Optimization of Composite Grouting Material Proportioning Based on Regression Analysis Method

Abstract

In order to improve the impermeability of ordinary “cement–water glass” double-liquid grouting material, this study prepares a composite grouting material with sodium bentonite, cement, and water glass as the main materials, water-reducing agent, and early strength agent as activators. In this study, orthogonal tests were conducted using an L₁₆(4⁵) orthogonal table and the test results were analyzed using regression analysis. Using bentonite instead of partial cement, the experimental factors of the bentonite ratio, glass slurry ratio, water-reducing agent admixture, and early strength agent admixture were used to study the laws of single-factor and two-factor interactions on the initial apparent viscosity, setting time, and compressive strength of composite grouting materials. After obtaining the optimal ratio of composite grouting materials, the mechanism of formation of material strength and seepage resistance properties was revealed. The results show that the mass ratio of bentonite to cement and the mass ratio of water glass to slurry have a large effect on the material properties, and there is an interactive effect of several factors at the same time. The optimal ratio of the material is 28.42% water, 36.58% cement, 6.20% bentonite, 0.04% water-reducing agent, 0.9% early strength agent, and 27.85% water glass. The microstructural analysis shows that the denseness of the stone body structure is enhanced, and thus the strength and impermeability of the material are improved. This composite grouting material improves injectability and seepage resistance and is more suitable for underpinning grouting of weakly cemented soft rocks, but the ratio optimization method and model used in this study can be applied to other materials.

1. Introduction

Water breakout in coal mines is the second most serious production accident after coal mine gas [1], and most water damage accidents in coal mines are safety accidents with serious casualties, huge losses, and bad nature [2]. Grouting technology can ensure sustainable mining of coal mines and avoid major water damage accidents. Therefore, improving the impermeability of materials is the key to maintaining the sustainable development of coal mines.

The 11,303 working face of Hongqingliang Coal Mine No. 3-1 is affected by the poor lithology of the bottom slab and the abundance of water in the bottom slab, which causes serious bottom drumming and sudden water in the bottom slab in the process of digging. This phenomenon is due to the sandy mudstone, which is a kind of weakly cemented soft rock with complex conditions of easy fracture, easy mudification, and easy hydrolysis, and the sudden water situation is encountered in the deep mining process. The currently used grouting materials are no longer suitable for this situation. In order to ensure sustainable mining in coal mines, there is an urgent need to improve the impermeability of grouting materials.

Grouting is a method of injecting a solidifiable liquid into fractured surrounding rock using physical or chemical methods to improve the mechanical properties and integrity of the surrounding rock, and grouting technology has become an effective way to reinforce fractured surrounding rock in roadways [3]. Among them, grouting material is the key component of grouting technology. Grouting materials are generally divided into three main categories, and they are inorganic materials based on cement-based grouting materials, organic materials based on polymer materials, and inorganic–organic composite grouting materials [4]. Cement-based grouting materials are the most widely used grouting materials in coal mining enterprises because they have the advantages of high strength, good permanence, no pollution, abundant sources, and low price [5,6,7,8,9]. Single cement slurry has problems such as weak water resistance, poor injectability, and low early strength. Compared with single cement slurry, cement–water glass double-liquid grouting material improves the early strength and enhances the resistance to dynamic water. However, when the groundwater action is strong, the effect of grout plugging and reinforcement is poor, and it is difficult to achieve a better grouting and reinforcement effect [10]. Therefore, it is important to improve the grouting materials.

Bentonite has good adsorption, swelling, and mud making properties. It was found through research that bentonite can fill the voids created by cement in the process of a reaction [11]. It can improve the water fixation capacity and stability of cement slurry [12], and can replace cement with little effect on the strength of the stone body [13]. However, the addition of bentonite will lead to low early strength of the slurry, water seepage will still occur when applied in actual projects, and the seepage resistance needs to be further improved. Due to the properties of bentonite such as good dispersion, small particle size, large specific surface area, and efficiency of incorporation into slurry, it can replace part of the cement in grouting materials and effectively improve the injectability of slurry and the impermeability of the stone body. Studies by Zhang G et al. [14,15] have shown that mixing a certain amount of clay into cement can effectively improve the stability and injectability of the slurry and increase the stone rate. Sheng Y et al. [16] found that bentonite can hydrate and swell in cement, plug the pores, and increase its impermeability, so that bentonite has the ability to improve the stability of the slurry and the resistance to deformation of the stone body. In order to manage the severe bottom bulge phenomenon caused by multiple water inrush situations in a weakly cemented soft rock (sandy mudstone) floor, it is optimal to mix bentonite, a barrier material, into the cement to improve its impermeability after considering various factors [17,18].

The composite slurry of bentonite, cement, and water glass has the advantages of a high stone rate, adjustable setting time, high early strength and compressive strength, low price, and no pollution compared to the cement–water glass double-liquid grouting material [19,20,21,22]. In particular, this material is a functional backfill material and has zero environmental impact. In order to study the composite grouting material formed by mixing bentonite, cement, water glass, and some admixtures suitable for a weakly cemented soft rock floor, the effects of the mass ratio of bentonite to cement (A), the mass ratio of water glass to slurry (B), the mass permillage of water-reducing agent to cement (C), and the mass percentage of the early strength agent to cement (D) on the viscosity (Y₁), setting time (Y₂), and 28d compressive strength (Y₃) of the grouting material were investigated by orthogonal experiments in this study. On this basis, the influence of each single-factor and multi-factor interaction on the viscosity, setting time, and 28d compressive strength of the grouting material was obtained by using regression model calculation and analysis, and the optimal material ratio was determined. This essay provides an effective method for the optimization of composite grouting material ratios.

2. Material Preparation and Experimental Protocol for Orthogonal Experiments

2.1. Reagents and Materials

2.1.1. Bentonite

Bentonite is a mineral clay with good bonding, water absorption, and mud making and swelling properties. The main component of bentonite is montmorillonite, which has the characteristics of a large specific surface area and strong water absorption capacity. Therefore, the appropriate amount of bentonite can improve the cement slurry’s ability to fix water and stability.

The bentonite used in this experiment is the sodium-based bentonite produced by Heilongjiang Donghai Bentonite Co. The physical properties of the bentonite are shown in

Table 1. Physical properties of bentonite.

Methylene Blue Index (g/100 g)	Colloid Index (%)	Expansion Multiple	Rate of Water Content (%)	Colloid Ratio (%)	Pulping Rate (m3/t)
18~30	66	>10	<10	100	≥16

, and the chemical properties are shown in

Table 2. Chemical composition of bentonite.

Composition	SiO2	Al2O3	TiO2	Fe2O3	MgO	CaO	K2O	Na2O	Ignition Loss
Content/%	63.66	16.64	0.59	4.93	1.69	0.88	1.37	4.93	10.79

.

The SEM and XRD analysis of bentonite is shown in Figure 1. The mineral composition of the bentonite obtained by XRD analysis is 50% quartz, 45.8% clay minerals, 1.9% potassium feldspar, and 2.3% plagioclase.

The distribution of bentonite particle size is shown in Figure 2. The analysis of Figure 2 shows that 85% of the bentonite particles are less than 20 μm, 95% of the particles are less than 40 μm, and the maximum particle size is 90 μm.

2.1.2. Cement

In this test, Swan brand P.O 42.5 ordinary silicate cement produced by Harbin Yatai Cement Company was used, and its chemical composition is shown in

Table 3. Chemical composition of cement.

Composition	CaO	SiO2	AL2O3	Fe2O3	SO3	MgO	F-CaO	Ignition Loss
Content/%	58.98	23.96	6.34	3.46	2.14	0.97	1.60	2.07

.

2.1.3. Water Glass

The setting time of slurry is adjusted by adjusting the amount of water glass added in the preparation of the cement slurry, and the setting time can be controlled between a few seconds and several minutes. Water glass is divided according to the chemical composition; the more common ones are sodium water glass, potassium water glass, lithium water glass, etc. Sodium water glass has the advantage of having a low cost and good effect, so it is more commonly used in grouting. The water glass in use was the liquid sodium water glass produced by Jingwen Pharmaceuticals, and the specific parameters are shown in

Table 4. Parameters of sodium water glass.

Baume Degree	Density	Modulus	pH Value
50° Bé	1.33	3.1	13

.

2.1.4. Water-Reducing Agent

The water-reducing agent can reduce the water consumption and improve the fluidity of the slurry while keeping the material slump basically unchanged. The water-reducing agent is polycarboxylic-acid-type high-efficiency water-reducing agent produced by a company in Suzhou, and its specific performance is shown in

Table 5. Water-reducing agent performance table.

Project	Standard	Result
Performance	Off-white or off-yellow powder	Qualified
Rate of water content (%)	3.0%	2.76
Fineness (%)	≤15.0	4.30
Fluidity of cement paste	≥240	240
PH	8.0 ± 1.0	8.46
Mortar water reduction rate (%)	≥20	21

.

2.1.5. Accelerating Admixture

The early strength agent can accelerate the cement hydration speed and improve the early strength of the material, but it can also play a certain role in water reduction. The early strength agent is selected as a compound early strength agent. This can have a multiple compound effect compared to a single type of early strength agent, which can better achieve the purpose of the experimental research.

2.2. Experimental Methods

Compound grouting materials are prepared with reference to DL/T 5150-2017 “Hydraulic Concrete Test Procedure”, and the viscosity test, setting time test, compressing test, and scanning electron microscope test are carried out [23].

(1)

Viscosity test

Viscosity is a physical quantity that measures the size of fluid viscosity, and its size directly affects the diffusion radius of the slurry and also determines the determination of parameters such as grouting pressure and flow rate [24]. In general, the initial viscosity of the slurry is required to be low, with good fluidity and high permeability. But under dynamic water conditions, the low-viscosity slurry is easily washed and diluted, which affects the grouting effect.

The test was conducted using the SNB-2 rotational viscometer manufactured by Shanghai Lunjie Electromechanical Instrument Co. (Shanghia, China).

(2)

Setting time test

According to GB/T 1346-2011 “standard consistency of cement water consumption, setting time, stability test method standard”, we recorded the setting time of the slurry [25].

(3)

Compressing test

We poured the prepared slurry into the prepared mold immediately. After waiting for the slurry to set, it was placed in a standard maintenance box for maintenance. After curing for 7 days and 28 days, the different ratios of the composite grouting material stone bodies were removed. According to GB/T 17671-2020 “cement sand strength test method (IOS method)” for compressive strength testing, the testing apparatus is a TAW-2000kN microcomputer-controlled electro-hydraulic servo three-axis testing machine [26].

(4)

Scanning electron microscopy test

Retests were conducted according to the optimal ratio obtained. After a certain age of maintenance of the composite grouting material nodules, they were cut into 4 mm thick slices. They were immersed in anhydrous ethanol to stop hydration and finally placed in an oven at 60 °C for 24 h to reach constant weight. The microstructure of the grouting material was characterized using a scanning electron microscope, JSW-5510LV, manufactured by Japan Electronics Co. (Tokyo, Japan).

2.3. Experimental Design

The main properties of the materials were investigated by orthogonal tests and single-factor and two-factor interaction analyses, mainly including the initial viscosity and setting time of the materials and the unconfined compressive strength and permeability pressure of the crystalline bodies, and finally the optimal ratio of the materials was determined. Afterwards, scanning electron microscopy tests were conducted to analyze the microstructure of the crystalline body and to investigate the mechanism of the hydration reaction of the cured slurry.

According to the results of a large number of basic tests in the early stage, the mass ratio of water to powder and granular material was specified as 0.65:1, and the water–solid ratio was determined as 0.65. In order to investigate the optimal ratio of composite grouting materials, the L₁₆(4⁵) orthogonal table was selected for the scheme design (

Table 6. Table of factor levels of orthogonal test.

Level (Code)	Factor
	A	B	C (‰)	D (%)
1	0.1	0.2	0.5	1
2	0.15	0.3	1	2
3	0.2	0.4	1.5	3
4	0.25	0.5	2	4

). The mass ratio of bentonite to cement (A), the mass ratio of water glass to slurry (B), the mass permillage of water-reducing agent to cement (C), and the mass percentage of early strength agent to cement (D) were studied as four main influencing factors. Four level values were taken for each factor, and the viscosity (Y₁), setting time (Y₂), and 28d compressive strength (Y₃) were selected as the experimental investigation indexes.

3. Results and Analysis

3.1. Test Results

According to the orthogonal test design, 16 sets of tests with different ratios of composite grouting materials were carried out, and the test results are shown in

Table 7. Test design and results.

Test Number	Factor LEVEL (code)				Test Result
	A	B	C	D	Y1 (mPa·s)	Y2 (s)	Y3 (MPa)
1#	1	1	1	1	615	17.62	11.71
2#	1	2	2	2	495	14.55	14.01
3#	1	3	3	3	456	38.01	20.71
4#	1	4	4	4	330	52.39	13.43
5#	2	1	2	3	569	25.12	11.36
6#	2	2	1	4	559	34.04	12.39
7#	2	3	4	1	432	32.89	10.53
8#	2	4	3	2	449	57.89	7.85
9#	3	1	3	4	580	22.53	14.32
10#	3	2	4	3	610	49.76	8.70
11#	3	3	1	2	664	34.27	9.50
12#	3	4	2	1	587	55.77	7.06
13#	4	1	4	2	780	41.26	7.01
14#	4	2	3	1	769	43.15	6.81
15#	4	3	2	4	754	56.15	10.32
16#	4	4	1	3	710	44.91	7.22

.

3.2. Analysis of Test Results

The DPS 7.05 mathematical and statistical software was applied to process the test data in Table 7, and the different models were fitted and analyzed by combining the same level sum, mean, extreme difference, and fitting error, and the quadratic model was found to be optimal. The regression models Y₁, Y₂, and Y₃ were obtained as in Equations (1)–(3), and the extreme difference analysis is shown in

Table 8. Analysis of extreme differences of test results.

Material Performance	Factor	Same Level				Average at the Same Level				Range
		K1	K2	K3	K4	${\overline{\mathbf{K}}}_1$	${\overline{\mathbf{K}}}_2$	${\overline{\mathbf{K}}}_3$	${\overline{\mathbf{K}}}_4$	$\mathbf{\Delta }\mathbf{K}$
Y1 (mPa·s)	A	1896	2009	2441	3013	474.0	502.3	610.3	753.3	279
	B	2544	2433	2306	2076	636.0	608.3	576.5	519	117
	C	2548	2405	2254	2152	637.0	601.3	563.5	538	99
	D	2403	2388	2345	2223	600.8	597	586.3	555.8	45
Y2 (s)	A	127	164	184	235	31.6	41.0	46.1	58.9	27.2
	B	247	206	147	111	61.6	51.4	36.8	27.7	33.9
	C	143	176	192	200	35.7	43.9	47.9	50.1	14.4
	D	163	174	178	195	40.9	43.5	44.5	48.8	7.9
Y3 (MPa)	A	60	42	40	31	15.0	10.5	9.9	7.8	7.1
	B	44	42	51	36	11.1	10.5	12.8	8.9	3.9
	C	41	43	50	40	10.2	10.7	12.4	9.9	2.5
	D	36	38	48	50	9.0	9.6	12.0	12.6	3.6

.

$${\mathrm{Y}}_1=634.17-48.62\mathrm{D}+17.06{\mathrm{A}}^2-7.97{\mathrm{B}}^2-12.96\mathrm{BC}+13.62\mathrm{BD}$$

(1)

$${(\mathrm{R}}^2=0.99,\mathrm{F}=73.08,\mathrm{P}=0.0001)$$

$${\mathrm{Y}}_2=3.75+18.62\mathrm{A}-10.92\mathrm{D}-6.48{\mathrm{A}}^2-3.29{\mathrm{C}}^2+8.03\mathrm{AC}+6.22\mathrm{BD}$$

(2)

$${(\mathrm{R}}^2=0.97,\mathrm{F}=24.27,\mathrm{P}=0.001)$$

$${\mathrm{Y}}_3=10.39-12.98\mathrm{A}+9.01\mathrm{B}+5.20\mathrm{D}+2.61{\mathrm{A}}^2-1.01{\mathrm{B}}^2+0.32{\mathrm{C}}^2-1.02\mathrm{AC}-2.12\mathrm{BD}+0.40\mathrm{CD}$$

(3)

$${(\mathrm{R}}^2=0.95,\mathrm{F}=6.09,\mathrm{P}=0.02)$$

We chose the goodness-of-fit test (R² test) to assess the significance of the model. The decidable coefficient R² represents the degree of difference between the calculated and true values of the model, and its value of 1 means that the two are in perfect agreement. The p-value represents the degree of significance of each factor in the test in the model, and if p < 0.05, it means that the factor is significant [27,28].

The F-value of each regression model is much larger than the p-value, which proves that the regression model is statistically significant. The p-values were all less than 0.05, which proved that the above regression models were fitted significantly. To further verify the accuracy of the regression equations, the level value data from 16 sets of orthogonal test protocols were substituted into the fitted equations and the calculated values obtained were compared with the experimental values by the model data results. As can be seen from Figure 3, the relative errors between the calculated values of all models and the experimental values are mostly within ±10%, indicating that the calculated values of the regression models are close to the actual values, and the model fitting effect can be considered reasonable and consistent.

3.3. Analysis of the Influence of Various Single Factors on Material Properties

The effect of each single factor on the performance of composite grouting material was plotted by the results of regression model data analysis, which is shown in Figure 4. Bentonite can absorb 8 to 15 times its own volume of water and can swell up to 30 times its own volume. And bentonite can be dispersed in water into colloidal and suspended forms. Therefore, A has a relatively large effect on Y₁. The bentonite, water glass, water-reducing agent, and early strength agent can all promote the hydration reaction of cement to a certain extent and affect the setting speed of cement slurry. Therefore, A, B, C, and D can all have influence on Y₂. The hydration product of cement is the main factor affecting the strength of the material, so A has a more significant effect on Y₃.

From the analysis of Figure 4, the degree of influence of each single factor on the performance of composite grouting material under comprehensive conditions and the optimal ratio can be seen in

Table 9. The degree of influence of each single factor on the material performance and the optimal ratio.

Material Performance	Degree of Influence	Optimal Ratio
Y1	A > B > C > D	A1B4C4D4
Y2	B > A > C > D	A1B1C1D3
Y3	A > B > D > C	A1B3C3D4

.

3.4. Analysis of the Effect of Interactions on Material Performance

After fitting the results of the orthogonal tests by regression analysis, it was found that the results were influenced not only by a single factor but also by the interaction between the factors.

3.4.1. Analysis of the Effect Law of Interaction on Y₁

Figure 5a,b show the effect of B and C interaction on material viscosity for A = 0.175 and D = 2.5. As can be seen from the figures, the viscosity value decreases as B increases and the increase in C also decreases the viscosity value. It can be seen that the viscosity value decreases significantly with the simultaneous addition of water glass and water-reducing agent. This means that the smaller the viscosity coefficient, the smaller the flow resistance and the smoother the slurry. However, when either B or C are small and the other factor is increased, the viscosity value changes are not obvious, which means that the slurry is still viscous. After one of B or C is raised to a large enough amount, raising the other factor (even if the amount is small) will result in a significant reduction in the viscosity value, which indicates a more fluid slurry.

Figure 5c,d show the effect of B and D interaction on the material viscosity for A = 0.175 and C = 1.25. As can be seen from the graph, the viscosity value decreases significantly as B increases. And the increase in D also causes a slight decrease in the viscosity value. It can be seen that the early strength agent does not play a big role in it. Just adjusting B can control the general viscosity of the slurry.

3.4.2. Analysis of the Effect Law of Interaction on Y₂

Figure 6a,b show the effect of A and C interaction on the material setting time for B = 0.35 and D = 2.5. As can be seen from the figures, when C is small (<1.0‰), the setting time shows a parabolic-like trend of first increasing and then decreasing as A increases. When C is large (>1.0‰), the setting time increases and then slightly decreases as A increases, and a similar phenomenon occurs when A is around 0.15. When both factors are increased, it will make the setting time longer.

Figure 6c,d show the effect of the interaction of B and D on the setting time of the material for A = 0.175 and C = 1.25. From the figures, it can be seen that the addition of the early strength agent decreases the setting time in all cases where B is small enough (<0.28). All cases with enough B (>0.3) will make the condensation time grow, and all increases in B will make the condensation time grow. When B < 0.3, the condensation time will be controlled within 40 s, and the condensation time will be less and less as D grows and B decreases.

3.4.3. Analysis of the Effect Pattern of Interaction on Y₃

Figure 7a,b show the effect of A and C interaction on the 28d compressive strength of the material for B = 0.35 and D = 2.5. As can be seen from the figure, with the increase in A, the 28 d compressive strength of the material generally shows a trend of first decreasing and then increasing. When A is small (<0.15), the compressive strength gradually increases with the increase in C. When A is relatively large (>0.15), the compressive strength gradually decreases. When A = 0.1, C increases from 0.5‰ to 2‰ and the compressive strength increases from 15 MPa to 21 MPa, respectively, and the overall compressive strength is great.

Figure 7c,d show the effect of B and D interaction on the 28 d compressive strength of the material for A = 0.175 and C = 1.25. As can be seen from the figure, when B is small (<3.0), the compressive strength increases with the increase in C. When B is larger (>3.0), C increases and the compressive strength decreases. When C is small (<2.5‰), with the increase in B, the compressive strength shows a trend of first increasing and then decreasing, and the maximum does not exceed 9 MPa. When C is larger (>2.5‰), with the increase in B, the compressive strength gradually decreases.

Figure 7e,f show the effect of C and D interaction on the 28 d compressive strength of the material for A = 0.175 and B = 0.35. As can be seen from the figure, when D is small (<2.5%), with the increase in C, the compressive strength decreases and then increases slightly, and all are less than 9. When D is larger (>2.5%), with the increase in C, the compressive strength gradually increases; generally, the higher the D, the greater the compressive strength. When D = 4.0% and C = 2.0‰, the compressive strength is 7.28 MPa at maximum. In addition, when C is small (<1.5‰), the change in compressive strength is not obvious. When C is larger (>1.5‰), the compressive strength increases more obviously, and its growth rate increases with the increase in C.

3.5. Optimization of Results and Applicability Analysis of Regression Analysis Method

In order to obtain the optimal ratio of composite grouting materials with comprehensive performance, multi-factor and multi-response co-optimization was performed using Matlab. According to the mechanical properties of sandy mudstone and actual engineering experience, the boundary conditions are set to 500 mPa-s ≤ Y₁ ≤ 600 mPa-s, 50 s ≤ Y₂ ≤ 60 s, Y₃ = max, and Y₃ ≥ 10 MPa. The regression analysis results and boundary conditions were substituted into the Matlab genetic algorithm and evolved for 500 iterations to obtain 15 optimized ratios, as shown in

Table 10. Optimized ratio (This “*” refers to the reassigned group number to distinguish it from the experimental group in the table above).

Number	Factor Level (Code)				Y1	Y2	Y3
	A	B	C	D
1 *	2.13	2.28	2.27	2.26	541.39	50.22	10.29
2 *	1.99	2.65	2.26	2.83	533.07	50.28	10.07
3 *	1.72	3.00	1.61	2.87	527.96	52.44	10.27
4 *	1.67	3.00	1.63	2.74	525.47	51.07	10.62
5 *	1.82	2.97	2.13	2.36	519.03	50.17	10.19
6 *	2.01	2.65	2.15	2.97	535.95	50.89	10.04
7 *	1.83	2.81	2.09	2.81	522.96	50.78	10.47
8 *	2.00	2.74	2.68	2.64	517.14	50.42	10.08
9 *	1.74	2.92	1.33	2.99	541.22	50.96	10.23
10 *	1.87	2.72	2.09	2.92	527.36	50.56	10.45
11 *	2.31	2.50	2.23	2.27	534.39	50.52	10.12
12 *	2.03	2.26	2.10	2.13	523.13	52.81	10.49
13 *	2.14	2.05	2.69	1.76	570.06	52.81	10.49
14 *	2.26	1.93	2.26	2.47	513.89	50.83	11.29
15 *	2.16	2.37	2.32	2.34	525.60	51.03	10.11

.

Considering the validity of the extreme difference analysis model, a multi-objective optimization of the optimized ratios was performed, and the optimal ratios were obtained as number 14*. We convert this into the ratio of A = 0.1695, B = 0.386, C = 1.13, and D = 2.47, which means 28.42% water, 36.58% cement, 6.20% bentonite, 0.04% water-reducing agent, 0.9% early strength agent, and 27.85% water glass.

Verification tests were conducted for the optimal ratio, and the errors between the measured and predicted values are detailed in

Table 11. Experimental results and prediction model for optimal proportional design.

Material Performance	Predicted Value	Actual Value	Error Rate (%)
Y1 (mPa·s)	525.6	518	1.47
Y2 (s)	51.03	56	−0.09
Y3 (MPa)	10.11	11.1	−0.09

. From Table 11, it is clear that the error values of measured and predicted values of viscosity, setting time, and 28d compressive strength of composite grouting materials are less than 2%, which indicates that the measured and predicted values are in good agreement and the model accuracy is high.

Finally, a comparative analysis of the cost and main technical properties of several currently common types of grouting materials was conducted, and the results are detailed in

Table 12. Comparison of the performance of grouting materials prepared with different slurries.

Material Performance	Cement Slurry [29]	Cement-Clay Slurry [29]	Cement-Water Glass Slurry [30]	Composite Grouting Material
7 d compressive strength (MPa)	2.30	1.16	13.25	8.3
28 d compressive strength (MPa)	9.00	2.07	17.75	11.1
Permeability pressure (MPa)	0.8	0.3	0.8	1.5
Cost (CNY/t)	400	148	600	280

. From Table 12, it can be seen that the composite grouting material can basically meet the needs of weakly cemented soft rock (sandy mudstone) substrate treatment in both early and late strength, and especially the seepage resistance has a greater advantage over ordinary cement-based grouting material. However, the slurry performance index requirements or ratios described in this study do not meet all possible grouting and water-plugging reinforcement conditions. It is recommended that different boundary conditions be set according to specific engineering characteristics while referring to the law of influence of various factors on slurry performance to determine the appropriate ratio for the actual working conditions.

4. Sample SEM Characterization

In order to investigate the principle of improving the compressive strength and seepage resistance of this composite grouting material, the composite grouting material stone body samples were analyzed.

The surface morphology of the composite grouting material stone body at the age of 7 d and 28 d is shown in Figure 8. Flocculent C-S-H gels, needle-like AFt crystals, interlaced laminated calcium hydroxide crystals CH, hydrated calcium aluminate C-A-H10 gels, and unreacted bentonite particles are observed in the crystalline bodies in Figure 8. Cement and water glass can sustain hydration and polymerization at different ages, while the ionized Na⁺ in bentonite can promote the hydration and curing effect of cement. With time, it can be observed in Figure 8b that the bentonite particles fill in between the crystalline skeleton, which can make the structure more dense and form a more solid system structure.

The strength of the composite grouting material comes from the cement, the hydration product C-S-H gels of water glass, AFt crystals, CH crystals, and C-A-H10 gels gathered to form crystals on the one hand. On the other hand, Ca²⁺ ions from the slurry and silicate ions are mutually attracted by electrostatic adsorption to form hydrated calcium silicate gels, which are attached to the surface of the particle plate. Eventually, a new reticular structure consisting of both amorphous silicate gel and crystalline clay minerals is formed. The cementation of multiple gels in the material can also consolidate the loose bentonite particles to produce a better micro-aggregate filling effect and improve the density of the material. By comparing the morphology of 7 d and 28 d samples in Figure 8, it can be seen that the number of hydrated calcium silicate and hydrated calcium aluminate particles has further increased, and the AFt generated in the early stage has basically been completely wrapped. The strength and the impermeability of the material are improved because the voids in the material basically disappear and the denseness of the structure is further enhanced.

Nair S et al. [31,32] studied that, in the presence of Ca(OH)₂, the specific surface area of AFt produced by the reaction is larger and swelling will occur more readily in the presence of water. Shi C et al. [33] showed that efficient water-reducing agents accelerate the nucleation of AFt but adsorb on the surface of AFt crystals and inhibit their growth. The preparation is mixed with a polycarboxylic acid high-efficiency water-reducing agent and early strength agent, thus avoiding the swelling damage of the composite grouting material when it encounters too much water during condensation and also improving the impermeability of the material to a certain extent.

5. Discussion

After several laboratory experiments and tests, this composite grouting material has been successfully applied in the 11,303 working face of the No. 3-1 coal seam in Hongqingliang coal mine, and it has achieved the ideal effect of pressure and seepage resistance. The engineering example proves that this material can be applied to weakly cemented soft rock.

The material has a lower strength than a typical cement–water glass slurry, but the permeability of the material is improved. However, this is only able to meet the engineering needs of weakly cemented soft rocks. In addition, the material has a denser crystalline structure and lower water absorption and swelling rate. In fact, we expect the grouting material for subgrade reinforcement and filling modification to have some swelling property. Some scholars have achieved effective compensation of microcracks produced by self-shrinkage and drying shrinkage of ground aggregates by incorporating modified bentonite, fly ash, nanoparticles, MgO, etc., and have also enhanced the compressive strength and durability of the material [34,35,36].

The SEM characterization analysis of this study can not fully explain the principle of the improved seepage resistance from the composite grouting material of bentonite, cement, and water glass system, and it is hoped that other scholars can systematically add to the principle of the improved seepage resistance from other aspects.

6. Conclusions

(1)

In this study, the results of the orthogonal experiments were analyzed based on regression analysis, and the primary and secondary relationships of the factors affecting the composite grouting material were obtained by single-factor analysis. It can be concluded from the primary and secondary relationships that the most important factors affecting the initial apparent viscosity and setting time are the mass ratio of bentonite to cement (A) and the mass ratio of water glass to slurry (B). The most important factors affecting the compressive strength are the mass ratio of bentonite to cement (A) and the mass permillage of water-reducing agent to cement (C).

(2)

Interaction effects on the performance of composite grouting materials are more complex. The following conclusions can be drawn from the multi-factor analysis. The amount of early strength agent has little effect on the viscosity, and the joint action of the water glass and water-reducing agent will make the slurry significantly smoother. When B < 0.3, the setting time will be controlled within 40 s, and with the addition of the early strength agent and the reduction in B, the setting time will become smaller and smaller. When C > 2.5‰ and B < 2.5, the compressive strength of the grouting cementing material is larger and greater than 10 MPa. Such a setting time and compressive strength interval is more suitable for industrial production.

(3)

Water glass and bentonite can promote the hydration of cement to a certain extent. The new mesh structure formed by the mixture of multiple substances participating in the hydration of cement is conducive to improving the compressive strength and impermeability of the material.