A method for predicting the aerodynamic forces of the rotor and fuselage of a tiltrotor aircraft

Abstract

This invention belongs to the field of unmanned aerial vehicle aerodynamics, specifically a method for predicting the aerodynamic forces of the rotor and fuselage of a tiltrotor aircraft. The method includes the following steps: First, key parameter variables and their ranges of variation are determined; then, a sampling space is established using the Latin hyperisotropic method; CFD numerical simulation is performed on this sampling space to calculate the flow field; a reduced-order flow field model is established using the POD intrinsic orthogonal decomposition method; and finally, the key parameters are transformed to solve for the fuselage downward load and rotor thrust. The intrinsic orthogonal decomposition method in this invention can extract a reduced-order model containing only key features from high-order complex flows, significantly reducing the computational load during the design process and greatly improving computational efficiency under varying parameters.

Description

Technical Field

[0001] This invention belongs to the field of unmanned aerial vehicle aerodynamics, specifically a method for predicting the aerodynamic forces of the rotor and fuselage of a tiltrotor aircraft. Background Technology

[0002] Traditional fixed-wing aircraft are characterized by high flight speed, large payload, and wide adaptability, but they have high site requirements and require well-constructed runways, making them unsuitable for mission deployment in relatively complex environments. Rotary-wing aircraft (helicopters, multi-rotor aircraft), on the other hand, are more suitable for small-scale, high-maneuverability operations.

[0003] Tiltrotors are a new type of rotorcraft that falls between helicopters and fixed-wing aircraft. They combine the advantages of both rotorcraft and propeller aircraft, giving them not only a much higher forward speed than conventional rotorcraft, but also vertical takeoff and landing and hovering capabilities that propeller aircraft lack. This allows them to meet the needs of various flight missions, greatly expanding the flight envelope of rotorcraft and fixed-wing aircraft, and making them extremely versatile.

[0004] Currently, relatively mature tiltrotor aircraft include the XV-15, BA-609, V-22, and V-280 tiltrotors. Among them, the V-280, jointly developed by Bell Helicopter and Lockheed Martin, is the latest technology demonstrator. Its principle primarily involves tilting the rotor nacelles at both ends of the tilting wing to achieve the conversion between helicopter and fixed-wing modes. In helicopter mode, the two nacelles are perpendicular to the ground, and the rotor plane is parallel to the ground. During helicopter mode flight, the two counter-rotating rotor systems maintain attitude stability. After climbing to a certain altitude in helicopter mode, the nacelles tilt forward, transitioning to high-speed level flight in fixed-wing mode. Through this method, the V-280 combines the characteristics of vertical takeoff and landing, hovering, and high-speed forward flight.

[0005] However, during the design process of tiltrotor aircraft, the influence of rotor hovering downward load needs to be considered. When the rotor radius, collective pitch, rotor center distance, and rotor center height are changed, the actual loads acting on the fuselage will change. Moreover, the parameters change very rapidly during the design phase, and conventional CFD simulation calculations are difficult to keep up with the actual optimization design iterations. Therefore, it is necessary to design a method to predict the rotor and fuselage aerodynamic forces of tiltrotor aircraft. Summary of the Invention

[0006] To address the aforementioned problems, the present invention aims to provide a method for predicting rotor thrust, power, and fuselage downward aerodynamic force under varying rotor parameters. By conducting a certain amount of simulation calculations, a reduced-order aerodynamic model is established to predict rotor thrust and fuselage downward load under different parameters, thereby assisting in the overall parameter design of tiltrotor aircraft and improving the speed and accuracy of overall parameter iteration.

[0007] The technical solution adopted by the present invention to achieve the above-mentioned objective is: a method for predicting the aerodynamic forces of the rotor and fuselage of a tiltrotor aircraft, comprising the following steps:

[0008] 1) Obtain the key parameter variables of the tiltrotor aircraft, and set the variation range of each key parameter variable based on the constraints formed by the configuration of the key parameter variables;

[0009] 2) Using simulation software, based on the key parameter variables of the tiltrotor aircraft and their variation range, the Latin hypersolution method is used to establish a sampling space based on the key parameter variables;

[0010] 3) Using fluid simulation software and CFD simulation method, obtain the flow field aerodynamic data of the flow field in the sampling space;

[0011] 4) Using simulation software and based on the aerodynamic data of the flow field, a flow field reduction model in the sampling space is established using the POD intrinsic orthogonal decomposition method;

[0012] 5) Based on the flow field reduction model, the velocity field and pressure field of the flow field are obtained, and the fuselage downward load and rotor thrust of the tiltrotor are obtained by integral calculation.

[0013] The key parameter variables include: rotor height, rotor spacing, rotor radius, and collective rotor pitch.

[0014] Step 2) includes the following steps:

[0015] 2-1) Using the range of change corresponding to each key parameter variable as the dividing target, divide the range of change corresponding to each key parameter variable into m non-overlapping intervals, so that each interval has the same probability;

[0016] 2-2) Randomly select a point in each interval of each key parameter variable;

[0017] 2-3) Map all randomly selected points to normally distributed sample points using the inverse function of the standard normal distribution;

[0018] 2-4) Then, randomly extract the sample points mapped in step 2-3) from each key parameter variable and form a sample vector; arrange the elements of each column of the sample vector in ascending order to form an X-dimensional sample space, i.e., the sampling space, composed of n key parameter variables.

[0019] Step 4) includes the following steps:

[0020] 4-1) Establish the optimization objective of the intrinsic orthogonal decomposition method, i.e., the orthogonal basis functions;

[0021] 4-2) Obtain the extreme values ​​of the optimization objective using the Lagrange multiplier method, and solve for the eigenvalues ​​and eigenvectors of the intrinsic orthogonal decomposition kernel function Z using the SVD decomposition method;

[0022] 4-3) The energy share method is used to screen eigenvalues ​​and eigenvectors and form an orthogonal basis, that is, a flow field price reduction model in the sampling space.

[0023] Step 4-1) specifically involves:

[0024] Based on the instantaneous aerodynamic data u extracted from the flow field in the sampling space defined by CFD numerical simulation method. i Where i = 1…N; find an orthogonal basis function on the objective function H of the intrinsic orthogonal decomposition. To maximize H, that is:

[0025]

[0026] Where u i exist The direction with the greatest, The value of is used as the optimal orthogonal basis for the orthogonal basis functions in the eigenorthogonal decomposition. u i For N instantaneous extractions of flow field aerodynamic data, This represents the inner product of the sampling space based on key parameter variables. It is a vector The 2-norm.

[0027] Step 4-2) specifically involves:

[0028] (1) Obtain the extreme value of the optimization objective using the Lagrange multiplier method, i.e.:

[0029]

[0030] (2) Let U={u1,u2,u3,...,u N Taking the partial derivative of formula (2), we get:

[0031]

[0032] (3) Obtain Let Z = UU, the extreme value of Z. T Soon to be sought The extreme values ​​are transformed into obtaining the eigenvalues ​​and eigenvectors of the POD kernel function Z, that is:

[0033]

[0034] (4) Decompose formula (4) using the SVD decomposition method, according to Z = UU T And Z=U T U, taking Z, yields:

[0035] U=ΛΣV

[0036] Where Λ represents an orthogonal matrix, Σ represents a diagonal matrix, and V represents an orthogonal matrix; the orthogonal matrices Λ and Σ contain the left and right singular vectors of matrix U, respectively, and the orthogonal matrix Λ is related to Z = UU. T The eigenvectors are equal, and the orthogonal matrices V and Z = U are equal. T The eigenvectors of U are equal, and the diagonal matrix Σ contains the eigenvalues ​​λ on the diagonal. i i = 1, 2, 3, ... N;

[0037] Based on the decomposed matrix, obtain the eigenvectors and eigenvalues ​​λ of Z. i .

[0038] Step 4-3) specifically involves:

[0039] a. Using the energy fraction method, the corresponding eigenvalues ​​λ are obtained through the solution. i The energy information contained therein, i.e. the magnitude of the eigenvalues, is used to characterize the data space of the entire model;

[0040] b. The eigenvalue λ i The energy occupied is used as the basis for sorting, that is, sorting is based on the size of the eigenvalues;

[0041] c. Take the eigenvalues ​​and eigenvectors of the first R orders, and determine the proportion I(r) of the first R order eigenvalues ​​in the total energy of the Nth order, i.e.:

[0042]

[0043] The closer the value of I(r) is to 1, the more complete the energy information contained in the eigenvector corresponding to the eigenvalue, and the stronger the ability of the reconstructed flow field reduction model to restore the flow field; for example, if it is greater than 0.99, then execute step d; if it is less than 0.9, then return to step c and reselect the eigenvalue and eigenvector.

[0044] d. The eigenvectors corresponding to the first R eigenvalues ​​obtained are used to form the basis of the final eigenorthogonal decomposition. As a flow field reduction model in the sampling space, the key parameter variables are input and the output is the flow field vector composed of orthogonal bases. Changing the key parameter variables changes the flow field in the sampling space.

[0045] The present invention has the following beneficial effects and advantages:

[0046] 1. The aerodynamic force prediction method of the present invention adopts the POD reduced-order model, which retains the accuracy of the CFD calculation method and improves the efficiency of aerodynamic force prediction under parameter changes.

[0047] 2. This invention employs the Latin hypercube method to minimize the sampling space, reduce the computational load during model building, and improve the computational efficiency of model generation.

[0048] 3. The Latin hypercube sampling method in this invention has the characteristic of uniform stratification, which can obtain more uniformly distributed sample values ​​with fewer samples, which is beneficial to the realization of the intrinsic orthogonal decomposition method.

[0049] 4. The intrinsic orthogonal decomposition method in this invention can extract a reduced-order model containing only key features from high-order complex flows, which greatly reduces the amount of computation generated in the design process and can greatly improve the computational efficiency under parameter changes.

[0050] 5. The energy share method in this invention can effectively screen out orthogonal basis vectors that meet the calculation or usage requirements by screening energy information in a quota manner, for subsequent calculations. Attached Figure Description

[0051] Figure 1 This is a schematic diagram of the tilt rotor mechanism in this embodiment;

[0052] Figure 2 Schematic diagram of key parameters of the present invention;

[0053] Figure 3 This is a schematic diagram of the Latin hypercube sampling method.

[0054] Figure 4 Flowchart of the POD intrinsic orthogonal decomposition method of the present invention. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0056] like Figure 1The diagram shown is of the tiltrotor mechanism for which the aerodynamic forces of the rotor and fuselage need to be calculated in this embodiment. This embodiment can also be replaced by the XV-15 tiltrotor aircraft, BA-609 tiltrotor aircraft, V-22 tiltrotor aircraft, V-280 tiltrotor aircraft, etc.

[0057] The specific method flow of this invention is as follows: Figure 4 The diagram shown is a flowchart of the POD intrinsic orthogonal decomposition of the present invention. The order reduction model method of the present invention first needs to determine the key parameter variables and their range of variation; then, the Latin hypersolution method is used to establish the sampling space of the sample; on the above sample space, CFD numerical simulation is performed to calculate and solve the flow field on the sampling space; the flow field order reduction model is established by the POD intrinsic orthogonal decomposition method; the key parameters are transformed to solve the fuselage downward load and rotor thrust.

[0058] like Figure 4 The diagram shown is a flowchart of the POD intrinsic orthogonal decomposition method of the present invention, wherein the present invention specifically includes the following steps:

[0059] 1) Obtain the key parameter variables of the tiltrotor aircraft, and set the variation range of each key parameter variable based on the constraints formed by the configuration of the key parameter variables;

[0060] Among them, such as Figure 2 The diagram shows the key parameter variables of a tiltrotor aircraft. These key parameters include: rotor height, rotor pitch, rotor radius, and collective rotor pitch. The range of variation represents the range within which these key parameter variables can be set according to the mechanical structure configuration of the tiltrotor aircraft.

[0061] 2) Using simulation software, based on the key parameter variables and their variation ranges of the tiltrotor aircraft, a sampling space based on the key parameter variables is established using the Latin hypersolution method; specifically, this includes the following steps:

[0062] 2-1) Using the range of change corresponding to each key parameter variable as the dividing target, divide the range of change corresponding to each key parameter variable into m non-overlapping intervals, so that each interval has the same probability;

[0063] 2-2) Randomly select a point in each interval of each key parameter variable;

[0064] 2-3) Map all randomly selected points to normally distributed sample points using the inverse function of the standard normal distribution;

[0065] 2-4) Randomly extract the mapped sample points from each key parameter variable in step 2-3) and form a sample vector; arrange the elements of each column of the sample vector in ascending order to form an X-dimensional sample space composed of n key parameter variables, i.e., the sampling space, as shown in the figure. Figure 3 As shown.

[0066] 3) Using fluid simulation software and CFD simulation method, obtain the flow field aerodynamic data of the flow field in the sampling space;

[0067] 4) Using simulation software and based on the aerodynamic data of the flow field, a flow field reduction model in the sampling space is established using the POD intrinsic orthogonal decomposition method;

[0068] The principle of the order reduction model method of this invention, namely the intrinsic orthogonal decomposition (POD), is to project a high-dimensional vector onto a low-order vector space through a set of optimal orthogonal bases, preserving the main features and essentially reconstructing the corresponding original high-order vector. The specific steps are as follows:

[0069] 4-1) Establish the optimization objective of the intrinsic orthogonal decomposition method, i.e., the orthogonal basis functions;

[0070] 4-2) Obtain the extreme values ​​of the optimization objective using the Lagrange multiplier method, and solve for the eigenvalues ​​and eigenvectors of the intrinsic orthogonal decomposition kernel function Z using the SVD decomposition method;

[0071] 4-3) The energy share method is used to screen eigenvalues ​​and eigenvectors and form an orthogonal basis, that is, a flow field price reduction model in the sampling space.

[0072] Step 4-1), specifically:

[0073] Based on the instantaneous aerodynamic data u extracted from the flow field in the sampling space defined by CFD numerical simulation method. i , where i = 1…N; represents N instantaneously extracted aerodynamic data points, determined through numerical simulation. The goal of POD is to find an orthogonal basis function that maximizes H, i.e.:

[0074]

[0075] Where u i exist The direction with the greatest, The value of is used as the optimal orthogonal basis for the orthogonal basis functions in the eigenorthogonal decomposition. u i For N instantaneous extractions of flow field aerodynamic data, This represents the inner product of the sampling space based on key parameter variables. It is a vector The 2-norm.

[0076] Step 4-2), specifically:

[0077] (1) Obtain the extreme value of the optimization objective using the Lagrange multiplier method, i.e.:

[0078]

[0079] (2) Let U={u1,u2,u3,...,u N Taking the partial derivative of formula (2), we get:

[0080]

[0081] (3) Obtain Let Z = UU, the extreme value of Z. T Soon to be sought The extreme values ​​are transformed into obtaining the eigenvalues ​​and eigenvectors of the POD kernel function Z, that is:

[0082]

[0083] (4) Decompose formula (4) using the SVD decomposition method, according to Z = UU T And Z=U T U, taking Z, yields:

[0084] U=ΛΣV

[0085] Where Λ represents an orthogonal matrix, Σ represents a diagonal matrix, and V represents an orthogonal matrix; the orthogonal matrices Λ and Σ contain the left and right singular vectors of matrix U, respectively, and the orthogonal matrix Λ is related to Z = UU. T The eigenvectors are equal, and the orthogonal matrices V and Z = U are equal. T The eigenvectors of U are equal, and the diagonal matrix Σ contains the eigenvalues ​​λ on the diagonal. i i = 1, 2, 3, ... N;

[0086] Based on the decomposed matrix, obtain the eigenvectors and eigenvalues ​​λ of Z. i .

[0087] Step 4-3), specifically:

[0088] a. At this point, the order of the POD basis corresponding to the eigenvalues ​​is still very high. Using the energy fraction method, the one-to-one corresponding eigenvalues ​​λ are obtained through the solution. i The energy information contained therein, i.e. the magnitude of the eigenvalues, is used to characterize the data space of the entire model;

[0089] b. The eigenvalue λ i The energy occupied is used as the basis for sorting, that is, sorting is based on the size of the eigenvalues;

[0090] c. Take the eigenvalues ​​and eigenvectors of the first R orders, and determine the proportion I(r) of the first R order eigenvalues ​​in the total energy of the Nth order, i.e.:

[0091]

[0092] Where I(r) represents the proportion of the total energy to the first R eigenvalues. The closer I(r) is to 1, the more complete the information contained in the eigenvector. In this embodiment, the value is 0.99. After executing step d, the eigenvectors corresponding to the first R eigenvalues ​​are the final POD basis.

[0093] d. The eigenvectors corresponding to the first R eigenvalues ​​obtained are used to form the basis of the final eigenorthogonal decomposition. As a flow field reduction model in the sampling space, the key parameter variables are input and the output is the flow field vector composed of orthogonal bases. Changing the key parameter variables changes the flow field in the sampling space.

[0094] 5) Based on the flow field reduction model, the velocity field and pressure field of the flow field are obtained, and the fuselage downward load and rotor thrust of the tiltrotor are obtained by integral calculation.

[0095] The reduced-order model method of this invention uses the basis vectors established by the intrinsic orthogonal decomposition method to complete the construction of the reduced-order flow field model. Then, through the reduced-order flow field model, the rotor thrust and fuselage downward load under different parameters can be quickly predicted.

[0096] The above description is merely an embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, extensions, etc., made within the spirit and principles of the present invention are included within the scope of protection of the present invention.

Claims

1. A method for predicting the aerodynamic forces of the rotor and fuselage of a tiltrotor aircraft, characterized in that, Includes the following steps: 1) Obtain the key parameter variables of the tiltrotor aircraft, and set the variation range of each key parameter variable based on the constraints formed by the configuration of the key parameter variables; The key parameter variables include: rotor height, rotor spacing, rotor radius, and rotor collective pitch; 2) Using simulation software, based on the key parameter variables of the tiltrotor aircraft and their variation range, the Latin hypersolution method is used to establish a sampling space based on the key parameter variables; 3) Using fluid simulation software and CFD simulation method, obtain the flow field aerodynamic data of the flow field in the sampling space; 4) Using simulation software and based on the aerodynamic data of the flow field, a flow field reduction model is established in the sampling space using the POD intrinsic orthogonal decomposition method; Step 4) includes the following steps: 4-1) Establish the optimization objective of the intrinsic orthogonal decomposition method, i.e., the orthogonal basis functions; 4-2) Obtain the extreme values ​​of the optimization objective using the Lagrange multiplier method, and solve for the eigenvalues ​​and eigenvectors of the intrinsic orthogonal decomposition kernel function Z using the SVD decomposition method; 4-3) The energy share method is used to screen eigenvalues ​​and eigenvectors and form an orthogonal basis, that is, a flow field price reduction model in the sampling space; 5) Based on the flow field reduction model, the velocity field and pressure field of the flow field are obtained, and the fuselage downward load and rotor thrust of the tiltrotor are obtained by integral calculation.

2. The method for predicting the aerodynamic forces of the rotor and fuselage of a tiltrotor aircraft according to claim 1, characterized in that, Step 2) includes the following steps: 2-1) Using the range of change corresponding to each key parameter variable as the dividing target, divide the range of change corresponding to each key parameter variable into m non-overlapping intervals, so that each interval has the same probability; 2-2) Randomly select a point in each interval of each key parameter variable; 2-3) Map all randomly selected points to normally distributed sample points using the inverse function of the standard normal distribution; 2-4) Then, randomly extract the sample points mapped in step 2-3) from each key parameter variable and form a sample vector; arrange the elements of each column of the sample vector in ascending order to form an X-dimensional sample space, i.e., the sampling space, composed of n key parameter variables.

3. The method for predicting the aerodynamic forces of the rotor and fuselage of a tiltrotor aircraft according to claim 1, characterized in that, Step 4-1) specifically involves: Based on the instantaneous aerodynamic data u extracted from the flow field in the sampling space defined by CFD numerical simulation method. i Where i = 1…N; find an orthogonal basis function on the objective function H of the intrinsic orthogonal decomposition. To make H reach its maximum value, that is: Where u i exist The direction with the greatest, The value of is used as the optimal orthogonal basis for the orthogonal basis functions in the eigenorthogonal decomposition. u i For N instantaneous extractions of flow field aerodynamic data, This represents the inner product of the sampling space based on key parameter variables. It is a vector The 2-norm.

4. The method for predicting the aerodynamic forces of the rotor and fuselage of a tiltrotor aircraft according to claim 1, characterized in that, Step 4-2) specifically involves: (1) Obtain the extreme value of the optimization objective using the Lagrange multiplier method, i.e.: (2) Let U={u1,u2,u3,...,u N Taking the partial derivative of formula (2), we get: (3) Obtain Let Z = UU, the extreme value of Z. T Soon to be sought The extreme values ​​are transformed into obtaining the eigenvalues ​​and eigenvectors of the POD kernel function Z, that is: (4) Decompose formula (4) using the SVD decomposition method, according to Z = UU T And Z=U T U, taking Z, yields: U=Λ∑V Where Λ represents an orthogonal matrix, Σ represents a diagonal matrix, and V represents an orthogonal matrix; the orthogonal matrices Λ and Σ contain the left and right singular vectors of matrix U, respectively, and the orthogonal matrix Λ is related to Z = UU. T The eigenvectors are equal, and the orthogonal matrices V and Z = U are equal. T The eigenvectors of U are equal, and the diagonal matrix Σ contains the eigenvalues ​​λ on the diagonal. i i = 1, 2, 3, ... N Based on the decomposed matrix, obtain the eigenvectors and eigenvalues ​​λ of Z. i .

5. The method for predicting the aerodynamic forces of the rotor and fuselage of a tiltrotor aircraft according to claim 1, characterized in that, Step 4-3) specifically involves: a. Using the energy fraction method, the corresponding eigenvalues ​​λ are obtained through the solution. i The energy information contained therein, i.e. the magnitude of the eigenvalues, is used to characterize the data space of the entire model; b. The eigenvalue λ i The energy occupied is used as the basis for sorting, that is, sorting is based on the size of the eigenvalues; c. Select the eigenvalues ​​and eigenvectors of the first R orders, and determine the proportion I(r) of the first R order eigenvalues ​​in the total energy of the Nth order, i.e.: The closer the value of I(r) is to 1, the more complete the energy information contained in the eigenvector corresponding to the eigenvalue, and the stronger the ability of the reconstructed flow field reduction model to restore the flow field. If it is greater than 0.99, then execute step d; if it is less than 0.9, then return to step c and reselect the eigenvalue and eigenvector. d. The eigenvectors corresponding to the first R eigenvalues ​​obtained are used to form the basis of the final eigenorthogonal decomposition. As a flow field reduction model in the sampling space, the key parameter variables are input and the output is the flow field vector composed of orthogonal bases. Changing the key parameter variables changes the flow field in the sampling space.

Images