Aircraft design stage controllability boundary analysis and optimization method based on LADRC

Abstract

The invention discloses an aircraft design stage controllability boundary analysis and optimization method based on LADRC, and relates to the technical field of aircraft design and control. The implementation process thought of the method is mainly as follows: a closed-loop system (LADRC control law + flight dynamics model) control performance boundary is taken as a constraint boundary, which parameters of an aircraft body cause that the closed-loop system cannot meet the control performance boundary are judged through simulation, and then the parameters of the aircraft body are continuously optimized. And finally, a closed-loop system control performance boundary is met. According to the method, the'black box 'characteristic of a traditional optimization algorithm is abandoned, a control performance short board is directly converted into a design improvement direction through human experience and data-driven explicit feedback, the engineering adaptation problem of a complex algorithm is avoided, and the interpretability and engineering practicability of design iteration are ensured.

Description

Technical Field

[0001] This invention relates to the field of aircraft design and control technology, and in particular to a method for analyzing and optimizing controllability boundaries in the early stages of aircraft design. Specifically, this method deeply integrates Linear Active Disturbance Rejection Control (LADRC) technology with aircraft dynamics models. Through a simulation framework, it quantitatively evaluates the synergistic effects of aerodynamic layout, steering efficiency, and control algorithms, providing an early prediction and optimization basis for the innovative design of aircraft with unconventional layouts. Background Art

[0002] Aircraft design is a highly complex systems engineering process, with the traditional process typically following a sequential approach of "aerodynamic structure first, control system second." In the early design phase, engineers often determine aerodynamic layout and control surface parameters based on empirical formulas or simplified dynamic models, deferring control system design to the detailed design phase. This disconnect makes it difficult to align the design with control requirements. When later testing reveals insufficient control efficiency or weak anti-disturbance capabilities, the only recourse is patchwork, such as reworking to increase control surface size or restrict the flight envelope.

[0003] Take the currently mainstream PID control algorithm as an example. Although widely used in traditional aircraft, it has significant limitations. First, PID relies on precise mathematical models, making parameter adjustment time-consuming and lacking robustness in highly nonlinear scenarios such as high-angle-of-attack stalls and transonic aerodynamic parameter jumps. Second, using PID as a benchmark for evaluating controllability in the early stages of a design often forces conservative rudder designs due to the algorithm's limited anti-disturbance capabilities. For example, during the prototype phase of a certain tailless flying-wing UAV, the PID control law was unable to suppress crosswind disturbances, ultimately increasing the vertical tail area by 23%, resulting in a reduction in range and stealth performance.

[0004] In recent years, active disturbance rejection control (ADRC) technology has emerged in the field of flight control due to its "model-insensitive and disturbance-self-compensating" characteristics. However, the industry generally positions it as a post-optimization tool for control systems, rarely forming a closed loop with the overall design. This "post-control" mentality has led many innovative configurations to underestimate the control potential in the early design stages, ultimately falling into the dilemma of repeated "configuration-control" iteration. How to quantify the performance boundaries of control algorithms in the early stages of design and convert them into optimization constraints for aerodynamic and structural parameters has become a key challenge in breaking through the bottleneck of innovative aircraft design. Summary of the Invention

[0005] To address the pain point of the separation between control and body parameters in traditional aircraft design, this invention proposes a new design paradigm - "pre-positioning" linear active disturbance rejection control (LADRC) technology to the aircraft conceptual design stage. Through the deep coupling of algorithms and models, the control potential is quantified at an early stage and the aerodynamic and structural parameters are reversely optimized, thereby realizing the innovative process of "controllability-driven design".

[0006] The core of the present invention is to break the timing barrier between design and control. In traditional methods, the control algorithm is only used as a "correction tool" in the later stage, while the present invention transforms it into a "decision-making basis" in the early stage of design. Specifically, by constructing a joint simulation framework of the LADRC control law and the six-degree-of-freedom dynamic model of the aircraft, the impact of nonlinear factors such as model disturbances and command tracking on the control performance is simulated in real time, and the LADRC's extended state observer (ESO) is used to dynamically compensate for unknown disturbances. This technical path enables designers to accurately evaluate the contribution of different aerodynamic layouts or structural parameters to the controllable boundaries before the prototype is manufactured, avoiding disruptive design modifications due to insufficient control capabilities in the later stage.

[0007] Furthermore, based on the strong anti-interference characteristics of LADRC, the design phase no longer relies solely on traditional static indicators (such as stability margins). Instead, multi-dimensional indicators such as time domain response (such as dynamic response under command tracking) and anti-interference robustness are considered to construct the controllability boundary within the flight envelope. At the same time, combined with a multi-objective collaborative mechanism, the control potential is directly mapped into design constraints (such as minimum control surface area and fuselage center of gravity position), guiding the iterative direction of the design of aerodynamic shaping, control surface size, center of gravity position, etc., achieving a coordinated improvement in aerodynamic efficiency and control robustness.

[0008] This approach offers a new approach for the innovative design of unconventional aircraft. By exploiting the control algorithm's anti-disturbance potential early in the design process, it can overcome the conservative design limitations of traditional PID control, unlocking more efficient aerodynamic configurations and control surface configurations, significantly improving the aircraft's mission adaptability and design freedom.

[0009] The technical solution of the present invention:

[0010] The LADRC-based controllability boundary analysis and optimization method for aircraft design phase is mainly based on the following ideas: using the control performance boundary of the closed-loop system (LADRC control law + flight dynamics model) as the constraint boundary, through simulation to determine which parameters of the aircraft body cause the closed-loop system to fail to meet the control performance boundary, and then continuously optimize these aircraft body parameters to ultimately meet the closed-loop system control performance boundary. The specific steps are as follows:

[0011] Step 1: Build a LADRC-flight dynamics model closed-loop system model

[0012] Step 1.1: Construction of flight dynamics model

[0013] Based on the six-degree-of-freedom kinematics and dynamics equations of the aircraft, a six-degree-of-freedom nonlinear dynamics model of the aircraft is constructed. The six-degree-of-freedom nonlinear dynamics model of the aircraft includes an actuator model, an aerodynamic model, an engine model, a weight model, a landing gear model, an attitude solution model, etc.

[0014] Step 1.2: LADRC control law construction

[0015] A linear active disturbance rejection control algorithm is used to design corresponding control law models for the pitch, roll and yaw channels respectively.

[0016] The flight dynamics model and the LADRC control law together constitute the LADRC-flight dynamics model closed-loop system model.

[0017] Step 2: Multi-scenario controllability boundary assessment

[0018] 1) Perform trim and small-disturbance linearization operations on the nominal model at typical flight state points, build a linear closed-loop system, and extract the stability margin parameters at the servo:

[0019] Static boundary: the stability margin of the control channel in a linear closed-loop simulation system;

[0020] 2) Conduct closed-loop command tracking simulations for the nominal model at typical flight state points to extract command tracking dynamic response parameters:

[0021] Dynamic boundary: aircraft dynamic response characteristics and control surface deflection characteristics under nominal state command control.

[0022] 3) Based on the nominal model of typical flight conditions, perform model parameter perturbations and corresponding Monte Carlo simulations to extract the dynamic response parameters under command tracking:

[0023] Robustness bounds: aircraft dynamic response characteristics and control surface deflection characteristics under command control under model parameter perturbations.

[0024] Step 3: Explicit feedback and iterative correction of design parameters

[0025] For the aircraft body, the factors affecting controllability mainly come from two aspects: aerodynamic layout and center of gravity position. Using the controllability assessment results as input, the aerodynamic layout or center of gravity position design is optimized to guide the iterative adjustment of the body plan.

[0026] If the evaluation results show that the closed-loop system stability margin is insufficient, or the command time domain tracking dynamic characteristics are poor, it is necessary to optimize the aerodynamic layout to improve the torque curve and reduce static instability. At the same time, combined with the center of gravity-focal position, moving the rear limit of the center of gravity forward can also reduce static instability and improve the stability margin characteristics. If the stability margin indicators cannot be met after the above modifications, consider using a servo with a larger bandwidth to improve stability.

[0027] If the evaluation results show that the required rudder angle is too large when tracking under nominal or parameter perturbation commands, approaching or exceeding the physical limit of the rudder angle, the rudder surface area should be increased to improve the rudder surface effectiveness.

[0028] If the evaluation results show that when tracking commands under nominal or parameter perturbation, the center point of the rudder surface deflection range is not near zero rudder deflection or is even far away from zero rudder deflection, it means that the body torque curve deviates too much from the baseline. This characteristic can be improved by adjusting the aerodynamic layout optimization methods such as adjusting the tail installation angle.

[0029] Step 4: Multiple rounds of closed-loop verification and design locking

[0030] After each round of vehicle parameter modifications, the vehicle dynamics model is re-established, followed by closed-loop system simulation. Multi-scenario controllability boundary data is extracted and compared against optimization targets. This process is repeated until all controllability indicators meet design requirements. The final output is a vehicle design that balances control robustness and aerodynamic efficiency.

[0031] The beneficial effects of the present invention are as follows: This method abandons the "black box" characteristics of traditional optimization algorithms and directly transforms control performance shortcomings into design improvement directions through manual experience and data-driven explicit feedback. This not only avoids the engineering adaptation difficulties of complex algorithms, but also ensures the interpretability and engineering practicality of design iterations. Specifically, it is reflected in:

[0032] Deeply integrate control algorithms with aircraft design at an early stage: Breaking through the traditional "design first, control later" sequential process, Linear Active Disturbance Rejection Control (LADRC) is introduced into the aircraft concept design phase for the first time, establishing a joint simulation framework for control, overall aerodynamics, and structure. LADRC's real-time disturbance compensation capability enables early prediction of the control potential of unconventional layouts, such as those without a vertical tail, to avoid disruptive design rework later due to insufficient control capabilities.

[0033] A multi-dimensional quantitative system for dynamic controllability boundaries: This system combines traditional static indicators such as stability margins with the addition of time-domain dynamic characteristic evaluation of command tracking under nominal and model parameter perturbations to construct a dynamic and static controllability evaluation dimension.

[0034] Explicit feedback-driven closed-loop iteration mechanism: The explicit iterative path of "simulation evaluation-design feedback-parameter correction" directly maps control performance shortcomings (such as insufficient rudder effectiveness and insufficient stability margin) into optimization directions for aerodynamic / structural parameters (such as horizontal tail area and wing sweep angle); through collaborative decision-making based on human experience and data, it replaces traditional black-box optimization algorithms to improve iteration efficiency and engineering interpretability.

[0035] Expanded design freedom driven by control potential: LADRC-based interference rejection unlocks radical designs that are not possible with traditional PID control.

[0036] Miniaturization of rudder surfaces: Reduce dependence on rudder efficiency through disturbance compensation, reduce rudder surface area, and improve aerodynamic efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 Flowchart of traditional design method.

[0038] Figure 2 Flow chart of the method of the present invention.

[0039] Figure 3 It is a closed-loop simulation system for the LADRC-aircraft dynamics model.

[0040] Figure 4 Iteratively revise the closed-loop flow chart for controllability assessment.

[0041] Figure 5 It is the stability margin at the pitch channel servo.

[0042] Figure 6 is the pitch angle tracking response characteristic.

[0043] Figure 7 is the elevator deflection response characteristic.

[0044] Figure 8 This is the stability margin at the pitch channel servo (after modification).

[0045] Figure 9 Pitch angle tracking response characteristics (after modification).

[0046] Figure 10 This is the elevator deflection response characteristic (after modification). DETAILED DESCRIPTION

[0047] The specific implementation process of the present invention is divided into the following steps, which realize the quantification and optimization iteration of the controllability boundary through the dynamic interaction between control and aircraft design:

[0048] Step 1: Build a LADRC-aircraft dynamics closed-loop simulation model

[0049] Step 1.1 Construction of aircraft dynamics model

[0050] Based on classic flight dynamics modeling, the aerodynamic, engine, weight, and center of gravity data, landing gear position, and compression characteristic parameters are used as the foundation for constructing the aerodynamic, engine, weight, and center of gravity models, respectively. The actuator model and attitude solution model are then integrated to construct a six-degree-of-freedom nonlinear dynamic model for the aircraft. The aircraft's initial position, initial velocity, initial attitude, and initial attitude angular rate parameters are then set based on typical flight conditions. The model's runtime can be set to 10ms.

[0051] Step 1.2 LADRC control law embedding

[0052] According to the LADRC control law design principle, a three-axis control law is constructed, and the extended state observer (ESO) in the linear active disturbance rejection control law is used to estimate and compensate for environmental disturbances and model uncertainties in real time. The initial control parameters are set; for most projects, ω is taken o ≈(3~5)ω c ; Then the linear active disturbance rejection control law (LADRC) and the aircraft dynamics model are embedded in the closed-loop simulation system, and simulation tests are performed to ensure that the closed-loop system can complete the simulation.

[0053] Step 2: Multi-scenario controllability boundary assessment

[0054] Step 2.1 Stability Margin Assessment

[0055] Stability margin is divided into amplitude margin and phase margin, which is a concept in linear control systems.

[0056] For the nominal model, the first step is to perform small-disturbance linearization on the nonlinear six-degree-of-freedom dynamic model of the aircraft under typical flight conditions, decouple the longitudinal channel and the lateral heading channel, and obtain the fourth-order models of the longitudinal and lateral heading directions respectively.

[0057] Taking the longitudinal direction as an example, the pitch control law is connected to the servo model and the longitudinal fourth-order linear model to form a closed-loop system. The connection is disconnected at the servo, and the linear transfer function from input to output at the servo disconnection point is obtained. The amplitude margin and phase margin values ​​are obtained using the Bode diagram analysis method, as shown in the following example: Figure 5 As shown; According to the Bode diagram, the amplitude margin of the channel is 21.3dB, which meets the amplitude margin requirement of 6dB, and the phase margin is 38deg, which does not meet the requirement of 45deg.

[0058] Step 2.2 Nominal state assessment

[0059] Taking the pitch channel as an example, the typical straight and level flight is taken as the initial state, and the pitch angle control command that changes with time is given, such as Figure 6The black curve is shown, and the red curve in the figure is the pitch angle tracking curve. It can be seen from the figure that the pitch angle tracking time is about 3s, the command tracking process is smooth, the overshoot is small, and there is no steady-state error; Figure 7 As can be seen from the elevator deflection response characteristic diagram, the red curve is the rudder response process under the nominal state. It can be seen that the elevator trim rudder deflection is about 10 degrees. During the process of tracking the change of the pitch angle command, the elevator deflection fluctuates around 10 degrees.

[0060] Step 2.3 Robustness Analysis

[0061] The random perturbation model parameter method is used, and then two hundred pitch angle command tracking simulations are carried out. The pitch angle tracking characteristics and rudder deflection characteristics under parameter perturbation are as follows: Figure 6 and Figure 7 As shown in the medium gray curve, the figure shows that under model parameter perturbations, the elevator varies up and down from a baseline of 10 degrees, with the entire elevator variation range being -4° to 32°. Pitch angle tracking divergence and elevator yaw oscillation divergence also occur, indicating insufficient robustness.

[0062] Step 3: Explicit feedback and iterative correction of design parameters

[0063] According to the above evaluation process, the phase margin does not meet the 45-degree index requirement. The reason is that the aircraft has insufficient pitch stability. The static stability can be improved by modifying the aerodynamic layout or moving the rear center of gravity forward.

[0064] At the same time, according to the evaluation results, the baseline trim elevator angle is 10 degrees. After being disturbed, the negative elevator deflection angle is small and the positive elevator deflection angle is large. Since the physical elevator deflection range is -35 to 35 degrees, it can be seen that the layout design does not fully utilize the deflection effect. Therefore, it is necessary to modify the tail installation angle to modify the body torque characteristics and reduce the trim deflection.

[0065] Step 4: Closed-loop verification and solution locking

[0066] According to the evaluation results, the center of gravity position is modified and the aerodynamic layout is optimized to reduce the degree of static instability and the downward body moment. After the scheme is modified, a new round of aerodynamic data and weight center of gravity data are calculated. The flight dynamics model is rebuilt based on the new data and re-evaluated. The evaluation results are as follows: Figures 8 to 10 As shown:

[0067] from Figure 10It can be seen that the phase margin at the same servo is 52 degrees when the scheme is optimized, which meets the index requirement of 45 degrees. The reference trim elevator angle is 0 degrees. After being disturbed, the trim elevator deflection range is -10 to 22 degrees. The result has been greatly improved compared with the previous round. At the same time, the command tracking process can also ensure a smooth process, small overshoot, and no steady-state error. After the model parameters are perturbed, the same pull-off target simulation is performed. The results show that the command tracking is normal under the perturbation of the model parameters, and no divergence occurs. Therefore, the robustness meets the requirements; the modified scheme meets the controllability requirements.

[0068] If the modified plan still does not meet the indicator requirements, the plan will continue to be optimized based on the evaluation results, and then iterative evaluation will be performed until all the above indicators meet the requirements and the plan can be closed.

Claims

1. The controllability boundary analysis and optimization method of the aircraft design phase based on LADRC is characterized by: The specific steps are as follows: Step 1: Build a LADRC-flight dynamics model closed-loop system model Step 1.1: Construction of flight dynamics model Based on the six-degree-of-freedom kinematics and dynamics equations of the aircraft, a six-degree-of-freedom nonlinear dynamics model of the aircraft is constructed. The six-degree-of-freedom nonlinear dynamics model of the aircraft includes an actuator model, an aerodynamic model, an engine model, a weight model, a landing gear model, an attitude solution model, etc. Step 1.2: LADRC control law construction Adopting the linear active disturbance rejection control algorithm, corresponding control law models are designed for the three channels of pitch, roll and yaw respectively; The flight dynamics model and the LADRC control law together constitute the LADRC-flight dynamics model closed-loop system model; Step 2: Multi-scenario controllability boundary assessment 1) Perform trim and small-disturbance linearization operations on the nominal model at typical flight state points, build a linear closed-loop system, and extract the stability margin parameters at the servo: Static boundary: the stability margin of the control channel in a linear closed-loop simulation system; 2) Conduct closed-loop command tracking simulations for the nominal model at typical flight state points to extract command tracking dynamic response parameters: Dynamic boundary: aircraft dynamic response characteristics and control surface deflection characteristics under nominal state command control; 3) Based on the nominal model of typical flight conditions, perform model parameter perturbations and corresponding Monte Carlo simulations to extract the dynamic response parameters under command tracking: Robustness bounds: aircraft dynamic response characteristics and control surface deflection characteristics under command control under model parameter perturbations; Step 3: Explicit feedback and iterative correction of design parameters For the aircraft itself, the factors that affect controllability mainly come from two aspects: aerodynamic layout and center of gravity position. Using the controllability assessment results as input, the aerodynamic layout or center of gravity position design is optimized to guide the iterative adjustment of the aircraft plan. If the evaluation results show that the closed-loop system stability margin is insufficient, or the command time domain tracking dynamic characteristics are poor, then it is necessary to optimize the aerodynamic layout to improve the torque curve and reduce static instability. At the same time, combined with the center of gravity-focal position, moving the center of gravity rear limit forward can also reduce static instability and improve the stability margin characteristics. If the stability margin indicators are still not met after the above modifications, consider using a servo with a larger bandwidth to improve stability. If the evaluation results show that the required rudder angle is too large when tracking the command under nominal or parameter perturbations, approaching or exceeding the physical limit of rudder angle, the rudder surface area should be increased to improve the rudder surface effectiveness; If the evaluation results show that the center of the rudder surface deflection range is not near zero rudder deflection or is even far away from zero rudder deflection during command tracking under nominal or parameter perturbation conditions, it indicates that the body torque curve deviates too much from the baseline. This characteristic can be improved by adjusting the aerodynamic layout, such as adjusting the tail mounting angle. Step 4: Multiple rounds of closed-loop verification and design locking After each round of aircraft body parameter modification, the aircraft dynamics model is re-performed, followed by closed-loop system simulation, extracting multi-scenario controllability boundary data and comparing them with optimization targets; This process needs to be executed repeatedly until all controllability indicators meet the design requirements, and the final output is an aircraft body design plan that takes into account both control robustness and aerodynamic efficiency.

Images