Method and system for detecting uniformity of subgrade compaction

Abstract

The application discloses a kind of evenness of subgrade compaction detection method and system, it is related to subgrade compaction technical field, can solve the problem of inaccurate compaction degree detection result in the process of evaluating subgrade compaction state at present stage, evenness determination reliability is low, comprising: obtaining acceleration signal generated in the process of multiple rolling subgrade by vibratory roller;For each acceleration signal corresponding to subgrade area, determine the mechanical vibration factor and impact factor of acceleration signal;According to the frequency spectrum data of acceleration signal corresponding to the same subgrade area, mechanical vibration factor and impact factor, determine the subgrade response factor of multiple frequency components in the acceleration signal corresponding to subgrade area;According to subgrade response factor, the acceleration signal corresponding to the latest rolling time of subgrade area is adaptively filtered and denoised, according to the acceleration signal after denoising, the compaction meter value of subgrade area is calculated, and according to the compaction meter value of all subgrade areas, the evenness of subgrade compaction is determined.

Description

Technical Field

[0001] This invention relates to the field of roadbed compaction technology, and specifically to a method and system for detecting the uniformity of roadbed compaction. Background Technology

[0002] In road construction, subgrade compaction is a crucial process for ensuring the stability and long-term durability of the pavement structure. Insufficient or uneven compaction leads to a decrease in the subgrade's bearing capacity, resulting in uneven pavement settlement, rutting, and cracking, severely impacting road lifespan and traffic safety. Currently, the industry commonly uses vibratory rollers for compaction, collecting vibration signals through acceleration sensors built into the rollers to calculate the Compaction Meter Value (CMV), thus indirectly assessing compaction quality. However, existing CMV-based detection methods are susceptible to various noise interferences in complex construction environments, primarily including: unsteady mechanical vibrations generated by the roller's engine and hydraulic system that do not change with the subgrade condition; and impact interference caused by localized unevenness on the subgrade surface. Consequently, these methods fail to accurately reflect the true compaction state of the subgrade, resulting in poor overall assessment of compaction uniformity and unreliable guidance for on-site compaction operations. Summary of the Invention

[0003] To address the technical problems of inaccurate compaction degree detection and low reliability of uniformity assessment in the current process of assessing roadbed compaction status due to the mixing of mechanical vibration and impact noise in the acceleration signal of vibratory rollers, the present invention aims to provide a method and system for detecting the uniformity of roadbed compaction. The specific technical solution adopted is as follows: In a first aspect, the present invention provides a method for detecting the uniformity of roadbed compaction, comprising: acquiring acceleration signals generated by a vibratory roller during multiple compaction cycles of the roadbed; wherein the roadbed is divided into multiple roadbed regions; determining a mechanical vibration factor and an impact factor for the acceleration signal corresponding to each roadbed region; wherein the mechanical vibration factor is used to characterize the noise component in the acceleration signal caused by engine vibration during multiple compaction cycles, and the impact factor is used to characterize the noise component in the acceleration signal caused by unevenness of the roadbed surface during multiple compaction cycles; determining roadbed response factors for multiple frequency components in the acceleration signal corresponding to the same roadbed region based on the spectral data, mechanical vibration factor, and impact factor of the acceleration signal corresponding to the same roadbed region; wherein the roadbed response factors are used to characterize the degree to which the frequency components are affected by the roadbed compaction state; performing adaptive filtering and denoising on the acceleration signal corresponding to the latest compaction cycle of the roadbed region based on the roadbed response factors to obtain a denoised acceleration signal; calculating the compaction meter value of the roadbed region based on the denoised acceleration signal, and determining the uniformity of roadbed compaction based on the compaction meter values ​​of all roadbed regions.

[0004] In one possible implementation, determining the mechanical vibration factor of the acceleration signal specifically includes: obtaining the basic mechanical vibration frequency of the vibratory roller and determining the target frequency range based on the basic mechanical vibration frequency; performing a spectral transformation on the acceleration signal to obtain spectral data; and determining the mechanical vibration factor based on the frequency amplitude within the target frequency range and the full-band amplitude in the spectral data.

[0005] In one possible implementation, determining the impact factor specifically includes: clustering the data points in the acceleration signal according to their amplitude to obtain two data clusters; and determining the impact factor based on the absolute value of the difference between the mean amplitudes of the two data clusters.

[0006] In one possible implementation, the roadbed response factors of multiple frequency components in the acceleration signal corresponding to the same roadbed area are determined based on the spectral data, mechanical vibration factor, and impact factor of the acceleration signal corresponding to the same roadbed area. Specifically, this includes: for each frequency component, determining the temporal roadbed response factor of the frequency component based on the amplitude sequence of the frequency component in the spectral data corresponding to the same roadbed area, combined with the mechanical vibration factor and impact factor corresponding to the roadbed area; determining the spatial roadbed response factor of the frequency component based on the spectral data of multiple roadbed areas corresponding to the roadbed area and a preset local range surrounding the roadbed area; and determining the roadbed response factor of the frequency component based on the temporal and spatial roadbed response factors.

[0007] In one possible implementation, for each frequency component, the time-based subgrade response factor is determined based on the amplitude sequence of the frequency component in the spectral data corresponding to the same subgrade area, combined with the mechanical vibration factor and impact factor corresponding to the subgrade area. Specifically, this includes: determining a symbolic parameter based on the amplitude sequence; wherein the symbolic parameter is used to characterize the rate of change of compaction trend; determining weighting coefficients for weighted calculation based on the mechanical vibration factor and impact factor; and performing weighted calculation based on the symbolic parameter and weighting coefficients to determine the time-based subgrade response factor.

[0008] In one possible implementation, the spatial subgrade response factor of the frequency component is determined based on the spectral data corresponding to the subgrade region and the spectral data corresponding to other subgrade regions adjacent to the subgrade region. Specifically, this includes: constructing a three-dimensional data space based on the spatial coordinates and compaction times of each subgrade region; wherein each data point in the three-dimensional data space corresponds to the spectral data of a subgrade region under one compaction; for each frequency component, determining the gradient vector of each subgrade region under the frequency component in the three-dimensional data space; and determining the spatial subgrade response factor of the frequency component based on the overall directional consistency of all gradient vectors, the dispersion of the magnitude of each gradient vector, and the temporal subgrade response factor of each subgrade region under the frequency component.

[0009] In one possible implementation, adaptive filtering and denoising are performed on the acceleration signal corresponding to the latest compaction cycle in the subgrade area based on the subgrade response factor. Specifically, this includes: determining the filtering coefficients for each frequency component in the acceleration signal corresponding to the latest compaction cycle based on the subgrade response factor of the frequency component; and performing adaptive filtering of the frequency component in the wavelet transform domain based on the filtering coefficients.

[0010] In one possible implementation, the uniformity of subgrade compaction is determined based on the compaction values ​​of all subgrade areas. Specifically, this includes: analyzing the spatial distribution of compaction values ​​of all subgrade areas, identifying low-compaction areas where the compaction values ​​are lower than a preset compaction threshold; and determining the uniformity of subgrade compaction based on the number of low-compaction areas, their area proportion, and the statistical dispersion index of compaction values ​​of all subgrade areas.

[0011] In one possible implementation, acquiring the acceleration signal generated by the vibratory roller during multiple compaction processes of the roadbed specifically includes: dividing the roadbed into multiple continuous roadbed regions; wherein the width of each roadbed region is consistent with the width of the vibratory roller's roller, and the length of each roadbed region is a preset value; and acquiring the acceleration signal generated by each roadbed region during each compaction process through an acceleration sensor installed inside the vibratory roller's roller, as the vibratory roller compacts the roadbed multiple times at a fixed speed.

[0012] Secondly, the present invention provides a roadbed compaction uniformity detection system, comprising: a signal acquisition module, a noise factor evaluation module, a response factor evaluation module, a signal denoising module, and a uniformity analysis module; the signal acquisition module is used to acquire acceleration signals generated by a vibratory roller during multiple roadbed compaction processes; wherein the roadbed is divided into multiple roadbed regions; the noise factor evaluation module is used to determine the mechanical vibration factor and impact factor of the acceleration signal for each roadbed region; wherein the mechanical vibration factor is used to characterize the noise component in the acceleration signal caused by engine vibration during multiple compaction processes, and the impact factor is used to characterize the noise component in the acceleration signal caused by uneven roadbed surface during multiple compaction processes. The system includes: a noise component caused by compaction; a response factor evaluation module, used to determine the subgrade response factor of multiple frequency components in the acceleration signal corresponding to the same subgrade area based on the spectral data, mechanical vibration factor, and impact factor of the acceleration signal corresponding to the same subgrade area; wherein, the subgrade response factor is used to characterize the degree to which the frequency components are affected by the subgrade compaction state; a signal denoising module, used to adaptively filter and denoise the acceleration signal corresponding to the latest compaction cycle in the subgrade area based on the subgrade response factor, to obtain the denoised acceleration signal; and a uniformity analysis module, used to calculate the compaction gauge value of the subgrade area based on the denoised acceleration signal, and to determine the uniformity of subgrade compaction based on the compaction gauge values ​​of all subgrade areas.

[0013] Thirdly, the present invention provides an electronic device, comprising: a processor and a memory; wherein the memory is used to store one or more programs, the one or more programs including computer-executable instructions, and when the electronic device is running, the processor executes the computer-executable instructions stored in the memory to cause the electronic device to perform the method for detecting the uniformity of roadbed compaction as described in the first aspect and any possible implementation thereof.

[0014] Fourthly, the present invention provides a computer-readable storage medium storing one or more programs, the one or more programs including instructions that, when executed by an electronic device of the present invention, cause the electronic device to perform a method for detecting the uniformity of subgrade compaction as described in the first aspect and any possible implementation thereof.

[0015] Fifthly, the present invention provides a computer program product containing instructions that, when executed on a computer, cause the electronic device of the present invention to perform the method for detecting the uniformity of subgrade compaction as described in the first aspect and any possible implementation thereof.

[0016] In a sixth aspect, the present invention provides a chip system applied to a device for detecting the uniformity of roadbed compaction; the chip system includes one or more interface circuits and one or more processors. The interface circuits and the processors are interconnected via wiring; the interface circuits are used to receive signals from the memory of the device for detecting the uniformity of roadbed compaction and to send the signals to the processors, the signals including computer instructions stored in the memory. When the processor executes the computer instructions, the device for detecting the uniformity of roadbed compaction performs the method for detecting the uniformity of roadbed compaction as described in the first aspect and any possible design thereof.

[0017] The present invention has the following beneficial effects: by systematically identifying and separating mechanical vibration noise and impact noise in acceleration signals, constructing a roadbed response factor that integrates spatiotemporal characteristics, and performing adaptive filtering accordingly, the measurement accuracy of compaction values ​​is significantly improved, thereby achieving accurate and reliable assessment of the uniformity of roadbed compaction, effectively overcoming the problem of distortion of detection results caused by noise interference in complex construction environments by traditional methods. Attached Figure Description

[0018] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram of the architecture of a roadbed compaction uniformity detection system provided in one embodiment of the present invention; Figure 2 This is a schematic diagram of the architecture of a noise factor evaluation module provided in one embodiment of the present invention; Figure 3 This is a schematic diagram of the architecture of a response factor evaluation module provided in one embodiment of the present invention; Figure 4 This is a schematic flowchart of a method for detecting the uniformity of roadbed compaction according to an embodiment of the present invention. Figure 5 This is a schematic flowchart of another method for detecting the uniformity of roadbed compaction provided in one embodiment of the present invention; Figure 6 This is a schematic flowchart of another method for detecting the uniformity of roadbed compaction provided in an embodiment of the present invention. Detailed Implementation

[0020] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the specific implementation methods, structures, features, and effects of the present invention will be described in detail below with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0022] The following description, in conjunction with the accompanying drawings, details the specific scheme of the method and system for detecting the uniformity of roadbed compaction provided by the present invention.

[0023] For example, such as Figure 1 The diagram shown illustrates the architecture of a roadbed compaction uniformity detection system (hereinafter referred to as the uniformity detection system) according to an embodiment of the present invention. The uniformity detection system 10 includes: a signal acquisition module 11, a noise factor evaluation module 12, a response factor evaluation module 13, a signal denoising module 14, and a uniformity analysis module 15. The modules are described below in sequence: (1) Signal acquisition module 11.

[0024] The signal acquisition module 11 is responsible for collecting acceleration signals during the compaction process of the vibratory roller, providing basic data support for the analysis of all subsequent modules.

[0025] Optionally, the signal acquisition module 11 is used to acquire the acceleration signal generated by the vibratory roller during multiple compaction processes of the roadbed. The roadbed is divided into multiple roadbed regions.

[0026] Specifically, the signal acquisition module 11 first divides the roadbed area: according to the engineering detection accuracy requirements, the complete roadbed to be detected is divided into multiple continuous roadbed areas (the width of each roadbed area is consistent with the width of the vibratory roller wheel, and the length is preset to 1 meter to ensure that the detection range of each area matches the coverage range of a single compaction by the roller). Secondly, the signal acquisition module 11 collects acceleration signals: by using the acceleration sensor pre-installed inside the roller of the vibratory roller, the roller is controlled to compact the roadbed multiple times at a fixed speed (to avoid speed changes causing additional interference to the vibration response), and the acceleration signals generated in each roadbed area during each compaction process are collected simultaneously (the signals include the combined information of the roller's own vibration and the feedback from the roadbed material). Finally, the signal acquisition module 11 performs data preprocessing and storage: it performs preliminary sorting of the collected acceleration signals (such as classifying and storing them according to "roadbed area number - number of compaction times" and removing obvious hardware fault signals).

[0027] The output of the signal acquisition module 11 is a classified acceleration signal, which will be directly transmitted to the noise factor evaluation module 12 and the response factor evaluation module 13, and serves as the data basis for all subsequent noise analysis and response factor calculation.

[0028] (2) Noise factor assessment module 12.

[0029] The noise factor evaluation module 12 is responsible for separating and quantifying two main noise components (mechanical vibration noise and impact noise) from the acceleration signal transmitted from module 11, and outputting the mechanical vibration factor and impact factor for subsequent weight calculation.

[0030] Optionally, the noise factor evaluation module 12 is used to determine the mechanical vibration factor and impact factor of the acceleration signal for each roadbed area. The mechanical vibration factor characterizes the noise component in the acceleration signal caused by engine vibration during multiple compaction cycles, while the impact factor characterizes the noise component in the acceleration signal caused by uneven roadbed surface during multiple compaction cycles.

[0031] For example, such as Figure 2 As shown, the noise factor evaluation module 12 may include two sub-modules: a mechanical vibration factor calculation sub-module 121 and an impact factor calculation sub-module 122. These two sub-modules are described below: (2.1) Mechanical vibration factor calculation submodule 121.

[0032] Optionally, the mechanical vibration factor calculation submodule 121 is used to determine the mechanical vibration factor of the acceleration signal.

[0033] Specifically, the mechanical vibration factor calculation submodule 121 quantifies the noise component caused by engine vibration in the acceleration signal. The implementation process is as follows: First, the engine speed during the compaction of the vibratory roller is obtained, and the basic mechanical vibration frequency (unit: Hz, the engine rotation frequency directly determines the core frequency of mechanical vibration noise) is calculated accordingly. Then, a preset frequency error range (usually 2Hz in engineering, which can be adjusted according to the roller model) is set, and the target frequency range of mechanical vibration noise is determined in combination with the aforementioned basic mechanical vibration frequency. Finally, the acceleration signal (i.e., the compaction signal) of a certain roadbed area transmitted by the signal acquisition module 11 is subjected to Fourier transform to obtain the spectrum data of the signal. Then, the ratio of the "sum of amplitudes of all frequencies within the target frequency range" to the "sum of amplitudes of all frequencies in the full frequency band" in the spectrum data is calculated. This ratio is the mechanical vibration factor of the acceleration signal of the roadbed area (the larger the ratio, the higher the proportion of mechanical vibration noise in the signal).

[0034] (2.2) Impact factor calculation submodule 122.

[0035] Optionally, the impact factor calculation submodule 122 is used to determine the impact factor.

[0036] Specifically, the impact factor calculation submodule 122 quantifies the impact noise component caused by uneven roadbed surface in the acceleration signal. The implementation process is as follows: First, extract all data points of a certain compaction signal in the roadbed area from the acceleration signal transmitted by the signal acquisition module 11. Then, using the amplitude difference of different data points as the distance metric, the k-means clustering algorithm is used to cluster all data points into two categories (one category is normal vibration data points, and the other category is impact noise data points—impact noise is generated by roadbed protrusions, and its amplitude is significantly greater than that of the normal signal). Finally, the mean amplitude of all data points in the two data clusters is calculated, and the absolute value of the difference between the two is obtained. The absolute value is normalized by the sigmoid function, and the normalized result is the impact factor of the acceleration signal in the roadbed area (the larger the value, the more significant the impact noise in the signal).

[0037] The noise factor evaluation module 12 outputs mechanical vibration factor and impact factor, which are then transmitted to the response factor evaluation module 13 as noise weight parameters in the response factor calculation process. This reduces the interference of noise on the response factor evaluation results by using negative correlation weights (the larger the factor, the smaller the weight).

[0038] (3) Response Factor Assessment Module 13.

[0039] The response factor evaluation module 13 is responsible for combining the acceleration signal spectrum data, mechanical vibration factor and impact factor to quantify the degree of influence of the roadbed compaction state on each frequency component of the acceleration signal, and finally outputting the roadbed response factor, which directly determines the priority of subsequent denoising.

[0040] Optionally, the response factor evaluation module 13 is used to determine the subgrade response factors of multiple frequency components in the acceleration signal corresponding to the same subgrade area based on the spectral data, mechanical vibration factor, and impact factor of the acceleration signal corresponding to the same subgrade area. The subgrade response factor characterizes the degree to which the frequency components are affected by the subgrade compaction state.

[0041] For example, such as Figure 3 As shown, the response factor evaluation module 13 may include two sub-modules: a time factor calculation sub-module 131 and a spatial factor calculation sub-module 132. These two sub-modules will be described in detail below: (3.1) Time factor calculation submodule 131.

[0042] Optionally, the time factor calculation submodule 131 is used to determine the time roadbed response factor of each frequency component based on the amplitude sequence of the frequency component in the spectrum data corresponding to the same roadbed area, combined with the mechanical vibration factor and impact factor corresponding to the roadbed area.

[0043] Specifically, the time factor calculation submodule 131 analyzes the variation of frequency components with the number of compaction cycles based on the time trend of multiple compaction cycles, quantifying the influence of the time dimension of compaction state on frequency. The specific process is as follows: Amplitude sequence extraction: From the acceleration signal spectrum data transmitted by the signal acquisition module 11, for a certain frequency component in a certain roadbed area, extract the amplitude sequence of that frequency under each rolling. Symbol parameter calculation: The symbol parameter is calculated based on the amplitude sequence. This parameter is used to characterize the rate of change of the subgrade compaction trend (positive value indicates that the compaction degree increases at a slower rate, negative value indicates that the increase rate increases at a faster rate, and it is directly related to the change of subgrade stiffness). Weighting coefficient calculation: Based on the mechanical vibration factor and impact factor of each compaction in the roadbed area transmitted by the noise factor evaluation module 12, the weighting coefficient is calculated (the larger the noise factor, the smaller the weighting coefficient, reducing the interference of noise on trend analysis). Time response factor calculation: The time subgrade response factor of this frequency component is calculated by a preset formula (the larger the value of the time subgrade response factor, the more significant the change of this frequency component with the number of compaction times, and the stronger the influence of the subgrade compaction state). For details of the calculation in conjunction with the preset formula, please refer to S603 below.

[0044] (3.2) Spatial factor calculation submodule 132.

[0045] Optionally, the spatial factor calculation submodule 132 is used to determine the spatial roadbed response factor of each frequency component based on the spectral data corresponding to multiple roadbed areas within a preset local range around the roadbed area.

[0046] Specifically, the spatial factor calculation submodule 132 starts from the spatial gradient characteristics of local areas and combines the spatial continuity of subgrade compaction to quantify the influence of the spatial dimension of compaction state on frequency components. The implementation process is as follows: Local range determination: Preset local range (usually 5 meters in engineering to ensure sufficient coverage of adjacent areas to reflect spatial continuity), and determine all roadbed areas within its local range with the target roadbed area as the center; Three-dimensional space construction: Using the spatial coordinates (x, y) of the roadbed area as the plane coordinates and the number of compaction times as the vertical coordinate, a three-dimensional data space is constructed, and the spectrum data of all roadbed areas within the local range transmitted by the signal acquisition module 11 is imported. Gradient vector and direction consistency analysis: For a certain frequency component, calculate the gradient vector of each roadbed region in three-dimensional space at that frequency (similar to the pixel gradient in an image, reflecting the rate of change of amplitude in space); calculate the sum vector of all gradient vectors (i.e., the overall vector), and solve the cosine similarity between each gradient vector and the overall vector (the greater the cosine similarity, the more consistent the direction of frequency amplitude change in that region is with the local overall trend, and the more likely it is to be a true reflection of the roadbed compaction state). Modulus length gradient factor calculation: Using the time subgrade response factors of each subgrade region output by the time factor calculation submodule 131 as weights, calculate the weighted variance of the modulus length of all gradient vectors in the local range (the smaller the difference in modulus length, the smoother the amplitude change in space, which is consistent with the continuous change characteristics of compaction). Then, normalize the weighted variance to obtain the modulus length gradient factor. Spatial response factor calculation: Calculate the directional gradient factor according to the preset formula, and multiply the directional gradient factor by the modulus gradient factor to obtain the spatial roadbed response factor of the frequency component. For details of the calculation in conjunction with the preset formula, please refer to S606 below.

[0047] Therefore, the response factor evaluation module 13 uses the product of the time-based and spatial-based subgrade response factors calculated by the two sub-modules as the subgrade response factor of the frequency component. The larger the product, the more significantly the frequency component is affected by the compaction state in both time trend and spatial variation, and it is a key frequency that reflects the true compaction state of the subgrade.

[0048] (4) Signal denoising module 14.

[0049] The signal denoising module 14 is responsible for accurately denoising the latest compaction acceleration signal collected by the module 11 based on the subgrade response factor output by the module 13, so as to obtain an acceleration signal that can truly reflect the subgrade compaction state.

[0050] Optionally, the signal denoising module 14 is used to adaptively filter and denoise the acceleration signal corresponding to the latest compaction cycle in the subgrade area according to the subgrade response factor, so as to obtain the denoised acceleration signal.

[0051] Specifically, the signal denoising module 14 first presets initial wavelet coefficients (usually 0.1 in engineering, but can be adjusted according to the signal noise intensity); secondly, for each frequency component in the latest compaction acceleration signal, it calculates the optimal wavelet coefficient for that frequency (the larger the subgrade response factor, the smaller the optimal wavelet coefficient—indicating that the frequency is a key component reflecting the compaction state, and the filtering intensity needs to be reduced to retain the signal; conversely, the filtering needs to be strengthened to remove noise); then, based on the calculated optimal wavelet coefficients, it performs wavelet transform domain adaptive filtering on each frequency component; finally, after filtering all frequency components, it integrates the filtering results of each frequency component to obtain the denoised acceleration signal of the latest compaction in the subgrade area.

[0052] The signal denoising module 14 outputs a denoised acceleration signal, which is then transmitted to the uniformity analysis module 15. Since the main noise interference has been removed from the signal, it can provide accurate vibration response data for subsequent CMV calculations, avoiding compaction degree assessment deviations caused by noise.

[0053] (5) Uniformity analysis module 15.

[0054] The uniformity analysis module 15 is responsible for calculating the compaction values ​​of each subgrade area based on the denoised acceleration signal, and comprehensively calculating the overall compaction uniformity of the subgrade.

[0055] Optionally, the uniformity analysis module 15 is used to calculate the compaction value of the subgrade area based on the denoised acceleration signal, and to determine the uniformity of subgrade compaction based on the compaction value of all subgrade areas.

[0056] Specifically, the uniformity analysis module 15 first uses the known CMV calculation method based on the denoised acceleration signal to obtain the compaction value of each subgrade area. Second, it presets a compaction threshold (set according to engineering design standards, such as the minimum CMV value required to meet subgrade bearing capacity), compares the CMV of each subgrade area with the threshold, and identifies low-compaction areas where the CMV is below the threshold. Then, it counts the number and area proportion of low-compaction areas, calculates the statistical dispersion index of CMV for all subgrade areas (such as variance and standard deviation—the smaller the dispersion, the smaller the CMV difference between areas, and the better the subgrade compaction uniformity). Finally, it combines the proportion of low-compaction areas with the CMV dispersion index to determine the uniformity level of subgrade compaction and generates a uniformity analysis report including a CMV spatial distribution map, low-compaction area location markings, uniformity index statistics, and optimization suggestions.

[0057] Therefore, the uniformity analysis module 15 outputs a roadbed compaction uniformity analysis report, which can provide a direct basis for decision-making in engineering construction adjustments. For example, it can formulate a compaction supplementation plan for the low compaction areas marked in the report to improve the overall compaction quality of the roadbed and avoid problems such as uneven settlement and cracks in the subsequent pavement.

[0058] The uniformity detection system 10 and its included modules have been described above.

[0059] For example, such as Figure 4 The diagram shown is a flowchart illustrating a method for detecting the uniformity of roadbed compaction according to an embodiment of the present invention, comprising the following steps: S401. Acquire the acceleration signal generated by the vibratory roller during multiple compaction processes of the roadbed. The roadbed is divided into multiple roadbed regions.

[0060] For example, this step can be performed by the signal acquisition module 11 in the uniformity detection system 10 described above, and specifically includes the following steps: (1) Divide the roadbed into multiple continuous roadbed regions. The width of each roadbed region is the same as the width of the vibratory roller, and the length of each roadbed region is a preset value.

[0061] Specifically, the signal acquisition module 11 divides the complete roadbed to be inspected into multiple continuous roadbed areas according to the engineering inspection accuracy requirements. The width of each roadbed area is consistent with the width of the vibratory roller, and the length of each roadbed area is a preset value L. For example, the preset value L ranges from 0.5 meters to 2 meters, and for example, an empirical value of 1 meter can be taken. Through the aforementioned division method, it is ensured that each roadbed area can be completely covered by the vibratory roller during the compaction process, and a spatial coordinate index is established.

[0062] (2) By using the acceleration sensor installed inside the roller of the vibratory roller, the acceleration signal generated in each roadbed area during each rolling process is collected during the process of the vibratory roller rolling the roadbed multiple times at a fixed speed.

[0063] For example, in conjunction with the foregoing description of the signal acquisition module 11, the fixed speed ranges from 2 km / h to 5 km / h; the sampling frequency of the accelerometer is not less than 1000 Hz, and its measurement axis is consistent with the radial direction of the roller.

[0064] Following this, the signal acquisition module 11 performs preliminary processing on the collected acceleration signals (e.g., storing them by category according to "subgrade area number - number of compactions", and removing obvious hardware fault signals). For example, the acceleration signal collected in the i-th subgrade area during the j-th compaction is labeled as A. ij (t), where t is the sampling time point, i is the subgrade area number, and j is the compaction number.

[0065] Therefore, the signal acquisition module 11 can establish an acceleration signal set organized by subgrade area and compaction subsystem, providing a structured data foundation for subsequent analysis.

[0066] S402. For the acceleration signal corresponding to each roadbed area, determine the mechanical vibration factor and impact factor of the acceleration signal.

[0067] Among them, the mechanical vibration factor is used to characterize the noise component in the acceleration signal caused by engine vibration during multiple compactions, and the impact factor is used to characterize the noise component in the acceleration signal caused by unevenness of the roadbed surface during multiple compactions. Specifically, the noise component caused by engine vibration during multiple compactions includes: the intensity of interference components that do not change with the compaction state, generated by the unstable operation of the roller's own power system (such as the engine) or by the vibration of mechanical components unrelated to the roadbed response.

[0068] For example, this step can be performed by the noise factor evaluation module 12 in the uniformity detection system 10 described above, and specifically includes the following steps: (1) Determine the mechanical vibration factor of the acceleration signal.

[0069] Optionally, the mechanical vibration factor calculation submodule 121 in the noise factor evaluation module 12 determines the mechanical vibration factor of the acceleration signal, specifically including: obtaining the basic mechanical vibration frequency of the vibratory roller, and determining the target frequency range based on the basic mechanical vibration frequency; then, performing a spectral transformation on the acceleration signal to obtain spectral data; finally, determining the mechanical vibration factor based on the frequency amplitude within the target frequency range and the full-band amplitude in the spectral data. It should be noted that the specific process of the mechanical vibration factor calculation submodule 121 determining the mechanical vibration factor according to the aforementioned sub-steps is described in S501-S503 below, and will not be repeated here.

[0070] In another possible implementation, the mechanical vibration factor calculation submodule 121 can also adaptively adjust the bandwidth based on the engine speed fluctuation characteristics when determining the target frequency range. By analyzing the engine speed fluctuation range during historical compaction processes and dynamically calculating the frequency error threshold, it can effectively adapt to the actual situation of engine speed changes during construction, avoiding the problem of missed or misjudged noise frequencies caused by fixed thresholds when speed fluctuations are large.

[0071] In another possible implementation, besides Fourier transform, wavelet transform is also suitable for the spectrum analysis requirements of this scheme. Compared to Fourier transform, which is mainly for stationary signals, wavelet transform, with its excellent time-frequency analysis capabilities, can more accurately handle non-stationary signals generated by speed fluctuations during rolling, thereby improving the accuracy of mechanical vibration frequency feature extraction.

[0072] Thus, the mechanical vibration factor calculation submodule 121 can calculate the noise component caused by engine vibration in the quantized acceleration signal.

[0073] (2) Determine the impact factor of the acceleration signal.

[0074] Optionally, the impact factor calculation submodule 122 in the noise factor evaluation module 12 determines the impact factor of the acceleration signal, specifically including: clustering the data points in the acceleration signal according to their amplitude to obtain two data clusters; then, determining the impact factor based on the absolute value of the difference between the mean amplitudes of the two data clusters. It should be noted that the specific process of the impact factor calculation submodule 122 determining the impact factor according to the aforementioned sub-steps is described in S504-S505 below, and will not be repeated here.

[0075] In another possible implementation, the impact factor calculation submodule 122 can also use the DBSCAN clustering algorithm or the hierarchical clustering algorithm to complete the data clustering when determining the impact factor. The DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm does not require a preset number of clusters and automatically identifies impact noise data points with abnormal amplitudes through a density threshold, making it particularly suitable for scenarios with a very small number of impact noise points. The hierarchical clustering algorithm, on the other hand, presents the amplitude correlation between data points intuitively by constructing a tree structure, which facilitates manual verification of the rationality of the clustering results.

[0076] Alternatively, in assessing data cluster differences, the impact factor calculation submodule 122 can expand the evaluation method of the amplitude mean difference to the amplitude median difference or the amplitude variance ratio. When there are anomalous data points with individual extreme amplitudes in the acceleration signal, the median difference can more robustly reflect the typical amplitude level of the data cluster; while the amplitude variance ratio, by calculating the variance ratio between the impact noise cluster and the normal signal cluster, can simultaneously characterize the difference in amplitude dispersion between the two types of data, thus providing richer dimensions for impact noise quantification.

[0077] Thus, the impact factor calculation submodule 122 can calculate the impact noise component caused by the unevenness of the roadbed surface in the quantized acceleration signal.

[0078] S403. Based on the spectral data, mechanical vibration factor, and impact factor of the acceleration signal corresponding to the same subgrade area, determine the subgrade response factors of multiple frequency components in the acceleration signal corresponding to the subgrade area. The subgrade response factors characterize the degree to which the frequency components are affected by the subgrade compaction state.

[0079] For example, this step can be performed by the response factor evaluation module 13 in the uniformity detection system 10 described above, specifically including: First, for each frequency component, based on the amplitude sequence of the frequency component in the spectrum data corresponding to the same roadbed area, combined with the mechanical vibration factor and impact factor corresponding to the roadbed area, the temporal roadbed response factor of the frequency component is determined; then, based on the spectrum data corresponding to multiple roadbed areas within a preset local range around the roadbed area, the spatial roadbed response factor of the frequency component is determined; finally, based on the temporal roadbed response factor and the spatial roadbed response factor, the roadbed response factor of the frequency component is determined. It should be noted that the specific process of the response factor evaluation module 13 in determining the roadbed response factor according to the aforementioned sub-steps is described in S601-S603 below, and will not be repeated here.

[0080] In another possible implementation, for the time dimension assessment, the response factor assessment module 13 can adopt an alternative method based on the goodness of fit of the amplitude trend or the rate of change of amplitude variance. The former is to fit the frequency amplitude sequence and calculate the goodness of fit, and then perform weighted correction by combining the mechanical vibration factor and the impact factor; the latter is to analyze the changing trend of the amplitude variance between adjacent compaction cycles and perform weighted calculation of the rate of change by combining the noise factor. Both can effectively characterize the response characteristics of frequency components changing with time.

[0081] In another possible implementation, regarding spatial dimension assessment, the response factor assessment module 13 can employ methods based on spatial correlation coefficients or neighborhood amplitude difference statistics. The former calculates the spatial correlation coefficients of frequency amplitudes between the target area and neighboring areas, and then performs weighted fusion by combining time-based roadbed response factors; the latter calculates the weighted results by statistically analyzing the dispersion of amplitude differences between the target area and its neighbors, and then combining these with time-based response factors. Both methods can accurately capture the continuous distribution characteristics of frequency components in the spatial dimension.

[0082] Therefore, the response factor evaluation module 13 integrates the two-dimensional evaluation results to obtain a roadbed response factor that can comprehensively characterize the degree of influence of the roadbed compaction state on the frequency components, providing a clear basis for subsequent targeted denoising.

[0083] S404. Based on the subgrade response factor, the acceleration signal corresponding to the latest compaction cycle in the subgrade area is subjected to adaptive filtering and denoising to obtain the denoised acceleration signal.

[0084] For example, this step can be performed by the signal denoising module 14 in the uniformity detection system 10 described above, and specifically includes the following steps: (1) For each frequency component in the acceleration signal corresponding to the latest compaction, determine the filtering coefficient for the frequency component based on the roadbed response factor of the frequency component.

[0085] Optionally, the signal denoising module 14 presets an initial wavelet coefficient β, the value of which ranges from 0.05 to 0.2, for example, an empirical value of 0.1; for the subgrade response factor F(ω), where ω represents the frequency component, the optimal wavelet coefficient K(ω) = β / F(ω) for the frequency component is calculated; where the value range of the subgrade response factor F(ω) is (0, 1], characterizing the degree to which the corresponding frequency component is affected by the subgrade compaction state.

[0086] (2) Based on the filtering coefficients, perform adaptive filtering of the frequency components in the wavelet transform domain.

[0087] In this step, an adaptive filtering algorithm in the wavelet transform domain is used, with the optimal wavelet coefficient K(ω) as the threshold parameter, to filter each frequency component in the latest compaction acceleration signal. For each layer coefficient after wavelet decomposition, coefficients less than K(ω) are set to zero, and coefficients greater than K(ω) are retained. Wavelet reconstruction is then performed to achieve adaptive denoising based on the roadbed response characteristics.

[0088] S405. Based on the denoised acceleration signal, calculate the compaction value of the subgrade area, and determine the uniformity of subgrade compaction based on the compaction value of all subgrade areas.

[0089] For example, this step can be performed by the uniformity analysis module 15 in the uniformity detection system 10 described above, and specifically includes the following steps: (1) Analyze the spatial distribution of compaction values ​​in all subgrade areas and identify low compaction areas where compaction values ​​are lower than the preset compaction threshold.

[0090] Specifically, the uniformity analysis module 15 first uses the known CMV calculation method based on the denoised acceleration signal to obtain the compaction value of each subgrade area.

[0091] Meanwhile, the uniformity analysis module 15 presets a compaction meter threshold, with an example value of 80 (the specific value can be adjusted according to the type of subgrade material, such as 75-85 for cohesive soil subgrade and 85-90 for sandy soil subgrade). The rule for the value is that it should not be lower than the minimum CMV value corresponding to the subgrade design bearing capacity.

[0092] (2) Determine the uniformity of subgrade compaction based on the number of low-compaction areas, the area ratio, and the statistical dispersion index of compaction values ​​of all subgrade areas.

[0093] In this step, the uniformity analysis module 15 counts the number and area proportion of low-compaction areas and calculates the statistical dispersion index of CMV for all subgrade areas. For example, the index may include variance and standard deviation. The smaller the dispersion of the index, the smaller the difference in CMV between areas, and the better the uniformity of subgrade compaction.

[0094] Finally, the uniformity analysis module 15 combines the proportion of low-compaction areas with the CMV dispersion index to determine the uniformity level of the subgrade compaction and generates a uniformity analysis report that includes a CMV spatial distribution map, low-compaction area location markings, uniformity index statistics, and optimization suggestions.

[0095] Based on the above technical solution, the embodiments of the present invention systematically identify and separate mechanical vibration noise and impact noise in acceleration signals, construct a roadbed response factor that integrates spatiotemporal characteristics, and perform adaptive filtering accordingly, which significantly improves the measurement accuracy of compaction values, thereby achieving accurate and reliable assessment of the uniformity of roadbed compaction and effectively overcoming the problem of distortion of detection results caused by noise interference in complex construction environments by traditional methods.

[0096] For example, in combination Figure 4 ,like Figure 5 The diagram shown is a flowchart illustrating another method for detecting the uniformity of roadbed compaction according to an embodiment of the present invention. In this method, for the acceleration signal corresponding to each roadbed region, the mechanical vibration factor and impact factor of the acceleration signal are determined, specifically including the following steps: S501. Obtain the basic mechanical vibration frequency of the vibratory roller and determine the target frequency range based on the basic mechanical vibration frequency.

[0097] Specifically, the mechanical vibration factor calculation submodule 121 obtains the real-time speed R of the vibratory roller engine through the engine speed sensor, calculates the basic mechanical vibration frequency f according to the formula f=R / 60, where R represents the engine speed per minute; with the basic mechanical vibration frequency f as the center frequency, a preset frequency error threshold τ is set, and the target frequency range is determined as [f-τ, f+τ], where the value of τ is in the range of 1-5Hz.

[0098] S502. Perform spectral transformation on the acceleration signal to obtain spectral data.

[0099] In this step, the mechanical vibration factor calculation submodule 121 processes the acquired acceleration signal A. ij (t) is subjected to a Fast Fourier Transform to obtain the corresponding spectrum data F. ij (ω), where ω represents the frequency, F ij (ω) represents the amplitude of the j-th compaction signal in the i-th subgrade area at frequency ω.

[0100] S503. Determine the mechanical vibration factor based on the frequency amplitude within the target frequency range and the full-band amplitude in the spectrum data.

[0101] For example, the mechanical vibration factor calculation submodule 121 first calculates the sum of amplitudes S1 of all frequencies within the target frequency range in the spectrum data, that is, the sum of amplitudes corresponding to each frequency in the interval [f-τ, f+τ], reflecting the total intensity of mechanical vibration noise. Then, it calculates the sum of amplitudes S2 of all frequencies across the entire frequency band in the spectrum data, that is, the sum of amplitudes of all frequency components contained in the acceleration signal, reflecting the total intensity of the signal.

[0102] Furthermore, the mechanical vibration factor calculation submodule 121 uses the ratio S1 / S2 of the aforementioned two calculation results as the mechanical vibration factor to quantify the proportion of mechanical vibration noise in the acceleration signal. The larger this factor is, the higher the proportion of fixed-frequency vibration interference energy in the signal that is unrelated to the compaction state and originates from the machine itself. The more significantly the signal data is affected by such interference, the lower its reliability should be in subsequent analysis.

[0103] S504. Cluster the data points in the acceleration signal according to their amplitude to obtain two data clusters.

[0104] Specifically, the impact factor calculation submodule 122 first acquires the acceleration signal A of a certain roadbed area during a certain compaction, output by the signal acquisition module 11. ij (t). To accurately identify the amplitude abrupt change characteristics caused by the impact and avoid interference from the zero-crossing point of periodic vibration, the acceleration signal A is first processed. ij (t) Perform a Hilbert transform to extract its envelope signal, or calculate its root mean square value within a sliding time window as a characteristic sequence representing the local energy of the signal. The data points of this envelope or energy sequence will then be used as the objects of cluster analysis.

[0105] Optionally, the impact factor calculation submodule 122 first acquires the acceleration signal of a certain roadbed area during a certain rolling process output by the signal acquisition module 11, and extracts all data points contained in the signal; Following this, the impact factor calculation submodule 122 uses the amplitude difference between different data points as the distance metric (i.e., the absolute value of the amplitude difference between two data points, reflecting the difference in signal strength between the data points), employs the k-means clustering algorithm, and sets the number of clusters to 2. The basis for setting the number of clusters to 2 is that there are only two core components in the acceleration signal: one is the normal vibration data points, which include the effective signals from the roller's own vibration and the roadbed feedback, with small amplitudes and concentrated distribution; the other is the impact noise data points, caused by the unevenness of the roadbed surface, with amplitudes significantly greater than the normal signal.

[0106] Finally, after performing k-means clustering, the impact factor calculation submodule 122 obtains two data clusters. One cluster (which can be denoted as C1) has relatively small amplitudes and typically has a higher proportion of data points; the other cluster (which can be denoted as C2) has relatively large amplitudes and typically has a lower proportion of data points. For example, in a certain acceleration signal, the amplitudes of the data points in cluster 1 are concentrated between 0.5 and 2 m / s². 2 The amplitude of data points in cluster 2 is concentrated between 5 and 8 m / s. 2 The two types of signal components were clearly separated.

[0107] S505. Determine the impact factor based on the absolute value of the difference between the mean amplitudes of the two data clusters.

[0108] In this step, the impact factor calculation submodule 122 calculates the mean values ​​μ1 and μ2 of the data point amplitudes in the two clusters C1 and C2 respectively, and then calculates the absolute difference between the two means D=|μ1-μ2|.

[0109] Next, the absolute difference D is normalized, and the impact factor R is obtained using the sigmoid function. ij =1 / [1+e -D The sigmoid function is used to characterize the degree of influence of impact noise in acceleration signals. It can be understood that the core function of the aforementioned sigmoid function is to map the absolute value of the difference to the interval [0, 1], ensuring that the range of values ​​for the impact factor is uniform.

[0110] Based on the above technical solution, this invention achieves accurate separation and feature extraction of the main noise sources by quantifying the influence of mechanical vibration noise and impact noise in acceleration signals, providing accurate noise feature parameters for subsequent signal denoising, effectively overcoming the problem of distortion in compaction state assessment caused by noise interference in traditional methods, and significantly improving the reliability of compaction value calculation and the accuracy of roadbed compaction uniformity detection.

[0111] For example, in combination Figure 4 ,like Figure 6 The diagram shown is a flowchart illustrating another method for detecting the uniformity of roadbed compaction according to an embodiment of the present invention. In this method, based on the spectral data, mechanical vibration factor, and impact factor of the acceleration signal corresponding to the same roadbed region, the roadbed response factor of multiple frequency components in the acceleration signal corresponding to the roadbed region is determined. Specifically, the method includes the following steps: S601. For each frequency component, based on the amplitude sequence of the frequency component in the spectrum data corresponding to the same roadbed area, and combined with the mechanical vibration factor and impact factor corresponding to the roadbed area, determine the time roadbed response factor of the frequency component.

[0112] For each subgrade area, time trend analysis requires spectral data from at least three consecutive compactions. When the number of compactions j is less than 3 (i.e., the first and second compactions), valid sign parameters cannot be calculated. In this case, a default initial time subgrade response factor value can be set (e.g., ...). =0.5), or at this stage, time-trend-based weights can be temporarily omitted, and spectral data can be used directly for subsequent spatial analysis. The following description is for the case where the number of compaction events j≥3.

[0113] For example, this step is performed by the time factor calculation submodule 131 in the response factor evaluation module 13, and specifically includes the following steps: (1) Determine the sign parameter based on the amplitude sequence. The sign parameter is used to characterize the rate of change of the compaction trend.

[0114] For example, the time factor calculation submodule 131 first determines the amplitude sequence of the frequency component being analyzed, specifically including: extracting the amplitude corresponding to the frequency from the spectral data of the i-th, i+1-th, and i+2-th compaction in the same roadbed area, respectively, to form the spectral amplitude sequence of three consecutive compactions. , , Where the subscripts i, i+1, and i+2 represent the crushing sequence number, Specifically, it refers to the spectral amplitude corresponding to frequency ω in the i-th compaction of the roadbed area. Therefore, , , It refers to three consecutive data points in a sequence that describe the change of the amplitude of the same frequency component over time (number of times it is rolled).

[0115] Furthermore, the symbolic parameter is obtained through the difference operation of the amplitude sequence, and its core function is to reflect the changing trend of the rate of increase in subgrade compaction. Specifically, it is obtained through... Calculation. Among them, the molecular part... This is the second difference of the amplitude sequence, reflecting the change in the rate of change of amplitude. A positive result indicates that the amplitude growth rate at that frequency is accelerating; a negative result indicates that the amplitude growth rate is slowing down. (The denominator is...) The first difference of the amplitude sequence reflects the direction of amplitude change under the current number of compaction cycles. If the result is positive, it indicates that the amplitude of the frequency increases with the number of compaction cycles, that is, the stiffness of the soft roadbed increases after being compacted, and the vibration amplitude fed back to the roller increases. If the result is negative, it indicates that the amplitude decreases with the number of compaction cycles, that is, it is mostly noise interference. The sign function outputs 1 when the input is positive, -1 when the input is negative, and 0 when the input is 0. When the final sign parameter is 1, it indicates that the frequency amplitude is increasing and the rate of increase is slowing down; when it is -1, it indicates that the amplitude change does not conform to the compaction law and may be dominated by noise.

[0116] It should be noted that when calculating the sign parameter, the denominator should be determined first. If the absolute value is less than a very small positive number ε (e.g., ε=0.01), then the amplitude is considered to have not changed effectively, and the sign parameter of this calculation is directly defined as 0; otherwise, the calculation is performed according to the original formula, and its sign function value is taken.

[0117] (2) Determine the weighting coefficients for weighted calculation based on the mechanical vibration factor and the impact factor.

[0118] In this step, the mechanical vibration factor of the i-th compaction... With impact factor The reciprocal of the sum is used as a weighting coefficient to reduce the contribution of noisy compaction cycles in the calculation.

[0119] (3) Perform weighted calculations based on symbolic parameters and weighting coefficients to determine the time-based roadbed response factor.

[0120] Optionally, the time factor calculation submodule 131 determines the time-based roadbed response factor according to the following formula: In the above formula, The time-based roadbed response factor represents the frequency component ω. The larger the value, the more significant the amplitude change of the frequency component conforms to the compaction law in the time dimension, and the stronger the influence of the roadbed compaction state. and These respectively represent the magnitude and , , The mechanical vibration factor and impact factor are calculated from the acceleration signal of the i-th compaction in the corresponding subgrade area. The summation range in the formula is from 1 to m-2, where m is the total number of compaction cycles in the subgrade area. By summing the calculation results of all compaction cycles i that meet the conditions, a time characteristic quantity characterizing the degree to which the frequency component ω is affected by the subgrade compaction process is obtained. .

[0121] Understandable That is, the weighting coefficient, where and The intensity of unsteady mechanical vibration disturbance and impact disturbance were quantified separately. The stronger the disturbance, the lower the reliability of the compaction data, and the smaller the weight should be assigned when assessing the compaction trend. Therefore, the weighting coefficient was designed as follows: This establishes a negative correlation between interference intensity and weight. Furthermore, combined with the aforementioned... and Calculations show that, The value of must be greater than zero. The value of is not less than zero, so the denominator in the weighting coefficient will not be equal to zero, and the calculation will not be meaningless.

[0122] S602. Based on the spectral data of multiple roadbed areas within a preset local range around the roadbed area, determine the spatial roadbed response factor of the frequency components.

[0123] For example, this step is performed by the spatial factor calculation submodule 132 in the response factor evaluation module 13, and specifically includes the following steps: (1) Construct a three-dimensional data space based on the spatial coordinates and compaction times of each roadbed area. In the three-dimensional data space, each data point corresponds to the spectral data of a roadbed area under one compaction.

[0124] In this step, the spatial factor calculation submodule 132 presets a local range γ=5 meters, establishes a spatial neighborhood centered on the current roadbed area, and constructs a three-dimensional coordinate system by combining the spatial coordinates (x,y) with the compaction level j. Each coordinate point is associated with the corresponding spectral data F. ij (ω).

[0125] (2) For each frequency component, determine the gradient vector of each roadbed region under the frequency component in the three-dimensional data space.

[0126] Specifically, for the frequency component ω being analyzed, the spatial factor calculation submodule 132 uses the central difference method to calculate the spectral amplitude F of each roadbed region in the spatial dimension. ij The rate of change of (ω) yields the gradient vector characterizing the spatial variation. , where j is the subgrade region number within a local area. The central difference method is common knowledge in this field, and the specific calculation process will not be elaborated further.

[0127] (3) Determine the spatial subgrade response factor of the frequency component based on the overall directional consistency of all gradient vectors, the dispersion of the magnitude of each gradient vector, and the time subgrade response factor of each subgrade region under the frequency component.

[0128] For example, the spatial factor calculation submodule 132 determines the spatial roadbed response factor of the frequency components through the following steps ac. : a. Calculate the direction gradient factor : In the above formula, The direction gradient factor representing the frequency component ω ranges from -1 to 1. The closer its value is to 1, the higher the proportion of areas with consistent direction and time dimension that conform to the compaction law within a local area. Let represent the gradient vector of the j-th roadbed region under the frequency component ω; M represents the number of roadbed regions within a preset local range; This represents the time-based roadbed response factor of the j-th roadbed region under the frequency component ω (calculated by step S601). The overall value represents the sum of the time-dependent subgrade response factors of all subgrade regions within a local area, and is a positive number.

[0129] It should be noted that in the formula Item is The weighting, specifically, means: by weighting and summing, the directional information of each data point within a preset local range is integrated to quantify the overall spatial consistency of the amplitude variation of the frequency component ω. Because is the time-dependent subgrade response factor for the j-th region at frequency ω. A larger value indicates that the region more closely matches the signal response pattern during subgrade compaction in the time dimension (i.e., less affected by noise and with a significant compaction response). Therefore, with... As a weight, regional data with more reliable responses in the time dimension are given a higher contribution to spatial analysis, ensuring that spatial analysis is based on reliable temporal evolution characteristics.

[0130] Therefore, by summing the weighted gradient vectors of all neighboring points, we obtain the overall vector. . The closer the modulus is to 1, the more consistent the gradient direction is at each point within the preset local range, and the larger the time response factor is at these points. This means that the amplitude of frequency ω exhibits a clear and coordinated gradual change direction within this spatial neighborhood (i.e., the preset local range), a characteristic that highly matches the continuous and gradual distribution of subgrade compaction in physical space. Conversely, if... If the modulus is very small, it means that the orientation is disordered or the proportion of reliable areas is low within the preset local range, which does not meet the expectation of continuous change in compaction.

[0131] b. Calculate the modulus gradient factor : First, the dispersion of the magnitude of each gradient vector is calculated, specifically by calculating the weighted variance of the magnitude of each gradient vector, with the weights being the time-based subgrade response factors for the corresponding subgrade region. Then, the weighted variance is normalized to obtain the modulus gradient factor. Its value range is (0,1). The method for calculating the weighted variance of the specific vector magnitude is common knowledge in the field of mathematics and will not be elaborated here.

[0132] c. Calculate the spatial roadbed response factor : In the above formula, the direction gradient factor is... With modulus gradient factor The physical meaning of multiplication is to simultaneously and collaboratively evaluate whether the frequency component ω in the spatial dimension conforms to the two key characteristics of continuous compaction change: consistent direction and gradual change.

[0133] Understandably, using multiplication instead of addition implies that the conditions of directional consistency and uniformity of change reinforce each other when determining the spatial response. Weakening of either condition (i.e., smaller values ​​for both parameters) will directly lead to a decrease in the final result. This significantly reduces the frequency response factor. This ensures that only frequency components that simultaneously satisfy directional coordination and gradual change can achieve a high spatial roadbed response factor, thereby accurately distinguishing spatially continuous signal components caused by the actual compaction state, and effectively separating them from spatially randomly distributed or isolated mechanical vibration noise and impact noise.

[0134] S603. Determine the subgrade response factor of the frequency component based on the time subgrade response factor and the spatial subgrade response factor.

[0135] For example, the response factor evaluation module 13 multiplies the temporal subgrade response factor and the spatial subgrade response factor, and then normalizes them so that their numerical range falls within the (0, 1] interval, and finally obtains the subgrade response factor that integrates spatiotemporal characteristics. This satisfies both the evolution law of the subgrade compaction process in time and the continuous distribution characteristics in space, providing an accurate frequency component evaluation basis for subsequent signal denoising.

[0136] Based on the above technical solution, this embodiment of the invention establishes a spatiotemporal joint evaluation model of roadbed compaction state by integrating time series analysis and spatial gradient field construction. This effectively overcomes the signal distortion problem caused by mechanical vibration and impact noise in traditional methods, significantly improves the accuracy and robustness of roadbed response factor calculation, and provides a reliable data foundation for subsequent accurate denoising and compaction uniformity determination.

[0137] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0138] Each embodiment in this specification is described in a progressive manner. Similar or identical parts between each embodiment can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A method for detecting the uniformity of roadbed compaction, characterized in that, The method includes: Acceleration signals generated by a vibratory roller during multiple compaction processes of the roadbed are acquired; wherein the roadbed is divided into multiple roadbed regions. For each roadbed area, the mechanical vibration factor and impact factor of the acceleration signal are determined; wherein, the mechanical vibration factor is used to characterize the noise component in the acceleration signal caused by engine vibration during multiple compaction, and the impact factor is used to characterize the noise component in the acceleration signal caused by uneven roadbed surface during multiple compaction. Based on the spectral data of the acceleration signal corresponding to the same subgrade area, the mechanical vibration factor, and the impact factor, the subgrade response factor of multiple frequency components in the acceleration signal corresponding to the subgrade area is determined; wherein, the subgrade response factor is used to characterize the degree to which the frequency components are affected by the subgrade compaction state; Based on the subgrade response factor, the acceleration signal corresponding to the latest compaction time in the subgrade area is adaptively filtered and denoised to obtain the denoised acceleration signal. Based on the denoised acceleration signal, the compaction value of the subgrade area is calculated, and the uniformity of subgrade compaction is determined based on the compaction values ​​of all subgrade areas.

2. The method for detecting the uniformity of roadbed compaction according to claim 1, characterized in that, Determining the mechanical vibration factor of the acceleration signal specifically includes: Obtain the basic mechanical vibration frequency of the vibratory roller, and determine the target frequency range based on the basic mechanical vibration frequency; The acceleration signal is subjected to spectral transformation to obtain spectral data; The mechanical vibration factor is determined based on the frequency amplitude within the target frequency range and the full-band amplitude in the spectrum data.

3. The method for detecting the uniformity of roadbed compaction according to claim 1, characterized in that, Determining the impact factor specifically includes: The data points in the acceleration signal are clustered according to their amplitude to obtain two data clusters; The impact factor is determined based on the absolute value of the difference between the mean amplitudes of the two data clusters.

4. The method for detecting the uniformity of roadbed compaction according to claim 2, characterized in that, Based on the spectral data of the acceleration signal corresponding to the same roadbed area, the mechanical vibration factor, and the impact factor, the roadbed response factor of multiple frequency components in the acceleration signal corresponding to the roadbed area is determined, specifically including: For each frequency component, the time-based roadbed response factor of the frequency component is determined based on the amplitude sequence of the frequency component in the spectrum data corresponding to the same roadbed area, combined with the mechanical vibration factor and the impact factor corresponding to the roadbed area. Based on the spectral data corresponding to multiple roadbed areas within a preset local range surrounding the roadbed area, the spatial roadbed response factor of the frequency component is determined; The subgrade response factor of the frequency component is determined based on the time subgrade response factor and the spatial subgrade response factor.

5. The method for detecting the uniformity of roadbed compaction according to claim 4, characterized in that, For each frequency component, based on the amplitude sequence of the frequency component in the spectral data corresponding to the same roadbed area, and in conjunction with the mechanical vibration factor and the impact factor corresponding to the roadbed area, the time-based roadbed response factor of the frequency component is determined, specifically including: Based on the amplitude sequence, a sign parameter is determined; wherein the sign parameter is used to characterize the rate of change of the compaction trend; The weighting coefficients used for the weighted calculation are determined based on the mechanical vibration factor and the impact factor. The time-based roadbed response factor is determined by performing a weighted calculation based on the symbolic parameters and the weighting coefficients.

6. The method for detecting the uniformity of roadbed compaction according to claim 4, characterized in that, Based on the spectral data corresponding to the roadbed region and the spectral data corresponding to other roadbed regions adjacent to the roadbed region, the spatial roadbed response factor of the frequency component is determined, specifically including: A three-dimensional data space is constructed based on the spatial coordinates and compaction times of each roadbed area; wherein, each data point in the three-dimensional data space corresponds to the spectral data of a roadbed area under one compaction. For each frequency component, the gradient vector of each roadbed region under the frequency component is determined in the three-dimensional data space; The spatial subgrade response factor of the frequency component is determined based on the overall directional consistency of all gradient vectors, the dispersion of the magnitude of each gradient vector, and the temporal subgrade response factor of each subgrade region under the frequency component.

7. The method for detecting the uniformity of roadbed compaction according to claim 1, characterized in that, Based on the subgrade response factor, adaptive filtering and denoising are performed on the acceleration signal corresponding to the latest compaction cycle in the subgrade area, specifically including: For each frequency component in the acceleration signal corresponding to the latest compaction, a filtering coefficient is determined based on the roadbed response factor of the frequency component. Based on the filtering coefficients, adaptive filtering in the wavelet transform domain is performed on the frequency components.

8. The method for detecting the uniformity of roadbed compaction according to claim 1, characterized in that, Based on the compaction gauge values ​​of all subgrade areas, the uniformity of subgrade compaction is determined, specifically including: Analyze the spatial distribution of compaction meter values ​​in all subgrade areas to identify low-compaction areas where compaction meter values ​​are below a preset compaction meter threshold; The uniformity of subgrade compaction is determined based on the number and area ratio of the low-compaction zones and the statistical dispersion index of compaction values ​​for all subgrade zones.

9. The method for detecting the uniformity of roadbed compaction according to any one of claims 1-8, characterized in that, Acquiring the acceleration signals generated by the vibratory roller during multiple compaction processes of the roadbed, specifically including: The roadbed is divided into multiple continuous roadbed regions; wherein the width of each roadbed region is the same as the width of the vibratory roller, and the length of each roadbed region is a preset value; Acceleration sensors installed inside the rollers of the vibratory roller collect acceleration signals generated in each roadbed area during each compaction process as the vibratory roller compacts the roadbed multiple times at a fixed speed.

10. A system for detecting the uniformity of roadbed compaction, characterized in that, The system includes: a signal acquisition module, a noise factor evaluation module, a response factor evaluation module, a signal denoising module, and a uniformity analysis module; The signal acquisition module is used to acquire the acceleration signal generated by the vibratory roller during multiple compaction processes of the roadbed; wherein the roadbed is divided into multiple roadbed regions; The noise factor evaluation module is used to determine the mechanical vibration factor and impact factor of the acceleration signal for each roadbed area; wherein, the mechanical vibration factor is used to characterize the noise component in the acceleration signal caused by engine vibration during multiple compaction, and the impact factor is used to characterize the noise component in the acceleration signal caused by uneven roadbed surface during multiple compaction. The response factor evaluation module is used to determine the roadbed response factor of multiple frequency components in the acceleration signal corresponding to the roadbed area based on the spectral data of the acceleration signal corresponding to the same roadbed area, the mechanical vibration factor, and the impact factor; wherein, the roadbed response factor is used to characterize the degree to which the frequency components are affected by the roadbed compaction state; The signal denoising module is used to adaptively filter and denoise the acceleration signal corresponding to the latest compaction time in the roadbed area according to the roadbed response factor, so as to obtain the denoised acceleration signal. The uniformity analysis module is used to calculate the compaction value of the subgrade area based on the denoised acceleration signal, and to determine the uniformity of subgrade compaction based on the compaction values ​​of all subgrade areas.

Images