Variable-camber trajectory optimization method for wing body fusion layout aircraft

Abstract

The invention provides a variable-camber trajectory optimization method for a wing body fusion layout aircraft, which combines variable-camber design and flight path planning based on the wing body fusion layout aircraft, covers the nonlinear optimal control problem of state constraint and control constraint, and improves the trajectory optimization accuracy. An optimal control problem of a continuous space is converted into nonlinear programming through a parameterization method, and a trajectory optimization problem is solved by adopting a pseudo-spectral method capable of obtaining relatively high solving precision with relatively low calculation cost in a direct method; according to the method, trajectory optimization is incorporated into multi-disciplinary optimization of the aircraft, so that subsystems are dynamically combined and matched, the influence of time history of the systems is considered, the optimal deformation trend / timetable during the whole cruise task is explored, and potential benefits of the whole flight task of the variable-camber wing body fusion layout aircraft are quantified.

Description

Technical Field

[0001] The present invention discloses a trajectory optimization method for variable curvature of a wing-body fusion layout aircraft, relates to the fields of aircraft aerodynamic design, aircraft trajectory optimization and neural network modeling, and is mainly applied to the design field of variable curvature of a wing-body fusion layout aircraft. Background Art

[0002] The conventional layout of a cylindrical fuselage plus wings has been developed for many years, but it is difficult to further improve the aerodynamic efficiency. The radical goal of green aviation "energy saving, emission reduction, and noise reduction" has prompted the comprehensive innovation of civil aviation technology. A series of unconventional layouts have become the common research direction of various countries. Among them, the Blended Wing Body (BWB) layout has attracted widespread attention due to its high cruise efficiency, light structural weight, high fuel efficiency, and low noise, and has become the mainstream solution for the next generation of subsonic civil aircraft.

[0003] Aircraft design is usually a compromise choice after considering a variety of flight conditions. For many years, people have been exploring the possibility of changing the wing shape during flight so that the aircraft always maintains good performance. Among them, the technology of continuously variable camber trailing edge has been widely studied. On the one hand, compared with other wing camber changing methods, the deformation of the trailing edge has less impact on the internal structural components of the wing and is easy to implement. On the other hand, through the optimization and comparative study of the wing with a completely deformed wing and the wing with only a deformed trailing edge, it was found that most of the improvement in aerodynamic performance can be obtained by deforming only the trailing edge of the wing.

[0004] The design of the variable curvature of the wing trailing edge enhances the aircraft's ability to adapt to various flight conditions. Combining it with the BWB layout aircraft, it is expected that the aircraft will maintain optimal flight performance throughout the entire flight range to give full play to the advantages of the high aerodynamic efficiency of the BWB layout. However, since the BWB layout aircraft cancels the traditional horizontal tail and vertical tail, its control surfaces are mainly arranged on the trailing edge of the fuselage, and the longitudinal lever arm is shorter than that of conventional layout aircraft. When the same pitch moment needs to be balanced, the BWB layout will pay a greater balance resistance penalty.

[0005] Further considering that the trailing edge control surfaces can only deflect continuously during dynamic flight, the BWB layout may not always be able to maintain the best aerodynamic shape to continuously obtain the optimal aerodynamic benefits. Therefore, further research is still needed to determine how much actual benefits the variable curvature wing can bring.

[0006] Traditional aircraft design adopts a serial design concept, where designers first select different key disciplines to design and optimize the aircraft according to the mission profile, and then optimize its flight trajectory. This approach essentially artificially separates the static and dynamic design variables that should be coupled together, ignoring the impact of the time history of each system state on the overall performance of the aircraft, and is likely to lose the overall optimal solution. Summary of the invention

[0007] In response to the above problems, considering that the overall performance of the aircraft in performing the mission is related to the mission profile and flight trajectory, a precise connection is established between the aircraft's trailing edge bending angle and the aircraft's aerodynamic characteristics. At the same time, trajectory optimization is incorporated into the aircraft's multidisciplinary optimization, so as to dynamically combine and match the subsystems and effectively improve the overall performance of the aircraft in performing the mission. This has very important significance and practical engineering value for the design of the aircraft.

[0008] Therefore, the present invention proposes a trajectory optimization method for a variable-camber aircraft. By combining deformation simulation with trajectory optimization, the influence of the time history of each system is considered, the optimal deformation trend / schedule during the entire cruise mission is explored, and the potential benefits of the entire flight mission of the variable-camber BWB layout aircraft are quantified.

[0009] In order to quantify the mission benefits that can be obtained by improving the BWB layout aircraft with variable camber technology, the present invention provides a variable camber trajectory optimization method for a wing-body blended layout aircraft, in order to provide certain technical support for the design of wing-body blended layout aircraft.

[0010] The core content of the present invention is as follows: a set of optimization methods combining variable curvature design with track planning is proposed.

[0011] Since trajectory optimization is a nonlinear optimal control problem covering state constraints and control constraints, we choose to transform the optimal control problem in continuous space into nonlinear programming through parameterization method, and then solve it by direct method combining nonlinear programming algorithm to obtain the optimal solution.

[0012] Specifically, the pseudo-spectral method is used in the direct method to solve the trajectory optimization problem, which can achieve higher solution accuracy at a lower computational cost.

[0013] To achieve the above object, the technical solution adopted by the present invention is: A method for optimizing the trajectory of a wing-body blended aircraft with variable curvature comprises the following steps: Step 1: Using the Kriging proxy model method, based on the neural network, a Kriging variable camber aerodynamic proxy model is established to establish a more accurate relationship between the wing trailing edge camber angle of the aircraft and the aerodynamic characteristics of the wing-body blended layout aircraft, including: Step 1.1: constructing an aerodynamic database of the variable curvature of the wing trailing edge of the aircraft, obtaining initial sample points, and the relationship between the input and output of the initial sample points; Step 1.2: Based on the relationship between the input and output of the initial sample point, a training set, a test set, and a validation set are allocated; Step 1.3: Based on the training set, the test set and the validation set, the neural network-based Kriging variable camber aerodynamic proxy model is trained to quickly evaluate the aerodynamic performance of the wing-body blended layout aircraft under given flight conditions and the wing trailing edge camber angle of the aircraft during trajectory optimization; Step 1.4: Check the accuracy of the Kriging variable curvature aerodynamic proxy model by cross-validation. If the accuracy does not meet the requirements, return to step 1.1; Step 2: A multidisciplinary analysis model for coupling loops in trajectory optimization is established using a parameterized method to incorporate trajectory optimization into aircraft multidisciplinary optimization, and to dynamically combine and match sub-models. The sub-models of the multidisciplinary analysis model include a flight dynamics model, an aircraft weight model, a fuel weight model, and a propulsion proxy model. Step 3: Based on the Kriging variable camber aerodynamic proxy model and the multidisciplinary analysis model, the Radau pseudo-spectral method is used to solve the cruise altitude control flight trajectory optimization, including: converting the continuous flight time interval of the wing-body fusion layout aircraft into a grid interval, solving the Radau pseudo-spectral differential matrix of the grid interval, converting the dynamic differential equation constraints of the continuous flight time optimal control problem of the wing-body fusion layout aircraft into algebraic constraints, and converting the continuous flight time optimal control problem of the wing-body fusion layout aircraft into a nonlinear programming problem; Step 4: Use optimization algorithm to solve.

[0014] In step 1, the method for obtaining the initial sample point and the relationship between the input and output of the initial sample point is: Step 1.1.1: Sampling in the design parameter space through experimental design and obtaining the initial sample points; Step 1.1.2: adopting a fixed-axis deflection method, realizing the curvature of the wing trailing edge of the aircraft through a parameterized method, and obtaining the curvature angle of the wing trailing edge of the aircraft at the initial sample point as the input of the initial sample point; Step 1.1.3: Generate a mesh after the curvature of the trailing edge of the aircraft is changed by using an inverse distance weighted moving mesh technique; Step 1.1.4: Calculate the grid of the aircraft trailing edge after the curvature changes by an aerodynamic solver to obtain the aerodynamic coefficient of the initial sample point as the output of the initial sample point.

[0015] In step 1, the expression of the Kriging variable camber aerodynamic proxy model is: , Among them, the symbol represents the proxy model, CL is the lift coefficient, CD is the drag coefficient, CM is the moment coefficient, mach is the Mach number, alt is the flight altitude, alpha is the angle of attack, and flapi is the trailing edge bending angle.

[0016] In step 2, the flight dynamics model, the aircraft weight model, the fuel weight model, and the propulsion proxy model are expressed in sequence as follows:

[0017] in, range For the voyage, alt is the flight altitude, V is the flight speed, is the trajectory angle, α is the angle of attack; L, D, T, m are lift, drag, thrust and weight, respectively. g is the acceleration due to gravity, symbol represents the time derivative of this term, is the total design weight of the aircraft, is the passenger weight, is the effective load weight, is the fuel weight, is the empty weight of the aircraft, is the fuel consumption rate per second, SFC is the unit fuel consumption rate, T max is the maximum thrust, mach is the Mach number.

[0018] In step 3, the calculation method for solving the optimization of the cruise altitude control flight trajectory using the Radau pseudo-spectral method is: Step 3.1: Convert the continuous flight time interval into the grid interval: Step 3.1.1: Set the continuous flight time interval Divide into K grid intervals, in, is the starting time of the continuous flight time interval, is the end time of the continuous flight time interval; Step 3.1.2: In each grid interval of the K grid intervals The time range can be converted into , the conversion formula is: , Wherein, k is the kth grid interval in the K grid intervals, , , is the distribution point within the grid interval; Step 3.2: Using the Radau pseudospectral method, the kth grid interval In the example, N LGR points are selected, and the N LGR points are , , , Among them, the N LGR points is a polynomial root, is an N-order orthogonal polynomial: ; Step 3.3: Use the N LGR points as nodes to construct a global interpolation polynomial of the state variable, and compensate the terminal points when interpolating the global interpolation polynomial of the state variable , the state variable is constructed using N+1 Lagrange interpolation polynomials: , in, is the state variable, is the Lagrange interpolation basis function; Step 3.4: Using the N LGR collocation points as nodes to construct a global interpolation polynomial of the control variable, the control variable is constructed using the N Lagrange interpolation polynomials as follows: , in, is the control variable; Step 3.5: Derivative the global interpolation polynomial of the state variable to obtain the derivative of the state variable: , in, is the Radau pseudospectral differential matrix at the kth grid interval, and the derivative of the state variable is discretized on the N LGR collocation points to obtain the discretized form of the state variable: ; Step 3.6: Use Lagrange integral to replace the integral term of the objective function to obtain the discretized form of the objective function: , Among them, J is the objective function, are the Lagrange polynomial coefficients; Step 3.7: Discretize the path constraint function C[S(t),U(t),t] on the N LGR points in the kth grid interval to obtain the discretized form of the path constraint function: .

[0019] In step 4, the optimization algorithm is a sequential quadratic programming (SQP) algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a schematic diagram of the process of the present invention.

[0021] Figure 2 It is a geometrical schematic diagram of the wing-body fusion layout aircraft in the present invention.

[0022] Figure 3 It is the surface pressure distribution and streamline diagram at the design point of the reference configuration in the present invention.

[0023] Figure 4 It is a schematic diagram of the division of the rudder surface of the reference configuration in the present invention.

[0024] Figure 5 It is a schematic diagram of the FFD control body of the reference configuration trailing edge deflection and the variable camber deflection effect in the present invention.

[0025] Figure 6 It is a schematic diagram of the relationship between the pseudo-spectral method and the original trajectory optimization problem in the present invention.

[0026] Figure 7 It is a schematic diagram of a characteristic curve of thrust and fuel consumption rate of a turbojet engine in the present invention.

[0027] Figure 8 It is a comparison of the trajectories before and after the change of curvature of the reference configuration during constant altitude flight in the present invention. DETAILED DESCRIPTION

[0028] The present invention will now be further described in detail in conjunction with a method embodiment. The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. The following embodiments are only descriptive and non-limiting, and cannot be used to limit the scope of protection of the present invention.

[0029] like Figure 1 As shown in the flowchart, the trajectory optimization of the wing-body blended layout aircraft using the variable camber technology consists of the following steps.

[0030] Step 1: Build an aerodynamic proxy model of the aircraft's wing trailing edge with variable camber.

[0031] The aircraft with variable camber of the wing trailing edge will adaptively change its aerodynamic shape according to different flight conditions, and the flight conditions are the state information of the aircraft, including flight altitude, flight speed, Mach number, trajectory angle, aircraft weight, etc. Therefore, in order to timely update and match the aerodynamic characteristics of the aircraft with different shapes during track analysis, the present invention adopts the Kriging proxy model method to establish the Kriging variable camber aerodynamic proxy model.

[0032] like Figure 2 The wing-body fusion layout aircraft shown in the figure has a span of 65.5 meters in the base configuration, a fuselage length of 41 meters, and a reference area of ​​849.51 square meters. Its design cruise state is: Mach number (mach) Ma = 0.84, lift coefficient C L =0.19, Reynolds number Re = 12.94 million. The torque at the design point is self-balancing, the cruise angle of attack is 3.12°, and the drag coefficient C D =90.4counts, lift-to-drag ratio is 21.02. Figure 3 The pressure cloud diagram and streamline distribution on the wing surface at the design point are shown. Figure 3 It can be seen that the pressure distribution in the transition area between the central body and the wing-body is basically in a shock wave-free state, there are weak shock waves in the outer wing section, and there is no separation in the streamlines of the entire surface.

[0033] Step 1.1: Construct an aerodynamic database of the aircraft's wing trailing edge camber.

[0034] like Figure 1 The aerodynamic database sampling process of the variable camber of the wing trailing edge shown in the left half. Through experimental design, sampling is performed in the design parameter space and the initial sample of the design parameters is obtained. Then, the variable camber of the trailing edge of the BWB configuration is realized by a parametric method using a fixed-axis deflection method, and the mesh of the wing trailing edge of the aircraft after the variable camber is automatically generated by the dynamic mesh technology, and provided to the aerodynamic solver for calculation to obtain the aerodynamic coefficient of the initial sample point.

[0035] The control surfaces used when the trailing edge of the wing of a standard shape aircraft changes camber are divided as follows Figure 4 As shown, six control surfaces are arranged along the span direction, including one control surface at the trailing edge of the central body and five control surfaces at the trailing edge of the wing. The present invention selects the deflection angle (rudder deflection angle) of each control surface to be in the range of [-4°, +4°], stipulates that the downward deflection is positive and the upward deflection is negative.

[0036] The camber of the aircraft's wing trailing edge is deflected by a fixed axis, and the parameterization of the trailing edge camber is achieved through the free-form deformation (FFD) method. The FFD method implements the geometric parameterization process based on the idea that an elastic body deforms after being subjected to force. The FFD control body is constructed through the FFD method. The constructed FFD control body and the effects before and after deflection are shown in the figure. Figure 5 As shown in Figure 2, it can be seen that the wing surface still maintains continuous and smooth characteristics after the camber is changed.

[0037] The basic idea of ​​the FFD parameterization method is to control deformation by parameterizing the deformation of the deformable object. It avoids parametric fitting of the geometric shape itself, greatly reduces the dependence on the data format and layout form of the initial configuration, and can smoothly describe the shape of curves, surfaces and three-dimensional geometric bodies, and is conveniently applied to overall and local shape design.

[0038] The mesh after the curvature of the aircraft trailing edge is automatically generated by the inverse distance weighted (IDW) dynamic mesh technology, and provided to the aerodynamic solver for calculation to obtain the aerodynamic coefficient of the initial sample point.

[0039] The aerodynamic solver uses a computational fluid dynamics (CFD) flow field simulation method based on the Reynolds-Averaged Navier-Stokes (RANS) equation, an implicit time advancement method, a Jameson-Schmidt-Turkel second-order central space discretization format, and a Shear Stress Transport turbulence model.

[0040] Step 1.2: In order to solve the problem that the nonlinearity between the shape and aerodynamic characteristics is high due to the influence of shock waves during the curvature of the aircraft's wing trailing edge, a more accurate connection between the curvature angle of the aircraft's wing trailing edge and the aerodynamic characteristics of the wing-body fusion layout aircraft is established based on the Kriging proxy model method.

[0041] Based on the relationship between the input and output of the initial sample point, the training set, the test set and the validation set are allocated. A person skilled in the art can select the allocation ratio of the training set, the test set and the validation set based on experience. For example, the allocation ratio of the training set, the test set and the validation set is 0.9:0.07:0.03. Among them, the input of the initial sample point is the bending angle of the trailing edge of the wing of the aircraft , the output is the aerodynamic coefficient.

[0042] Based on the training set, the test set and the validation set, the neural network-based Kriging variable camber aerodynamic proxy model is trained to achieve rapid evaluation of aerodynamic performance under given flight conditions and shapes during trajectory optimization. The expression of the Kriging variable camber aerodynamic proxy model is: (1), symbol Represents the proxy model, .

[0043] Among them, C L is the lift coefficient, CD is the drag coefficient, C M is the torque coefficient, mach is the Mach number, alt is the flight altitude, alpha is the angle of attack, flap i The trailing edge bending angle.

[0044] During the analysis process, it is necessary to update and match the performance of each subsystem in a timely manner, and build each subsystem model through empirical formulas or proxy model methods to improve optimization efficiency.

[0045] When solving the trajectory using the pseudo-spectral method, each system needs to provide information about the derivative of the output relative to the input. The surrogate modeling toolbox (SMT) is used to construct the Kriging variable-camber aerodynamic proxy model, so that the corresponding aerodynamic performance can be obtained according to the input rudder deflection angle combination, and the derivatives of each aerodynamic performance relative to each rudder surface can also be obtained. The SMT toolbox is used to construct the Kriging variable-camber aerodynamic proxy model to quickly obtain the maximum thrust, specific fuel consumption rate, and its derivatives relative to the altitude and Mach number at a given altitude and Mach number.

[0046] Step 1.3: Check the accuracy of the Kriging variable curvature aerodynamic proxy model by cross-validation. If the accuracy does not meet the requirements shown in Table 1, repeat steps 1.1 to 1.2, add the initial sample points, and retrain the Kriging variable curvature aerodynamic proxy model until the accuracy meets the requirements. Finally, the optimal configuration of the variable curvature deflection angle and aerodynamic coefficient is obtained.

[0047] In the present invention, the requirements that the Kriging variable camber aerodynamic proxy model needs to meet are shown in Table 1. According to the method of the present invention, the accuracy of the performance of the Kriging variable camber aerodynamic proxy model meets the requirements of Table 1, and a relatively accurate relationship between the wing trailing edge camber angle of the aircraft and the aerodynamic characteristics of the wing-body fusion configuration can be obtained.

[0048] Table 1 Accuracy requirements that the Kriging variable curvature aerodynamic proxy model must meet Mean Absolute Error (MAE) Mean relative error (MRE) <![CDATA[ C L ]]> < 5e-3 <5% <![CDATA[ C D ]]> < 2e-4 <5% <![CDATA[ C M ]]> < 1e-3 <5% The performance evaluation results of the Kriging variable curvature proxy model in the present invention are shown in Table 2, and the accuracy meets the requirements.

[0049] Table 2 Evaluation results of the Kriging variable curvature proxy model performance.

[0050] Mean Absolute Error (MAE) Mean relative error (MRE) <![CDATA[ C L ]]> 2e-3 < 5e-3 3.44% <![CDATA[ C D ]]> 1.5e-4 < 2e-4 2.91% <![CDATA[ C M ]]> 9.6e-4 < 1e-3 2.56% Step 2: Establish other multidisciplinary analysis models for coupling loops in trajectory optimization. The specific steps are as follows: Step 2.1: Build a flight dynamics model.

[0051] The present invention only considers the situation that the aircraft flies in the plumb plane, and assumes that the engine installation angle is 0 and the thrust is in the same direction as the flight trajectory. range For the voyage, alt is the flight altitude, velocity is the flight speed (abbreviated as V ), gam is the trajectory angle (abbreviated as ); alpha is the angle of attack (abbreviated as α), L, D, T, m are lift, drag, thrust and weight, respectively. g is the acceleration due to gravity, and the symbol "" represents the time derivative of this term. The invention does not consider the existence of a wind field. The mathematical expression of the unsteady motion equation of the aircraft flying in the plumb plane is: (2) Step 2.2: Build a weight model.

[0052] The aircraft design gross weight includes the passenger weight, payload weight, fuel weight, and empty weight. The equation for the aircraft design gross weight is: (3) in, is the aircraft design gross weight, is the passenger weight, is the effective load weight, is the fuel weight, It is the empty machine weight.

[0053] Among them, the aircraft design gross weight , g is the acceleration due to gravity.

[0054] Step 2.3: Build the fuel weight model.

[0055] Since the fuel weight decreases over time during flight, the fuel consumption rate per second can be obtained by multiplying the thrust by the fuel consumption per unit thrust: (4) Among them, SFC is the specific fuel consumption.

[0056] The relationship between the weight model and the fuel weight model is as follows: .

[0057] Step 2.4: Build a propulsion agent model.

[0058] Two high bypass ratio turbojet engines actually used in engineering are used, and the engines are simplified, ignoring the difference in fuel consumption and the influence of throttling coefficient at different speeds. It is assumed that the engine is always running at 95% speed during the cruise mission trajectory optimization process of the present invention. At this time, the maximum thrust T max , Unit fuel consumption rate SFC Only with height alt ,Mach number mach This paper constructs the propulsion agent model based on the engine speed characteristic parameter table: (5) in, ，h The time unit is expressed in hours.

[0059] The characteristic curve of single engine thrust and fuel consumption rate obtained based on the propulsion proxy model is as follows: Figure 7 shown.

[0060] In this step, the equations of the flight dynamics model provide the relationships between parameters such as the aircraft's range, flight altitude, flight speed, and aircraft weight. The fuel weight in the aircraft weight model is constrained by the fuel weight model. The propulsion proxy model provides the constraint relationship between thrust, Mach number, and flight altitude, as well as the constraint relationship between fuel consumption rate and Mach number and flight altitude.

[0061] In addition, technicians in this field are aware that the flight altitude, flight speed, and Mach number of an aircraft satisfy certain constraints. At the same speed, the Mach number of an aircraft will vary due to the different speeds of sound in the air at the altitude; the higher the altitude, the lower the speed of sound, which results in a higher Mach number, so the speed of high-altitude flight will be reduced to avoid impact.

[0062] Therefore, there is a parameter constraint relationship between the Kriging variable camber aerodynamic proxy model in step 1 and the multidisciplinary analysis model in step 2.

[0063] The present invention incorporates trajectory optimization into the multidisciplinary optimization of aircraft by establishing a multidisciplinary analysis model, and transforms the optimal control problem in continuous space into nonlinear programming through a parameterization method.

[0064] Step 3: Pseudospectral method is used to solve the optimization of cruise altitude control flight trajectory.

[0065] In the present invention, reference Figure 1After establishing the Kriging variable camber aerodynamic proxy model and the multidisciplinary analysis model in step 1 and step 2, the objective function value of the aircraft performance is solved by the method of step 3 based on the dynamically generated Kriging variable camber aerodynamic proxy model and the multidisciplinary analysis model, and the dynamic differential equation constraints of the continuous time optimal control problem are converted into algebraic constraints. The potential benefits of the entire flight mission of the variable camber BWB layout aircraft are quantified by considering the influence of the time history of each aircraft system.

[0066] In the present invention, there is no requirement for order limitation in establishing the Kriging variable camber aerodynamic proxy model and the multidisciplinary analysis model in step 1 and step 2, and step 3 performs trajectory optimization based on the models in step 1 and step 2. At the same time, according to the flight state of the aircraft, the parameter values ​​of the models in step 1 and step 2 are dynamically changed, and step 3 performs dynamic trajectory optimization based on the changed Kriging variable camber aerodynamic proxy model and the multidisciplinary analysis model.

[0067] In the present invention, the aircraft cruise mission is set to fly 15000km at a fixed altitude of 11.5km. This case is used to study the performance of the aircraft when flying at the designed altitude. The initial Mach number is set to Ma=0.84, and the Mach number constraint range is Ma=0.8~0.86. The optimization problem is shown in Table 3, where the reference configuration only uses the central body trailing edge bending angle flap 0 Balance, while in the variable camber configuration all the rudder surfaces on the trailing edge participate in the balance.

[0068] Table 3 Optimization problem of variable curvature trajectory of fixed-altitude flight

[0069] The specific steps are as follows: Step 3.1: Time interval conversion.

[0070] The continuous flight time interval of the wing-body fusion layout aircraft Divide into K grid intervals.

[0071] in, is the starting time of the continuous flight time interval, is the end time of the continuous flight time interval.

[0072] In the kth grid interval The time range can be converted into ,in, , k is the kth grid interval in the K grid intervals, is the distribution point within the grid interval.

[0073] The conversion formula is: (6) Step 3.2: Discretize state variables and control variables.

[0074] Using the Radau pseudospectral method, the kth grid interval interval In the example, the state variables and control variables are discretized at Legendre Gauss-Radau (LGR) points. Select N LGR points ,in, , is an N-order orthogonal polynomial: (7) The global interpolation polynomials of the state variables and the control variables are constructed with these discrete collocation points as nodes. Since the LGR collocation points include the starting point but not the end point, in order to ensure the complete mapping between the physical continuous time interval and the time interval in the pseudospectral method, the terminal collocation points are compensated when constructing the global interpolation polynomials of the state variables. , so there is no need to consider the terminal state integral constraint. Therefore, the state variable is constructed using N+1 Lagrange interpolation polynomials as follows: (8) in, is the state variable, is the Lagrange interpolation basis function.

[0075] The dynamic equation is only valid on the LGR collocation points, so the interpolation nodes of the control variables only include the N LGR collocation points. The control variables are constructed using N Lagrange interpolation polynomials as follows: (9) in, is the control variable, is the Lagrange interpolation basis function.

[0076] Step 3.3: Discretize the state variables. Since the state variables are approximated by the N+1 Lagrange interpolation polynomials in the pseudo-spectral method, the derivatives of the state variables can be obtained by taking the derivatives of the N LGR collocation points: (10) in is the Radau pseudospectral differential matrix at the kth grid interval. Discretizing the above formula at the N LGR collocation points yields: (11) Step 3.4: Discretize the objective function and the path constraint function. By replacing the integral term of the objective function with the Lagrange integral, the objective function can be discretized into: (12) in, J is the objective function, are the Lagrange polynomial coefficients.

[0077] The path constraint function C[S(t),U(t),t] is discretized at the N LGR points in the kth grid interval: (13) The present invention utilizes differential matrices to transform the dynamic differential equation constraints of the continuous flight time optimal control problem of the wing-body fusion layout aircraft into algebraic constraints, and transforms the BWB aircraft continuous time optimal control problem into a nonlinear programming problem.

[0078] Step 4: Finally, the optimization algorithm is used to solve the problem.

[0079] The present invention can be programmed in a MATLAB environment and solved in combination with the SNOPT toolkit, or can be solved using the SNOPT toolkit through a Python software package based on other compilers. The SNOPT is a toolkit for quickly solving large-scale nonlinear optimization problems, and its main core algorithm is the sequential quadratic programming (SQP) algorithm. The optimization algorithm in the present invention is not limited to the sequential quadratic programming (SQP) algorithm, and personnel in the relevant technical field can select an algorithm for solving nonlinear programming problems as needed, such as the interior point method.

[0080] The results of the fixed altitude flight are compared in Figure 8 Compared with the 79.49kg consumed by the baseline configuration flight, the variable camber configuration only consumed 77.45kg of fuel, saving 2.6% of fuel consumption. After adding the variable camber technology, the thrust required for the entire journey is reduced, and the altitude and speed changes are smoother, which improves the stability of the flight. Overall, the variable camber technology configuration has a higher lift-to-drag ratio for the entire journey. After adding the variable camber technology, the overall performance of the aircraft in performing the mission is significantly improved, and it can be considered that this improvement only includes the improvement of the aerodynamic performance by the trailing edge variable camber technology.

[0081] To sum up, in the variable curvature trajectory optimization method for a wing-body fusion layout aircraft proposed in the present invention, the optimization results of the embodiments show that the method of the present invention achieves the design intent, combines the variable curvature design of the wing-body fusion layout aircraft with the trajectory planning, and intuitively and clearly quantifies the potential benefits that can be brought by using the variable curvature technology on the wing-body fusion layout aircraft.

Claims

1. A method for optimizing the trajectory of a wing-body fusion layout aircraft with variable curvature, characterized in that: The steps are as follows: Step 1: Using the Kriging proxy model method, based on the neural network, a Kriging variable camber aerodynamic proxy model is established to establish a more accurate relationship between the wing trailing edge camber angle of the aircraft and the aerodynamic characteristics of the wing-body blended layout aircraft, including: Step 1.1: constructing an aerodynamic database of the variable curvature of the wing trailing edge of the aircraft, obtaining initial sample points, and the relationship between the input and output of the initial sample points; Step 1.2: Based on the relationship between the input and output of the initial sample point, a training set, a test set, and a validation set are allocated; Step 1.3: Based on the training set, the test set and the validation set, the neural network-based Kriging variable camber aerodynamic proxy model is trained to quickly evaluate the aerodynamic performance of the wing-body blended layout aircraft under given flight conditions and the wing trailing edge camber angle of the aircraft during trajectory optimization; Step 1.4: Check the accuracy of the Kriging variable curvature aerodynamic proxy model by cross-validation. If the accuracy does not meet the requirements, return to step 1.1; Step 2: A multidisciplinary analysis model for coupling loops in trajectory optimization is established using a parameterized method to incorporate trajectory optimization into aircraft multidisciplinary optimization, and to dynamically combine and match sub-models. The sub-models of the multidisciplinary analysis model include a flight dynamics model, an aircraft weight model, a fuel weight model, and a propulsion proxy model. Step 3: Based on the Kriging variable camber aerodynamic proxy model and the multidisciplinary analysis model, the Radau pseudo-spectral method is used to solve the cruise altitude control flight trajectory optimization, including: converting the continuous flight time interval of the wing-body fusion layout aircraft into a grid interval, solving the Radau pseudo-spectral differential matrix of the grid interval, converting the dynamic differential equation constraints of the continuous flight time optimal control problem of the wing-body fusion layout aircraft into algebraic constraints, and converting the continuous flight time optimal control problem of the wing-body fusion layout aircraft into a nonlinear programming problem; Step 4: Use optimization algorithm to solve.

2. The trajectory optimization method for variable curvature of a wing-body blended layout aircraft according to claim 1, characterized in that: The method for obtaining the initial sample point and the relationship between the input and output of the initial sample point is: Step 1.1.1: Sampling in the design parameter space through experimental design and obtaining the initial sample points; Step 1.1.2: adopting a fixed-axis deflection method, realizing the curvature of the wing trailing edge of the aircraft through a parameterized method, and obtaining the curvature angle of the wing trailing edge of the aircraft at the initial sample point as the input of the initial sample point; Step 1.1.3: Generate a mesh after the curvature of the trailing edge of the aircraft is changed by using an inverse distance weighted moving mesh technique; Step 1.1.4: Calculate the grid of the aircraft trailing edge after the curvature changes by an aerodynamic solver to obtain the aerodynamic coefficient of the initial sample point as the output of the initial sample point.

3. The trajectory optimization method for variable curvature of a wing-body blended layout aircraft according to claim 1, characterized in that: The expression of the Kriging variable camber aerodynamic proxy model is: Among them, the symbol Represents the proxy model, , C L is the lift coefficient, C D is the drag coefficient, C M is the torque coefficient, mach is the Mach number, alt is the flight altitude, alpha is the angle of attack, is the bending angle of the trailing edge of the wing of the aircraft.

4. The method for optimizing the trajectory of a wing-body blended aircraft with variable curvature according to claim 1, characterized in that: The flight dynamics model, the aircraft weight model, the fuel weight model, and the propulsion proxy model are expressed in sequence as follows: ， in, range For the voyage, alt is the flight altitude, V is the flight speed, is the trajectory angle, α is the angle of attack; L, D, T, m are lift, drag, thrust and weight, respectively. g is the acceleration due to gravity, symbol represents the time derivative of this term, is the total design weight of the aircraft, The weight of the occupant, is the effective load weight, is the fuel weight, is the empty weight of the aircraft, is the fuel consumption rate per second, SFC is the unit fuel consumption rate, T max is the maximum thrust, mach is the Mach number.

5. The method for optimizing the trajectory of a wing-body blended aircraft with variable curvature according to claim 1, characterized in that: The calculation method for solving the optimization of the cruise altitude control flight trajectory using the Radau pseudo-spectral method is: Step 3.1: Convert the continuous flight time interval into the grid interval: Step 3.1.1: Divide the continuous flight time interval into K grid intervals, in, is the starting time of the continuous flight time interval, is the end time of the continuous flight time interval; Step 3.1.2: In each grid interval of the K grid intervals The time range can be converted into , the conversion formula is: ， Wherein, k is the kth grid interval in the K grid intervals, , , is the distribution point within the grid interval; Step 3.2: Using the Radau pseudospectral method, the kth grid interval In the example, N LGR points are selected, and the N LGR points are , Among them, the N LGR points is a polynomial The root, is an N-order orthogonal polynomial: ； Step 3.3: Use the N LGR points as nodes to construct a global interpolation polynomial of the state variable, and compensate the terminal points when interpolating the global interpolation polynomial of the state variable , the state variable is constructed using N+1 Lagrange interpolation polynomials: ， in, is the state variable, is the Lagrange interpolation basis function; Step 3.4: Use the N LGR collocation points as nodes to construct a global interpolation polynomial of the control variable. The control variable is constructed using the N Lagrange interpolation polynomials as follows: ， in, is the control variable; Step 3.5: Derivative the global interpolation polynomial of the state variable to obtain the derivative of the state variable: ， in, is the Radau pseudospectral differential matrix at the kth grid interval, and the derivative of the state variable is discretized on the N LGR collocation points to obtain the discretized form of the state variable: ； Step 3.6: Use Lagrange integral to replace the integral term of the objective function to obtain the discretized form of the objective function: ， in, J is the objective function, are the Lagrange polynomial coefficients; Step 3.7: Discretize the path constraint function C[S(t),U(t),t] on the N LGR points in the kth grid interval to obtain the discretized form of the path constraint function: 。 6. The method for optimizing the trajectory of a wing-body blended aircraft with variable curvature according to claim 1, characterized in that: The optimization algorithm is a sequential quadratic programming (SQP) algorithm.

Images