Aircraft training data packet iterative correction method and system based on test flight data feedback

Abstract

The invention provides an aircraft training data packet iterative correction method and system based on test flight data feedback, and the method comprises the steps: collecting multiple types of test flight data in real time, and carrying out the preprocessing of the collected test flight data; based on the preprocessed test flight data, constructing a test flight-simulation mapping matrix; through the test flight-simulation mapping matrix, the multi-working-condition deviation degree of the test flight data is obtained through calculation; based on the multi-working-condition deviation degree, iterative correction is carried out on the test flight data with the multi-working-condition deviation degree exceeding the standard through a weighted gradient descent algorithm; and verifying the test flight data subjected to iterative correction. According to the method and the system provided by the invention, accurate matching of multiple types of test flight data and data packet parameters is realized by establishing'working condition-parameter 'two-dimensional association, and cross-working-condition correction errors are avoided; and in combination with parameter weights and physical constraints, multi-parameter collaborative iterative correction is realized, and the deviation degree is ensured to be rapidly reduced to 1.5% or below.

Description

Technical Field

[0001] This invention relates to the field of flight simulation technology, and in particular to an iterative correction method and system for aircraft training data packets based on flight test data feedback. Background Technology

[0002] The realism of C919 flight simulator training depends on the consistency between the training data package and the characteristics of the actual aircraft, and flight test data is the only direct basis for reflecting the characteristics of the actual aircraft. Currently, the initial parameters of the C919 training data package come from wind tunnel tests and theoretical calculations (such as the lift coefficient CL being set to 1.2 based on wind tunnel data). However, due to the influence of actual factors such as air viscosity and structural elasticity during actual flight tests, the measured CL under the same operating conditions may be 1.25, leading to simulation deviations.

[0003] Current correction methods require engineers to manually extract flight test data (e.g., exporting an Excel spreadsheet from the flight test data recorder) after the test flights are completed, compare the data packet parameters one by one, and make corrections. This is not only time-consuming but also cannot cover parameter differences under multiple operating conditions (such as high-altitude takeoff and night approach). As the number of C919 test flights increases (more than 1,000 cumulative test flights), the massive amount of test flight data cannot be used efficiently, and discrepancies between the data packets and the actual aircraft characteristics persist for a long time. Summary of the Invention

[0004] The embodiments of the present invention provide an iterative correction method and system for aircraft training data packets based on flight test data feedback, which is used to solve the technical problems existing in the prior art.

[0005] To achieve the above objectives, the present invention adopts the following technical solution.

[0006] The iterative correction method for aircraft training data packets based on flight test data feedback includes:

[0007] S1 collects various types of flight test data in real time and preprocesses the collected flight test data;

[0008] S2 Based on the preprocessed flight test data, a flight test-simulation mapping matrix is ​​constructed; the multi-condition deviation of the flight test data is calculated through the flight test-simulation mapping matrix.

[0009] S3 is based on multi-condition deviation and uses a weighted gradient descent algorithm to iteratively correct flight test data with excessive multi-condition deviation.

[0010] S4 verifies the flight test data after iterative corrections;

[0011] The validated flight test data is used to upgrade the aircraft simulator.

[0012] Preferably, in step S2, the flight test-simulation mapping matrix is ​​a two-dimensional mapping matrix M, which has the dimensions of 8 operating conditions × 12 core parameters, and sets the matrix elements M. j,k Let k be the weight of the k-th parameter under the j-th working condition;

[0013] In step S2, the process of calculating the multi-condition deviation of the flight test data through the flight test-simulation mapping matrix includes:

[0014] Based on the pre-processed flight test data, the corresponding operating condition parameters in the flight test data are matched according to the pre-made operating condition labels;

[0015] For each matched operating condition parameter k, the formula is used...

[0016]

[0017] Calculate the relative deviation between the test flight value Ptest,k and the data packet value Psim,k. ;

[0018] Through

[0019]

[0020] Calculate each matrix element M j,k The overall deviation under the j-th working condition;

[0021] The overall deviation of a certain working condition is weighted according to the frequency of occurrence of that working condition, and then calculated using the formula...

[0022]

[0023] Calculate the global deviation of the flight test data; where: Let be the training frequency for the j-th working condition.

[0024] Preferably, the execution process of step S3 specifically includes:

[0025] Based on the physical performance of the actual aircraft and flight training standards, set constraints for correcting flight test parameters;

[0026] Initialize the parameters in the flight test data that need to be corrected;

[0027] For a specific parameter that needs to be corrected Through the formula

[0028]

[0029] Calculate the global deviation for this parameter The partial derivatives are used to reflect the direction of the influence of parameter changes on the deviation;

[0030] Through

[0031]

[0032] This parameter along the negative gradient direction Adjust the parameter to make it so that Change in the direction of decreasing deviation It is an empirical coefficient;

[0033] If the adjusted parameter Pk+1 exceeds the constraint range, it will be truncated to the boundary value.

[0034] Through

[0035]

[0036] Calculate the adjusted parameter P k +1 global deviation If the global deviation Dglobal(P) k +1) ≤ preset threshold δ th , or the number of iterations k ≥ k max If the iteration fails, stop; otherwise, let k = k+1 and return to step S2.

[0037] Preferably, step S3 further includes:

[0038] For multiple parameters that need to be corrected due to coupling relationships, a multivariate gradient descent method is used to perform simultaneous correction of multiple parameters.

[0039] Preferably, the execution process of step S4 specifically includes:

[0040] S41 Collect the parameters output by the aircraft simulator, calculate the static deviation between the collected parameters output by the aircraft simulator and the test flight data after iterative correction. If the static deviation is less than the first preset threshold, execute sub-step S42; otherwise, return to execute step S2.

[0041] S42 collects dynamic response curves of the aircraft simulator by running dynamic test cases;

[0042] S43 compares the dynamic response curve of the collected aircraft simulator with the dynamic curve of the test flight data after iterative correction, and obtains the dynamic deviation index by calculation. If the dynamic deviation index exceeds the second preset threshold and the third preset threshold, the execution process of the aircraft training data packet iterative correction method is completed.

[0043] In a second aspect, the present invention provides an iterative correction system for aircraft training data packets based on flight test data feedback, which performs the above-described method, comprising:

[0044] The data acquisition module is used to: collect various types of flight test data in real time and preprocess the collected flight test data;

[0045] The data analysis and correction module is used for:

[0046] Based on the preprocessed flight test data, a flight test-simulation mapping matrix is ​​constructed; the multi-condition deviation of the flight test data is calculated using the flight test-simulation mapping matrix.

[0047] Based on the multi-condition deviation, the flight test data with excessive multi-condition deviation is iteratively corrected using a weighted gradient descent algorithm.

[0048] The verification module is used for:

[0049] The test flight data, after iterative corrections, were verified.

[0050] Update the version of the fully verified flight test data to obtain the latest version of the flight test data;

[0051] The latest flight test data is sent to the aircraft simulator, enabling the simulator to update its software and allowing the data acquisition module to collect new flight test data in real time.

[0052] Preferably, the data analysis and correction module is further configured to:

[0053] The new flight test data was compared with the original flight test data, using a formula...

[0054]

[0055] The deviation D of the new test flight data relative to the original test flight data is calculated. If the deviation D is less than the fourth preset threshold, a verification pass message is sent to the verification module. Otherwise, the following iterative correction process is performed: a test flight-simulation mapping matrix is ​​constructed; the multi-condition deviation of the new test flight data is calculated using the test flight-simulation mapping matrix; based on the multi-condition deviation of the new test flight data, the data with excessive multi-condition deviation of the new test flight data are iteratively corrected using a weighted gradient descent algorithm.

[0056] As can be seen from the technical solutions provided by the embodiments of the present invention above, the present invention provides an iterative correction method and system for aircraft training data packets based on flight test data feedback. The method includes: real-time acquisition of multiple types of flight test data; preprocessing the acquired flight test data; constructing a flight test-simulation mapping matrix based on the preprocessed flight test data; calculating the multi-condition deviation of the flight test data through the flight test-simulation mapping matrix; iteratively correcting the flight test data with excessive multi-condition deviation using a weighted gradient descent algorithm based on the multi-condition deviation; and verifying the flight test data after iterative correction. The method and system provided by this invention establish a two-dimensional correlation between "operating condition and parameters" to achieve accurate matching between multiple types of flight test data and data packet parameters, avoiding errors in cross-operating condition correction; combining parameter weights and physical constraints, it achieves multi-parameter collaborative iterative correction, ensuring that the deviation is quickly reduced to below 1.5%; through dual verification of "simulation verification + flight test feedback", it establishes a long-term iterative process, so that the data packet parameters continuously approach the characteristics of the real aircraft as flight test data accumulates; and it uses the 3σ criterion to remove anomalies, Kalman filtering to remove noise, and standardization transformation to ensure the usability and consistency of flight test data.

[0057] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and will become apparent from the description or may be learned by practice of the invention. Attached Figure Description

[0058] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0059] Figure 1 The flowchart of the iterative correction method for aircraft training data packets based on flight test data feedback provided by the present invention is shown below.

[0060] Figure 2 The present invention provides a logic block diagram of an aircraft training data packet iterative correction system based on flight test data feedback. Detailed Implementation

[0061] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0062] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in this specification means the presence of the stated features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. It should be understood that when we say an element is “connected” or “coupled” to another element, it can be directly connected or coupled to the other element, or there may be intermediate elements. Furthermore, “connected” or “coupled” as used herein can include wireless connections or couplings. The term “and / or” as used herein includes any and all combinations of one or more of the associated listed items.

[0063] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. It should also be understood that terms such as those defined in general dictionaries should be understood to have the same meaning as in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless defined as herein.

[0064] To facilitate understanding of the embodiments of the present invention, the following will provide further explanation and description with reference to the accompanying drawings and several specific embodiments. These embodiments do not constitute a limitation on the embodiments of the present invention.

[0065] This invention provides a method and system for iterative correction of aircraft training data packets based on flight test data feedback, to solve the following technical problems existing in the prior art:

[0066] The closest implementation to this invention is the "Static Correction Method for Data Packets Based on Wind Tunnel Data", which is applied to the simulation training system of the Airbus A350 aircraft (Source: Airbus A350 Flight Simulation DataPackage Calibration Guide).

[0067] The core process of this solution is as follows:

[0068] Collect wind tunnel test data (such as drag coefficient CD at different Mach numbers);

[0069] Technicians use specialized tools to open the data package and manually replace the original parameters with wind tunnel data (e.g., changing CD from 0.02 to 0.022).

[0070] The revised version was only tested using a ground simulator and was not verified using actual aircraft flight test data.

[0071] There is no iterative mechanism; corrections are only made once when the wind tunnel data is updated.

[0072] Its limitations are:

[0073] It relies on wind tunnel data (which differs from the actual aircraft environment) and does not utilize flight test data for feedback.

[0074] Manual correction is inefficient and cannot adapt to multiple working conditions;

[0075] Without closed-loop iteration, the accuracy of the corrected parameters cannot be continuously improved (the deviation D is consistently >4%).

[0076] Limited data sources: relying on wind tunnel data rather than actual flight test data, resulting in significant deviations between the corrected parameters and actual flight characteristics.

[0077] Inefficient correction: Manual comparison and modification require 2-3 days to correct parameters for a single operating condition, and more than 1 week for correction of multiple operating conditions.

[0078] No operational condition association: The mapping between "flight test conditions and data packet parameters" has not been established, which can easily lead to cross-operation condition correction errors (such as using level flight data to correct stall parameters).

[0079] Lack of closed-loop iteration: After correction, the test was only conducted through simulation without being verified with new flight test data, thus preventing continuous optimization.

[0080] See Figure 1 This invention provides an iterative correction method for aircraft training data packets based on flight test data feedback, comprising the following steps:

[0081] S1 collects various types of flight test data in real time and preprocesses the collected flight test data;

[0082] S2 Based on the preprocessed flight test data, a flight test-simulation mapping matrix is ​​constructed; the multi-condition deviation of the flight test data is calculated through the flight test-simulation mapping matrix.

[0083] S3 is based on multi-condition deviation and uses a weighted gradient descent algorithm to iteratively correct flight test data with excessive multi-condition deviation.

[0084] S4 verifies the flight test data after iterative corrections;

[0085] The validated flight test data is used to upgrade the aircraft simulator's software, making the environment simulation and control feedback of the aircraft simulator closer to real-world scenarios.

[0086] In a preferred embodiment provided by the present invention, step S1 specifically includes:

[0087] This embodiment uses the C919 model as an example:

[0088] Data collected: Key data from the C919's actual flight test process, categorized by type as follows:

[0089] Aerodynamic parameters: lift coefficient C L Drag coefficient C D Pitch moment coefficient Cm (collected by wing surface pressure sensors);

[0090] Attitude parameters: altitude (GPS sensor, accuracy ±0.5m), airspeed (Pitto sensor, accuracy ±1km / h), pitch angle (gyroscope, accuracy ±0.1°).

[0091] Control response parameters: lever input (displacement sensor, range -50~+50mm), throttle position (angle sensor, 0~100%).

[0092] Environmental parameters: atmospheric density (air velocity tube + temperature sensor, accuracy ±0.01 kg / m³), wind speed (weather radar, accuracy ±0.5 m / s).

[0093] Data acquisition equipment: The C919 test flight dedicated data recorder (model: CFDR-2000) is used, with a sampling frequency of 100Hz and data storage format of ARINC 717 (civil aviation standard test flight data format).

[0094] Operating condition labeling: Automatically adds operating condition labels to the data (such as "takeoff phase - high-altitude airport - altitude 3000m"), and makes judgments based on a combination of altitude, speed, and throttle position (such as altitude < 1000m and throttle > 80% is judged as takeoff operating condition).

[0095] (2) Data preprocessing

[0096] Outlier removal: Using the 3σ criterion, for each parameter P, calculate the mean μ and standard deviation σ, and remove outliers with values ​​greater than σ (such as airspeed sudden change to 0 caused by sensor failure).

[0097] Noise filtering: Kalman filtering (filter coefficient α=0.8) is used to suppress high-frequency noise (such as CL fluctuations caused by airflow disturbances), the formula is:

[0098] in Pt is the filtered value, and Pt is the current sampled value. This is the filtered value from the previous moment;

[0099] Data standardization: Convert parameters to a format consistent with the data packet (e.g., convert "stick displacement (-50~+50mm)" in flight test data to "standardized control input (-1~+1)" in the data packet). The conversion formula is:

[0100]

[0101] Where d is the flight test displacement, d min =−50mm,d max =+50mm, where u is the standardized control value.

[0102] In a preferred embodiment of the present invention, the process of calculating the multi-condition deviation of flight test data through the flight test-simulation mapping matrix includes:

[0103] Based on the pre-processed flight test data, the corresponding operating condition parameters in the flight test data are matched according to the pre-made operating condition labels;

[0104] For each matched operating condition parameter k, the formula is used...

[0105]

[0106] Calculate the relative deviation between the test flight value Ptest,k and the data packet value Psim,k;

[0107] Through

[0108]

[0109] Calculate each matrix element M j,k The overall deviation under the j-th working condition;

[0110] The overall deviation of a certain working condition is weighted according to the frequency of occurrence of that working condition, and then calculated using the formula...

[0111]

[0112] Calculate the global deviation of the flight test data; where: Let be the training frequency for the j-th working condition.

[0113] Taking the C919 aircraft as an example, based on the C919 flight manual and test flight subjects, a two-dimensional mapping matrix M of "operating condition-parameter" is established (dimension: 8 operating conditions × 12 core parameters). The matrix element Mj,k is the weight of the k-th parameter under the j-th operating condition (weight range 0.1~0.9, the higher the weight, the greater the influence). An example is shown below:

[0114]

[0115] Case Study: Deviation Analysis of Cruise Condition (10000m, 0.78Ma)

[0116] Flight test data: CL,test=1.25, CD,test=0.023;

[0117] Data packet parameters: CL,sim=1.2, CD,sim=0.02;

[0118] Single parameter deviation: δCL=∣1.2−1.25∣ / 1.25=0.04, δCD=∣0.02−0.023∣ / 0.023≈0.13;

[0119] Operating condition deviation: Cruise (13.2%), far exceeding the threshold of 1.5%, requires correction.

[0120] In step S3, a weighted gradient descent algorithm is used to iteratively correct the parameters that have exceeded the deviation limit. The core is to minimize the global deviation Dglobal while satisfying the aircraft's physical constraints. The specific steps are as follows:

[0121] Based on the physical performance of the actual aircraft and flight training standards, set constraints for correcting flight test parameters;

[0122] Initialize the parameters in the flight test data that need to be corrected;

[0123] For a specific parameter that needs to be corrected Through the formula

[0124]

[0125] Calculate the global deviation for this parameter The partial derivatives are used to reflect the direction of the influence of parameter changes on the deviation;

[0126] Through

[0127]

[0128] This parameter along the negative gradient direction Adjust the parameter to make it so that Change in the direction of decreasing deviation;

[0129] If the adjusted parameter P k If +1 exceeds the constraint range, it will be truncated to the boundary value;

[0130] Through

[0131]

[0132] Calculate the global deviation of the adjusted parameter Pk+1. If the global deviation Dglobal(P) k +1)≤δ th (where δ) th (e.g., a preset threshold, such as 1.5%), or the number of iterations k ≥ k maxIf the iteration fails, stop; otherwise, let k = k+1 and return to step S2.

[0133] Taking the C919 model as an example:

[0134] Set modified constraints

[0135] Based on the physical performance of the C919 aircraft and civil aviation training standards, the constraint range for parameter correction is set as follows:

[0136] Aerodynamic parameters: C L ∈[0.8,1.5], C D ∈[0.015,0.03], Cm∈[−0.2,0.3];

[0137] Control parameters: control stick response coefficient K∈[0.5,2.0] (to ensure control sensitivity is within a safe range).

[0138] Initialize correction parameters

[0139] The number of iterations k=0, and the maximum number of iterations k max =50;

[0140] Learning rate η = 0.05 (adaptive adjustment: η = 0.1 when >, η = 0.05 when <);

[0141] Initial parameters: P0=[C L,sim C D,sim ,...,K sim (Current parameters of the data packet).

[0142] Iterative correction process

[0143] Calculate the bias gradient: for each parameter P k Calculate Dglobal for P k The partial derivatives (gradient) reflect the direction of the effect of parameter changes on the deviation:

[0144] Where ΔP = 0.001 (a tiny increment).

[0145] Update parameters: Adjust parameters along the negative gradient direction to ensure they change in the direction of decreasing bias.

[0146] Constraint verification: If Pk+1 exceeds the constraint range, truncate it to the boundary value (e.g., when Pk+1=1.55, truncate it to 1.5).

[0147] Deviation verification: Calculate the updated If k < or k = kmax, stop the iteration; otherwise, k = k + 1 and return to step 1.

[0148] Case Study: Cruise Condition C L Iterative correction

[0149] Initial parameter: C L,0 =1.2, Dglobal,0=13.2%;

[0150] First iteration: ∇DC L =0.8 (the deviation decreases as the parameter increases), C L,1 =1.2−0.1×0.8=1.28, Dglobal,1=8.5%;

[0151] Second iteration: ∇DC L =0.4, C L,2 =1.28−0.05×0.4=1.26, Dglobal,2=3.2%;

[0152] 3rd iteration: ∇DC L =0.1, C L,3 =1.26−0.05×0.1=1.255,Dglobal,3=1.4% (<1.5%, stop iteration);

[0153] Final correction value: C L,final =1.255 (the deviation from the test flight value of 1.25 is only 0.4%).

[0154] In some improved embodiments, a multi-parameter collaborative correction process is also included:

[0155] For parameters that have a coupling relationship (such as C) L (Compared with Cm), multivariate gradient descent is used to synchronously update parameters to avoid new biases caused by single-parameter corrections. For example, correcting C... L Simultaneously adjust Cm (because changes in lift will affect the pitching moment) to ensure the physical correlation between parameters.

[0156] The execution process of step S4 specifically includes:

[0157] S41 Collect the parameters output by the aircraft simulator, calculate the static deviation between the collected parameters output by the aircraft simulator and the test flight data after iterative correction. If the static deviation is less than the first preset threshold, execute sub-step S42; otherwise, return to execute step S2.

[0158] S42 collects dynamic response curves of the aircraft simulator by running dynamic test cases;

[0159] S43 compares the dynamic response curve of the collected aircraft simulator with the dynamic curve of the test flight data after iterative correction, and obtains the dynamic deviation index by calculation. If the dynamic deviation index exceeds the second preset threshold and the third preset threshold, the execution process of the aircraft training data packet iterative correction method is completed.

[0160] Taking the C919 model as an example:

[0161] Simulator static verification

[0162] Load the corrected data package on the C919 simulator and run standard test cases (such as maintaining an altitude of 10,000m during the cruise segment).

[0163] Collect parameters output by the simulator (such as C) L C D ), calculate the static deviation from the flight test data (requirement <1.5%).

[0164] If the static deviation meets the standard, proceed to dynamic verification; otherwise, return to the correction step to adjust the parameters.

[0165] Simulator dynamic verification

[0166] Run dynamic test cases (such as pitch response after pushing the lever 10°) and collect the dynamic response curves of the simulator (such as pitch angle change over time).

[0167] Compare the dynamic curves with the flight test data to calculate dynamic deviation indicators (such as step response delay <50ms, overshoot <5%).

[0168] Once the dynamic deviation meets the standard, the data packet is marked as "Pending Flight Test Verification" version (e.g., V1.1).

[0169] Flight test data closed-loop verification

[0170] Write the data packet parameters of the "to be tested and verified" version into the C919 real aircraft flight test data acquisition system;

[0171] During the next test flight, new test flight data will be collected under the same operating conditions (such as new C-mode cruise conditions). L =1.252);

[0172] Comparing the new flight test data with the revised parameters (C) L,sim =1.255), the deviation was calculated as D = |1.255−1.252| / 1.252≈0.24% (<1.5%), and the verification was successful;

[0173] If the verification passes, the data package will be officially updated to V1.1; otherwise, return to the deviation analysis step and start the next iteration.

[0174] In some improved embodiments, a closed-loop iteration mechanism is also included, specifically:

[0175] Establish a long-term iterative process of "correction-verification-correction" to ensure that data packets are continuously optimized as test flight data accumulates:

[0176] Regular iteration: After every 3 test flights, a global deviation analysis and correction is automatically triggered (cycle approximately 1 month).

[0177] Triggering iteration: If the deviation of a certain condition during a test flight is greater than 1, the parameter correction for that condition will be triggered immediately.

[0178] Version management: Create an archive for each iteration (including correction time, test flight data source, and deviation changes), and support backtracking to any historical version (e.g., V1.0 → V1.1 → V1.2).

[0179] Secondly, the present invention provides an iterative correction system for aircraft training data packets based on flight test data feedback, comprising:

[0180] The data acquisition module 201 is used to: acquire multiple types of flight test data in real time and preprocess the acquired flight test data;

[0181] Data analysis and correction module 202 is used for:

[0182] Based on the preprocessed flight test data, a flight test-simulation mapping matrix is ​​constructed; the multi-condition deviation of the flight test data is calculated using the flight test-simulation mapping matrix.

[0183] Based on the multi-condition deviation, the flight test data with excessive multi-condition deviation is iteratively corrected using a weighted gradient descent algorithm.

[0184] Verification module 203 is used for:

[0185] The test flight data, after iterative corrections, were verified.

[0186] Update the version of the fully verified flight test data to obtain the latest version of the flight test data;

[0187] The latest flight test data is sent to the aircraft simulator, enabling the simulator to update its software and allowing the data acquisition module to collect new flight test data in real time.

[0188] The data analysis and correction module is also used for:

[0189] The new flight test data was compared with the original flight test data, using a formula...

[0190]

[0191] The deviation D of the new test flight data relative to the original test flight data is calculated. If the deviation D is less than the fourth preset threshold, the verification module sends a verification pass message. Otherwise, the following iterative correction process is executed: a test flight-simulation mapping matrix is ​​constructed; the multi-condition deviation of the new test flight data is calculated using the test flight-simulation mapping matrix; based on the multi-condition deviation of the new test flight data, the data with excessive multi-condition deviation of the new test flight data are iteratively corrected using a weighted gradient descent algorithm.

[0192] The present invention also provides two embodiments to illustrate the corrective effects of the method provided by the present invention.

[0193] Example 1:

[0194] Case Study: C919 Approach Condition Correction Effect

[0195] Existing technology: C-correction based on wind tunnel data L =1.3, flight test measured C L =1.38, deviation 6.4%;

[0196] This invention: Through three rounds of iterative correction, C L From 1.3→1.36→1.375→1.382, the final deviation was 0.8%, and the RMSE of the simulated approach trajectory and the test flight trajectory decreased from 5m to 1.2m.

[0197] Example 2:

[0198] Experimental verification data: Ten rounds of iterative corrections were performed on five core operating conditions of the C919 (takeoff, cruise, approach, crosswind landing, and single-engine failure). The results are as follows:

[0199]

[0200] Key algorithm parameters:

[0201] Gradient descent: learning rate η = 0.05 - 0.1 (adaptive), maximum number of iterations 50, convergence threshold <;

[0202] Kalman filtering: process noise covariance Q = 0.01, measurement noise covariance R = 0.02, initial covariance P0 = 0.1;

[0203] Mapping matrix: The operating condition weights are set based on the C919 training outline (cruise 0.3, takeoff 0.2, approach 0.2, others 0.3).

[0204] Software and hardware tools:

[0205] Data acquisition hardware: C919 flight test data recorder (CFDR-2000), pressure sensor (accuracy ±0.1kPa), gyroscope (accuracy ±0.01°);

[0206] Software tools: Python 3.11 (Pandas library for processing flight test data), MATLAB 2023 (gradient descent algorithm implementation), C++ 20 (data packet parameter modification interface development), Qt 6.5 (visualized deviation analysis interface).

[0207] In summary, this invention provides a method and system for iterative correction of aircraft training data packets based on flight test data feedback. The method includes: real-time acquisition of multiple types of flight test data; preprocessing the acquired flight test data; constructing a flight test-simulation mapping matrix based on the preprocessed flight test data; calculating the multi-condition deviation of the flight test data using the flight test-simulation mapping matrix; iteratively correcting flight test data with excessive multi-condition deviation using a weighted gradient descent algorithm based on the multi-condition deviation; and verifying the iteratively corrected flight test data. The method and system provided by this invention achieve precise matching between multiple types of flight test data and data packet parameters by establishing a two-dimensional "condition-parameter" association, avoiding cross-condition correction errors; combine parameter weights and physical constraints to achieve multi-parameter collaborative iterative correction, ensuring the deviation quickly decreases to below 1.5%; establish a long-term iterative process through dual verification of "simulation verification + flight test feedback," allowing data packet parameters to continuously approach the characteristics of the actual aircraft as flight test data accumulates; and employ 3σ criteria for anomaly removal, Kalman filtering for noise reduction, and standardization transformation to ensure the usability and consistency of the flight test data. The specific advantages are shown in the table below:

[0208]

[0209] Those skilled in the art will understand that the accompanying drawings are merely schematic diagrams of one embodiment, and the modules or processes shown in the drawings are not necessarily essential for implementing the present invention.

[0210] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of the present invention.

[0211] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, for apparatus or system embodiments, since they are basically similar to method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. The apparatus and system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0212] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An iterative correction method for aircraft training data packets based on flight test data feedback, characterized in that, include: S1 collects various types of flight test data in real time and preprocesses the collected flight test data; S2 Based on the preprocessed flight test data, a flight test-simulation mapping matrix is ​​constructed; the multi-condition deviation of the flight test data is calculated through the flight test-simulation mapping matrix. S3 is based on multi-condition deviation and uses a weighted gradient descent algorithm to iteratively correct flight test data with excessive multi-condition deviation. S4 verifies the flight test data after iterative corrections; The validated flight test data is used to upgrade the aircraft simulator.

2. The aircraft training data packet iterative correction method according to claim 1, characterized in that, In step S2, the flight test-simulation mapping matrix is ​​a two-dimensional mapping matrix M. The dimensions of this two-dimensional mapping matrix M are 8 operating conditions × 12 core parameters, and the matrix elements M are set. j,k Let k be the weight of the k-th parameter under the j-th working condition; In step S2, the process of calculating the multi-condition deviation of the flight test data through the flight test-simulation mapping matrix includes: Based on the pre-processed flight test data, the corresponding operating condition parameters in the flight test data are matched according to the pre-made operating condition labels; For each matched operating condition parameter k, the formula is used... Calculate the relative deviation between the test flight value Ptest,k and the data packet value Psim,k. ; Through Calculate each matrix element M j,k The overall deviation under the j-th working condition; The overall deviation of a certain working condition is weighted according to the frequency of occurrence of that working condition, and then calculated using the formula... Calculate the global deviation of the flight test data; where: Let be the training frequency for the j-th working condition.

3. The aircraft training data packet iterative correction method according to claim 2, characterized in that, The execution process of step S3 specifically includes: Based on the physical performance of the actual aircraft and flight training standards, set constraints for correcting flight test parameters; Initialize the parameters in the flight test data that need to be corrected; For a specific parameter that needs to be corrected Through the formula Calculate the global deviation for this parameter The partial derivatives are used to reflect the direction of the influence of parameter changes on the deviation; Through This parameter along the negative gradient direction Adjust the parameter to make it so that Change in the direction of decreasing deviation It is an empirical coefficient; If the adjusted parameter Pk+1 exceeds the constraint range, it will be truncated to the boundary value. Through Calculate the adjusted parameter P k +1 global deviation If the global deviation Dglobal(P) k +1) ≤ preset threshold δ th , or the number of iterations k ≥ k max If the iteration fails, stop; otherwise, let k = k+1 and return to step S2.

4. The aircraft training data packet iterative correction method according to claim 3, characterized in that, Step S3 also includes: For multiple parameters that need to be corrected due to coupling relationships, a multivariate gradient descent method is used to perform simultaneous correction of multiple parameters.

5. The aircraft training data packet iterative correction method according to claim 3, characterized in that, The execution process of step S4 specifically includes: S41 Collect the parameters output by the aircraft simulator, calculate the static deviation between the collected parameters output by the aircraft simulator and the test flight data after iterative correction. If the static deviation is less than the first preset threshold, execute sub-step S42; otherwise, return to execute step S2. S42 collects dynamic response curves of the aircraft simulator by running dynamic test cases; S43 compares the dynamic response curve of the collected aircraft simulator with the dynamic curve of the test flight data after iterative correction, and obtains the dynamic deviation index by calculation. If the dynamic deviation index exceeds the second preset threshold and the third preset threshold, the execution process of the aircraft training data packet iterative correction method is completed.

6. An iterative correction system for aircraft training data packets based on flight test data feedback, characterized in that, include: The data acquisition module is used to: collect various types of flight test data in real time and preprocess the collected flight test data; The data analysis and correction module is used for: Based on the preprocessed flight test data, a flight test-simulation mapping matrix is ​​constructed; the multi-condition deviation of the flight test data is calculated using the flight test-simulation mapping matrix. Based on the multi-condition deviation, the flight test data with excessive multi-condition deviation is iteratively corrected using a weighted gradient descent algorithm. The verification module is used for: The test flight data, after iterative corrections, were verified. Update the version of the fully verified flight test data to obtain the latest version of the flight test data; The latest flight test data is sent to the aircraft simulator, enabling the simulator to update its software and allowing the data acquisition module to collect new flight test data in real time.

7. The aircraft training data packet iterative correction system according to claim 6, characterized in that, The data analysis and correction module is also used for: The new flight test data was compared with the original flight test data, using a formula... The deviation D of the new test flight data relative to the original test flight data is calculated. If the deviation D is less than the fourth preset threshold, a verification pass message is sent to the verification module. Otherwise, the following iterative correction process is performed: a test flight-simulation mapping matrix is ​​constructed; the multi-condition deviation of the new test flight data is calculated using the test flight-simulation mapping matrix; based on the multi-condition deviation of the new test flight data, the data with excessive multi-condition deviation of the new test flight data are iteratively corrected using a weighted gradient descent algorithm.

Images