A parameter characterization method for the gradation and shape of coarse-grained soil in compaction construction

Abstract

The present invention discloses a parameter characterization method for the gradation and shape of coarse-grained soil in compaction construction, comprising the following steps: S1, obtaining coarse-grained soil; S2, combining a plurality of coarse-grained soils of different gradations; preliminarily screening the gradation parameters to obtain representative gradation parameters; S3, performing three-dimensional scanning to obtain the particle shape of the coarse-grained soil; preliminarily screening the shape parameters to obtain representative shape parameters; S4, preparing several groups of coarse-grained soil, characterizing them by using representative gradation parameters and representative shape parameters; performing vibration compaction tests respectively; S5, based on the vibration compaction test data, using a machine learning algorithm, establishing a regression model between gradation parameters and shape parameters and dry density; identifying key parameters that affect the compaction effect by using the SHAP value obtained by the regression model. The present invention screens out key parameters that have a significant impact on the compaction effect, can reasonably select suitable materials and parameters during the construction process, and optimize the compaction process.

Description

Technical Field

[0001] The invention belongs to the technical field of civil engineering and traffic infrastructure construction, and in particular relates to a parameter characterization method for gradation and shape of coarse-grained soil in compaction construction. Background Art

[0002] Coarse-grained soil is the main filling material for highway and railway subgrades, and its compaction characteristics are significantly affected by particle gradation and shape. However, the key parameters used to characterize the compaction effect of coarse-grained soil have not yet been clearly defined, making it difficult to accurately identify the main factors affecting the compaction effect and effectively guide the scientific selection of construction equipment. This limitation results in high energy consumption, low construction efficiency, and difficulty in ensuring compaction quality during the compaction process.

[0003] Grading parameters include fractal dimension, curvature coefficient, non-uniformity coefficient, maximum particle size, etc., while shape indicators include overall shape, concavity and convexity, sphericity, etc. Although some studies have explored the influence of some parameters on compaction effect, they have not yet fully covered all relevant parameters, and existing studies are mostly limited to the influence of a single parameter, lacking in-depth discussion on the comprehensive effect of grading and shape.

[0004] Therefore, how to select the key parameters that play a decisive role in the compaction effect from a variety of gradation and shape parameters, and establish a systematic and scientific selection method for the gradation and shape of coarse-grained soil particles that takes the compaction effect into consideration has become a problem that needs to be solved in engineering. Summary of the invention

[0005] The purpose of the present invention is to provide a parameter characterization method for the gradation and shape of coarse-grained soil in compaction construction, so as to solve the problem raised in the background technology of how to screen out the key parameters that play a decisive role in the compaction effect from a variety of gradation and shape parameters, and to establish a systematic and scientific selection method for the gradation and shape of coarse-grained soil particles taking into account the compaction effect.

[0006] To achieve the above object, the present invention provides a parameter characterization method for coarse-grained soil gradation and shape in compaction construction, comprising the following steps:

[0007] S1. Obtaining coarse-grained soil used in compaction construction;

[0008] S2. Screening and grading the obtained coarse-grained soil, and reorganizing the graded coarse-grained soil to combine more than 15 coarse-grained soils with different gradations; using 20 gradation parameters to characterize the combined coarse-grained soils with different gradations, and obtaining a gradation data group corresponding to the number of gradation parameters; using a hierarchical clustering method and a principal component analysis method to preliminarily screen the gradation parameters, and select less than 20 representative gradation parameters;

[0009] S3. Randomly perform three-dimensional scanning on 100 to 300 coarse-grained soils to obtain the particle shape of the coarse-grained soil; characterize the coarse-grained soils that have been three-dimensionally scanned with 25 shape parameters to obtain a shape data group corresponding to the number of shape parameters; use a hierarchical clustering method and a principal component analysis method to preliminarily screen the shape parameters and select less than 25 representative shape parameters;

[0010] S4. Prepare several groups of coarse-grained soils with different gradations and shapes, and characterize the prepared groups of coarse-grained soils using representative gradation parameters and representative shape parameters obtained through preliminary screening; conduct vibration compaction tests respectively, and record the dry density changes of the coarse-grained soils with different gradations and shapes during the vibration compaction process; the dry density changes are used to characterize the compaction effect of the coarse-grained soil;

[0011] S5. Based on the vibration compaction test data, a machine learning algorithm is used to establish a regression model between grading parameters, shape parameters and dry density; the SHAP value obtained by the regression model is used to identify the key parameters affecting the compaction effect.

[0012] In a specific embodiment, the machine learning algorithm includes a ridge regression model, a support vector machine model, a decision tree model, an XGBoost model and a LightGBM model. By comparing the computational efficiency and accuracy of the five models, the optimal model is selected for subsequent analysis.

[0013] In a specific embodiment, when selecting the optimal model, the mean absolute error, root mean square error and correlation coefficient (R 2 ), first ensure that the selected regression model satisfies the mean absolute error ≤ 0.05 and the root mean square error ≤ 0.1, and then select the correlation coefficient (R 2 ) The largest regression model is the optimal model;

[0014] The SHAP value obtained by the optimal model is used to identify the key parameters affecting the compaction effect.

[0015] In a specific implementation, when establishing a regression model between grading parameters, shape parameters and dry density, the entire vibration compaction test data is randomly divided into a training set and a test set, and a five-fold cross validation is used to adjust the parameters. The hyperparameters of the regression model are optimized using a Bayesian optimization algorithm. In order to quantitatively evaluate the accuracy of the model correction results, the root mean square error is used as the loss function of the model training. The particle characteristic values ​​of the regression model are representative grading parameters and representative shape parameters, and the target characteristic value is dry density.

[0016] In a specific implementation, in step S2, the 20 grading parameters include:

[0017] Maximum particle sizeD max ; Minimum particle size D min ; Particle size with a cumulative passing rate of 10% D 10 ; Particle size with a cumulative passing rate of 50% D 50 ; Particle size with a cumulative passing rate of 60% D 60 ; Average particle size D mean ; The proportion of particles with the largest particle size R max ; Proportion of particles with the smallest particle size R min ; Unevenness coefficient C u ; Curvature coefficient C c ; Particle size range width D max-min ; Maximum and minimum particle size ratio D max / min ; Particle size distribution index G d ; Fractal dimension D f ; Standard deviation of particle size distribution σ ; Grading curve fitting degree r 2 ; Slope of particle size distribution curve S ; Asymmetry of particle distribution A g ; Particle size cumulative curve area ratio A r ; Particle size distribution flatness F d .

[0018] In a specific implementation, in step S3, 25 shape parameters are used to characterize the three-dimensionally scanned coarse-grained soil, and the 25 shape parameters include:

[0019] Overall shape parameters ; Overall shape parameters ; Overall shape parameters ; Overall shape parameters in , ; Overall shape parameters ; Overall shape parameters ; Overall shape parameters ; Overall shape parameters ; Overall shape parameters ; Overall shape parameters ; Overall shape parameters ; Overall shape parameters ; Overall shape parameters ; Sphericity parameter ; Sphericity parameter ; Sphericity parameter ; Sphericity parameter Where n=1.6075; sphericity parameter ; Sphericity parameter ; Sphericity parameter ; Sphericity parameter ; Sphericity parameter ; Concavity parameter ; Concavity parameter ; Concavity parameter .

[0020] In a specific embodiment, in step S4, the prepared coarse-grained soil includes coarse-grained soil of five typical filler types in the shape of flakes, strips, blocks, smooth and angular shapes; a vibration compaction test is carried out by applying a vibration load to the coarse-grained soil using a UTM-250 multifunctional testing machine.

[0021] Compared with the prior art, the present invention has the following beneficial effects:

[0022] The present invention can accurately identify and quantify the influence of each grade and shape parameter on the compaction effect by establishing a quantitative relationship model of compaction effect based on particle gradation and shape characteristics. The quantitative relationship model of compaction effect established by the present invention analyzes the contribution of different particle characteristics to dry density and screens out key parameters that have a significant influence on the compaction effect. By accurately identifying these key parameters, it is possible to reasonably select suitable materials and parameters during the construction process, optimize the compaction process, thereby improving the compaction efficiency and ensuring a better compaction effect.

[0023] In addition to the above-described objects, features and advantages, the present invention has other objects, features and advantages. The present invention is further described in detail below. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The drawings constituting a part of this application are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0025] Figure 1 is a particle gradation curve diagram of an embodiment of the present invention;

[0026] Figure 2 It is a flow chart of integration and screening of gradation parameters and shape parameters according to an embodiment of the present invention;

[0027] Figure 3It is a flow chart of machine learning regression analysis according to an embodiment of the present invention. DETAILED DESCRIPTION

[0028] The embodiments of the present invention are described in detail below. The specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0029] Example 1

[0030] A parameter characterization method for coarse-grained soil gradation and shape in compaction construction of the present invention comprises the following steps:

[0031] S1. Obtain coarse-grained soil for compaction.

[0032] S2. Screen and grade the obtained coarse-grained soil. In this embodiment, sieves with apertures of 30 mm, 25 mm, 20 mm, 15 mm, 10 mm, 5 mm, and 2 mm are used to screen the coarse-grained soil. Coarse-grained soil with a particle size range of 2 mm to 30 mm is screened to obtain particle size groups with particle size ranges of 2 mm to 5 mm, 5 mm to 10 mm, 10 mm to 15 mm, 15 mm to 20 mm, 20 mm to 25 mm, and 25 mm to 30 mm.

[0033] Using the obtained particle size groups, 15 coarse-grained soils with different gradations were recombined.

[0034] 20 gradation parameters were used to characterize the 15 coarse-grained soils with different gradations, and 20 gradation data sets corresponding to the number of gradation parameters were obtained. Each gradation data set contained 15 data related to the same gradation parameter.

[0035] Hierarchical clustering method and principal component analysis method were used to screen the grading parameters and select less than 20 representative grading parameters.

[0036] Hierarchical clustering method was used to cluster and group 20 gradation parameters.

[0037] By using the hierarchical clustering algorithm, the Euclidean distances between various grading parameters are first calculated.

[0038] The gradation parameters are gradually classified according to the similarity values. In the clustering process, the gradation parameters with higher correlation are classified into the same clustering group. The gradation parameters in the same clustering group have high similarity and can be regarded as parameters reflecting the same physical phenomenon or similar characteristics, which can better represent similar shape characteristics.

[0039] For the grading parameters in each cluster group, the correlation analysis between the two shape parameters was performed respectively, and the correlation coefficient between the parameters was calculated.

[0040] When the correlation between the gradation parameters within a cluster group is high, that is, the correlation coefficient exceeds 0.7, it can be considered that the gradation parameters in the cluster group have a high degree of repetition and redundancy, and it is necessary to screen out the most representative gradation parameter. At the same time, for the secondary parameters with a correlation of gradation parameters within the cluster lower than 0.4, they are still considered as representative parameters to avoid missing important information. For shape features with the same correlation coefficient, the principal component analysis (PCA) method is used to select the main features, and the principal component with a larger load in each cluster group is selected as the representative gradation parameter of the shape feature of the cluster group, thereby providing more accurate feature variables for subsequent training.

[0041] Through the above method, redundant features can be effectively reduced, the independence and representativeness of the selected grading parameters can be ensured, the interference of redundant information can be avoided, and the significant feature differences between each clustering group can be retained.

[0042] Finally, a group of representative grading parameter groups are screened out, which includes multiple grading parameters. The grading parameters in this grading parameter group can effectively characterize the characteristics of particle grading and avoid redundancy caused by the high correlation between grading parameters.

[0043] The grading parameters in the selected representative grading parameter groups will be used to describe 15 particle grading curves in subsequent steps to further analyze the impact of particle grading on compaction quality and provide more accurate characteristic variables for subsequent training.

[0044] The 20 grading parameters include:

[0045] Maximum particle size D max ; Minimum particle size D min ; Particle size with a cumulative passing rate of 10% D 10 ; Particle size with a cumulative passing rate of 50% D 50 ; Particle size with a cumulative passing rate of 60% D 60 ; Average particle size D mean ; The proportion of particles with the largest particle size R max ; Proportion of particles with the smallest particle size R min ; Unevenness coefficient C u ; Curvature coefficient C c ; Particle size range width D max-min ; Maximum and minimum particle size ratio D max / min ; Particle size distribution indexG d ; Fractal dimension D f ; Standard deviation of particle size distribution σ ; Grading curve fitting degree r 2 ; Slope of particle size distribution curve S ; Asymmetry of particle distribution A g ; Particle size cumulative curve area ratio A r ; Particle size distribution flatness F d .

[0046] The calculation formulas for 20 grading parameters are shown in Table 1.

[0047] Table 1. Grading parameters

[0048]

[0049] Table 1: Maximum particle size D max ; Minimum particle size D min ; The cumulative pass rates are 10%, 30%, 50%, 60%, 90% ( D 10 , D 30 , D 50 , D 60 and D 90 Average particle size D mean ; Particle size within the screening range D i Cumulative pass rate P i ; R max is the proportion of particles with the largest particle size; R min is the proportion of particles with the smallest particle size; the sum of all particle sizes D total ; Unevenness coefficient C u ; Curvature coefficient C c ; Particle size range width D max-min ; Maximum and minimum particle size ratio D max / min ; Particle size distribution index G d ; Fractal dimension Df ; Cumulative percentage of particle size D P ( D ); standard deviation of particle size distribution σ ; Grading curve fitting degree r 2 ; Slope of particle size distribution curve S ; P 1. P 2 is the particle size D 1. D 2 cumulative pass rate; actual measured particle size distribution points y i ; Corresponding points of the fitting curve y fit ; Average value of particle size distribution ; Asymmetry of particle distribution A g ; Particle size cumulative curve area ratio A r ; Actual cumulative pass rate P obs ; Ideal cumulative pass rate P ideal ; Particle size distribution flatness F d .

[0050] S3. Randomly perform 3D scanning on 200 coarse-grained soils to obtain the particle shape of the coarse-grained soil; characterize the coarse-grained soils that have been 3D scanned using 25 shape parameters, and obtain 25 shape data sets corresponding to the number of shape parameters. Shape parameters include overall shape (such as the length of the three principal axes, aspect ratio, flatness, etc.) to describe the geometric structure of the particles; sphericity, which is used to characterize the degree to which the particles tend to be spherical; and concavity, which reflects the angular characteristics of the particle surface.

[0051] The 200 coarse-grained soils obtained include flake, strip, block, smooth and angular particles. The true shape of the particles is obtained by 3D particle scanning technology. The corresponding shape parameters are calculated by combining 3D image processing software and Python code to establish a coarse particle shape database.

[0052] Hierarchical clustering method and principal component analysis method were used to screen the shape parameters and select less than 25 representative shape parameters.

[0053] In order to reduce the impact of the correlation between shape parameters on subsequent training, a hierarchical clustering method is used to group all shape parameters. By calculating the similarity between the parameters, using appropriate distance metrics and clustering algorithms, the shape parameters are divided into multiple groups according to similarity. The shape parameters in each group have a high correlation and can better represent similar gradation characteristics. For the shape parameters in each group, the correlation analysis between the two shape parameters is performed separately, the correlation coefficient between the parameters is calculated, and the parameters with larger or smaller correlation are screened out. This can reduce redundant features and ensure that each parameter provides independent information in subsequent analysis. For shape features with the same correlation coefficient, the principal component analysis method is used to select the main features, and the principal component with a larger load in each group is selected as the main parameter of the shape feature of this group.

[0054] The 25 shape parameters include:

[0055] Overall shape parameters ; Overall shape parameters ; Overall shape parameters ; Overall shape parameters in , ; Overall shape parameters ; Overall shape parameters ; Overall shape parameters ; Overall shape parameters ; Overall shape parameters ; Overall shape parameters ; Overall shape parameters ; Overall shape parameters ; Overall shape parameters ; Sphericity parameter ; Sphericity parameter ; Sphericity parameter ; Sphericity parameter Where n=1.6075; sphericity parameter ; Sphericity parameter ; Sphericity parameter ; Sphericity parameter ; Sphericity parameter ; Concavity parameter ; Concavity parameter ; Concavity parameter .

[0056] In the above formula: the lengths of the longest side, the second side, and the short side of the circumscribed cuboid L , I ,and S ;volume V Surface area S A ; The diameter of a sphere of equal volumed s ; Volume of the circumscribed convex hull V con ; The diameter of the largest inscribed sphere d i ; The diameter of the smallest circumscribed sphere d c ;No. i The radius of the inscribed sphere of the local corners r i ; Total number of corner areas N c .

[0057] S4. Prepare several groups of coarse-grained soils with different gradations and shapes. For example, use five shapes of coarse-grained soils, such as flakes, strips, blocks, smooth and angular, to prepare 6 groups of coarse-grained soils with different gradations. Each group of tests uses different filler types and gradation combinations, for a total of 30 groups of tests.

[0058] The representative gradation parameters and representative shape parameters obtained by preliminary screening were used to characterize the 30 groups of coarse-grained soils.

[0059] The UTM-250 multifunctional testing machine was used to apply vibration load to coarse-grained soil for vibration compaction test. The test equipment is a steel cylindrical container with a diameter of 160 mm and a height of 250 mm. The samples with a total mass of 7 kg were naturally piled into the cylindrical container by layered filling. During the test, the vibration frequency was set to 20 Hz, the vibration amplitude was 10 kN, a sinusoidal waveform load was applied, and the number of vibrations was 1000 times to simulate the vibration compaction process in actual engineering. Vibration compaction tests were carried out on 30 groups of coarse-grained soils, and the dry density changes of coarse-grained soils with different gradations and shapes during the vibration compaction process were recorded respectively; the dry density changes were used to characterize the compaction effect of coarse-grained soils.

[0060] The prepared coarse-grained soil includes five typical filler types of coarse-grained soil in the shapes of flakes, strips, blocks, smooth and angular shapes.

[0061] S5. Based on the vibration compaction test data, a machine learning algorithm is used to establish a regression model between grading parameters, shape parameters and dry density; the SHAP value obtained by the regression model is used to identify the key parameters affecting the compaction effect.

[0062] The machine learning algorithms include ridge regression model, support vector machine model, decision tree model, XGBoost model and LightGBM model. By comparing the computational efficiency and accuracy of the five models, the optimal model is selected for subsequent analysis.

[0063] When establishing the regression model between gradation parameters, shape parameters and dry density, the entire vibration compaction test data was randomly divided into a training set (70%) and a test set (30%), and a five-fold cross validation was used to adjust the parameters, and the Bayesian optimization algorithm was used to optimize the hyperparameters of the regression model. In order to quantitatively evaluate the accuracy of the model correction results, the root mean square error was used as the loss function of the model training; the particle characteristic values ​​of the regression model were representative gradation parameters and representative shape parameters, and the target characteristic value was dry density.

[0064] When selecting the optimal model, the mean absolute error, root mean square error, and correlation coefficient (R 2 ), first ensure that the selected regression model satisfies the mean absolute error ≤ 0.05 and the root mean square error ≤ 0.1, and then select the correlation coefficient (R 2 ) is the optimal model, indicating that this model has better prediction accuracy. Then, the optimal model is used for SHAP value analysis to deeply explore the synergy of particle shape and gradation and their comprehensive influence on compaction effect. Through SHAP value, the influence degree and law of each parameter on compaction effect are analyzed, so as to identify the key parameters affecting compaction effect.

[0065] The above contents are further detailed descriptions of the present invention in combination with specific preferred embodiments, and it cannot be determined that the specific implementation of the present invention is limited to these descriptions. For ordinary technicians in the technical field to which the present invention belongs, several simple deductions and substitutions can be made without departing from the concept of the present invention, which should be regarded as belonging to the protection scope of the present invention.

Claims

1. A parameter characterization method for coarse-grained soil gradation and shape in compaction construction, characterized in that: The steps include: S1. Obtaining coarse-grained soil used in compaction construction; S2, screening and grading the obtained coarse-grained soil, and reorganizing the graded coarse-grained soil to combine more than 15 coarse-grained soils with different gradations; using 20 gradation parameters to characterize the combined coarse-grained soils with different gradations, and obtaining a gradation data group corresponding to the number of gradation parameters; using a hierarchical clustering method and a principal component analysis method to preliminarily screen the gradation parameters, and select less than 20 representative gradation parameters; S3. Randomly perform three-dimensional scanning on 100 to 300 coarse-grained soils to obtain the particle shape of the coarse-grained soil; characterize the coarse-grained soils that have been three-dimensionally scanned with 25 shape parameters to obtain a shape data group corresponding to the number of shape parameters; use a hierarchical clustering method and a principal component analysis method to preliminarily screen the shape parameters and select less than 25 representative shape parameters; S4. Prepare several groups of coarse-grained soils with different gradations and shapes, and characterize the prepared groups of coarse-grained soils using representative gradation parameters and representative shape parameters obtained through preliminary screening; conduct vibration compaction tests respectively, and record the dry density changes of the coarse-grained soils with different gradations and shapes during the vibration compaction process; the dry density changes are used to characterize the compaction effect of the coarse-grained soil; S5. Based on the vibration compaction test data, a regression model between gradation parameters, shape parameters and dry density was established using a machine learning algorithm; The SHAP values ​​obtained by the regression model are used to identify the key parameters affecting the compaction effect; In step S2, the 20 grading parameters include: maximum particle size D max ; Minimum particle size D min ; Particle size with a cumulative passing rate of 10% D 10 ; Particle size with a cumulative passing rate of 50% D 50 ; Particle size with a cumulative passing rate of 60% D 60 ; Average particle size D mean ; The proportion of particles with the largest particle size R max ; Proportion of particles with the smallest particle size R min ; Unevenness coefficient C u ; Curvature coefficient C c ; Particle size range width D max-min ; Maximum and minimum particle size ratio D max / min ; Particle size distribution index G d ; Fractal dimension D f ; Standard deviation of particle size distribution σ ; Grading curve fitting degree r 2 ; Slope of particle size distribution curve S ; Asymmetry of particle distribution A g ; Particle size cumulative curve area ratio A r ; Particle size distribution flatness F d ; In step S3, the 25 shape parameters include: Overall shape parameters ; Overall shape parameters ; Overall shape parameters ; Overall shape parameters in , ; Overall shape parameters ; Overall shape parameters ; Overall shape parameters ; Overall shape parameters ; Overall shape parameters ; Overall shape parameters ; Overall shape parameters ; Overall shape parameters ; Overall shape parameters ; Sphericity parameter ; Sphericity parameter ; Sphericity parameter ; Sphericity parameter Where n=1.6075; sphericity parameter ; Sphericity parameter ; Sphericity parameter ; Sphericity parameter ; Sphericity parameter ; Concavity parameter ; Concavity parameter ; Concavity parameter ; In the above formula: the length of the longest side, the second side and the short side of the circumscribed cuboid L , I ,and S ;volume V Surface area S A ; The diameter of a sphere of equal volume d s ; Volume of the circumscribed convex hull V con ; The diameter of the largest inscribed sphere d i ; The diameter of the smallest circumscribed sphere d c ;No. i The radius of the inscribed sphere of the local corners r i ; Total number of corner areas N c .

2. The parameter characterization method for coarse-grained soil gradation and shape in compaction construction according to claim 1, characterized in that: The machine learning algorithms include ridge regression model, support vector machine model, decision tree model, XGBoost model and LightGBM model. By comparing the computational efficiency and accuracy of the five models, the optimal model is selected for subsequent analysis.

3. The parameter characterization method for coarse-grained soil gradation and shape in compaction construction according to claim 2, characterized in that: When selecting the optimal model, calculate the mean absolute error, root mean square error, and correlation coefficient R for each regression model 2 First, ensure that the selected regression model satisfies the mean absolute error ≤ 0.05 and the root mean square error ≤ 0.1, and then select the correlation coefficient R 2 The largest regression model is the optimal model; The SHAP value obtained by the optimal model is used to identify the key parameters affecting the compaction effect.

4. The parameter characterization method for coarse-grained soil gradation and shape in compaction construction according to claim 1, characterized in that: When establishing the regression model between grading parameters, shape parameters and dry density, the entire vibration compaction test data was randomly divided into a training set and a test set, and five-fold cross validation was used to adjust the parameters. The Bayesian optimization algorithm was used to optimize the hyperparameters of the regression model. In order to quantitatively evaluate the accuracy of the model correction results, the root mean square error was used as the loss function of the model training. The particle characteristic values ​​of the regression model were representative grading parameters and representative shape parameters, and the target characteristic value was dry density.

5. The parameter characterization method for coarse-grained soil gradation and shape in compaction construction according to claim 1, characterized in that: In step S4, the prepared coarse-grained soil includes five typical filler types of coarse-grained soil in the shape of flakes, strips, blocks, smooth and angular shapes; a vibration compaction test is carried out by applying a vibration load to the coarse-grained soil using a UTM-250 multifunctional testing machine.

Images