Artificial intelligence-based geopolymer recycled concrete proportion optimization method and system

Abstract

The invention discloses a geopolymer recycled concrete proportion optimization method and system based on artificial intelligence, and the method comprises the steps: obtaining soil test data and finite element simulation data, determining the neat paste strength fp and the design range of geopolymer components, calculating a deviation coefficient, correcting the finite element simulation data, obtaining a neat paste index library and a mortar index library, and obtaining a geopolymer recycled concrete proportion optimization result. Matching a database to obtain a first mortar design parameter set and a second mortar design parameter set, screening the second mortar design parameter set to obtain a first design formula set, and constructing a geopolymer recycled concrete proportioning optimization model based on artificial intelligence according to the first design formula set, and inputting a to-be-optimized ratio into the geopolymer recycled concrete ratio optimization model to output an optimized ratio. The method not only can improve the ratio optimization speed and effect of the geopolymer recycled concrete, but also has better interpretability, and can be directly applied to a ratio optimization system of the geopolymer recycled concrete.

Description

Technical Field

[0001] The present invention relates to the field of formula optimization, and in particular to an artificial intelligence-based geopolymer recycled concrete mix ratio optimization method and system. Background Art

[0002] The recycling of construction waste is a hot topic in the current research on sustainable development in the construction field. Geopolymer recycled concrete, as a new type of environmentally friendly material, combines the characteristics and advantages of recycled aggregate and geopolymer concrete, and has excellent mechanical properties, durability and environmental friendliness.

[0003] However, the composition of geopolymer recycled concrete is complex. The traditional concrete mix design method relies on empirical trial mixing, which requires the preparation of a large number of samples and complex geotechnical tests. It is not only inefficient but also difficult to achieve optimal performance. As a numerical analysis method for solving approximate solutions to complex physical problems, finite element simulation technology can combine a small amount of geotechnical test data with a large amount of finite element simulation data to effectively predict material properties and optimize mix design. At the same time, the development of machine learning, data processing and optimization algorithms has provided powerful tools for dealing with complex nonlinear problems. Although some geopolymer recycled concrete mix optimization methods have been invented, there are still problems such as huge data, low efficiency and insufficient optimization effect. By combining finite element simulation and artificial intelligence technology, a large amount of effective data is obtained, the data is accurately processed, and a scientific and efficient mix optimization model is established. A geopolymer recycled concrete mix optimization method and system based on artificial intelligence is designed to overcome the shortcomings of the existing comparison method and achieve accurate and efficient optimization of geopolymer recycled concrete mix, which is of great significance to the resource utilization of construction waste and the development of green building materials. Summary of the invention

[0004] The purpose of the present invention is to provide a method and system for optimizing the mix ratio of geopolymer recycled concrete based on artificial intelligence.

[0005] To achieve the above object, the present invention is implemented according to the following technical solutions:

[0006] The present invention comprises the following steps:

[0007] Construct a finite element model of geopolymer recycled concrete with different aggregate size distributions, conduct numerical simulation analysis to obtain the first simulation data, and determine the target strength f of geopolymer recycled concrete. aim , according to the concrete target strength f aim The net slurry design strength f is determined by the first simulation data p scope;

[0008] Prepare slurries of different geopolymer components to obtain first slurry parameters, and calculate the first slurry parameters to determine the design range of the geopolymer components;

[0009] According to the design range of the geopolymer component, finite element simulation and geotechnical test of the pure slurry are performed to obtain a pure slurry index library, and fine aggregates with different volume fractions are added based on the design range of the component group to perform finite element simulation and geotechnical test of mortar to obtain a mortar index library;

[0010] The slurry design strength f p Range matching the mortar index library to obtain a first mortar design parameter set, matching the mortar index library according to the first mortar design parameter set to obtain a second mortar design parameter set, screening the second mortar design parameter set, and determining a first design formula set according to the screening result;

[0011] A geopolymer recycled concrete mix ratio optimization model is constructed according to the first design formula, and the mix ratio to be optimized is input into the geopolymer recycled concrete mix ratio optimization model to obtain the optimized mix ratio of the geopolymer recycled concrete.

[0012] Further, determine the design strength f of the slurry p A range of methods, including:

[0013] Conduct geotechnical tests to obtain basic material parameters, and construct finite element models of geopolymer recycled concrete with different aggregate particle size distributions based on the material parameters to perform numerical simulation analysis to obtain the first simulation data;

[0014] Determine multiple groups of commonly used net slurry strengths, prepare multiple groups of geopolymer recycled concrete specimens with different aggregate particle size distributions, conduct geotechnical tests and obtain the first geotechnical data;

[0015] Match the first geotechnical data with the first simulation data according to the principle of consistency of the net slurry strength and aggregate particle size gradation, calculate the specimen strength deviation of the matching data to determine the simulation measured deviation coefficient, and correct the first simulation data according to the simulation measured deviation coefficient to obtain the geopolymer recycled concrete index library;

[0016] Determine the target strength f of geopolymer recycled concrete based on the original mix ratio aim , according to the concrete target strength f aim and concrete framework theory to determine the coarse aggregate strength f A and mortar strength f m scope;

[0017] According to the coarse aggregate strength f A The range of is matched to the coarse aggregate contribution strength of the geopolymer recycled concrete index library, and the corresponding coarse aggregate volume fraction V is obtained. A , fine aggregate volume fraction V S and mortar strength f m scope;

[0018] The volume fraction of coarse aggregate VA , fine aggregate volume fraction V S The net paste strength function is input with mortar strength to obtain the range of net paste design strength.

[0019] Further, the method for determining the design range of the geopolymer component includes:

[0020] Prepare basic slurries with different geopolymer components to measure first slurry parameters; the first slurry parameters include geopolymer components and fluidity b; the geopolymer components include liquid-to-solid ratio L / S, alkali activator concentration C and alkali activator modulus M;

[0021] According to the geopolymer composition and fluidity b, the design range of the geopolymer composition is determined, and the evaluation matrix X is constructed. The scheme set is A = (A1, A2, ..., A m ), the target set is B=(B1,B2,…,B m ), m is the number of items, and the evaluation matrix X is expressed as:

[0022]

[0023] Among them, plan A i and target B j The corresponding evaluation value is recorded as x ij , which is also the corresponding geopolymer component. The design target of the geopolymer component is quality-oriented. The evaluation matrix X is standardized to obtain the matrix V, where v ij =(x ij -min x j ) / (maxx j -min x j ) is the standard matrix V in solution A i and target B j The evaluation value of

[0024] Calculate the entropy and weight of the j-item target, the expression is:

[0025]

[0026]

[0027] where e j is the entropy value of the j-item target, b j is the control constant representing the fluidity corresponding to the j-th target, w j is the weight of the j-item target;

[0028] According to the weight w j Calculate the positive ideal solution corresponding to the best solution Negative ideal solution corresponding to the worst solution Calculate the Euclidean distance d of the model based on the ideal solution iGet the relative closeness, the expression is:

[0029]

[0030] where η i For plan A i With positive ideal solution The closeness of is the positive ideal solution of the j-item target, is the negative ideal solution of the j-th target, is the Euclidean distance between the evaluation solution and the ideal solution;

[0031] The design range of geopolymer components is determined based on the schemes corresponding to the maximum degree of proximity and the average degree of proximity.

[0032] Furthermore, the method for obtaining the mortar index library includes:

[0033] Prepare multiple groups of slurries with different geopolymer components according to the design range of geopolymer components, and conduct geotechnical tests to obtain the second slurry parameters;

[0034] According to the basic material parameters and the design range of geopolymer components, finite element models with different liquid-to-solid ratios are constructed to perform finite element simulation to obtain second simulation data; the second simulation data includes net slurry strength, net slurry plastic viscosity and net slurry yield stress; the net slurry strength is obtained by nonlinear analysis using ANSYS software; the net slurry rheological parameters are obtained by analyzing the net slurry discrete model established using PFC3D software;

[0035] Based on the same geopolymer component, the net slurry strength deviation coefficient is determined according to the second net slurry parameters and the second simulation data, and the expression is:

[0036]

[0037] where ξ fp is the net pulp strength deviation coefficient, n is the data volume, f pFi is the simulated strength of the ith group of pure slurry, f pti is the test strength of the ith group of pure slurry;

[0038] According to the above method, the deviation coefficient of plastic viscosity of the net slurry ξ is calculated. ηp and the net slurry yield stress deviation coefficient ξ τp , the second simulation data is corrected according to the deviation coefficient, and the corrected data and the corresponding liquid-solid ratio L / S, alkali activator concentration C and alkali activator modulus M are combined to form a clean slurry index library;

[0039] According to the design range of geopolymer components, fine aggregates with different volume fractions are added to the pure paste of different geopolymer components to prepare multiple groups of mortars, and geotechnical tests are carried out to obtain the mortar measurement parameters;

[0040] According to the basic material parameters, the design range of geopolymer components and the volume fraction V of fine aggregate in mortar S-m , construct finite element models with different liquid-to-solid ratios to perform finite element simulation to obtain third simulation data; the third simulation data includes mortar strength, mortar plastic viscosity, mortar yield stress and mortar slump; the mortar strength is obtained by nonlinear analysis using ANSYS software; the mortar plastic viscosity and yield stress are obtained by analyzing a mortar discrete model established using PFC3D software; the mortar slump is obtained by lattice Boltzmann method simulation based on Herschel-Bulkley model;

[0041] Based on the same geopolymer composition and fine aggregate volume fraction V in mortar S-m The above method is used to calculate the deviation coefficient according to the mortar parameters and the third simulation data, and the third simulation data is corrected according to the deviation coefficient to obtain the mortar index library.

[0042] Furthermore, the method for obtaining the second mortar design parameter set includes:

[0043] Set the net slurry design strength f p Range and corresponding coarse aggregate volume fraction V A and the volume fraction of fine aggregate in mortar V S-m To search the data set, the improved reptile search algorithm is used to search the slurry index library according to the search data set, and the first mortar design parameter set is matched;

[0044] The Sobol sequence is used to initialize the population position, and the expression is:

[0045]

[0046] Where X i,j is the initial population position of individual i in j-dimensional space, [u bs , l bs ] is the value range of the global solution, D(Y) is the value range of x=(x1,x2,…,x n ) is the total variance of the function Y = f(x) of the independent variable, n is the population size, D i is the variance generated by the i-th independent variable;

[0047] Update the population position in the encirclement phase and the population position in the hunting phase according to the number of iterations;

[0048] Set sensitive parameters to control the search accuracy, add control factors to control the decreasing amplitude of the evolution factor, the percentage difference between the optimal solution and the current solution, and the evolution factor expression are:

[0049]

[0050] Where Pi,j is the percentage difference between the optimal solution and the current solution, α=0.1 is a sensitive parameter, x best j (t) is the j-dimensional position of the current best solution, r3 is a random integer between [-1,1], ε is a minimum value to prevent the denominator from being 0, E s (t) is the evolution factor, δ is the [1,3] optimization index;

[0051] Calculate the fitness according to the population position and save the optimal solution, continue to calculate the fitness of the total population and update the position of the total population, and iterate until the maximum number of iterations is reached and then output the optimal solution, and obtain the first mortar design parameter set according to the optimal solution;

[0052] Calculate the similarity between the first mortar design parameter set and the mortar index library, and convert the design strength f p , liquid-to-solid ratio L / S, alkali activator concentration C, alkali activator modulus M, coarse aggregate volume fraction V A and the volume fraction of fine aggregate in mortar V S-m The characteristic vector composed of the first mortar design parameter set and the characteristic vector corresponding to the mortar index library are calculated, and the data group with the highest similarity is selected to form the second mortar design parameter set.

[0053] Further, the method for determining the first design formula set includes:

[0054] According to the volume fraction V of coarse aggregate mortar A Calculate the mortar film thickness H m , according to the mortar film thickness H m The physical parameters of mortar and aggregate determine the plastic viscosity threshold η of mortar m0 and mortar yield stress threshold τ m0 , set the mortar slump target value S m0 ;

[0055] The hypersphere classifier is used to classify the mortar slump S m , mortar plastic viscosity η m and mortar yield stress τ m To classify;

[0056] Set information set H = (h i ,q i ), h i =(h1,h2,…,h j ) represents an element in the set, q i Represents the class label. Assuming that the data is located inside the hypersphere (a, R), a is the center of the hypersphere, and R is the radius of the hypersphere. It is related to the data threshold. The hypersphere function expression is optimized by combining the Lagrangian formula. The expression is:

[0057]

[0058] Where L is the objective function, k(h i ,h j ) is the kernel function used to measure the element h i 、h j The similarity of i , α j is the Lagrangian factor, C is a constant, and also serves as the upper bound of the Lagrangian factor;

[0059] The Gaussian kernel function is used to calculate the distance between the sample data point and the center of the hypersphere, and to determine the classification. The expression is:

[0060]

[0061] f(h)=sgn[R-d1(h)]

[0062] Where d1 is the distance between the sample data point h and the center of the hypersphere, k(h,h) is the Gaussian kernel function that calculates the similarity between h and itself, and k(h i ,h) Calculate the sample point h i The similarity with point h, f(h) is a classification function used to determine the classification of the sample data; the classification is that the sample data is located inside the corresponding hypersphere, and the sample data is located outside the corresponding hypersphere;

[0063] The slump of mortar S m The value is located in the corresponding sphere external data group, mortar plastic viscosity η m The value is located in the corresponding sphere external data group, mortar yield stress τ m The data sets whose values ​​lie inside the corresponding spheres are intersected to obtain the comprehensive data set;

[0064] The comprehensive data set and the corresponding geopolymer concrete mix proportions are combined to form the first design formulation set of geopolymer recycled concrete.

[0065] Furthermore, the method for obtaining the optimized ratio includes:

[0066] The first design formula set of geopolymer recycled concrete is divided into a training set and a test set;

[0067] Constructing a geopolymer recycled concrete mix ratio optimization model, which includes an input layer, a hidden layer and an output layer;

[0068] Support vector machine (SVM) is used as the base model, and Gaussian kernel function is used to map data into high-dimensional feature space. The nonlinearly related data in the first design formula set are processed, and the optimal plane is found. The grid search training model is used to select the optimal parameters.

[0069] The geopolymer recycled concrete mix ratio objective function is used to evaluate the quality of each generation of solutions, and the solutions and model weights that meet the requirements of environmental friendliness, high cost performance and excellent mortar working performance are selected. Multiple iterations are performed until the evaluation value of the geopolymer recycled concrete mix ratio objective function no longer decreases.

[0070] The objective function expression of geopolymer recycled concrete mix ratio is:

[0071]

[0072] Where y(f aim ) is the geopolymer recycled concrete mix ratio objective function, w1 and w2 are economic weights, w3 is the environmental weight, w4 is the work performance weight, ρ1 is the replacement rate of recycled aggregate in coarse aggregate, k is the number of commercial crushed stone particle size categories in coarse aggregate, ρ 2i is the proportion of the i-th type of gravel, PRICE i is the price of the i-th type of crushed stone, l1(L / S,C) is the alkali activator concentration function, which means the alkali activator concentration C corresponding to the liquid-to-solid ratio L / S, V p is the mortar volume fraction, PRICE C is the price of alkali activator, ρ0 is the standard recycled aggregate replacement rate, μ is the carbon emission impact coefficient, l2(ρ1V A ,C CO2 ) is a function of CO2 concentration, indicating the reduction in carbon emissions from replacing crushed stone with recycled aggregate;

[0073] The mean square loss function is used to evaluate the economic cost difference between the predicted value and the standard, the cross entropy loss function is used to measure the environmental impact difference between the predicted value and the standard value, L1 and L2 regularization is used to prevent overfitting and improve the generalization ability of the model, and the test set is used to evaluate the geopolymer recycled concrete mix optimization model;

[0074] The mix ratio to be optimized is input into the geopolymer recycled concrete mix ratio optimization model to obtain the optimized mix ratio of geopolymer recycled concrete.

[0075] Furthermore, the geopolymer recycled concrete comprises: coarse aggregate, fine aggregate and pure paste; the coarse aggregate is replaced by recycled coarse aggregate instead of commercial crushed stone, and the replacement rate is ρ1; alkali activator and aluminosilicate raw material are added to the pure paste, and the mass ratio of the alkali activator and the aluminosilicate raw material is represented by the liquid-solid ratio L / S; the alkali activator is a mixture of solid sodium hydroxide, sodium silicate solution and water; the aluminosilicate raw material is composed of granular blast furnace slag and metakaolin.

[0076] Secondly, the artificial intelligence-based geopolymer recycled concrete mix optimization system includes:

[0077] Simulation module: connected to the data acquisition module; used for finite element simulation of the geopolymer recycled concrete, paste and mortar;

[0078] Data acquisition module: connected to the simulation module, the data processing module and the intelligent supervision module; used to obtain geotechnical test data and finite element simulation data, and pre-process the geotechnical test data and the simulation data; the geotechnical test data includes the first slurry parameter, the second slurry parameter and the mortar measurement parameter; the finite element simulation data includes the first simulation data, the second simulation data and the third simulation data;

[0079] Data processing module: connects the data acquisition module, the optimization model module and the intelligent supervision module; is used to determine the net slurry design strength f p The range is used to determine the liquid-to-solid ratio L / S design range, which is used to correct the finite element simulation data of the mortar and mortar to obtain a mortar index library and a mortar index library, which are used to match the first mortar design parameter set and the second mortar design parameter set, and are used to screen the mortar rheological parameters to determine the first design formula set of geopolymer recycled concrete;

[0080] Optimization model module: connected to the data processing module and the intelligent supervision module; used to construct a geopolymer recycled concrete mix ratio optimization model according to the first design formula set of the geopolymer recycled concrete, and input the geopolymer recycled concrete mix ratio to be optimized into the geopolymer recycled concrete mix ratio optimization model to obtain the optimized geopolymer recycled concrete mix ratio;

[0081] Intelligent supervision module: connecting the data acquisition module, the data processing module and the optimization model module; used for storing, viewing and managing the geotechnical test data, the finite element simulation data, the first design formula set and the optimized mix ratio of the geopolymer recycled concrete.

[0082] The beneficial effects of the present invention are:

[0083] The present invention is a method and system for optimizing the mix ratio of geopolymer recycled concrete based on artificial intelligence. Compared with the prior art, the present invention has the following technical effects:

[0084] The present invention can improve the data preprocessing capabilities and enhance the model adaptability in the optimization of geopolymer recycled concrete mix ratio through finite element simulation, correction of simulation data, matching of mortar index library, screening of second mortar design parameter set, and construction of model steps, thereby improving the optimization efficiency and accuracy of geopolymer recycled concrete mix ratio. The optimization technology of geopolymer recycled concrete mix ratio can improve the scientific nature of the formula and the optimization work efficiency, and can achieve the optimization of geopolymer recycled concrete mix ratio, providing strong support for the selection of building material formulas, and is of great significance to the resource utilization of construction waste and the development of green building materials. It can adapt to different artificial intelligence-based geopolymer recycled concrete mix ratio optimization systems and different users' artificial intelligence-based geopolymer recycled concrete mix ratio optimization systems. Terminal ratio optimization needs, have a certain universality. BRIEF DESCRIPTION OF THE DRAWINGS

[0085] Figure 1 The present invention is a flowchart of the steps of the method for optimizing the mix ratio of geopolymer recycled concrete based on artificial intelligence. DETAILED DESCRIPTION

[0086] The present invention is further described below by means of specific embodiments. The illustrative embodiments and descriptions of the present invention are used to explain the present invention but are not intended to limit the present invention.

[0087] The method and system for optimizing the mix ratio of geopolymer recycled concrete based on artificial intelligence of the present invention include the following steps:

[0088] like Figure 1 As shown, in this embodiment, the following steps are included:

[0089] Construct a finite element model of geopolymer recycled concrete with different aggregate size distributions, conduct numerical simulation analysis to obtain the first simulation data, and determine the target strength f of geopolymer recycled concrete. aim , according to the concrete target strength f aim The net slurry design strength f is determined by the first simulation data p scope;

[0090] Prepare slurries of different geopolymer components to obtain first slurry parameters, and calculate the first slurry parameters to determine the design range of the geopolymer components;

[0091] According to the design range of the geopolymer component, finite element simulation and geotechnical test of the pure slurry are performed to obtain a pure slurry index library, and fine aggregates with different volume fractions are added based on the design range of the component group to perform finite element simulation and geotechnical test of mortar to obtain a mortar index library;

[0092] The slurry design strength f pRange matching the mortar index library to obtain a first mortar design parameter set, matching the mortar index library according to the first mortar design parameter set to obtain a second mortar design parameter set, screening the second mortar design parameter set, and determining a first design formula set according to the screening result;

[0093] A geopolymer recycled concrete mix ratio optimization model is constructed according to the first design formula, and the mix ratio to be optimized is input into the geopolymer recycled concrete mix ratio optimization model to obtain the optimized mix ratio of the geopolymer recycled concrete.

[0094] In this embodiment, the net slurry design strength f is determined p A range of methods, including:

[0095] Conducting geotechnical tests to obtain basic material parameters, constructing a finite element model of geopolymer recycled concrete with different aggregate particle size gradations according to the material parameters, and performing numerical simulation analysis to obtain first simulation data; the first simulation data includes specimen strength, coarse aggregate contribution strength, mortar contribution strength, net slurry set strength, and aggregate particle size gradation; the aggregate particle size gradation includes coarse aggregate volume fraction and fine aggregate volume fraction;

[0096] Determine multiple groups of commonly used net slurry strengths, prepare multiple groups of geopolymer recycled concrete specimens with different aggregate particle size distributions, conduct geotechnical tests to obtain first geotechnical data; the first geotechnical data includes specimen strength, coarse aggregate contribution strength, mortar contribution strength, and net slurry set strength;

[0097] Match the first geotechnical data with the first simulation data according to the principle of consistency of the net slurry strength and aggregate particle size gradation, calculate the specimen strength deviation of the matching data to determine the simulation measured deviation coefficient, and correct the first simulation data according to the simulation measured deviation coefficient to obtain the geopolymer recycled concrete index library;

[0098] Determine the target strength f of geopolymer recycled concrete based on the original mix ratio aim , according to the concrete target strength f aim and concrete framework theory to determine the coarse aggregate strength f A and mortar strength f m scope;

[0099] According to the coarse aggregate strength f A The range of is matched to the coarse aggregate contribution strength of the geopolymer recycled concrete index library, and the corresponding coarse aggregate volume fraction V is obtained. A , fine aggregate volume fraction V S and mortar strength f m scope;

[0100] The volume fraction of coarse aggregate V A , fine aggregate volume fraction V SThe mortar strength is input into the net slurry strength function to obtain the range of net slurry design strength, which is expressed as:

[0101] f m =a1V S-m 2 +a2V S-m +a3f p 2 +a4f p +a5

[0102] where f p is the design strength of the pure slurry, a1, a2, a3, a4, a5 are the strength coefficients of the pure slurry, obtained by fitting the first simulation data, V S-m =V S / (1-V A ) is the volume fraction of fine aggregate in mortar;

[0103] In the actual evaluation, a set of geopolymer recycled concrete mix ratios were randomly obtained: liquid-solid ratio L / S = 1.1, alkali activator concentration C was 27%, alkali activator modulus M was 1.8, coarse aggregate volume fraction V A The volume fraction of fine aggregate is 40%. S %, the replacement rate of recycled aggregate ρ1 is 20%;

[0104] Prepare standard pure slurry, prepare geopolymer recycled concrete according to the formula, and measure the mortar plastic viscosity η m is 4.56Pa·s, and the yield stress of mortar is τ m is 120.61Pa, slump S m is 175mm, concrete strength f cu The plastic viscosity of the mortar is less than the corresponding plastic viscosity threshold of 5.66Pa·s, the slump is less than the slump target value of 180mm, and the concrete strength is low. It is necessary to adjust the proportion of geopolymer recycled concrete to ensure that the rheological parameters of the mortar meet the working performance requirements and improve the concrete strength.

[0105] The concrete strength f aim It is set to 55MPa, and the mortar strength f is determined by combining the concrete framework theory m 48.4-55MPa, coarse aggregate strength f A is 50-70MPa, thus determining the design strength f of the slurry p It is 45-53MPa.

[0106] In this embodiment, the method for determining the design range of the geopolymer composition includes:

[0107] Prepare basic slurries with different geopolymer components to measure first slurry parameters; the first slurry parameters include geopolymer components and fluidity b; the geopolymer components include liquid-to-solid ratio L / S, alkali activator concentration C and alkali activator modulus M;

[0108] According to the geopolymer composition and fluidity b, the design range of the geopolymer composition is determined, and the evaluation matrix X is constructed. The scheme set is A = (A1, A2, ..., A m ), the target set is B=(B1,B2,…,B m ), m is the number of items, and the evaluation matrix X is expressed as:

[0109]

[0110] Among them, plan A i and target B j The corresponding evaluation value is recorded as x ij , which is also the corresponding geopolymer component. The design target of the geopolymer component is quality-oriented. The evaluation matrix X is standardized to obtain the matrix V, where v ij =(x ij -min x j ) / (maxx j -min x j ) is the standard matrix V in solution A i and target B j The evaluation value of

[0111] Calculate the entropy and weight of the j-item target, the expression is:

[0112]

[0113] where e j is the entropy value of the j-item target, b j is the control constant representing the fluidity corresponding to the j-th target, w j is the weight of the j-item target;

[0114] According to the weight w j Calculate the positive ideal solution corresponding to the best solution Negative ideal solution corresponding to the worst solution Calculate the Euclidean distance d of the model based on the ideal solution i Get the relative closeness, the expression is:

[0115]

[0116] where η i For plan A i With positive ideal solution The closeness of is the positive ideal solution of the j-item target, is the negative ideal solution of the j-th target, is the Euclidean distance between the evaluation solution and the ideal solution;

[0117] Determine the design range of geopolymer components based on the schemes corresponding to the maximum degree of closeness and the average degree of closeness;

[0118] In the actual evaluation, the basic slurry of the geopolymer component is prepared to determine its fluidity Can determine the basis According to the schemes corresponding to the maximum proximity and the average proximity, the design range of the liquid-solid ratio L / S is determined to be 0.8-1.6, the range of the alkali activator concentration C is 27%-35%, and the range of the alkali activator modulus M is 1.4-1.8.

[0119] In this embodiment, the method for obtaining the mortar index library includes:

[0120] Prepare multiple groups of slurries with different geopolymer components according to the design range of geopolymer components, and conduct geotechnical tests to obtain the second slurry parameters;

[0121] According to the basic material parameters and the design range of geopolymer components, finite element models with different liquid-to-solid ratios are constructed to perform finite element simulation to obtain second simulation data; the second simulation data includes net slurry strength, net slurry plastic viscosity and net slurry yield stress; the net slurry strength is obtained by nonlinear analysis using ANSYS software; the net slurry rheological parameters are obtained by analyzing the net slurry discrete model established using PFC3D software;

[0122] Based on the same geopolymer component, the net slurry strength deviation coefficient is determined according to the second net slurry parameters and the second simulation data, and the expression is:

[0123]

[0124] where ξ fp is the net pulp strength deviation coefficient, n is the data volume, f pFi is the simulated strength of the ith group of pure slurry, f pti is the test strength of the ith group of pure slurry;

[0125] According to the above method, the deviation coefficient of plastic viscosity of the net slurry ξ is calculated. ηp and the net slurry yield stress deviation coefficient ξ τp , the second simulation data is corrected according to the deviation coefficient, and the corrected data and the corresponding liquid-solid ratio L / S, alkali activator concentration C and alkali activator modulus M are combined to form a clean slurry index library;

[0126] According to the design range of geopolymer components, fine aggregates with different volume fractions are added to the pure paste of different geopolymer components to prepare multiple groups of mortars, and geotechnical tests are carried out to obtain the mortar measurement parameters;

[0127] According to the basic material parameters, the design range of geopolymer components and the volume fraction V of fine aggregate in mortarS-m , construct finite element models with different liquid-to-solid ratios to perform finite element simulation to obtain third simulation data; the third simulation data includes mortar strength, mortar plastic viscosity, mortar yield stress and mortar slump; the mortar strength is obtained by nonlinear analysis using ANSYS software; the mortar plastic viscosity and yield stress are obtained by analyzing a mortar discrete model established using PFC3D software; the mortar slump is obtained by lattice Boltzmann method simulation based on Herschel-Bulkley model;

[0128] Based on the same geopolymer composition and fine aggregate volume fraction V in mortar S-m , the above method is used to calculate the deviation coefficient according to the mortar parameters and the third simulation data, and the third simulation data is corrected according to the deviation coefficient to obtain a mortar index library; the mortar index library data includes mortar strength f m , mortar slump S m , mortar plastic viscosity η m 、Mortar yield stress τ m , liquid-to-solid ratio L / S, alkali activator concentration C, alkali activator modulus M, coarse aggregate volume fraction V A and the volume fraction of fine aggregate in mortar V S-m ;

[0129] In the actual evaluation, the same basic neat slurry was used to prepare neat slurry and mortar according to different alkali activator concentrations C and alkali activator moduli M corresponding to a liquid-to-solid ratio of 1.1. Geotechnical tests and finite element simulations were carried out to obtain the neat slurry index library and mortar index library.

[0130] In this embodiment, the method for obtaining the second mortar design parameter set includes:

[0131] Set the net slurry design strength f p Range and corresponding coarse aggregate volume fraction V A and the volume fraction of fine aggregate in mortar V S-m To search the data set, the improved reptile search algorithm is used to search the slurry index library according to the search data set, and the first mortar design parameter set is matched;

[0132] The Sobol sequence is used to initialize the population position, and the expression is:

[0133]

[0134] Where X i,j is the initial population position of individual i in j-dimensional space, [u bs , l bs ] is the value range of the global solution, D(Y) is the value range of x=(x1,x2,…,x n) is the total variance of the function Y = f(x) of the independent variable, n is the population size, D i is the variance generated by the i-th independent variable;

[0135] Update the population position in the encirclement phase and the population position in the hunting phase according to the number of iterations;

[0136] Set sensitive parameters to control the search accuracy, add control factors to control the decreasing amplitude of the evolution factor, the percentage difference between the optimal solution and the current solution, and the evolution factor expression are:

[0137]

[0138] Where P i,j is the percentage difference between the optimal solution and the current solution, α=0.1 is a sensitive parameter, x best j (t) is the j-dimensional position of the current best solution, r3 is a random integer between [-1,1], ε is a minimum value to prevent the denominator from being 0, E s (t) is the evolution factor, δ is the [1,3] optimization index;

[0139] Calculate the fitness according to the population position and save the optimal solution, continue to calculate the fitness of the total population and update the position of the total population, and iterate until the maximum number of iterations is reached and then output the optimal solution. According to the optimal solution, obtain the first mortar design parameter set; the first mortar design parameter set includes the net mortar design strength f p , liquid-to-solid ratio L / S, alkali activator concentration C, alkali activator modulus M, fine aggregate volume fraction V in mortar S-m , volume fraction of coarse aggregate V A ;

[0140] Calculate the similarity between the first mortar design parameter set and the mortar index library, and convert the design strength f p , liquid-to-solid ratio L / S, alkali activator concentration C, alkali activator modulus M, coarse aggregate volume fraction V A and the volume fraction of fine aggregate in mortar V S-m The first mortar design parameter set is composed of characteristic vectors, and the similarity between the first mortar design parameter set and the characteristic vector corresponding to the mortar index library is calculated, and the data group with the highest similarity is selected to form the second mortar design parameter set; the second mortar design parameter set includes mortar strength f m , mortar slump S m , mortar plastic viscosity η m 、Mortar yield stress τ m , net slurry design strength f p , liquid-to-solid ratio L / S, alkali activator concentration C, alkali activator modulus M, fine aggregate volume fraction V in mortar S-m , volume fraction of coarse aggregate V A ;

[0141] In actual evaluation, according to the design strength f p Matching the pure paste index library for 45-53MPa yields: liquid-solid ratio 1.1, alkali activator concentration 31%-35%, alkali activator modulus 1.6-1.8, coarse aggregate volume fraction 30%-40%, fine aggregate volume fraction in mortar 44.61%-45.71%;

[0142] The similarity between the first mortar design parameter set and the mortar index library is calculated based on the similarity, and the mortar strength f is determined based on the similarity: m 43-54MPa, mortar plastic viscosity η m 5.34-11.14Pa·s, mortar yield stress τ m is 82-99.1Pa, and the slump is S cu 197-215 mm.

[0143] In this embodiment, the method for determining the first design formula set includes:

[0144] According to the volume fraction V of coarse aggregate mortar A Calculate the mortar film thickness H m , according to the mortar film thickness H m The physical parameters of mortar and aggregate determine the plastic viscosity threshold η of mortar m0 and mortar yield stress threshold τ m0 , set the mortar slump target value S m0 ;

[0145] The hypersphere classifier is used to classify the mortar slump S m , mortar plastic viscosity η m and mortar yield stress τ m To classify;

[0146] Set information set H = (h i ,q i ), h i =(h1,h2,…,h j ) represents an element in the set, q i Represents the class label. Assuming that the data is located inside the hypersphere (a, R), a is the center of the hypersphere, and R is the radius of the hypersphere. It is related to the data threshold. The hypersphere function expression is optimized by combining the Lagrangian formula. The expression is:

[0147]

[0148] Where L is the objective function, k(h i ,h j ) is the kernel function used to measure the element h i 、h jThe similarity of i , α j is the Lagrangian factor, C is a constant, and also serves as the upper bound of the Lagrangian factor;

[0149] The Gaussian kernel function is used to calculate the distance between the sample data point and the center of the hypersphere, and to determine the classification. The expression is:

[0150]

[0151] f(h)=sgn[R-d1(h)]

[0152] Where d1 is the distance between the sample data point h and the center of the hypersphere, k(h,h) is the Gaussian kernel function that calculates the similarity between h and itself, and k(h i ,h) Calculate the sample point h i The similarity with point h, f(h) is a classification function used to determine the classification of the sample data; the classification is that the sample data is located inside the corresponding hypersphere, and the sample data is located outside the corresponding hypersphere;

[0153] The slump of mortar S m The value is located in the corresponding sphere external data group, mortar plastic viscosity η m The value is located in the corresponding sphere external data group, mortar yield stress τ m The data sets whose values ​​lie inside the corresponding spheres are intersected to obtain the comprehensive data set;

[0154] The comprehensive data set and the corresponding geopolymer concrete mix ratio are combined into a first design formula set of geopolymer recycled concrete; the first design formula set includes the geopolymer concrete target strength f aim , liquid-to-solid ratio L / S, alkali activator concentration C, alkali activator modulus M, coarse aggregate mortar volume fraction V A , fine aggregate volume fraction V S , mortar plastic viscosity η m 、Mortar yield stress τ m ;

[0155] In the actual evaluation, the hypersphere classifier is used to narrow the ratio range to a data set, including (alkali activator concentration C, alkali activator modulus M, liquid-solid ratio L / S, coarse aggregate mortar volume fraction V A , Volume fraction of fine aggregate in mortar V S-m , mortar strength f m , mortar plastic viscosity η m 、Mortar yield stress τ m , slump is S m): 31%, 1.6, 1.1, 40%, 45%, 45.9MPa, 5.34Pa·s, 99.71Pa, 197mm; 35%, 1.6, 1.1, 35%, 45.71%, 52.6MPa, 11.14Pa·s, 82Pa, 245mm; 35%, 1.6, 1.1, 40%, 45%, 52.6MPa, 11.14Pa·s, 82Pa, 205mm and other groups of data, the above multiple groups of data are combined into the first design formula set of geopolymer recycled concrete.

[0156] In this embodiment, the method for obtaining the optimized ratio includes:

[0157] The first design formula set of geopolymer recycled concrete is divided into a training set and a test set;

[0158] Construct a geopolymer recycled concrete mix optimization model. The geopolymer recycled concrete mix optimization model includes an input layer, a hidden layer, and an output layer. The training set is input to the input layer, and the output layer is used to output the target strength f of the geopolymer recycled concrete. aim , liquid-to-solid ratio L / S, alkali activator concentration C, alkali activator modulus M, coarse aggregate mortar volume fraction V A and fine aggregate volume fraction V S , the replacement rate of recycled aggregate in coarse aggregate ρ1;

[0159] Support vector machine (SVM) is used as the base model, and Gaussian kernel function is used to map data to high-dimensional feature space. The nonlinearly related data in the first design formula set are processed and the optimal plane is found. The grid search training model is used to select the optimal parameters. The geopolymer recycled concrete mix ratio objective function is used to evaluate the quality of each generation of solutions. Solutions and model weights that meet the requirements of environmental friendliness, high cost performance and excellent mortar working performance are selected. Multiple iterations are performed until the evaluation value of the geopolymer recycled concrete mix ratio objective function no longer decreases.

[0160] The objective function expression of geopolymer recycled concrete mix ratio is:

[0161]

[0162]

[0163] Where y(f aim ) is the geopolymer recycled concrete mix ratio objective function, w1 and w2 are economic weights, w3 is the environmental weight, w4 is the work performance weight, ρ1 is the replacement rate of recycled aggregate in coarse aggregate, k is the number of commercial crushed stone particle size categories in coarse aggregate, ρ 2i is the proportion of the i-th type of gravel, PRICE iis the price of the i-th type of crushed stone, l1(L / S,C) is the alkali activator concentration function, which means the alkali activator concentration C corresponding to the liquid-to-solid ratio L / S, V p is the mortar volume fraction, PRICE C is the price of alkali activator, ρ0 is the standard recycled aggregate replacement rate, μ is the carbon emission impact coefficient, l2(ρ1V A ,C CO2 ) is a function of CO2 concentration, indicating the reduction in carbon emissions from replacing crushed stone with recycled aggregate;

[0164] The mean square loss function is used to evaluate the economic cost difference between the predicted value and the standard, the cross entropy loss function is used to measure the environmental impact difference between the predicted value and the standard value, L1 and L2 regularization is used to prevent overfitting and improve the generalization ability of the model, and the test set is used to evaluate the geopolymer recycled concrete mix optimization model;

[0165] Inputting the geopolymer recycled concrete mix ratio to be optimized into the geopolymer recycled concrete mix ratio optimization model to obtain the optimized geopolymer recycled concrete mix ratio;

[0166] In the actual evaluation, the first design formula set was obtained according to the mix ratio to be optimized and the target strength. The target strength and the first design formula set were input into the geopolymer recycled concrete mix ratio optimization model to obtain the optimal mix ratio: alkali activator concentration C is 31%, alkali activator modulus M is 1.6, liquid-solid ratio L / S is 1.1, and coarse aggregate mortar volume fraction V is 1.6. A The volume fraction of fine aggregate in mortar is 35%. S-m is 44.61%, and the replacement rate of recycled aggregate in coarse aggregate is 30%;

[0167] According to the optimized ratio, geopolymer recycled concrete was prepared. The mortar strength f m is 45.9MPa, concrete strength f cu is 55.3MPa, and the plastic viscosity of mortar is η m is 5.34Pa·s, mortar yield stress τ m is 99.71Pa, slump S m is 215 mm; the corresponding plastic viscosity threshold minη m is 3.41Pa·s, and the yield stress threshold maxτ m is 161Pa, and the slump target value is S m0 160mm;

[0168] The strength of geopolymer recycled concrete prepared according to the optimized ratio reaches the target strength of 55MPa, and the working performance of mortar (rheological parameters, slump) meets the requirements.

[0169] In this embodiment, the geopolymer recycled concrete comprises: coarse aggregate, fine aggregate and pure paste; the coarse aggregate is replaced by recycled coarse aggregate, and the replacement rate is ρ1; an alkali activator and aluminosilicate raw materials are added to the pure paste, and the mass ratio of the alkali activator and the aluminosilicate raw materials is represented by the liquid-solid ratio L / S; the alkali activator is a mixture of solid sodium hydroxide, sodium silicate solution and water; the aluminosilicate raw materials are composed of granulated blast furnace slag and metakaolin;

[0170] In the actual evaluation, the aluminosilicate raw material consists of granulated blast furnace slag and metakaolin in a mass ratio of 2:3; granulated blast furnace slag is a highly active acidic slag, and its main chemical composition is 30.23wt% SiO2, 13.72wt% Al2O3 and 44.06wt% CaO; kaolin is obtained by calcining kaolin at 700°C for 24 hours, and its main chemical composition is 52.53wt% SiO2 and 45.42wt% Al2O3; the solute content of sodium silicate solution is 34.89%, and the modulus is 3.28; sodium hydroxide is flaky sodium hydroxide solid with a purity of 99.8%; the alkali activator concentration C is the ratio of the total mass of water glass solute and sodium hydroxide solid to the mass of the alkali activator, the alkali activator modulus M is the molar ratio of SiO2 to Na2O in the alkali activator, and the liquid-solid ratio L / S is the mass ratio of the alkali activator to the mineral raw material.

[0171] Secondly, the artificial intelligence-based geopolymer recycled concrete mix optimization system includes:

[0172] Simulation module: connected to the data acquisition module; used for finite element simulation of the geopolymer recycled concrete, paste and mortar;

[0173] Data acquisition module: connected to the simulation module, the data processing module and the intelligent supervision module; used to obtain geotechnical test data and finite element simulation data, and pre-process the geotechnical test data and the simulation data; the geotechnical test data includes the first slurry parameter, the second slurry parameter and the mortar measurement parameter; the finite element simulation data includes the first simulation data, the second simulation data and the third simulation data;

[0174] Data processing module: connects the data acquisition module, the optimization model module and the intelligent supervision module; is used to determine the net slurry design strength f p The range is used to determine the liquid-to-solid ratio L / S design range, which is used to correct the finite element simulation data of the mortar and mortar to obtain a mortar index library and a mortar index library, which are used to match the first mortar design parameter set and the second mortar design parameter set, and are used to screen the mortar rheological parameters to determine the first design formula set of geopolymer recycled concrete;

[0175] Optimization model module: connected to the data processing module and the intelligent supervision module; used to construct a geopolymer recycled concrete mix ratio optimization model according to the first design formula set of the geopolymer recycled concrete, and input the geopolymer recycled concrete mix ratio to be optimized into the geopolymer recycled concrete mix ratio optimization model to obtain the optimized geopolymer recycled concrete mix ratio;

[0176] Intelligent supervision module: connecting the data acquisition module, the data processing module and the optimization model module; used for storing, viewing and managing the geotechnical test data, the finite element simulation data, the first design formula set and the optimized mix ratio of the geopolymer recycled concrete.

[0177] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for optimizing the mix ratio of geopolymer recycled concrete based on artificial intelligence, characterized in that: The following steps are involved: S1. Construct a finite element model of geopolymer recycled concrete with different aggregate size distributions, conduct numerical simulation analysis to obtain the first simulation data, and determine the target strength f of geopolymer recycled concrete. aim , according to the concrete target strength f aim The net slurry design strength f is determined by the first simulation data p scope; S2, preparing slurries of different geopolymer components to obtain first slurry parameters, and determining the design range of the geopolymer components according to the first slurry parameters; S3, performing finite element simulation and geotechnical test of the net slurry according to the design range of the geopolymer component to obtain a net slurry index library, and adding fine aggregates with different volume fractions based on the design range of the component group to perform finite element simulation and geotechnical test of mortar to obtain a mortar index library; S4, the slurry design strength f p Range matching the mortar index library to obtain a first mortar design parameter set, matching the mortar index library according to the first mortar design parameter set to obtain a second mortar design parameter set, screening the second mortar design parameter set, and determining a first design formula set of geopolymer recycled concrete according to the screening result; S5. Constructing a geopolymer recycled concrete mix ratio optimization model according to the first design formula, and inputting the mix ratio to be optimized into the geopolymer recycled concrete mix ratio optimization model to obtain an optimized mix ratio of the geopolymer recycled concrete.

2. The method for optimizing the mix ratio of geopolymer recycled concrete based on artificial intelligence according to claim 1, characterized in that: Determine the design strength f of the slurry p A range of methods, including: Conduct geotechnical tests to obtain basic material parameters, and construct finite element models of geopolymer recycled concrete with different aggregate particle size distributions based on the material parameters to perform numerical simulation analysis to obtain the first simulation data; Determine multiple groups of commonly used net slurry strengths, prepare multiple groups of geopolymer recycled concrete specimens with different aggregate particle size distributions, conduct geotechnical tests and obtain the first geotechnical data; Match the first geotechnical data with the first simulation data according to the principle of consistency of the net slurry strength and aggregate particle size gradation, calculate the specimen strength deviation of the matching data to determine the simulation measured deviation coefficient, and correct the first simulation data according to the simulation measured deviation coefficient to obtain the geopolymer recycled concrete index library; Determine the target strength f of geopolymer recycled concrete based on the original mix ratio aim , according to the concrete target strength f aim and concrete framework theory to determine the coarse aggregate strength f A and mortar strength f m scope; According to the coarse aggregate strength f A The range of the coarse aggregate contribution strength of the geopolymer recycled concrete index library is matched to obtain the corresponding coarse aggregate volume fraction V A , fine aggregate volume fraction V S and mortar strength f m scope; The volume fraction of coarse aggregate V A , fine aggregate volume fraction V S The net paste strength function is input with mortar strength to obtain the range of net paste design strength.

3. The method for optimizing the mix ratio of geopolymer recycled concrete based on artificial intelligence according to claim 1, characterized in that: The method for determining the design range of the geopolymer composition comprises: Prepare basic slurries with different geopolymer components to measure first slurry parameters; the first slurry parameters include geopolymer components and fluidity b; the geopolymer components include liquid-to-solid ratio L / S, alkali activator concentration C and alkali activator modulus M; According to the geopolymer composition and fluidity b, the design range of the geopolymer composition is determined, and the evaluation matrix X is constructed. The scheme set is A = (A1, A2, ..., A m ), the target set is B=(B1,B2,…,B m ), m is the number of items, and the evaluation matrix X is expressed as: Among them, plan A i and target B j The corresponding evaluation value is recorded as x ij , which is also the corresponding geopolymer component. The design target of the geopolymer component is quality-oriented. The evaluation matrix X is standardized to obtain the matrix V, where v ij =(x ij -min x j ) / (max x j -minx j ) is the standard matrix V in solution A i and target B j The evaluation value of Calculate the entropy and weight of the j-item target, the expression is: where e j is the entropy value of the j-item target, b j is the control constant representing the fluidity corresponding to the j-th target, w j is the weight of the j-item target; According to the weight w j Calculate the positive ideal solution corresponding to the best solution Negative ideal solution corresponding to the worst solution Calculate the Euclidean distance d of the model based on the ideal solution i Get the relative closeness, the expression is: where η i For plan A i With positive ideal solution The closeness of is the positive ideal solution of the j-item target, is the negative ideal solution of the j-th target, is the Euclidean distance between the evaluation solution and the ideal solution; The design range of geopolymer components is determined based on the schemes corresponding to the maximum degree of proximity and the average degree of proximity.

4. The method for optimizing the mix ratio of geopolymer recycled concrete based on artificial intelligence according to claim 1, characterized in that: The method for obtaining the mortar index library comprises: Prepare multiple groups of slurries with different geopolymer components according to the design range of geopolymer components, and conduct geotechnical tests to obtain the second slurry parameters; According to the basic material parameters and the design range of geopolymer components, finite element models with different liquid-to-solid ratios are constructed to perform finite element simulation to obtain second simulation data; the second simulation data includes net slurry strength, net slurry plastic viscosity and net slurry yield stress; the net slurry strength is obtained by nonlinear analysis using ANSYS software; the net slurry rheological parameters are obtained by analyzing the net slurry discrete model established using PFC3D software; Based on the same geopolymer component, the net slurry strength deviation coefficient is determined according to the second net slurry parameters and the second simulation data, and the expression is: where ξ fp is the net pulp strength deviation coefficient, n is the data volume, f pFi is the simulated strength of the ith group of pure slurry, f pti is the test strength of the ith group of pure slurry; According to the above method, the deviation coefficient of plastic viscosity of the net slurry ξ is calculated. ηp and the net slurry yield stress deviation coefficient ξ τp , the second simulation data is corrected according to the deviation coefficient, and the corrected data and the corresponding liquid-solid ratio L / S, alkali activator concentration C and alkali activator modulus M are combined to form a clean slurry index library; According to the design range of geopolymer components, fine aggregates with different volume fractions are added to the pure paste of different geopolymer components to prepare multiple groups of mortars, and geotechnical tests are carried out to obtain the mortar measurement parameters; According to the basic material parameters, the design range of geopolymer components and the volume fraction V of fine aggregate in mortar S-m , construct finite element models with different liquid-to-solid ratios to perform finite element simulation to obtain third simulation data; the third simulation data includes mortar strength, mortar plastic viscosity, mortar yield stress and mortar slump; the mortar strength is obtained by nonlinear analysis using ANSYS software; the mortar plastic viscosity and yield stress are obtained by analyzing a mortar discrete model established using PFC3D software; the mortar slump is obtained by lattice Boltzmann method simulation based on Herschel-Bulkley model; Based on the same geopolymer composition and fine aggregate volume fraction V in mortar S-m The above method is used to calculate the deviation coefficient according to the mortar parameters and the third simulation data, and the third simulation data is corrected according to the deviation coefficient to obtain the mortar index library.

5. The method for optimizing the mix ratio of geopolymer recycled concrete based on artificial intelligence according to claim 1, characterized in that: The method for obtaining the second mortar design parameter set comprises: Set the net slurry design strength f p Range and corresponding coarse aggregate volume fraction V A and the volume fraction of fine aggregate in mortar V S-m To search the data set, the improved reptile search algorithm is used to search the slurry index library according to the search data set, and the first mortar design parameter set is matched; The Sobol sequence is used to initialize the population position, and the expression is: Where X i,j is the initial population position of individual i in j-dimensional space, [u bs , l bs ] is the value range of the global solution, D(Y) is the value range of x=(x1,x2,…,x n ) is the total variance of the function Y = f(x) of the independent variable, n is the population size, D i is the variance generated by the ith independent variable; Update the population position in the encirclement phase and the population position in the hunting phase according to the number of iterations; Set sensitive parameters to control the search accuracy, add control factors to control the decreasing amplitude of the evolution factor, the percentage difference between the optimal solution and the current solution, and the evolution factor expression are: Where P i,j is the percentage difference between the optimal solution and the current solution, α=0.1 is a sensitive parameter, x bestj (t) is the j-dimensional position of the current best solution, r3 is a random integer between [-1,1], ε is a minimum value to prevent the denominator from being 0, E s (t) is the evolution factor, δ is the [1,3] optimization index; Calculate the fitness according to the population position and save the optimal solution, continue to calculate the fitness of the total population and update the position of the total population, and iterate until the maximum number of iterations is reached and then output the optimal solution, and obtain the first mortar design parameter set according to the optimal solution; Calculate the similarity between the first mortar design parameter set and the mortar index library, and convert the design strength f p , liquid-to-solid ratio L / S, alkali activator concentration C, alkali activator modulus M, coarse aggregate volume fraction V A and the volume fraction of fine aggregate in mortar V S-m The similarity between the first mortar design parameter set and the corresponding characteristic vector of the mortar index library is calculated, and the data group with the highest similarity is selected to form the second mortar design parameter set.

6. The method for optimizing the mix ratio of geopolymer recycled concrete based on artificial intelligence according to claim 1, characterized in that: The method for determining the first design formula set comprises: According to the volume fraction V of coarse aggregate mortar A Calculate the mortar film thickness H m , according to the mortar film thickness H m The physical parameters of mortar and aggregate determine the plastic viscosity threshold η of mortar m0 and mortar yield stress threshold τ m0 , set the mortar slump target value S m0 ; The hypersphere classifier is used to classify the mortar slump S m , mortar plastic viscosity η m and mortar yield stress τ m To classify; Set information set H = (h i ,q i ), h i =(h1,h2,…,h j ) represents an element in the set, q i Represents the class label. Assuming that the data is located inside the hypersphere (a, R), a is the center of the hypersphere, and R is the radius of the hypersphere. It is related to the data threshold. The hypersphere function expression is optimized by combining the Lagrangian formula. The expression is: Where L is the objective function, k(h i ,h j ) is the kernel function used to measure the element h i 、h j The similarity of i , α j is the Lagrangian factor, C is a constant, and also serves as the upper bound of the Lagrangian factor; The Gaussian kernel function is used to calculate the distance between the sample data point and the center of the hypersphere, and to determine the classification. The expression is: f(h)=sgn[R-d1(h)] Where d1 is the distance between the sample data point h and the center of the hypersphere, k(h,h) is the Gaussian kernel function that calculates the similarity between h and itself, and k(h i ,h) Calculate the sample point h i The similarity with point h, f(h) is a classification function used to determine the classification of the sample data; the classification is that the sample data is located inside the corresponding hypersphere, and the sample data is located outside the corresponding hypersphere; The slump of mortar S m The value is located in the corresponding sphere external data group, mortar plastic viscosity η m The value is located in the corresponding sphere external data group, mortar yield stress τ m The data sets whose values ​​are inside the corresponding spheres are intersected to obtain the comprehensive data set; The comprehensive data set and the corresponding geopolymer concrete mix proportions are combined to form the first design formulation set of geopolymer recycled concrete.

7. The method for optimizing the mix ratio of geopolymer recycled concrete based on artificial intelligence according to claim 1, characterized in that: The method for obtaining the optimized ratio comprises: The first design formula set of geopolymer recycled concrete is divided into a training set and a test set; Constructing a geopolymer recycled concrete mix ratio optimization model, which includes an input layer, a hidden layer and an output layer; Support vector machine (SVM) is used as the base model, and Gaussian kernel function is used to map data into high-dimensional feature space. The nonlinearly related data in the first design formula set are processed, and the optimal plane is found. The grid search training model is used to select the optimal parameters. The geopolymer recycled concrete mix ratio objective function is used to evaluate the quality of each generation of solutions, and the solutions and model weights that meet the requirements of environmental friendliness, high cost performance and excellent mortar working performance are selected. Multiple iterations are performed until the evaluation value of the geopolymer recycled concrete mix ratio objective function no longer decreases. The objective function expression of geopolymer recycled concrete mix ratio is: Where y(f aim ) is the geopolymer recycled concrete mix ratio objective function, w1 and w2 are economic weights, w3 is the environmental weight, w4 is the work performance weight, ρ1 is the replacement rate of recycled aggregate in coarse aggregate, k is the number of commercial crushed stone particle size categories in coarse aggregate, ρ 2i is the proportion of the i-th type of gravel, PRICE i is the price of the i-th type of crushed stone, l1(L / S,C) is the alkali activator concentration function, which means the alkali activator concentration C corresponding to the liquid-to-solid ratio L / S, V p is the mortar volume fraction, PRICE C is the price of alkali activator, ρ0 is the standard recycled aggregate replacement rate, μ is the carbon emission impact coefficient, l2(ρ1V A ,C CO2 ) is a function of CO2 concentration, indicating the reduction in carbon emissions from replacing crushed stone with recycled aggregate; The mean square loss function is used to evaluate the economic cost difference between the predicted value and the standard, the cross entropy loss function is used to measure the environmental impact difference between the predicted value and the standard value, L1 and L2 regularization is used to prevent overfitting and improve the generalization ability of the model, and the test set is used to evaluate the geopolymer recycled concrete mix optimization model; The mix ratio to be optimized is input into the geopolymer recycled concrete mix ratio optimization model to obtain the optimized mix ratio of geopolymer recycled concrete.

8. The method for optimizing the mix ratio of geopolymer recycled concrete based on artificial intelligence according to claim 1, characterized in that: The geopolymer recycled concrete comprises: coarse aggregate, fine aggregate and pure paste; the coarse aggregate is replaced by recycled coarse aggregate in a replacement rate of ρ1; an alkali activator and aluminosilicate raw materials are added to the pure paste, and the mass ratio of the alkali activator and the aluminosilicate raw materials is represented by a liquid-solid ratio L / S; the alkali activator is a mixture of solid sodium hydroxide, sodium silicate solution and water; the aluminosilicate raw materials are composed of granular blast furnace slag and metakaolin.

9. An artificial intelligence-based geopolymer recycled concrete mix ratio optimization system for executing the method according to any one of claims 1 to 8, comprising: Simulation module: connected to data acquisition module; Used for conducting finite element simulation of the geopolymer recycled concrete, paste and mortar; Data acquisition module: connects the simulation module, data processing module and intelligent supervision module; Used to obtain geotechnical test data and finite element simulation data, and preprocess the geotechnical test data and the simulation data; the geotechnical test data includes a first slurry parameter, a second slurry parameter and a mortar measurement parameter; the finite element simulation data includes a first simulation data, a second simulation data and a third simulation data; Data processing module: connects the data acquisition module, optimization model module and intelligent supervision module; Used to determine the net slurry design strength f p The range is used to determine the liquid-to-solid ratio L / S design range, which is used to correct the finite element simulation data of the mortar and mortar to obtain a mortar index library and a mortar index library, which are used to match the first mortar design parameter set and the second mortar design parameter set, and are used to screen the mortar rheological parameters to determine the first design formula set of geopolymer recycled concrete; Optimization model module: connects the data processing module and the intelligent supervision module; Used to construct a geopolymer recycled concrete mix ratio optimization model according to the first design formula set of geopolymer recycled concrete, and input the geopolymer recycled concrete mix ratio to be optimized into the geopolymer recycled concrete mix ratio optimization model to obtain the optimized geopolymer recycled concrete mix ratio; Intelligent supervision module: connecting the data acquisition module, data processing module and optimization model module; Used to store, view and manage the geotechnical test data, the finite element simulation data, the first design formula set and the optimized mix ratio of geopolymer recycled concrete.

Images