Online adaptive planning method for the roll angle of a 180-degree flip of a symmetrical aircraft

Abstract

The present invention discloses an online adaptive planning method for the roll angle of a plane-symmetric aircraft flipping 180 degrees, and the method comprises the following steps: S1, establishing an air density model; S2, establishing a roll control torque coefficient model; S3, establishing a roll control torque model; S4, setting an adjustment coefficient; S5, establishing an amplitude model of a roll angle instruction; S6, establishing a roll angle change frequency model; S7, establishing a roll angle instruction model; S8, establishing an angle instruction execution model. The method uses the relationship between sine and cosine curves and angles and angular velocity to solve practical engineering problems such as the discontinuity of the first-order derivative of the angle instruction, the instruction deviation caused by the integral error, and the complexity of the planning strategy; and uses the physical property that the roll torque is closely related to the aircraft response speed to solve the strategic problem that the fixed rotation period and rotation direction will reduce the maximum range capability from the two dimensions of increasing the roll angle change rate and reducing the rotation angle.

Description

Technical Field

[0001] The invention belongs to the field of aircraft guidance and control, and relates to an aircraft roll angle instruction planning method, and in particular to an online adaptive planning method for the roll angle of a plane-symmetric aircraft flipping 180 degrees. Background Art

[0002] Symmetrical aircraft often adopt aerodynamic layout schemes with large aspect ratio wings to improve the maximum range capability. Due to the interference between the lifting lugs and the wings, the mounting scheme adopts an inverted hanging method with the lifting lugs on top and the main wing on the bottom to ensure the integrity of the main wing. After the aircraft safely leaves the aircraft, an active 180-degree roll command is executed. After the main wing is fully rotated, the advantage of the large lift-to-drag ratio of the upper wing is brought into play, improving the maximum range capability of the symmetric aircraft.

[0003] In order to quickly stabilize the roll angle of the aircraft to 180°, it is necessary to reasonably plan the active roll command, shorten the active roll cycle, increase the flight time of the upper wing, and thus improve the maximum range capability of the aircraft. At present, the more common roll angle command planning method is the S-curve planning method with fixed time and fixed rotation direction. This method does not consider the impact of the roll torque on the response speed of the aircraft's roll angle velocity. Selecting the same roll time within the delivery envelope will delay the active roll time; this method also does not consider the characteristics of the aircraft flipping that can be achieved by clockwise or counterclockwise rotation. Both rotate in a fixed direction. When the initial value of the roll angle is opposite to the expected value, the amplitude of the aircraft rotation will increase. Therefore, the use of the S-curve planning strategy with fixed time and fixed rotation direction will result in a loss of the aircraft's range.

[0004] At the same time, when using the S-curve for roll angle command planning, there will be the following practical engineering problems: first, the first-order derivative is discontinuous and the maximum angular velocity change is a constant or a point, which can easily cause response overshoot at the switching point; second, the S-curve is a planning angular velocity strategy, and the angle command is obtained after integration. The influence of the step size will cause integration error, and it is necessary to rely on the autopilot to eliminate the static error introduced by design defects; third, the S-curve can also plan the angle according to the rolling capability of the aircraft, but the rotation cycle needs to be divided into several segments, and the logic is more complicated.

[0005] In summary, the existing angle planning methods have problems in engineering applications such as response overshoot, complex logic and deviation in planned angle instructions, and in strategy, there is a problem of loss of maximum range capability. Summary of the invention

[0006] The purpose of the present invention is to provide an online adaptive planning method for the roll angle of a plane-symmetric aircraft flipping 180 degrees. The method uses sine and cosine curves and the relationship between angle and angular velocity to solve practical engineering problems such as the discontinuity of the first-order derivative of the angle command, the command deviation caused by the integral error, and the complexity of the planning strategy; and uses the physical property that the roll torque is closely related to the response speed of the aircraft to solve the strategic problem that a fixed rotation period and rotation direction will reduce the maximum range capability from the two dimensions of increasing the roll angle change rate and reducing the rotation angle.

[0007] The objective of the present invention is achieved through the following technical solutions:

[0008] An online adaptive planning method for the roll angle of a plane-symmetric aircraft flipping 180 degrees comprises the following steps:

[0009] Step S1, establishing an air density model:

[0010] According to the altitude H0 at the time of delivery, the air density ρ0 is calculated by using a table lookup method;

[0011] Step S2: Establishing a rolling control torque coefficient model:

[0012] Given the maximum rudder deflection angle δ of the rolling channel rm The Mach number ma0, angle of attack α0 and sideslip angle β0 at the time of starting the roll command are used. The three-layer forward feedback BP neural network algorithm is used to calculate the Mach number ma0, angle of attack α0, sideslip angle β0 and the maximum rudder deflection angle δ of the roll channel. rm As input, taking wind tunnel test data as sample, the maximum rolling control moment coefficient m x0 As output, the maximum rolling control moment coefficient m is fitted. x0 Function:

[0013] m x0 =f(ma0,α0,β0,δ rm )

[0014] Step S3: Establishing a rolling control torque model:

[0015] Using the air density ρ0 obtained in step S1 and the maximum rolling control moment coefficient m obtained in step S2 x0 The maximum roll control torque M at the time of starting the roll command is calculated by the following formula: xδ0 :

[0016]

[0017] Where v is the speed of the aircraft, s is the characteristic area, and l is the length of the aircraft;

[0018] Step S4, setting the adjustment coefficient:

[0019] Considering the deviation between the actual flight aerodynamic characteristics of the aircraft and the wind tunnel data and the actuator error, the adjustment coefficient k is set;

[0020] Step S5: Establishing the amplitude model of the roll angle command:

[0021] Set the initial roll angle to γ dn , the expected roll angle γ up If it is π or -π, the amplitude A0 at the time of starting the rolling command can be calculated according to the following formula:

[0022]

[0023] Step S6: Establish a roll angle change frequency model:

[0024] The maximum roll control torque M at the time of starting the roll command xδ0 Substituting the amplitude A0 into the following formula, we can find the roll angle change frequency when the roll command is started.

[0025]

[0026] In the formula, J x is the polar moment of inertia;

[0027] Step S7: Establishing a roll angle instruction model:

[0028] The amplitude A0 and the roll angle change frequency at the time of starting the roll command Substitute into the following formula to obtain the roll angle command γ c (t):

[0029]

[0030] Step S8, establishing a rudder angle instruction execution model:

[0031] The roll channel adopts PID control, and the integrated control computer controls the roll angle command γ c (t) Calculate the roll channel rudder deflection angle δ r , to achieve active rolling action, namely:

[0032] δ r =k p (γ c -γ)+k i ∫(γ c -γ)dt+k d w x

[0033] In the formula, k p is the proportionality coefficient, k i is the integration coefficient, k dis the differential coefficient, γ is the rolling angle of the aircraft, w x is the aircraft rolling angular velocity.

[0034] Compared with the prior art, the present invention has the following advantages:

[0035] The invention utilizes the derivative relationship between angle and angular velocity and the derivative relationship between cosine and sine to plan the roll angle instruction of a plane-symmetric aircraft flipping 180 degrees as a cosine function with respect to time t, and after planning, both the roll angle and the roll angular velocity are continuous functions; the rotation direction can be adaptively planned according to the initial value of the roll angle at the rotation moment; the maximum roll control moment coefficient of the aircraft is calculated online in real time by utilizing the BP neural network fitting function according to the altitude, Mach number, angle of attack, sideslip angle and maximum available roll rudder deflection angle at the rotation moment, and the aircraft capability is fully utilized without increasing the storage pressure of the integrated control computer; there is no requirement for the integral step length, and the angle planning without static error can be realized without improving the computing power of the integrated control computer; after determining the frequency of the roll angular velocity change, the method can determine the rotation period, and the angle instruction smoothing can be realized without segmentation processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 A flowchart of an online adaptive planning method for a 180-degree roll angle of a plane-symmetric aircraft;

[0037] Figure 2 This is a comparison chart of sine and cosine curves;

[0038] Figure 3 This is a comparison chart of the sine and cosine curves after planning;

[0039] Figure 4 To plan the roll angle command and roll angle velocity curve;

[0040] Figure 5 The roll angle command variation curve with a step size of 0.005 in Example 1;

[0041] Figure 6 The roll angle command variation curve with a step size of 0.0005 in Example 1;

[0042] Figure 7 The roll angular velocity variation curves of the two planning methods in Example 1;

[0043] Figure 8 The rolling rudder control command change curves of the two planning methods in Example 1;

[0044] Fig. 9 is the roll angle command and response change curve in Example 2;

[0045] Fig.10is the roll angular velocity command and response variation curve in Example 2;

[0046] Fig.11 is the rolling rudder angle variation curve in Example 2;

[0047] Fig.12 The roll angle command and response change curve in Example 3;

[0048] Fig.13 is the roll angular velocity command and response change curve in Example 3;

[0049] Fig.14 is the rolling rudder angle variation curve in Example 3;

[0050] Fig.15 This is the roll angle command and response change curve in Example 4. DETAILED DESCRIPTION

[0051] The technical solution of the present invention is further described below in conjunction with the accompanying drawings, but is not limited thereto. Any modification or equivalent replacement of the technical solution of the present invention without departing from the spirit and scope of the technical solution of the present invention should be included in the protection scope of the present invention.

[0052] This embodiment provides an online adaptive planning method for the roll angle of a plane-symmetric aircraft flipping 180 degrees, such as Figure 1 The method shown is as follows:

[0053] Design an angle command plan based on the cosine curve, let:

[0054]

[0055] Where A is the amplitude, is the frequency of the aircraft's roll angular velocity change.

[0056] By taking the derivative of formula (1), we can get:

[0057]

[0058] make A=1, draw yt, xt curves, see Figure 2 From the curve trend, we can see that In the time, x goes from 0-1-0, and y goes from -A-0-A. In order to make y go from 0 to A, y(t) can be replanned as y1(t), that is:

[0059]

[0060] Derivative of formula (3) is:

[0061]

[0062] Draw y1-t, x1-t curves, see Figure 3 .exist During the time, y1(t) changes from 0-A and x1(t) changes from 0-1-0.

[0063] If the roll angle command γ = y1, the expected roll angle value is γ m , the initial value of the rolling angle is γ0, then:

[0064]

[0065] Since the range of the roll angle is [-180°, 180°], the aircraft can be flipped by rotating clockwise or counterclockwise. The expected roll angle is γ m The initial value γ0 of the roll angle at the start time of the roll command can be used for online calculation, and thus the amplitude A in formula (5) can be calculated, that is:

[0066]

[0067] From the above steps, we can know that when planning the roll angle command, it can be planned as a cosine curve about t, see Figure 4 , taking half a period, the roll angle can be changed from γ0 to γ m , the rolling angular velocity is from 0-w max -0.

[0068] Derivative formula (5) yields:

[0069]

[0070] Among them, w x is the rolling angular velocity, Characterizes the frequency of change of the aircraft's roll angular velocity.

[0071] Deriving formula (7), we get:

[0072]

[0073] From the aircraft dynamics equation, we know that:

[0074]

[0075] Among them, M x is the rolling moment, J x is the polar moment of inertia.

[0076] From formula (9), we can see that when t = 0, is the maximum value. Through the analysis of aerodynamic characteristics of aircraft wind tunnel tests, it is known that the roll control moment coefficient is linearly related to the rudder deflection angle. Therefore, when the maximum available rudder deflection angle of the roll channel is determined, the maximum roll moment M can be calculated. xmax ,Right now:

[0077]

[0078] The rolling moment of the aircraft M x The rolling damping moment M of the aircraft itself xx The rolling control torque M generated by the actuator xδ ,Right now:

[0079]

[0080] in, is the rotation derivative with dimension 1, is the dynamic pressure, v is the aircraft speed, s is the characteristic area, l is the aircraft length, δ x is the rolling rudder angle, is the dimensionless rolling angular velocity, The rolling damping moment coefficient can be abbreviated as is the roll control moment coefficient; roll damping moment coefficient and roll control moment coefficient It can be calculated from the relevant data of wind tunnel test, and then formula (11) can be transformed into:

[0081]

[0082] According to the wind tunnel data of the plane-symmetric aircraft, the rolling control moment coefficient It is related to the Mach number, angle of attack, sideslip angle and roll rudder deflection angle. Taking into account the storage space and computing pressure of the aircraft integrated control computer, a three-layer forward feedback BP neural network algorithm is adopted. With the Mach number, angle of attack, sideslip angle and roll rudder deflection angle as input, the wind tunnel test data as samples and the roll control moment coefficient as output, a function for calculating the roll control moment coefficient is fitted. This function can calculate the roll control moment coefficient at the active roll moment without loading the wind tunnel test data into the integrated control computer, thereby reducing the storage and computing pressure of the integrated control computer.

[0083] The calculation process of the three-layer forward feedback BP neural network algorithm is as follows: design the input layer to have 4 layers (roll rudder deflection angle, Mach number, angle of attack, sideslip angle), the hidden layer to have 9 layers, and the output layer to have 1 layer (roll control torque coefficient). Use the wind tunnel test data to give the student and teacher samples, and train the weights from the input layer to the hidden layer, the weights from the hidden layer to the output layer, the thresholds from the input layer to the hidden layer, and the thresholds from the hidden layer to the output layer. The function fitted by the three-layer forward feedback BP neural network algorithm is:

[0084] m xδ =f(ma,α,β,δ x ) (13)

[0085] Among them, m xδ is the roll control moment coefficient, ma is the Mach number, α is the angle of attack, β is the sideslip angle, δ x is the rolling rudder angle.

[0086] It can be seen from formula (11) that the rolling moment of the aircraft is related to the dynamic pressure q. In order to obtain the dynamic pressure at the rolling moment, it is necessary to calculate the air density ρ0 using a table lookup method based on the altitude H0 at the time of launch.

[0087] Given the maximum available rudder deflection angle of the roll channel and the Mach number, angle of attack and sideslip angle at the time of release, the roll control torque coefficient M in the current state can be calculated using formula (13): xδmax :

[0088]

[0089] Through and According to the data comparison and analysis, when the absolute value of the roll rudder angle is greater than 10°, the roll damping moment coefficient is smaller than the roll control moment coefficient, so it is ignored. Considering the deviation between the actual flight aerodynamic characteristics of the aircraft and the wind tunnel test data and the actuator error, redundant design is required. Now set the adjustment coefficient k and transform formula (14) into:

[0090] M x ′=kM xδmax (15)

[0091] Substituting formula (15) into formula (10) and setting the adjustment coefficient k, the maximum frequency of the aircraft roll angle change can be calculated: Right now:

[0092]

[0093] The specific implementation steps are as follows:

[0094] Step S1, establishing an air density model:

[0095] According to the altitude H0 at the time of release, the air density ρ0 is calculated by using a table lookup method.

[0096] Step S2: Establishing a rolling control torque coefficient model:

[0097] Given the maximum rudder deflection angle δ of the rolling channel rmand the Mach number ma0, angle of attack α0 and sideslip angle β0 at the time of starting the roll command. Using formula (17), the maximum roll control torque coefficient m can be calculated: x0 :

[0098] m x0 =f(ma0,α0,β0,δ rm ) (17)

[0099] Step S3: Establishing a rolling control torque model:

[0100] The air density ρ0 and the maximum rolling control moment coefficient m obtained in step S1 and step S2 are used. x0 , use formula (18) to calculate the maximum rolling control torque M at the time of starting the rolling command xδ0 :

[0101]

[0102] Step S4, setting the adjustment coefficient:

[0103] Taking into account the deviation between the actual flight aerodynamic characteristics of the aircraft and the wind tunnel data and the actuator error, the adjustment coefficient k is set to 0.8.

[0104] Step S5: Establishing the amplitude model of the roll angle command:

[0105] Set the initial roll angle to γ dn , the expected roll angle γ up is π or -π, the amplitude A0 can be calculated according to formula (19):

[0106]

[0107] Step S6: Establish a roll angle change frequency model:

[0108] Substituting formula (18) and formula (19) into formula (20), we can calculate the rolling angle change frequency:

[0109]

[0110] Step S7: Establishing a roll angle instruction model:

[0111] Substituting formula (19) and formula (20) into formula (21), the roll angle command γ can be obtained: c (t), that is:

[0112]

[0113] Step S8: Establishing the rudder angle instruction execution model:

[0114] The roll channel adopts PID control, and the integrated control computer controls the roll angle command γ c (t) Calculate the roll channel rudder deflection angle δ r , to achieve active rolling action, namely:

[0115] δ r =k p (γ c -γ)+k i ∫(γ c -γ)dt+k d w x (twenty two)

[0116] In the formula, k p is the proportionality coefficient, k i is the integration coefficient, k d is the differential coefficient, γ is the rolling angle of the aircraft, w x is the aircraft rolling angular velocity.

[0117] Example 1

[0118] Set the launch conditions: altitude H = 10000m, Mach number Ma = 0.51, and simulation time is set to 4s. Given the maximum roll angle change rate The planning method of the present invention and the S-curve planning method are used to perform single feature point simulation comparison. The simulation curves are shown in FIG. Figure 5 From the simulation curve, it can be seen that there is a static error between the roll angle command planned by the S-curve planning method and the expected roll angle command. After analysis, it is caused by the integral step of the Euler method. The Euler integral step is set from 0.005 to 0.0005. The simulation curve is shown in Figure 6 From the simulation curve, it can be seen that the S-curve planning method is greatly affected by the integration step length. However, considering the computing power of the integrated control computer, the planning method of the present invention is not affected by the integration step length, and can plan the roll angle instruction without static error, thereby reducing the demand for the computing power of the integrated control computer.

[0119] Figure 7 The comparison curve of active roll command roll angular velocity planned by the planning method of the present invention and the S-curve planning method is shown. It can be seen from the curve that the peak value of the roll angular velocity of the present invention is smaller than that of the S-curve, which is more beneficial to the stable control of the aircraft. Figure 8 The comparison curve of active roll rudder control instructions planned by the planning method of the present invention and the S-curve planning method is shown. It can be seen from the curve that the changing trends of the two are consistent in the starting and reaching stages of active roll, but in the intermediate process, the rudder control instructions of the planning method of the present invention are smoother, which can reduce the requirements on the actuator.

[0120] Example 2

[0121] The planning method of the present invention is used to perform six-degree-of-freedom simulation verification, and the delivery conditions are set: the altitude is H = 10000m, the target altitude H0 = 0m, the delivery ground speed Vd = 550Km / h, and the range S = 8000m. When executing the active roll command, the Mach number ma0 = 0.51, the angle of attack α0 = -2.784°, and the sideslip angle β0 = -0.511°. Given the maximum roll rudder deflection angle δ max =-10°, the planned rolling angular velocity change rate The initial value of the roll angle is γ dn =0.0004rad. Set the adjustment factor k=0.8.

[0122] The simulation results are as follows Figure 9-11 As shown, from Fig. 9 It can be seen that the roll angle can track the command in a timely and stable manner without overshoot. The command planning time is 1.55s, the response time is 1.86s, and both the command and the response reach the expected angle without static error. Fig.10 It can be seen that the rolling angular velocity of the aircraft is basically consistent with the planned rolling angular velocity; Fig.11 It can be seen that the maximum rudder angle of the roll channel is -8.076°.

[0123] Example 3

[0124] The planning method of the present invention is used to perform six-degree-of-freedom simulation verification, and the launch conditions are set: the altitude is H = 3000m, the target altitude H0 = 0m, the launch ground speed Vd = 900Km / h, and the range S = 5000m. When executing the active roll command, the Mach number ma0 = 0.757, the angle of attack α0 = -0.923°, and the sideslip angle β0 = -0.866°. Given the maximum roll rudder deflection angle δ max =-10°, the planned rolling angular velocity change rate The initial value of the roll angle is γ dn =0.0033rad. Set adjustment factor k=0.8.

[0125] The simulation results are as follows Figure 12-14 As shown, from Fig.12 It can be seen that the roll angle can track the command in a timely and stable manner without overshoot, the command planning time is 0.6s, and the response time is 0.75s; Fig.13 It can be seen that the rolling angular velocity of the aircraft is basically consistent with the planned rolling angular velocity; Fig.14 It can be seen that the maximum rudder angle of the roll channel is -9.129°.

[0126] From the simulation results, the active rolling time of Example 2 is significantly longer than that of Example 3, but the required rudder deflection angle is similar, which shows that the planning method of the present invention can realize real-time online active rolling cycle according to different delivery conditions.

[0127] Example 4

[0128] Select the delivery conditions of Example 2, and the initial value of the rolling angle is γ dn =-0.0878rad, simulation results are shown in Fig.15 , the expected instruction γ up = -π. From the simulation results, it can be seen that the planning method of the present invention can determine the clockwise or counterclockwise rolling strategy online according to the initial rolling angle, thereby meeting the requirements of active rolling online adaptive planning.

Claims

1. An online adaptive planning method for the roll angle of a 180-degree flip of a plane-symmetric aircraft, characterized in that The method comprises the following steps: Step S1, establishing an air density model: According to the altitude H0 at the time of delivery, the air density ρ0 is calculated by using a table lookup method; Step S2: Establishing a rolling control torque coefficient model: Given the maximum rudder deflection angle δ of the rolling channel rm The Mach number ma0, angle of attack α0 and sideslip angle β0 at the time of starting the roll command are used. The three-layer forward feedback BP neural network algorithm is used to calculate the Mach number ma0, angle of attack α0, sideslip angle β0 and the maximum rudder deflection angle δ of the roll channel. rm As input, taking wind tunnel test data as sample, the maximum rolling control moment coefficient m x0 As output, the maximum rolling control moment coefficient m is fitted. x0 Function: m x0 =f(ma0,α0,β0,δ rm ) Step S3: Establishing a rolling control torque model: Using the air density ρ0 obtained in step S1 and the maximum rolling control moment coefficient m obtained in step S2 x0 The maximum roll control torque M at the time of starting the roll command is calculated by the following formula: xδ0 : In the formula, v is the speed of the aircraft, s is the characteristic area, and l is the length of the aircraft; Step S4, setting the adjustment coefficient: Considering the deviation between the actual flight aerodynamic characteristics of the aircraft and the wind tunnel data and the actuator error, the adjustment coefficient k is set; Step S5: Establishing the amplitude model of the roll angle command: Set the initial roll angle to γ dn , the expected roll angle γ up If it is π or -π, the amplitude A0 at the time of starting the rolling command can be calculated according to the following formula: Step S6: Establish a roll angle change frequency model: The maximum roll control torque M at the time of starting the roll command xδ0 Substituting the amplitude A0 into the following formula, we can find the roll angle change frequency when the roll command is started. In the formula, J x is the polar moment of inertia; Step S7: Establishing a roll angle instruction model: The amplitude A0 and the roll angle change frequency at the time of starting the roll command Substitute into the following formula to obtain the roll angle command γ c (t): Step S8, establishing a rudder angle instruction execution model: The roll channel adopts PID control, and the integrated control computer controls the roll angle command γ c (t) Calculate the roll channel rudder deflection angle δ r , to achieve active rolling action, namely: d r =k p (c c -c)+k i ∫(γ c -c)dt+k d w x In the formula, k p is the proportionality coefficient, k i is the integration coefficient, k d is the differential coefficient, γ is the rolling angle of the aircraft, w x is the aircraft's roll angular velocity.

2. The method for online adaptive planning of the roll angle of a plane-symmetric aircraft flipping 180 degrees according to claim 1 is characterized in that The k=0.8.

Images