A method and device for active vibration isolation of a reactor vibration isolation system

Abstract

This invention relates to an active vibration isolation method and apparatus for a reactor vibration isolation system, belonging to the technical field of reactor vibration isolation systems. The active vibration isolation method of this invention acquires vibration data of the reactor vibration isolation system in real time, enabling timely detection of performance defects and generation of repair plans. Simultaneously, in conjunction with the active vibration isolation apparatus of this invention, after generating the repair plan, the type and proportion of flexible material within the isolator's accommodating space can be adjusted in real time according to the repair plan, thereby achieving real-time repair of vibration defects.

Description

Technical Field

[0001] This invention relates to the field of reactor vibration isolation system technology, and in particular to an active vibration isolation method and device for a reactor vibration isolation system. Background Technology

[0002] In reactor vibration isolation systems, Chinese patent application CN119153221B discloses a vibration analysis method, program product, and electronic equipment for reactor vibration isolation systems. This patent employs passive vibration isolation technology. By analyzing the vibration characteristics of the vibration isolation support, a finite element model and modal analysis are established to optimize the structural parameters and damping materials of the support, thereby improving passive vibration isolation performance from a design perspective and solving the problem of balancing damping effect and vibration isolation performance. However, in the passive vibration isolation technology of the aforementioned patent, the designed target vibration isolation support cannot be adjusted in real time. When it ages or the vibration source changes, the entire support needs to be replaced.

[0003] In the face of changes in operating conditions or external disturbances, there is a need for an active vibration isolation method that can be adjusted in real time. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides an active vibration isolation method and apparatus for a reactor vibration isolation system.

[0005] An active vibration isolation method for a reactor vibration isolation system includes the following steps: S1: Real-time acquisition of vibration data of the reactor vibration isolation system under working conditions, extraction of feature parameters based on the vibration data and construction of vibration feature vector; S2: Analyze the vibration defect type of the reactor isolation system and quantify the performance deviation based on the vibration feature vector; S3: Establish a correspondence matrix between the vibration defect type and the physical properties of the filling material, and determine the physical properties of the filling material based on the vibration defect type; S4: Construct an adaptive neural fuzzy reasoning system; S5: Based on the adaptive neural fuzzy inference system, determine the final repair scheme according to the performance deviation and the physical properties of the filling material. The final repair scheme includes the type and ratio of the filling material. S6: According to the final repair plan, retrieve the corresponding proportion of filling material from the material pool and inject it into the vibration isolation cavity of the vibration isolator to achieve defect repair.

[0006] Preferably, step S1 includes: S11: Collect the vibration data in real time using sensors; S12: Preprocess the vibration data; S13: Extract vibration parameters from the vibration data and construct a vibration feature vector.

[0007] Preferably, in step S11, a piezoelectric accelerometer with a sensitivity of 0.1V / g and a frequency response range of 0.1Hz~1000Hz is selected, with a sampling frequency of 2560Hz and a sampling time of 60 seconds, to collect vibration data of the reactor vibration isolation system under different working conditions.

[0008] Preferably, step S12 includes: S121: Wavelet decomposition is performed using wavelet transform, and the wavelet coefficients are soft-thresholded using the thresholding rule to remove high-frequency noise interference. S122: The Z-score normalization method is used to normalize the denoised vibration data to eliminate the influence of different dimensions and orders of magnitude. S123: Perform data cleaning on the normalized vibration data, identify outlier data points, and ensure data reliability.

[0009] Preferably, step S13 includes: S131: Wavelet transform is used to analyze the preprocessed vibration data to obtain characteristic parameters such as vibration amplitude, frequency, attenuation coefficient, and spectral characteristics; S132: The fast Fourier transform algorithm is used to convert the preprocessed vibration data from the time domain to the frequency domain and perform frequency domain analysis to extract frequency characteristic parameters such as the dominant frequency, secondary frequency and their amplitude; at the same time, the root mean square value, peak factor and other statistical characteristic parameters of the time domain signal are calculated. S133: Combining feature importance ranking and correlation analysis based on decision tree, feature parameters are screened according to their contribution to vibration signal identification, and a vibration feature vector is constructed.

[0010] Preferably, step S2 includes: S21: Compare the feature parameters in the vibration feature vector with the parameter template in the healthy state. When the feature parameters exceed the set threshold range, it is determined that a performance defect has occurred, and the performance deviation is quantitatively evaluated. S22: Obtain the historical fault dataset of the reactor vibration isolation system. The historical fault dataset includes input data and output data. The input data includes vibration feature vectors when performance defects occur. The output data includes vibration defect types. S23: Construct a vibration defect type identification model, and use historical datasets to train and optimize the vibration defect type identification model to obtain an optimized vibration defect type identification model; S24: Input the vibration feature vector collected in real time into the optimized vibration defect type identification model to obtain the vibration defect type.

[0011] Preferably, the vibration defect type identification model is constructed using algorithms such as support vector machine, decision tree, and random forest.

[0012] Preferably, step S4 includes: S41: Establish a material property database, which includes material parameters such as the physical properties, mechanical properties, and damping characteristics of the filling material; S42: Through finite element simulation and experimental testing, the performance of different materials with the same physical properties is verified by using them alone and in different combinations. The defect repair capabilities of the filling material under different types and ratios are obtained, and a filling material dataset is formed. S43: Based on the filling material dataset, construct an adaptive neural fuzzy inference system, establish a nonlinear mapping relationship between material parameters and defect repair capability, and use a hybrid learning algorithm to optimize the weights and fuzzy rule parameters of the adaptive neural fuzzy inference system.

[0013] Preferably, step S5 includes: S51: Input the performance deviation and the physical properties of the filling material into the adaptive neural fuzzy inference system. The adaptive neural fuzzy inference system will initially match the types of filling materials that can make up for the performance gap and adapt to the defect type based on the nonlinear mapping relationship. S52: Based on minimizing performance deviation, the adaptive neural fuzzy inference system uses its fuzzy inference function to perform intelligent inference and calculation of the proportion parameters for the initially matched material types, outputs the optimal proportion scheme for the material type, and forms a preliminary repair scheme. S53: Perform vibration isolation performance simulation verification on the preliminary repair plan to confirm whether the preliminary repair plan can effectively compensate for the current performance deviation of the system; if the simulation verification meets the standard, it is directly determined as the final repair plan; if it does not meet the standard, return to step S51 until the simulation verification meets the standard.

[0014] An active vibration isolation device for a reactor vibration isolation system, comprising: A vibration isolator, comprising a housing, a movable plate, and a vibration isolation cavity, wherein the vibration isolation cavity is an accommodating space formed by the housing and the movable plate, for filling with a flexible material; A reactor, wherein the reactor is disposed on the movable plate; A material library, which is connected to the vibration isolation cavity, contains a variety of filling materials; A processor, the processor being configured to execute the active vibration isolation method for a reactor vibration isolation system as described in any one of claims 1-9, and output a final repair solution; A controller, connected to the material library, is used to control the material library to retrieve the corresponding filling material according to the final repair plan, and to fill the vibration isolation cavity with the filling material.

[0015] The technical solution of the present invention has the following advantages compared with the prior art: The active vibration isolation method for reactor vibration isolation system described in this invention can promptly detect performance defects and generate repair plans by acquiring the vibration data of the reactor vibration isolation system in real time. At the same time, in conjunction with the active vibration isolation device for reactor vibration isolation system described in this invention, after generating the repair plan, the type and ratio of flexible materials in the vibration isolator accommodating space can be adjusted in real time according to the repair plan, thereby realizing real-time repair of vibration defects. Attached Figure Description

[0016] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings.

[0017] Figure 1 This is a flowchart of the active vibration isolation method of the reactor vibration isolation system of the present invention.

[0018] Figure 2 This is a schematic diagram of the active vibration isolation device of the reactor vibration isolation system of the present invention.

[0019] Explanation of reference numerals in the accompanying drawings: 101, outer casing; 102, movable plate; 103, vibration isolation cavity; 2, reactor; 3, material library; 4, processor; 5, controller. Detailed Implementation

[0020] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.

[0021] This application provides an active vibration isolation method for a reactor vibration isolation system, including the following steps: S1: Real-time acquisition of vibration data of the reactor vibration isolation system under working conditions, extraction of feature parameters based on the vibration data and construction of vibration feature vector.

[0022] In one specific embodiment, step S1 includes: S11: Collect the vibration data in real time using sensors; S12: Preprocess the real-time vibration data; S13: Extract vibration parameters from the real-time vibration data and construct a real-time vibration feature vector.

[0023] In an optional embodiment, step S11 includes: selecting a piezoelectric accelerometer with a sensitivity of 0.1V / g and a frequency response range of 0.1Hz~1000Hz, setting the sampling frequency to 2560Hz, and the sampling time to 60 seconds, to collect vibration data of the reactor vibration isolation system under different operating conditions.

[0024] In an optional embodiment, step S12 includes: S121: Wavelet transform is used to perform wavelet decomposition, and the wavelet coefficients are soft-thresholded using the threshold rule to remove high-frequency noise interference; specifically, the Dobessi fourth-order wavelet is selected, and the decomposition level is 5. S122: The Z-score normalization method is used to normalize the denoised vibration data to eliminate the influence of different dimensions and orders of magnitude. S123: Perform data cleaning on the normalized vibration data, identify outlier data points, and ensure data reliability.

[0025] In an optional embodiment, step S13 includes: S131: Wavelet transform is used to analyze the preprocessed vibration data to obtain characteristic parameters such as vibration amplitude, frequency, attenuation coefficient, and spectral characteristics; S132: The fast Fourier transform algorithm is used to convert the preprocessed vibration data from the time domain to the frequency domain and perform frequency domain analysis to extract frequency characteristic parameters such as the main frequency, secondary frequency and their amplitude; at the same time, the root mean square value, peak factor and other statistical characteristic parameters of the time domain signal are calculated to comprehensively reflect the energy and waveform characteristics of the vibration signal. S133: Combining feature importance ranking and correlation analysis based on decision tree, feature parameters are screened according to their contribution to vibration signal identification, and a vibration feature vector is constructed.

[0026] S2: Analyze the vibration defect type of the reactor isolation system based on the vibration feature vector and quantify the performance deviation.

[0027] In one specific embodiment, step S2 includes: S21: Compare the feature parameters in the vibration feature vector with the parameter template in the healthy state. When the feature parameters exceed the set threshold range, it is determined that a performance defect has occurred, and the performance deviation is quantitatively evaluated. The performance deviation is the deviation between the feature parameters and the parameter template in the healthy state. S22: Obtain the historical fault dataset of the reactor vibration isolation system. The historical fault dataset includes input data and output data. The input data includes vibration feature vectors when performance defects occur. The output data includes vibration defect types. S23: Construct a vibration defect type identification model, and use historical datasets to train and optimize the vibration defect type identification model to obtain an optimized vibration defect type identification model; S24: Input the real-time vibration feature vector into the optimized vibration defect type identification model to obtain the vibration defect type.

[0028] In one specific embodiment, the historical fault dataset is constructed and improved based on historical fault cases and expert experience to provide knowledge support for the identification and diagnosis of performance defects.

[0029] In an optional embodiment, step S23 includes: S231: constructing a vibration defect type identification model using a support vector machine; S232: training the vibration defect type identification model using a historical dataset; S233: optimizing the kernel function type and penalty coefficient of the vibration defect type identification model using methods such as cross-validation and grid search to obtain an optimized vibration defect type identification model. In a specific embodiment, this application uses a support vector machine to construct the vibration defect type identification model, employs a radial basis function kernel function, and obtains the optimal penalty coefficient C=10 and kernel function parameter g=0.1 through 5-fold cross-validation and grid search. The final optimized vibration defect type identification model achieves a defect identification accuracy of 95%.

[0030] In an optional embodiment, step S23 includes: S231: constructing a vibration defect type identification model using a decision tree; S232: training the vibration defect type identification model using a historical dataset; S233: optimizing the maximum depth and minimum number of leaf node samples of the vibration defect type identification model using methods such as cross-validation and grid search to obtain an optimized vibration defect type identification model.

[0031] In an optional embodiment, step S23 includes: S231: constructing a vibration defect type identification model using a random forest; S232: training the vibration defect type identification model using a historical dataset; S233: optimizing the number of decision trees and the size of the feature subset of the vibration defect type identification model using methods such as cross-validation and grid search to obtain an optimized vibration defect type identification model.

[0032] By optimizing the vibration defect type identification model, the accuracy and generalization ability of defect identification can be improved.

[0033] S3: Establish a correspondence matrix between the vibration defect type and the physical properties of the filling material, and determine the physical properties of the filling material based on the vibration defect type.

[0034] Different types of defect characteristics require different filling materials with varying physical properties. For example, for stiffness degradation defects, high-modulus and high-strength filling materials are selected to compensate for the stiffness loss of the vibration isolation system and restore its vibration transmission characteristics through the high stiffness characteristics of the materials. For insufficient damping defects, high-damping materials are selected as filling materials to increase the energy dissipation capacity of the vibration isolation system, suppress the resonance amplification effect, and improve the vibration isolation effect of the system.

[0035] S4: Construct an adaptive neural fuzzy reasoning system.

[0036] In one specific embodiment, step S4 includes: S41: Establish a material property database, which includes parameters such as the physical properties, mechanical properties, and damping characteristics of the filling material; S42: Through finite element simulation and experimental testing, the performance of different materials with the same physical properties is verified by using them alone and in different combinations. The defect repair capabilities of the filling material under different types and ratios are obtained, and a filling material dataset is formed. S43: Based on the filling material dataset, construct an adaptive neural fuzzy inference system, establish a nonlinear mapping relationship between material parameters and defect repair capability, and use a hybrid learning algorithm to optimize the weights and fuzzy rule parameters of the adaptive neural fuzzy inference system.

[0037] S5: Based on the adaptive neural fuzzy inference system, determine the final repair scheme according to the performance deviation and the physical properties of the filling material. The repair scheme includes the type and ratio of the filling material.

[0038] In one specific embodiment, step S5 includes: S51: Input the performance deviation and the physical properties of the filling material into the adaptive neural fuzzy inference system. The adaptive neural fuzzy inference system will initially match the types of filling materials that can make up for the performance gap and adapt to the defect type based on the nonlinear mapping relationship. S52: Based on minimizing performance deviation, the adaptive neural fuzzy inference system uses its fuzzy inference function to perform intelligent inference and calculation of the proportion parameters for the initially matched material types, outputs the optimal proportion scheme for the material type, and forms a preliminary repair scheme. S53: Perform vibration isolation performance simulation verification on the preliminary repair plan to confirm whether the preliminary repair plan can effectively compensate for the current performance deviation of the system; if the simulation verification meets the standard, it is directly determined as the final repair plan; if it does not meet the standard, return to step S51 until the simulation verification meets the standard.

[0039] S6: According to the final repair plan, retrieve the corresponding proportion of filling material from the material pool and inject it into the vibration isolation cavity of the vibration isolator to achieve defect repair.

[0040] This application also provides an active vibration isolation device for a reactor vibration isolation system, comprising: A vibration isolator, comprising a housing 101, a movable plate 102, and a vibration isolation cavity 103, wherein the vibration isolation cavity 103 is an accommodating space formed by the housing 101 and the movable plate 102, for filling with flexible material; Reactor 2, wherein the reactor 2 is disposed on the movable plate 102; Material library 3, which is connected to the vibration isolation cavity 103, contains a variety of flexible filling materials; Processor 4 is used to execute the active vibration isolation method of the reactor vibration isolation system as described above, and output the final repair solution; The controller 5 is connected to the material library 3 and is used to control the material library 3 to retrieve the corresponding filling material according to the final repair plan and fill the filling material into the vibration isolation cavity 103.

[0041] In this embodiment, the flexible filling material in the vibration isolation cavity 103 can be used for vibration resistance. By adjusting the ratio and type of the flexible filling material in the vibration isolation cavity 103 in real time, the vibration isolation effect of the vibration isolator can be effectively adjusted.

[0042] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. An active vibration isolation method for a reactor vibration isolation system, characterized in that, Includes the following steps: S1: Real-time acquisition of vibration data of the reactor vibration isolation system under working conditions, extraction of feature parameters based on the vibration data and construction of vibration feature vector; S2: Analyze the vibration defect type of the reactor isolation system and quantify the performance deviation based on the vibration feature vector; S3: Establish a correspondence matrix between the vibration defect type and the physical properties of the filling material, and determine the physical properties of the filling material based on the vibration defect type; S4: Construct an adaptive neural fuzzy reasoning system; S5: Based on the adaptive neural fuzzy inference system, determine the final repair scheme according to the performance deviation and the physical properties of the filling material. The final repair scheme includes the type and ratio of the filling material. S6: According to the final repair plan, retrieve the corresponding proportion of filling material from the material pool and inject it into the vibration isolation cavity of the vibration isolator to achieve defect repair.

2. The active vibration isolation method for a reactor vibration isolation system according to claim 1, characterized in that, Step S1 includes: S11: Collect the vibration data in real time using sensors; S12: Preprocess the vibration data; S13: Extract vibration parameters from the vibration data and construct a vibration feature vector.

3. The active vibration isolation method for a reactor vibration isolation system according to claim 2, characterized in that, In step S11, a piezoelectric accelerometer with a sensitivity of 0.1V / g and a frequency response range of 0.1Hz~1000Hz is selected, the sampling frequency is 2560Hz, and the sampling time is 60 seconds to collect vibration data of the reactor vibration isolation system under different working conditions.

4. The active vibration isolation method for a reactor vibration isolation system according to claim 2, characterized in that, Step S12 includes: S121: Wavelet decomposition is performed using wavelet transform, and the wavelet coefficients are soft-thresholded using the thresholding rule to remove high-frequency noise interference. S122: The Z-score normalization method is used to normalize the denoised vibration data to eliminate the influence of different dimensions and orders of magnitude. S123: Perform data cleaning on the normalized vibration data, identify outlier data points, and ensure data reliability.

5. The active vibration isolation method for a reactor vibration isolation system according to claim 2, characterized in that, Step S13 includes: S131: Wavelet transform is used to analyze the preprocessed vibration data to obtain characteristic parameters such as vibration amplitude, frequency, attenuation coefficient, and spectral characteristics; S132: The fast Fourier transform algorithm is used to convert the preprocessed vibration data from the time domain to the frequency domain and perform frequency domain analysis to extract frequency characteristic parameters such as the dominant frequency, secondary frequency and their amplitude; at the same time, the root mean square value, peak factor and other statistical characteristic parameters of the time domain signal are calculated. S133: Combining feature importance ranking and correlation analysis based on decision tree, feature parameters are screened according to their contribution to vibration signal identification, and a vibration feature vector is constructed.

6. The active vibration isolation method for a reactor vibration isolation system according to claim 1, characterized in that, Step S2 includes: S21: Compare the feature parameters in the vibration feature vector with the parameter template in the healthy state. When the feature parameters exceed the set threshold range, it is determined that a performance defect has occurred, and the performance deviation is quantitatively evaluated. S22: Obtain the historical fault dataset of the reactor vibration isolation system. The historical fault dataset includes input data and output data. The input data includes vibration feature vectors when performance defects occur. The output data includes vibration defect types. S23: Construct a vibration defect type identification model, and use historical datasets to train and optimize the vibration defect type identification model to obtain an optimized vibration defect type identification model; S24: Input the vibration feature vector collected in real time into the optimized vibration defect type identification model to obtain the vibration defect type.

7. The active vibration isolation method for a reactor vibration isolation system according to claim 6, characterized in that, The vibration defect type identification model is constructed using algorithms such as support vector machine, decision tree, and random forest.

8. The active vibration isolation method for a reactor vibration isolation system according to claim 1, characterized in that, Step S4 includes: S41: Establish a material property database, which includes material parameters such as the physical properties, mechanical properties, and damping characteristics of the filling material; S42: Through finite element simulation and experimental testing, the performance of different materials with the same physical properties is verified by using them alone and in different combinations. The defect repair capabilities of the filling material under different types and ratios are obtained, and a filling material dataset is formed. S43: Based on the filling material dataset, construct an adaptive neural fuzzy inference system, establish a nonlinear mapping relationship between material parameters and defect repair capability, and use a hybrid learning algorithm to optimize the weights and fuzzy rule parameters of the adaptive neural fuzzy inference system.

9. The active vibration isolation method for a reactor vibration isolation system according to claim 1, characterized in that, Step S5 includes: S51: Input the performance deviation and the physical properties of the filling material into the adaptive neural fuzzy inference system. The adaptive neural fuzzy inference system will initially match the types of filling materials that can make up for the performance gap and adapt to the defect type based on the nonlinear mapping relationship. S52: Based on minimizing performance deviation, the adaptive neural fuzzy inference system uses its fuzzy inference function to perform intelligent inference and calculation of the proportion parameters for the initially matched material types, outputs the optimal proportion scheme for the material type, and forms a preliminary repair scheme. S53: Perform vibration isolation performance simulation verification on the preliminary repair plan to confirm whether the preliminary repair plan can effectively compensate for the current performance deviation of the system; if the simulation verification meets the standard, it is directly determined as the final repair plan; if it does not meet the standard, return to step S51 until the simulation verification meets the standard.

10. An active vibration isolation device for a reactor vibration isolation system, characterized in that, include: A vibration isolator, comprising a housing, a movable plate, and a vibration isolation cavity, wherein the vibration isolation cavity is an accommodating space formed by the housing and the movable plate, for filling with a flexible material; A reactor, wherein the reactor is disposed on the movable plate; A material library, which is connected to the vibration isolation cavity, contains a variety of filling materials; A processor, the processor being configured to execute the active vibration isolation method for a reactor vibration isolation system as described in any one of claims 1-9, and output a final repair solution; A controller, connected to the material library, is used to control the material library to retrieve the corresponding filling material according to the final repair plan, and to fill the vibration isolation cavity with the filling material.

Images