-12-30 | Intelligent recognition method for abnormal flight state of aviation aircraft |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211725975.1A CN116010889A (en) | 2022-12-30 | 2022-12-30 | Intelligent recognition method for abnormal flight state of aviation aircraft |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116010889A true CN116010889A (en) | 2023-04-25 |
Family
ID=86036839
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211725975.1A Pending CN116010889A (en) | 2022-12-30 | 2022-12-30 | Intelligent recognition method for abnormal flight state of aviation aircraft |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116010889A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116700070A (en) * | 2023-05-17 | 2023-09-05 | 北京锐士装备科技有限公司 | Safety supervision method and system for flight state of unmanned aerial vehicle |
| CN116861300A (en) * | 2023-09-01 | 2023-10-10 | 中国人民解放军海军航空大学 | Automatic auxiliary labeling method and device for complex maneuvering data set of military machine type |
| CN117853827A (en) * | 2024-03-07 | 2024-04-09 | 安徽省大气探测技术保障中心 | Sampling pump working state operation monitoring system and method for atmospheric chamber gas monitoring |
| CN117972454A (en) * | 2023-12-26 | 2024-05-03 | 中国人民解放军海军航空大学 | VMSD-TICC-based flight phase division method and division terminal |
-
2022
- 2022-12-30 CN CN202211725975.1A patent/CN116010889A/en active Pending
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116700070A (en) * | 2023-05-17 | 2023-09-05 | 北京锐士装备科技有限公司 | Safety supervision method and system for flight state of unmanned aerial vehicle |
| CN116700070B (en) * | 2023-05-17 | 2024-01-30 | 北京锐士装备科技有限公司 | Safety supervision method and system for flight state of unmanned aerial vehicle |
| CN116861300A (en) * | 2023-09-01 | 2023-10-10 | 中国人民解放军海军航空大学 | Automatic auxiliary labeling method and device for complex maneuvering data set of military machine type |
| CN116861300B (en) * | 2023-09-01 | 2024-01-09 | 中国人民解放军海军航空大学 | Automatic auxiliary labeling method and device for complex maneuvering data set of military machine type |
| CN117972454A (en) * | 2023-12-26 | 2024-05-03 | 中国人民解放军海军航空大学 | VMSD-TICC-based flight phase division method and division terminal |
| CN117853827A (en) * | 2024-03-07 | 2024-04-09 | 安徽省大气探测技术保障中心 | Sampling pump working state operation monitoring system and method for atmospheric chamber gas monitoring |
| CN117853827B (en) * | 2024-03-07 | 2024-05-14 | 安徽省大气探测技术保障中心 | Sampling pump working state operation monitoring system and method for atmospheric chamber gas monitoring |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116010889A (en) | Intelligent recognition method for abnormal flight state of aviation aircraft | |
| CN108594788B (en) | Airplane actuator fault detection and diagnosis method based on deep random forest algorithm | |
| US20240071236A1 (en) | Unsupervised anomaly detection for autonomous vehicles | |
| CN109976378B (en) | Trajectory control method for unmanned aerial vehicle perching and landing maneuver under wind disturbance | |
| CN112198890B (en) | Aircraft attitude control method, system and device based on reinforcement learning | |
| CN112859898B (en) | Aircraft trajectory prediction method based on two-channel bidirectional neural network | |
| Zhang et al. | Aircraft damage identification and classification for database-driven online flight-envelope prediction | |
| Kong et al. | Bayesian deep learning for aircraft hard landing safety assessment | |
| CN106980262A (en) | Self-adaptive flight device robust control method based on Kernel recursive least square algorithm | |
| CN117648866A (en) | Technical method for landing boundary sliding at final approach stage of civil aircraft test flight | |
| CN112947073A (en) | Model prediction control-based shipboard aircraft sliding track tracking method | |
| CN110779526B (en) | Path planning method, device and storage medium | |
| CN111008549B (en) | UUV platform DVL signal distortion reconstruction method based on sample entropy and IFOA-GRNN | |
| Juang et al. | Intelligent automatic landing system using fuzzy neural networks and genetic algorithm | |
| Dutta et al. | Machine-learning based rotor fault diagnosis in a multicopter with strain data | |
| Jianhong et al. | Synthesis cascade estimation for aircraft system identification | |
| Leško et al. | Flight phase classification for small unmanned aerial vehicles | |
| Ochoa et al. | Multicopter failure diagnosis through supervised learning and statistical trajectory prediction | |
| CN113110558B (en) | Hybrid propulsion unmanned aerial vehicle demand power prediction method | |
| CN115293225A (en) | Pilot flat drift ejector rod cause analysis method and device | |
| CN114023111A (en) | Air traffic situation complexity evaluation method oriented to autonomous operation of aircraft | |
| CN117786375B (en) | Method for evaluating control quality of guiding following flight in instrument flight mode | |
| CN112232567A (en) | Method for predicting influence probability of unknown intention flight path on air route | |
| Nobahari et al. | Wind compensation in trajectory tracking of a fixed wing uav using a nonlinear model predictive controller based on the particle swarm optimization | |
| CN114118537B (en) | Airspace flight carbon emission combined prediction method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |