Vibration data processing and evaluation method for large-scale ultra-precise engineering site

Abstract

The invention discloses a method for processing and evaluating vibration data of a large-scale ultra-precise engineering site. The method comprises the following steps: step 1, acquiring original data; step 2, preprocessing the obtained original data; step 3, performing filtering processing on the frequency signal components; 4, frequency domain calculus is carried out for acceleration or speed signals collected in the vibration signals; 5, adopting a random decrement method or an NExT method to carry out modal identification preprocessing on the vibration data processed in the step 4, and then adopting an ITD method to carry out modal parameter identification; step 6, on the basis of modal analysis and vibration main characteristic analysis and evaluation, based on wavelet transform, decomposing and reconstructing the vibration signal; and step 7, evaluating the overall vibration signal and the main modal component distribution signal after multi-band evaluation by adopting a one-third octave spectrum in combination with a VC curve.

Description

Technical Field

[0001] The present invention relates to the field of vibration control technology, and more specifically to a set of technical methods for vibration data processing and evaluation of large-scale ultra-precision engineering sites, which can be applied, for example, to site vibration testing and evaluation of large-scale ultra-precision engineering projects such as electronic industrial plants and large scientific projects. Background Art

[0002] Large-scale ultra-precision projects (such as chip manufacturing, high-end optical equipment manufacturing, precision measurement, etc.) have extremely high requirements for the site vibration environment. In these projects, even tiny vibrations may affect the performance, processing precision and measurement accuracy of the equipment, and thus affect product quality and production efficiency. For example, the nano-level processing technology in the chip manufacturing process is extremely sensitive to vibration, and even extremely small vibrations may cause problems such as chip circuit defects.

[0003] Site vibration may be caused by a variety of factors, including surrounding traffic (such as vehicle driving, rail transit operation), industrial production activities (such as factory equipment operation), natural environmental factors (such as wind load, seismic wave propagation) and the operation of the site's own equipment. The vibrations from these different sources are superimposed on each other, making the vibration signals complex and diverse. In large-scale ultra-precision engineering sites, long-term monitoring involving many measuring points is required to comprehensively evaluate the vibration situation. Long-term monitoring generates a large amount of data, and many measuring points mean high data dimensions, which brings huge challenges to data processing and analysis. Different working conditions may exist in different areas of the site, such as different operating conditions of production equipment in different workshops, changes in traffic flow in different time periods, etc., resulting in different characteristics of vibration data under different working conditions, which increases the difficulty of data processing. In the process of conventional data collection, the characteristics of the collection equipment, environmental factors, etc. may cause trend items and zero drift in the data, affecting the accuracy of the data and the reliability of subsequent analysis results. For example, temperature changes may cause the sensor to drift slowly, causing the measured data to deviate from the true value. Factors such as electromagnetic interference in the environment and the sensor's own noise make the collected data contain a large number of burr components, which obscures the true characteristics of the vibration signal, easily misleading the analysis results, and making it difficult to accurately obtain useful information from the vibration signal. Vibration signals contain multiple frequency components, but existing methods have difficulty in effectively separating and processing signals in different frequency ranges, and cannot accurately obtain the characteristics of vibration signals within a specific frequency range, which is not conducive to in-depth analysis of the impact of vibration on ultra-precision engineering. In terms of modal identification, when dealing with complex vibration data of large ultra-precision engineering sites, traditional methods may have problems such as inaccurate modal parameter identification and inability to effectively extract the main modal components, and cannot fully and accurately reflect the vibration modal characteristics of the site. There is a lack of effective comprehensive evaluation methods for the overall vibration signal and the distribution signal of the main modal components, making it difficult to comprehensively and intuitively evaluate the impact of site vibration on ultra-precision engineering, and unable to provide sufficient basis for engineering decision-making. Summary of the Invention

[0004] In view of the above problems, the present invention proposes a complete set of technical methods for processing and evaluating vibration data of large-scale ultra-precision engineering sites. The technology aims to comprehensively and intuitively evaluate the impact of site vibration on ultra-precision engineering and can provide a basis for engineering decision-making.

[0005] More specifically, according to one aspect of the present invention, there is provided a method for processing and evaluating vibration data of large-scale ultra-precision engineering sites, including:

[0006] Step 1, actually conduct vibration tests on-site to obtain raw data;

[0007] Step 2, preprocess the obtained raw data;

[0008] Step 3, respectively use frequency-domain low-pass and band-pass filtering, and frequency-domain high-pass and band-stop filtering to filter the high- and low-pass frequency signal components to obtain the vibration signal characteristics in a specified frequency range;

[0009] Step 4, for the acceleration or velocity signals collected in the vibration signal after preprocessing in Step 2, conduct frequency-domain calculus to achieve the conversion between acceleration and velocity, and perform batch analysis and processing;

[0010] Step 5, use the random decrement method or the NExT method to perform modal identification preprocessing on the vibration data that has been processed in Step 4, and then use the ITD method to identify modal parameters and extract different modal components in the site vibration;

[0011] Step 6: Based on modal analysis and the analysis and evaluation of the main vibration characteristics in Step 3, decompose the vibration signal based on wavelet transform to obtain the time-frequency domain components of the vibration signal under the main modal components and achieve signal decomposition; and

[0012] Step 7: Use the one-third octave spectrum combined with the VC curve to evaluate the overall vibration signal and the distribution signals of the main modal components after multi-band evaluation.

[0013] According to the implementation scheme of the present invention, in Step 2, the preprocessing includes removing the trend term in the data by using the moving average method or the least square method.

[0014] According to the implementation scheme of the present invention, in Step 2, the preprocessing further includes smoothing the data after removing the trend term by using the five-point cubic smoothing method or the five-point moving average method.

[0015] According to an embodiment of the present invention, in step five, the random reduction method includes intercepting multiple sub-samples from the vibration response signal, calculating the average value of the sub-samples, and obtaining an approximate free decay response signal through multiple interception and averaging operations; the NExT method includes calculating the cross-correlation function between different measuring points and extracting the free decay part in the cross-correlation function.

[0016] According to the implementation scheme of the present invention, in step five, the modal parameter identification using the ITD method includes: defining the modal order according to the characteristics and experience of the engineering structure, then solving the free vibration response of the structure, and finally solving the characteristic equation based on the least squares method to obtain the modal parameters of the structure, accurately extracting different modal components in the site vibration, and accurately reflecting the vibration modal characteristics of the site.

[0017] According to the implementation scheme of the present invention, step six includes selecting a suitable wavelet basis and determining the scale of wavelet decomposition, decomposing the vibration signal into detail signals and approximate signals at different scales, obtaining the time-frequency domain components of the vibration signal under the main modal components, and deeply analyzing the characteristics of the vibration signal in time and frequency; during wavelet reconstruction, the wavelet coefficients obtained by decomposition are used to restore the signal according to the wavelet reconstruction algorithm, and then the reconstructed signal is evaluated by FFT, the error between the reconstructed signal and the original signal in the frequency domain is calculated, and the accuracy of the reconstruction is verified.

[0018] According to the implementation scheme of the present invention, step seven includes: first, calculating the one-third octave band spectrum of the vibration signal, dividing the frequency range into multiple one-third octave bands, calculating parameters such as the vibration energy or amplitude in each band, analyzing the distribution of vibration energy in different frequency bands, and determining whether there is vibration anomaly in a specific frequency band; at the same time, according to the vibration requirements of the engineering site and the vibration standards allowed by the equipment, drawing the VC curve, comparing the one-third octave band spectrum of the actual vibration signal with the VC curve, and evaluating whether the site vibration meets the requirements of ultra-precision engineering.

[0019] This invention has many beneficial effects in the processing and evaluation of vibration data in large-scale ultra-precision engineering sites:

[0020] 1) Improved data accuracy

[0021] Effectively handle zero drift and trend items, remove trend items through sliding average method or least square method, solve the problem of data zero drift, make long-term monitoring data and disturbed data more accurate and reliable, avoid data deviation caused by zero drift, and turn invalid data into valid data, providing a more accurate basis for subsequent analysis.

[0022] Remove noise interference and use the five-point cubic smoothing method or the five-point sliding average method to smooth the data, effectively removing burrs, noise and other components in the data, making the data closer to the actual vibration situation, reducing the interference of noise on the extraction of vibration signal features, and improving the quality and analyzability of the data.

[0023] 2) Frequency component processing optimization

[0024] Accurately obtain specific frequency characteristics. With the help of frequency domain low-pass, band-pass, high-pass and band-stop filtering, and reasonable selection of cut-off frequency, it is possible to accurately obtain the vibration signal characteristics of the specified frequency range, which is helpful for in-depth analysis of the impact of vibrations of different frequency components on ultra-precision engineering. For example, it is possible to specifically study specific frequency vibrations that have a greater impact on equipment operation, providing a more targeted basis for engineering design and equipment protection.

[0025] Flexible signal conversion and analysis, frequency domain calculus processing realizes the conversion of acceleration and velocity, and can eliminate frequency components outside the specified frequency band. It can quickly and effectively evaluate large quantities of data in different domains, facilitate the analysis of vibration data in different physical quantity domains according to actual needs, and better understand the essential characteristics of vibration.

[0026] 3) Improved accuracy of modal analysis

[0027] Accurate modal identification preprocessing, using random reduction method or NExT method for modal identification preprocessing, by reasonably intercepting the amplitude and multiples of the input signal standard deviation and superimposing the input signal sub-samples, can more effectively extract modal related information, laying the foundation for subsequent accurate modal parameter identification.

[0028] Accurate modal parameter identification and the application of the ITD method in the effective definition of modal orders, free vibration response solution, and characteristic equation solution based on the least squares method make it possible to accurately extract different modal components in site vibration and accurately reflect the vibration modal characteristics of the site. This helps to gain a deeper understanding of the inherent structure and laws of site vibration and provide key parameters for the dynamic analysis and optimal design of engineering structures.

[0029] 4) Comprehensive evaluation is more comprehensive

[0030] In-depth analysis of time-frequency domain signals, signal decomposition and reconstruction based on wavelet transform, by selecting appropriate wavelet basis and decomposition scale, can obtain the time-frequency domain components of vibration signals under the main modal components, deeply explore the characteristics of vibration signals in time and frequency, and more comprehensively understand the complex characteristics of vibration signals, providing multi-dimensional information for evaluating the impact of vibration on ultra-precision engineering.

[0031] The comprehensive evaluation method is effective. It uses one-third octave band spectrum combined with VC curve for evaluation, and comprehensively considers factors such as frequency distribution and vibration amplitude. It can comprehensively and intuitively evaluate the impact of site vibration on ultra-precision engineering, providing a sufficient and scientific basis for engineering decision-making, and helping to formulate reasonable vibration control strategies and engineering design plans to ensure the normal operation and product quality of ultra-precision engineering.

[0032] Based on the following detailed description of specific embodiments of the present invention in conjunction with the accompanying drawings, those skilled in the art will become more aware of the above and other objects, advantages and features of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 A schematic flow chart of a method for processing and evaluating vibration data for a large ultra-precision engineering site according to an embodiment of the present invention;

[0034] Figure 2 This is a diagram showing the effect of eliminating trend items (sliding average method) in a method for processing and evaluating vibration data of a large ultra-precision engineering site according to an embodiment of the present invention;

[0035] Figure 3 A schematic diagram of low-pass and high-pass filtering effects in a method for processing and evaluating vibration data of a large ultra-precision engineering site according to an embodiment of the present invention;

[0036] Figure 4 A schematic diagram of stripping free vibration signals using a random reduction method in a method for processing and evaluating vibration data for a large ultra-precision engineering site according to an embodiment of the present invention;

[0037] Figure 5 Schematic diagram of the cross-correlation function calculated by the NExT method in the method for processing and evaluating vibration data of large ultra-precision engineering sites according to an embodiment of the present invention;

[0038] Figure 6 It is a free vibration response fitting schematic diagram of the ITD method in the vibration data processing and evaluation method for large ultra-precision engineering sites according to an embodiment of the present invention;

[0039] Figure 7 is a wavelet decomposition diagram of a vibration signal used in a method for processing and evaluating vibration data of a large ultra-precision engineering site according to an embodiment of the present invention;

[0040] Figure 8 is a spectrum analysis diagram of a vibration signal after wavelet decomposition in a method for processing and evaluating vibration data of a large ultra-precision engineering site according to an embodiment of the present invention;

[0041] Fig. 9is a schematic diagram of wavelet reconstruction of vibration signals in a method for processing and evaluating vibration data of a large ultra-precision engineering site according to an embodiment of the present invention; and

[0042] Fig.10 It is a schematic diagram of one-third octave band spectrum and VC curve evaluation in a vibration data processing and evaluation method for large ultra-precision engineering sites according to an implementation scheme of the present invention. DETAILED DESCRIPTION

[0043] The present invention is further described in detail below through specific embodiments in conjunction with the accompanying drawings. The shown contents are used to fully illustrate the contents of the present invention, but are not used to limit the present invention.

[0044] It should be understood that the models and tools involved in the present invention, such as the sliding average method, the least squares method, the five-point cubic smoothing method, the five-point moving average method, the random decrement method, the NExT method, the ITD method, the wavelet transform, etc., are themselves known. Therefore, the present invention focuses on how to combine and apply the above-mentioned various tools or models to design the process of the present invention for the vibration data processing and evaluation method of large-scale ultra-precision engineering sites.

[0045] Figure 1 Schematic diagram of a process for processing and evaluating vibration data of a large ultra-precision engineering site according to an embodiment of the present invention. Figure 1 According to an embodiment of the present invention, a method for processing and evaluating vibration data for a large ultra-precision engineering site may include:

[0046] First, data collection and preliminary preparation are carried out. Specifically, vibration data collection is carried out, and appropriate vibration sensors (such as accelerometers, etc.) are arranged at large ultra-precision engineering sites. The number and location of measurement points are determined according to the site scale and monitoring requirements to ensure that the site vibration conditions can be fully reflected. The sensor must have sufficient accuracy and sensitivity to accurately capture weak vibration signals.

[0047] Set appropriate data acquisition parameters, including sampling frequency (needs to be reasonably selected according to the site vibration frequency range, generally several times higher than the highest frequency of the vibration signal, such as 10kHz or higher sampling frequency may be selected for high-frequency vibration), collection time (considering the long-term variation characteristics of site vibration, the collection time may last for several hours or even days), etc., to obtain complete and accurate original vibration data.

[0048] In addition, the collected raw vibration data can be sorted to ensure the integrity and accuracy of the data and check whether there are missing data or abnormal values. If the data format is not convenient for subsequent processing, the format can be converted to meet the requirements of the analysis software or algorithm.

[0049] Second, pre-process the acquired data, including:

[0050] 1) Use the sliding average method or the least squares method to remove the trend item in the data. Taking the sliding average method as an example, select an appropriate sliding window length (such as selecting a window length of 10-100 data points based on the frequency of change and trend characteristics of the vibration data), calculate the average value of the data in the window as the correction value of the central data point of the window, and process the entire data sequence in turn to eliminate the slow change trend in the data and solve the zero drift problem. Figure 2 The figure is a diagram showing the effect of eliminating trend items by using the sliding average method in the vibration data processing and evaluation method for large ultra-precision engineering sites according to the embodiment of the present invention. For the least squares method, a suitable polynomial function (such as a first-order or second-order polynomial) is fitted according to the distribution characteristics of the data to approximate the data trend, and then the fitted trend item is subtracted from the original data to obtain the vibration data after removing the trend.

[0051] This method can be applied to processing long-term monitoring data and disturbed data generated by long-term monitoring of large ultra-precision engineering sites, with many measuring points and complex working conditions, so as to turn invalid data into valid data.

[0052] 2) Use the five-point cubic smoothing method or the five-point sliding average method to smooth the data without trend items. The five-point cubic smoothing method uses a specific weighting coefficient to perform weighted average calculation on the data point and its adjacent points, and calculates each point in the data sequence in turn, reducing the burrs and noise components in the data, making the data smoother and closer to the real vibration signal, improving data quality, and reducing the misleading of subsequent analysis caused by noise interference.

[0053] The five-point moving average rule simply calculates the average of each data point and the two points before and after it as the smoothed value of the point, that is, the entire data series is processed in the same way.

[0054] Third, frequency component processing may specifically include filtering methods and frequency domain calculus methods.

[0055] Filtering includes determining the frequency range that needs attention according to engineering requirements and selecting the appropriate filtering method (low-pass, band-pass, high-pass, band-stop). When performing frequency domain low-pass and band-pass filtering, set the cutoff frequency (for example, for low-pass filtering, if you want to filter out high-frequency signals above 100Hz, set the cutoff frequency to 100Hz), convert the time domain data to the frequency domain through Fourier transform, and then filter the signal in the frequency domain to retain or remove the signal components within the specified frequency range. Finally, convert the filtered frequency domain signal back to the time domain through inverse Fourier transform. For frequency domain high-pass and band-stop filtering, also set the cutoff frequency according to the frequency range that needs to be filtered (for example, if the high-pass filter is to filter out low-frequency signals below 5Hz, the cutoff frequency is set to 5Hz), and perform similar frequency domain processing operations to obtain the vibration signal characteristics of specific frequency components.

[0056] In this way, the vibration signal characteristics in the specified frequency range can be obtained, and targeted processing of vibration signals with different frequency components can be achieved, which is helpful for in-depth analysis of the high and low frequency distribution and characteristics of the main vibration components of the site. Figure 3 The figure is a schematic diagram of the effects of low-pass and high-pass filtering in a method for processing and evaluating vibration data of a large ultra-precision engineering site according to an embodiment of the present invention.

[0057] The frequency domain calculus processing is to perform frequency domain calculus operations on the acceleration or velocity signal in the vibration signal (after preprocessing). After the signal is converted to the frequency domain by Fourier transform, according to the frequency domain properties of calculus (such as the relationship between the Fourier transform of the acceleration signal and the Fourier transform of the velocity signal, where is the imaginary unit and is the angular frequency), the frequency domain signal is multiplied or divided to achieve the conversion of acceleration and velocity.

[0058] At the same time, you can specify to eliminate frequency components outside the positive and negative frequency bands. For example, if you only focus on the positive frequency range of 0-100 Hz, set the Fourier coefficients corresponding to the frequency components outside this range to 0, and then use the inverse Fourier transform to obtain the processed time domain signal, thereby realizing fast and effective evaluation of large quantities of data in different physical quantity domains.

[0059] Fourth, perform modal analysis, including modal identification preprocessing and modal parameter identification.

[0060] Modal identification preprocessing can use random reduction method or NExT method to perform modal identification preprocessing on the vibration data that has been processed previously.

[0061] Taking the random reduction method as an example, multiple sub-samples are intercepted from the vibration response signal (e.g., each sub-sample has a length of 1000-5000 data points), the average value of the sub-samples is calculated, and an approximate free decay response signal is obtained through multiple interception and averaging operations. In this process, it is necessary to reasonably intercept the amplitude and the multiple of the standard deviation of the input signal (e.g., select 2-5 times the standard deviation according to the statistical characteristics of the signal) and superimpose the input signal sub-samples to improve the preprocessing effect and provide more accurate data for modal parameter identification. Figure 4 It is a schematic diagram of stripping free vibration signals by random reduction method in a vibration data processing and evaluation method for large ultra-precision engineering sites according to one embodiment of the present invention.

[0062] For the NExT method, by calculating the cross-correlation function between different measuring points and extracting the free decay part in the cross-correlation function, it is also necessary to reasonably set the relevant parameters (such as the time delay range of the cross-correlation function calculation) to obtain effective modal identification preprocessing results. Figure 5 It is a schematic diagram of the cross-correlation function calculated by the NExT method in the vibration data processing and evaluation method for large-scale ultra-precision engineering sites according to one embodiment of the present invention.

[0063] The ITD method can be used for modal parameter identification. First, according to the characteristics and experience of the engineering structure, the modal order is effectively defined (for example, for simple structures, 2-5 orders may be selected, and for complex structures, 10 orders or more may be selected), and then the free vibration response of the structure is solved (by analyzing and calculating the preprocessed signal, and using relevant mathematical models and algorithms to obtain the free vibration response signal), and finally the characteristic equation is solved based on the least squares method to obtain the modal parameters such as the modal frequency, damping ratio and modal vibration shape of the structure, accurately extract the different modal components in the site vibration, and accurately reflect the vibration modal characteristics of the site. Figure 6 It is a schematic diagram of free vibration response fitting of the ITD method in a vibration data processing and evaluation method for large-scale ultra-precision engineering sites according to an embodiment of the present invention.

[0064] Fifth, signal decomposition and reconstruction.

[0065] First, the vibration signal can be decomposed based on wavelet transform. You can select a suitable wavelet basis (such as wavelets in the Daubechies wavelet system, etc., and select a wavelet basis with suitable time-frequency characteristics according to the characteristics of the signal and analysis requirements), determine the scale of wavelet decomposition (such as selecting 3-8 scales, determined according to the signal frequency range and analysis resolution requirements), decompose the vibration signal into detail signals and approximate signals at different scales, obtain the time-frequency domain components of the vibration signal under the main modal components, and deeply analyze the characteristics of the vibration signal in time and frequency.

[0066] During wavelet reconstruction, the wavelet coefficients obtained by decomposition are used to restore the signal according to the wavelet reconstruction algorithm, and then the reconstructed signal is evaluated by FFT (Fast Fourier Transform), and the error (such as mean square error, etc.) in the frequency domain between the reconstructed signal and the original signal (or the reference signal after previous processing) is calculated to verify the accuracy of the reconstruction and ensure the quality of the signal processing process.

[0067] Figure 7-9 A schematic diagram of wavelet decomposition and reconstruction in a method for processing and evaluating vibration data of a large ultra-precision engineering site according to an embodiment of the present invention is shown, specifically including Figure 7 Wavelet decomposition diagram of vibration signal; Figure 8 Spectral analysis diagram of the vibration signal after wavelet decomposition; and Fig. 9 Schematic diagram of wavelet reconstruction of vibration signal.

[0068] Finally, a comprehensive evaluation is performed, including the use of one-third octave spectrum combined with VC curve to evaluate the overall vibration signal and the main modal component distribution signal after multi-band evaluation. First, calculate the one-third octave spectrum of the vibration signal, divide the frequency range into multiple one-third octave bands (such as divided according to international standards), calculate the vibration energy or amplitude and other parameters in each band, analyze the distribution of vibration energy in different frequency bands, and determine whether there is vibration abnormality in a specific frequency band.

[0069] At the same time, according to the vibration requirements of the engineering site and the allowable vibration standards of the equipment, VC curves (such as VC-A, VC-B and other different standard curves) are drawn, and the one-third octave band spectrum of the actual vibration signal is compared with the VC curve to evaluate whether the site vibration meets the requirements of ultra-precision engineering. If it exceeds the standard range, the reasons are further analyzed and improvement measures are proposed to provide a comprehensive and scientific basis for engineering decision-making.

[0070] Embodiments of the present invention may:

[0071] (1) Provide full life cycle vibration data services for the planning, construction and renovation of large ultra-precision engineering sites. Conduct site vibration environment assessment and prediction in the early stage of the project to provide a basis for site selection and layout; monitor vibration data in real time during the construction process to guide construction process improvements and ensure that vibration during the construction process meets requirements; during the project operation phase, continuously monitor and evaluate vibration conditions to promptly identify potential problems and propose solutions.

[0072] (2) Develop customized vibration data processing and evaluation solutions based on the characteristics and needs of different types of ultra-precision engineering (such as chip manufacturing, high-end optical instrument production, precision metrology laboratories, etc.). For chip manufacturing plants, focus on the impact of high-frequency vibration on key processes such as chip lithography, optimize data processing and evaluation methods, and provide targeted vibration control recommendations.

[0073] (3) Expand the application of this technology in emerging fields, such as micro-electromechanical system (MEMS) manufacturing, quantum computing equipment environmental monitoring, etc. As the requirements for environmental vibration in these fields continue to increase, research on how to adapt existing technologies to meet the special needs of new fields and provide technical support for the development of emerging industries.

[0074] The above description of the embodiments is to facilitate the understanding and application of the present invention by those skilled in the art. It is obvious that those skilled in the art can easily make various modifications to these embodiments and apply the general principles described herein to other embodiments without creative work. Therefore, the present invention is not limited to the embodiments herein, and improvements and modifications made by those skilled in the art based on the disclosure of the present invention without departing from the scope of the present invention should be within the scope of protection of the present invention.

Claims

1. A method for processing and evaluating vibration data for large-scale ultra-precision engineering sites, characterized in that: include: Step 1: Conduct vibration test on site to obtain raw data; Step 2: preprocessing the obtained raw data; Step 3: filter the high and low pass frequency signal components by using frequency domain low pass and band pass filtering, frequency domain high pass and band stop filtering, respectively, to obtain the vibration signal characteristics in the specified frequency range; Step 4: Perform frequency domain calculus on the acceleration or velocity signal collected from the vibration signal after preprocessing in step 2 to achieve the conversion between acceleration and velocity, and perform batch analysis and processing; Step 5: Use random reduction method or NExT method to pre-process the vibration data processed in step 4 for modal identification, and then use ITD method to identify modal parameters and extract different modal components in site vibration; Step 6: Based on modal analysis and step 3, the vibration signal is decomposed and reconstructed based on wavelet transform; as well as Step 7: Use the one-third octave spectrum combined with the VC curve to evaluate the overall vibration signal and the main modal component distribution signal after multi-band evaluation.

2. A method for processing and evaluating vibration data for large ultra-precision engineering sites according to claim 1, characterized in that: In step 2, the preprocessing includes removing trend items in the data using a sliding average method or a least squares method.

3. A method for processing and evaluating vibration data for large ultra-precision engineering sites according to claim 2, characterized in that: In step 2, the preprocessing further includes smoothing the data with trend removed by using a five-point cubic smoothing method or a five-point sliding average method.

4. A method for processing and evaluating vibration data for large-scale ultra-precision engineering sites according to claim 1, characterized in that: In step five, the random reduction method includes intercepting multiple sub-samples from the vibration response signal, calculating the average value of the sub-samples, and obtaining an approximate free decay response signal through multiple interception and averaging operations; the NExT method includes calculating the cross-correlation function between different measuring points and extracting the free decay part in the cross-correlation function.

5. According to the method for processing and evaluating vibration data for large ultra-precision engineering sites as described in claim 1, in step 5, using the ITD method to identify modal parameters includes: According to the characteristics and experience of engineering structures, the modal order is defined, and then the free vibration response of the structure is solved. Finally, the characteristic equation is solved based on the least squares method to obtain the modal parameters of the structure, accurately extract the different modal components in the site vibration, and accurately reflect the vibration modal characteristics of the site.

6. A method for processing and evaluating vibration data for large-scale ultra-precision engineering sites according to claim 1, characterized in that: Step six includes selecting a suitable wavelet basis and determining the scale of wavelet decomposition, decomposing the vibration signal into detail signals and approximate signals at different scales, obtaining the time-frequency domain components of the vibration signal under the main modal components, and deeply analyzing the characteristics of the vibration signal in time and frequency; during wavelet reconstruction, the wavelet coefficients obtained by decomposition are used to restore the signal according to the wavelet reconstruction algorithm, and then the reconstructed signal is evaluated by FFT, the error between the reconstructed signal and the original signal in the frequency domain is calculated, and the accuracy of the reconstruction is verified.

7. A method for processing and evaluating vibration data for large-scale ultra-precision engineering sites according to claim 1, characterized in that: Step seven includes: first, calculating the one-third octave band spectrum of the vibration signal, dividing the frequency range into multiple one-third octave bands, calculating parameters such as the vibration energy or amplitude in each band, analyzing the distribution of vibration energy in different frequency bands, and determining whether there is vibration anomaly in a specific frequency band; at the same time, according to the vibration requirements of the engineering site and the vibration standards allowed by the equipment, drawing the VC curve, comparing the one-third octave band spectrum of the actual vibration signal with the VC curve, and evaluating whether the site vibration meets the requirements of ultra-precision engineering.

Images