Vertical fin load ground calibration and flight actual measurement method in T-shaped empennage
By conducting ground calibration tests on vertical and horizontal tail loads and using an elastic net sparse regression algorithm, the problem of the coupling effect between vertical tail load and horizontal tail load in T-tail fins was solved, enabling accurate measurement of vertical tail load and providing data support for aircraft structural safety assessment.
Patent Information
- Application Number
- CN202511965587.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-24
- Publication Date
- 2026-03-27
AI Technical Summary
The lack of effective methods in the current technology to distinguish and calibrate the coupling effect between the vertical tail load and the horizontal tail load in the T-tail fin leads to inaccurate measurement accuracy of the vertical tail load. Traditional multivariate regression methods cannot eliminate invalid strain signals and redundant collinearity, which affects the accurate calculation of loads in flight.
Ground calibration tests with vertical and horizontal tail loading were conducted, and the elastic net sparse regression algorithm was used to establish calibration equations for the clean load, coupled strain, and coupled load of the vertical tail, respectively. By separating the loading test and using a sparsity modeling strategy, invalid features were eliminated to improve calibration accuracy. During flight, the total load of the vertical tail was accurately reconstructed based on the decoupled strain of the ground calibration equations.
It enables precise differentiation and measurement of vertical tail loads in T-tail fins, improves the physical accuracy of load identification and the robustness of the model, and provides higher-quality data support for aircraft structural safety assessment and fatigue analysis.
Smart Images
Figure CN121734683A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of aircraft structural load measurement technology in flight testing, specifically involving a ground calibration and flight measurement method for the vertical tail load in a T-tail wing. Background Technology
[0002] T-tail configurations are widely used in large aircraft and regional jets, with the horizontal stabilizer mounted high atop the vertical stabilizer. This configuration experiences complex aerodynamic loads under specific flight conditions, and is particularly susceptible to adverse conditions such as flutter. The periodic loads generated by flutter are transmitted through the tail structure, potentially threatening the integrity and safety of the aircraft's tail section. Therefore, accurate measurement of tail structure loads during flight testing is crucial.
[0003] However, load measurement of the T-tail structure presents significant challenges. Because the horizontal stabilizer is mounted at the tip of the vertical stabilizer, the aerodynamic forces acting on the horizontal stabilizer are transmitted to the fuselage via the vertical stabilizer. This results in the measurement signal on the vertical stabilizer being a coupled result of the horizontal stabilizer load. Specifically, when a strain gauge is installed at the root of the vertical stabilizer to measure strain, the obtained total strain includes both the strain generated by the aerodynamic forces acting on the vertical stabilizer itself (clean strain) and the strain component caused by the load transmission from the horizontal stabilizer (coupled strain). This load coupling effect makes it difficult to distinguish between the vertical stabilizer's own load and the load transmitted from the horizontal stabilizer using only the original measurement on the vertical stabilizer, thus affecting the accuracy of the vertical stabilizer load measurement.
[0004] Currently, the need for accurate measurement of vertical tail loads on T-tail fins is increasingly urgent, but existing technologies lack effective methods for decoupling vertical tail load measurement. Regarding ground calibration, there is no mature method specifically designed to distinguish and calibrate the vertical tail load after coupling with the horizontal tail load, making it impossible to directly extract the true vertical tail load from the measurement signals during flight. Secondly, when establishing the calibration equation for vertical tail load and strain, traditional multivariate regression methods often directly use all measured strain signals for fitting, failing to eliminate invalid strain signals and redundant collinear strain signals. Thirdly, in flight measurements, if the total strain at the root of the vertical tail is directly substituted into the vertical tail load calibration equation to calculate the load using the traditional method, the vertical tail coupling strain caused by the horizontal tail load is mixed in. This component does not follow the direct correspondence between the vertical tail's own load and strain, thus leading to significant deviations when using the original calibration relationship to extrapolate the vertical tail load. Summary of the Invention
[0005] Purpose of the invention: This paper presents a ground calibration and flight measurement method for vertical tail load in a T-tail wing. A clean load calibration equation for the vertical tail is established through a ground calibration test with the vertical tail loaded. Calibration equations for coupled strain and coupled load of the vertical tail are established through a ground calibration test with the horizontal tail loaded, enabling accurate differentiation between the load on the vertical tail itself and the load transferred from the horizontal tail. During calibration modeling, an elastic net sparse regression algorithm is introduced to perform feature selection and redundancy suppression on signals from multiple strain measurement points. This effectively removes invalid features with minimal impact on load estimation and alleviates modeling instability caused by multicollinearity, thereby improving the accuracy and robustness of the calibration equations. During flight, based on the ground calibration equations, coupled components in the vertical tail strain are first removed, and then the clean load and coupled load are calculated separately, ultimately achieving accurate reconstruction of the total load on the vertical tail. This method integrates decoupled calibration test design and sparse modeling strategies, improving the accuracy and reliability of flight load measurements and providing higher-quality data support for aircraft structural safety assessment and fatigue analysis.
[0006] Technical solution: A ground calibration method for vertical tail load in a T-tail fin includes: Vertical tail loading ground calibration test and horizontal tail loading ground calibration test; The vertical tail loading ground calibration test specifically includes: applying a known clean shear force perpendicular to the vertical tail surface to the vertical tail, and converting the clean shear force into a clean bending moment at the root of the vertical tail and a clean torque on the torsional rigid axis of the vertical tail. Measure the corresponding clean strain at the root of the tail under the clean load. Combine the clean strain and clean load of the tail to establish the calibration equation for the clean load of the tail: clean shear force = f1 (clean strain of tail), clean bending moment = f2 (clean strain of tail), clean torque = f3 (clean strain of tail). The horizontal tail loading ground calibration test specifically includes: applying known shear forces perpendicular to the surfaces of the left and right horizontal tails to the left and right horizontal tails respectively, and measuring the horizontal tail strain at the root of the horizontal tail and the vertical tail coupling strain at the root of the vertical tail due to the horizontal tail load. Using horizontal tail strain and vertical tail coupled strain, a vertical tail coupled strain calibration equation is established: Vertical tail coupled strain = f 4 (Horizontal Tail Strain); The horizontal tail shear load is converted into the coupled bending moment borne by the root of the vertical tail and the coupled torque borne by the torsional rigid bearing of the vertical tail. Using the horizontal tail strain and the vertical tail coupled load, the vertical tail coupled load calibration equation is established: Coupled bending moment = f 5 (flat tail strain) and coupling torque = f 6 (flat tail strain).
[0007] Furthermore, the clean load of the vertical tail refers to the condition where the vertical tail itself bears the load, while the horizontal tail does not bear the load. The clean strain of the vertical tail is the structural response of the vertical tail root caused by the clean load of the vertical tail.
[0008] Furthermore, vertical tail coupled load refers to the situation where the vertical tail itself does not bear the load, but the horizontal tail bears the load. Due to the special configuration of the T-tail, the horizontal tail load is transferred to the fuselage through the vertical tail. At this time, the load at the root of the vertical tail caused by the horizontal tail load is called the vertical tail coupled load. The vertical tail coupled strain is the structural response at the root of the vertical tail caused by the vertical tail coupled load.
[0009] Furthermore, the sum of the clean strain and the coupled strain of the vertical tail is the total strain of the vertical tail, and the sum of the clean load and the coupled load of the vertical tail is the total load of the vertical tail.
[0010] Furthermore, the vertical tail clean load calibration equation, the vertical tail coupled strain calibration equation, and the vertical tail coupled load calibration equation all exhibit a multiple linear regression relationship. In the process of identifying the multiple linear regression coefficients of the calibration equations, a machine learning algorithm for elastic net sparse regression is first used to calculate the elastic net sparse regression coefficients. Then, invalid strain features with zero elastic net sparse regression coefficients are removed, leaving the valid strain features with non-zero elastic net sparse regression coefficients. Finally, the least squares method is used to estimate the valid strain features to solve for the multiple linear regression coefficients in the calibration equations, thus obtaining the calibration equation model.
[0011] Furthermore, the objective function of the machine learning algorithm for sparse regression in elastic networks is:
[0012] in, Y For the dependent variable vector, X For all independent variable vectors, This is the sparse regression coefficient vector of the elastic network. and For regularization parameters, Let S be the sum of squared residuals from the sparse regression of the elastic network. Let L1 norm be the regression coefficient vector. The L2 norm of the regression coefficient vector; The objective function of the multiple linear regression in the calibration equation is:
[0013] in, Y For the dependent variable vector, X eff This represents the vector of effective independent variables selected through elastic net sparse regression. β eff This is the vector of multiple linear regression coefficients for the calibration equation. This is the sum of squared residuals for the calibration equation, which is a multiple linear regression.
[0014] Furthermore, it also includes: the regularization parameter in the objective function of the machine learning algorithm for sparse regression of elastic networks. l 1 and l 2. Optimize the design, specifically by evaluating different cross-validation methods. l 1 and l The calibration equation model predicts the mean square error under two parameter combinations, and selects the parameter that minimizes the mean square error. l 1 and l 2. Parameter values.
[0015] A method for flight-based measurement of vertical tail load in a T-tail fin, the method utilizing the aforementioned determined vertical tail clean load calibration equation, vertical tail coupled strain calibration equation, and vertical tail coupled load calibration equation, the measurement method comprising: Step 1: Collect the total strain of the vertical tail and the strain of the horizontal tail, and substitute the horizontal tail strain into the vertical tail coupled strain calibration equation: vertical tail coupled strain = f4(horizontal tail strain), to obtain the vertical tail coupled strain response in flight. Step 2: Eliminate the vertical tail coupled strain flight response from the total vertical tail strain flight response: Vertical tail clean strain = vertical tail total strain - vertical tail coupled strain, to obtain the vertical tail clean strain response in flight. Step 3: Substitute the clean strain of the vertical tail into the vertical tail clean load calibration equation: clean shear force = f1 (clean strain of vertical tail), clean bending moment = f2 (clean strain of vertical tail), clean torque = f3 (clean strain of vertical tail), to obtain the response of the vertical tail clean load in flight. Step 4: Substitute the horizontal tail strain into the vertical tail coupled load calibration equation: coupled bending moment = f5 (horizontal tail strain) and coupled torque = f6 (horizontal tail strain) to obtain the vertical tail coupled load response in flight; Step 5: Based on the principle of linear superposition, add the clean load of the vertical tail to the coupled load of the vertical tail: Total load of the vertical tail = Clean load of the vertical tail + Coupled load of the vertical tail, and obtain the response of the total load of the vertical tail in flight.
[0016] Furthermore, it also includes dynamically reconstructing the aircraft tail load distribution based on the total vertical tail load data obtained by linear superposition, and outputting it to the aircraft health monitoring system for real-time safety assessment; and using the decoupled load data to generate a load-time history report for predicting structural fatigue life and identifying abnormal operating conditions.
[0017] Beneficial effects: 1) This invention constructs a systematic decoupling calibration method based on separate loading test, and establishes three types of calibration equations for vertical tail clean load, vertical tail coupled strain and vertical tail coupled load respectively. For the first time, it realizes accurate modeling and separation of multi-source load in T-tail, effectively solves the load aliasing problem caused by structural coupling in traditional methods, and improves the physical accuracy of vertical tail load identification.
[0018] 2) In the process of constructing the calibration model, this invention introduces the elastic net sparse regression algorithm. By combining L1 and L2 regularization to screen and constrain high-dimensional strain features, it can suppress the fitting instability caused by multicollinearity, and automatically remove invalid or redundant signals, thereby improving the regression accuracy and generalization performance of the model and enhancing the adaptability of the method to complex load conditions.
[0019] 3) This invention establishes a complete decoupled calculation process from ground calibration to flight measurement. Based on the measured strain data in flight, it performs step-by-step decoupled calculation and linear superposition, which can dynamically reconstruct the clean load, coupled load and total load of the vertical tail, improve the real-time performance, integrity and engineering usability of load measurement, and provide reliable data support for the structural safety assessment and life prediction of T-tail aircraft. Attached Figure Description
[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. 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.
[0021] Figure 1 This is a flowchart of the ground calibration method for the vertical tail load in a T-tail fin; Figure 2 This is a flowchart of the flight test method for vertical tail load in a T-tail fin. Detailed Implementation
[0022] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0023] The features and illustrative embodiments of various aspects of the present invention will now be described in detail. Numerous specific details are set forth in the following detailed description to provide a thorough understanding of the invention. However, it will be apparent to those skilled in the art that the invention may be practiced without requiring some of these specific details. The following description of embodiments is merely intended to provide a better understanding of the invention by illustrating examples of the invention. The invention is by no means limited to any specific setups and methods set forth below, but covers any improvements, substitutions, and modifications to structures, methods, and devices without departing from the spirit of the invention. Well-known structures and techniques are not shown in the drawings and the following description to avoid unnecessarily obscuring the invention.
[0024] In the description of this invention, it should be noted that the directions or positional relationships indicated by terms such as "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer" are based on the directions or positional relationships shown in the accompanying drawings and are only for the convenience of describing and simplifying the invention, and should not be construed as limiting the invention. Furthermore, the use of ordinal numbers (e.g., "first and second," etc.) is for distinguishing objects and is not limited to this order, and should not be construed as indicating or implying relative importance.
[0025] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly, encompassing both direct connection and indirect connection via an intermediate medium. Those skilled in the art can understand the specific meaning of these terms in this invention based on the specific circumstances.
[0026] It should be noted that, unless otherwise specified, the embodiments of the present invention and the features thereof can be combined with each other, and the various embodiments can be referenced and cited in each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.
[0027] The present invention will be further described in detail below with reference to the embodiments and accompanying drawings, but the embodiments of the present invention are not limited thereto.
[0028] A ground calibration method for vertical tail load in a T-shaped tail fin, as follows: Figure 1 As shown, it includes: Vertical tail loading ground calibration test and horizontal tail loading ground calibration test; In the vertical tail loading ground calibration test, a known clean shear force perpendicular to the vertical tail surface is applied to the vertical tail to simulate the actual aerodynamic conditions the vertical tail experiences during flight. At the same time, the clean shear force is converted into a clean bending moment at the root of the vertical tail and a clean torque on the torsional rigid axis of the vertical tail. The corresponding clean strain response at the root of the vertical tail is measured under the clean load. Combining the clean strain and clean load data of the vertical tail, the vertical tail clean load calibration equation is established: clean shear force = f1 (clean strain of the vertical tail), clean bending moment = f2 (clean strain of the vertical tail), clean torque = f3 (clean strain of the vertical tail). In the horizontal stabilizer loading ground calibration test, known shear forces perpendicular to the surfaces of the left and right horizontal stabilizers were applied to the left and right horizontal stabilizers respectively to simulate the actual aerodynamic conditions experienced by the horizontal stabilizers in flight. The strain response at the root of the horizontal stabilizer was measured, as well as the coupled strain response at the root of the vertical stabilizer caused by the horizontal stabilizer load. Using the horizontal stabilizer strain and the coupled strain data of the vertical stabilizer, a calibration equation for the coupled strain of the vertical stabilizer was established: Coupled strain of the vertical stabilizer = f 4 (Horizontal Tail Strain); In the horizontal tail loading ground calibration test, the horizontal tail shear load is converted into the coupled bending moment borne by the root of the vertical tail and the coupled torque borne by the torsional rigid bearing of the vertical tail. Using the horizontal tail strain and vertical tail coupled load data, the vertical tail coupled load calibration equation is established: coupled bending moment = f5 (horizontal tail strain) and coupled torque = f6 (horizontal tail strain).
[0029] The clean load of the vertical tail refers to the condition where the vertical tail itself bears the load, while the horizontal tail does not bear the load. The clean strain of the vertical tail is the structural response of the vertical tail root caused by the clean load of the vertical tail. The coupled load of the vertical tail refers to the condition where the vertical tail itself does not bear the load, while the horizontal tail bears the load. Due to the special configuration of the T-tail, the load of the horizontal tail is transferred to the fuselage through the vertical tail. At this time, the load at the root of the vertical tail caused by the load of the horizontal tail is called the coupled load of the vertical tail. The coupled strain of the vertical tail is the structural response of the vertical tail root caused by the coupled load of the vertical tail.
[0030] According to the linear elastic theory of mechanics of materials, the clean load of the vertical tail and the clean strain of the vertical tail exhibit a multiple linear regression relationship in the vertical tail clean load calibration equation; the coupled strain of the vertical tail and the strain of the horizontal tail exhibit a multiple linear regression relationship in the vertical tail coupled strain calibration equation; and the coupled load of the vertical tail and the strain of the horizontal tail exhibit a multiple linear regression relationship. When identifying and modeling the multiple linear regression parameters of the above calibration equations, a machine learning algorithm of sparse regression of elastic networks is used to screen the strain features involved in the modeling according to their importance, removing invalid strain signals with little effect on the calibration equations and redundant strain signals with multicollinearity. Then, the effective strain features after screening are used to solve the regression coefficients using the least squares method, finally obtaining a calibration equation model with high accuracy and good generalization performance.
[0031] The machine learning algorithm for sparse regression of elastic networks incorporates both L1-norm and L2-norm penalties in its objective function, which can be expressed as:
[0032] in, Y For the dependent variable vector, X For all independent variable vectors, β This is the sparse regression coefficient vector of the elastic network. l 1 and l 2 is the regularization parameter. Let S be the sum of squared residuals from the sparse regression of the elastic network. Let L1 norm be the regression coefficient vector. Let L2 be the norm of the regression coefficient vector; when l When 2 = 0, the model degenerates into a Lasso regression with only L1 regularization; when l When λ1 = 0, it degenerates into a ridge regression with only L2 regularization; when λ1 = λ2 = 0, it becomes ordinary least squares regression without regularization constraints; through elastic net sparse regularization, some... β By shrinking the model to zero, automatic selection of strain features is achieved, while mitigating the impact of multicollinearity on the model. Therefore, compared with unregularized ordinary regression methods, elastic net sparse regression yields a sparser and more robust model, offering significant advantages in feature selection and modeling accuracy.
[0033] The objective function of the multiple linear regression in the calibration equation is:
[0034] in, Y For the dependent variable vector, X eff This represents the vector of effective independent variables selected through elastic net sparse regression. β eff This is the vector of multiple linear regression coefficients for the calibration equation. This is the sum of squared residuals for the calibration equation, which is a multiple linear regression.
[0035] In the process of sparse regression modeling of elastic networks, regularization parameters in the objective function of the machine learning algorithm for sparse regression of elastic networks are also included. l 1 and l 2. Optimize the design, specifically by evaluating different cross-validation methods. l 1 and l The calibration equation model predicts the mean square error under two parameter combinations, and selects the parameter that minimizes the mean square error. l 1 and l2. Parameter values; This optimization process ensures more accurate feature selection, effectively suppresses overfitting, and improves the stability and robustness of the calibration equation, thereby maintaining high accuracy under complex load conditions and solving the model degradation problem caused by fixed parameters in traditional regression.
[0036] A method for flight measurement of vertical tail load in a T-tail fin, as follows: Figure 2 As shown, flight response data of total vertical tail strain and horizontal tail strain were collected during the flight test. Then, the clean load, coupled load, and total load of the vertical tail were predicted sequentially according to the following steps: Step 1: Substitute the horizontal tail strain into the vertical tail coupled strain calibration equation: vertical tail coupled strain = f4(horizontal tail strain), to obtain the vertical tail coupled strain response in flight; Step 2: Eliminate the vertical tail coupled strain flight response from the total vertical tail strain flight response: Vertical tail clean strain = vertical tail total strain - vertical tail coupled strain, to obtain the vertical tail clean strain response in flight. Step 3: Substitute the clean strain of the vertical tail into the vertical tail clean load calibration equation: clean shear force = f1 (clean strain of vertical tail), clean bending moment = f2 (clean strain of vertical tail), clean torque = f3 (clean strain of vertical tail), to obtain the response of the vertical tail clean load in flight. Step 4: Substitute the horizontal tail strain into the vertical tail coupled load calibration equation: coupled bending moment = f5 (horizontal tail strain) and coupled torque = f6 (horizontal tail strain) to obtain the vertical tail coupled load response in flight; Step 5: Based on the principle of linear superposition, add the clean load of the vertical tail to the coupled load of the vertical tail: Total load of the vertical tail = Clean load of the vertical tail + Coupled load of the vertical tail, and obtain the response of the total load of the vertical tail in flight.
[0037] After calculating the total load on the vertical tail, a real-time application step is also included: based on the total load data of the vertical tail obtained by linear superposition, the load distribution of the aircraft tail is dynamically reconstructed and output to the aircraft health monitoring system for real-time safety assessment; this step uses the decoupled load data to generate a load-time history report, which is used to predict structural fatigue life and identify abnormal operating conditions (such as flutter), thereby providing immediate decision support for the flight safety of T-tail aircraft.
[0038] The above description is merely a specific embodiment of this application, but the scope of protection of this application 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 this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.
Claims
1. A ground calibration method for vertical tail load in a T-shaped tail fin, characterized in that, include: Vertical tail loading ground calibration test and horizontal tail loading ground calibration test; The vertical tail loading ground calibration test specifically includes: applying a known clean shear force perpendicular to the vertical tail surface to the vertical tail, and converting the clean shear force into a clean bending moment at the root of the vertical tail and a clean torque on the torsional rigid axis of the vertical tail. Measure the corresponding clean strain at the root of the tail under the clean load. Combine the clean strain and clean load of the tail to establish the calibration equation for the clean load of the tail: clean shear force = f1 (clean strain of tail), clean bending moment = f2 (clean strain of tail), clean torque = f3 (clean strain of tail). The horizontal tail loading ground calibration test specifically includes: applying known shear forces perpendicular to the surfaces of the left and right horizontal tails to the left and right horizontal tails respectively, and measuring the horizontal tail strain at the root of the horizontal tail and the vertical tail coupling strain at the root of the vertical tail due to the horizontal tail load. Using horizontal tail strain and vertical tail coupled strain, a vertical tail coupled strain calibration equation is established: Vertical tail coupled strain = f 4 (Horizontal Tail Strain); The horizontal tail shear load is converted into the coupled bending moment borne by the root of the vertical tail and the coupled torque borne by the torsional rigid bearing of the vertical tail. Using the horizontal tail strain and the vertical tail coupled load, the vertical tail coupled load calibration equation is established: Coupled bending moment = f 5 (flat tail strain) and coupling torque = f 6 (flat tail strain).
2. The method according to claim 1, characterized in that, The clean load of the vertical tail refers to the condition where the vertical tail itself bears the load, while the horizontal tail does not bear the load. The clean strain of the vertical tail is the structural response of the vertical tail root caused by the clean load of the vertical tail.
3. The method according to claim 2, characterized in that, Vertical tail coupled load refers to the situation where the vertical tail itself does not bear the load, but the horizontal tail bears the load. Due to the special configuration of the T-tail, the horizontal tail load is transferred to the fuselage through the vertical tail. At this time, the load caused by the horizontal tail load at the root of the vertical tail is called the vertical tail coupled load. The vertical tail coupled strain is the structural response of the vertical tail root caused by the vertical tail coupled load.
4. The method according to claim 3, characterized in that, The sum of the clean strain and the coupled strain of the vertical tail is the total strain of the vertical tail, and the sum of the clean load and the coupled load of the vertical tail is the total load of the vertical tail.
5. The method according to claim 4, characterized in that, The vertical tail clean load calibration equation, the vertical tail coupled strain calibration equation, and the vertical tail coupled load calibration equation all exhibit a multiple linear regression relationship. In the process of identifying the multiple linear regression coefficients of the calibration equation, a machine learning algorithm of elastic net sparse regression is first used to calculate the elastic net sparse regression coefficients. Then, invalid strain features with zero elastic net sparse regression coefficients are removed, leaving the valid strain features with non-zero elastic net sparse regression coefficients. Finally, the least squares method is used to estimate the valid strain features to solve for the multiple linear regression coefficients in the calibration equation, thus obtaining the calibration equation model.
6. The method according to claim 5, characterized in that, The objective function of the machine learning algorithm for sparse regression in elastic networks is: in, Y For the dependent variable vector, X For all independent variable vectors, This is the sparse regression coefficient vector of the elastic network. and For regularization parameters, Let S be the sum of squared residuals from the sparse regression of the elastic network. Let L1 norm be the regression coefficient vector. The L2 norm of the regression coefficient vector; The objective function of the multiple linear regression in the calibration equation is: in, Y For the dependent variable vector, X eff This represents the vector of effective independent variables selected through elastic net sparse regression. β eff This is the vector of multiple linear regression coefficients for the calibration equation. This is the sum of squared residuals for the calibration equation, which is a multiple linear regression.
7. The method according to claim 6, characterized in that, Also includes: The regularization parameters λ1 and λ2 in the objective function of the machine learning algorithm for sparse regression of elastic networks are optimized. Specifically, the mean square error of the calibration equation model prediction under different combinations of λ1 and λ2 parameters is evaluated by cross-validation, and the parameter values of λ1 and λ2 that minimize the mean square error are selected.
8. A method for flight-based measurement of vertical tail load in a T-tail fin, characterized in that, The method is performed using the vertical tail clean load calibration equation, the vertical tail coupled strain calibration equation, and the vertical tail coupled load calibration equation determined by any one of claims 1-7, and the experimental method includes: Step 1: Collect the total strain of the vertical tail and the strain of the horizontal tail, and substitute the horizontal tail strain into the vertical tail coupled strain calibration equation: vertical tail coupled strain = f4(horizontal tail strain), to obtain the vertical tail coupled strain response in flight. Step 2: Eliminate the vertical tail coupled strain flight response from the total vertical tail strain flight response: Vertical tail clean strain = vertical tail total strain - vertical tail coupled strain, to obtain the vertical tail clean strain response in flight. Step 3: Substitute the clean strain of the vertical tail into the vertical tail clean load calibration equation: clean shear force = f1 (clean strain of vertical tail), clean bending moment = f2 (clean strain of vertical tail), clean torque = f3 (clean strain of vertical tail), to obtain the response of the vertical tail clean load in flight. Step 4: Substitute the horizontal tail strain into the vertical tail coupled load calibration equation: coupled bending moment = f5 (horizontal tail strain) and coupled torque = f6 (horizontal tail strain) to obtain the vertical tail coupled load response in flight; Step 5: Based on the principle of linear superposition, add the clean load of the vertical tail to the coupled load of the vertical tail: Total load of the vertical tail = Clean load of the vertical tail + Coupled load of the vertical tail, and obtain the response of the total load of the vertical tail in flight.
9. The method according to claim 8, characterized in that, It also includes dynamically reconstructing the aircraft tail load distribution based on the total vertical tail load data obtained by linear superposition, and outputting it to the aircraft health monitoring system for real-time safety assessment; and using the decoupled load data to generate load-time history reports for predicting structural fatigue life and identifying abnormal operating conditions.