Airplane anti-skid brake control method and system based on brake torque dynamic prediction

Abstract

The invention provides an aircraft anti-skid brake control method and system based on brake torque dynamic prediction, and belongs to the field of aircraft brake control, the method comprises the following steps: obtaining test data of multiple groups of brake control parameters, including brake pressure, brake temperature and aircraft wheel speed; the method comprises the following steps of: performing an anti-skid brake system dynamic test on an aircraft brake wheel on a test bed to obtain a plurality of brake disc friction coefficients corresponding to a plurality of groups of brake control parameter test data; a neural network model is constructed, the brake control parameters in the test state serve as the input of the neural network model, the brake disc friction coefficient serves as the output of the neural network model, and a dynamic soft measurement model of the brake disc friction coefficient of the aircraft anti-skid brake system is established; based on the brake disc friction coefficient dynamic soft measurement model, a to-be-predicted brake control parameter is input to obtain a predicted brake disc friction coefficient, and based on the predicted brake disc friction coefficient and the brake pressure, the brake torque used for controlling the anti-skid brake of the airplane is obtained.

Description

Technical Field

[0001] The present application belongs to the technical field of aircraft brakes, and in particular relates to an aircraft anti-skid brake control method and system based on dynamic prediction of braking torque. Background Art

[0002] The aircraft brake system is a relatively independent subsystem on the aircraft. Its function is to bear the static weight and dynamic impact load of the aircraft and absorb the kinetic energy of the aircraft when landing, so as to achieve the braking and control of the aircraft's takeoff, landing and taxiing. At present, the aircraft anti-skid brake system widely adopts hydraulic actuation mode, which introduces hydraulic pressure to the brake device to generate positive pressure on the brake device, and control the aircraft deceleration under the friction of the brake disc.

[0003] The aircraft anti-skid brake system is controlled by pressure control. It is usually assumed that the friction coefficient of the brake disc is a constant value, so the brake torque value can be indirectly obtained by controlling the pressure. However, the actual situation is that the friction coefficient of the brake disc is a nonlinear and dynamically changing parameter, and there are many factors that affect the friction coefficient of the brake disc. However, how to determine the dynamic change of the friction coefficient of the brake disc during the landing deceleration process of the aircraft to ensure that the torque can be accurately controlled during the dynamic anti-skid braking process of the aircraft, thereby ensuring that the aircraft will not have a brake drag phenomenon during the braking process, preventing the aircraft from having a tire blowout during braking, and improving the safety of the aircraft's braking deceleration, the prior art does not provide any method. Summary of the invention

[0004] The purpose of the present application is to provide an aircraft anti-skid braking control method and system based on dynamic prediction of braking torque to solve or alleviate at least one problem in the background technology.

[0005] The technical solution of the present application is: an aircraft anti-skid braking control method based on dynamic prediction of braking torque, comprising:

[0006] obtaining test data of multiple sets of brake control parameters, the brake control parameters including brake pressure, brake temperature and wheel speed;

[0007] By conducting an anti-skid brake system dynamic test on the aircraft brake wheel on the test bench, a plurality of brake disc friction coefficients corresponding to a plurality of sets of brake control parameter test data are obtained;

[0008] A neural network model was constructed, with the brake pressure, brake temperature and wheel speed under the test state as the input of the neural network model, and the brake disc friction coefficient as the output of the neural network model, to establish a dynamic soft measurement model of the brake disc friction coefficient of the aircraft anti-skid brake system;

[0009] Based on the dynamic soft measurement model of the brake disc friction coefficient of the aircraft anti-skid brake system, the brake pressure, brake temperature and wheel speed to be predicted are input to obtain the predicted brake disc friction coefficient. Based on the predicted brake disc friction coefficient and brake pressure, the brake torque used to control the aircraft anti-skid brake is obtained.

[0010] Preferably, the process of obtaining test data of multiple sets of brake control parameters, the brake control parameters including brake pressure P, brake temperature T and wheel speed V, comprises:

[0011] S11, determining M brake control parameter test data within the range of variation of brake pressure P, brake temperature T and wheel speed V respectively;

[0012] S12, combining multiple brake control parameter test data to obtain M×M×M groups of test states;

[0013] S13, measuring the corresponding M×M×M brake disc friction coefficient μ values ​​by performing an anti-skid brake system dynamic test on the aircraft brake wheels on the test bench for the M×M×M groups of test states.

[0014] Preferably, the process of determining M brake control parameter test data within the brake pressure P variation range includes:

[0015] Determine the range of brake pressure P as P 1 ≤P≤P 2 , P 1 is the minimum pressure, P 2 is the maximum pressure;

[0016] Calculate the minimum pressure P 1 With maximum pressure P 2 The difference ΔP = P 2 -P 1 ;

[0017] In [P 1 ,P 2 ] select M points within the range of P 1 , P 2

[0018] Preferably, the process of determining M brake control parameter test data within the brake temperature T variation range includes:

[0019] Determine the range of brake temperature T as T 1 ≤T≤T 2 , T 1 is the minimum temperature, T 2 is the maximum temperature;

[0020] Calculate the minimum temperature T 1 With the maximum temperature T 2 The difference ΔT = T 2 -T 1 ;

[0021] In [T 1 ,T 2 ] Select M points within the range of T 1 , T 2 .

[0022] Preferably, the process of determining M brake control parameter test data within the range of wheel speed V includes:

[0023] Determine the range of wheel speed V as V 1 ≤V≤V 2 , V 1 is the minimum speed, V 2 is the maximum speed;

[0024] Calculate the minimum speed V 1 With the maximum speed V 2 The difference ΔV = V 2 -V 1 ;

[0025] In [V 1 ,V 2 ] select M points within the range of V 1 , V 2 .

[0026] Preferably, the neural network model includes a BP neural network model, a radial basis function neural network model, a recursive neural network model and a convolutional neural network model.

[0027] Preferably, when the neural network model is a BP neural network model, the number of hidden layer variables of the BP neural network model ranges from 4 to 10, and the number of layers of the BP neural network model includes 3 to 5 layers.

[0028] Preferably, the process of establishing a dynamic soft-sensing model of the friction coefficient of the brake disc of the aircraft anti-skid brake system comprises:

[0029] S31, determine the input and output relationship of the neural network model:

[0030] Where: x j (j=1,2,…,N) is the input signal, τ i is the threshold value, ω jiis the connection weight from neuron j to neuron i, f() is the transfer function, input signal X = [P, T, V], output signal y i is the friction coefficient of the brake disc μ;

[0031] S32, substitute the M×M×M groups of test conditions and the corresponding M×M×M brake disc friction coefficient μ values ​​into the relation: Training in

[0032] S33, after the training is completed, save the connection weight ω obtained through training ji And transfer function f(), so as to establish a dynamic soft measurement model of aircraft brake disc friction coefficient.

[0033] In addition, the present application provides an aircraft anti-skid brake control system for implementing any of the above-described aircraft anti-skid brake control methods based on dynamic prediction of braking torque, comprising: an anti-skid brake control box, an electro-hydraulic pressure servo valve, a pressure sensor, a brake temperature sensor, and a wheel speed sensor;

[0034] The pressure sensor measures the brake pressure at the outlet of the electro-hydraulic pressure servo valve and feeds back to the anti-skid brake control box, the wheel speed sensor measures the wheel speed and feeds back to the anti-skid brake control box, and the brake temperature sensor measures the temperature of the brake device and feeds back to the anti-skid brake control box;

[0035] The anti-skid brake control box calculates the brake disc friction coefficient according to the measured brake pressure, wheel speed, and brake temperature, and obtains the brake pressure used for anti-skid braking of the aircraft according to the brake disc friction coefficient.

[0036] Preferably, the process of the anti-skid brake control box calculating the brake pressure for performing anti-skid braking of the aircraft according to the friction coefficient of the brake disc comprises:

[0037] The real-time braking torque under the current state is calculated by using the brake disc friction coefficient and brake pressure predicted under the current state;

[0038] Compare the real-time braking torque in the current state with the preset braking torque corresponding to the braking command. When the real-time braking torque is greater than the preset braking torque, the braking pressure is reduced. At the same time, the anti-skid brake control box obtains the updated brake disc friction coefficient again through the updated brake pressure, wheel speed and brake temperature, and obtains the updated braking torque through calculation, and performs system closed-loop control accordingly.

[0039] The aircraft anti-skid brake control method and system provided in the present application target the infinite combinations of three variables that affect the friction coefficient of the aircraft brake disc. Only a limited sample size of test is needed to obtain the brake disc friction coefficient μ under any state, which greatly reduces the number of tests, saves test cycles and expenses. During operation, the anti-skid brake system measures the current real-time brake pressure, brake temperature and wheel speed, and can calculate the real-time brake disc friction coefficient through a soft measurement model, and finally can conveniently obtain the current real brake torque value, thereby realizing precise control of the aircraft anti-skid brake. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solution provided by the present application, the following is a brief introduction to the accompanying drawings. Obviously, the accompanying drawings described below are only some embodiments of the present application.

[0041] Figure 1 This is a schematic diagram of the aircraft anti-skid braking control method based on dynamic prediction of braking torque in the present application.

[0042] Figure 2 Schematic diagram of the neural network model in this application.

[0043] Figure 3 This is a schematic diagram of the aircraft anti-skid brake control system of the present application. DETAILED DESCRIPTION

[0044] In order to make the purpose, technical solutions and advantages of the implementation of this application clearer, the technical solutions in the embodiments of this application will be described in more detail below in conjunction with the drawings in the embodiments of this application.

[0045] The purpose of this application is to obtain the real-time friction coefficient of the brake disc at different times during the aircraft anti-skid braking process, and then control the braking torque in real time according to the braking pressure and the friction coefficient of the brake disc, to ensure that the braking command and the real-time braking torque of the aircraft are in a truly proportional relationship, and to ensure that the relationship between the braking command and the braking torque of the aircraft during deceleration meets the pilot's braking expectations.

[0046] To this end, the present application provides an aircraft anti-skid braking control method based on dynamic prediction of braking torque, the method comprising the following steps:

[0047] S10, obtaining test data of multiple groups of brake control parameters, the brake control parameters including brake pressure P, brake temperature T and wheel speed V, including:

[0048] S11, determining M brake control parameter test data within the range of variation of brake pressure P, brake temperature T and wheel speed V respectively;

[0049] S12 combines the multiple brake control parameter test data to obtain M×M×M groups of test states;

[0050] S13, performing an anti-skid brake system dynamic test on the aircraft brake wheels for M×M×M groups of test states on an inertia test bench, and measuring the corresponding M×M×M brake disc friction coefficient μ values.

[0051] In this application, the process of selecting M points of brake pressure P is as follows:

[0052] Determine the range of brake pressure P as P 1 ≤P≤P 2 , P 1 is the minimum pressure, P 2 is the maximum pressure;

[0053] Calculate the minimum pressure P 1 With maximum pressure P 2 The difference ΔP = P 2 -P 1 ;

[0054] In [P 1 ,P 2 ] select M points within the range of P 1 , P 2 .

[0055] In this application, the process of selecting M points of brake temperature T is as follows:

[0056] Determine the range of brake temperature T as T 1 ≤T≤T 2 , T 1 is the minimum temperature, T 2 is the maximum temperature;

[0057] Calculate the minimum temperature T 1 With the maximum temperature T 2 The difference ΔT = T 2 -T 1 ;

[0058] In [T 1 ,T 2 ] Select M points within the range of T 1 , T 2 .

[0059] In this application, the process of selecting M points of wheel speed V is as follows:

[0060] Determine the range of wheel speed V as V 1≤V≤V 2 , V 1 is the minimum speed, V 2 is the maximum speed;

[0061] Calculate the minimum speed V 1 With the maximum speed V 2 The difference ΔV = V 2 -V 1 ;

[0062] In [V 1 ,V 2 ] select M points within the range of V 1 , V 2 .

[0063] For example, in some embodiments of the present application, each brake control parameter variable among the brake pressure P, brake temperature T and wheel speed V can be measured at 6 points in the entire variation range at a rate of 20% each time, as follows:

[0064] The range of brake pressure P is P 1 ≤P≤P 2 , then calculate P 1 With P 2 The difference is as follows, ΔP = P 2 -P 1 , in [P 1 ,P 2 ] Select 6 points within the range of P 1 , P 1 +0.2ΔP, P 1 +0.4ΔP, P 1 +0.6ΔP, P 1 +0.8ΔP, P 2 ;

[0065] The same method is used to select the same 6 points for the brake temperature T, and the range of the brake temperature T is T 1 ≤T≤T 2 , then calculate T 1 With T 2 The difference is as follows, ΔT = T 2 -T 1 , in [T 1 ,T 2 ] Select 6 points within the range of T 1 , T 1 +0.2ΔT, T 1 +0.4ΔT, T 1 +0.6ΔT, T 1 +0.8ΔT, T 2 ;

[0066] The same method is used to select the same 6 points for the wheel speed V. The range of the wheel speed V is V 1 ≤V≤V 2 , then calculate V 1 With V 2 The difference is as follows, ΔV = V 2 -V 1 , in [V 1 ,V 2 ] Select 6 points within the range of V 1 、V 1 +0.2ΔV, V 1 +0.4ΔV, V 1 +0.6ΔV, V 1 +0.8ΔV, V 2 .

[0067] S20, performing an anti-skid brake system dynamic test on the aircraft brake wheel on an inertia test bench to obtain a plurality of brake disc friction coefficients μ corresponding to a plurality of groups of brake control parameter test data.

[0068] For example, in this embodiment of the present application, after combining each value of the above-mentioned brake pressure P, brake temperature T, and wheel speed V, 6×6×6=216 groups of test conditions can be obtained. All 216 groups of test conditions are physically tested and the brake disc friction coefficient μ values ​​under different conditions are measured.

[0069] S30, constructing a neural network model, taking the brake pressure P, brake temperature T and wheel speed V in step 10 as inputs of the neural network model, taking the brake disc friction coefficient μ in step 20 as output of the neural network model, and establishing a dynamic soft measurement model of the brake disc friction coefficient of the aircraft anti-skid brake system;

[0070] like Figure 2 As shown, the neural network model in this application includes BP neural network model, radial basis function neural network model, recursive neural network model and convolutional neural network model, etc. Among them, if the neural network adopted is BP neural network model, the number of hidden layer variables of BP neural network model ranges from 4 to 10, and the number of layers of BP neural network model includes 3 to 5 layers.

[0071] In this application, a dynamic soft measurement model of the friction coefficient of the brake disc of the aircraft anti-skid brake system is established, which specifically includes:

[0072] S31, determine the input and output relationship of the neural network model as:

[0073]

[0074] Where: xj (j=1,2,…,N) is the input signal, τ i is the threshold value, ω ji is the connection weight from neuron j to neuron i, f() is the transfer function, input signal X = [P, T, V], output signal y i is the friction coefficient of the brake disc μ;

[0075] S32, substitute the M×M×M groups of test conditions and the corresponding M×M×M brake disc friction coefficient μ values ​​into the relation: Training in

[0076] S33, after the training is completed, save the connection weight ω obtained through training ji And transfer function f(), so as to establish a dynamic soft measurement model of aircraft brake disc friction coefficient.

[0077] For example, in this embodiment of the present application, a neural network model with brake pressure P, brake temperature T, and wheel speed V as input and the measured brake disc friction coefficient μ as output adopts a three-layer BP network structure, in which the number of input layer variables is 3, the number of hidden layer (middle layer) variables is 8, and the number of output layer variables is 1. The input and output relationship of each layer is: Where: x j (j=1,2,…,n) is the input signal; τ i is the threshold value; ω ji is the connection weight from neuron j to neuron i; f() is the transfer function.

[0078] The input parameter X = [P, T, V], the output layer parameter is y, that is, the brake disc friction coefficient μ.

[0079] Substitute the 216 sets of input values ​​and results measured in the experiment into The BP neural network toolbox function trainbp in MATLAB software is used to train the BP network. After the training is completed, the connection weights ω obtained by the training are saved. ji and transfer function f().

[0080] In the prediction process, for any set of input X 0 =[P 0 ,T 0 ,V 0 ], and bring it into the trained neural network system to obtain the y corresponding to this input. This method only needs to conduct a limited number of samples for the infinite combinations of the three variables brake pressure P, brake temperature T, and wheel speed V to obtain the brake disc friction coefficient μ under any state.

[0081] S40, based on the dynamic soft measurement model of the brake disc friction coefficient of the aircraft anti-skid brake system, the brake pressure P, brake temperature T and wheel speed V to be predicted are input to obtain the predicted brake disc friction coefficient μ, and the brake torque used to control the aircraft anti-skid brake can be obtained based on the predicted brake disc friction coefficient μ and the brake pressure.

[0082] like Figure 3 As shown, on this basis, the present application also provides an aircraft anti-skid brake control system based on dynamic prediction of brake torque, the control system includes: an anti-skid brake control box 11, an electro-hydraulic pressure servo valve 12, a pressure sensor 13, a brake temperature sensor 14, and a wheel speed sensor 15. Among them, the anti-skid brake control box 11 receives the driver's operation command signal from the command sensor, controls the electro-hydraulic pressure servo valve 12 to output the corresponding brake pressure according to the preset brake control signal, the pressure sensor 13 measures the outlet pressure of the electro-hydraulic pressure servo valve 12 and feeds it back to the anti-skid brake control box 11, the wheel speed sensor 15 measures the wheel speed and feeds it back to the anti-skid brake control box 11, and the brake temperature sensor 15 measures the temperature of the brake device and feeds it back to the anti-skid brake control box 11.

[0083] The anti-skid brake control box 11 calculates the real-time brake torque in the current state through the predicted brake disc friction coefficient μ and brake pressure P in the current state, and compares the real-time brake torque in the current state with the preset brake torque corresponding to the brake command. When the real-time brake torque is greater than the preset brake torque, the brake pressure is reduced. At the same time, the anti-skid brake control box 11 obtains the updated brake disc friction coefficient again through the updated brake pressure, wheel speed and brake temperature, and obtains the updated brake torque through calculation, and performs system closed-loop control accordingly.

[0084] The aircraft anti-skid brake control method and system provided in the present application target the infinite combinations of three variables that affect the friction coefficient of the aircraft brake disc. Only a limited sample size of test is needed to obtain the brake disc friction coefficient μ under any state, which greatly reduces the number of tests, saves test cycles and expenses. During operation, the anti-skid brake system measures the current real-time brake pressure, brake temperature and wheel speed, and can calculate the real-time brake disc friction coefficient through a soft measurement model, and finally can conveniently obtain the current real brake torque value, thereby realizing precise control of the aircraft anti-skid brake.

[0085] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. An aircraft anti-skid braking control method based on dynamic prediction of braking torque, characterized in that: include: obtaining test data of multiple sets of brake control parameters, the brake control parameters including brake pressure, brake temperature and wheel speed; By conducting an anti-skid brake system dynamic test on the aircraft brake wheel on the test bench, a plurality of brake disc friction coefficients corresponding to a plurality of sets of brake control parameter test data are obtained; A neural network model was constructed, with the brake pressure, brake temperature and wheel speed under the test state as the input of the neural network model, and the brake disc friction coefficient as the output of the neural network model, to establish a dynamic soft measurement model of the brake disc friction coefficient of the aircraft anti-skid brake system; Based on the dynamic soft measurement model of the brake disc friction coefficient of the aircraft anti-skid brake system, the brake pressure, brake temperature and wheel speed to be predicted are input to obtain the predicted brake disc friction coefficient. Based on the predicted brake disc friction coefficient and brake pressure, the brake torque used to control the aircraft anti-skid brake is obtained.

2. The aircraft anti-skid braking control method based on dynamic prediction of braking torque according to claim 1, characterized in that: The process of obtaining test data of multiple sets of brake control parameters, wherein the brake control parameters include brake pressure P, brake temperature T and wheel speed V, comprises: S11, determining M brake control parameter test data within the range of variation of brake pressure P, brake temperature T and wheel speed V respectively; S12, combining multiple brake control parameter test data to obtain M×M×M groups of test states; S13, measuring the corresponding M×M×M brake disc friction coefficient μ values ​​by performing an anti-skid brake system dynamic test on the aircraft brake wheels on the test bench for the M×M×M groups of test states.

3. The aircraft anti-skid braking control method based on dynamic prediction of braking torque according to claim 2, characterized in that: The process of determining M brake control parameter test data within the brake pressure P variation range includes: Determine the change range of the brake pressure P as P1≤P≤P2, where P1 is the minimum pressure and P2 is the maximum pressure; Calculate the difference between the minimum pressure P1 and the maximum pressure P2 ΔP=P2-P1; Select M points in the range [P1,P2] and they are P1, P2.

4. The aircraft anti-skid braking control method based on dynamic prediction of braking torque according to claim 2, characterized in that: The process of determining M brake control parameter test data within the brake temperature T variation range includes: Determine the range of the brake temperature T as T1≤T≤T2, where T1 is the lowest temperature and T2 is the highest temperature; Calculate the difference between the lowest temperature T1 and the highest temperature T2, ΔT = T2 - T1; Select M points in the range [T1, T2], namely T1, T2.

5. The aircraft anti-skid braking control method based on dynamic prediction of braking torque according to claim 2, characterized in that: The process of determining M brake control parameter test data within the range of wheel speed V includes: Determine the range of the wheel speed V as V1≤V≤V2, V1 is the lowest speed, and V2 is the highest speed; Calculate the difference ΔV between the lowest speed V1 and the highest speed V2 = V2-V1; Select M points in the range [V1, V2], namely V1, V2.

6. The aircraft anti-skid braking control method based on dynamic prediction of braking torque according to any one of claims 3 to 5, characterized in that: The neural network models include BP neural network model, radial basis function neural network model, recursive neural network model and convolutional neural network model.

7. The aircraft anti-skid braking control method based on dynamic prediction of braking torque according to claim 6, characterized in that: When the neural network model is a BP neural network model, the number of hidden layer variables of the BP neural network model ranges from 4 to 10, and the number of layers of the BP neural network model includes 3 to 5 layers.

8. The aircraft anti-skid braking control method based on dynamic prediction of braking torque according to claim 7, characterized in that: The process of establishing a dynamic soft-sensing model of friction coefficient of brake disc of aircraft anti-skid brake system includes: S31, determine the input and output relationship of the neural network model: Where: x j (j=1,2,…,N) is the input signal, τ i is the threshold value, ω ji is the connection weight from neuron j to neuron i, f() is the transfer function, input signal X = [P, T, V], output signal y i is the friction coefficient of the brake disc μ; S32, substitute the M×M×M groups of test conditions and the corresponding M×M×M brake disc friction coefficient μ values ​​into the relation: Training in S33, after the training is completed, save the connection weight ω obtained through training ji And transfer function f(), so as to establish a dynamic soft measurement model of aircraft brake disc friction coefficient.

9. An aircraft anti-skid brake control system for implementing the aircraft anti-skid brake control method based on dynamic prediction of braking torque as claimed in any one of claims 1 to 8, characterized in that: include: Anti-skid brake control box, electro-hydraulic pressure servo valve, pressure sensor, brake temperature sensor, wheel speed sensor; The pressure sensor measures the brake pressure at the outlet of the electro-hydraulic pressure servo valve and feeds back to the anti-skid brake control box, the wheel speed sensor measures the wheel speed and feeds back to the anti-skid brake control box, and the brake temperature sensor measures the temperature of the brake device and feeds back to the anti-skid brake control box; The anti-skid brake control box calculates the brake disc friction coefficient according to the measured brake pressure, wheel speed, and brake temperature, and obtains the brake pressure used for anti-skid braking of the aircraft according to the brake disc friction coefficient.

10. The aircraft anti-skid brake control system according to claim 9, characterized in that: The process of the anti-skid brake control box calculating the brake pressure for performing anti-skid braking of the aircraft according to the friction coefficient of the brake disc includes: The real-time braking torque under the current state is calculated by using the brake disc friction coefficient and brake pressure predicted under the current state; Compare the real-time braking torque in the current state with the preset braking torque corresponding to the braking command. When the real-time braking torque is greater than the preset braking torque, the braking pressure is reduced. At the same time, the anti-skid brake control box obtains the updated brake disc friction coefficient again through the updated brake pressure, wheel speed and brake temperature, and obtains the updated braking torque through calculation, and performs system closed-loop control accordingly.

Images