Millisecond-level vibration active control method based on variable stiffness

Abstract

The invention discloses a millisecond-level vibration active control method based on variable stiffness, and relates to the field of steering engine vibration control, and the method comprises the steps: defining a plurality of discrete stiffness modes, and starting a periodic timer; in each control period, collecting or generating a time domain vibration signal comprising a plurality of sampling points; applying a window function to the time domain vibration signal and then performing fast Fourier transform to obtain frequency spectrum data; performing peak detection and frequency interpolation calculation on the frequency spectrum data, and identifying a dominant excitation frequency with the highest energy ratio; comparing the dominant excitation frequency with a preset mode switching threshold value, and selecting a target rigidity mode capable of maximizing frequency separation; according to the selected target rigidity mode, querying and obtaining a corresponding target control voltage; and outputting the target control voltage to the piezoelectric actuator by using the digital signal processor through the digital-to-analog converter, so that the system rigidity is switched to a target rigidity mode. The problem that high-frequency and time-varying vibration is difficult to restrain in the prior art is solved.

Description

Technical Field

[0001] This invention relates to the field of servo motor vibration control, and in particular to a millisecond-level active vibration control method based on variable stiffness. Background Technology

[0002] As a key actuator in advanced weaponry, the dynamic performance, response speed, and control accuracy of servo motors directly determine the tactical and technical specifications of the equipment. However, in actual operation, factors such as the nonlinear transmission within the servo motor, its structural flexibility, and severe external aerodynamic loads can induce complex vibrations that are high-frequency, broadband, and time-varying. These vibrations are a major cause of decreased servo motor output accuracy, deterioration of dynamic characteristics, and even structural fatigue failure.

[0003] Currently, vibration suppression research both domestically and internationally is mainly divided into passive control and active control. Passive control technologies (such as optimized structures and high-damping materials) are mature, but they suffer from narrow tuning bandwidth and poor adaptability when dealing with broadband, time-varying vibrations. Traditional active control technologies (such as PID algorithms) acquire vibration information through sensors and apply reaction forces. However, traditional control algorithms inevitably have time delays in the closed-loop "sensing-computation-execution" process. When facing transient high-frequency vibrations of up to hundreds of hertz, these millisecond or even microsecond-level cumulative delays are sufficient to cause phase lag in the control force output, resulting not only in poor suppression effects but also potentially inducing resonance and causing vibration deterioration.

[0004] In summary, existing vibration suppression technologies generally suffer from common bottlenecks such as slow response speed and large control delay, making it difficult to meet the stringent requirements of next-generation servos for ultra-real-time vibration control. Summary of the Invention

[0005] To address the aforementioned shortcomings in existing technologies, this invention provides a millisecond-level active vibration control method based on variable stiffness, which solves the problem that traditional technologies struggle to suppress high-frequency, time-varying vibrations.

[0006] To achieve the above-mentioned objectives, the technical solution adopted by this invention is: a millisecond-level active vibration control method based on variable stiffness, comprising the following steps:

[0007] S1: Initialize the control system, which includes a digital signal processor, a digital-to-analog converter, and a piezoelectric actuator;

[0008] S2: Define multiple discrete stiffness modes, start a periodic timer, and proceed to step S3;

[0009] S3: In each control cycle, acquire or generate a time-domain vibration signal containing multiple sampling points;

[0010] S4: Apply a window function to the time-domain vibration signal and then perform a fast Fourier transform to transform it from the time domain to the frequency domain to obtain the spectrum data;

[0011] S5: Perform peak detection and frequency interpolation calculation on the spectrum data to identify the dominant excitation frequency with the highest energy proportion;

[0012] S6: Using a multimodal control decision algorithm, the dominant excitation frequency is compared with a preset mode switching threshold, and the target stiffness mode that maximizes frequency separation is selected.

[0013] S7: Based on the selected target stiffness mode, query and obtain its corresponding target control voltage;

[0014] S8: The digital signal processor outputs the target control voltage to the piezoelectric actuator through a digital-to-analog converter, so that the system stiffness switches to the target stiffness mode;

[0015] S9: Determine whether the current control cycle has reached the preset cycle threshold. If yes, complete the millisecond-level active vibration control based on variable stiffness; otherwise, return to step S3.

[0016] Furthermore, the window function in S4 is the Hanning window, and its calculation formula is as follows:

[0017]

[0018] in, For the Hanning window function, Here, N is the index of the sampling point, and N is the total number of points in the Fast Fourier Transform.

[0019] Furthermore, the Fast Fourier Transform in S4 is implemented based on a real-number FFT library optimized by the TI FPU (Floating-Point Unit), and uses a rotation factor table pre-calculated during the initialization phase.

[0020] Furthermore, the frequency interpolation calculation in S5 adopts the three-point parabolic interpolation algorithm, and the calculation formula is as follows:

[0021]

[0022] in, The detected peak amplitude The frequency index at which it is located. and Indexes and The amplitude at that point, For frequency correction factor, As the dominant excitation frequency, Where N is the sampling frequency, and N is the total number of points in the Fast Fourier Transform.

[0023] Furthermore, the control logic of the multimodal control decision algorithm in S6 is as follows:

[0024] When the dominant excitation frequency <Modal switching threshold If so, then the target stiffness mode is selected as the maximum stiffness mode;

[0025] When the dominant excitation frequency ≥ Modal switching threshold If so, then the target stiffness mode is selected as the minimum stiffness mode;

[0026] The mode switching threshold lie in and Between, among them, The natural frequency corresponding to the minimum stiffness mode. This represents the natural frequency corresponding to the maximum stiffness mode.

[0027] Further, the target control voltage output by the digital-to-analog converter in S6 includes:

[0028] Floating-point target control voltage Linear mapping and conversion to 8-bit unsigned integer code Its formula is:

[0029]

[0030] in, This is the reference voltage for the digital-to-analog converter.

[0031] The beneficial effects of this invention are:

[0032] (1) The method proposed in this invention has a fast response speed. The 4096-point FFT analysis and processing time based on DSP FPU optimization is less than 1ms, and the control decision time is less than 100µs, which is much less than the control cycle of 100ms. It realizes "ultra-real-time" vibration identification and control and overcomes the time delay bottleneck of traditional technology.

[0033] (2) The method proposed in this invention has high frequency identification accuracy. By combining FFT, Hann window function and three-point parabolic interpolation algorithm, the relative error of frequency identification is no more than 10% (in practice, it can reach within 0.1Hz), providing accurate and reliable input for subsequent control decisions;

[0034] (3) The method proposed in this invention has a significant vibration suppression effect. It adopts a dual-mode variable stiffness control strategy based on frequency separation, which can actively make the system's natural frequency avoid the external excitation frequency, effectively avoid resonance, and the vibration amplitude attenuation rate can reach more than 70%.

[0035] (4) This invention provides a clear software modular implementation scheme based on DSP and C language. The algorithm logic is clear, the computational overhead is small, and it is easy to implement and promote in embedded systems. Attached Figure Description

[0036] Figure 1 This is a flowchart of a millisecond-level active vibration control method based on variable stiffness according to the present invention.

[0037] Figure 2 This is a schematic diagram of the system hardware connection architecture in a specific embodiment of the present invention.

[0038] Figure 3 This is a diagram showing the dual-modal frequency-modal mapping relationship in a specific embodiment of the present invention.

[0039] Figure 4 This is a comparison diagram of the vibration control effect before and after in a specific embodiment of the present invention. Detailed Implementation

[0040] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0041] like Figure 1 As shown, a millisecond-level active vibration control method based on variable stiffness includes the following steps:

[0042] S1: Initialize the control system, which includes a digital signal processor (DSP), a digital-to-analog converter (DAC), and a piezoelectric actuator;

[0043] like Figure 2 As shown, the hardware core is a DSP28335 processor, which connects to a TLV5620 DAC chip via an SPI-A interface. The DAC output voltage drives the piezoelectric actuator. On the software side, the Config_Init() function is called to initialize system parameters.

[0044] S2: Define multiple discrete stiffness modes, start a periodic timer, and proceed to step S3;

[0045] Define two stiffness modes:

[0046] Minimum stiffness mode (MODE-MIN): corresponding to the natural frequency Actuator voltage .

[0047] Maximum stiffness mode (MODE-MAX): corresponding to the natural frequency Actuator voltage .

[0048] Configure real-time frequency identification parameters: Call the FFT_Init() function to set the number of FFT points. Signal sampling frequency Pre-compute and store the rotation factor table required for FFT to improve the efficiency of subsequent calculations.

[0049] Start the periodic timer: Call the TIM1_Init() function to configure CPU timer 1 to generate a periodic interrupt every 100ms (i.e., 10Hz). The interrupt service routine sets the global flag g_timer_flag=1 to trigger the control flow in the main loop.

[0050] S3: In each control cycle, acquire or generate a time-domain vibration signal containing multiple sampling points;

[0051] In the main loop, after detecting that g_timer_flag=1 (i.e., the 100ms period has arrived), the flag is first cleared. In this embodiment, to verify the algorithm, the SIGNAL_GenerateHarmonic() function is called to generate a simulated test signal with 4096 points. This signal is composed of three harmonic components (e.g., 60Hz, 90Hz, 150Hz) and noise superimposed to simulate complex external excitation.

[0052] S4: Apply a window function to the time-domain vibration signal and then perform a fast Fourier transform to transform it from the time domain to the frequency domain to obtain the spectrum data;

[0053] To suppress spectral leakage in the FFT, the Hann window function is applied to the 4096-point signal obtained in S3.

[0054] The window function is the Hanning window, and the calculation formula is:

[0055]

[0056] in, For the Hanning window function, Here, N is the index of the sampling point, and N is the total number of points in the Fast Fourier Transform.

[0057] Perform an N-point Fast Fourier Transform (FFT): Call the TI FPU-optimized real-number FFT library function RFFT_f32(&g_fft_ctrl.rfft_handle) to perform a 4096-point FFT transform on the windowed signal. Then, call RFFT_f32_mag(&g_fft_ctrl.rfft_handle) to calculate the one-sided magnitude spectrum, obtaining 2049 points of frequency domain magnitude data.

[0058] The Fast Fourier Transform is implemented using a real-number FFT library optimized by the TI FPU (Floating-Point Unit) and uses a twiddle factor table pre-computed during the initialization phase.

[0059] S5: Perform peak detection and frequency interpolation calculation on the spectrum data to identify the dominant excitation frequency with the highest energy proportion;

[0060] By traversing the amplitude spectrum obtained from S4, a three-point local maximum detection algorithm is used to find the peak. For the detected main peak with the strongest energy, a three-point parabolic interpolation algorithm is applied to refine the frequency. Even if the true frequency lies between two frequency points, this method can accurately identify the dominant excitation frequency through interpolation. .

[0061] Frequency interpolation is calculated using the three-point parabolic interpolation algorithm, and the formula is as follows:

[0062]

[0063] in, The detected peak amplitude The frequency index at which it is located. and Indexes and The amplitude at that point, For frequency correction factor, As the dominant excitation frequency, Where N is the sampling frequency, and N is the total number of points in the Fast Fourier Transform.

[0064] S6: Using a multimodal control decision algorithm, the dominant excitation frequency is compared with a preset mode switching threshold, and the target stiffness mode that maximizes frequency separation is selected.

[0065] The control logic of the multimodal control decision algorithm is as follows:

[0066] When the dominant excitation frequency <Modal switching threshold If so, then the target stiffness mode is selected as the maximum stiffness mode;

[0067] When the dominant excitation frequency ≥ Modal switching threshold If so, then the target stiffness mode is selected as the minimum stiffness mode;

[0068] The mode switching threshold lie in and Between, among them, The natural frequency corresponding to the minimum stiffness mode. This represents the natural frequency corresponding to the maximum stiffness mode.

[0069] By calling the Controller_Update() function, the dominant stimulus frequency identified by S5 is updated. With preset mode switching threshold Comparison:

[0070] like (If the excitation frequency is close to 52.4 Hz), then select the target stiffness mode. The maximum stiffness mode (MODE-MAX) is used to switch the system's natural frequency to 68.5Hz, away from the excitation.

[0071] like (If the excitation is close to 68.5Hz), then select the target stiffness mode. The minimum stiffness mode (MODE-MIN) is used to switch the system's natural frequency to 52.4Hz, away from the excitation.

[0072] This strategy aims to maximize the separation distance between the system's natural frequency and the excitation frequency.

[0073] S7: Based on the selected target stiffness mode, query and obtain its corresponding target control voltage;

[0074] Based on the target stiffness mode selected in S6 Query the corresponding target control voltage from the configuration module. .like If it is MODE-MIN, then If it is MODE-MAX, then .

[0075] S8: The digital signal processor outputs the target control voltage to the piezoelectric actuator through a digital-to-analog converter, so that the system stiffness switches to the target stiffness mode;

[0076] This is done by calling the DAC_SetVoltage(V_target) function. This function first sets the floating-point voltage... (0.0V or 1.0V) converted to 8-bit digital code (0 or 134), then the DSP sends the digital code $D$ to the TLV5620DAC chip via the SPI-A interface. The DAC outputs the corresponding analog voltage to the piezoelectric actuator, causing the system stiffness to switch to the target mode within 100ms. Subsequently, the system waits for the next timer interrupt and repeats steps S3 to S9 in a loop.

[0077] The digital-to-analog converter outputs a target control voltage, including:

[0078] Floating-point target control voltage Linear mapping and conversion to 8-bit unsigned integer code Its formula is:

[0079]

[0080] in, This is the reference voltage for the digital-to-analog converter.

[0081] S9: Determine whether the current control cycle has reached the preset cycle threshold. If yes, complete the millisecond-level active vibration control based on variable stiffness; otherwise, return to step S3.

[0082] In one embodiment of the present invention, given as follows Figure 2 The control system shown simulates external vibration with a dominant excitation frequency of 60Hz through a signal generation module. The system accurately identifies the dominant excitation frequency through FFT analysis. .

[0083] like Figure 3 As shown, due to The system determines that the excitation frequency is in the orange region, which is closer to the natural frequency of the maximum stiffness mode (68.5Hz). Therefore, the controller selects the target mode. This is the minimum stiffness mode (MODE-MIN). System output. By switching the system's natural frequency to 52.4Hz, a distance of 7.6Hz is created between it and the 60Hz excitation frequency, thus avoiding resonance.

[0084] like Figure 4 As shown, the effect before and after control is compared. In the "uncontrolled" state (i.e., the system remains at its natural frequency of 68.5Hz), the 60Hz excitation is close to its resonance point, resulting in a vibration amplitude (RMS) as high as 3.75mm. In the "controlled" state after activating the present invention, the system actively switches to 52.4Hz, and the vibration amplitude (RMS) rapidly decreases to 0.88mm, with an amplitude suppression rate of 76.5%, effectively avoiding resonance damage.

[0085] In summary, this numerical example fully demonstrates the effectiveness of the method proposed in this invention.

[0086] Those skilled in the art will recognize that the embodiments described herein are intended to help the reader understand the principles of the invention, and should be understood that the scope of protection of the invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations based on the technical teachings disclosed in this invention without departing from the spirit of the invention, and these modifications and combinations are still within the scope of protection of the invention.

Claims

1. A millisecond-level active vibration control method based on variable stiffness, characterized in that, Includes the following steps: S1: Initialize the control system, which includes a digital signal processor, a digital-to-analog converter, and a piezoelectric actuator; S2: Define multiple discrete stiffness modes, start a periodic timer, and proceed to step S3; S3: In each control cycle, acquire or generate a time-domain vibration signal containing multiple sampling points; S4: Apply a window function to the time-domain vibration signal and then perform a fast Fourier transform to transform it from the time domain to the frequency domain to obtain the spectrum data; S5: Perform peak detection and frequency interpolation calculation on the spectrum data to identify the dominant excitation frequency with the highest energy proportion; S6: Using a multimodal control decision algorithm, the dominant excitation frequency is compared with a preset mode switching threshold, and the target stiffness mode that maximizes frequency separation is selected. S7: Based on the selected target stiffness mode, query and obtain its corresponding target control voltage; S8: The digital signal processor outputs the target control voltage to the piezoelectric actuator through a digital-to-analog converter, so that the system stiffness switches to the target stiffness mode; S9: Determine whether the current control cycle has reached the preset cycle threshold. If yes, complete the millisecond-level active vibration control based on variable stiffness; otherwise, return to step S3.

2. The millisecond-level active vibration control method based on variable stiffness according to claim 1, characterized in that, The window function in S4 is the Hanning window, and its calculation formula is as follows: ； in, For the Hanning window function, Here, N is the index of the sampling point, and N is the total number of points in the Fast Fourier Transform.

3. The millisecond-level active vibration control method based on variable stiffness according to claim 1, characterized in that, The Fast Fourier Transform in S4 is implemented based on a real-number FFT library optimized by the TI FPU (Floating-Point Unit), and uses a twiddle factor table pre-calculated during the initialization phase.

4. The millisecond-level active vibration control method based on variable stiffness according to claim 1, characterized in that, The frequency interpolation calculation in S5 uses a three-point parabolic interpolation algorithm, and the calculation formula is as follows: ； in, The detected peak amplitude The frequency index at which it is located. and Indexes and The amplitude at that point, For frequency correction factor, As the dominant excitation frequency, Where N is the sampling frequency, and N is the total number of points in the Fast Fourier Transform.

5. The millisecond-level active vibration control method based on variable stiffness according to claim 1, characterized in that, The control logic of the multimodal control decision algorithm in S6 is as follows: When the dominant excitation frequency <Modal switching threshold If so, then the target stiffness mode is selected as the maximum stiffness mode; When the dominant excitation frequency ≥ Modal switching threshold If so, then the target stiffness mode is selected as the minimum stiffness mode; The mode switching threshold lie in and Between, among them, The natural frequency corresponding to the minimum stiffness mode. This represents the natural frequency corresponding to the maximum stiffness mode.

6. The millisecond-level active vibration control method based on variable stiffness according to claim 1, characterized in that, The target control voltage output by the digital-to-analog converter in S6 includes: Floating-point target control voltage Linear mapping and conversion to 8-bit unsigned integer code Its formula is: ； in, This is the reference voltage for the digital-to-analog converter.

Images