Vibration frequency analysis method and system for pavement compaction
Through the vibration frequency analysis method, the road roller signals are collected and processed, and the problem of relying on experience in traditional road compaction quality control is solved, real-time monitoring and data visualization are realized, and construction quality and data accuracy are improved.
Patent Information
- Application Number
- CN202411006634.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-25
- Publication Date
- 2024-11-15
AI Technical Summary
Traditional pavement compaction quality control methods rely on experience, resulting in large errors, poor quality standards, and inability to record and trace in real time, which cannot meet the needs of modern mechanized construction.
Through the vibration frequency analysis method, the vibration signal of the roller is collected, the regularized least squares method and SG smoothing algorithm are used to process the signals, and variational mode decomposition and Hilbert transformation are performed to obtain the time-frequency distribution characteristics of the vibration signal, reduce noise interference and Improve data accuracy.
Real-time monitoring and data visualization of compaction quality are realized, errors are reduced, quality control requirements of modern construction are met, and data support is provided with strong adaptability and good noise reduction performance.
Smart Images

Figure CN118960943A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of vibration spectrum analysis, and in particular to a vibration frequency analysis method and system for road compaction. Background Art
[0002] The control of the compaction quality of asphalt mixture is crucial to the quality of pavement materials. According to my country's current "Highway Roadbed Construction Specifications" (JTG3610-2019), compaction quality detection is mainly based on random sampling after construction. However, there are several problems with traditional compaction quality control methods: compaction quality is mainly based on experience, and compaction parameters cannot be visualized in data. Construction under the "empirical control" method may lead to defects such as large errors and substandard quality; field tests with destructive characteristics are time-consuming and labor-intensive, and cannot meet the requirements of modern mechanized construction; the compaction quality of the entire construction process cannot be recorded in real time, and data traceability is extremely poor. Therefore, traditional control methods not only cannot effectively control the compaction quality of the construction area, but also cannot meet the increasingly stringent control requirements for roadbed and pavement construction quality.
[0003] According to the analysis of the "machine-material" coupling model of asphalt mixture vibration compaction, due to the interaction between the vibrating steel wheel and the asphalt mixture during the compaction process, the vibration feedback on the vibrating steel wheel will change dynamically with the change of the compaction state of the asphalt mixture. Therefore, the compaction process can be monitored in real time through the vibration feedback signal, and the relationship between the vibration spectrum and the harmonic method can be studied to obtain the compaction degree. The vibration frequency of the roller mainly affects the compaction quality of the asphalt mixture, and when the vibration frequency is slightly higher than the natural frequency of the asphalt mixture, a better compaction effect can be obtained; the vibration amplitude mainly affects the compaction thickness of the asphalt mixture. When the thickness of the rolling layer is thin, high-frequency, low-amplitude vibration parameters should be selected.
[0004] Intelligent compaction technology obtains parameters such as vibration amplitude and vibration frequency through vibration sensors, and performs signal processing and analysis. Since the vibration compaction of asphalt mixture is a nonlinear and non-stationary process, and the construction environment is complex and the operation mode is changeable, the feedback signal is a nonlinear and non-stationary signal, containing a lot of noise, clutter and other interference. Therefore, the processing and analysis of vibration feedback signals is one of the keys to intelligent compaction technology. Summary of the invention
[0005] In view of the above problems, the present invention provides a vibration frequency analysis method for road compaction.
[0006] To achieve the above object, the present invention provides a vibration frequency analysis method for road compaction, comprising:
[0007] Collect vibration signals of road roller;
[0008] Obtaining all local maximum points and local minimum points of the vibration signal;
[0009] By introducing linear or nonlinear fitting in different domains through regularized least squares method, it is more in line with the actual situation on site. While restoring the smoothness of data, the SG smoothing algorithm reduces the high-frequency noise mixed in the vibration signal and avoids overfitting of data.
[0010] The vibration signal determines the fitness value of each initial component through chaotic mapping;
[0011] Determine whether the fitness value of the initial component is correct;
[0012] If not, it will be updated by the sine algorithm and then iterated;
[0013] If so, you can continue to update and iterate;
[0014] Iterate to find the optimal solution of the fitness value of the initial component, use Gauss-Cauchy mixed mutation perturbation to select the optimal k and α, and then determine k and α and decompose them by VMD;
[0015] The vibration signal is decomposed into a fundamental signal and multiple harmonic signals;
[0016] All the IMF components obtained by decomposing the vibration signal are subjected to spectrum analysis by applying Hilbert transform, and the instantaneous frequency and instantaneous amplitude of each IMF component are respectively calculated to obtain the Hilbert spectrum of the vibration signal.
[0017] Preferably, the vibration signal is collected by a vibration sensor arranged on the road roller.
[0018] Preferably, the collected vibration signal is preprocessed, and the preprocessing includes filtering and removing trend items.
[0019] Preferably, the filtering process is performed through an FIR filter, and the FIR filter parameter settings include: the passband lower cutoff frequency and upper cutoff frequency are 30 Hz and 300 Hz respectively, the amplitude attenuation parameter at the passband cutoff frequency is fixed to 6 db, the sampling frequency is 1024 Hz, and the filter order is 29.
[0020] Preferably, the vibration signal trend term removal is performed using a regularized least squares method.
[0021] It is preferred to restore the smoothness of the data while reducing the high-frequency noise mixed into the vibration signal:
[0022] The vibration signal is subjected to SG sliding average processing.
[0023] Preferably, determining whether the sub-signal is an IMF component includes that the sub-signal needs to simultaneously satisfy the following conditions: within the entire time range, the number of local extreme points and zero-crossing points of the sub-signal must be equal or differ by at most one, and at any time point, the mean of the upper envelope of the local maximum value and the lower envelope of the local minimum value of the sub-signal is zero.
[0024] Preferably, assuming that the sub-signal function is c i (t), then after Hilbert transform we get:
[0025]
[0026] Where: t is time; is the integral component.
[0027] Preferably, the formula obtained by analyzing the Hilbert transform is used to obtain the center frequency of the IMF component, and the center frequency of the IMF component is:
[0028]
[0029] The Hilbert spectrum of the vibration signal is:
[0030]
[0031] Where: Re is the real part, a k (t) is the kth instantaneous amplitude, ω k (t) is the kth instantaneous frequency.
[0032] The present invention also provides a vibration spectrum analysis system for intelligent compaction technology, comprising:
[0033] A collection module, used to collect vibration signals of the roller;
[0034] An acquisition module, used for acquiring all local maximum points and local minimum points of the vibration signal;
[0035] The calculation module introduces linear or nonlinear fitting in different domains through the regularized least squares method, which is more in line with the actual situation on site. The SG smoothing algorithm restores the smoothness of the data while reducing the high-frequency noise mixed in the vibration signal and avoiding overfitting of the data.
[0036] The iteration module determines whether the initial value is correct. If it is correct, it will continue to update and iterate. If the initial value is incorrect, it will be updated by the sine algorithm and then iterated. Iterate to get the optimal solution of the fitness value of the initial component, use the Gauss-Cauchy mixed mutation perturbation to select the optimal k and α, and then determine k and α and decompose them by VMD.
[0037] A decomposition module, used for performing variational mode decomposition of the vibration signal to obtain sub-signals;
[0038] The analysis module is used to apply Hilbert transform to perform spectrum analysis on all the IMF components obtained by decomposing the vibration signal, calculate the center frequency of each IMF component respectively, and obtain the Hilbert spectrum of the vibration signal.
[0039] Compared with the prior art, the present invention has the following beneficial effects:
[0040] The present invention processes vibration signals by optimizing k and α in the variational mode decomposition method, and performs Hilbert transform on the IMF components obtained by decomposition, thereby obtaining the time-frequency distribution characteristics of the vibration signals and reducing errors such as spectrum leakage and fence effect. It has the advantages of strong adaptability and good noise reduction performance, and provides good data support for subsequent compaction state perception and intelligent compaction quality evaluation. BRIEF DESCRIPTION OF THE DRAWINGS
[0041] Figure 1 It is a flow chart of the vibration frequency analysis method for road surface compaction of the present invention. DETAILED DESCRIPTION
[0042] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0043] Reference Figure 1 The present invention provides a vibration frequency analysis method for road compaction, comprising:
[0044] Collect vibration signals of road roller;
[0045] Obtain all local maximum and local minimum points of the vibration signal;
[0046] By introducing linear or nonlinear fitting in different domains through regularized least squares method, it is more in line with the actual situation on site. While restoring the smoothness of data, the SG smoothing algorithm reduces the high-frequency noise mixed in the vibration signal and avoids overfitting of data.
[0047] The vibration signal determines the fitness value of each initial component through chaotic mapping;
[0048] Determine whether the fitness value of the initial component is correct;
[0049] If not, it will be updated by the sine algorithm and then iterated;
[0050] If so, you can continue to update and iterate;
[0051] Iterate to find the optimal solution of the fitness value of the initial component, use Gauss-Cauchy mixed mutation perturbation to select the optimal k and α, and then determine k and α and decompose them by VMD;
[0052] The vibration signal is decomposed into a fundamental signal and multiple harmonic signals;
[0053] Hilbert transform is applied to all IMF components obtained by decomposing the vibration signal for spectrum analysis, and the instantaneous frequency and instantaneous amplitude of each IMF component are calculated respectively to obtain the Hilbert spectrum of the vibration signal.
[0054] In this embodiment, the vibration signal is collected by a vibration sensor arranged on the roller. The collected vibration spectrum is not only related to the vibration transmitted at the installation position, but also to the structure where the installation position is located. Therefore, it is very important to correctly select the installation position. The principle of selecting the installation position of the vibration sensor is the shortest vibration transmission path, the maximum stiffness, and the rigid connection. The collection position must be the same each time to minimize the influence of different positions on the collection results. Based on the above principles, when the present application conducts an outdoor vibration compaction test, the vibration sensor is installed on the transmission shaft of the roller steel wheel, directly above the vibration eccentric structure. At this time, the acquisition path of the vibration signal is shorter and the signal is more balanced, thereby minimizing the impact on the test effect. The operating frequency of the roller is about 50 / 67Hz, so the fundamental frequency of the vibration compaction feedback signal should be about 40Hz-100Hz, and the corresponding harmonic multiple frequency should be about 80Hz, about 120Hz, and so on. Taking into account the frequency range of the vibration compaction feedback signal and the measurement principle of the acceleration sensor, the present application selects acceleration as the parameter for vibration compaction measurement. The factors that affect the vibration signal acquisition process mainly include sampling frequency and accuracy.
[0055] Specifically, the sampling frequency refers to the number of sampled signals obtained per unit time, indicating the sampling speed. During the vibration signal measurement process, the setting of the sampling frequency must comply with the sampling theorem requirements and be limited by the highest signal frequency of the sampled object. According to Shannon's sampling theorem, during the A / D signal conversion process, in order to restore the vibration signal as losslessly as possible, the sampling frequency should be greater than twice the highest frequency in the analog signal, that is:
[0056] f s ≥2f m ;
[0057] Where: f s is the sampling frequency, f m is the highest frequency in the original signal. Generally speaking, fs The larger the value, the denser the sampling points, and the closer the obtained digital signal is to the original signal. s The larger it is, the higher the requirements for the sampling system, and the conversion circuit must have a faster conversion speed. In order to take into account the requirements of memory storage and computing workload, it is generally sufficient to ensure that the signal is not lost or the vibration signal information is not distorted to meet actual needs. The sampling theorem is the basis for achieving correct sampling. In practical applications, the sampling frequency is generally taken as 8 times or more of the highest frequency of the signal. When determining the sampling frequency, it is necessary to correctly estimate the highest frequency in the original signal. According to the operating frequency range and harmonic frequency multiples of the vibratory roller, the sampling frequency set in this application is 1024Hz.
[0058] In this embodiment, the present application also includes preprocessing the collected vibration signal, and the preprocessing includes filtering and removing trend items.
[0059] The filtering process is performed by using an FIR filter.
[0060] Specifically, the filtering process uses a digital filter to process the quantized digital signal sequence, thereby changing the relative proportion of the frequency components in the input signal and effectively removing noise and clutter. Compared with analog filters, digital filters have flexible parameter design, high precision, good stability, and can well optimize the characteristics of in-band fluctuations, out-of-band suppression, transition band, etc., and can be designed into a linear phase. Therefore, in practical applications, digital filters are usually selected. According to the time domain characteristics of the unit impulse response function, the classic digital filter can be divided into two categories: infinite impulse response (IIR) filter and finite impulse response (FIR) filter. In the vibration compaction detection of asphalt pavement, clear and accurate vibration feedback is required. The filtering process of the vibration signal removes the interference without affecting the normal signal waveform, which requires the filter to have good amplitude-frequency characteristics and strict linear phase characteristics. The IIR filter must adopt a recursive structure, and the pole position must be within the unit circle, otherwise the system will be unstable. In addition, in this structure, due to the rounding of the sequence during the calculation process, this finite word length effect sometimes introduces parasitic oscillations. On the contrary, FIR filters mainly use non-recursive structures. There is no stability problem in theory or in actual finite precision operations, and the operation error is also small. In addition, FIR filters can use fast Fourier transform algorithms, and the operation speed can be much faster under the same order. The essence of FIR filter design is to find a physically realizable system function H(z) so that its frequency response H(e jω) approximates the ideal frequency response required by the filter. The methods for designing FIR filters include window function method, frequency sampling method and equiripple approximation method, among which the window function method is simple and practical, and is easy to implement on a digital signal processor. Therefore, the present application adopts the window function method to design the FIR digital filter. Since the vibration compaction feedback signal contains strong clutter and noise signals, and the processing result requires a high amplitude accuracy, a window function with a main lobe as narrow as possible and a small side lobe amplitude should be selected. The above requirements are mainly met by the Hanning window and the Hamming window. The present application designs a FIR bandpass filter based on the Hanning window, and the filter parameters are set as follows: the lower limit cutoff frequency and the upper limit cutoff frequency of the passband are 30Hz and 300Hz respectively, the amplitude attenuation parameter at the passband cutoff frequency is fixed to 6db, the sampling frequency is 1024Hz, and the filter order is 29. After the filter design is completed, in order to verify whether its performance indicators meet the requirements, the filterDesigner command in the Matlab software can be called to call out the toolbox to calculate the frequency domain characteristics of the digital filter. If the requirements are not met, the window function type or length can be adjusted according to the specific situation until the requirements are met.
[0061] Furthermore, the trend term is a long-term cumulative system error caused by certain factors in signal acquisition, such as environmental interference of the sensor, and is mainly reflected in the integration link. In vibration measurement, the integral of acceleration over time is velocity, and the integral of velocity over time is displacement, so trend terms are often encountered in vibration test systems. Trend terms can cause the original data to be offset, directly affecting the accuracy of the signal, and thus affecting the accuracy of data processing. In order to more accurately understand the spectral characteristics of the vibration signal, it is necessary to remove the trend term in advance before signal processing. Since the elimination of trend terms is mainly for low-frequency components in the signal, the order of the polynomial is usually relatively small. This application selects the regularized least squares method to eliminate trend terms, and introduces linear or nonlinear fitting in different domains, which is more in line with the actual situation on site and avoids overfitting of the data.
[0062] Specifically, the trend term is removed using the regularized least squares method;
[0063] Then, the data collected by the acceleration sensor, due to the interference of external irregular factors, the final curve time domain contains a lot of burrs. In order to improve the smoothness of the real vibration curve, it is necessary to perform SG sliding average processing on the sampled data, thereby reducing the impact of interference signals on the real data. The present invention uses SG sliding average processing, the algorithm is simple, and the effect is good. The calculation formula is:
[0064] Y (2m+1) =X (2m+1)*k ·A k +E (2m+1)
[0065] In the formula, Y(2m+1) is the 2m+1th weighted average, X(2m+1) is the actual value of the kth acquisition point, and k is the moving span period; the advantage of the SG sliding average method is that different window widths can be arbitrarily selected at any position on the same curve to meet the needs of different smoothing filters; especially when processing time series data, it has obvious advantages for sequence processing at different stages. It works well for processing non-periodic and nonlinear noise samples.
[0066] In this example, the collected vibration signal is subjected to improved variational modal decomposition, and the steps include:
[0067] Step 1: Determine all local maximum and local minimum points of the vibration signal.
[0068] Step 2 introduces linear or nonlinear fitting in different domains through regularized least squares method, which is more in line with the actual situation on site. The SG smoothing algorithm restores the smoothness of the data while reducing the high-frequency noise mixed in the vibration signal and avoiding overfitting of the data.
[0069] Step 3: The vibration signal is used to determine the fitness value of each initial component through chaotic mapping; whether the fitness value of the initial component is correct is determined; if not, it is updated and iterated after being updated by the sine algorithm; if so, the update and iteration can be continued.
[0070] Step 4: Iterate to find the optimal solution of the fitness value of the initial component, use Gauss-Cauchy mixed variation perturbation to select the optimal k and α, determine k and α, and then decompose by VMD. At this time, k IMF components are obtained, which usually include the fundamental component and its harmonic components.
[0071] Step 5: Separate c1(t) from x(t) to obtain a residual component r1(t) without high-frequency components.
[0072] r1(t)=x(t)-c1(t);
[0073] The residual component r1(t) contains the portion of the signal that has not been decomposed into its components.
[0074] Step 6: r1(t) is a monotone signal or has only one pole. At this time, r1(t) can be regarded as the component of the original signal that is not removed by the clutter, and the improved VMD decomposition is completed. Finally, the vibration signal x(t) can be expressed as follows after decomposition:
[0075]
[0076] In the formula, c i(t) is each IMF component, and r1(t) is the residual term. From the decomposition process, it can be seen that the improved VMD decomposition method uses iterative search for the optimal solution of the variational model to determine the center frequency and bandwidth of each decomposed component. It is a completely non-recursive model. The model searches for a set of modal components and their respective center frequencies, and each mode is smooth after demodulation to baseband. The experimental results show that this method is more robust in terms of sampling and noise.
[0077] The essence of the improved VMD is to decompose the original signal to obtain different harmonic components, thereby decomposing the fluctuation or trend of the signal step by step.
[0078] To determine whether a sub-signal is an IMF component, the sub-signal must simultaneously satisfy the following conditions: within the entire time range, the number of local extreme points and zero-crossing points of the sub-signal must be equal or differ by at most one, and at any time point, the mean of the upper envelope of the local maximum value and the lower envelope of the local minimum value of the sub-signal is zero.
[0079] In this embodiment, after obtaining each IMF component of the signal x(t) by using the improved variational mode decomposition, Hilbert transform (HTT) can be applied to each IMF component for spectral analysis to calculate the instantaneous frequency and instantaneous amplitude of each IMF component to obtain the Hilbert spectrum of the signal.
[0080] For each function c i (t) is transformed by Hilbert to obtain:
[0081]
[0082] Where t represents time, Represents the integral component. Construct analytical signal
[0083]
[0084] And the Hilbert spectrum of the signal x(t) is obtained as:
[0085]
[0086] The residual term r1(t) is omitted here, Re represents the real part, and a i (t) represents the i-th instantaneous amplitude, ω i (t) represents the i-th instantaneous frequency. H(ω,t) is the distribution of instantaneous amplitude in the frequency-time plane, which accurately describes the variation of signal amplitude with time and frequency in all frequency bands. The HHT transform ensures that the frequency of each harmonic component is obtained after the Hilbert transform of each IMF component by improving the VMD decomposition.
[0087] After the vibration feedback signal is decomposed by improved VMD, the obtained IMF components are the components that make up the original signal. At the same time, improved VMD is an adaptive, completely non-recursive modal variation and signal processing method. This technology has the advantage of being able to determine the number of modal decompositions. Its adaptability is reflected in the fact that according to the actual situation, the number of modal decompositions of the given sequence is determined through chaotic mapping and Gaussian-Cauchy mixed variation perturbation. In the subsequent search and solution process, the center frequency and limited bandwidth of each mode can be adaptively matched, and the effective separation of the intrinsic modal component (IMF) and the frequency domain division of the signal can be achieved, and then the effective decomposition components of the given signal can be obtained, and finally the optimal solution to the variation problem is obtained. It overcomes the problems of endpoint effect and modal component aliasing in the EMD method, and has a more solid mathematical theoretical foundation. It can reduce the non-stationarity of time series with high complexity and strong nonlinearity, and decompose to obtain relatively stable subsequences containing multiple different frequency scales. It is suitable for non-stationary sequences. The core idea of improved VMD is to construct and solve variational problems.
[0088] The present invention designs a FIR digital bandpass filter based on the Hanning window to address the large amount of noise and clutter interference contained in the original signal; the filtering effect shows that the FIR filter can effectively filter out high-frequency noise components, and its linear phase characteristics ensure that the vibration signal after filtering will not be distorted. The SG sliding average method is used for smoothing; the vibration signal preprocessing can effectively suppress clutter interference and retain the original waveform trend of the original signal without over-processing the original signal, laying a good foundation for the next step of vibration signal processing and analysis. The Hilbert-Huang transform (HHT) is a commonly used time-frequency analysis method. It separates the effective signal from the clutter to be eliminated through the variational mode decomposition method, and has unique advantages in nonlinear and non-stationary signal processing. Therefore, the present invention processes the vibration compaction feedback signal through the variational mode decomposition method. In view of the modal aliasing problem of the HHT transform, the vibration feedback signal is decomposed into a group of IMF components based on the variational mode decomposition method; the IMF components obtained by decomposition are Hilbert transformed respectively to obtain the time-frequency distribution characteristics of the vibration signal; according to the excitation frequency of the roller, the fundamental component in the IMF component is found. Variational mode decomposition method is an adaptive, completely non-recursive mode variation and signal processing method, which overcomes the problems of endpoint effect and modal component aliasing in the EMD method, and has a more solid mathematical theoretical basis, can reduce the non-stationarity of time series with high complexity and strong nonlinearity, and decompose to obtain relatively stable subsequences containing multiple different frequency scales, which is suitable for non-stationary sequences. The research of this invention provides good data support for subsequent compaction state perception and intelligent compaction quality evaluation.
[0089] The present invention also provides a vibration spectrum analysis system for intelligent compaction technology, comprising:
[0090] A collection module, used to collect vibration signals of the roller;
[0091] An acquisition module, used for acquiring all local maximum points and local minimum points of the vibration signal;
[0092] The calculation module introduces linear or nonlinear fitting in different domains through the regularized least squares method, which is more in line with the actual situation on site. The SG smoothing algorithm restores the smoothness of the data while reducing the high-frequency noise mixed in the vibration signal and avoiding overfitting of the data;
[0093] The iteration module determines whether the initial value is correct. If it is correct, it will continue to update and iterate. If the initial value is incorrect, it will be updated by the sine algorithm and then iterated. Iterate to get the optimal solution of the fitness value of the initial component, use the Gauss-Cauchy mixed mutation perturbation to select the optimal k and α, and then determine k and α and decompose them by VMD.
[0094] A decomposition module, used for performing variational mode decomposition of the vibration signal to obtain sub-signals;
[0095] The analysis module is used to apply Hilbert transform to perform spectrum analysis on all the IMF components obtained by decomposing the vibration signal, calculate the center frequency of each IMF component respectively, and obtain the Hilbert spectrum of the vibration signal.
[0096] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. A vibration frequency analysis method for road compaction, characterized in that: include: Collect vibration signals of road roller; Obtaining all local maximum points and local minimum points of the vibration signal; By introducing linear or nonlinearity in different domains through regularized least squares method, it is more in line with the actual situation on site. While restoring the smoothness of data, the SG smoothing algorithm reduces the high-frequency noise mixed in the vibration signal and avoids overfitting of data. The vibration signal determines the fitness value of each initial component through chaotic mapping; Determine whether the fitness value of the initial component is correct; If not, it will be updated by the sine algorithm and then iterated; If so, you can continue to update and iterate; Iterate to find the optimal solution of the fitness value of the initial component, use Gauss-Cauchy mixed mutation perturbation to select the optimal k and α, and then determine k and α and decompose them by VMD; All the IMF components obtained by decomposing the vibration signal are subjected to spectrum analysis by applying Hilbert transform, and the instantaneous frequency and instantaneous amplitude of each IMF component are respectively calculated to obtain the Hilbert spectrum of the vibration signal.
2. The vibration frequency analysis method for road compaction according to claim 1, characterized in that: The vibration signal is collected by a vibration sensor arranged on the road roller.
3. The vibration frequency analysis method for road compaction according to claim 2, characterized in that: The collected vibration signal is preprocessed, and the preprocessing includes filtering and removing trend items.
4. The vibration frequency analysis method for road compaction according to claim 3, characterized in that: The filtering process is performed by an FIR filter, and the FIR filter parameter settings include: the passband lower limit cutoff frequency and the upper limit cutoff frequency are 30 Hz and 300 Hz respectively, the amplitude attenuation parameter at the passband cutoff frequency is fixed to 6 db, the sampling frequency is 1024 Hz, and the filter order is 29.
5. The vibration frequency analysis method for road compaction according to claim 4, characterized in that: The vibration signal trend term removal method uses a regularized least square method to remove the trend term.
6. The vibration frequency analysis method for road compaction according to claim 5, characterized in that: While restoring the smoothness of the data, it also reduces the high-frequency noise mixed into the vibration signal.
7. The vibration frequency analysis method for road compaction according to claim 6, characterized in that: Determine whether the sub-signal is a component of the original signal, including that the sub-signal must simultaneously satisfy the following conditions: within the entire time range, the number of local extreme points and zero-crossing points of the sub-signal must be equal or differ by at most one, and at any time point, the mean of the upper envelope of the local maximum value and the lower envelope of the local minimum value of the sub-signal is zero.
8. The vibration frequency analysis method for road compaction according to claim 7, characterized in that: Assume that the sub-signal function is c i (t), then after Hilbert transform we get: Where: t is time; is the integral component.
9. The vibration frequency analysis method for road compaction according to claim 8, characterized in that: The formula obtained by analyzing the Hilbert transform is used to obtain the center frequency of the IMF component. The center frequency of the IMF component is: The Hilbert spectrum of the vibration signal is: Where: Re is the real part, a k (t) is the kth instantaneous amplitude, ω k (t) is the kth instantaneous frequency.
10. A vibration spectrum analysis system for intelligent compaction technology, characterized in that: include: A collection module, used to collect vibration signals of the roller; An acquisition module, used for acquiring all local maximum points and local minimum points of the vibration signal; The calculation module introduces linear or nonlinear fitting in different domains through the regularized least squares method, which is more in line with the actual situation on site. The SG smoothing algorithm restores the smoothness of the data while reducing the high-frequency noise mixed in the vibration signal and avoiding overfitting of the data; The iteration module determines whether the initial value is correct. If it is correct, it will continue to update and iterate. If the initial value is incorrect, it will be updated by the sine algorithm and then iterated. Iterate to get the optimal solution of the fitness value of the initial component, use the Gauss-Cauchy mixed mutation perturbation to select the optimal k and α, and then determine k and α and decompose them by VMD. A decomposition module, used for performing variational mode decomposition of the vibration signal to obtain sub-signals; The analysis module is used to apply Hilbert transform to perform spectrum analysis on all the IMF components obtained by decomposing the vibration signal, calculate the center frequency of each IMF component respectively, and obtain the Hilbert spectrum of the vibration signal.
Citation Information
Cited By
Road condition quality detection method and system for highway engineering supervision
CN119964390A
Vibration signal coordinate offset correction method, system, equipment and medium
CN120252943A
Intersection water area alluvial site earth surface ultra-low frequency micro-vibration time-varying effect prediction method
CN121165163A
Highway subgrade compaction quality evaluation method and system based on mutual energy feedback
CN121544130A
