Variable geometry aircraft fault detection model generation method and fault detection method

Abstract

The application provides a morphing aircraft fault detection model generation method and a fault detection method. The morphing aircraft fault detection model generation method comprises the following steps: obtaining a state control feedback matrix corresponding to each wingspan change rate based on an initial value of each wingspan change rate, an initial value of an aircraft state variable, and a preset flight linear change parameter model of the morphing aircraft; generating flight simulation data based on to-be-simulated data, wherein the to-be-simulated data comprises the initial value of the aircraft state variable, the wingspan change rate, and the state control feedback matrix, and each set of flight simulation data comprises flight state parameters at multiple time points; generating sample data based on the flight simulation data, wherein the sample data comprises the flight simulation data and a fault label corresponding to the flight simulation data; and obtaining the morphing aircraft fault detection model by training based on multiple sets of sample data. The application can effectively detect faults of the morphing aircraft.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of aircraft fault detection, and in particular to a morphing aircraft fault detection model generation method and a fault detection method. BACKGROUND

[0002] Morphing aircrafts have attracted extensive attention because they can change their shape structure (extensible wings) in real time according to different flight stages and task conditions to adapt to today's more complex flight environment. This adaptability enhances the performance of the aircraft, but also brings great challenges to the design of the control system and increases the risk of aircraft failure. However, there is no effective method for detecting the failure of the morphing aircraft in the prior art. SUMMARY

[0003] The present application provides a morphing aircraft fault detection model generation method and a fault detection method to solve the defect that there is no effective detection of the failure of the morphing aircraft in the prior art, and to achieve effective detection of the failure of the morphing aircraft.

[0004] The present application provides a morphing aircraft fault detection model generation method, comprising:

[0005] Based on the initial values of each wing span change rate, the aircraft state variable and the preset flight linear change parameter model of the morphing aircraft, a state control feedback matrix corresponding to each wing span change rate is obtained;

[0006] Flight simulation data is generated based on the to-be-simulated data, the to-be-simulated data including the initial values of the aircraft state variables, the wing span change rates and the state control feedback matrix, and each set of flight simulation data including flight state parameters at multiple time points;

[0007] Sample data is generated based on the flight simulation data, the sample data including the flight simulation data and a fault label corresponding to the flight simulation data, the fault label being a fault when the wing span change rate corresponding to the state control feedback matrix included in the flight simulation data is inconsistent with the wing span change rate included in the flight simulation data, and the fault label being no fault when the wing span change rate corresponding to the state control feedback matrix included in the flight simulation data is consistent with the wing span change rate included in the flight simulation data;

[0008] A morphing aircraft fault detection model is obtained by training based on multiple sets of sample data.

[0009] According to the morphing aircraft fault detection model generation method provided by the application, the flight state parameters include flight speed, angle of attack, pitch angle, pitch angle speed and flight height, and the aircraft state variables include flight speed variation, angle of attack variation, pitch angle variation, pitch angle speed variation and flight height variation.

[0010] According to the morphing aircraft fault detection model generation method provided by the application, the flight linear variation parameter model is:

[0011] ;

[0012] Wherein, , is the wing span variation rate, is the flight speed, is the state control feedback matrix, is the aircraft state variable, represents interference in the flight process of the aircraft, 、 、 、 is a preset fitting matrix.

[0013] According to the morphing aircraft fault detection model generation method provided by the application, the fitting matrix in the flight linear variation parameter model is obtained by linearizing the nonlinear dynamic equation of the deformable aircraft at each working point in the flight envelope and then fitting and simplifying.

[0014] According to the morphing aircraft fault detection model generation method provided by the application, the morphing aircraft fault detection model is constructed based on a decision tree algorithm.

[0015] The application further provides a morphing aircraft fault detection method, comprising:

[0016] Obtaining flight state parameters of an aircraft to be detected, and inputting the flight state parameters into a morphing aircraft fault detection model;

[0017] Obtaining a fault detection result output by the morphing aircraft fault detection model;

[0018] The morphing aircraft fault detection model is generated based on the morphing aircraft fault detection model generation method according to any one of the above.

[0019] The application further provides a morphing aircraft fault detection model generation device, comprising:

[0020] The controller determining module is configured to determine a state control feedback matrix corresponding to each of the wing span change rates based on the initial values of the aircraft state variables and a preset flight linear variation parameter model of the morphing aircraft.

[0021] The simulation module is configured to generate flight simulation data based on to-be-simulated data, the to-be-simulated data including the initial values of the aircraft state variables, the wing span change rates and the state control feedback matrix, and each set of the flight simulation data including flight state parameters at multiple time points.

[0022] The data set generating module is configured to generate sample data based on the flight simulation data, the sample data including the flight simulation data and a fault label corresponding to the flight simulation data, the fault label being a fault when the wing span change rate corresponding to the state control feedback matrix included in the flight simulation data is inconsistent with the wing span change rate included in the flight simulation data, and the fault label being no fault when the wing span change rate corresponding to the state control feedback matrix included in the flight simulation data is consistent with the wing span change rate included in the flight simulation data.

[0023] The training module is configured to train a morphing aircraft fault detection model based on multiple sets of the sample data.

[0024] The present application also provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the morphing aircraft fault detection model generation method and / or the morphing aircraft fault detection method as described above.

[0025] The present application also provides a non-transitory computer-readable storage medium having a computer program stored thereon, and the computer program is executable on a processor to implement the morphing aircraft fault detection model generation method and / or the morphing aircraft fault detection method as described above.

[0026] The present application also provides a computer program product including a computer program, and the computer program is executable on a processor to implement the morphing aircraft fault detection model generation method and / or the morphing aircraft fault detection method as described above.

[0027] The application provides a variable-geometry aircraft fault detection model generation method and a fault detection method. BRIEF DESCRIPTION OF DRAWINGS

[0028] In order to more clearly illustrate the technical solutions in the application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are some embodiments of the application, and other drawings can be obtained by those skilled in the art without any creative effort on the basis of these drawings.

[0029] Figure 1 FIG. 1 is a flowchart of a variable-geometry aircraft fault detection model generation method provided by the application.

[0030] Figure 2 FIG. 2 is a system stability control effectiveness schematic diagram of a flight linear change parameter model of a variable-geometry aircraft in the variable-geometry aircraft fault detection model generation method provided by the application. Figure 1 .

[0031] Figure 3 FIG. 3 is a system stability control effectiveness schematic diagram of a flight linear change parameter model of a variable-geometry aircraft in the variable-geometry aircraft fault detection model generation method provided by the application. Figure 2 .

[0032] Figure 4 FIG. 4 is a system stability control effectiveness schematic diagram of a flight linear change parameter model of a variable-geometry aircraft in the variable-geometry aircraft fault detection model generation method provided by the application. Figure 3 .

[0033] Figure 5 FIG. 5 is a system stability control effectiveness schematic diagram of a flight linear change parameter model of a variable-geometry aircraft in the variable-geometry aircraft fault detection model generation method provided by the application. Figure 4 .

[0034] Figure 6 FIG. 6 is a system stability control effectiveness schematic diagram of a flight linear change parameter model of a variable-geometry aircraft in the variable-geometry aircraft fault detection model generation method provided by the application. Figure 5 .

[0035] Figure 7A schematic diagram of a variable shape aircraft fault detection model training process is provided.

[0036] Figure 8 A schematic diagram of a model confusion matrix in a variable shape aircraft fault detection model generation method is provided.

[0037] Figure 9 A schematic diagram of a variable shape aircraft fault detection method is provided.

[0038] Figure 10 A schematic diagram of a variable shape aircraft fault detection model generation device is provided.

[0039] Figure 11 A schematic diagram of an electronic device is provided. DETAILED DESCRIPTION

[0040] To make the objectives, technical solutions, and advantages of the present application clearer, the technical solutions in the present application will be described below in conjunction with the accompanying drawings in the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0041] The present application provides a variable shape aircraft fault detection model generation method. Figures 1-8 The variable shape aircraft fault detection model generation method includes the following steps: Figure 1

[0042] S110, based on each wing span change rate, the initial value of the aircraft state variable, and the preset flight linear change parameter model of the variable shape aircraft, a state control feedback matrix corresponding to each wing span change rate is obtained;

[0043] S120, flight simulation data is generated based on the to-be-simulated data, the to-be-simulated data includes the initial value of the aircraft state variable, the wing span change rate, and the state control feedback matrix, and each set of flight simulation data includes flight state parameters at multiple time points;

[0044] S130, sample data is generated based on the flight simulation data, the sample data includes the flight simulation data and a fault label corresponding to the flight simulation data, when the wing span change rate corresponding to the state control feedback matrix included in the flight simulation data is inconsistent with the wing span change rate included in the flight simulation data, the fault label is fault, and when the wing span change rate corresponding to the state control feedback matrix included in the flight simulation data is consistent with the wing span change rate included in the flight simulation data, the fault label is no fault;

[0045] ​S140, training based on multiple sets of sample data to obtain a variable-geometry aircraft fault detection model.

[0046] The method provided by the application realizes variable-geometry aircraft fault detection in a data-driven manner to adapt to the special needs of variable-geometry aircraft. Data-driven is a thinking mode of making decisions and taking actions based on data. Its advantage lies in that through collection, analysis, processing and application of data, more scientific and accurate decisions and prediction results can be obtained. In order to realize the above functions, the application first establishes a flight linear variation parameter (LPV) model of the variable-geometry aircraft based on flight dynamics principles and linearization methods. Specifically, the flight linear variation parameter model is as follows:

[0047] ;

[0048] wherein, , is the wing span variation rate, is the flight speed, is the state control feedback matrix, is the aircraft state variable, represents the disturbance in the flight of the aircraft, , , , is a preset fitting matrix.

[0049] The fitting matrix in the flight linear variation parameter model is obtained by fitting and simplifying the linearization processing of the nonlinear dynamics equation of the morphing aircraft at each working point in the flight envelope. The construction process of the model is described as follows:

[0050] By studying the specific influence of wing stretching on the aerodynamic performance of the aircraft, the aerodynamic analysis tool is used for simulation, and the aerodynamic parameters of the aircraft under diversified flight environments, various working states and different wing span deformation rates are calculated in advance, and the data are imported into the Matlab environment, and the data fitting technology is used to derive the key aerodynamic parameters such as lift coefficient , drag coefficient , and pitching moment coefficient , and the calculation formula is as follows:

[0051] ; (1.1)

[0052] wherein, represents the flight speed, represents the angle of attack, represents the pitch rate, represents the elevator deflection angle, Represents the average geometric chord length of the wing. After research, it was found that the above aerodynamic parameters are significantly affected by the wingspan deformation and flight speed changes. Therefore, the least squares fitting method is used to establish the relationship between aerodynamic parameters and wingspan deformation rate. and Mach number The relationship between them is as follows:

[0053] ; (1.2)

[0054] From the results of formula (1.1) and formula (1.2), it can be seen that the aerodynamic parameters of the variable span aircraft are significantly different from those of general aircraft. The most critical difference lies in the aerodynamic parameters and wingspan deformation rate of this aircraft. The extension and retraction of its wings can directly affect its aerodynamic characteristics.

[0055] After establishing the aerodynamic parameter expressions related to wing extension and retraction, the longitudinal nonlinear dynamic model of the aircraft's variable-profile flight is further constructed as follows:

[0056] ; (1.3)

[0057] in, represents the pitch angle, h represents the flight altitude, represents the engine thrust, represents the moment of inertia, is the acceleration due to gravity, Indicates the mass of the aircraft. ,resistance and pitching moment Respectively expressed as:

[0058] ; (1.4)

[0059] in, is the dynamic pressure, is the wing reference area.

[0060] Analyzing equations (1-3) and (1-4), we can see that wing deformation is factored into the dynamic model of a variable-shape aircraft. This indicates that during morphing flight, the wing's expansion and contraction will significantly alter the aircraft's aerodynamic characteristics, stimulating additional dynamic responses and thus affecting the overall flight trajectory. This interaction between wing deformation and flight dynamics is a core feature of variable-shape aircraft design and is crucial to controller design.

[0061] The Jacobian linearization technology based on LPV modeling is adopted to linearize the nonlinear dynamic equation (1.3) of the variable-span aircraft at each working point in the flight envelope, and through fitting and simplifying, the LPV model suitable for variable-shape flight is successfully constructed:

[0062] (1.5)

[0063] wherein, state variable ; respectively represent the flight speed change, the angle of attack change, the pitch angle change, the pitch angle speed change and the flight height change, the input variable , represents the throttle lever control change. The matrix coefficients are respectively

[0064] (1.6)

[0065] (1.7)

[0066] (1.8)

[0067] (1.9)

[0068] Considering the modeling error, external disturbance and additional uncertainty caused by the wing stretching, the LPV model (1.5) of the variable-span aircraft variable flight is rewritten as

[0069] (1.10)

[0070] wherein, d represents the external disturbance in the flight process, represents the state characteristic change caused by the deformation, represents the input characteristic change caused by the deformation, d, , which can be defined in advance, for example, can be defined as .

[0071] Let the composite disturbance be , then formula (1.10) can be simplified as

[0072] (1.11)

[0073] For the constructed LPV model, the state feedback design is performed on the closed-loop control system by the pole placement method, and the state feedback control is introduced: the expression of the state feedback control law is:

[0074] ; (1.12)

[0075] Substituting into equation (1.11), the closed-loop system is transformed into

[0076] ; (1.13)

[0077] Therefore, if a state feedback gain matrix is ​​designed , which makes Asymptotically stable, then when hour, By calculating the optimal gain matrix, the closed-loop poles are assigned to the specified positions to ensure the stability of the closed-loop system. In other words, the state feedback gain matrix includes the control parameters used to control the flight stability of the shape-shifting aircraft. Its principle is based on the complete controllability of the system state. That is, if the original system state is completely controllable, then there must be a state feedback gain matrix. , making the closed-loop system controllable and Configurable to any desired pole position.

[0078] Based on the flight linear variation parameter model provided by the present invention, the dynamic response process of the aircraft closed-loop system can be simulated, thereby realizing the simulation of the aircraft flight parameters. The simulation can be realized by Matlab. First, the parameters are initialized, the specific values ​​of the state matrix and the input matrix are input, and the controllability discriminant matrix is ​​used. Determine whether the system is controllable. Then, use the pole placement method to design the state feedback controller, and let the vector , the eigenvalue of the closed-loop system represents the behavioral characteristics of the system. If the eigenvalue is in the left half plane (the real part is negative), it means that the system is stable. The initial setting is: -6, -5+i, -5-i, -3+2i, -3-2i, then set the wingspan deformation rate The value of , using the solver function Design the corresponding state feedback gain matrix to configure the poles of the closed-loop system to The specified position and calculate the state matrix of the closed-loop system and eigenvalues .

[0079] Then, initialize the variables. Set the state variables , input variables , and introduce the initial small perturbation , to simulate the interference of external environment and the uncertain factors caused by wing stretch deformation when the aircraft is flying, and then based on the state space equation of discrete time linear time invariant system (LTI) and by taking the approximate integral method, the dynamic response process of the closed loop system is simulated through loop iteration.

[0080] The flight linear variation parameter model of the variable shape aircraft provided by the application combines the wing span deformation rate The time response curves of the five state parameters (speed variation, angle of attack variation, pitch angle variation, pitch angle speed variation and flight height variation) are shown in Figures 2-6 . Figure 2 is the time response curve of the speed variation, Figure 3 is the time response curve of the angle of attack variation, Figure 4 is the time response curve of the pitch angle variation, Figure 5 is the time response curve of the pitch angle speed variation, Figure 6 is the time response curve of the flight height variation. Overall, the dynamic performance of the flight linear variation parameter model of the variable shape aircraft established by the application is very excellent, and no matter how the value changes, the five state parameters can reach the steady state within a short time (within 3s) and have a fast response speed; after reaching the steady state, the state is basically unchanged, and the anti-interference performance is relatively strong; the overall overshoot is not too large, and the stability of the system is good.

[0081] As can be seen from the foregoing description, the flight linear variation parameter model of the variable shape aircraft provided by the application can simulate the flight state parameters of the variable shape aircraft under stable control through the wing span deformation rate, the initial value of the aircraft state variable and the preset flight linear variation parameter model of the variable shape aircraft. Based on this, the application first obtains the state control feedback matrix corresponding to each wing span deformation rate through each wing span deformation rate, each initial value of the aircraft state variable and the preset flight linear variation parameter model of the variable shape aircraft. At this time, the state control feedback matrix obtained can simulate the flight state parameters under the fault-free state, and therefore, the sample data with the fault-free fault label can be constructed. When there is a fault, the controller of the variable shape aircraft cannot identify the change of the wing span deformation rate, and therefore, the correspondence between the wing span deformation rate and the state control feedback matrix can be disturbed, and the sample data with the fault fault label can be constructed. For example, for the wing span deformation rate =0.1, the control state feedback matrix under the fault-free state is obtained through the flight linear variation parameter model, and the control state feedback matrix can control the aircraft to stably fly when the wing span deformation rate =0.1. However, when the wing span deformation rate =0.3, the control state feedback matrix Controlling the aircraft, therefore the wing span variation rate =0.3, the control state feedback matrix is used to simulate the aircraft flight state parameters in the fault state.

[0082] In the above manner, a plurality of sets of sample data can be constructed for model training. The specific process of constructing sample data for model training can be expressed as follows:

[0083] First, the parameters of the closed-loop system transfer function are set and initialized, and the wing span variation rate is selected as =0.1, =0.3, =0.6, =0.9, representing four different fault conditions (one normal state and three different fault states), according to different values, four different state matrices are obtained, and the function is used to calculate the state feedback gain to obtain different state feedback gain matrices , , , , so that the poles of the closed-loop system are located at the preset positions .

[0084] Then, the state variable and the input variable are initialized, and random noise is added to the initial state vector to simulate the uncertainty in the actual system. Then, the experiment uses a loop iteration method to constantly update the state vector to simulate the dynamic behavior of the aircraft. In order to fully test the system performance, 50 simulation experiments are defined, and in each simulation, data is saved every 20 steps. Finally, all simulation data is saved to a file. Since each simulation produces 500 rows of data, after 50 iterations, a total of 25,000 rows of flight simulation data are generated.

[0085] Next, the generated dataset is preprocessed. After reading the flight simulation data in the file, the dataset is divided into a training set and a test set, and the total number of rows of the dataset and the proportional allocation of the test set and the training set are specified. Then, in order to ensure the stability and comparability of the data, the training set and the test set are subjected to z-score standardization processing, which makes the processed data have a mean of 0 and a standard deviation of 1, which helps the subsequent training and evaluation of the algorithm. After that, the experiment specifies columns 1-5 as feature columns (containing key features related to fault detection: flight speed, angle of attack, pitch angle, pitch angle velocity, and flight altitude), and column 6 as the label column (representing the fault state of the aircraft). Finally, the labels and features of the processed training set and test set are saved to a new Excel file.

[0086] The fault detection model is trained on the constructed sample data to obtain a morphing aircraft fault detection model, which is a classification model that outputs a classification result of having a fault or not.

[0087] As Figure 7 shown, in one possible implementation, the morphing aircraft fault detection model is constructed based on a decision tree algorithm, that is, the decision tree algorithm model is trained on the constructed sample data to obtain a morphing aircraft fault detection model for detecting morphing aircraft faults.

[0088] The training process of the decision tree algorithm model can include:

[0089] In a loop iteration manner, the feature data and label data of the training set are read respectively and stored in the initialized empty matrices "train_features_data" and "train_labels_data"; then, in the same way, the feature data and label data of the test set are read respectively and stored in the initialized empty matrices "test_features_data" and "test_labels_data".

[0090] Secondly, the function is used to construct the decision tree model of the training set. Various parameter indicators are added in the function to improve the classification accuracy of the model.

[0091] Next, the classification accuracy of the training set is calculated and the result is output, if the accuracy does not meet the experimental requirements, the The function parameters set the maximum tree depth (MaxNumSplits) to 100, the minimum parent node sample size (MinParentSize) to 10, the minimum number of samples per leaf node (MinLeafSize) to 10, the Gini index (gdi) for the splitting criterion (SplitCriterion), and "on" for "Prune" to perform post-pruning on the tree. After model construction, the model was learned and trained using the training set data, and the classification accuracy of the training set was calculated to ensure that the model accurately fits the training data. The trained decision tree model was then applied to the test set data for classification and prediction. The classification accuracy for the normal state and the four fault states is shown in Table 1. The classification accuracy for all five states was above 98.5%, and the average classification accuracy for the fault state reached 99.8%. This fully demonstrates the excellent performance of the decision tree model in detecting and identifying variable-length faults.

[0092] Table 1

[0093]

[0094] Then, the K-fold cross-validation method was used to evaluate the model performance.

[0095] First, through loop iteration, multiple unclassified feature and label files are read one by one and stacked vertically in sample order to form a feature matrix and label vector.

[0096] Next, the constructed decision tree model is Fold cross validation. The experiment defines the number of folds 5. Use the decision tree model to train the data.

[0097] Then, the performance of the model is evaluated. The true positives ( ), false positives ( )、True counterexample( ) and false counterexamples ( ), and then calculate the accuracy, precision, recall, Value and other performance indicators.

[0098] Finally, use Function to calculate the confusion matrix and use The function is visualized and the classification results are added to each cell to more intuitively show the classification performance of the model. From the preprocessed data set, 5000 rows of feature data are randomly selected from each group for testing to verify whether the model can accurately diagnose each fault. The span deformation rate corresponding to the span length change command of the control system is calculated. The data sets when the expected values of the decision tree model take different values are input into the decision tree model respectively, and the corresponding confusion matrix is obtained, as shown in Table 1. Figure 8 The accuracy, precision, recall and F1 value of the model under different conditions are calculated, as shown in Table 2. It can be seen that the classification accuracy of the model is high, the comprehensive performance is good, and the prediction effect is good, and the wing span deformation fault of the variable shape aircraft can be quickly and accurately detected and identified.

[0099] Table 2

[0100]

[0101] In summary, in an implementation manner of the method provided by the application, through accurate decision tree algorithm design, reliable system verification and comprehensive performance evaluation, after detecting that the aircraft is in a fault state, fault identification is performed on different fault states of the variable shape aircraft, the classification accuracy of each type of fault reaches 98.5% or more, and the constructed model not only has high accuracy and stability, but also provides a reliable idea and method for fault diagnosis and safety guarantee related to the variable span of the variable shape aircraft.

[0102] Based on the variable shape aircraft fault detection model generation method described above, the application further provides a variable shape aircraft fault detection method, as shown in Figure 9 The variable shape aircraft fault detection method comprises the following steps:

[0103] S210, obtaining flight state parameters of a to-be-detected aircraft, and inputting the flight state parameters into a variable shape aircraft fault detection model;

[0104] S220, obtaining a fault detection result output by the variable shape aircraft fault detection model;

[0105] The variable shape aircraft fault detection model is generated based on the variable shape aircraft fault detection model generation method provided by the application.

[0106] The variable shape aircraft fault detection model generation device provided by the application is described below, and the variable shape aircraft fault detection model generation device described below can be correspondingly referred to the variable shape aircraft fault detection model generation method described above. As shown in Figure 10 The variable shape aircraft fault detection model generation device provided by the application comprises:

[0107] The controller determination module 1010 is configured to obtain a state control feedback matrix corresponding to each wing span change rate based on the initial values of each wing span change rate and the aircraft state variable and the preset flight linear change parameter model of the variable shape aircraft.

[0108] The simulation module 1020 is configured to generate flight simulation data based on to-be-simulated data, the to-be-simulated data including initial values of aircraft state variables, a wing span change rate, and a state control feedback matrix, and each set of flight simulation data including flight state parameters at multiple time points.

[0109] The data set generation module 1030 is configured to generate sample data based on the flight simulation data, the sample data including the flight simulation data and a fault label corresponding to the flight simulation data, the fault label being fault when the wing span change rate corresponding to the state control feedback matrix included in the flight simulation data is inconsistent with the wing span change rate included in the flight simulation data, and the fault label being no fault when the wing span change rate corresponding to the state control feedback matrix included in the flight simulation data is consistent with the wing span change rate included in the flight simulation data.

[0110] The training module 1040 is configured to train the variable-geometry aircraft fault detection model based on the multiple sets of sample data.

[0111] Figure 11 An example of an entity structure diagram of an electronic device is shown in FIG. 1. Figure 11As shown, the electronic device can include a processor 1110, a communications interface 1120, a memory 1130, and a communications bus 1140, wherein the processor 1110, the communications interface 1120, and the memory 1130 complete mutual communication through the communications bus 1140. The processor 1110 can invoke a logical instruction in the memory 1130 to execute a variable-geometry aircraft fault detection model generation method and / or a variable-geometry aircraft fault detection method. The variable-geometry aircraft fault detection model generation method includes: obtaining a state control feedback matrix corresponding to each wingspan change rate based on an initial value of each wingspan change rate and an aircraft state variable and a preset flight linear change parameter model of a variable-geometry aircraft; generating flight simulation data based on to-be-simulated data, the to-be-simulated data including an initial value of an aircraft state variable, a wingspan change rate, and a state control feedback matrix, each set of flight simulation data including flight state parameters at multiple time points; generating sample data based on the flight simulation data, the sample data including the flight simulation data and a fault label corresponding to the flight simulation data, the fault label being fault when the wingspan change rate corresponding to the state control feedback matrix included in the flight simulation data is inconsistent with the wingspan change rate included in the flight simulation data, and the fault label being no fault when the wingspan change rate corresponding to the state control feedback matrix included in the flight simulation data is consistent with the wingspan change rate included in the flight simulation data; and training to obtain a variable-geometry aircraft fault detection model based on multiple sets of sample data. The variable-geometry aircraft fault detection method includes: obtaining flight state parameters of a to-be-detected aircraft, inputting the flight state parameters to the variable-geometry aircraft fault detection model, and obtaining a fault detection result output by the variable-geometry aircraft fault detection model, wherein the variable-geometry aircraft fault detection model is generated based on the variable-geometry aircraft fault detection model generation method described above.

[0112] In addition, the logical instruction in the memory 1130 described above can be implemented in the form of a software function unit and sold or used as an independent product, and can be stored in a computer-readable storage medium. Based on such understanding, the technical solutions of the present application or parts of the present application that essentially contribute to the prior art or parts of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the various embodiments of the present application. The aforementioned storage medium includes a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0113] In another aspect, the present application also provides a computer program product, which comprises a computer program, the computer program being stored on a non-transitory computer readable storage medium, and the computer program being executable by a processor to cause a computer to perform the morphing aircraft fault detection model generation method and / or the morphing aircraft fault detection method provided by the above-mentioned methods, the morphing aircraft fault detection model generation method comprising: obtaining a state control feedback matrix corresponding to each wingspan change rate based on an initial value of each wingspan change rate, an initial value of an aircraft state variable, and a preset flight linear change parameter model of the morphing aircraft; generating flight simulation data based on to-be-simulated data, the to-be-simulated data comprising the initial value of the aircraft state variable, the wingspan change rate, and the state control feedback matrix, each set of flight simulation data comprising flight state parameters at multiple time points; generating sample data based on the flight simulation data, the sample data comprising the flight simulation data and a fault label corresponding to the flight simulation data, the fault label being fault when the wingspan change rate corresponding to the state control feedback matrix included in the flight simulation data is inconsistent with the wingspan change rate included in the flight simulation data, and the fault label being no fault when the wingspan change rate corresponding to the state control feedback matrix included in the flight simulation data is consistent with the wingspan change rate included in the flight simulation data; and training the morphing aircraft fault detection model based on multiple sets of sample data. The morphing aircraft fault detection method comprises: obtaining flight state parameters of a to-be-detected aircraft, and inputting the flight state parameters to the morphing aircraft fault detection model; and obtaining a fault detection result output by the morphing aircraft fault detection model, wherein the morphing aircraft fault detection model is generated based on the morphing aircraft fault detection model generation method mentioned above.

[0114] In another aspect, the present application also provides a non-transitory computer readable storage medium having stored thereon a computer program, which, when executed by a processor, implements the morphing aircraft fault detection model generation method and / or the morphing aircraft fault detection method provided by the above method, the morphing aircraft fault detection model generation method comprising: obtaining a state control feedback matrix corresponding to each wingspan change rate based on the initial value of each wingspan change rate, the initial value of the aircraft state variable, and a preset flight linear change parameter model of the morphing aircraft; generating flight simulation data based on to-be-simulated data, the to-be-simulated data comprising the initial value of the aircraft state variable, the wingspan change rate, and the state control feedback matrix, each set of flight simulation data comprising flight state parameters at multiple time points; generating sample data based on the flight simulation data, the sample data comprising the flight simulation data and a fault label corresponding to the flight simulation data, the fault label being fault when the wingspan change rate corresponding to the state control feedback matrix included in the flight simulation data is inconsistent with the wingspan change rate included in the flight simulation data, and the fault label being no fault when the wingspan change rate corresponding to the state control feedback matrix included in the flight simulation data is consistent with the wingspan change rate included in the flight simulation data; and training based on multiple sets of sample data to obtain a morphing aircraft fault detection model. The morphing aircraft fault detection method comprises: obtaining flight state parameters of a to-be-detected aircraft, and inputting the flight state parameters to the morphing aircraft fault detection model; and obtaining a fault detection result output by the morphing aircraft fault detection model, wherein the morphing aircraft fault detection model is generated based on the morphing aircraft fault detection model generation method described above.

[0115] The device embodiments described above are merely illustrative, wherein the units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, i.e., they can be located in one place or distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the present embodiment scheme according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0116] From the above description of the embodiments, those skilled in the art can clearly understand that the embodiments can be implemented by means of software and necessary universal hardware platforms, and of course, can also be implemented by hardware. Based on such understanding, the above technical solutions, essentially or in terms of contribution to the prior art, can be embodied in the form of a software product. The computer software product can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, etc., and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0117] Finally, it should be noted that the above examples are only used to illustrate the technical solutions of the present application, and are not intended to limit the same; although the present application has been described in detail with reference to the foregoing examples, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing examples, or make equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for generating a fault detection model for a variable-shape aircraft, characterized in that: include: Based on each wingspan change rate, the initial value of the aircraft state variable and a preset flight linear change parameter model of the deformable aircraft, a state control feedback matrix corresponding to each wingspan change rate is obtained; Generating flight simulation data based on the data to be simulated, the data to be simulated including initial values ​​of the aircraft state variables, a wingspan change rate, and a state control feedback matrix for closed-loop control, each set of the flight simulation data including flight state parameters at multiple moments; generating sample data based on the flight simulation data, the sample data including the flight simulation data and a fault label corresponding to the flight simulation data, wherein when the wingspan change rate corresponding to the state control feedback matrix for closed-loop control included in the flight simulation data is inconsistent with the wingspan change rate included in the flight simulation data, the fault label is fault; and when the wingspan change rate corresponding to the state control feedback matrix for closed-loop control included in the flight simulation data is consistent with the wingspan change rate included in the flight simulation data, the fault label is no fault; A deformable aircraft fault detection model is obtained by training based on multiple groups of sample data.

2. The method for generating a fault detection model for a deformable aircraft according to claim 1, wherein: The flight state parameters include flight speed, angle of attack, pitch angle, pitch angular velocity and flight altitude, and the aircraft state variables include flight speed change, angle of attack change, pitch angle change, pitch angular velocity change and flight altitude change.

3. The method for generating a fault detection model for a deformable aircraft according to claim 1, wherein: The flight linear variation parameter model is: Among them, A=A0+ξA1+VA2, ξ is the wingspan change rate, V is the flight speed, K is the state control feedback matrix, x is the aircraft state variable, D represents the interference during the aircraft flight, and B, A0, A1, and A2 are preset fitting matrices.

4. The method for generating a fault detection model for a deformable aircraft according to claim 3, wherein: The fitting matrix in the flight linear variation parameter model is obtained by linearizing the nonlinear dynamic equations of each working point of the deformable aircraft in the flight envelope, and then fitting and simplifying them.

5. The method for generating a fault detection model for a deformable aircraft according to claim 1, wherein: The deformable aircraft fault detection model is constructed based on a decision tree algorithm.

6. A method for detecting faults in a deformable aircraft, characterized in that: include: Obtaining flight state parameters of the aircraft to be detected, and inputting the flight state parameters into a deformable aircraft fault detection model; Obtaining a fault detection result output by the deformable shape aircraft fault detection model; The deformable-shape aircraft fault detection model is generated based on the deformable-shape aircraft fault detection model generation method according to any one of claims 1 to 5.

7. A device for generating a fault detection model for a deformable aircraft, characterized in that: include: A controller determination module is configured to obtain a state control feedback matrix corresponding to each wingspan change rate based on each wingspan change rate, an initial value of an aircraft state variable, and a preset flight linear change parameter model of a deformable aircraft; a simulation module for generating flight simulation data based on data to be simulated, wherein the data to be simulated includes initial values ​​of the aircraft state variables, a wingspan change rate, and a state control feedback matrix for closed-loop control, wherein each set of the flight simulation data includes flight state parameters at multiple moments; a data set generation module, configured to generate sample data based on the flight simulation data, wherein the sample data includes the flight simulation data and a fault label corresponding to the flight simulation data, wherein when the wingspan change rate corresponding to the state control feedback matrix for closed-loop control included in the flight simulation data is inconsistent with the wingspan change rate included in the flight simulation data, the fault label is fault; and when the wingspan change rate corresponding to the state control feedback matrix for closed-loop control included in the flight simulation data is consistent with the wingspan change rate included in the flight simulation data, the fault label is no fault; The training module is used to obtain a deformable aircraft fault detection model through training based on multiple groups of sample data.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the method for generating a fault detection model for a deformable aircraft according to any one of claims 1 to 5 and / or the method for detecting a fault for a deformable aircraft according to claim 6 are implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for generating a fault detection model for a deformable aircraft according to any one of claims 1 to 5 and / or the method for detecting a fault for a deformable aircraft according to claim 6 are implemented.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the method for generating a fault detection model for a deformable aircraft according to any one of claims 1 to 5 and / or the method for detecting a fault for a deformable aircraft according to claim 6 are implemented.

Images