Active control method of multi-harmonic and multi-channel global vibration of helicopter

Abstract

A multi-harmonic, multi-channel, global active vibration control method for helicopters involves constructing independent bandpass tracking filter banks to separate the harmonic components in the acceleration response error signal at each control point on the helicopter's fuselage. Each harmonic component is independently controlled by selecting a convergence factor. During the control process, the reference input signal is reconstructed while updating the weights of the Fx-LMS adaptive controller with global vibration minimization as the control objective. This method achieves significant attenuation of the acceleration response at all measurement points on the helicopter's fuselage under multi-harmonic interference from the helicopter's rotor blades, effectively improving the vibration level of the fuselage. The present invention separates and independently controls the fundamental and harmonic components of the vibration signal by constructing several independent adaptive filters, with global vibration minimization as the control objective, to update the controller weights.

Description

Technical Field

[0001] The present invention relates to a technology in the field of helicopter control, in particular to a method for active multi-harmonic multi-channel global vibration control of a helicopter. Background Art

[0002] Due to the unique characteristics of helicopter rotor dynamics, vibration loads with a frequency of kNΩ (k = 1, 2, 3, ..., where N is the number of rotor blades and Ω is the rotor speed) are transmitted through the rotor to the fuselage structure. Consequently, fuselage vibration manifests as a superposition of steady-state and harmonic vibrations of various orders of rotor loads, causing severe vibration problems for the fuselage. Existing active control technologies for helicopter structural response can effectively control vibration components at the main pass-through frequency and have been successfully applied to a range of aircraft models. However, due to limitations in controller computing power and actuator operating frequency bands, it is difficult to simultaneously suppress other higher-order harmonic components.

[0003] The existing technology only performs vibration control on single-frequency noise and is not applicable to vibration control of multi-frequency noise signals and mutual interference coupling between control channels in multi-input multi-output systems. When the reference input signal contains multiple frequency components, if only a single error signal containing multiple frequency components is used as the feedback signal for all control channels to update the weights of the control filter, the control effect will be reduced. In addition, in multi-input multi-output systems, there is channel coupling between control channels, which may cause the control system to diverge in severe cases.

[0004] In addition, existing technologies are also not applicable to multi-harmonic disturbance vibration control. When the reference input signal contains multiple frequency components, if only a single error signal containing multiple frequency components is used as the feedback signal of all control channels to update the weight of the control filter, the control effect will be reduced. Summary of the Invention

[0005] In order to solve the problem that the existing adaptive filtering Fx-LMS algorithm relies only on a single broadband secondary channel identification model to control multi-frequency vibrations, and can only select a small convergence factor, resulting in a slow system convergence speed and poor control effect, the present invention proposes a helicopter multi-harmonic multi-channel global vibration active control method. By constructing several independent adaptive filters, the fundamental frequency and harmonic components in the vibration signal are separated and independently controlled, with the global vibration minimization as the control goal, and the controller weight update is achieved.

[0006] The present invention is achieved through the following technical solutions:

[0007] The present invention relates to a multi-harmonic multi-channel global vibration active control method for a helicopter. An independent bandpass tracking filter group is constructed to separate the various harmonic components in the acceleration response error signal at each to-be-controlled point on the fuselage of the helicopter to be controlled. Each harmonic component is independently controlled by selecting a convergence factor. During the control process, a reference input signal is reconstructed while updating the weight of an Fx-LMS adaptive controller with global vibration minimization as the control target. This method achieves significant attenuation of the acceleration response of all measuring points on the fuselage under multi-harmonic interference of the helicopter rotor, thereby effectively improving the vibration level of the fuselage.

[0008] The bandpass tracking filter group specifically includes: a plurality of bandpass filters with adjustable center frequencies, which filter the helicopter fuselage fundamental frequency signal and high-order frequency harmonic signal components respectively, so that the error signal containing all frequency components is bandpass filtered to obtain a group of signal components with a single frequency component.

[0009] The independent control means that each signal component having a single frequency component after passing through the bandpass filter is independently controlled, and the signal components do not interfere with each other during the control process.

[0010] The global vibration minimization means that in a multi-input multi-output system, when multiple actuators operate in coordination, the root mean square value of the acceleration response of multiple sensor measurement points on the helicopter fuselage is minimized.

[0011] The present invention relates to a system for implementing the above-mentioned method, comprising: a tracking filter unit, a control channel identification model unit, a reconstructed reference input signal unit, and a multi-harmonic controller unit, wherein: the tracking filter unit performs filtering processing based on an error response signal measured by a sensor to obtain a group of signal components having a single frequency component, namely, a sub-error signal; the control channel identification model unit performs processing on a secondary channel transfer function between an actuator and a sensor measuring point based on the principle of adaptive filtering to obtain filter weights that can reflect the secondary channel transfer function model; the reconstructed reference input signal unit performs signal reconstruction based on the sub-error signal and an identification frequency response function model to obtain a reconstructed reference input signal related to the input signal; and the multi-harmonic controller unit updates controller weights based on the reconstructed reference input signal and the error response signal using a gradient descent method to obtain a controller voltage signal.

[0012] Technical Effects

[0013] The present invention effectively separates multiple frequency components in the error signal through a band-pass tracking filter, and each control channel performs identification within the corresponding sub-frequency band, which can more effectively improve the accuracy and precision of identification and effectively enhance the vibration control effect. At the same time, with the goal of minimizing global vibration, the reconstructed interference input signal is obtained from the respective identified frequency response functions and error signals, avoiding the problem in the prior art of using a single error signal containing multiple frequency components as the feedback signal for all control channels to update the weights of the control filter, modulating multiple uncorrelated frequency components in the error signal, effectively improving the vibration control effect, and accelerating the convergence speed of the control system. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 is a flowchart of the present invention;

[0015] Figure 2 is a flowchart of an embodiment;

[0016] Figure 3 is a schematic diagram of the reconstruction process of the multi-harmonic reference input signal in the present invention;

[0017] FIG. 4(a) is a result diagram of the frequency response identification process of the control channel of actuator 1 for control point 1 within the frequency range of 15 - 25 Hz at the main passing frequency according to the present invention;

[0018] FIG. 4(b) is a result diagram of the frequency response identification process of the control channel of actuator 1 for control point 1 within the frequency range of 30 - 50 Hz at the second harmonic frequency according to the present invention;

[0019] FIG. 5(a) is a time-domain diagram before and after control of measurement point 1 of the helicopter airframe measured experimentally under the excitation of vertical rotor load (two harmonics);

[0020] FIG. 5(b) is a frequency-domain diagram before and after control of measurement point 1 of the helicopter airframe measured experimentally under the excitation of vertical rotor load (two harmonics);

[0021] Figure 6 is a time-domain diagram before and after control of measurement points 1 to 4 of the helicopter airframe measured experimentally under variable working condition excitation (variable amplitude, variable frequency, variable phase). DETAILED DESCRIPTION OF THE EMBODIMENTS

[0022] As Figure 1 and Figure 2 shown, the present embodiment relates to a multi-harmonic multi-channel global vibration active control method for a helicopter, specifically including:

[0023] Step 1, constructing a band-pass filter: According to the characteristics of the rotor interference frequency, K-order band-pass tracking filters B k (b) are designed for the corresponding frequency band ranges for each harmonic frequency range respectively, specifically: B1(b), B2(b),..., Bk (Bi), where: the frequency band range of B1(Bi) is 15 - 25 Hz, the frequency band range of B2(Bi) is 30 - 50 Hz, and the frequency band range of B3(Bi) is 50 - 75 Hz.

[0024] Step 2, System Identification: Collect the control output voltages of each active actuator and the acceleration responses at each body point to be measured respectively. Use the normalized least mean square algorithm to identify the helicopter vibration active control system to be processed. According to the main passing frequency and the k-th order harmonic frequency, identify the corresponding secondary channel models respectively. k identification models are required between each active actuator and each body point to be measured. Convert all the identified control channels obtained, that is, the secondary channel frequency response function model, into the pulse sequence corresponding to the frequency response.

[0025] All the identified control channels mentioned above are Where: is the identified control channel model of the active actuator j for the body point i to be measured in the relevant frequency band k, such as: is the identified control channel model of the active actuator 1 for the body point 1 to be measured in the main passing frequency range, is the identified control channel model of the active actuator 2 for the body point 1 to be measured in the second harmonic frequency range.

[0026] Step 3, Error Signal Measurement: Collect the acceleration response error signals at each body point to be measured and input them into the band-pass tracking filter B k (Bi) obtained in Step 1 to obtain each sub-error signal Where: e i (Bi) is the acceleration response error signal at each body point to be measured.

[0027] Step 4, Reconstruct the Reference Input Signal: Obtain the reference input signal from the frequency response functions identified respectively in Step 2 and the sub-error signals measured in Step 3 Where: is the input of the control channel of the active actuator j in the relevant frequency band k, such as: is the input of the control channel of the active actuator 1 in the main passing frequency range.

[0028] Step 5, Update the Controller Weights: According to the sub-error response signals obtained in Step 3 and the reconstructed reference input signals obtained in Step 4, update the weights with the minimum global vibration as the control target according to the gradient descent method to obtain the sub-controller output voltage signal.

[0029] The updated weights mentioned above refer to: Where: i = 1, 2,..., P, is the convergence factor corresponding to the j-th controller in the relevant frequency band k, is the control factor corresponding to the j-th controller within the relevant frequency band k, B k (Bi) is the band-pass filter constructed in step 1, e i (Bi) is the acceleration response error signal at each control point to be measured in step 3, P is the number of control points to be measured is the filtered -x signal obtained by the secondary channel filter after the reference input signal reconstructed in step 4 passes through step 2, and the filter length is L

[0030] The convergence factor mentioned above satisfies where: λ is the eigenvalue of the autocorrelation matrix of the reference input signal. The larger the convergence factor, the faster the convergence speed of the control system, but it will cause the system to be unstable. Therefore, a smaller initial value is selected during selection, and it is gradually increased according to the actual convergence speed and stability

[0031] Step 6, obtain the total control signal according to the outputs of each sub - controller: Each sub - controller adjusts its own weight for control output, and the total control signal is the sum of the outputs of all sub - controllers, that is where: k = 1, 2,..., K is the input of the j - th control channel of the active actuator within the relevant frequency band k is the reconstructed reference input signal obtained in step 4 is the controller weight obtained in step 5

[0032] Step 7, take the total controller signal obtained in step 6 as the input signal of the active actuator at the next moment, drive the active actuator to generate active force through the control channel, and excite a corresponding response signal on the fuselage of the helicopter to be processed. Superimpose the output signal of the control channel with the acceleration response signal at the control point of the fuselage to obtain the error response signal, and continuously loop through steps 3 to step 6 until the root mean square of the acceleration responses of all measurement points on the fuselage reaches the minimum value under the coordinated operation of multiple actuators

[0033] After specific actual experiments, under the helicopter active control system test bench, set the convergence factor to be 0.00015 in the main rotor passing frequency range and 0.00025 in the second - order harmonic frequency band range of the rotor. Through the above method, the acceleration responses of all measurement points on the fuselage are significantly reduced, the vibration of the fuselage is effectively improved, and the attenuation of the acceleration responses at each measurement point reaches 74%

[0034] Figure 4(a) shows the frequency response identification process for the control channel of actuator 1 at controlled point 1, with a main pass frequency of 19 Hz and an identification frequency range of 15-25 Hz. Figure 4(b) shows the frequency response identification process for the control channel of actuator 1 at controlled point 1, with a second-order harmonic frequency of 38 Hz and an identification frequency range of 30-50 Hz. During the identification process, the expected signal and the actual output signal gradually approached each other, the identification error rapidly decreased, and the identification accuracy reached over 95%. It should be noted that the frequency response identification process and results for the remaining channels are similar to this result.

[0035] Figure 5(a) shows a time-domain plot of helicopter airframe measurement point 1 before and after control, as measured in the experiment under vertical second-order harmonic rotor load excitation. Compared with the traditional FxLMS method, this method achieves faster control convergence and better control effectiveness. After control stabilizes, the control decay rate reaches 74% using this method, while the control decay rate using the traditional FxLMS method is only 57%.

[0036] Figure 5(b) shows the frequency domain plot of helicopter airframe measurement point 1 before and after control, as measured in the experiment under vertical second-order harmonic rotor load excitation. Compared with the traditional FxLMS method, this method has better control capabilities for multiharmonic interference, with significant attenuation of all interference frequency line spectrum components.

[0037] like Figure 6 Figure 2 shows the time domain plots of the helicopter body at measurement points 1 to 4 before and after control, measured during the experiment under variable excitation conditions (variable amplitude, frequency, and phase). The rotor disturbance is a vertical second-order harmonic excitation load. The controller is turned on at 20 seconds. After the vibration attenuation stabilizes, the rotor disturbance excitation load amplitude increases by 40% at 50 seconds. At 80 seconds, the rotor disturbance excitation load phase changes by 90°. At 110 seconds, the rotor disturbance excitation load frequency increases by 10%. At 140 seconds, the rotor disturbance excitation load phase changes by -45°. Finally, at 170 seconds, the rotor disturbance excitation load amplitude decreases by 40%. At 210 seconds, the controller is turned off.

[0038] Compared to existing FxLMS technology, where the acceleration response of each controlled point fluctuates significantly before the vibration response gradually decays, this method rapidly converges to a stable acceleration response, with minimal fluctuations when the disturbance excitation changes. Acceleration response fluctuations are most pronounced when the frequency changes, but this method quickly suppresses vibration fluctuations and converges to stability. Results from variable disturbance excitation conditions indicate that this method offers improved control of multiharmonic disturbances and exhibits greater stability and robustness.

[0039] The above-mentioned specific implementation can be partially adjusted in different ways by those skilled in the art without departing from the principles and purpose of the present invention. The scope of protection of the present invention shall be based on the claims and shall not be limited by the above-mentioned specific implementation. All implementation schemes within its scope shall be subject to the constraints of the present invention.

Claims

1. A helicopter multi-harmonic multi-channel global vibration active control method, characterized in that: An independent bandpass tracking filter bank is constructed to separate the harmonic components in the acceleration response error signal at each control point on the helicopter's fuselage. Each harmonic component is independently controlled by selecting a convergence factor. During the control process, the reference input signal is reconstructed while the weights of the Fx-LMS adaptive controller are updated with the global vibration minimization as the control objective. This significantly attenuates the acceleration response at all measurement points on the helicopter's fuselage under multi-harmonic interference from the helicopter rotor, effectively improving the vibration level of the fuselage. The global vibration minimization mentioned above means that in a multi-input multi-output system, when multiple actuators are operating in coordination, the root mean square value of the acceleration response of multiple sensor measurement points on the helicopter fuselage is minimized, specifically including: Step 1: Construct a bandpass filter: According to the rotor interference frequency characteristics, design a K-order bandpass tracking filter for each frequency band for each harmonic frequency range. , specifically: ,in: The frequency band range is 15-25Hz, The frequency band range is 30-50Hz, The frequency band ranges from 50-75Hz; Step 2, system identification: The control output voltage of each active actuator and the acceleration response at each control point on the fuselage are collected separately. The normalized least mean square algorithm is used to identify the active vibration control system of the helicopter to be processed. The corresponding secondary channel model is identified based on the main pass frequency and the k-order higher-order harmonic frequency. K identification models are required between each active actuator and each test point on the fuselage. All identified control channels, i.e., the secondary channel frequency response function models, are converted into pulse trains corresponding to the frequency response. Step 3, error signal measurement: collect the acceleration response error signal at each control point of the fuselage and input it into the bandpass tracking filter obtained in step 1 , and obtain each sub-error signal ,in: is the acceleration response error signal at each point to be controlled; Step 4: Reconstruct the reference input signal: Obtain the reference input signal from the frequency response functions identified in step 2 and the sub-error signals measured in step 3. ,in: The control channel input of active actuator j within the relevant frequency band k, such as: It is the control channel input of active actuator 1 within the main pass frequency range; Step 5: Update the controller weights: Based on the sub-error response signal obtained in step 3 and the reconstructed reference input signal obtained in step 4, update the weights using the gradient descent method with the global vibration minimum as the control objective to obtain the sub-controller output voltage signal. Step 6: Get the total control signal based on the output of each sub-controller: Each sub-controller controls the output by adjusting its own weight. The total control signal is the sum of the outputs of all sub-controllers, that is, ,in: , is the control channel input of active actuator j within the relevant frequency band k, is the reconstructed reference input signal obtained in step 4, is the controller weight obtained in step 5; Step 7: Use the total controller signal obtained in step 6 as the input signal of the active actuator at the next moment, drive the active actuator to generate active force through the control channel, and stimulate the corresponding response signal on the fuselage of the helicopter to be processed, and use the output signal of the control channel as the input signal of the active actuator. The error response signal is superimposed on the acceleration response signal at the fuselage control point to obtain the error response signal. Steps 3 to 6 are continuously executed in a loop until the root mean square of the acceleration response of all measuring points on the fuselage reaches the minimum value under the coordinated operation of multiple actuators.

2. The helicopter multi-harmonic multi-channel global vibration active control method according to claim 1 is characterized in that: The bandpass tracking filter bank specifically includes: a plurality of bandpass filters with adjustable center frequencies, which filter the helicopter fuselage fundamental frequency signal and the high-order frequency multiplication signal components respectively, so that the error signal containing all frequency components is bandpass filtered to obtain a group of signal components with a single frequency component; The independent control means that each signal component having a single frequency component after passing through the bandpass filter is independently controlled, and the signal components do not interfere with each other during the control process.

3. The helicopter multi-harmonic multi-channel global vibration active control method according to claim 1 is characterized in that: All identification control channels are ,in: is the control channel model for the identification of the fuselage test point i by the active actuator j within the relevant frequency band k, such as: is the control channel model for the identification of the fuselage test point 1 by the active actuator 1 within the main pass frequency range, It is a control channel model for the active actuator 2 to identify the fuselage test point 1 within the second-order harmonic frequency range.

4. The helicopter multi-harmonic multi-channel global vibration active control method according to claim 1, characterized in that: The update weights are: ,in: is the convergence factor corresponding to the j-th controller within the relevant frequency band k, The control factor corresponding to the j-th controller within the relevant frequency band k is, The bandpass filter constructed in step 1, is the acceleration response error signal at each control point measured in step 3, P is the number of control points, The filtered-x signal after the reference input signal is reconstructed in step 4 and passed through the secondary channel filter obtained in step 2, with a filter length of L. .

5. The helicopter multi-harmonic multi-channel global vibration active control method according to claim 1, characterized in that: The convergence factor satisfies ,in: Referring to the eigenvalue of the autocorrelation matrix of the reference input signal, the larger the convergence factor, the faster the control system converges, but it may cause system instability. Therefore, a smaller initial value is selected when selecting it, and it is gradually increased according to the actual convergence speed and stability.

6. A system for implementing the helicopter multi-harmonic multi-channel global vibration active control method according to any one of claims 1 to 5, characterized in that: include: A tracking filter unit, a control channel identification model unit, a reconstructed reference input signal unit and a multi-harmonic controller unit, wherein: the tracking filter unit performs filtering processing on the error response signal measured by the sensor to obtain a group of signal components with a single frequency component, namely, a sub-error signal; the control channel identification model unit performs secondary channel transfer function processing between the actuator and the sensor measurement point according to the adaptive filtering principle to obtain filter weights that can reflect the secondary channel transfer function model; the reconstructed reference input signal unit performs signal reconstruction based on the sub-error signal and the identification frequency response function model to obtain a reconstructed reference input signal related to the input signal; the multi-harmonic controller unit updates the controller weights according to the reconstructed reference input signal and the error response signal according to the gradient descent method to obtain a controller voltage signal.

Images