Method and system for predicting mixture ratio and strength of solidified soil based on SEM image analysis

Abstract

The present application relates to a solidified soil proportioning and strength prediction method and system based on SEM image analysis, and belongs to the technical field of solidified soil strength testing; comprising the following steps: S1: physical index determination of disturbed soil; S2: sample preparation and curing under multiple working conditions; S3: calculation of SEM image fractal dimension; S4: unconfined compressive strength test; S5: establishment of microstructure inversion model and macroscopic strength prediction model; S6: proportioning optimization and strength prediction: using the microstructure inversion model and the macroscopic strength prediction model to optimize the proportioning of the solidified soil or to predict the strength; the method of the present application establishes a quantitative relationship between the microstructure (fractal dimension) and the macroscopic strength through SEM image analysis, makes up for the shortcomings of the prior art, and provides a brand-new technical path for the resource utilization of disturbed soil.

Description

Technical Field

[0001] This invention relates to a method and system for predicting the mix proportion and strength of solidified soil based on SEM image analysis, belonging to the field of solidified soil strength testing technology. Background Technology

[0002] Solidification technology is widely used in soft soil foundation treatment and the resource utilization of excavated soil from engineering projects. With the rapid development of urban rail transit, high-rise buildings, and other engineering projects, a large amount of disturbed soil generated from foundation pit excavation and cast-in-place pile construction has become construction waste that urgently needs to be treated. Traditional treatment methods (such as off-site transportation and landfill) are not only costly but also prone to environmental pollution. Using cement solidification technology to transform disturbed soil into usable materials for engineering projects is an effective way to achieve resource utilization.

[0003] Currently, the mix design of solidified soil mainly relies on experience or single strength tests, which has the following shortcomings:

[0004] Micro-macro correlation is missing: the mix proportions are optimized only through macro strength tests, without considering the influence of microstructure (such as pore distribution and cementation state) on strength;

[0005] The testing cycle is long: the unconfined compressive strength test requires curing for more than 28 days, which is difficult to meet the needs of rapid construction in engineering.

[0006] Blindly optimizing the mix proportion: Too much cement content leads to wasted costs, while too little content results in insufficient strength.

[0007] In existing technologies, methods for predicting the strength of cement-stabilized soil are mainly divided into two categories:

[0008] Macro-level empirical models: For example, patent application publication number CN117690511A discloses a method for predicting the strength of cement-solidified silt. This method obtains an activity index by measuring physical indicators such as the clay content and plasticity index of the silt, performs preliminary strength prediction based on the activity index, and then corrects the model parameters through 4-6 sets of indoor tests to finally obtain the prediction formula. Although this method can achieve preliminary prediction, it still has limitations:

[0009] Microstructure correlation is missing: relying solely on macroscopic physical indicators of silt (such as clay content and plasticity index) does not reveal the mechanism by which microstructure affects strength;

[0010] The testing cycle is still relatively long: the pre-test correction phase requires 4 to 6 sets of indoor tests, and the curing cycle is long, which makes it difficult to meet the needs of rapid construction of the project.

[0011] Limited applicability: It is mainly applicable to silty soils and is not suitable for other types of disturbed soils (such as silty clay).

[0012] Microstructure analysis methods: The microstructure of solidified soil is analyzed by techniques such as scanning electron microscopy (SEM), backscattered electron imaging (BSE), and X-ray diffraction (XRD). However, existing studies are mostly at the level of qualitative description and lack quantitative correlation models between microstructure parameters and macroscopic strength.

[0013] Therefore, there is an urgent need to establish a method for optimizing the mix proportion of solidified soil that combines microstructure analysis and macro strength prediction, so as to achieve efficient and accurate mix design and strength prediction. Summary of the Invention

[0014] The purpose of this invention is to provide a method and system for predicting the mix proportion and strength of solidified soil based on SEM image analysis. By establishing a quantitative relationship between microstructure (fractal dimension) and macroscopic strength through SEM image analysis, this invention overcomes the shortcomings of existing technologies and provides a new technical path for the resource utilization of disturbed soil.

[0015] The method for predicting the mix proportion and strength of solidified soil based on SEM image analysis as described in this invention includes the following steps:

[0016] S1: Determination of physical parameters of disturbed soil: Determination of basic physical parameters of disturbed soil selected at the construction site;

[0017] S2: Multi-condition sample preparation and curing: Prepare and cure solidified soil samples with different cement content C, moisture content M and curing age t.

[0018] S3: SEM image fractal dimension calculation: SEM images are obtained by scanning electron microscopy experiments on samples of various ages, and the fractal dimension D of the SEM images is calculated;

[0019] S4: Unconfined compressive strength test: Unconfined compressive strength test is performed on specimens at all ages;

[0020] S5: Model Establishment: Based on the fractal dimension D, moisture content M, curing age t, and unconfined compressive strength q of the solidified soil. u Based on experimental data, a microstructure inversion model and a macro-intensity prediction model were established;

[0021] S6: Mix proportion optimization and strength prediction: Optimize the mix proportion or predict the strength of the solidified soil using the microstructure inversion model and macro strength prediction model.

[0022] Preferably, the multi-condition sample preparation and curing in step S2 includes the following sub-steps:

[0023] S21: Select silty clay from the construction site after excavation of the foundation pit or drilling of cast-in-place piles, and determine its dry density. The moisture content w0, void ratio e0, and compression modulus Es were determined by drying the undisturbed soil after treatment.

[0024] S22: Add water to the predetermined moisture content using a spray method and let it stand;

[0025] S23: Add cement according to the cement dosage, mix evenly, and compact into a cylindrical sample;

[0026] S24: Cured in a constant temperature and humidity environment until the set age.

[0027] Preferably, in step S3, the fractal dimension D is calculated from the SEM image using box counting: The fractal dimension is calculated using the built-in FracLab program in the MATLAB toolbox by covering the SEM image with boxes of different scales. It is assumed that the area of ​​the SEM image is equal to a square with a side length of 1. Then, a square with a side length of 1 is used to cover the SEM image, and this square is scaled multiple times. Each scaling change causes the two-dimensional planar image to change from a single size to a larger size. Each time an identical small square is covered, the number of small squares containing white areas under different conditions is counted, and this number is recorded as follows: ;

[0028] .

[0029] Preferably, the expression for the microstructure inversion model in step S5 is:

[0030] ;

[0031] Where p0, p1, p2, p3, p4, p5, p6, p7, p8, and p9 are the polynomial partial coefficients obtained based on the fitting of experimental data; D is the fractal dimension, and q u The value represents the unconfined compressive strength, M represents the moisture content, and t represents the curing age.

[0032] Preferably, the expression for the macro-intensity prediction model in step S5 is:

[0033] ;

[0034] ;

[0035] ;

[0036] ;

[0037] Positive values ​​from the calculation results are taken as the predicted values ​​of unconfined compressive strength.

[0038] Preferably, in step S6, the measured micro-fractal dimension D, moisture content M, and curing age t are substituted into the macro-strength prediction model to predict the unconfined compressive strength, thereby evaluating the effect of the solidified soil mix design and reducing the number of unconfined compressive strength tests for solidified soil.

[0039] Preferably, in step S6, samples are taken from the solidified soil construction site, and the measured unconfined compressive strength, moisture content M, and curing age t are substituted into the microstructure inversion model to calculate the micro fractal dimension D, thereby quantitatively evaluating the solidification effect, uniformity, or damage state.

[0040] Preferably, for silty clay, when the optimized cement content C = 15%, moisture content M = 10%, and curing age t = 28 days, the unconfined compressive strength q u ≥7MPa, water stability residual strength retention rate ≥75%.

[0041] Preferably, the coefficient of determination R of the microstructure inversion model and the macro intensity prediction model is... 2 Defined as:

[0042] ;

[0043] ;

[0044] ;

[0045] Where SST is the total sum of squares, SSE is the residual sum of squares, and q u,meas q represents the measured value of unconfined compressive strength. u,pred This is the predicted value of the unconfined compressive strength. This represents the average value of the measured unconfined compressive strength. To ensure prediction accuracy, R... 2 ≥0.9.

[0046] The SEM image analysis-based cement-stabilized soil mix proportion and strength prediction system of the present invention, applied to the SEM image analysis-based cement-stabilized soil mix proportion and strength prediction method, includes:

[0047] The sample preparation module is used to prepare solidified soil samples with different cement content, moisture content, and curing age.

[0048] The data acquisition module is used to acquire SEM images of the specimen and calculate the fractal dimension D, as well as the unconfined compressive strength q of the specimen. u ;

[0049] The model building module establishes a two-way prediction model between the unconfined compressive strength qu of solidified soil and its fractal dimension D, moisture content M, and curing age t. This model includes a macroscopic strength prediction model and a microscopic structure inversion model. It uses microscopic features to predict macroscopic performance, or vice versa. It quantitatively evaluates the microscopic solidification effect, uniformity, or damage state through macroscopic performance.

[0050] The application module is used to optimize the mix proportions or predict the strength of solidified soil using the macroscopic strength prediction model and the microstructure inversion model.

[0051] Compared with existing technologies, the present invention provides a method and system for predicting the mix proportion and strength of solidified soil based on SEM image analysis, which has the following advantages:

[0052] 1. Breakthrough establishment of micro-macro linkages

[0053] This invention extracts fractal dimension through scanning electron microscopy (SEM) image analysis, establishing a quantitative relationship model between the microstructure (pore distribution, cementation state) and macroscopic compressive strength of solidified soil, revealing the essential laws governing strength formation at the microscopic mechanism level. It addresses the shortcomings of existing technologies that rely solely on macroscopic physical indicators of silt (such as plasticity index and clay content) for prediction, neglecting the influence of microstructure on strength and failing to explain the intrinsic reasons for strength differences. This invention fills this key technological gap.

[0054] 2. Significantly shortened testing cycle

[0055] This invention utilizes the fractal dimension of SEM images to predict the macroscopic strength of solidified soil 28 days in advance (e.g., after 7 days of curing), shortening the test cycle by approximately 75% and effectively meeting the needs of rapid construction in engineering projects.

[0056] 3. The accuracy of the optimized proportions has been significantly improved.

[0057] This invention optimizes the mix proportions based on microstructural parameters (fractal dimension), enabling precise control of key parameters such as cement content and moisture content, thus avoiding cost waste or insufficient strength caused by excessive or insufficient cement. Existing technologies rely on empirical corrections of macroscopic physical indicators, resulting in a high degree of blindness in mix proportion optimization and making it difficult to achieve precise control at the microscopic level.

[0058] 4. Wide range of applications

[0059] This invention is applicable to various types of disturbed soil (such as silty clay, muddy soil, etc.) generated in engineering projects such as foundation pit excavation and cast-in-place pile drilling. Different soil types can be adapted by adjusting the parameters of the microstructure model.

[0060] 5. Dual optimization of experimental efficiency and cost

[0061] This invention rapidly predicts macroscopic strength using microscopic fractal dimension, which can significantly reduce the number of traditional unconfined compressive strength tests and save testing time and manpower and material costs.

[0062] In summary, this invention achieves efficient, accurate, and widely applicable optimization of solidified soil mix design and strength prediction through the quantitative correlation between microstructure and macro strength, providing a novel technical path for the resource utilization of disturbed soil. Attached Figure Description

[0063] Figure 1 This is a flowchart of the overall method of the present invention;

[0064] Figure 2 This is a flowchart of the multi-condition sample preparation and curing process in this invention;

[0065] Figure 3 This is a comparison of the unconfined compressive strength of solidified soil with different cement and water contents after 7 days of curing in this invention.

[0066] Figure 4 This is a graph showing the relationship between the unconfined compressive strength and moisture content of solidified soil at different ages in this invention.

[0067] Figure 5 This is a graph showing the correlation between fractal dimension and unconfined compressive strength of solidified soil in this invention.

[0068] Figure 6 The images shown are scanning electron microscope (SEM) images at different magnifications for a water content of 20%, a cement content of 18%, and a curing period of 7 days, as described in this invention. (a) represents a scanning electron microscope image at a magnification of 500; (b) represents a scanning electron microscope image at a magnification of 2000; (c) represents a scanning electron microscope image at a magnification of 5000; and (d) represents a scanning electron microscope image at a magnification of 10000. Detailed Implementation

[0069] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0070] Example 1:

[0071] like Figure 1 As shown in the figure, this embodiment discloses a method for predicting the mix proportion and strength of solidified soil based on SEM image analysis, including the following steps:

[0072] S1: Determination of physical parameters of disturbed soil: Determination of basic physical parameters of disturbed soil selected at the construction site;

[0073] S2: Multi-condition sample preparation and curing: Prepare and cure solidified soil samples with different cement content C, moisture content M and curing age t.

[0074] The dry density of silty clay was measured at the construction site after excavation of the foundation pit or drilling of cast-in-place piles. Physical properties such as moisture content w0, void ratio e0, and compression modulus Es.

[0075] Solidified soil samples were prepared according to the standard "Standard for Geotechnical Testing Methods" (GB / T 50123-2019) to prepare Ø50 mm × 50 mm cylindrical samples. The undisturbed soil samples were first crushed and sieved, and then placed in an oven and dried at 80℃ for 12 hours.

[0076] After drying, the soil samples were divided into batches and water was added evenly to the predetermined moisture content using a spray method. They were then sealed and left to stand for 24 hours to ensure that the moisture was fully and evenly distributed.

[0077] Table 1 Mix design of solidified soil samples

[0078]

[0079] The cured soil samples were weighed and divided into 5 groups, each with a mass of 1400g. Cement content gradients were set at C = 6%, 9%, 12%, 15%, and 18%, and moisture content gradients at M = 10% and 20%. Based on the mix design in Table 1 (where C is the cement content and M is the moisture content), the corresponding mass of cement was added to each group, and the mixture was mechanically stirred until the soil particles, cement, and water were uniformly mixed.

[0080] like Figure 2 As shown, the multi-condition sample preparation and curing process includes the following steps: mixing, molding, compaction, demolding, and curing, specifically as follows: The well-mixed solidified soil is divided into groups, and 6 parallel samples are prepared for each group. The uniformly mixed soil material is weighed, with each sample weighing 200g, and then placed into the mold in sequence. A lower pad is placed at the bottom of the mold, and a pad is laid on the contact surface between the pad and the soil sample; after the soil sample is placed, an upper pad is placed (again with a pad), leaving a certain amount of compression space.

[0081] The upper pad was then pressed into the mold at a uniform speed using a compactor to compact the soil sample. Finally, the sample was demolded using a demolding machine to obtain a cylindrical test specimen with a diameter of 50 mm and a height of 50 mm.

[0082] The height of the sample was measured using vernier calipers and its weight was recorded. If the sample mass deviation exceeded 15% of the standard, it should be discarded and a new sample prepared. The sample was covered with plastic wrap and marked before being placed in a constant temperature and humidity curing room for curing. The temperature was controlled at 25℃ and the humidity at 90%. The curing ages (t) were set at 7 days, 14 days, and 28 days, respectively.

[0083] S3: SEM image fractal dimension calculation: First, SEM images were obtained by scanning electron microscopy experiments on samples of various ages. Then, the images to be processed were binarized using the built-in FracLab program in the MATLAB toolbox (black areas are pores and white areas are solids). The fractal dimension D of the SEM image was then calculated using the box counting method.

[0084] Box counting calculates the fractal dimension by covering a SEM image with boxes of different scales. Assuming the area of ​​the SEM image is equal to a square with a side length of 1, the image is covered with a box of side length 1, and this box is scaled multiple times according to a certain similarity ratio. Each scaling change makes the two-dimensional image... Each time the small squares are covered, the number of small squares containing white areas under different conditions can be counted and recorded as follows: .

[0085] .

[0086] S4: Unconfined compressive strength test: Unconfined compressive strength test is performed on specimens at all ages;

[0087] Scanning electron microscopy (SEM) and unconfined compressive strength tests were conducted on specimens at various ages to obtain SEM images and stress-strain curves of solidified soil specimens under different working conditions (C, M, t). The SEM images and unconfined compressive strength q were compared and analyzed. u The correlation.

[0088] With cement content C as the abscissa and unconfined compressive strength q as the ordinate u Using the vertical axis as the ordinate, we fit the evolution curves under different moisture content conditions in Table 1.

[0089] In this embodiment, the disturbed soil was taken from the excavation site of the foundation pit of a school dormitory building. The soil sample had a w0 of 53%. =1.56g / cm 3 e0 = 0.73, Es = 6.5 MPa. The specimens were prepared according to the method of the present invention, and the unconfined compressive strength test was completed. The results are as follows:

[0090] Figure 3 The results show that in the early stage of cement-stabilized soil curing (7 days), as the cement content increases from 6% to 18%, although the unconfined strength of cement-stabilized soil is approximately positively correlated with the cement content, the low moisture content condition is significantly unfavorable to the growth trend of compressive strength of cement-stabilized soil in the early stage of curing.

[0091] Figure 4The effect of moisture content on the average unconfined compressive strength of cement-stabilized soil at different curing ages was shown, with the cement content maintained at 15% throughout the experiment. As the curing age increased from 7 days to 28 days, the average peak stress of the 10% moisture content sample increased from 3.108 MPa to 7.201 MPa, an increase of 131.69%. In contrast, the average peak stress of the 20% moisture content sample increased from 3.411 MPa to 4.869 MPa, an increase of only 42.74%. This indicates that while low moisture content conditions are not conducive to promoting the compressive strength of cement-stabilized soil in the early stages of curing, it exhibits better compressive strength when the standard curing time (28 days) is reached.

[0092] Figure 5 The study revealed a significant negative correlation between the fractal dimension of SEM images of solidified soil and its unconfined compressive strength. Based on the MATLAB data analysis platform, the method of this invention proposes a polynomial mathematical model to link the unconfined compressive strength of solidified soil with the fractal dimension of its SEM images.

[0093] SEM scans at different magnifications were performed on solidified soil samples with a moisture content of 20%, a cement content of 18%, and a curing period of 7 days.

[0094] like Figure 6 As shown in (a), magnified 500 times: the whole structure presents a flocculent aggregate structure, with soil particles and cement hydration products wrapping each other to form large aggregates, but there are a few pores and microcracks between the aggregates;

[0095] like Figure 6 As shown in (b), magnified 2000 times: the morphological characteristics of cement hydration products can be clearly observed. Needle-shaped ettringite and amorphous villous CSH gel are attached to the surface of soil particles. The cementation characteristics are obvious in some areas, and many micropores are visible and unevenly distributed.

[0096] like Figure 6 As shown in (c), magnified 5000 times: it shows that the fiber interweaving between soil particles is relatively dense, and the cement hydration products have not completely wrapped the soil particles;

[0097] like Figure 6 As shown in (d), magnified 10,000 times: Under high magnification, a clear hydrated calcium silicate gel fiber network structure and needle-like ettringite can be seen, and there are nanoscale pores between the gel bodies.

[0098] Based on the results of the unconfined compressive strength test, SEM image analysis revealed that the presence of large pores is the main reason for the low initial strength. Cement hydration products only form cement in localized areas, failing to create a continuous spatial network structure and thus unable to effectively transfer stress. The fractal dimension D reflects the roughness of the pore surface. Studies show that for every 0.01 decrease in fractal dimension, the 28-day strength of the solidified soil can increase by approximately 8%, providing a quantitative indicator for precise mix design.

[0099] S5: Model Establishment: Based on the fractal dimension D, moisture content M, curing age t, and unconfined compressive strength q of the solidified soil. u Based on experimental data, a microstructure inversion model and a macro-intensity prediction model were established;

[0100] S6: Mix proportion optimization and strength prediction: Use microstructure inversion model and macro strength prediction model to optimize the mix proportion or predict the strength of solidified soil.

[0101] (1) Microstructure inversion model

[0102] Input: Unconfined compressive strength (q) u + Material state (M, t);

[0103] Output: Microscopic fractal dimension (D);

[0104] (1)

[0105] In formula (1), D represents the fractal dimension of the SEM image, M represents the moisture content, t represents the curing period (days), and q u This represents the unconfined compressive strength. p0, p1, p2, p3, p4, p5, p6, p7, p8, and p9 require fitting experimental data using the MATLAB data analysis platform to obtain specific values. In this embodiment, the first four columns of data in Table 2 are fitted to complete the parameter calibration. These partial coefficients themselves do not have independent, universal physical meanings; they are mainly used to quantitatively characterize the effect of each factor (M, D, t) on the unconfined compressive strength q. u The direction and sensitivity of the impact.

[0106] With q u Taking the partial coefficients as an example, p0 mainly limits the upper limit of the two-dimensional fractal dimension, p1 characterizes the direction of influence, and p2 and p3 are q u The effective weight, p2 (positive value), means that the longer the curing time t, the greater the unconfined compressive strength q. u The greater the influence on the fractal dimension D, the higher the water content M, and the greater the unconfined compressive strength q. uThe smaller the influence on the fractal dimension D, the more they reflect that age amplifies the effect of intensity, while moisture content weakens it. Similarly, p4 and p5 characterize the effects of M and t on D, respectively, p6 characterizes the combined effect of M and t, and p7, p8, and p9 mainly characterize nonlinear effects. Based on the MATLAB data analysis platform, according to Figure 5 Twenty sets of data from four operating conditions (M10-7d, M10-28d, M20-7d, M20-28d) were analyzed using multiple linear regression, with the coefficient of determination R0 being used as the basis for the results. 2 Find the optimal coefficient with an accuracy approaching 1.

[0107] In formula (1), after the coefficients are calibrated, q is calculated given D. u At that time, q u It is a predicted value; given q u When finding D, q u These are measured values.

[0108] Coefficient of determination R for microstructure inversion model and macro intensity prediction model 2 Defined as:

[0109] ;

[0110] ;

[0111] ;

[0112] Where SST is the total sum of squares, SSE is the residual sum of squares, and q u,meas q represents the measured value of unconfined compressive strength. u,pred This is the predicted value of the unconfined compressive strength. This represents the average value of the measured unconfined compressive strength. To ensure prediction accuracy, R... 2 ≥0.9.

[0113] When R 2 When the coefficient is 0.99, the coefficients generated by the MATLAB data analysis platform are: p0=1.918; p1=252.7985; p2=1.3785; p3=-9.5234; p4=0.1818; p5=0.0175; p6=-1.4749; p7=3.9162; p8=-2.8866; p9=1.7366. Substituting these values ​​into the formula yields:

[0114] (2)

[0115] (2) Macroeconomic intensity prediction model

[0116] Input: Microscopic fractal dimension (D) + material state (M, t);

[0117] Output: Unconfined compressive strength (q) u );

[0118] From formula (1), it can be seen that if we want to predict the unconfined compressive strength q of the solidified soil from D, M, and t, u (Positive value), a quadratic equation needs to be solved.

[0119] (3)

[0120] (4)

[0121] in, , ;

[0122] ;

[0123] Substituting into formula (2), we get: , ,

[0124] ;

[0125] Thus, the predicted value q of the unconfined compressive strength can be calculated. u .

[0126] Table 2 Model Data and Error Analysis Table

[0127]

[0128] Table 2 data demonstrates that the model has good predictive accuracy within a specific parameter range (e.g., moisture content 10%-20%, age 7-28 days), with a coefficient of determination R0. 2 The coefficient of determination (R²) is 0.99. The model was validated using 20 sets of experimental data, and the results show that the model has a small prediction error and a high R². 2 With a value ≥0.9, it exhibits high prediction accuracy. For out-of-boundary conditions, predictions can be made through interpolation or empirical correction. The two-way (macro-micro) closed loop formed by this model can guide the early-stage mix design optimization and also serve the rapid diagnosis and back-analysis of later-stage engineering quality.

[0129] Early prediction of intensity: By utilizing the fractal dimension of SEM images after 7 days of curing, the intensity at 28 days can be accurately predicted with a prediction error of <5%, shortening the test cycle by 75% compared to traditional methods;

[0130] Quantitative assessment of solidification effect: By inverting the fractal dimension D, the uniformity of solidified soil can be quantitatively assessed (standard deviation of D value < 0.02 indicates uniformity), providing a new method for engineering quality testing;

[0131] Optimized mix design: Based on microstructure parameters, the cement content can be optimized from the traditional empirical value of 18% to 15%, reducing costs by 20% while ensuring strength.

[0132] Example 2:

[0133] The SEM image analysis-based cement-stabilized soil mix proportion and strength prediction system of the present invention, applied to the SEM image analysis-based cement-stabilized soil mix proportion and strength prediction method described in Example 1, includes:

[0134] The sample preparation module is used to prepare solidified soil samples with different cement content, moisture content, and curing age.

[0135] The data acquisition module is used to acquire SEM images of the specimen and calculate the fractal dimension D, as well as the unconfined compressive strength q of the specimen. u ;

[0136] The model building module establishes the unconfined compressive strength q of the solidified soil. u A two-way prediction model relating fractal dimension D, moisture content M, and curing age t, including a macro-intensity prediction model and a micro-structure inversion model;

[0137] The application module is used to optimize the mix proportions or predict the strength of solidified soil using the macroscopic strength prediction model and the microstructure inversion model.

[0138] Forward prediction (design stage): For a newly designed mix proportion, by substituting the measured microscopic fractal dimension D, moisture content M, and curing age t into the macroscopic strength prediction model, the macroscopic unconfined compressive strength q can be quickly predicted. u This allows for the evaluation of the effectiveness of solidified soil mix design and reduces the need for unconfined compressive strength tests on solidified soil.

[0139] Inverse inversion (evaluation phase): By sampling the solidified soil at the construction site, and through the measured unconfined compressive strength q... u By substituting the moisture content M and curing age t into the microstructure inversion model, its D can be calculated, thereby quantitatively assessing its curing effect, uniformity, or damage state.

[0140] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for predicting the mixture and strength of a solidified soil based on SEM image analysis, characterized by, The method comprises the following steps: S1: disturbance soil physical index determination: basic physical parameter determination is performed on the disturbance soil selected from the construction site; S2: multi-working condition sample preparation and curing: solidified soil samples with different cement contents C, water contents M and curing ages t are prepared and cured; S3: SEM image fractal dimension calculation: a scanning electron microscope test is performed on the samples of each age to obtain SEM images, and the fractal dimension D of the SEM images is calculated; S4: unconfined compressive strength test: unconfined compressive strength tests are performed on the samples of each age; S5: Model establishment: based on the fractal dimension D, water content M, curing age t and unconfined compressive strength q of the solidified soil, the microstructure inversion model and the macroscopic strength prediction model are established; u experimental data of the unconfined compressive strength q of the solidified soil. S6: proportioning optimization and strength prediction: the microstructure inversion model and the macroscopic strength prediction model are used to optimize the proportioning of the solidified soil or predict the strength of the solidified soil; The expression of the microstructure inversion model in step S5 is: ； wherein p0, p1, p2, p3, p4, p5, p6, p7, p8, p9 are polynomial coefficients obtained by fitting experimental data; D is the fractal dimension, q u is the unconfined compressive strength, M is the moisture content, and t is the curing age. The expression of the macroscopic strength prediction model in step S5 is: ； ； ； ； The positive value in the calculation result is taken as the predicted value of the unconfined compressive strength.

2. The method of claim 1, wherein the method is characterized by: The multi-working condition sample preparation and curing in step S2 comprises the following sub-steps: S21: Select silty clay disturbed after foundation pit excavation or cast-in-place pile drilling at construction site, determine its dry density, moisture content w0, void ratio e0 and compression modulus Es, and dry the undisturbed soil after treatment ; S22: water is added to a predetermined water content by spraying and standing; S23: cement is added according to the cement content, and stirred uniformly, and then compacted into a cylindrical sample; S24: curing is performed in a constant temperature and humidity environment until the set age.

3. The method of claim 1, wherein the method is characterized by: The fractal dimension D in the step S3 is calculated by a box counting method: through the built-in program FracLab in the MATLAB toolbox, the fractal dimension is calculated by the method of covering the SEM image with square boxes of different scales, it is assumed that the area of the SEM image is equal to a square with a side length of 1, then the SEM image is covered with a square box with a side length of 1, and the square box is scaled multiple times, each scaling makes the two-dimensional plane image covered by the same small square boxes, and the number of small square boxes containing the white area under different conditions is counted as ; 。 4. The method of claim 1, wherein the method is characterized by: In step S6, the measured micro-fractal dimension D, water content M and curing age t are substituted into the macroscopic strength prediction model to predict the unconfined compressive strength, so as to evaluate the proportioning effect of the solidified soil and reduce the unconfined compressive strength test of the solidified soil.

5. The method of claim 1, wherein the method is characterized by: In step S6, the solidified soil is sampled from the construction site, the measured unconfined compressive strength, water content M and curing age t are substituted into the microstructure inversion model, and the micro-fractal dimension D is calculated inversely, so as to quantitatively evaluate the solidification effect, uniformity or damage state.

6. The method of claim 1, wherein the method is characterized by: For silty clay, when the optimized cement content C = 15%, moisture content M = 10%, curing age t = 28d, the unconfined compressive strength q u ≥ 7MPa, the water stability residual strength retention rate ≥ 75%.

7. The method of claim 1, wherein the method further comprises: determining a soil type of the soil sample based on the SEM image analysis. The microstructure inversion model and the macroscopic strength prediction model determine the coefficient R 2 is defined as: ； ； ； where SST is the total sum of squares, SSE is the residual sum of squares, q u,meas is the measured value of unconfined compressive strength, q u,pred is the predicted value of unconfined compressive strength, is the average value of measured values of unconfined compressive strength, R 2 ≥ 0.

9.

8. A system for predicting the mixture and strength of a solidified soil based on SEM image analysis, applied to the method for predicting the mixture and strength of a solidified soil based on SEM image analysis according to any one of claims 1 to 7, characterized in that, It comprises: A sample preparation module for preparing solidified soil samples with different cement contents, water contents and curing ages; a data acquisition module for acquiring SEM images of the sample and calculating the fractal dimension D, and acquiring the unconfined compressive strength q of the sample u ; The model establishing module establishes the unconfined compressive strength q of the solidified soil u A bidirectional prediction model between the fractal dimension D, the moisture content M, and the curing age t, including a macroscopic strength prediction model and a microscopic structure inversion model, is used to predict the macroscopic performance by the microscopic characteristics or vice versa; the macroscopic performance is used to quantitatively evaluate the microscopic solidification effect, uniformity, or damage state. An application module for using the macroscopic strength prediction model and the microstructure inversion model to optimize the proportioning of the solidified soil or predict the strength of the solidified soil.

Images