An intelligent optimization control method for aerodynamic force of aircraft with local surface protrusion

Abstract

The present invention discloses an intelligent optimization control method for aerodynamic force of a local surface protrusion aircraft. First, an aerodynamic force model is established to predict the aerodynamic force generated by a protrusion at any position. On the basis of the established aerodynamic force model, the optimal protrusion position is selected according to the aerodynamic force index required by the actual task to complete the aerodynamic force optimization control. The intelligent optimization algorithm is used to find the optimal control solution under the given index. Given the required yaw moment and pitch moment, the optimal solution that can meet the conditions and minimize the axial force is found. The present invention adopts the intelligent optimization algorithm ant colony algorithm and particle swarm algorithm as the optimization algorithm, which has high calculation efficiency, good robustness and high precision.

Description

Technical Field

[0001] The invention relates to an intelligent optimization control method for aerodynamic force of an aircraft with a local surface protrusion, and belongs to the technical field of aircraft aerodynamic force modeling. Background Art

[0002] Aircraft generally use air rudders to control their attitude, but air rudders will limit the improvement of aircraft aerodynamic performance. At low speeds and large angles of attack, the efficiency of the rudder surface will decrease. At high speeds, the rudder surface will increase resistance and face severe aerodynamic heating. As the demand for stealth performance increases, the rudder surface is a source of radar reflection, which restricts stealth performance. Therefore, rudderless control technology has become a hot spot in attitude control and one of the important development directions of attitude control in the future.

[0003] In the future, it will be possible to generate arbitrarily protrusions on the surface of an aircraft through local deformation of smart materials, and control the attitude of the aircraft in real time. This is a new type of rudderless control technology. Protrusions (such as centimeter-level hemispherical protrusions) can change the aerodynamic torque by changing the local pressure distribution, thereby controlling the attitude of the aircraft. According to the flight state of the aircraft in real time, it is determined how much torque the protrusion needs to generate, thereby determining the position of the protrusion. An aerodynamic model of the protrusion is established based on machine learning, and on this basis, an aerodynamic optimization control scheme is established. Traditional optimization schemes have slow responses and poor robustness for complex models. In addition, the input model is a machine learning prediction model, which is relatively complex and difficult to solve with traditional optimization schemes. Summary of the invention

[0004] The purpose of the present invention is to overcome the above-mentioned shortcomings of the prior art and to provide a method for intelligent optimization control of aerodynamic forces of an aircraft with local surface protrusions. In view of the problems of long time consumption and poor robustness in online optimization of the optimal protrusion position under any given aerodynamic index, the method establishes an intelligent optimization control scheme for any given aerodynamic force based on an intelligent optimization algorithm.

[0005] The technical solution of the present invention is:

[0006] The present invention discloses an intelligent optimization control method for aerodynamic force of an aircraft with a local surface protrusion, comprising:

[0007] S1, aerodynamic model of aircraft with local convexity on loading surface;

[0008] S2, judging the optimization category according to the optimization index; if the optimization category is judged to be a dual-channel control model, proceed to step S3; if the optimization category is judged to be a single-channel unconstrained model, proceed to step S4; if the optimization category is judged to be a single-channel constrained model, proceed to step S5;

[0009] S3, setting optimization parameters; using the particle swarm algorithm to solve the optimal coordinates of the two-point control according to the aerodynamic model and the optimization parameters; proceeding to step S8;

[0010] S4, setting optimization parameters; using the ant colony algorithm to solve the single-point control optimal point coordinates according to the aerodynamic model and the optimization parameters; proceeding to step S6;

[0011] S5, setting optimization parameters; solving the single-point control optimal point coordinates using a physical model according to the aerodynamic model and the optimization parameters;

[0012] S6. Obtaining a value range of the pitch moment or the yaw moment according to the coordinates of the optimal point of single-point control;

[0013] S7, judging whether the optimization index is within the value range, if yes, using the ant colony algorithm to solve the coordinates of the optimal point of single-point control, if not, using the ant colony algorithm to solve the coordinates of the optimal point of double-point control;

[0014] S8. End.

[0015] In the above-mentioned intelligent optimization control method for aerodynamic force of an aircraft, the optimization category is determined according to the optimization index, and the specific steps are as follows:

[0016] (1) If the optimization index is the yaw moment or pitch moment of a channel, the optimization category is a single-channel unconstrained model;

[0017] (2) If the optimization index is the yaw moment or pitch moment of one channel, and the optimization index of the other channel is 0, then the optimization category is a single-channel constrained model;

[0018] (3) If the optimization index is the yaw moment and pitch moment of two channels, the optimization category is a dual-channel control model.

[0019] In the above-mentioned intelligent optimization control method for aerodynamic force of an aircraft, the specific method for setting the optimization parameters is as follows:

[0020] (4) Determine the optimization category as a single-channel unconstrained model and set the single-channel unconstrained optimization parameters;

[0021] When the constraint parameter is the optimization index of the yaw moment, the single-channel optimization parameter calculation formula is:

[0022] minF(x)

[0023] stdmyg=k1

[0024] Among them, x is the aerodynamic model coordinate, F is the axial force, dmyg is the yaw moment, min is the minimum value, and k1 is the optimization index of the yaw moment;

[0025] When the constraint parameter is the optimization index of the pitch moment, the single-channel optimization parameter calculation formula is:

[0026] minF(x)

[0027] stdmzg=k2

[0028] Among them, dmzg is the pitching moment, and k2 is the optimization index of the pitching moment;

[0029] (5) Determine the optimization category as a single-channel constrained model and set the single-channel constrained optimization parameters;

[0030] When the constraint parameter is the optimization index of the yaw moment, the single-channel optimization parameter calculation formula is:

[0031] minF(x)

[0032] stdmzg=k1

[0033] dmyg=0

[0034] When the constraint parameter is the optimization index of the pitch moment, the single-channel optimization parameter calculation formula is:

[0035] minF(x)

[0036] stdmzg=0

[0037] dmyg=k2

[0038] (6) Determine the optimization category as a dual-channel model and set dual-channel optimization parameters;

[0039] The method for setting the dual-channel optimization parameters is as follows:

[0040] minF(x)

[0041] stdmyg=k1

[0042] dmzg=k2

[0043] Among them, F is the axial force, dmyg is the yaw moment, and dmzg is the pitch moment.

[0044] In the above-mentioned intelligent optimization control method for aerodynamic force of an aircraft, the particle swarm algorithm is used to solve the optimal coordinates of the double-point control, and the specific method is:

[0045] (1) Initializing the initial coordinate points and initial velocity of the particle swarm;

[0046] (2) calculating a dual-point objective function of the initial coordinate point according to the aerodynamic model;

[0047] (3) selecting the local optimal point coordinates and the global optimal point coordinates at the initial moment according to the dual-point objective function of the initial coordinate point;

[0048] (4) Similarly, based on the coordinate point and velocity at time t, calculate the coordinate point and velocity at time t+1; where t=0…N, N is the maximum number of iterations;

[0049] (5) calculating the dual-point objective function of the coordinate point at time t+1 according to the aerodynamic model, and selecting the local optimal point coordinates and the global optimal point coordinates at time t+1 according to the dual-point objective function;

[0050] (6) Determine whether the maximum number of iterations has been reached or the value of the global optimal point coordinates remains unchanged. If so, exit the iteration and output the global optimal point coordinates as the optimal point coordinates of the two-point control. Otherwise, return to step (4).

[0051] In the above-mentioned intelligent optimization control method for aerodynamic force of an aircraft, the ant colony algorithm is used to solve the coordinates of the optimal point of single-point control, and the specific steps are:

[0052] (1) Initialize the transfer probability, pheromone volatility coefficient and the coordinate point of the ant colony at the initial moment;

[0053] (2) obtaining a single point objective function of the ant colony coordinate point at the initial moment according to the aerodynamic model;

[0054] (3) According to the single point objective function of the ant colony coordinate point at the initial moment, the local optimal point coordinates and the global optimal point coordinates at the initial moment are selected;

[0055] (4) Calculate pheromones based on the pheromone volatility coefficient and calculate the transfer probability based on the pheromones;

[0056] (5) Similarly, the ant colony coordinates at time t+1 are calculated based on the ant colony coordinates and transition probabilities at time t; where t=0…N-1, and N is the maximum number of iterations;

[0057] (6) calculating the single point objective function of the ant colony coordinate point at time t+1 according to the aerodynamic model;

[0058] (7) According to the single point objective function at time t+1, select the coordinates of the local optimal point and the global optimal point at time t+1;

[0059] (8) Determine whether the maximum number of iterations has been reached or the global optimal point coordinates remain unchanged. If so, output the global optimal point coordinates as the single-point control optimal point coordinates; otherwise, return to step (5).

[0060] In the above-mentioned intelligent optimization control method for aerodynamic force of an aircraft, the physical model is used to solve the coordinates of the optimal point of single-point control. The specific method is:

[0061] By using the optimization control input model and the physical characteristics of the aircraft, grid points are divided in the area near the center of pressure coordinates where dmzg=0, and coordinate points that satisfy the pitch moment being less than a given error e1 are found by traversing point by point and the variation trend of the pitch moment near the center of pressure;

[0062] Or: using the optimization control input model and the physical characteristics of the aircraft, grid points are divided in the area near the symmetry axis of the aircraft where dmyg=0, and the coordinate points that satisfy the yaw moment less than the error e2 are found by traversing point by point and the change trend of the yaw moment near the symmetry axis;

[0063] According to the coordinate points, the ant colony algorithm is used to solve the coordinates of the optimal point control of the single point.

[0064] In the above-mentioned intelligent optimization control method for aerodynamic force of an aircraft, the value range of the pitch moment or the yaw moment is obtained according to the coordinates of the optimal point of single-point control, and the specific method is:

[0065] According to the aerodynamic model, the pitch moment or yaw moment corresponding to the coordinates of the single-point control optimal point is obtained, and the pitch moment or yaw moment is determined as the maximum value of the value range, and the minimum value of the value range is 0.

[0066] In the above-mentioned intelligent optimization control method for aerodynamic force of an aircraft, the ant colony algorithm is used to solve the coordinates of the optimal point of double-point control. The specific method is:

[0067] (1) Initialize the transfer probability and pheromone volatility coefficient;

[0068] (2) obtaining a dual-point objective function of two randomly specified initial ant colony coordinate points according to the aerodynamic model;

[0069] (3) According to the dual-point objective function, the local optimal point coordinates and the global optimal point coordinates at the initial moment are selected;

[0070] (4) Calculate pheromones based on the pheromone volatility coefficient and calculate the transfer probability based on the pheromones;

[0071] (5) Similarly, the ant colony coordinates at time t+1 are calculated based on the ant colony coordinates and transition probabilities at time t; where t=0…N-1, and N is the maximum number of iterations;

[0072] (6) calculating the double-point objective function of the ant colony coordinate point at time t+1 according to the aerodynamic model;

[0073] (7) According to the dual-point objective function at time t+1, select the coordinates of the local optimal point and the global optimal point at time t+1;

[0074] (8) Determine whether the maximum number of iterations has been reached or the global optimal point coordinates remain unchanged. If so, output the global optimal point coordinates as the single-point control optimal point coordinates; otherwise, return to step (5).

[0075] In the above-mentioned aircraft aerodynamic intelligent optimization control method, the dual-point objective function is:

[0076]

[0077] Where T(x) is the dual-point objective function, F(x i ) is x i The axial force at the coordinate point, σ is the penalty factor, L i is the normalization coefficient, P i is the penalty function, x i are the coordinate points of the aerodynamic model;

[0078] Normalization coefficient, the formula is:

[0079]

[0080] Where, L i is the normalization coefficient, P i is the penalty function;

[0081] The penalty function of the yaw moment is:

[0082]

[0083] Among them, dmyg is the yaw moment, k1 is the optimization index of the yaw moment, and e1 is the small error of the yaw moment;

[0084] The penalty function of the pitching moment is:

[0085]

[0086] Among them, dmzg is the pitching moment, k2 is the optimization index of the pitching moment, and e2 is the small error of the pitching moment.

[0087] In the above-mentioned intelligent optimization control method for aerodynamic force of an aircraft, the coordinates of the local optimal point and the coordinates of the global optimal point are calculated by the following formula:

[0088] The coordinates of the local optimal point are given by:

[0089] Pbest k =argmin(T(x i ))

[0090] Where i is the particle number, i = 1, 2, ... m, m is the size of the particle group, argmin(T(x i)) is to make the objective function T(x i ) when the variable x takes the minimum value i The value of

[0091] The global optimal point coordinates are as follows:

[0092] Gbest=argmin(Gbest,Pbest k )

[0093] Among them, argmin(Gbest,Pbest k ) is to take Gbest and Pbest k The smaller of the two.

[0094] In the above-mentioned aircraft aerodynamic intelligent optimization control method, the coordinate point and velocity of the aerodynamic model at time t+1 are calculated using the formula:

[0095]

[0096]

[0097] Among them, ω is the convergence parameter, c1, c2 are learning factors, r1, r2 are two random numbers between 0 and 1; X i is the coordinate point of the ith particle, is the coordinate point of the kth particle at the tth iteration, Pbest is the iteration speed of the kth particle at the tth iteration; k is the local optimal point coordinate of the current particle, and Gbest is the global optimal point coordinate.

[0098] In the above-mentioned intelligent optimization control method for aerodynamic force of an aircraft, pheromones are calculated according to the volatility coefficient of pheromones, and the transfer probability is calculated according to the pheromones. The specific method is as follows:

[0099] The pheromone calculation formula is:

[0100]

[0101] Among them, τ ij is the pheromone between i and j, ρ is the pheromone volatility coefficient, For pheromone increment, Q is a constant, L k is the length of the path currently traveled by the ant.

[0102] The transfer probability is calculated as follows:

[0103]

[0104] Among them, α, β are importance factors, τij (t) is the pheromone between i and j, η ij =1 / d ij , d ij is the distance between i and j;

[0105]

[0106] Among them, x is the coordinate point.

[0107] In the above-mentioned intelligent optimization control method for aerodynamic force of an aircraft, the single-point objective function is formulated as follows:

[0108]

[0109] Where T(x) is the single point objective function with penalty function added, F(x) is the axial force at the x coordinate point, σ is the penalty factor, and L i is the normalization coefficient, P i is the penalty function, and x is the coordinate point of the aerodynamic model.

[0110] In the above-mentioned intelligent optimization control method for aerodynamic force of an aircraft, the ant colony coordinate point at time t+1 is calculated according to the ant colony coordinate point at time t and the transfer probability. The specific method is:

[0111]

[0112] Among them, x(t) is the coordinate at time t, x(t+1) is the coordinate at time t+1, and step is the given step size. i is the current iteration number.

[0113] The beneficial effects of the present invention and the prior art are:

[0114] 1) The present invention adopts the intelligent optimization algorithm Ant Colony Algorithm (ACO) and Particle Swarm Algorithm (PSO) as the optimization algorithm. Compared with the traditional optimization control method, the intelligent optimization algorithm has high computational efficiency, good robustness and high precision. For complex models obtained based on machine learning, the intelligent optimization algorithm is easier to implement.

[0115] 2) The present invention can load and import different types of aerodynamic models. As the models are updated, different optimization schemes can be obtained, thereby improving the flexibility of optimization control.

[0116] 3) The present invention adopts an intelligent optimization algorithm for the optimization control of complex models, which can better solve the complexity problem of the model. The algorithm has higher accuracy and scalability, laying the foundation for subsequent optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0117] Figure 1 This is a flow chart of the aerodynamic intelligent optimization control method of the present invention;

[0118] Figure 2 It is a schematic diagram of the yaw moment control area under the pitch moment constraint of the present invention;

[0119] Figure 3 It is a schematic diagram of the pitch moment control area under the constrained yaw moment of the present invention. DETAILED DESCRIPTION

[0120] The present invention is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0121] like Figure 1 As shown, the present invention discloses an intelligent optimization control method for aerodynamic force of an aircraft with a local surface protrusion, comprising:

[0122] The present invention discloses an intelligent optimization control method for aerodynamic force of an aircraft with a local surface protrusion, comprising:

[0123] Step 1, loading the aerodynamic model of the aircraft with local bulge on the surface;

[0124] Step 2: Determine the optimization category according to the optimization index; if the optimization category is determined to be a dual-channel control model, proceed to step 3; if the optimization category is determined to be a single-channel unconstrained model, proceed to step 4; if the optimization category is determined to be a single-channel constrained model, proceed to step 5;

[0125] Step 3, set the optimization parameters; according to the aerodynamic model and the optimization parameters, use the particle swarm algorithm to solve the optimal coordinates of the double-point control; go to step 8;

[0126] Step 4, set the optimization parameters; according to the aerodynamic model and the optimization parameters, use the ant colony algorithm to solve the single-point control optimal point coordinates; go to step 6;

[0127] Step 5: Set the optimization parameters; according to the aerodynamic model and the optimization parameters, use the physical model to solve the single-point control optimal point coordinates;

[0128] Step 6: Obtain the value range of the pitch moment or the yaw moment according to the coordinates of the optimal point of single-point control;

[0129] Step 7: Determine whether the optimization index is within the value range. If yes, use the ant colony algorithm to solve the coordinates of the optimal point for single-point control. If not, use the ant colony algorithm to solve the coordinates of the optimal point for double-point control.

[0130] Step 8. End.

[0131] Determine the optimization category based on the optimization index. The specific steps are:

[0132] (1) If the optimization index is the yaw moment or pitch moment of a channel, the optimization category is a single-channel unconstrained model;

[0133] (2) If the optimization index is the yaw moment or pitch moment of one channel and the optimization index of the other channel is 0, the optimization category is a single-channel constrained model;

[0134] (3) If the optimization index is the yaw moment and pitch moment of the two channels, the optimization category is the dual-channel control model.

[0135] Set the optimization parameters according to the optimization category. The specific method is as follows:

[0136] (7) Determine the optimization category as a single-channel unconstrained model and set the single-channel unconstrained optimization parameters;

[0137] When the constraint parameter is the optimization index of the yaw moment, the single-channel optimization parameter calculation formula is:

[0138] min F(x)

[0139] stdmyg=k1

[0140] Where x is the aerodynamic model coordinate, F is the axial force, dmyg is the yaw moment, min is the minimum value, and k1 is the optimization index of the yaw moment;

[0141] When the constraint parameter is the optimization index of the pitch moment, the single-channel optimization parameter calculation formula is:

[0142] min F(x)

[0143] stdmzg=k2

[0144] Among them, dmzg is the pitching moment, and k2 is the optimization index of the pitching moment;

[0145] (8) Determine the optimization category as a single-channel constrained model and set the single-channel constrained optimization parameters;

[0146] When the constraint parameter is the optimization index of the yaw moment, the single-channel optimization parameter calculation formula is:

[0147] min F(x)

[0148] stdmzg=k1

[0149] dmyg=0

[0150] When the constraint parameter is the optimization index of the pitch moment, the single-channel optimization parameter calculation formula is:

[0151] min F(x)

[0152] stdmzg=0

[0153] dmyg=k2

[0154] (9) Determine the optimization category as a dual-channel model and set dual-channel optimization parameters;

[0155] Set dual-channel optimization parameters by:

[0156] min F(x)

[0157] stdmyg=k1

[0158] dmzg=k

[0159] Among them, F is the axial force, dmyg is the yaw moment, and dmzg is the pitch moment.

[0160] The particle swarm algorithm is used to solve the optimal coordinates of the two-point control. The specific method is as follows:

[0161] (1) Initialize the initial coordinate points and initial velocity of the particle swarm;

[0162] (2) Calculate the double-point objective function of the initial coordinate point according to the aerodynamic model;

[0163] (3) According to the dual-point objective function of the initial coordinate point, select the local optimal point coordinates and the global optimal point coordinates at the initial moment;

[0164] (4) Similarly, based on the coordinate point and velocity at time t, calculate the coordinate point and velocity at time t+1; where t=0…N, N is the maximum number of iterations;

[0165] (5) According to the aerodynamic model, the double-point objective function of the coordinate point at time t+1 is calculated, and the coordinates of the local optimal point and the global optimal point at time t+1 are selected according to the double-point objective function;

[0166] (6) Determine whether the maximum number of iterations has been reached or the value of the global optimal point coordinates remains unchanged. If so, exit the iteration and output the global optimal point coordinates as the optimal point coordinates of the two-point control. Otherwise, return to step (4).

[0167] The ant colony algorithm is used to solve the coordinates of the optimal point control point. The specific steps are as follows:

[0168] (1) Initialize the transfer probability, pheromone volatility coefficient and the coordinate point of the ant colony at the initial moment;

[0169] (2) According to the aerodynamic model, the single point objective function of the ant colony coordinate point at the initial moment is obtained;

[0170] (3) According to the single point objective function of the ant colony coordinate point at the initial moment, the local optimal point coordinates and the global optimal point coordinates at the initial moment are selected;

[0171] (4) Calculate pheromones based on the pheromone volatility coefficient and calculate the transfer probability based on the pheromones;

[0172] (5) Similarly, the ant colony coordinates at time t+1 are calculated based on the ant colony coordinates and transition probabilities at time t; where t=0…N-1, and N is the maximum number of iterations;

[0173] (6) Calculate the single point objective function of the ant colony coordinate point at time t+1 according to the aerodynamic model;

[0174] (7) According to the single point objective function at time t+1, select the coordinates of the local optimal point and the global optimal point at time t+1;

[0175] (8) Determine whether the maximum number of iterations has been reached or the global optimal point coordinates remain unchanged. If so, output the global optimal point coordinates as the single-point control optimal point coordinates; otherwise, return to step (5).

[0176] The physical model is used to solve the coordinates of the optimal point of single-point control. The specific method is:

[0177] By using the optimized control input model and the physical characteristics of the aircraft, the grid points are divided in the area near the center of pressure coordinates where dmzg=0, and the coordinate points that satisfy the pitch moment less than the given error e1 are found by traversing point by point and the change trend of the pitch moment near the center of pressure.

[0178] Or: using the optimized control input model and the physical characteristics of the aircraft, grid points are divided in the area near the symmetry axis of the aircraft where dmyg = 0, and the coordinate points that satisfy the yaw moment less than the error e2 are found by traversing point by point and the change trend of the yaw moment near the symmetry axis;

[0179] According to the coordinate points, the ant colony algorithm is used to solve the coordinates of the optimal point control of the single point.

[0180] The value range of the pitch moment or yaw moment is obtained according to the coordinates of the optimal point of single-point control. The specific method is:

[0181] According to the aerodynamic model, the pitch moment or yaw moment corresponding to the coordinates of the single-point control optimal point is obtained, and the pitch moment or yaw moment is determined as the maximum value of the value range, and the minimum value of the value range is 0.

[0182] The ant colony algorithm is used to solve the optimal coordinates of the two-point control. The specific method is as follows:

[0183] (1) Initialize the transfer probability and pheromone volatility coefficient;

[0184] (2) According to the aerodynamic model, a dual-point objective function of two randomly specified initial ant colony coordinate points is obtained;

[0185] (3) According to the dual-point objective function, select the coordinates of the local optimal point and the global optimal point at the initial moment;

[0186] (4) Calculate pheromones based on the pheromone volatility coefficient and calculate the transfer probability based on the pheromones;

[0187] (5) Similarly, the ant colony coordinates at time t+1 are calculated based on the ant colony coordinates and transition probabilities at time t; where t=0…N-1, and N is the maximum number of iterations;

[0188] (6) According to the aerodynamic model, calculate the double-point objective function of the ant colony coordinate point at time t+1;

[0189] (7) According to the dual-point objective function at time t+1, select the coordinates of the local optimal point and the global optimal point at time t+1;

[0190] (8) Determine whether the maximum number of iterations has been reached or the global optimal point coordinates remain unchanged. If so, output the global optimal point coordinates as the single-point control optimal point coordinates; otherwise, return to step (5).

[0191] The double-point objective function is:

[0192]

[0193] Where T(x) is the dual-point objective function, F(x i ) is x i The axial force at the coordinate point, σ is the penalty factor, L i is the normalization coefficient, P i is the penalty function, x i are the coordinate points of the aerodynamic model;

[0194] Normalization coefficient, the formula is:

[0195]

[0196] Where, L i is the normalization coefficient, P i is the penalty function;

[0197] The penalty function of the yaw moment is:

[0198]

[0199] Among them, dmyg is the yaw moment, k1 is the optimization index of the yaw moment, and e1 is the small error of the yaw moment;

[0200] The penalty function of the pitching moment is:

[0201]

[0202] Among them, dmzg is the pitching moment, k2 is the optimization index of the pitching moment, and e2 is the small error of the pitching moment.

[0203] The coordinates of the local optimal point and the global optimal point are calculated as follows:

[0204] The coordinates of the local optimal point are given by:

[0205] Pbest k =argmin(T(x i ))

[0206] Where i is the particle number, i = 1, 2, ... m, m is the size of the particle group, argmin(T(x i )) is to make the objective function T(x i ) when the variable x takes the minimum value i The value of

[0207] The global optimal point coordinates are as follows:

[0208] Gbest=argmin(Gbest,Pbest k )

[0209] Among them, argmin(Gbest,Pbest k ) is to take Gbest and Pbest k The smaller of the two.

[0210] Calculate the coordinate points and velocity of the aerodynamic model at time t+1 using the formula:

[0211]

[0212]

[0213] Among them, ω is the convergence parameter, c1, c2 are learning factors, r1, r2 are two random numbers between 0 and 1; X i is the coordinate point of the ith particle, is the coordinate point of the kth particle at the tth iteration, Pbest is the iteration speed of the kth particle at the tth iteration; k is the local optimal point coordinate of the current particle, and Gbest is the global optimal point coordinate.

[0214] The pheromone is calculated based on the pheromone volatility coefficient, and the transfer probability is calculated based on the pheromone. The specific method is:

[0215] Pheromones, the calculation formula is:

[0216]

[0217] Among them, τ ij is the pheromone between i and j, ρ is the pheromone volatility coefficient, For pheromone increment, Q is a constant, L k is the length of the path currently traveled by the ant.

[0218] The transition probability is calculated as:

[0219]

[0220] Among them, α, β are importance factors, τ ij (t) is the pheromone between i and j, η ij =1 / d ij , d ij is the distance between i and j;

[0221]

[0222] Among them, x is the coordinate point.

[0223] Single point objective function, the formula is:

[0224]

[0225] Where T(x) is the single point objective function with penalty function added, F(x) is the axial force at the x coordinate point, σ is the penalty factor, and L i is the normalization coefficient, P i is the penalty function, and x is the coordinate point of the aerodynamic model.

[0226] According to the ant colony coordinate point and transition probability at time t, the ant colony coordinate point at time t+1 is calculated as follows:

[0227]

[0228] Among them, x(t) is the coordinate at time t, x(t+1) is the coordinate at time t+1, and step is the given step size. i is the current iteration number.

[0229] Example 1

[0230] Single channel without constraints: The ant colony algorithm (ACO) is used to solve the value range of the yaw moment dmyg and the pitch moment dmzg, and then select the optimization scheme. Some parameters of the ant colony algorithm are selected as follows: ant scale 20, number of iterations 20, pheromone volatility coefficient 0.9, and transfer probability 0.2. Finally, the value range and coordinates of the yaw moment dmyg and the pitch moment dmzg are obtained as follows:

[0231] para min min_pos max max_pos dmyg 0.225 [3548.583,261.140] 213.37 [1201.475,405.011] dmzg 563.9 [1202.145,176.645] 135.932 [2808.880,603.327]

[0232] According to the range of values ​​obtained by the solution, the selection of optimization scheme can be given. Given some examples and their results, the optimized response time is calculated as follows:

[0233]

[0234]

[0235] Single channel with constraints: Given the thresholds of the yaw moment dmyg and the pitch moment dmzg, it is believed that when the absolute value of the yaw moment dmyg is less than 1 and the absolute value of the pitch moment dmzg is less than 5, there is no control effect on the torque. According to the physical model analysis, the center of mass is approximately at x = 2250, at which time dmzg ≈ 0, so the optimization is performed near the center of mass, and the region of constrained single channel control of the yaw moment is established and its value range is solved. Near z = 0, dmyg ≈ 0, and it is believed that the optimization at this time does not affect the yaw moment, so a single channel constrained control region of the pitch moment is established near the center and its value range is solved.

[0236] The single-channel control area is subdivided by grid construction, and the input model is called to calculate the aircraft area that can be controlled by single channel, and its boundary is obtained by curve fitting, such as Figure 2 , as shown in Figure 3.

[0237] On the basis of solving the region, solving the maximum value of the single-channel control in the corresponding region is the basis for selecting the constrained single-channel control optimization scheme, that is, judging single-channel control or multi-point control. Using the ant colony algorithm to find the maximum value in the region, the results are as follows:

[0238] para min max dmyg 10.295 130.6221429 dmzg -340.43 41.465

[0239] When a constrained single-channel optimization problem is selected, it is necessary to compare the indicator with the maximum value to determine whether the indicator can be achieved by single-channel control. If it cannot be achieved, dual-point control is used.

[0240] Dual-channel control: The particle swarm algorithm (PSO) is used to solve the value range of the yaw moment dmyg and pitch moment dmzg that can be used for dual-channel control using a single channel, and then determine the optimization solution for dual-channel control.

[0241] Given a certain working condition, the solution and error are as follows:

[0242]

[0243]

[0244] According to the results and errors, except for individual cases, the error can be controlled within 10%. Therefore, the particle swarm algorithm can find the optimal coordinates for dual-channel control, and minimize the sum of the axial forces while satisfying the two indicators of pitch moment and yaw moment to meet the control requirements.

[0245] The above is a detailed introduction to the aerodynamic intelligent optimization control method of the local protrusion on the surface of the aircraft of the present invention. This article uses specific cases to illustrate the principles and implementation methods of the present invention. The above specific case description is only used to help understand the method and core idea of ​​the present invention. It should be pointed out that any simple modification made to the above examples without departing from the technical essence of the present invention belongs to the technical scope of the present invention.

[0246] The contents of the invention not described in detail are well known to those skilled in the art.

Claims

1. A method for intelligent optimization control of aerodynamic force of a locally raised surface aircraft, characterized in that: include: S1, aerodynamic model of aircraft with local convexity on loading surface; S2. Determine the optimization category according to the optimization index; If the optimization category is determined to be a dual-channel control model, proceed to step S3; if the optimization category is determined to be a single-channel unconstrained model, proceed to step S4; if the optimization category is determined to be a single-channel constrained model, proceed to step S5; S3, setting optimization parameters; using the particle swarm algorithm to solve the optimal coordinates of the two-point control according to the aerodynamic model and the optimization parameters; proceeding to step S8; S4, setting optimization parameters; using the ant colony algorithm to solve the single-point control optimal point coordinates according to the aerodynamic model and the optimization parameters; proceeding to step S6; S5, setting optimization parameters; solving the single-point control optimal point coordinates using a physical model according to the aerodynamic model and the optimization parameters; S6. Obtaining a value range of the pitch moment or the yaw moment according to the coordinates of the optimal point of single-point control; S7, judging whether the optimization index is within the value range, if yes, using the ant colony algorithm to solve the coordinates of the optimal point of single-point control, if not, using the ant colony algorithm to solve the coordinates of the optimal point of double-point control; S8, end; The specific method for setting the optimization parameters is as follows: (1) Determine the optimization category as a single-channel unconstrained model and set the single-channel unconstrained optimization parameters; When the constraint parameter is the optimization index of the yaw moment, the single-channel optimization parameter calculation formula is: minF(x) stdmyg=k1 Among them, x is the aerodynamic model coordinate, F is the axial force, dmyg is the yaw moment, min is the minimum value, and k1 is the optimization index of the yaw moment; When the constraint parameter is the optimization index of the pitch moment, the single-channel optimization parameter calculation formula is: minF(x) stdmzg=k2 Among them, dmzg is the pitching moment, and k2 is the optimization index of the pitching moment; (2) Determine the optimization category as a single-channel constrained model and set the single-channel constrained optimization parameters; When the constraint parameter is the optimization index of the yaw moment, the single-channel optimization parameter calculation formula is: minF(x) stdmzg=k1 dmyg=0 When the constraint parameter is the optimization index of the pitch moment, the single-channel optimization parameter calculation formula is: minF(x) stdmzg=0 dmyg=k2 (3) Determine the optimization category as a dual-channel model and set dual-channel optimization parameters; The method for setting the dual-channel optimization parameters is as follows: minF(x) stdmyg=k1 dmzg=k2 Among them, F is the axial force, dmyg is the yaw moment, and dmzg is the pitch moment.

2. The method for intelligent optimization control of aerodynamic force of a partially raised surface aircraft according to claim 1, characterized in that: Determine the optimization category based on the optimization index. The specific steps are: (1) If the optimization index is the yaw moment or pitch moment of a channel, the optimization category is a single-channel unconstrained model; (2) If the optimization index is the yaw moment or pitch moment of one channel, and the optimization index of the other channel is 0, then the optimization category is a single-channel constrained model; (3) If the optimization index is the yaw moment and pitch moment of two channels, the optimization category is a dual-channel control model.

3. The method for intelligent optimization control of aerodynamic force of a partially raised surface aircraft according to claim 1, characterized in that: The particle swarm algorithm is used to solve the optimal coordinates of the double-point control, and the specific method is as follows: (1) Initializing the initial coordinate points and initial velocity of the particle swarm; (2) calculating a dual-point objective function of the initial coordinate point according to the aerodynamic model; (3) selecting the local optimal point coordinates and the global optimal point coordinates at the initial moment according to the dual-point objective function of the initial coordinate point; (4) Similarly, based on the coordinate point and velocity at time t, calculate the coordinate point and velocity at time t+1; where t=0…N, N is the maximum number of iterations; (5) calculating the dual-point objective function of the coordinate point at time t+1 according to the aerodynamic model, and selecting the local optimal point coordinates and the global optimal point coordinates at time t+1 according to the dual-point objective function; (6) Determine whether the maximum number of iterations has been reached or the value of the global optimal point coordinates remains unchanged. If so, exit the iteration and output the global optimal point coordinates as the optimal point coordinates of the two-point control. Otherwise, return to step (4).

4. The method for intelligent optimization control of aerodynamic force of a partially raised surface aircraft according to claim 1, characterized in that: The specific steps of using the ant colony algorithm to solve the coordinates of the optimal point control point are as follows: (1) Initialize the transfer probability, pheromone volatility coefficient and the coordinate point of the ant colony at the initial moment; (2) obtaining a single point objective function of the ant colony coordinate point at the initial moment according to the aerodynamic model; (3) According to the single point objective function of the ant colony coordinate point at the initial moment, the local optimal point coordinates and the global optimal point coordinates at the initial moment are selected; (4) Calculate pheromones based on the pheromone volatility coefficient and calculate the transfer probability based on the pheromones; (5) Similarly, the ant colony coordinates at time t+1 are calculated based on the ant colony coordinates and transition probabilities at time t; where t=0…N-1, and N is the maximum number of iterations; (6) calculating the single point objective function of the ant colony coordinate point at time t+1 according to the aerodynamic model; (7) According to the single point objective function at time t+1, select the coordinates of the local optimal point and the global optimal point at time t+1; (8) Determine whether the maximum number of iterations has been reached or the global optimal point coordinates remain unchanged. If so, output the global optimal point coordinates as the single-point control optimal point coordinates; otherwise, return to step (5).

5. The method for intelligent optimization control of aerodynamic force of a partially raised surface aircraft according to claim 1, characterized in that: The specific method of solving the single-point control optimal point coordinates by using the physical model is as follows: By using the physical model, grid points are divided in the area near the center of pressure coordinates where dmzg=0, and coordinate points that satisfy the pitching moment being less than a given error e1 are found by traversing point by point and the change trend of the pitching moment near the center of pressure; Or: using the physical model, grid points are divided in the area near the symmetry axis of the aircraft where dmyg=0, and the coordinate points satisfying the yaw moment less than the error e2 are found by traversing point by point and the change trend of the yaw moment near the symmetry axis; According to the coordinate points, the ant colony algorithm is used to solve the coordinates of the optimal point control of the single point.

6. The method for intelligent optimization control of aerodynamic force of a partially raised surface aircraft according to claim 1, characterized in that: The specific method for obtaining the value range of the pitch moment or the yaw moment according to the coordinates of the optimal point of single-point control is as follows: According to the aerodynamic model, the pitch moment or yaw moment corresponding to the coordinates of the single-point control optimal point is obtained, and the pitch moment or yaw moment is determined as the maximum value of the value range, and the minimum value of the value range is 0.

7. The method for intelligent optimization control of aerodynamic force of a partially raised surface aircraft according to claim 1, characterized in that: The specific method of using the ant colony algorithm to solve the optimal coordinates of the double-point control is as follows: (1) Initialize the transfer probability and pheromone volatility coefficient; (2) obtaining a dual-point objective function of two randomly specified initial ant colony coordinate points according to the aerodynamic model; (3) According to the dual-point objective function, the local optimal point coordinates and the global optimal point coordinates at the initial moment are selected; (4) Calculate pheromones based on the pheromone volatility coefficient and calculate the transfer probability based on the pheromones; (5) Similarly, the ant colony coordinates at time t+1 are calculated based on the ant colony coordinates and transition probabilities at time t; where t=0…N-1, and N is the maximum number of iterations; (6) calculating the double-point objective function of the ant colony coordinate point at time t+1 according to the aerodynamic model; (7) According to the dual-point objective function at time t+1, select the coordinates of the local optimal point and the global optimal point at time t+1; (8) Determine whether the maximum number of iterations has been reached or the global optimal point coordinates remain unchanged. If so, output the global optimal point coordinates as the single-point control optimal point coordinates; otherwise, return to step (5).

8. The method for intelligent optimization control of aerodynamic force of a partially raised surface aircraft according to claim 4, characterized in that: The dual-point objective function is: Where T(x) is the dual-point objective function, F(x i ) is x i The axial force at the coordinate point, σ is the penalty factor, L i is the normalization coefficient, P i is the penalty function, x i are the coordinate points of the aerodynamic model; Normalization coefficient, the formula is: Where, L i is the normalization coefficient, P i is the penalty function; The penalty function of the yaw moment is: Among them, dmyg is the yaw moment, k1 is the optimization index of the yaw moment, and e1 is the small error of the yaw moment; The penalty function of the pitching moment is: Among them, dmzg is the pitching moment, k2 is the optimization index of the pitching moment, and e2 is the small error of the pitching moment.

9. The method for intelligent optimization control of aerodynamic force of a partially raised surface aircraft according to claim 4, characterized in that: The local optimal point coordinates and the global optimal point coordinates are calculated as follows: The coordinates of the local optimal point are given by: Pbest k =argmin(T(x i )) Where i is the particle number, i = 1, 2, ... m, m is the size of the particle group, argmin(T(x i )) is to make the objective function T(x i ) when the variable x takes the minimum value i The value of The global optimal point coordinates are as follows: Gbest=argmin(Gbest,Pbest k ) Among them, argmin(Gbest,Pbest k ) is to take Gbest and Pbest k The smaller of the two.

10. The method for intelligent optimization control of aerodynamic force of a partially raised surface aircraft according to claim 4, characterized in that: The formula for calculating the coordinate points and speed of the aerodynamic model at time t+1 is: Among them, ω is the convergence parameter, c1, c2 are learning factors, r1, r2 are two random numbers between 0 and 1; X i is the coordinate point of the ith particle, is the coordinate point of the kth particle at the tth iteration, V k t Pbest is the iteration speed of the kth particle at the tth iteration; k is the local optimal point coordinate of the current particle, and Gbest is the global optimal point coordinate.

11. The method for intelligent optimization control of aerodynamic force of a partially raised surface aircraft according to claim 6, characterized in that: The pheromone is calculated based on the pheromone volatility coefficient, and the transfer probability is calculated based on the pheromone. The specific method is: The pheromone calculation formula is: Among them, τ ij is the pheromone between i and j, ρ is the pheromone volatility coefficient, For pheromone increment, Q is a constant, L k is the length of the path currently traveled by the ant; The transfer probability is calculated as follows: Among them, α, β are importance factors, τ ij (t) is the pheromone between i and j, η ij =1 / d ij , d ij is the distance between i and j; Among them, x is the coordinate point.

12. The method for intelligent optimization control of aerodynamic force of a vehicle with local surface protrusions according to claim 5, characterized in that: The single point objective function is formulated as follows: Where T(x) is the single point objective function with penalty function added, F(x) is the axial force at the x coordinate point, σ is the penalty factor, and L i is the normalization coefficient, P i is the penalty function, and x is the coordinate point of the aerodynamic model.

13. The method for intelligent optimization control of aerodynamic force of a vehicle with local surface protrusions according to claim 5 or 8, characterized in that: According to the ant colony coordinate point and transition probability at time t, the ant colony coordinate point at time t+1 is calculated as follows: Among them, x(t) is the coordinate at time t, x(t+1) is the coordinate at time t+1, and step is the given step size. i is the current iteration number.

Images