Self-adaptive pipeline vibration suppression system and method based on multi-mode coupling analysis

Abstract

The invention provides a self-adaptive pipeline vibration suppression system and method based on multi-mode coupling analysis. The system comprises a sensing module, a vibration analysis module and a control execution module, multiple sensing signals in the vibration process are obtained through measurement of a multi-sensor array, the vibration analysis module comprises a wavelet decomposition feature extraction unit, a prediction network unit and a judgment unit, and the wavelet decomposition feature extraction unit extracts feature data of the multiple sensing signals and sends the feature data to the control execution module; the prediction network unit obtains the probability distribution of various vibration modes according to the characteristic data, the judgment unit determines the dominant vibration frequency according to the probability distribution of various vibration modes, and the fuzzy adaptive unit of the control execution module dynamically distributes the output value of an actuator according to the dominant vibration frequency. And the hybrid execution unit controls the piezoelectric actuator and the magneto-rheological damper to perform hybrid suppression on vibration according to the dynamically distributed output value of the actuator. The system can accurately identify the vibration source, improve the control responsivity, and improve the overall high temperature resistance of the system.

Description

Technical Field

[0001] The present application belongs to the technical field of safety control of high-temperature and high-pressure pipeline systems in thermal power plants, and specifically relates to an adaptive pipeline vibration suppression system and method based on multimodal coupling analysis. Background Art

[0002] In thermal power plants, piping systems are essential for transporting media such as steam, water, and fuel. However, due to a variety of factors, including the flow of media within the pipelines, equipment operation, and the external environment, pipeline vibration is common. Excessive pipeline vibration can lead to a range of problems, such as loose connections and seal failure, which can cause media leaks. This not only wastes energy but also poses a threat to equipment and personnel safety. Furthermore, long-term vibration can accelerate fatigue damage to the pipelines and their associated components, shortening their service life and increasing equipment maintenance costs and downtime.

[0003] Currently, the main methods for controlling pipeline vibration in thermal power plants include strengthening the pipeline support structure and optimizing the pipeline layout. However, these methods are often not ideal for vibration problems under some complex working conditions.

[0004] The existing technology has the following defects:

[0005] 1. Insufficient vibration source tracing accuracy: Traditional FFT spectrum analysis cannot separate multi-source vibration coupling components (such as fluid excitation and structural resonance), resulting in poorly targeted control measures;

[0006] 2. Control response lag: The passive damping device has a response delay of more than 200ms under variable operating conditions and cannot suppress transient impact vibrations;

[0007] 3. Poor reliability in high temperature environments: The performance of existing piezoelectric actuators degrades by more than 50% in environments above 450°C. Summary of the Invention

[0008] The technical problem solved by this application is: how to provide a pipeline vibration suppression system that can accurately analyze the main vibration source, promptly control the response and adapt to high temperature environments.

[0009] The present application provides an adaptive pipeline vibration suppression system based on multimodal coupling analysis, the system comprising:

[0010] A sensing module, comprising a multi-sensor array, wherein the multi-sensor array is installed in the pipeline and measures a plurality of sensor signals during the vibration process;

[0011] a vibration analysis module comprising a wavelet decomposition feature extraction unit, a prediction network unit, and a judgment unit; the wavelet decomposition feature extraction unit is used to extract feature data of multiple sensor signals; the prediction network unit is used to obtain a probability distribution of various vibration modes based on the feature data; and the judgment unit determines a dominant vibration frequency based on the probability distribution of various vibration modes;

[0012] A control execution module includes a fuzzy adaptive unit and a hybrid execution unit. The fuzzy adaptive unit dynamically allocates the actuator output value according to the dominant vibration frequency. The hybrid execution unit controls the piezoelectric actuator and the magnetorheological damper to perform hybrid vibration suppression according to the dynamically allocated actuator output value.

[0013] Optionally, the multi-sensor array includes a three-axis MEMS accelerometer, a fiber Bragg grating strain sensor and a non-invasive ultrasonic flowmeter, the three-axis MEMS accelerometer outputs a voltage signal, the fiber Bragg grating strain sensor outputs a wavelength shift signal, and the non-invasive ultrasonic flowmeter outputs a pulse frequency signal.

[0014] Optionally, the perception module also includes an adaptive anti-aliasing filter, a notch filter and a temperature compensation unit. The adaptive anti-aliasing filter is used to filter multiple sensing signals, the notch filter is used to filter out power frequency interference from the voltage signal, and the temperature compensation unit is used to perform temperature compensation on the wavelength offset signal.

[0015] Optionally, the perception module further includes an edge computing unit, which performs data caching, format standardization, and preliminary anomaly detection on the pulse frequency signal after filtering, the voltage signal after filtering out power frequency interference, and the wavelength offset signal after temperature compensation.

[0016] Optionally, the method for extracting feature data of multiple sensor signals by the wavelet decomposition feature extraction unit includes:

[0017] The db4 wavelet basis is used to decompose the multi-layer wavelet packet of each sensor signal to obtain several sub-bands;

[0018] The energy entropy feature vector of each sub-band is extracted as feature data.

[0019] Optionally, the prediction network unit is a deep residual network.

[0020] Optionally, the method for the judgment unit to determine the dominant vibration frequency according to the probability distribution of various vibration modes includes:

[0021] Based on the gradient weighted class activation mapping technology, the contribution of each sub-band in the feature map to the classification results of various vibration modes is reversely analyzed to generate a contribution ranking list;

[0022] The sub-band corresponding to the energy entropy eigenvector with the largest weight is taken as the dominant vibration frequency.

[0023] Optionally, the method for the fuzzy adaptive unit to dynamically allocate the actuator output value according to the dominant vibration frequency includes:

[0024] Define the sliding surface: According to the current dominant vibration frequency f0, define the sliding surface Where λ = 2πf0, e(t) is the vibration amplitude error, is the error derivative;

[0025] Construct the control output equation of the actuator output value:

[0026] Where u(t) is the output of the controller, e(t) is the error signal, and K p is the proportional gain coefficient, K d is the differential gain coefficient, Time derivative of the error, K smc is the sliding mode gain, sgn(s) is the sign function;

[0027] According to the real-time vibration frequency f and amplitude error e, the proportional gain coefficient K is updated in real time based on the fuzzy rule base. p and differential gain coefficient K d .

[0028] Optionally, the method for hybrid execution unit to control the piezoelectric actuator and the magnetorheological damper according to the dynamically allocated actuator output value to perform hybrid vibration suppression includes:

[0029] Using a magnetorheological damper to suppress low-frequency vibrations below a first threshold;

[0030] A piezoelectric actuator is used to suppress high-frequency vibrations having a frequency higher than a first threshold.

[0031] The present application also discloses a suppression method of an adaptive pipeline vibration suppression system based on multimodal coupling analysis, the suppression method comprising:

[0032] Use a multi-sensor array to measure and obtain various sensor signals during pipeline vibration;

[0033] A wavelet decomposition feature extraction unit is used to extract feature data of a plurality of sensor signals, a prediction network unit is used to obtain probability distributions of various vibration modes based on the feature data, and a judgment unit is used to determine a dominant vibration frequency based on the probability distributions of the various vibration modes;

[0034] A fuzzy adaptive unit is used to dynamically distribute the actuator output value according to the dominant vibration frequency, and a hybrid execution unit is used to control the piezoelectric actuator and the magnetorheological damper according to the dynamically distributed actuator output value to perform hybrid vibration suppression.

[0035] The present application provides an adaptive pipeline vibration suppression system and method based on multimodal coupling analysis, which has the following technical effects:

[0036] The system can accurately identify the vibration source and use the rapid response of the magnetorheological damper to improve the control responsiveness. Through the mutual cooperation of the piezoelectric actuator and the magnetorheological damper, the exposure time of the piezoelectric actuator in the high temperature environment is reduced. The magnetorheological damper is resistant to high temperatures, thereby improving the overall high temperature resistance of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 This is a schematic diagram of an adaptive pipeline vibration suppression system based on multimodal coupling analysis according to the first embodiment of the present application;

[0038] Figure 2 This is a schematic diagram of the installation status of the adaptive pipeline vibration suppression system based on multimodal coupling analysis of Example 1 of the present application.

[0039] The correspondence between the reference numerals and component names is as follows:

[0040] 10-Perception module, 11-Multi-sensor array, 111-Three-axis MEMS accelerometer, 112-Fiber Bragg grating strain sensor, 113-Non-invasive ultrasonic flowmeter, 12-Adaptive anti-aliasing filter, 13-Notch filter, 14-Temperature compensation unit, 15-Edge computing unit, 20-Vibration analysis module, 21-Wavelet decomposition feature extraction unit, 22-Prediction network unit, 23-Judgment unit, 30-Control execution module, 31-Fuzzy adaptive unit, 32-Hybrid execution unit, 321-Piezoelectric actuator, 322-Magnetorheological damper. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solutions and advantages of this application more clearly understood, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0042] Before describing the various embodiments of the present application in detail, the technical concept of the present application is briefly described first: the current vibration suppression methods have problems such as inaccurate vibration tracing, delayed control response, and poor reliability in high-temperature environments. To this end, the present application provides an adaptive pipeline vibration suppression system and method based on multimodal coupling analysis. The system uses a multi-sensor array of a sensing module to measure and obtain multiple sensor signals during pipeline vibration. The vibration analysis module determines the dominant vibration frequency from the multiple sensor signals. Finally, a piezoelectric actuator and a magnetorheological damper are used to perform mixed suppression of the vibration. The system can accurately identify the vibration source and improve the control responsiveness by utilizing the rapid response of the magnetorheological damper. The interaction between the piezoelectric actuator and the magnetorheological damper reduces the exposure time of the piezoelectric actuator to the high-temperature environment. The magnetorheological damper has high-temperature resistance, thereby improving the overall high-temperature resistance of the system. The specific principles of the adaptive pipeline vibration suppression system and method based on multimodal coupling analysis of the present application are described below in conjunction with more embodiments.

[0043] Specifically, if Figure 1 and Figure 2 As shown, the adaptive pipeline vibration suppression system based on multimodal coupling analysis provided in this embodiment 1 includes a sensing module 10, a vibration analysis module 20, and a control execution module 30. The sensing module 10 includes a multi-sensor array 11. The multi-sensor array 11 is installed in the pipeline 100 and measures various sensor signals during vibration. The vibration analysis module 20 includes a wavelet decomposition feature extraction unit 21, a prediction network unit 22, and a judgment unit 23. The wavelet decomposition feature extraction unit 21 is used to extract feature data from the various sensor signals. The prediction network unit 22 is used to obtain the probability distribution of various vibration modes based on the feature data. The judgment unit 23 determines the dominant vibration frequency based on the probability distribution of various vibration modes. The control execution module 30 includes a fuzzy adaptive unit 31 and a hybrid execution unit 32. The fuzzy adaptive unit 31 dynamically allocates actuator output values based on the dominant vibration frequency. The hybrid execution unit 32 controls the piezoelectric actuator 321 and the magnetorheological damper 322 based on the dynamically allocated actuator output values to perform hybrid vibration suppression.

[0044] In one or more embodiments, the multi-sensor array 11 includes a triaxial MEMS accelerometer 111, a fiber Bragg grating (FBG) strain sensor 112, and a non-invasive ultrasonic flowmeter 113. The triaxial MEMS accelerometer 111 outputs a voltage signal, the fiber Bragg grating (FBG) strain sensor 112 outputs a wavelength shift signal, and the non-invasive ultrasonic flowmeter 113 outputs a pulse frequency signal. For example, the triaxial MEMS accelerometer 111 outputs an analog voltage signal (with a range of ±50g), which is converted into a digital signal via a high-precision ADC module (24-bit resolution, sampling rate ≥5kHz). The fiber Bragg grating (FBG) strain sensor 112 outputs a wavelength shift signal, which is converted into a digital strain value via an optoelectronic conversion module (including an FBG demodulator with a wavelength resolution of ±1pm) and transmitted via an RS-485 interface. The non-invasive ultrasonic flowmeter 113 outputs a pulse frequency signal (positively correlated with the steam pulsation pressure), and the frequency data is collected in real time via an FPGA counter module. Among them, all sensor data is aggregated to the edge computing gateway unit via the CAN bus (transmission rate 1Mbps, compliant with ISO11898 standard), ensuring low latency (<10ms) and high reliability.

[0045] In one or more embodiments, the sensing module 10 also includes an adaptive anti-aliasing filter 12, a notch filter 13 and a temperature compensation unit 14. The adaptive anti-aliasing filter 12 is used to filter a variety of sensing signals, the notch filter 13 is used to filter out power frequency interference from the voltage signal, and the temperature compensation unit 14 is used to perform temperature compensation on the wavelength offset signal. Among them, the original signals output by all sensors need to be filtered to avoid high-frequency noise interference and aliasing effects. The cutoff frequency is automatically adjusted according to the current signal spectrum characteristics (such as the main vibration frequency). For example, when the vibration frequency is 200Hz, the cutoff frequency is set to 250Hz (following the Nyquist theorem). The notch filter is a 50Hz notch filter (stopband width +2Hz, attenuation>40dB), which is only for MEMS accelerometer signals that are susceptible to power grid interference. Since the fiber Bragg grating strain sensor 112 is affected by temperature, wavelength drift will occur, so the temperature compensation unit 14 needs to be adjusted. The calculation formula of the temperature compensation unit 14 is as follows:

[0046] V corr =V raw ×[1+α(T-T0)]

[0047] Where V corr is the compensated sensor voltage signal, V raw is the raw sensor voltage signal, α is the temperature sensitivity coefficient, T is the current sensor ambient temperature, and T0 is the reference calibration temperature (factory calibration temperature). The triaxial MEMS accelerometer has a built-in temperature compensation chip, requiring no additional processing. Ultrasonic flowmeters utilize non-contact measurement and output a frequency signal, so temperature effects are negligible and require no compensation.

[0048] In one or more embodiments, the perception module 10 also includes an edge computing unit 15, which performs data caching, format standardization, and preliminary anomaly detection on the pulse frequency signal after filtering, the voltage signal after filtering out the power frequency interference, and the wavelength offset signal after temperature compensation.

[0049] In one or more embodiments, the method for extracting feature data of multiple sensor signals by the wavelet decomposition feature extraction unit includes: using the db4 wavelet basis to decompose the multi-layer wavelet packet of each sensor signal to obtain several sub-bands; extracting the energy entropy feature vector of each sub-band as feature data.

[0050] For example, the vibration signal is decomposed into 6 layers of wavelet packets using the db4 wavelet basis to ensure that the frequency resolution is less than 5 Hz within the 500 Hz frequency band. After decomposition, 64 sub-bands are generated (6 layers of decomposition correspond to 26 = 64 nodes). The energy entropy eigenvector F = [E1, E2, ... E64] of each sub-band is extracted using the following formula:

[0051]

[0052] Among them, P ij is the normalized energy ratio of the jth sampling point in the ith sub-band, and N is the number of sampling points.

[0053] In one or more embodiments, the prediction network unit 22 is a deep residual network. Exemplarily, the deep residual network adopts the ResNet-18 architecture: it contains 18 layers of depth and is composed of multiple residual blocks. Each residual block contains two layers of convolutional layers and a jump connection structure to solve the problem of vanishing gradients in deep networks. The input layer receives a 64-dimensional energy entropy feature vector, and the output layer generates vibration mode classification probabilities through a Softmax function. Training process: supervised learning is performed using a labeled vibration dataset (including fluid excitation, structural resonance, transient impact and other modes). The optimization target is the cross entropy loss function, and the Adam optimizer is used to dynamically adjust the learning rate (initial value 0.001).

[0054] Furthermore, the method by which the judgment unit 23 determines the dominant vibration frequency according to the probability distribution of various vibration modes includes: based on the gradient weighted class activation mapping technology, reversely analyzing the contribution of each sub-band in the feature map to the classification results of various vibration modes, generating a contribution ranking list, and taking the sub-band corresponding to the energy entropy feature vector with the largest weight as the dominant vibration frequency.

[0055] In one or more embodiments, the method by which the fuzzy adaptive unit 31 dynamically allocates the actuator output value according to the dominant vibration frequency includes:

[0056] Define the sliding surface: According to the current dominant vibration frequency f0, define the sliding surface Where λ = 2πf0, e(t) is the vibration amplitude error, is the error derivative;

[0057] Construct the control output equation of the actuator output value:

[0058] Where u(t) is the output of the controller, e(t) is the error signal, and K p is the proportional gain coefficient, K d is the differential gain coefficient, Time derivative of the error, K smc is the sliding mode gain, sgn(s) is the sign function;

[0059] According to the real-time dominant vibration frequency f and amplitude error e, the proportional gain coefficient K is updated in real time based on the fuzzy rule base. p and differential gain coefficient K d .

[0060] For example, according to the real-time dominant vibration frequency f and amplitude error e, fuzzification is performed into levels such as "low frequency", "medium frequency", "high frequency" and "small error", "medium error", and "large error". Fuzzy rule base: 25 rules are formulated to dynamically adjust K p With K d ,For example:

[0061] Rule 1: If f is low frequency and e is a large error, increase K p to 1.2, K d to 0.1.

[0062] Rule 2: If f is high frequency and e is small error, reduce K p to 0.8, K d To 0.05.

[0063] Then defuzzification is performed: the fuzzy output is converted into accurate parameter values using the centroid method.

[0064] When the system detects a change in vibration frequency or a sudden stress change (such as a sudden stress rate > 20 MPa / s), the fuzzy logic updates K in real time. p With K d , ensuring that the sliding surface converges quickly.

[0065] In one or more embodiments, a magnetorheological damper is used to suppress low-frequency vibrations below a first threshold, and a piezoelectric actuator is used to suppress high-frequency vibrations above the first threshold. For example, the first threshold is 100 Hz. The response time of the magnetorheological damper is less than or equal to 3 ms, and the output is greater than or equal to 2000 N. The piston diameter of the magnetorheological damper is Φ45 ± 0.1 mm, the magnetic gap width is 0.8 ± 0.05 mm, and the number of coil turns is 2 × 350 ± 5. The displacement of the piezoelectric actuator is ± 50 μm, and the phase adjustment accuracy is less than or equal to 9°.

[0066] The adaptive piping vibration suppression system based on multimodal coupling analysis was used to verify a 1000MW supercritical unit. The verification results are shown in Table 1 below:

[0067] index Before governance After governance Improvement rate Vibration speed RMS (mm / s) 12.3 3.8 69.1％ Pipeline stress amplitude (MPa) 45.6 13.2 71.1％ Control energy consumption (kW) 2.4 1.1 54.2％

[0068] Table 1

[0069] The test data for a power plant are as follows: (1) The vibration attenuation rate in the full frequency band of 0.1-500Hz is ≥69%; (2) The accuracy of fault mode recognition is 98.7%; (3) The actuator life in high temperature environment is >20,000 hours.

[0070] In one or more embodiments, a suppression method of an adaptive pipeline vibration suppression system based on multimodal coupling analysis includes the following steps:

[0071] Use a multi-sensor array to measure and obtain various sensor signals during pipeline vibration;

[0072] A wavelet decomposition feature extraction unit is used to extract feature data of various sensor signals, a prediction network unit is used to obtain probability distributions of various vibration modes based on the feature data, and a judgment unit is used to determine the dominant vibration frequency based on the probability distributions of various vibration modes;

[0073] The fuzzy adaptive unit is used to dynamically allocate the actuator output value according to the dominant vibration frequency, and the hybrid execution unit is used to control the piezoelectric actuator and magnetorheological damper according to the dynamically allocated actuator output value to perform hybrid vibration suppression.

[0074] The detailed process of each step of the suppression method can be referred to the description of the above embodiment and will not be repeated here.

[0075] The adaptive pipeline vibration suppression system and method based on multimodal coupling analysis provided in this embodiment can accurately assess and control steam pipeline vibration, enhance pipeline system reliability, and increase pipeline service life. This embodiment has broad application prospects in thermal power plants, petrochemicals, and other fields, significantly reducing equipment maintenance costs and downtime, and providing important economic and social benefits.

[0076] The above describes in detail the specific implementation methods of the present application. Although some embodiments have been shown and described, those skilled in the art should understand that these embodiments can be modified and improved without departing from the principles and spirit of the present application, the scope of which is defined by the claims and their equivalents. These modifications and improvements should also be within the scope of protection of the present application.

Claims

1. An adaptive pipeline vibration suppression system based on multimodal coupling analysis, characterized in that: The system comprises: A sensing module, comprising a multi-sensor array, wherein the multi-sensor array is installed in the pipeline and measures a plurality of sensor signals during the vibration process; a vibration analysis module comprising a wavelet decomposition feature extraction unit, a prediction network unit, and a judgment unit; the wavelet decomposition feature extraction unit is used to extract feature data of multiple sensor signals; the prediction network unit is used to obtain a probability distribution of various vibration modes based on the feature data; and the judgment unit determines a dominant vibration frequency based on the probability distribution of various vibration modes; A control execution module includes a fuzzy adaptive unit and a hybrid execution unit. The fuzzy adaptive unit dynamically allocates the actuator output value according to the dominant vibration frequency. The hybrid execution unit controls the piezoelectric actuator and the magnetorheological damper to perform hybrid vibration suppression according to the dynamically allocated actuator output value.

2. The adaptive pipeline vibration suppression system based on multimodal coupling analysis according to claim 1 is characterized in that: The multi-sensor array includes a three-axis MEMS accelerometer, a fiber Bragg grating strain sensor and a non-invasive ultrasonic flowmeter. The three-axis MEMS accelerometer outputs a voltage signal, the fiber Bragg grating strain sensor outputs a wavelength shift signal, and the non-invasive ultrasonic flowmeter outputs a pulse frequency signal.

3. The adaptive pipeline vibration suppression system based on multimodal coupling analysis according to claim 2 is characterized in that: The sensing module also includes an adaptive anti-aliasing filter, a notch filter and a temperature compensation unit. The adaptive anti-aliasing filter is used to filter multiple sensing signals, the notch filter is used to filter out power frequency interference from the voltage signal, and the temperature compensation unit is used to perform temperature compensation on the wavelength offset signal.

4. The adaptive pipeline vibration suppression system based on multimodal coupling analysis according to claim 2 is characterized in that: The perception module also includes an edge computing unit, which performs data caching, format standardization and preliminary anomaly detection on the pulse frequency signal after filtering, the voltage signal after filtering out power frequency interference, and the wavelength offset signal after temperature compensation.

5. The adaptive pipeline vibration suppression system based on multimodal coupling analysis according to claim 1 is characterized in that: The method for extracting characteristic data of a plurality of sensor signals by the wavelet decomposition feature extraction unit includes: The db4 wavelet basis is used to decompose the multi-layer wavelet packet of each sensor signal to obtain several sub-bands; The energy entropy feature vector of each sub-band is extracted as feature data.

6. The adaptive pipeline vibration suppression system based on multimodal coupling analysis according to claim 5 is characterized in that: The prediction network unit is a deep residual network.

7. The adaptive pipeline vibration suppression system based on multimodal coupling analysis according to claim 5 is characterized in that: The method for the judgment unit to determine the dominant vibration frequency according to the probability distribution of various vibration modes includes: Based on the gradient weighted class activation mapping technology, the contribution of each sub-band in the feature map to the classification results of various vibration modes is reversely analyzed to generate a contribution ranking list; The sub-band corresponding to the energy entropy eigenvector with the largest weight is taken as the dominant vibration frequency.

8. The adaptive pipeline vibration suppression system based on multimodal coupling analysis according to claim 1 is characterized in that: The method for the fuzzy adaptive unit to dynamically distribute the actuator output value according to the dominant vibration frequency includes: Define the sliding surface: According to the current dominant vibration frequency f0, define the sliding surface Where λ = 2πf0, e(t) is the vibration amplitude error, is the error derivative; Construct the control output equation of the actuator output value: Where u(t) is the output of the controller, e(t) is the error signal, and K p is the proportional gain coefficient, K d is the differential gain coefficient, Time derivative of the error, K smc is the sliding mode gain, sgn(s) is the sign function; According to the real-time vibration frequency f and amplitude error e, the proportional gain coefficient K is updated in real time based on the fuzzy rule base. p and differential gain coefficient K d .

9. The adaptive pipeline vibration suppression system based on multimodal coupling analysis according to claim 1 is characterized in that: The method for hybrid vibration suppression by controlling the piezoelectric actuator and the magnetorheological damper according to the dynamically allocated actuator output value by the hybrid execution unit includes: Using a magnetorheological damper to suppress low-frequency vibrations below a first threshold; A piezoelectric actuator is used to suppress high-frequency vibrations having a frequency higher than a first threshold.

10. A suppression method for an adaptive pipeline vibration suppression system based on multi-modal coupling analysis according to any one of claims 1 to 9, characterized in that: The inhibition method comprises: Use a multi-sensor array to measure and obtain various sensor signals during pipeline vibration; A wavelet decomposition feature extraction unit is used to extract feature data of a plurality of sensor signals, a prediction network unit is used to obtain probability distributions of various vibration modes based on the feature data, and a judgment unit is used to determine a dominant vibration frequency based on the probability distributions of the various vibration modes; A fuzzy adaptive unit is used to dynamically distribute the actuator output value according to the dominant vibration frequency, and a hybrid execution unit is used to control the piezoelectric actuator and the magnetorheological damper according to the dynamically distributed actuator output value to perform hybrid vibration suppression.

Images