A self-compensating geopolymer cement and a method for regulating the Mg content limit based on a BP neural network

Abstract

The application discloses a novel self-compensating geopolymer cement and a method for regulating a Mg content limit based on a BP neural network, and belongs to the field of building materials. The material is based on MATLAB to quickly and accurately determine the limit Mg content, thereby determining the optimal mixing amount of different magnesia clay, saving time cost and raw material cost. Through self-compensation, the drying shrinkage of the geopolymer cement is reduced, and the volume shrinkage and cracking problems in the hydration process of the geopolymer cement are solved. Combined with the special layered and fibrous microstructure of the clay, the brittleness of the geopolymer cement is improved. The free Mg 2+ complexing agent in the geopolymer remains, and the alkali bleeding of the geopolymer cement in the later hydration stage is reduced. It is a cementing material which can quickly determine the proportion through simulation calculation according to the material structure content and be used for self-compensating shrinkage building structure.

Description

Technical Field

[0001] This invention discloses a self-compensating geopolymer cement and a method for controlling the Mg content limit based on a BP neural network. It is a cementitious material that can quickly determine the mix ratio based on the phase content of the material through simulation calculation and can be used in self-compensating shrinkage building structures. Specifically, it relates to the field of building material simulation calculation. Background Technology

[0002] Alkali-activated slag-fly ash cement is a widely studied cementitious material in the building materials field in recent years. This cementitious material outperforms ordinary Portland cement in terms of compressive strength, frost resistance, and high-temperature resistance, especially in early-stage strength. Furthermore, because its raw materials, fly ash and slag, are derived from industrial waste, it not only reduces engineering costs but is also an environmentally friendly green building material. However, the main reason for its limited application is the inherent drawbacks of geopolymer cement, such as easy drying shrinkage, high brittleness, and efflorescence. Firstly, alkali metal oxides, such as calcium oxide and magnesium oxide, are commonly used to improve the drying shrinkage performance of materials. CN104710121B These alkali metal oxidation processes require the calcination of natural rocks. Besides, calcium sulfoaluminate is a commonly used expanding agent; however, sulfoaluminate needs to react with C3A in ordinary silicate cement systems to form expansive ettringite, while the C3A content is low in geopolymer cement systems. Secondly, the most common method to improve brittleness is fiber toughening, but for geopolymer cement, being in an alkali-activated system, all fibers—organic, inorganic, or metal fibers like steel—are easily corroded and extremely unstable. Finally, improving efflorescence usually involves controlling the alkali content of the reaction system, but for geopolymer cement systems, alkali solution is required to activate materials like slag for hydration and hardening, thus inevitably exceeding the alkali content limit. Considering all these factors, magnesian layered or fibrous clay is the best material to simultaneously solve all these problems.

[0003] Magnesian clay is rich in Mg-O bonds and free Mg. 2+ Under alkaline stimulation, the Mg-O bonds break, forming a new phase—a hydrotalcite-like phase (Ht)—with the hydration products of the geopolymer cement system. The theoretical density of the hydrotalcite phase is 1.89 g / cm³. 3 It has a density less than that of the main hydration product, calcium silicate hydrate (CSH), which is 2.16 g / cm³. 3 Therefore, volume expansion compensates for contraction. In addition, the excess alkali activator's OH... - Just with Mg 2+The generated magnesium hydroxide has a volume 1.5 times that of the original, solving both the problem of geopolymer shrinkage and the technical issue of easy efflorescence in geopolymers. Furthermore, due to its unique layered structure, it expands rapidly when added to water, forming a gel containing a large water network structure, exhibiting good elasticity, thixotropy, dispersibility, suspension, and thickening properties, thereby increasing the toughness of the geopolymer cement. For fibrous magnesium clay, it can form "defect connections" at the microscopic level, resisting the propagation of stress, thus playing a toughening role. However, the Mg content must be controlled within a reasonable range to ensure a proper balance between expansion and contraction, preventing over-expansion and poor stability. Therefore, based on simulation calculation technology, comprehensively controlling the overall Mg content in raw materials is also a key technology to ensure the performance of self-compensating geopolymer cement.

[0004] Due to differences in region and production methods, the composition of raw materials may vary, resulting in different properties of the prepared self-compensating geopolymer cement. By constructing a neural network prediction model using MATLAB, a suitable shrinkage compensation value and a reasonable range of Mg content can be found. At the same time, when using different types of magnesian clay, the optimal admixture can be calculated by applying this limit of Mg content. This reduces a lot of experimental processes, lowers time costs, and reduces raw material costs. More importantly, it precisely controls the performance and quality of the geopolymer cement.

[0005] In summary, the present invention provides a method for rapidly and accurately preparing a novel self-compensating, low-brittle cementitious material by means of a self-compensating polymer cement and a method for regulating the Mg content limit based on a BP neural network. Summary of the Invention

[0006] The technical problem to be solved:

[0007] 1. This invention provides a self-compensating geopolymer cement and a method for regulating the Mg content limit based on a BP neural network, which enables faster and more accurate determination of the proportions, saving time and raw material costs.

[0008] 2. This invention provides a self-compensating geopolymer cement and a method for regulating the Mg content limit based on a BP neural network, which aims to further reduce the drying shrinkage of geopolymer cement.

[0009] 3. This invention provides a self-compensating geopolymer cement and a method for regulating the Mg content limit based on a BP neural network, which aims to further reduce the brittleness of the geopolymer cement.

[0010] 4. This invention provides a self-compensating geopolymer cement and a method for regulating the Mg content limit based on a BP neural network, which aims to further reduce efflorescence in the later stage of hydration of geopolymer cement.

[0011] Technical solution:

[0012] To achieve the above technical requirements, the following technical solution is used: Utilizing MATLAB and a BP neural network, the influence of different types of magnesium clay on the performance of self-compensating geopolymer cementitious materials is simulated and calculated based on the different chemical compositions of magnesium. The dosage is then rapidly adjusted to precisely prepare a new type of self-compensating, low-brittleness geopolymer cement, including the following steps:

[0013] 1. Self-compensating geopolymer mortar is made by mixing 60-80 parts slag, 10-300 parts fly ash, 0-5 parts magnesia clay, 0-5 parts alkali activator, and 25-35 parts water, then filling the mixture into a 40*40*160mm mold and curing it to the corresponding age to determine its strength and shrinkage value.

[0014] 2. Determine the MgO structural content (X) in the magnesian clay used, and collect the raw material ratio and performance data sets;

[0015] 3. Construct a feedforward artificial neural network, which includes one hidden layer, four input features (mass ratio of slag to fly ash, alkali activator, magnesia clay, and water quality), and two output features (flexural strength and shrinkage value). Call the transig function to randomly divide the data set into training and test sets at a ratio of 4:1. Input the data into the program for training and testing to obtain a linear function.

[0016] 4. Based on the linear function, when the magnesian clay content is 5-10 parts, input the prediction program, perform reverse normalization, analyze the output characteristic values, and determine the limiting magnesian clay content as Y.

[0017] 5. The optimal Mg content for magnesian clay is calculated to be Z = X * Y * 0.6;

[0018] 6. Determine the optimal Mg content for other types of magnesian clay.

[0019] The positive effects of this invention:

[0020] Compared with existing technologies, this invention provides a novel self-compensating geopolymer cement and a method for regulating the Mg content limit based on a BP neural network, with the following positive effects:

[0021] 1. Based on the BP neural network model, the limit Mg content can be determined quickly and accurately, thereby determining the optimal admixture of different magnesian clays, saving time and raw material costs.

[0022] 2. Self-compensation reduces the drying shrinkage of geopolymer cement, thus solving the problems of volume shrinkage and cracking during the hydration process of geopolymer cement.

[0023] 3. The unique layered and fibrous microstructure of clay improves the brittleness of geopolymer cement.

[0024] 4. Utilizing free Mg2+ The residual alkali activator in the complexed geopolymer reduces efflorescence in the later stage of hydration of geopolymer cement. Attached Figure Description

[0025] Figure 1 Self-compensating polymer cement mortar test block.

[0026] Figure 2 : Fitting effect of neural network regulation of Mg content. Detailed Implementation

[0027] The present invention is further illustrated by the following embodiments, which are not intended to limit the invention in any way. Any modifications or alterations made to the present invention that are easily implemented by those skilled in the art without departing from the technical solutions of the present invention shall fall within the scope of the claims of the present invention.

[0028] Example:

[0029] 1. Mix 60-80 parts slag, 10-300 parts fly ash, 0-5 parts laponite clay (MgO content 28.3%), 0-5 parts alkali activator, and 25-35 parts water. Pour the mixture into a 40*40*160mm mold and cure to the appropriate age. Measure the strength and shrinkage value. See attached sample for self-compensating polymer cement mortar test blocks. Figure 1 .

[0030] 2. Collect 60 sets of data and divide them into training and testing datasets in a 4:1 ratio, resulting in matrices 1 and 2. The fitting effect is attached. Figure 2 .

[0031]

[0032] Matrix 1

[0033]

[0034] Matrix 2

[0035] 3. Input 5-10 parts of laponite clay into the prediction program. After inverse normalization, analyze the output eigenvalues ​​to determine the optimal laponite content as 4 parts. At this point, the strength value is 10.1 MPa and the shrinkage value is 20 × 10⁻⁶ MPa. -6 The calculated limit for Mg content is 0.68.

[0036] 4. Prepare self-compensating geopolymer cement by using another type of magnesian clay to extract lithium saponite (MgO content 32.6%). The admixture should be 3.5 parts. After curing to the corresponding age, the strength and shrinkage values ​​are measured to be 10.8 MPa and 16 × 10⁻⁶ MPa, respectively. -6 The experimental values ​​are similar to those of laponite magnesian clay.

Claims

1. A method for controlling the Mg content limit in self-compensating geopolymer cement based on a BP neural network, characterized in that: Self-compensating polymer cement comprises the following components: slag, fly ash, magnesian clay, alkali activator, and water. The magnesian clay is a high-magnesium clay with a layered or fibrous structure. The magnesium content limit of the self-compensating polymer cement is predicted using a BP neural network model constructed with MATLAB. The specific steps for predicting the magnesium content limit are as follows: (1) Collect raw material ratio and performance data sets; (2) Construct a feedforward artificial neural network containing one hidden layer; four input features: slag to fly ash mass ratio, alkali activator, magnesia clay and water mass, and two output features: flexural strength and shrinkage value. Call the transig function to randomly divide the data set into training set and test set in a 4:1 ratio. Input the data into the program for training and testing to obtain a linear function. (3) Based on the linear function, when the clay content is 5-10 parts, input the prediction program, perform reverse normalization, analyze the output characteristic value, and determine the limit clay content as Y; (4) The limit Mg content of clay is calculated as Z = X * Y * 0.6, where X is the structural content of MgO in magnesian clay; (5) Determine the optimal dosage of other magnesian clays based on the Z value.

2. The method according to claim 1, characterized in that: The mass fractions of the components in the self-compensating geopolymer cement mortar are: 60-80 parts slag, 10-300 parts fly ash, 0-5 parts magnesia clay, 0-5 parts alkali activator, and 25-35 parts water.

3. The method according to claim 1 or 2, characterized in that: The magnesian clay is laponite, lithium saponite, or sepiolite.

4. The method according to claim 3, characterized in that: The laponite is layered with a particle size of 5nm-25nm, the lithium saponite is layered with a particle size of 5nm-25nm, the sepiolite is filamentous with a particle size of 100nm-300nm, and the MgO structural content X in the magnesian clay is 22.0%-35.0%.

Images