Airplane control surface oscillation fault monitoring method based on multi-scale wavelet transform

Abstract

The invention discloses an aircraft control surface oscillation fault monitoring method based on multi-scale wavelet transformation, and the method comprises the steps: carrying out the calculation through a PID (Proportion Integration Differentiation) control law through an autopilot according to an input flight path command, and obtaining a control surface control instruction deltadepth; a control surface deflection prediction value delta est is generated based on the nonlinear model and the control surface control instruction, the control surface deflection prediction value and an actual deflection value delta mea collected by the sensor module are subtracted to generate a residual signal Rn, and the residual signal is decomposed into a plurality of sub-band signals through wavelet decomposition; wavelet energy calculation is carried out on the sub-bands possibly containing the control surface oscillation fault signals, whether a fault occurs or not is judged according to the magnitude of an energy comparison threshold value, and the judgment threshold value is determined through Monte Carlo simulation. According to the method, instruction signals are separated by generating residual errors, wavelet transform processing is carried out on residual error signals, interference of noise and irrelevant frequency signals on monitoring performance is reduced, meanwhile, small-amplitude oscillation signals can be effectively monitored according to sub-band wavelet energy serving as a fault criterion, and the monitoring coverage rate of oscillation faults is remarkably increased.

Description

Technical Field

[0001] The invention relates to a technology in the field of aircraft manufacturing, in particular to an aircraft control surface oscillation fault monitoring method based on multi-scale wavelet transform. Background Art

[0002] For aircraft using fly-by-wire flight control systems, during flight, the extremely complex working environment may cause serious non-command oscillation problems during takeoff, landing or cruising. The oscillation failure of the main flight control system control surfaces such as ailerons, elevators and rudders is mainly caused by the pseudo-sinusoidal command signals generated by electronic components in the fault mode, which causes the main flight control system control surfaces to produce corresponding periodic vibrations, that is, causing the aircraft to have a control surface oscillation failure. After the control surface oscillation failure occurs, the impact on the aircraft is mainly reflected in the impact on the body structure and the impact on the flight control quality. At present, the commonly used algorithms for monitoring control surface oscillation are: sensor signal-based methods, data-driven methods and model-based methods. Among them, the sensor signal-based method requires the introduction of hardware redundancy; the data-based method is sensitive to noise and it is difficult to cover all frequency band signals at the same time; the model-based method requires the construction of a specific model based on the flight principle and component parameters. Summary of the invention

[0003] In view of the above-mentioned deficiencies in the prior art, the present invention proposes an aircraft control surface oscillation fault monitoring method based on multi-scale wavelet transform. The command signal is separated by generating a residual, and the residual signal is processed by wavelet transform, thereby reducing the interference of noise and irrelevant frequency signals on the monitoring performance. At the same time, based on the sub-band wavelet energy as the fault criterion, the oscillation signal with small amplitude can be effectively monitored, and the monitoring coverage of oscillation faults can be significantly improved.

[0004] The present invention is achieved through the following technical solutions:

[0005] The present invention relates to a method for monitoring aircraft control surface oscillation fault based on multi-scale wavelet transform, comprising:

[0006] Step 1: Receive the control command δ generated by the autopilot desired ;

[0007] The control instructions of the control surface are obtained by the autopilot through the PID control law according to the input track command.

[0008] Step 2: Generate the predicted value of rudder deflection δ based on the nonlinear model and rudder control instructions est , and compare the predicted deflection value of the rudder surface with the actual deflection value δ collected by the sensor module mea The residual signal R is generated by n , and decompose the residual signal into several sub-band signals through wavelet decomposition;

[0009] The predicted value of the deflection of the rudder surface is obtained by constructing a nonlinear electro-hydraulic servo actuation system dynamics model based on the working principle and parameter information of the electro-hydraulic servo actuation system, and predicting it according to the input rudder control command. The nonlinear electro-hydraulic servo actuation system dynamics model includes: an autopilot, an electro-hydraulic servo valve unit, a symmetrical hydraulic cylinder unit, a load and rudder surface unit, and a sensor unit, wherein: the autopilot compares the expected position and the actual feedback position according to the rudder surface deflection command information of the aircraft, and generates a servo valve drive signal; the electro-hydraulic servo valve unit converts the current into mechanical force through the torque motor according to the servo valve drive signal, drives the valve core to move, and adjusts the flow rate of high-pressure oil to the hydraulic cylinder; a pressure difference is formed between the hydraulic cylinders on both sides of the symmetrical hydraulic cylinder unit due to the different flow rates, thereby pushing the piston rod connected to the rudder surface load, driving the rudder surface, and converting the hydraulic flow rate into a force acting on the rudder surface; the load and rudder surface unit are subjected to hydraulic thrust, and the rudder surface deflection is completed under the action of the internal hydraulic load and the external aerodynamic load; finally, the sensor unit detects the deflection amount of the rudder surface in real time and feeds it back to the autopilot.

[0010] The sub-band signal generates a residual signal through the predicted value of the rudder deflection and the actual deflection value, and after sampling at a frequency of 40 Hz, the discrete residual signal is captured through a time window with a length of 8 steps, and 8 sub-band signals with frequency distribution from low to high are obtained through wavelet decomposition.

[0011] Step 3: Calculate the wavelet energy of the sub-band that may contain the rudder oscillation fault signal, and judge whether the fault occurs by judging the size of the energy comparison threshold, wherein the judgment threshold is determined by Monte Carlo simulation;

[0012] The wavelet energy calculation selects all sub-bands within the frequency band of 0.5-10 Hz, calculates the sum of the wavelet energies within 8 time steps according to the wavelet amplitudes of the sub-bands, and determines whether the sum of the wavelet energies exceeds a threshold determined by Monte Carlo simulation as a criterion for the occurrence of a fault.

[0013] The Monte Carlo simulation is to perform simulation under normal working conditions without faults, and record the maximum value E of the wavelet energy in the 0.5-10 Hz frequency band in the residual signal after each simulation. max ,After accumulating 1000 simulation samples, the mean and standard deviation are calculated, the Z critical value is queried according to the confidence level, the confidence interval is calculated, and the upper limit of the confidence interval is used as the judgment threshold of the wavelet energy.

[0014] The simulation refers to: according to the constructed electro-hydraulic servo actuation system, the hydraulic pressure difference ΔP and the actuator damping coefficient K are calculated. d Uncertainty is introduced into the parameter setting to improve the robustness and applicability of the model.

[0015] The present invention relates to an aircraft rudder oscillation fault monitoring system based on multi-scale wavelet transform for realizing the above method, comprising: a signal preprocessing module, a feature extraction module and an oscillation fault diagnosis module, wherein: the signal preprocessing module receives the expected rudder deflection angle output by the autopilot and the actual rudder deflection angle measured by the sensor, generates a predicted value of the rudder deflection angle according to a pre-built nonlinear actuator model and the expected rudder deflection angle, and obtains a residual signal by subtracting the predicted value from the sensor measurement value; the feature extraction module performs multi-scale wavelet decomposition on the processed residual signal, and calculates the wavelet energy of the sub-frequency bands where the fault signal may be distributed; the fault diagnosis module compares the sub-frequency band wavelet energy with the judgment threshold corresponding to each frequency band, and if the threshold is exceeded, it is considered that there is a rudder oscillation problem in the servo actuator system. Technical Effects

[0016] The present invention performs multi-scale wavelet decomposition on the residual signal, and calculates the wavelet energy of the frequency band where the rudder oscillation signal may be distributed in the decomposition result, and diagnoses whether the fault occurs by comparing the wavelet energy and judging the threshold value. Compared with the prior art, the present invention significantly improves the fault coverage of the traditional amplitude oscillation counting method, and verifies the performance of the improved rudder oscillation fault monitor through Monte Carlo simulation. This solution not only improves the monitoring accuracy, but also takes into account the practicality and reliability of the system, providing a strong guarantee for the safety of the fly-by-wire flight control system. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is the principle diagram of the present invention;

[0018] Figure 2 is a schematic diagram of a hydraulic servo actuator of the present invention;

[0019] Figure 3 It is a structural block diagram of the rudder surface oscillation monitoring device of the present invention;

[0020] Figure 4 It is a logical schematic diagram of the control surface oscillation monitoring device of the present invention;

[0021] Figure 5 Schematic diagram of sample data for determining simulation thresholds for Monte Carlo simulation;

[0022] Figure 6 It is a schematic diagram of residual curve and monitoring situation example when solid control surface oscillation failure occurs;

[0023] In the figure: (a) from top to bottom are the measured and estimated values ​​of the control surface deflection in the case of solid fault, the residual signal generated by comparison, and the curve of the control surface oscillation monitor to determine whether to issue an early warning; (b) from top to bottom are the wavelet energy curves of sub-band 1 and sub-band 2 after wavelet decomposition;

[0024] Figure 7 This is a schematic diagram of the residual curve and monitoring situation when a liquid control surface oscillation failure occurs. DETAILED DESCRIPTION

[0025] The present embodiment relates to an aircraft control surface oscillation fault monitoring method based on multi-scale wavelet transform, which generates residuals by comparing and predicting signal sensor acquisition signals, performs data processing and fault diagnosis on the residuals, identifies control surface oscillation faults, improves the coverage of control surface oscillation faults, and ensures real-time performance.

[0026] like Figure 1 The figure shows possible fault sources and fault forms of the aircraft control surface oscillation fault. The fly-by-wire flight control system includes a flight control computer and a control surface servo actuator. In other embodiments, the fly-by-wire flight control system may include more or less flight control electronic equipment.

[0027] The flight control computer includes a monitoring channel and a control channel. The monitoring channel is used to detect various faults in the servo loop, and the detection of the control surface oscillation fault is also completed in the monitoring channel; the control channel can calculate the control surface control instructions based on the input signal according to the control law and output them to the servo actuator control electronics to control the aircraft control surface.

[0028] The rudder oscillation fault is mainly caused by the failure of electronic or mechanical components, such as the generation of false signals (such as sine waves). According to the different manifestations of the fault, it can be divided into two categories:

[0029] Liquid form fault: the oscillation signal is superimposed on the original control signal.

[0030] Solid-state failure: The oscillating signal completely replaces the original control signal.

[0031] These two types of oscillation signals are transmitted to the control surface through the servo control loop, causing the control surface to oscillate. The sources of oscillation of the control surface include: analog input / output interface (ADC), hydraulic piston rod position sensor, servo current amplifier, control surface sensor, and flight control computer. These components generate oscillation pseudo-command signals, which are transmitted to the control surface through the actuator servo control loop, causing the control surface to oscillate.

[0032] like Figure 2 As shown in the figure, the working principle of the hydraulic servo actuator is as follows: p(t) is the rod displacement of the servo actuator, and the first-order differential is used to obtain v(t), which is the displacement rate of the actuator rod. After receiving the input electrical signal, the servo valve changes the valve core displacement to control the flow in the cavity, so that the pressure at the hydraulic cavity A and B is P respectively. A and P BThe difference between the two is the hydraulic pressure difference ΔP, S is the surface area of ​​the actuator piston, and the force acting on the piston rod when multiplied by the pressure difference pushes the piston rod to produce displacement, driving the rudder surface to deflect. The transformation relationship from the actual rod displacement p(t) to the actual servo actuator rudder surface deflection δ can be obtained by interpolation table lookup, which is generally related to the actual installation angle.

[0033] like Figure 3 As shown, an aircraft rudder oscillation fault monitoring system based on multi-scale wavelet transform involved in this embodiment includes: a rudder servo control loop and a rudder oscillation monitor. The rudder servo control loop is an actuation system when the fly-by-wire flight control system of a civil aircraft performs trajectory tracking, and is shown here in a simplified flow chart; the function of the rudder oscillation monitor is to determine the occurrence of a fault based on the autopilot expected deflection signal and the rudder sensor acquisition signal when a rudder oscillation fault occurs in the actuator or sensor during the operation of the rudder servo control loop.

[0034] The rudder servo control loop includes: an autopilot, an electro-hydraulic servo valve unit, a symmetrical hydraulic cylinder unit, a load and rudder unit, and a sensor unit, wherein: the autopilot compares the expected position and the actual feedback position according to the rudder deflection instruction information of the aircraft, and generates a servo valve drive signal; the electro-hydraulic servo valve unit converts the current into mechanical force through the torque motor according to the servo valve drive signal, drives the valve core to move, and adjusts the flow of high-pressure oil to the hydraulic cylinder; the symmetrical hydraulic cylinder unit pushes the piston rod connected to the rudder load according to the pressure difference of the hydraulic cylinders on both sides, drives the rudder, and converts the hydraulic flow into a force acting on the rudder. The load and rudder unit are subjected to hydraulic thrust, and the rudder deflection is completed under the action of the internal hydraulic load and the external aerodynamic load; finally, the sensor unit detects the rudder deflection amount in real time and feeds it back to the autopilot.

[0035] The rudder surface oscillation monitor comprises: a signal preprocessing module, a feature extraction module and an oscillation fault diagnosis module, wherein: the signal preprocessing module receives the expected rudder surface deflection angle output by the autopilot and the actual rudder surface deflection angle measured by the sensor, generates a predicted value of the rudder surface deflection angle according to a pre-built nonlinear actuator model and the expected rudder surface deflection angle, and obtains a residual signal by subtracting the predicted value from the sensor measurement value; the feature extraction module performs multi-scale wavelet decomposition on the processed residual signal, and calculates the wavelet energy of the sub-frequency bands where the fault signal may be distributed; the fault diagnosis module compares the sub-frequency band wavelet energy with the judgment threshold corresponding to each frequency band, and if the threshold is exceeded, it is considered that there is a rudder surface oscillation problem in the servo actuator system.

[0036] The oscillation fault sources of the control surface considered in the present invention are the oscillation fault of the rod displacement sensor signal and the oscillation fault of the servo actuator input signal. In addition, the noise interference received by the sensor during operation, the disturbance impact received during normal operation of the aircraft and the model uncertainty problem of the servo actuator are also considered.

[0037] like Figure 4 As shown, when the actuator receives a control command for deflecting the rudder surface, it deflects the rudder surface and measures the actual rod displacement with a sensor. At the same time, the monitor calculates the estimated rod displacement based on the control command and the pre-built actuator dynamics model.

[0038] A residual signal is generated based on the estimated and measured values, and wavelet decomposition is performed on the residual signal. Wavelet decomposition decomposes the signal into low-frequency and high-frequency components at different levels, and sub-band signals with frequency components around 0.5-10Hz are selected. Wavelet energy calculation is performed on this part of the signal, and the wavelet energy in the frequency band is compared with a pre-set threshold to determine the fault situation and decide whether to trigger an early warning or continue operation.

[0039] like Figure 5 As shown in the figure, the sample data for determining the simulation threshold of Monte Carlo simulation is recorded, and the maximum wavelet energy of the two sub-bands is recorded for each simulation, showing the distribution of wavelet energy. The initial conditions of the simulation are that the system works normally, there is no rudder oscillation fault, the sensor is affected by white noise with a variance of 0.0005, and the model uncertainty is calculated by setting the model hydraulic pressure difference ΔP in the random range [16,30] and the actuator damping coefficient K d [6.8,10], the number of simulations is 1000.

[0040] Calculate the sample mean based on the statistical results And the standard deviation σ1=0.0038, σ2=0.0444, the upper limit of the 99.9999% confidence space is used as the monitoring threshold, and finally E is determined th1 =0.0411, E th2 =0.3894 is used as the wavelet energy threshold for the two sub-bands.

[0041] Through Figure 6 The specific experiment shown is that when a civil aircraft executes a trajectory tracking command, a solid fault or liquid fault of 1-10Hz is injected into the sensor module. When a solid oscillation fault occurs, the residual curve and monitoring situation example are shown. The rudder oscillation type is a solid fault, which is injected into the sensor output signal to simulate a sensor fault. The fault injection time is 40s, the amplitude is 1mm, and the fault frequency is 5Hz. The above method can detect the oscillation fault problem in the servo actuation loop within three sampling steps after the fault occurs.

[0042] like Figure 6 (a) and Figure 6As shown in (b), when the solid fault of the sensor occurs at 40s, the measured value and estimated value of the rudder deflection deviate. Since the solid fault is an oscillation signal replacing the original signal, the residual and wavelet energy can both reflect whether the oscillation fault occurs at this time. The energy of the two frequency bands exceeds the threshold, the rudder oscillation monitor triggers an early warning, and the fault signal is set from 0 to 1 at 40s. It can be seen that the rudder oscillation monitoring method that introduces wavelet energy calculation on the basis of the model-based method can effectively detect the oscillation fault in the system.

[0043] like Figure 7 As shown in the figure, when a liquid oscillation fault occurs, the residual curve and monitoring situation are examples. The rudder oscillation type is a liquid fault, which is injected into the sensor output signal to simulate the sensor fault. The fault injection time is 40s, the amplitude is 1mm, and the fault frequency is 5Hz.

[0044] like Figure 7 As shown in (a), from top to bottom, they are the measured value and estimated value of the control surface deflection in the case of liquid failure, the residual signal generated by comparison, and the curve of the control surface oscillation monitor to determine whether to issue an early warning; Figure 6 (b) shows, from top to bottom, the wavelet energy curves of sub-band 1 and sub-band 2 after wavelet decomposition.

[0045] like Figure 7 As shown in Figure 2, when the sensor liquid failure occurs at 40s, the measured value and estimated value of the rudder deflection amount deviate. Since the amplitude of the liquid failure is small at this time, it is difficult to distinguish it based on the residual signal alone. However, the wavelet energy amplitude of frequency band 2 obviously exceeds the judgment threshold E th2 =0.3894, the rudder oscillation monitor triggers an early warning, and the fault signal is set from 0 to 1 in 40s. It can be seen that the rudder oscillation monitoring method based on the model-based method by introducing wavelet energy calculation can effectively improve the fault coverage of fault detection.

[0046] As shown in Table 1, this is the minimum fault amplitude that can be detected by the rudder oscillation fault monitor, demonstrating the detection capability of the minimum fault amplitude.

[0047] Table 1 Fault location Fault type Frequency range Noise intensity Working conditions Minimum monitoring amplitude Actuator Liquid 1-10Hz 0.0005 Tracking 0.09mm sensor Liquid 1-10Hz 0.0005 Tracking 0.05mm

[0048] Since solid fault detection is independent of amplitude, only liquid fault conditions are considered here. Through simulation experiments, it is found that when the frequency of the oscillation signal is 0.5-10 Hz, and white noise interference with a variance of 0.0005 is considered, when a fault occurs when the aircraft is performing trajectory tracking, the oscillation fault monitoring device designed in the present invention can detect servo actuator faults with a minimum amplitude of 0.09 mm and sensor faults with a minimum amplitude of 0.05 mm, and has good fault monitoring coverage.

[0049] Compared with the prior art, the present invention introduces multi-scale wavelet decomposition in the feature extraction module, which greatly improves the robustness of the monitor against noise and interference. The introduction of wavelet energy as a fault diagnosis indicator in the fault diagnosis module can effectively detect early small oscillations and has a higher coverage of faults than traditional methods.

[0050] The above-mentioned specific implementation can be partially adjusted in different ways by those skilled in the art without departing from the principle and purpose of the present invention. The protection scope of the present invention shall be based on the claims and shall not be limited by the above-mentioned specific implementation. Each implementation scheme within its scope shall be subject to the constraints of the present invention.

Claims

1. A method for monitoring aircraft control surface oscillation faults based on multi-scale wavelet transform, characterized in that: include: Step 1: The autopilot uses the PID control law to solve the control command δ according to the input track command desired ; Step 2: Generate the predicted value of rudder deflection δ based on the nonlinear model and rudder control instructions est , and compare the predicted deflection value of the rudder surface with the actual deflection value δ collected by the sensor module mea The residual signal R is generated by n , and decompose the residual signal into several sub-band signals through wavelet decomposition; Step 3: Based on the wavelet energy calculation of the sub-band that may contain the rudder oscillation fault signal, the energy comparison threshold is determined to determine whether the fault occurs, wherein the judgment threshold is determined by Monte Carlo simulation.

2. The method for monitoring aircraft control surface oscillation fault according to claim 1, characterized in that: The predicted value of the deflection of the rudder surface is obtained by constructing a nonlinear electro-hydraulic servo actuation system dynamics model based on the working principle and parameter information of the electro-hydraulic servo actuation system, and predicting it according to the input rudder control command. The nonlinear electro-hydraulic servo actuation system dynamics model includes: an autopilot, an electro-hydraulic servo valve unit, a symmetrical hydraulic cylinder unit, a load and rudder surface unit, and a sensor unit, wherein: the autopilot compares the expected position and the actual feedback position according to the rudder surface deflection command information of the aircraft, and generates a servo valve drive signal; the electro-hydraulic servo valve unit converts the current into mechanical force through the torque motor according to the servo valve drive signal, drives the valve core to move, and adjusts the flow of high-pressure oil to the hydraulic cylinder; the symmetrical hydraulic cylinder unit pushes the piston rod connected to the rudder surface load according to the pressure difference between the hydraulic cylinders on both sides, drives the rudder surface, and converts the hydraulic flow into a force acting on the rudder surface. The load and rudder surface unit are subjected to hydraulic thrust, and the rudder surface deflection is completed under the action of the internal hydraulic load and the external aerodynamic load; finally, the sensor unit detects the deflection amount of the rudder surface in real time and feeds it back to the autopilot.

3. The method for monitoring aircraft control surface oscillation fault according to claim 1, characterized in that: The sub-band signal generates a residual signal through the predicted value of the rudder deflection and the actual deflection value, and after sampling at a frequency of 40 Hz, the discrete residual signal is captured through a time window with a length of 8 steps, and 8 sub-band signals with frequency distribution from low to high are obtained through wavelet decomposition.

4. The method for monitoring aircraft control surface oscillation fault according to claim 1, characterized in that: The wavelet energy calculation selects all sub-bands within the frequency band of 0.5-10 Hz, calculates the sum of the wavelet energies within 8 time steps according to the wavelet amplitudes of the sub-bands, and determines whether the sum of the wavelet energies exceeds a threshold determined by Monte Carlo simulation as a criterion for the occurrence of a fault.

5. The method for monitoring aircraft control surface oscillation fault according to claim 1, characterized in that: The Monte Carlo simulation is to perform simulation under normal working conditions without faults, and record the maximum value E of the wavelet energy in the 0.5-10 Hz frequency band in the residual signal after each simulation. max ,After accumulating 1000 simulation samples, the mean and standard deviation are calculated, the Z critical value is queried according to the confidence level, the confidence interval is calculated, and the upper bound of the confidence interval is used as the judgment threshold of the wavelet energy.

6. The method for monitoring aircraft control surface oscillation fault according to claim 1, characterized in that: The simulation refers to: according to the constructed electro-hydraulic servo actuation system, the hydraulic pressure difference ΔP and the actuator damping coefficient K are calculated. d Uncertainty is introduced into the parameter setting to improve the robustness and applicability of the model.

7. An aircraft control surface oscillation fault monitoring system based on multi-scale wavelet transform according to any one of the methods of claims 1 to 6, characterized in that: include: A signal preprocessing module, a feature extraction module and an oscillation fault diagnosis module, wherein: the signal preprocessing module receives the expected rudder surface deflection angle output by the autopilot and the actual rudder surface deflection angle measured by the sensor, generates a predicted value of the rudder surface deflection angle according to a pre-built nonlinear actuator model and the expected rudder surface deflection angle, and obtains a residual signal by subtracting the predicted value from the sensor measurement value; the feature The extraction module performs multi-scale wavelet decomposition on the processed residual signal and calculates the wavelet energy of the sub-frequency bands where the fault signal may be distributed; the fault diagnosis module compares the sub-frequency band wavelet energy with the judgment threshold corresponding to each frequency band. If the threshold is exceeded, it is considered that there is a rudder oscillation problem in the servo actuator system.

Images