A design method and grouting material for deep coal-bearing strata

Abstract

This application discloses a design method and grouting material for deep coal-bearing strata, comprising the following steps: S1 Data acquisition based on material pre-tests, correlation analysis, and orthogonal experiments; S2 Data expansion based on orthogonal experimental data to establish a dataset; S3 Creating a prediction model based on the RBF algorithm; S4 Optimizing the prediction model based on the PSO-RBF algorithm; S5 Proportion design based on the RBF-PSO-EWM algorithm; S6 Obtaining the proportion of grouting material for deep coal-bearing strata. This application proposes a PSO-RBF-EWM multi-objective optimization grouting material proportion design method and applies this method to design the comprehensive optimal proportion of grouting material for deep coal-bearing strata. The grouting material with the obtained optimal proportion has excellent characteristics such as low water-cement ratio, green environmental protection, and economic efficiency, and is suitable for the reinforcement characteristics of deep coal-bearing strata.

Description

Technical Field

[0001] This application relates to the field of grouting material proportioning design technology, and in particular to a design method and grouting material for deep coal-bearing strata. Background Technology

[0002] Coal resources, as a crucial energy resource, are gradually being depleted due to socio-economic development and increasing energy demand, particularly shallow coal reserves. As coal mining depths increase, the geological conditions of deep coal-bearing strata become more severe and complex. The presence of high temperatures, high stress, and high permeability pressure makes the surrounding rock of roadways prone to severe deformation and damage under mining disturbances, while also making support more difficult. These problems seriously impact the long-term stability of roadways. Grouting, as one of the main methods of roadway reinforcement, involves mixing various materials and solid waste in specific proportions to improve grouting material performance and reduce costs. In recent years, with the development of computer and artificial intelligence technologies, machine learning has been widely applied in civil engineering due to its advantages in solving complex problems such as multi-factor nonlinearity, massive data volumes, and multi-objective optimization in engineering, while also overcoming the shortcomings of traditional linear and nonlinear regression models.

[0003] Currently, cement-based materials are commonly used for grouting reinforcement in coal mines. However, encountering unfavorable conditions such as aquifers, fracture zones, and cavities during the grouting process increases the demand for grouting materials and consequently, reinforcement costs. Furthermore, cement production consumes significant resources and emits large amounts of greenhouse gases, polluting the environment. Traditional grouting materials require a higher water-cement ratio to maintain high fluidity, but a high water-cement ratio can lead to a significant reduction in strength. Additionally, grout with a high water-cement ratio is prone to water separation, and the soft rock surrounding the tunnel softens upon contact with water, reducing the strength and stability of the rock mass. In various material mix design methods, traditional mix design methods involve a large workload and high costs in laboratory testing. Summary of the Invention

[0004] To address the aforementioned problems, this application provides a design method and grouting material for deep coal-bearing strata. The low water-ash ratio grouting material of this application is environmentally friendly, economical, and efficient, suitable for the reinforcement characteristics of deep coal-bearing strata. The technical solution is as follows:

[0005] The first aspect of this application provides a method for designing grouting materials for deep coal-bearing strata, comprising the following steps: S1 Data acquisition based on material pre-tests, correlation analysis, and orthogonal tests; S2 Data expansion based on orthogonal test data to establish a dataset; S3 Creating a prediction model based on the RBF algorithm; S4 Optimizing the prediction model based on the PSO-RBF algorithm; S5 Proportion design based on the RBF-PSO-EWM algorithm; S6 Obtaining the proportion of grouting materials for deep coal-bearing strata.

[0006] For example, in a method for designing grouting materials for deep coal-bearing formations provided in one embodiment, the pre-test in S1 includes: taking strength and water-cement ratio tests as the main test contents, first determining the optimal mixing ratio of ordinary Portland cement and sulfoaluminate cement, adding different auxiliary materials and additives to the mixed cement, casting into standard cylindrical samples and curing, and conducting uniaxial compressive strength tests; adding different additives to cement with different water-cement ratios and stirring thoroughly, observing the state and fluidity of the grout to determine a suitable water-cement ratio, and recording the test results and proportion ranges of different auxiliary materials and additives during the test; the correlation analysis in S1 includes: performing correlation calculations based on the test results of the pre-test, quantitatively analyzing the correlation between different materials and indicators, and between materials, and determining the base material and additives of the grouting material based on the correlation results; the orthogonal test in S1 includes: setting strength, fluidity, porosity, viscosity, and contact angle as orthogonal test indicators based on the results of the correlation analysis and the proportion range of the pre-test materials, and the materials determined by the correlation analysis as orthogonal test factors, with the level range and gradient of each factor determined based on the pre-test results.

[0007] For example, in a method for designing grouting materials for deep coal-bearing formations provided in one embodiment, in step S1, the correlation coefficient R is calculated based on equation (1):

[0008]

[0009] Where, x i y i They are different variables, These are the averages of n different variables.

[0010] For example, in a method for designing grouting materials for deep coal-bearing formations provided in one embodiment, the data expansion based on orthogonal experimental data in step S2 includes the following steps:

[0011] Step 1: Using multiple linear regression analysis combined with orthogonal experimental data, multiple linear regression index analysis models for five indicators—strength, flowability, porosity, viscosity, and contact angle—were established according to equation (2):

[0012] y = β0 + β1x1 + β2x2 + ... + β p x p +ε Equation (2);

[0013] Where β1, β2, ..., β p β is the regression coefficient, β0 is the regression constant, x1, x2, ..., x p Let be the independent variable, y be the predicted value, and ε be the random error term;

[0014] The second step is to use the orthogonal experimental data factors as independent variables in the multiple linear regression index analysis model for interpolation processing, and then input the independent variables into the prediction model to increase the amount of data for the independent variables in order to achieve data expansion.

[0015] Step 3: Data Validation: Calculate R using equation (3) by combining the orthogonal experimental test data and extended data for the five indicators. 2 A comparison was performed to verify the validity of the extended data:

[0016]

[0017] in, This is the actual value. y is the actual average value. i These are predicted values.

[0018] For example, in a deep coal-bearing strata grouting material design method provided in one embodiment, the creation of a prediction model based on the RBF algorithm in S3 includes the following steps: Parameter setting: setting the mixed cement ratio, water-cement ratio, fly ash content, and mineral powder content as input data, and setting strength, fluidity, porosity, viscosity, and contact angle as output data; Data setting: 80% of the data in the dataset is used for model training, and 20% of the data is used for model testing; Accuracy and error analysis: calculating RMSE according to formula (4) and MAE according to formula (5) to test the prediction accuracy of the model:

[0019]

[0020]

[0021] For example, in a method for designing grouting materials for deep coal-bearing formations provided in one embodiment, the optimization of the prediction model based on the PSO-RBF algorithm in step S4 includes the following steps: Step 1: Establish the correspondence between RBF parameters and PSO; Step 2: Initialize the particle swarm and randomly generate an initial solution space according to the set PSO algorithm parameters to ensure the diversity of the initial solutions; Step 3: Map the information of all particles to the RBF neural network to construct a neural network model; Step 4: Calculate the fitness function value of the PSO algorithm and update the particle extrema, velocity, and position; Step 5: Determine whether the number of iterations meets the preset termination condition; Step 6: If the number of iterations meets the preset termination condition, terminate directly; otherwise, repeat steps 4 and 5 until the preset termination condition is met; Step 7: Record the updated global extrema, terminate the PSO algorithm, and construct the RBF neural network; Step 8: Optimize the model and verify: Input the data of the five indicators into the RBF and PSO-RBF models respectively in the extended data to perform R... 2 The accuracy of the optimized model was verified by comparing RMSE and MAE.

[0022] For example, in a deep coal-bearing strata grouting material design method provided in one embodiment, the PSO algorithm parameters set in the second step are as follows: PSO initialization particle n=30, initial inertial weight w=0.8, learning factors c1=2 and c2=2 for adjusting particle flight trajectory and state, random factors r1=0.5 and r2=0.6, spatial dimension D=15, and maximum number of iterations is 150.

[0023] For example, in a method for designing grouting materials for deep coal-bearing formations provided in one embodiment, the proportioning design based on the RBF-PSO-EWM algorithm in step S5 includes the following steps: Step 1: Input the material's various factors and all horizontal parallel test proportions into the index prediction model to predict five indicators: strength, fluidity, viscosity, porosity, and contact angle; Step 2: Set the index trend and optimize the index data using the entropy weight method (EWM) of the multi-index optimization algorithm; Step 3: Perform weighted summation on the five indicators of strength, fluidity, viscosity, porosity, and contact angle for each proportion, and select the proportion with the highest weighted total score as the proportion of the grouting material for deep coal-bearing formations.

[0024] For example, in a deep coal-bearing strata grouting material design method provided in one embodiment, the second step of optimizing the index data by combining the multi-index optimization algorithm entropy weight method (EWM) includes the following steps: Step 1: Data standardization: First, the dimensions of each index are dedimensionalized. For positive indices, the calculation is performed according to formula (6):

[0025]

[0026] The negative index is calculated according to formula (7):

[0027]

[0028] The optimal index is calculated according to equation (8):

[0029]

[0030] Step 2: Calculate the coefficient of variation of the indicators: Calculate the proportion of the j-th indicator in the i-th scheme according to equation (9):

[0031]

[0032] Step 3: Calculate the information entropy of the index: Calculate the information entropy of the j-th index according to equation (10):

[0033]

[0034] Step 4: Calculate the weight of each indicator according to equation (11):

[0035]

[0036] The second aspect of this application provides a grouting material for deep coal-bearing formations. According to the design method for the grouting material for deep coal-bearing formations described above, the proportions of the grouting material are as follows: sulfoaluminate cement: ordinary silicate cement: fly ash: mineral powder = 15-16%: 54-55%: 9.5-11%: 18-21.5%, with a water-cement ratio of 0.25-0.28, a water-reducing agent content of 1.1-1.25% of the total mass, and a thickener content of 0.03-0.06% of the total mass.

[0037] This application provides a design method for grouting materials in deep coal-bearing strata, and the beneficial effects of the grouting materials are as follows: First, the base material and proportion range of the grouting material are determined through preliminary experiments; a dataset is created using orthogonal experiments and data augmentation, and an index prediction model is established through algorithm comparison and optimization; then, the prediction model is combined with the entropy weight method to propose a PSO-RBF-EWM multi-objective optimization grouting material proportion design method, and the method is applied to design the comprehensive optimal proportion of grouting materials in deep coal-bearing strata. The grouting material with the obtained optimal proportion has excellent characteristics such as low water-cement ratio, green environmental protection, and economic efficiency, and is suitable for the reinforcement characteristics of deep coal-bearing strata. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in the embodiments of this specification or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0039] Figure 1 This is a graph showing the results of the correlation analysis;

[0040] Figure 2 This is a data expansion flowchart;

[0041] Figure 3a It is a dataset of intensity indicators;

[0042] Figure 3b It is a dataset of liquidity indicators;

[0043] Figure 3c It is a viscosity index dataset;

[0044] Figure 3d It is a porosity index dataset;

[0045] Figure 3e It is a contact angle index dataset;

[0046] Figure 4a This is the result of extended precision of the strength index data;

[0047] Figure 4b This is the result of extended precision of the fluidity index data;

[0048] Figure 4c This is the result of extended precision of viscosity index data;

[0049] Figure 4d This is the result of extended accuracy of the contact angle index data;

[0050] Figure 4e This is the result of extended accuracy of the porosity index data;

[0051] Figure 5 This is a flowchart for optimizing the prediction model;

[0052] Figure 6a This is a comparison chart of data optimized by the intensity algorithm;

[0053] Figure 6b This is a comparison chart of data optimized by the mobility algorithm;

[0054] Figure 6c This is a comparison chart of viscosity algorithm optimization data;

[0055] Figure 6d This is a comparison chart of contact angle algorithm optimization data;

[0056] Figure 6e This is a comparison chart of data optimized by the porosity algorithm;

[0057] Figure 7 This is a comparison chart of RBF optimization accuracy and error;

[0058] Figure 8 It refers to the information entropy and weight of the indicator. Detailed Implementation

[0059] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0060] Unless otherwise defined, the technical or scientific terms used in this disclosure shall have the ordinary meaning understood by one of ordinary skill in the art to which this disclosure pertains. The terms “first,” “second,” and similar terms used in this disclosure do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as “comprising” or “including” mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as “connected” or “linked” are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as “upper,” “lower,” “left,” and “right” are used only to indicate relative positional relationships, and these relative positional relationships may change accordingly when the absolute position of the described objects changes.

[0061] The first aspect of this application provides a method for designing grouting materials for deep coal-bearing strata, comprising the following steps:

[0062] S1 collects data based on material pre-experiments, correlation analysis, and orthogonal experiments;

[0063] S2 expands the data based on orthogonal experimental data to establish a dataset;

[0064] S3 creates a prediction model based on the RBF algorithm;

[0065] S4 optimizes the prediction model based on the PSO-RBF algorithm;

[0066] S5 is designed based on the RBF-PSO-EWM algorithm;

[0067] S6 obtains the grouting material ratio for deep coal-bearing strata.

[0068] This application first determined the base material and proportion range of the grouting material through preliminary experiments; then, a dataset was created using orthogonal experiments and data augmentation, and an index prediction model was established through algorithm comparison and optimization; finally, the prediction model was combined with the entropy weight method to propose the PSO-RBF-EWM multi-objective optimization grouting material proportion design method, and the method was applied to design the comprehensive optimal proportion of grouting material for deep coal-bearing strata. The grouting material with the obtained optimal proportion has excellent characteristics such as low water-cement ratio, green environmental protection, and economic efficiency, and is suitable for the reinforcement characteristics of deep coal-bearing strata.

[0069] For example, in one embodiment of the deep coal-bearing strata grouting material design method:

[0070] The preliminary test in S1 includes: with strength and water-cement ratio as the main test contents, firstly, the optimal ratio of ordinary Portland cement and sulfoaluminate cement is determined, different auxiliary materials and additives are added to the mixed cement, and standard cylindrical samples with a diameter of 50 mm and a height of 100 mm are poured and cured for 7 days to conduct uniaxial compressive strength tests; different additives are added to cement with different water-cement ratios and thoroughly stirred, and the state and fluidity of the slurry are observed to determine the appropriate water-cement ratio. During the test, the test results and proportion range of different auxiliary materials and additives are recorded.

[0071] The auxiliary materials include nano-silica, silica fume, fly ash, silica fume-fly ash, mineral powder, and talc; the additives include water-reducing agents, defoamers, and thickeners.

[0072] The correlation analysis in S1 includes: calculating the correlation based on the test results of the pre-test using formula (1), quantitatively analyzing the correlation between different materials and indicators, and between materials, and determining the base material and additives of the grouting material based on the correlation results;

[0073]

[0074] Where, x i y i They are different variables, Let X and Y be the average values ​​of n distinct variables. When the value of X increases (decreases) and the value of Y increases (decreases), the two vectors X and Y are positively correlated, with a correlation coefficient between 0.0 and +1.0; when the value of X increases (decreases) and the value of Y decreases (increases), the two vectors X and Y are negatively correlated, with a correlation coefficient between -1.0 and 0.0.

[0075] like Figure 1 As shown, based on the correlation analysis results, fly ash, mineral powder, ordinary silicate cement and sulfoaluminate cement were determined as the base materials for grouting materials, and water-reducing agents and thickeners were used as additives.

[0076] The orthogonal experiment in S1 includes: based on the results of correlation analysis and the proportion range of the pre-test materials, strength, fluidity, porosity, viscosity, and contact angle are set as orthogonal test indicators. The materials determined by correlation analysis are the orthogonal test factors, and the level range and gradient of each factor are determined based on the pre-test results. The factors and levels are shown in Table 1, and the test results are shown in Table 2. Table 2 shows that the maximum strength of all samples in the experiment was 52.09 MPa, the minimum was 28.88 MPa, and the average was 39.73 MPa; the average fluidity of the slurry was 23.18 mm, the average porosity was 5.28%, and the maximum contact angle was 93.45° and the minimum was 23.35°.

[0077] Table 1. Factor Levels for Orthogonal Design Experiments

[0078]

[0079] Table 2 Results of the orthogonal experiment on the index

[0080]

[0081] For example, in a method for designing grouting materials for deep coal-bearing formations provided in one embodiment, a large amount of training sample data is needed to improve the prediction accuracy and performance of the machine learning algorithm. Based on orthogonal experimental data, data expansion is performed to establish a dataset. The expansion method is as follows: Figure 2 As shown, the data expansion based on orthogonal experimental data in S2 includes the following steps:

[0082] Step 1: Using multiple linear regression analysis combined with orthogonal experimental data, multiple linear regression index analysis models for five indicators—strength, flowability, porosity, viscosity, and contact angle—were established according to equation (2):

[0083] y = β0 + β1x1 + β2x2 + ... + β p x p +ε Equation (2);

[0084] Where β1, β2, ..., β p β is the regression coefficient, β0 is the regression constant, x1, x2, ..., x p Let be the independent variable, y be the predicted value, and ε be the random error term;

[0085] Specifically, the multiple linear regression analysis model for the five indicators—strength, fluidity, porosity, viscosity, and contact angle—is as follows:

[0086]

[0087] Step 2: The orthogonal experimental data factors are used as independent variables in the multiple linear regression index analysis model for interpolation processing. Then, the independent variables are input into the prediction model to increase the amount of data for each independent variable, thus achieving data expansion. Through data expansion, 350 sets of data are added for each index, including the five index datasets for strength, fluidity, porosity, viscosity, and contact angle. Figures 3a-3e As shown. By Figures 3a-3e It can be seen that the strength and viscosity data are mainly concentrated between 34 and 42 MPa and 1000 and 3500 mPa.s; the flowability, porosity and contact angle data have a wide range, with the main data distributed between 23 and 24 mm, 4.5 and 5.5%, and 56 and 62°, respectively.

[0088] Step 3: Data Validation: Calculate R using equation (3) by combining the orthogonal experimental test data and extended data for the five indicators. 2A comparison was performed to verify the validity of the extended data:

[0089]

[0090] in, This is the actual value. y is the actual average value. i These are predicted values.

[0091] The results of the extended accuracy verification of the five indicators—strength, flowability, porosity, viscosity, and contact angle—are as follows: Figures 4a-4e As shown. By Figures 4a-4e It can be seen that the data expansion model has high accuracy, with R values ​​for the five metrics. 2 The values ​​were 0.94, 0.86, 0.96, 0.90, and 0.91, respectively. By combining linear regression analysis with linear interpolation, the dataset established met the data requirements of the indicator prediction model, providing a data foundation for the model's establishment.

[0092] For example, in a method for designing grouting materials for deep coal-bearing formations provided in one embodiment, the step S3 of creating a prediction model based on the RBF algorithm includes the following steps:

[0093] Parameter settings: Set the mixing ratio of cement, water-cement ratio, fly ash content, and mineral powder content as input data, and set the strength, fluidity, porosity, viscosity, and contact angle as output data;

[0094] Data setup: 80% of the data in the dataset is used for model training, and 20% is used for model testing; of which 280 sets of data are used for model training and 70 sets of data are used for model testing.

[0095] Accuracy and Error Analysis: RMSE was calculated according to Equation (4), and MAE was calculated according to Equation (5) to test the model's prediction accuracy.

[0096]

[0097]

[0098] For example, in a method for designing grouting materials for deep coal-bearing formations provided in one embodiment, in step S4, the prediction model is optimized based on the PSO-RBF algorithm. The Particle Swarm Optimization (PSO) algorithm is selected to optimize the prediction model. The parameters are set as follows: in the PSO initialization, the particle n = 30, the initial value of the inertia weight w = 0.8, the learning factors c1 = 2 and c2 = 2 for adjusting the particle flight trajectory and state, the random factors r1 = 0.5 and r2 = 0.6, the spatial dimension D = 15, and the maximum number of iterations is 150.

[0099] The detailed optimization flowchart is as follows: Figure 5 As shown, Figure 5 The flowchart shown is for optimizing RBF using PSO. The specific optimization steps are as follows:

[0100] Step 1: Establish the correspondence between RBF parameters and PSO;

[0101] Step 2: Initialize the particle swarm and randomly generate an initial solution space according to the set PSO algorithm parameters to ensure the diversity of the initial solutions;

[0102] Step 3: Map the information of all particles onto the RBF neural network to build the neural network model;

[0103] Step 4: Calculate the fitness function value of the PSO algorithm and update the particle extrema, velocity, and position;

[0104] Step 5: Determine if the number of iterations meets the preset termination condition;

[0105] Step 6: If the number of iterations meets the preset termination condition, the process ends directly; otherwise, repeat steps 4 and 5 until the preset termination condition is met.

[0106] Step 7: Record the updated global extremum, end the PSO algorithm, and construct the RBF neural network;

[0107] Step 8: Model Optimization and Validation: Input the data of the five indicators into the RBF and PSO-RBF models respectively in the expanded data to perform R-tests. 2 RMSE and MAE comparison tests were conducted to verify the accuracy of the optimized model. The verification results are as follows: Figures 6a-6e , Figure 7 As shown. Generally, Figures 6a-6e The closer the data is to the diagonal, the higher the prediction accuracy and the smaller the error. Overall, the PSO-RBF prediction data is closer to the diagonal than the RBF prediction data, while over 60% of the RBF prediction data is more dispersed and farther from the diagonal. For example... Figure 7 As shown, the average prediction accuracy (R) of the five indicators of the PSO-RBF prediction model is... 2 The accuracy was improved by 7.93%, with RMSE and MAE decreasing by an average of 13.24% and 25.68%, respectively. In summary, this verifies the feasibility and applicability of the PSO algorithm in optimizing the RBF network model and effectively improves the accuracy of indicator predictions.

[0108] For example, in a method for designing grouting materials for deep coal-bearing formations provided in one embodiment, the proportioning design based on the RBF-PSO-EWM algorithm in step S5 includes the following steps:

[0109] Step 1: In order to conduct scientific and comprehensive formula design, input all material factors and all horizontal parallel test formulas into the index prediction model to predict five indices: strength, flowability, viscosity, porosity, and contact angle.

[0110] Step 2: Set the indicator trend and optimize the indicator data using the multi-indicator optimization algorithm Entropy Weight Method (EWM). Specifically, the data optimization process includes the following steps:

[0111] Step 1: Data Standardization: First, the various indicators are dedimensionalized. Assume there are m indicators, X1, X2, ..., X... m , where X i ={x1, x2, ..., x n Because each indicator has a different order of magnitude, they need to be converted to the same range before comparison. Strength, flowability, and contact angle are positive indicators, while porosity and viscosity are negative indicators.

[0112] For positive indices, calculate according to formula (6):

[0113]

[0114] The negative index is calculated according to formula (7):

[0115]

[0116] The optimal index is calculated according to equation (8):

[0117]

[0118] Step 2: Calculate the coefficient of variation of the indicators: Calculate the proportion of the j-th indicator in the i-th scheme according to equation (9):

[0119]

[0120] Step 3: Calculate the information entropy of the index: Calculate the information entropy of the j-th index according to equation (10):

[0121]

[0122] Step 4: Calculate the weight of each indicator according to equation (11):

[0123]

[0124] The information entropy and weights of different indicators were calculated using the Entropy Weight Method (EWM) multi-index optimization algorithm, and the results are as follows: Figure 8 As shown. By Figure 8It can be seen that the information entropy of the five indicators, from largest to smallest, is: strength, porosity, flowability, contact angle, and viscosity. Viscosity has the lowest information entropy among the five indicators, but the data variation is the largest, therefore it is assigned the largest weight of 0.266. Strength data shows relatively small variations and stable changes, therefore it has a smaller weight of 0.140.

[0125] Step 3: For each mix proportion, the five indicators of strength, fluidity, viscosity, porosity and contact angle are weighted and summed. The mix proportion with the highest weighted total score is selected as the optimal mix proportion for grouting materials in deep coal-bearing strata.

[0126] The second aspect of this application provides a grouting material for deep coal-bearing formations. According to the design method for the grouting material for deep coal-bearing formations described above, the proportions of the grouting material are as follows: sulfoaluminate cement: ordinary silicate cement: fly ash: mineral powder = 15.0-16%: 54-55%: 9.5-11%: 18-21.5%, with a water-cement ratio of 0.25-0.28, a water-reducing agent content of 1.1-1.25% of the total mass, and a thickener content of 0.03-0.06% of the total mass.

[0127] Although the embodiments of this application have been disclosed above, they are not limited to the applications listed in the specification and embodiments. They can be applied to various fields suitable for this application. For those skilled in the art, other modifications can be easily made. Therefore, without departing from the general concept defined by the claims and their equivalents, this application is not limited to the specific details and the illustrations shown and described herein.

Claims

1. A method for designing grouting materials for deep coal-bearing strata, characterized in that, Includes the following steps: S1 Data collection is based on material pre-experiments, correlation analysis, and orthogonal experiments; S2 expands the dataset based on orthogonal experimental data; S3. Creating a prediction model based on the RBF algorithm; S4 optimizes the prediction model based on the PSO-RBF algorithm, specifically including the following steps: Step 1: Establish the correspondence between RBF parameters and PSO; Step 2: Initialize the particle swarm and randomly generate an initial solution space according to the set PSO algorithm parameters to ensure the diversity of the initial solutions; Step 3: Map the information of all particles onto the RBF neural network to build the neural network model; Step 4: Calculate the fitness function value of the PSO algorithm and update the particle extrema, velocity, and position; Step 5: Determine if the number of iterations meets the preset termination condition; Step 6: If the number of iterations meets the preset termination condition, the process ends directly; otherwise, repeat steps 4 and 5 until the preset termination condition is met. Step 7: Record the updated global extremum, end the PSO algorithm, and construct the RBF neural network; Step 8: Model Optimization and Validation: Input the data of the five indicators into the RBF and PSO-RBF models respectively in the expanded data to perform R-tests. 2 RMSE and MAE comparison tests were conducted to verify the accuracy of the optimized model; S5 uses the RBF-PSO-EWM algorithm for proportioning design, specifically including the following steps: Step 1: Input all material factors and the proportions of all horizontal parallel tests into the index prediction model to predict five indices: strength, flowability, viscosity, porosity, and contact angle. Step 2: Set the indicator trend and optimize the indicator data using the Entropy Weight Method (EWM) multi-indicator optimization algorithm. This includes the following steps: Step 1: Data Standardization: First, perform dimensionless processing on each indicator. For positive indicators, calculate as follows: ； The negative index is calculated using the following formula: ； The specified optimal index is calculated using the following formula: ； Step 2: Calculate the coefficient of variation of the indicators: Calculate the proportion of the j-th indicator in the i-th scheme according to the following formula: ； Step 3: Calculate the information entropy of the indicator ... according to the following formula. j Information entropy of the indicator: ； Step 4: Calculate the weight of each indicator according to the following formula: ； Step 3: For each mix proportion, the five indicators of strength, fluidity, viscosity, porosity and contact angle are weighted and summed. The mix proportion with the highest weighted total score is selected as the mix proportion of grouting material for deep coal-bearing strata. S6 Obtain the grouting material ratio for deep coal-bearing strata.

2. The design method for grouting materials in deep coal-bearing strata according to claim 1, characterized in that, The pre-test in S1 includes: with strength and water-cement ratio tests as the main test contents, firstly, the optimal mixing ratio of ordinary Portland cement and sulfoaluminate cement is determined, different auxiliary materials and additives are added to the mixed cement, and standard cylindrical samples are poured and cured for uniaxial compressive strength test; different additives are added to cement with different water-cement ratios and stirred thoroughly, and the state and fluidity of the slurry are observed to determine the appropriate water-cement ratio, and the test results and proportion ranges of different auxiliary materials and additives are recorded during the test; The correlation analysis in S1 includes: calculating the correlation based on the test results of the pre-test, quantitatively analyzing the correlation between different materials and indicators, and between materials, and determining the base material and additives of the grouting material based on the correlation results; The orthogonal experiment in S1 includes: based on the results of correlation analysis and the proportion range of the pre-test materials, setting strength, flowability, porosity, viscosity, and contact angle as orthogonal test indicators, and the materials determined by correlation analysis as orthogonal test factors. The level range and gradient of each factor are determined based on the pre-test results.

3. The design method for grouting materials in deep coal-bearing strata according to claim 2, characterized in that, In S1, the correlation coefficient R is calculated based on equation (1): Equation (1); in, x i 、y i They are different variables, , They are respectively n The average of 1,000 different variables.

4. The design method for grouting materials in deep coal-bearing strata according to claim 1, characterized in that, The data expansion based on orthogonal experimental data in S2 includes the following steps: Step 1: Using multiple linear regression analysis combined with orthogonal experimental data, establish multiple linear regression index analysis models for five indicators: strength, flowability, porosity, viscosity, and contact angle, according to equation (2): Equation (2); in, β 1 ，β 2 ，…，β p For regression coefficients, β 0 For the regression constant, x 1 ，x 2 ，…，x p As the independent variable, y For predicted values, This is the random error term; The second step is to use the orthogonal experimental data factors as independent variables in the multiple linear regression index analysis model for interpolation processing, and then input the independent variables into the prediction model to increase the amount of data for the independent variables in order to achieve data expansion. Step 3: Data validation: Calculate R using equation (3) by combining the orthogonal experimental test data and extended data for the five indicators. 2 A comparison was performed to verify the validity of the extended data: Equation (3); in, This is the actual value. This is the actual average value. These are predicted values.

5. The design method for grouting materials in deep coal-bearing strata according to claim 4, characterized in that, The creation of the prediction model based on the RBF algorithm in S3 includes the following steps: Parameter settings: Set the mixing ratio of cement, water-cement ratio, fly ash content, and mineral powder content as input data, and set the strength, fluidity, porosity, viscosity, and contact angle as output data; Data setup: 80% of the data in the dataset is used for model training, and 20% is used for model testing; Accuracy and Error Analysis: RMSE was calculated according to Equation (4), and MAE was calculated according to Equation (5) to test the model's prediction accuracy. Equation (4); Equation (5).

6. The design method for grouting materials in deep coal-bearing strata according to claim 1, characterized in that, The PSO algorithm parameters set in the second step of S4 are: particle n = 30 in PSO initialization, and initial value of inertial weight. w = 0.8, a learning factor that adjusts the particle's flight trajectory and state. c 1 = 2、 c 2 = 2, random factor r 1 = 0.5、 r 2 = 0.6, spatial dimension D = 15, the maximum number of iterations is 150.

7. A grouting material for deep coal-bearing formations, characterized in that, The grouting material design method for deep coal-bearing formations according to any one of claims 1-6 yields a grouting material mix ratio of: sulfoaluminate cement: ordinary silicate cement: fly ash: mineral powder = 15~16%: 54~55%: 9.5~11%: 18~21.5%, a water-cement ratio of 0.25~0.28, a water-reducing agent content of 1.1~1.25% of the total mass, and a thickener content of 0.03~0.06% of the total mass.

Images