Method for predicting performance of rap steel slag asphalt mixture based on microstructure parameters

Abstract

The application discloses a RAP steel slag asphalt mixture performance prediction method based on a microstructure parameter, and comprises the following steps: (1) performing RAP steel slag recycled asphalt mixture mix proportion design, and testing road performance parameters thereof; (2) performing image processing through a brightness gradient and a marked watershed algorithm, and calculating microstructure parameters, i.e., an average contact point number, an average shape factor, an aggregate shape factor variance, an average minimum distance, a total minimum distance, an aggregate minimum distance variance, a direction angle variance and a direction angle mean value; (3) using ridge and lasso functions in matlab to establish a ridge regression and a lasso regression model of the microstructure parameters and the road performance parameters; and (4) inputting the microstructure parameters into the ridge regression equation and the lasso regression equation to predict the road performance parameters. The recycled mixture microstructure parameters identified through image recognition can effectively predict the road performance of the mixture.

Description

Technical Field

[0001] This invention belongs to the field of highway asphalt pavement technology, and more specifically, relates to a method for predicting the performance of RAP steel slag asphalt mixture based on microstructural parameters. Background Technology

[0002] With the continuous development of road engineering in my country, natural aggregate resources are decreasing and prices are rising. A large amount of reclaimed concrete (RAP) cannot be efficiently utilized, impacting the ecological environment and causing resource waste. Increasing research is shifting towards how to more efficiently utilize recycled concrete and industrial solid waste to replace natural aggregates in road construction. Steel slag possesses excellent physical properties, with abundant angularity, high roughness, and high strength, making it suitable for road construction after processing. Recycled asphalt mixtures with added steel slag exhibit better performance indicators, playing a dual role in controlling production costs and mitigating environmental problems. The increase in stiffness of RAP mixtures is mainly due to the increased hardness of aged asphalt, but it also reduces the toughness of recycled asphalt mixtures. The improvement in mixture performance by steel slag is mainly due to the interlocking structure formed when used as aggregate, its rough surface providing better internal friction, and better asphalt adhesion compared to natural aggregates. However, the incorporation of steel slag may affect the water stability of the mixture. The volume expansion of free calcium oxide contained in steel slag after hydration adversely affects the skeletal stability of recycled asphalt mixtures. This is closely related to the production process and aging methods of steel slag. The advantages and disadvantages of incorporating RAP and steel slag into the asphalt mixture need to be considered in a balanced way. While increasing the RAP content improves the stiffness of the mixture, it reduces the overall workability. The interlocking structure formed by the steel slag in the aggregate and the internal friction of the aggregate are also affected. Balancing these characteristics of RAP and steel slag requires in-depth research into the microstructure and road performance of RAP-steel slag recycled asphalt mixtures. RAP-steel slag recycled asphalt mixtures that meet road performance requirements will save significant resources, reduce engineering costs, and alleviate environmental problems to some extent.

[0003] Current research on RAP (Recycled Asphalt Powder) steel slag asphalt mixtures mainly focuses on road performance testing, with limited research on its microstructure. Existing microstructure studies evaluate its properties at the microscopic level without establishing a connection with road performance parameters. Therefore, this invention provides a method for predicting the performance of RAP steel slag asphalt mixtures based on microstructure parameters. Summary of the Invention

[0004] Objective of this invention: To provide a method for predicting the performance of RAP (Recycled Asphalt Powder) steel slag asphalt mixtures based on microstructural parameters. When preparing different RAP steel slag recycled asphalt mixtures, researchers can input microstructural parameters into the regression equation to predict the corresponding road performance parameters.

[0005] Technical Solution: A method for predicting the performance of RAP steel slag asphalt mixtures based on microstructural parameters. This method establishes ridge regression and cable regression equations between microstructural parameters and road performance to predict the road performance of RAP steel slag recycled asphalt mixtures. The method includes the following steps:

[0006] (1) Conduct mix design of RAP steel slag recycled asphalt mixture and test its road performance parameters: dynamic stability, freeze-thaw splitting residual strength ratio, water immersion Marshall residual stability, and low temperature fracture failure strain.

[0007] (2) Image processing is performed using the brightness gradient and marker watershed algorithm to calculate the following microstructure parameters: average number of contact points, average shape factor, aggregate shape factor variance, average minimum spacing, total minimum spacing, aggregate minimum spacing variance, orientation angle variance, and orientation angle mean.

[0008] (3) Use the ridge and lasso functions in MATLAB to establish ridge regression and lasso regression models for microstructural parameters and road performance parameters;

[0009] (4) Input the microstructural parameters into the ridge regression equation and the lasso regression equation to predict the road performance parameters.

[0010] Furthermore, in step (1), the mix design and road performance testing of RAP steel slag recycled asphalt mixture include:

[0011] (101) Test the properties of raw materials to ensure they meet the specifications. Analyze the gradation of RAP and adjust the aggregate gradation curve;

[0012] (102) Conduct standard Marshall compaction test, design 5 samples with different asphalt contents, and determine the optimal asphalt content of the selected RAP steel slag recycled asphalt mixture through Marshall test.

[0013] (103) Conduct road performance tests on RAP steel slag recycled asphalt mixtures, testing road performance parameters including: dynamic stability, freeze-thaw splitting residual strength ratio, water immersion Marshall residual stability, and low-temperature fracture strain. The calculation formulas are as follows:

[0014] Dynamic stability:

[0015]

[0016] Where: DS—dynamic stability of asphalt mixture, cycles / mm;

[0017] d1, d2 — Deformation amounts at times t1 and t2, respectively, in mm;

[0018] C1—Experimental machine type coefficient, taken as 1.0;

[0019] C2—Specimen coefficient, 1.0 for laboratory-prepared specimens with a width of 300 mm;

[0020] N—the reciprocating rolling speed of the experimental wheel.

[0021] Freeze-thaw splitting residual strength ratio:

[0022] R T1 =0.006287P T1 / h1 (2)

[0023] R T2 =0.006287P T2 / h2 (3)

[0024]

[0025] In the formula: R T1 R T2 —Splitting tensile strength of a single specimen in the non-freeze-thaw group and the freeze-thaw group, in MPa;

[0026] P T1 P T2 —Maximum experimental load value of a single specimen in the non-freeze-thaw group and the freeze-thaw group, in N;

[0027] h1, h2 — Height of each specimen in the non-freeze-thaw group and the freeze-thaw group, respectively, in mm;

[0028] R T1 R T2 —The average splitting tensile strength of the effective specimens in the non-freeze-thaw group and the freeze-thaw group, in MPa;

[0029] TSR—Freeze-thaw splitting test strength ratio, %.

[0030] Water immersion Marshall residue stability:

[0031]

[0032] Where: MS0—residual stability of the Marshall specimen after immersion in water, %;

[0033] MS—Stability of Marshall specimen after immersion in water for 30 min, kN;

[0034] MS1—Stability of Marshall specimen after immersion in water for 48 hours, kN;

[0035] Low-temperature fracture strain:

[0036]

[0037] In the formula: ε B —The maximum flexural tensile strain at specimen failure, με;

[0038] h—Height of the specimen across the cross section, mm;

[0039] L—span of the specimen, mm;

[0040] d—mid-span deflection at specimen failure, mm.

[0041] Furthermore, in step (2), image processing is performed using the brightness gradient and marker watershed algorithm to calculate the mesoscopic structure parameters, including:

[0042] (201) Quartzite was used to replace basalt as the new aggregate in RAP steel slag recycled asphalt mixture, and four standard Marshall specimens were prepared for each RAP steel slag recycled asphalt mixture.

[0043] (202) The Marshall specimen was cut at a spacing of 1.2 cm using a high-precision cutting machine to obtain a complete disc-shaped specimen;

[0044] (203) Use anhydrous ethanol and a brush to clean the cut surface of the specimen, and then use a high-definition digital camera to take horizontal images of the cut surface of the specimen. 20 clear and uniformly lit photos were selected for each RAP steel slag recycled asphalt mixture.

[0045] (204) Use the rgb2gray function to convert the original slice image into an 8-bit grayscale image, and then determine the actual size of the pixel according to the image resolution and sample size;

[0046] (205) Use the Remove Outliers function to adjust the radius and threshold to remove overly bright and dark areas in a grayscale image;

[0047] (206) The MedianFilter algorithm is used to set a certain radius to eliminate random noise in the image and obtain a clearer grayscale image;

[0048] (207) The Adaptive histogram equalization algorithm was used to enhance the contrast between aggregates and asphalt mortar. The equalized image was used for the segmentation of each phase of the recycled asphalt mixture.

[0049] (208) Using the Adaptive Thresholding algorithm, a larger gray threshold is set to select the sample region in the image to obtain sample region A, thus initially removing the interference of non-sample regions on image segmentation;

[0050] (209) Based on the grayscale histogram of the equalized image, determine the grayscale range of the asphalt mortar, and use the AdaptiveThresholding algorithm again to select the area in sample area A that is lower than the grayscale median of the asphalt mortar, thus obtaining the asphalt mortar and pore area B.

[0051] (210) Use the Dilation function to dilate the asphalt mortar and the pore region B. Set the dilation type to Disc and select the specific precision value according to the image processing effect.

[0052] (211) Use the Arithmetic function to edit A1=AB and remove the asphalt mortar and pores in the sample area to obtain the aggregate area;

[0053] (212) Use the Fill small spots function to remove the holes inside the aggregate, set the fill unit to Disc, and adjust the pixel size of the fill unit according to the image processing effect to obtain the complete aggregate area A2.

[0054] (213) Using the Image gradient function, after setting the gradient type and smoothing coefficient, extract the brightness difference boundary I from the grayscale data in the aggregate area A2;

[0055] (214) Repeatedly use the Adaptive Thresholding algorithm to extract particles of different components in region A2 to ensure that the extraction accuracy does not completely cover the target particle area. If necessary, the lasso-Autotrace edges tool in AVIZO can be used for manual correction. After correction, the steel slag aggregate region C, RAP aggregate region D and new aggregate region E are initially obtained.

[0056] (215) Using the Maker-based watershed function, the brightness gradient boundary I extracted by the Image gradient function is used as the marker watershed algorithm to extend the boundaries of each phase. After extending regions C, D, and E, the complete steel slag aggregate region C1, RAP aggregate region D1, and quartzite aggregate region E1 are obtained.

[0057] (216) The three types of coarse aggregate image data segmented above are stored as RAP, steel slag, and new aggregate datasets, respectively. Specific data are calculated based on the corresponding microstructure parameters, including: average number of contact points (x1), average shape factor (x2), aggregate shape factor variance (x3), average minimum spacing (x4), total minimum spacing (x5), aggregate minimum spacing variance (x6), orientation angle variance (x7), and orientation angle mean (x8). These microstructure parameters are defined as follows:

[0058] Average number of contact points: The contact threshold is set to 0.54 mm. The total number of contact points of all coarse aggregates is determined by a window search algorithm, and then the average number of contact points is obtained.

[0059] Shape factor: Shape factor = (Crofton perimeter)² / (4π × aggregate area) where the formula for calculating Crofton perimeter is as follows:

[0060] Crofton perimeter=(D(X,0°)+D(X,45°)+D(X,90°)+D(X,135°)) / 4(7)

[0061] D(X,a)=∫N a (X)dx (8)

[0062] In the formula: D(X,a) represents the number of intersections between all parallel lines in direction a and the aggregate boundary.

[0063] N a The number of intersections between the boundary and the straight line in direction a.

[0064] Minimum aggregate spacing: Minimum aggregate spacing is the shortest distance from any point on the edge of the target aggregate to any point on the edge of its nearest aggregate.

[0065] Orientation angle: The aggregate orientation angle is the angle between the aggregate's moment of inertia vector and the coordinate axis. The angle of the moment of inertia vector is in the range of [-90°, 90°]. The formula for calculating the moment of inertia is as follows:

[0066] I x =∫y 2 dA (9)

[0067] I y =∫x 2 dA (10)

[0068] The calculated moment of inertia vector is as follows:

[0069] Furthermore, in step (3), a regression analysis model is established between microstructural parameters and road performance, and key microstructural parameters affecting different road performance are selected, including:

[0070] (301) Use MATLAB to standardize the data of microstructure parameters and road performance parameters respectively. Set the eight microstructure parameters as independent variables x1 to x8, and set the four road performance parameters of dynamic stability, freeze-thaw splitting residual strength ratio, water immersion Marshall residual stability, and low temperature fracture failure strain as different dependent variables. Scale the independent and dependent variables so that the mean of each group of data is 0 and the standard deviation is 1.

[0071] (302) Using the ridge and lasso functions in MATLAB and setting the regularization term to 0.1, calculate the four sets of regression equations corresponding to the four road performance parameters;

[0072] Furthermore, in step (4), based on the influence of different aggregates on the microstructure, and combined with regression analysis, the synergistic recycling design of RAP and steel slag is optimized, including:

[0073] (401) Prepare other RAP steel slag recycled asphalt mixture specimens that need to be tested for road performance, obtain cross-sectional images of the mixture and calculate the corresponding microstructure parameters by the method in step (2);

[0074] (402) Standardize the microstructure parameter data and input them as independent variables into the ridge regression and lasso regression equations obtained in step (3). Take the average value of the two regression equations to calculate the road performance parameters required for the mixture.

[0075] Beneficial effects: Compared with the prior art, the significant advantages of this invention are:

[0076] 1. This invention utilizes quartzite to replace basalt as a new aggregate, which can accurately distinguish RAP, steel slag and new aggregate during image processing.

[0077] 2. The equivalent diameter of the aggregate calculated in this invention is the Fret diameter, which better reflects the shape characteristics of the aggregate compared to the equivalent radius calculation methods used in other technologies.

[0078] 3. This invention can predict the road performance parameters of RAP steel slag recycled asphalt mixture by calculating microstructural parameters, saving a lot of road performance testing work.

[0079] 4. The three aggregate datasets segmented from the image in this invention can not only be used to predict road performance parameters, but also to explore the effects of RAP and steel slag on microstructure in depth. Attached Figure Description

[0080] Figure 1 This is a technical roadmap of the present invention;

[0081] Figure 2 It is the grayscale histogram of the image after equalization;

[0082] Figure 3 This is a schematic diagram of image segmentation for quartzite aggregate;

[0083] Figure 4 This is a schematic diagram of image segmentation for steel slag aggregate;

[0084] Figure 5 This is a schematic diagram of RAP aggregate image segmentation. Detailed Implementation

[0085] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0086] In this embodiment, the RAP steel slag recycled asphalt mixture is used for both the lower and upper base layers, replacing the conventional ATB-25 asphalt mixture used in the lower or upper base layers.

[0087] like Figure 1 As shown, the method for predicting the performance of RAP steel slag asphalt mixtures based on microstructural parameters includes the following steps:

[0088] (1) Conduct mix design of RAP steel slag recycled asphalt mixture and test its road performance parameters: dynamic stability, freeze-thaw splitting residual strength ratio, water immersion Marshall residual stability, and low temperature fracture failure strain.

[0089] (101) Test the properties of raw materials to ensure they meet the specifications. Analyze the gradation of RAP (see Table 1), and adjust the aggregate gradation (see Table 2).

[0090] Table 1:

[0091]

[0092] Table 2:

[0093]

[0094]

[0095] (102) A standard Marshall compaction test was conducted, and five samples with different asphalt contents were designed. The optimal asphalt content of the selected RAP steel slag recycled asphalt mixture was determined by the Marshall test. When the RAP content was 20% and the steel slag content was 20% and 40%, the optimal asphalt-aggregate ratio was 3.54% and 3.65%, respectively. When the RAP content was 40% and the steel slag content was 20% and 40%, the optimal asphalt-aggregate ratio was 3.3% and 3.56%, respectively.

[0096] (103) Conduct road performance tests on RAP steel slag recycled asphalt mixtures, testing road performance parameters including: dynamic stability, freeze-thaw splitting residual strength ratio, water immersion Marshall residual stability, and low-temperature fracture strain. The calculation formulas are as follows:

[0097] Dynamic stability:

[0098]

[0099] Where: DS—dynamic stability of asphalt mixture, cycles / mm;

[0100] d1, d2 — Deformation amounts at times t1 and t2, respectively, in mm;

[0101] C1—Experimental machine type coefficient, taken as 1.0;

[0102] C2—Specimen coefficient, 1.0 for laboratory-prepared specimens with a width of 300 mm;

[0103] N—the reciprocating rolling speed of the experimental wheel.

[0104] Freeze-thaw splitting residual strength ratio:

[0105] R T1 =0.006287P T1 / h1 (2)

[0106] R T2 =0.006287P T2 / h2 (3)

[0107]

[0108] In the formula: R T1 R T2 —Splitting tensile strength of a single specimen in the non-freeze-thaw group and the freeze-thaw group, in MPa;

[0109] P T1 P T2 —Maximum experimental load value of a single specimen in the non-freeze-thaw group and the freeze-thaw group, in N;

[0110] h1, h2 — Height of each specimen in the non-freeze-thaw group and the freeze-thaw group, respectively, in mm;

[0111] R T1 R T2 —The average splitting tensile strength of the effective specimens in the non-freeze-thaw group and the freeze-thaw group, in MPa;

[0112] TSR—Freeze-thaw splitting test strength ratio, %.

[0113] Water immersion Marshall residue stability:

[0114]

[0115] Where: MS0—residual stability of the Marshall specimen after immersion in water, %;

[0116] MS—Stability of Marshall specimen after immersion in water for 30 min, kN;

[0117] MS1—Stability of Marshall specimen after immersion in water for 48 hours, kN;

[0118] Low-temperature fracture strain:

[0119]

[0120] In the formula: ε B —The maximum flexural tensile strain at specimen failure, με;

[0121] h—Height of the specimen across the cross section, mm;

[0122] L—span of the specimen, mm;

[0123] d—mid-span deflection at specimen failure, mm.

[0124] The test results of the above road performance parameters are shown in Table 3.

[0125] Table 3:

[0126]

[0127] (2) Image processing is performed using the brightness gradient and marker watershed algorithm to calculate the following microstructure parameters: average number of contact points, average shape factor, aggregate shape factor variance, average minimum spacing, total minimum spacing, aggregate minimum spacing variance, orientation angle variance, and orientation angle mean.

[0128] (201) Quartzite was used to replace basalt as the new aggregate in RAP steel slag recycled asphalt mixture, and four standard Marshall specimens of four RAP steel slag recycled asphalt mixtures were prepared.

[0129] (202) The Marshall specimen was cut at a spacing of 1.2 cm using a high-precision cutting machine to obtain a complete disc-shaped specimen;

[0130] (203) Use anhydrous ethanol and a brush to clean the cross-section of the specimen, and then use a high-definition digital camera to take horizontal images of the cross-section of the specimen. For each RAP steel slag recycled asphalt mixture, select 20 clear and uniformly lit photos, for a total of 80 cross-section images.

[0131] (204) Use the rgb2gray function to convert the original slice image into an 8-bit grayscale image, and then determine the actual size of the pixel according to the image resolution and sample size;

[0132] (205) Use the Remove Outliers function to adjust the radius and threshold to remove overly bright and dark areas in a grayscale image;

[0133] (206) The Median Filter algorithm is used, and a radius of 3 pixels is set to eliminate random noise in the image and obtain a clearer grayscale image;

[0134] (207) The Adaptive histogram equalization algorithm was used to enhance the contrast between aggregates and asphalt mortar. The equalized image was used for the segmentation of each phase of the recycled asphalt mixture.

[0135] (208) Using the Adaptive Thresholding algorithm, the gray threshold is set to 26, and the sample region in the image is selected to obtain the sample region A, thus initially removing the interference of the non-sample region on image segmentation;

[0136] (209) Based on the grayscale histogram of the equalized image, determine the grayscale range of the asphalt mortar, and then use the AdaptiveThresholding algorithm again to select the regions in sample region A that are lower than the median grayscale value of the asphalt mortar, thus obtaining the asphalt mortar and pore region B. The grayscale histogram of the equalized image is shown below. Figure 2 ;

[0137] (210) Use the Dilation function to expand the asphalt mortar and the pore region B. Set the expansion type to Disc and the pixel value of the expansion unit to 5.

[0138] (211) Use the Arithmetic function to edit A1=AB and remove the asphalt mortar and pores in the sample area to obtain the aggregate area;

[0139] (212) Use the Fill small spots function to remove the holes inside the aggregate, set the fill unit to Disc, and adjust the pixel value of the fill unit to 5000 to obtain the complete aggregate area A2.

[0140] (213) Use the Image gradient function, select Canny Deriche as the gradient type, and set the smoothing coefficient to 30 to extract the brightness difference boundary I from the grayscale data in the aggregate area A2;

[0141] (214) Repeatedly use the Adaptive Thresholding algorithm to extract particles of different components in region A2 to ensure that the extraction accuracy does not completely cover the target particle area. If necessary, the lasso-Autotrace edges tool in AVIZO can be used for manual correction. After correction, the steel slag aggregate region C, RAP aggregate region D and new aggregate region E are initially obtained.

[0142] (215) Using the Maker-based watershed function, the brightness gradient boundary I extracted by the Image gradient function is used as the marker watershed algorithm to extend the boundaries of each phase. After extending regions C, D, and E, the complete steel slag aggregate region C1, RAP aggregate region D1, and quartzite aggregate region E1 are obtained; the segmented image is shown below. Figure 3 , Figure 4 , Figure 5 ;

[0143] (216) The three types of coarse aggregate image data segmented above are stored as RAP, steel slag, and new aggregate datasets, respectively. Specific data are calculated based on the corresponding microstructure parameters, including: average number of contact points (x1), average shape factor (x2), aggregate shape factor variance (x3), average minimum spacing (x4), total minimum spacing (x5), aggregate minimum spacing variance (x6), orientation angle variance (x7), and orientation angle mean (x8). These microstructure parameters are defined as follows:

[0144] Average number of contact points: The contact threshold is set to 0.54 mm. The total number of contact points of all coarse aggregates is determined by a window search algorithm, and then the average number of contact points is obtained.

[0145] Shape factor: Shape factor = (Crofton perimeter)² / (4π × aggregate area) where the formula for calculating Crofton perimeter is as follows:

[0146] Crofton perimeter=(D(X,0°)+D(X,45°)+D(X,90°)+D(X,135°)) / 4(7)

[0147] D(X,a)=∫N a (X)dx (8)

[0148] In the formula: D(X,a) represents the number of intersections between all parallel lines in direction a and the aggregate boundary.

[0149] N a The number of intersections between the boundary and the straight line in direction a.

[0150] Minimum aggregate spacing: Minimum aggregate spacing is the shortest distance from any point on the edge of the target aggregate to any point on the edge of its nearest aggregate.

[0151] Orientation angle: The aggregate orientation angle is the angle between the aggregate's moment of inertia vector and the coordinate axis. The angle of the moment of inertia vector is in the range of [-90°, 90°]. The formula for calculating the moment of inertia is as follows:

[0152] I x =∫y 2dA (9)

[0153] I y =∫x 2 dA (10)

[0154] The calculated moment of inertia vector is as follows:

[0155] The microstructure parameters calculated based on the above definition are shown in Table 4;

[0156] Table 4:

[0157]

[0158] (3) Use the ridge and lasso functions in MATLAB to establish ridge regression and lasso regression models for microstructural parameters and road performance parameters;

[0159] (301) Use MATLAB to standardize the data of microstructure parameters and road performance parameters respectively. Set the eight microstructure parameters as independent variables x1 to x8, and set the four road performance parameters of dynamic stability, freeze-thaw splitting residual strength ratio, water immersion Marshall residual stability, and low temperature fracture failure strain as different dependent variables. Scale the independent and dependent variables so that the mean of each group of data is 0 and the standard deviation is 1.

[0160] (302) Set the regularization terms of the ridge and lasso functions in MATLAB to 0.1, and calculate the four sets of regression equations corresponding to the four road performance parameters, as follows:

[0161] Dynamic stability:

[0162] Ridge(y)=-0.2273x1-0.0977x2+0.1206x3+0.2884x4-0.1633x5+0.4402x6+0.1679x7-0.0295x8

[0163] Lasso(y)=-0.2686+0.1908x4-0.1075x5+0.4834x6

[0164] Freeze-thaw splitting residual strength ratio:

[0165] Ridge(y)=-0.0810x1+0.0867x2-0.4370x3-0.1340x4-0.2741x5-0.6080x6+0.0315x7-0.0422x8

[0166] Lasso(y)=-0.3302x3-0.1981x5-0.6203x6

[0167] Water immersion Marshall residue stability

[0168] Ridge(y)=0.1374x1-0.0206x2-0.3059x3-0.2555x4+0.1072x5-0.5841x6-0.2290x7-0.0049x8

[0169] Lasso(y) = 0.2183x5 - 0.8101x6

[0170] Destructive strain

[0171] Ridge(y)=-0.0946x1-0.1412x2-0.4638x3-0.0912x4-0.0746x5-0.2088x6+0.2935x7+0.3042x8

[0172] Lasso(y)=-0.6468x3+0.1871x7+0.2041x8

[0173] (4) Input the microstructural parameters into the ridge regression equation and the lasso regression equation to predict the road performance parameters.

[0174] (401) Prepare other RAP steel slag recycled asphalt mixture specimens that need to be tested for road performance. Obtain cross-sectional images of the mixture and calculate the corresponding microstructure parameters using the method in step (2). See Table 5.

[0175] Table 5:

[0176] Average shape factor 1.77 Shape factor variance 0.19 Average minimum spacing 0.51 Total minimum spacing 40.89 Minimum Spacing Variance 0.17 Direction angle variance -2.74 Mean of direction angle 2213.84

[0177] (402) Standardize the microstructure parameter data and input them as independent variables into the ridge regression and lasso regression equations obtained in step (3). Take the average value of the two regression equations to calculate the road performance parameters required for the mixture, as shown in Table 6.

[0178] Table 6:

[0179]

[0180]

[0181] The above embodiments are merely illustrative of the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made to the technical solutions based on the technical concept proposed in this invention shall fall within the scope of protection of this invention.

Claims

1. A method for predicting the performance of RAP steel slag asphalt mixtures based on microstructural parameters, characterized in that: The following steps are included: (1) Design the mix proportion of RAP steel slag recycled asphalt mixture and test its road performance parameters: dynamic stability, freeze-thaw splitting residual strength ratio, water immersion Marshall residual stability, and low temperature fracture failure strain; (2) Perform image processing through brightness gradient and marker watershed algorithm to calculate the microstructure parameters: average number of contact points, average shape factor, aggregate shape factor variance, average minimum spacing, total minimum spacing, aggregate minimum spacing variance, orientation angle variance, and orientation angle mean. (3) Use the ridge and lasso functions in MATLAB to establish ridge regression and lasso regression models for mesostructural parameters and road performance parameters; (4) Input the mesostructural parameters into the ridge regression equation and lasso regression equation to predict the road performance parameters; In step (2), image processing is performed by the brightness gradient and marker watershed algorithm to calculate the mesostructural parameters, including: (201) Using quartzite to replace basalt as the new aggregate in RAP steel slag recycled asphalt mixture, four standard Marshall specimens of each RAP steel slag recycled asphalt mixture are prepared; (202) Using a high-precision cutting machine at 1.2 Marshall specimens were cut at intervals of cm to obtain complete disc-shaped specimens; (203) the cut surfaces of the specimens were cleaned with anhydrous ethanol and a brush, and then the cross-sectional images of the specimens were taken horizontally using a high-definition digital camera. Twenty clear and uniformly lit photos were selected for each type of RAP steel slag recycled asphalt mixture; (204) the original slice images were converted into 8-bit grayscale images using the rgb2gray function, and then the actual size of the pixels was determined according to the image resolution and sample size; (205) the Remove Outliers function was used to adjust the radius and threshold to remove overly bright and dark areas in the grayscale image; (206) the Median Filter algorithm was used to set a certain radius to eliminate random noise in the image and obtain a clearer grayscale image; (207) the Adaptive histogramequalization algorithm was used to enhance the contrast between aggregates and asphalt mortar, and the equalized image was used for the segmentation of each phase of the recycled asphalt mixture; (208) the Adaptive Thresholding algorithm was used to enhance the contrast between aggregates and asphalt mortar. Algorithm, set gray threshold to select sample area in image to obtain sample area A, initially remove non-sample area interference to image segmentation; (209) According to gray histogram of equalized image, determine gray range of asphalt mortar, and again use Adaptive Thresholding algorithm to select area in sample area A below gray value of asphalt mortar to obtain asphalt mortar and pore area B; (210) Use Dilation function to dilate asphalt mortar and pore area B, set dilation type to Disc, select specific precision value according to image processing effect; (211) Use Arithmetic function to edit A1=AB, remove asphalt mortar and pore in sample area to obtain aggregate area; (212) Use Fill small spots function to remove pores inside aggregate, set fill unit to Disc, adjust pixel size of fill unit according to image processing effect to obtain complete aggregate area A2; (213) Use Image gradient function, after setting gradient type and smoothing coefficient, extract brightness difference boundary I in gray data of aggregate area A2;(214) The Adaptive Thresholding algorithm was repeatedly used in region A2 to extract particles of different components to ensure that the extraction accuracy did not completely cover the target particle area. The lasso-Auto traceedges tool in AVIZO was used for manual correction. After correction, the steel slag aggregate region C, RAP aggregate region D and new aggregate region E were initially obtained. (215) The Maker-based watershed function was used to extend the boundary of each phase by using the brightness gradient boundary I extracted by the Image gradient function as the marker watershed algorithm. After extending regions C, D and E, the complete steel slag aggregate region C1, RAP aggregate region D1 and quartzite aggregate region E1 were obtained. (216) The three coarse aggregate image data of the above segmentation were stored as RAP, steel slag and new aggregate datasets respectively, and the specific data were calculated according to the corresponding microstructure parameters, including: the average number of contact points. Average shape factor aggregate shape factor variance Average minimum spacing Total minimum spacing Minimum aggregate spacing variance Direction angle variance Mean of direction angle The definitions of these microstructural parameters are as follows: Average number of contact points: The contact threshold is set to 0.54 mm. The total number of contact points of all coarse aggregates is determined by a window search algorithm, and then the average number of contact points is obtained; Shape factor: Shape factor = ( ) 2 / (4π × aggregate area) where The calculation formula is as follows: （7） In the formula: :all The number of intersections between parallel lines and aggregate boundaries; Boundary and Number of intersections of directional lines; Minimum aggregate spacing: The minimum aggregate spacing is the shortest distance from any point on the edge of the target aggregate to any point on the edge of its nearest aggregate; Orientation angle: The aggregate orientation angle is the angle between the aggregate's moment of inertia vector and the coordinate axis. The angle of the moment of inertia vector is in the range of [-90°, 90°]. The formula for calculating the moment of inertia is as follows: The calculated moment of inertia vector is as follows: .

2. The method for predicting the performance of RAP steel slag asphalt mixture based on microstructural parameters according to claim 1, characterized in that, In step (1), the mix design and road performance test of RAP steel slag recycled asphalt mixture include: (101) testing the performance of raw materials to ensure that they meet the specifications; analyzing the gradation of RAP and adjusting the aggregate gradation curve; (102) conducting standard Marshall compaction tests, designing 5 samples with different asphalt contents, and determining the optimal asphalt content of the selected RAP steel slag recycled asphalt mixture through Marshall tests; (103) conducting road performance tests of RAP steel slag recycled asphalt mixture, testing road performance parameters, including: dynamic stability, freeze-thaw splitting residual strength ratio, water immersion Marshall residual stability, and low-temperature fracture failure strain, calculated as follows: Dynamic stability: Where: DS—dynamic stability of asphalt mixture, cycles / mm; d1, d2—deformation at times t1 and t2 respectively, mm; C1—testing machine type coefficient, taken as 1.0; C2—specimen coefficient, 1.0 for laboratory-prepared specimens with a width of 300mm; N—reciprocating rolling speed of the testing wheel; freeze-thaw splitting residual strength ratio: Where: RT1, RT2—splitting tensile strength of a single specimen in the non-freeze-thaw group and the freeze-thaw group, MPa; PT1, PT2—maximum experimental load value of a single specimen in the non-freeze-thaw group and the freeze-thaw group, N; h1, h2—height of each specimen in the non-freeze-thaw group and the freeze-thaw group, mm; RT1, RT2—average splitting tensile strength of effective specimens in the non-freeze-thaw group and the freeze-thaw group, MPa; TSR—freeze-thaw splitting strength ratio, %; immersion Marshall residual stability: Where: MS0—residual stability of the Marshall specimen after immersion in water, %; MS—stability of the Marshall specimen after immersion in water for 30 min, kN; MS1—stability of the Marshall specimen after immersion in water for 48 h, kN; low-temperature fracture strain: In the formula: —The maximum flexural tensile strain at specimen failure h—height of the cross section of the specimen, mm; L—span of the specimen, mm; d—mid-span deflection at specimen failure, mm.

3. The method for predicting the performance of RAP steel slag asphalt mixture based on microstructural parameters according to claim 2, characterized in that, In step (3), the establishment of ridge regression and lasso regression models for microstructural parameters and road performance parameters using the ridge and lasso functions in MATLAB includes: (301) Standardizing the data of microstructural parameters and road performance parameters using MATLAB, and setting the eight microstructural parameters as independent variables. The four road performance parameters, namely dynamic stability, freeze-thaw splitting residual strength ratio, water immersion Marshall residual stability, and low-temperature fracture failure strain, are set as different dependent variables. The independent and dependent variables are scaled so that the average value of each group of data is 0 and the standard deviation is 1. (302) The ridge and lasso functions in MATLAB are used and the regularization term is set to 0.1 to calculate the four sets of regression equations corresponding to the four road performance parameters.

4. The method for predicting the performance of RAP steel slag asphalt mixture based on microstructural parameters according to claim 3, characterized in that, In step (4), the microstructure parameters of RAP steel slag recycled asphalt mixture are input into the ridge regression and lasso regression equations to obtain the corresponding road performance parameters, including: (401) preparing other RAP steel slag recycled asphalt mixture specimens that need to be tested for road performance, obtaining cross-sectional images of the mixture through the method in step (2) and calculating the corresponding microstructure parameters; (402) standardizing the microstructure parameter data, inputting it as an independent variable into the ridge regression and lasso regression equations obtained in step (3), and taking the average value of the two regression equations to calculate the required road performance parameters of the mixture.

Images