Easily-grouted dry-shrinkage-resistant earthen ruins crack repair compatible slurry admixture proportion optimization method

Abstract

The invention provides an earthen archaeological site crack repair compatible slurry admixture proportion optimization method which is easy to fill and resistant to dry shrinkage, and the method comprises the following steps: collecting data of mortar compressive strength (KY) and admixture proportion to form a data set, taking parameters influencing the size of the mortar KY as analysis objects, and carrying out quantitative optimization to obtain a mortar compressive strength data set; a grey wolf optimization-gradient boosting tree (GWO-GBDT) model is used for processing and determining a mortar KY prediction result, and a KY analysis function is established; the method comprises the following steps: collecting at least one slurry data of carbon emission data, cost data and energy consumption data of materials used in the mortar slurry, and establishing at least one of a carbon emission analysis function, a cost analysis function or an energy consumption analysis function of the mortar slurry; and establishing an admixture proportion optimization model of the pourable mortar based on the analytical function. According to the method, the performance of the mortar material and the environmental protection property and economical efficiency of the material are considered, a user can conveniently select a suitable decision scheme according to own requirements under the background of green development of the current engineering construction industry, and an efficient, convenient and scientific method is provided for material selection and performance analysis of earthen site repair engineering.

Description

Technical Field

[0001] The invention relates to the technical field of building materials, and in particular to an optimal method for mixing proportions of compatible slurry admixtures for repairing earthen ruins cracks that is easy to irrigate and resistant to shrinkage. Background Art

[0002] Earthen sites are cultural relics that contain valuable historical information. However, under the influence of natural and human factors, these sites are often covered with soil damage of varying sizes, including cracks and fissures, which will seriously affect the safety of the soil. Pouring mortar is a common method to repair such damage, and being able to develop mortar with superior performance is an important part of site maintenance projects.

[0003] However, the method of measuring mortar performance through experiments not only causes a lot of waste, but also fails to intuitively analyze the impact of admixtures on mortar performance, which is not in line with the current development trend of digitalization. In addition to engineering performance, green development is also receiving more and more attention. How to make the pourable mortar used for the maintenance of earthen sites achieve the minimum resource consumption on the basis of meeting performance requirements is also the focus of current research in the field of building materials. However, there are currently few digital optimization cases for this type of mortar, especially scientific analysis methods that take into account the combination of mortar performance with low carbon and energy saving.

[0004] The invention patent with application number 202210742732.2 discloses a method for evaluating the compatibility of fissure grouting materials of earthen sites and the soil of the site. The steps are: calculate the drought index of the soil of the site, determine the climate level according to the drought index, and determine the evaluation index according to the climate level; assign subjective weights to each evaluation index, use the test values ​​to calculate the objective weights, and average the subjective weights and objective weights to obtain the combined weights; use the TOPSIS method to process the test values ​​of each evaluation index of the grouting slurry and the site soil to calculate the relative progress; determine the compatibility evaluation grade and the improvement direction of the grouting slurry performance: determine the compatibility evaluation grade according to the relative progress, and give the improvement direction of the grouting slurry material according to the weighted and unweighted ranking of the performance differences. The above invention comprehensively considers the common evaluation indicators of different climate levels of earthen sites, and assigns weights to both subjective and objective weights, gives the improvement direction of the material, and provides a good reference value for compatibility evaluation. However, the above invention only discusses the performance of slurry with a single ratio, and does not consider the influence of different admixture amounts on the performance of slurry, which makes it impossible to be directly applied to the optimization of the ratio of slurry for earthen site restoration, and the invention's method of obtaining data through a large number of experiments will waste a lot of manpower, material resources and time costs. In addition, the use of TOPSIS to approximate the ideal solution is highly sensitive to data. If there are outliers or missing values, it will have a great impact on the decision-making results, and it is not suitable for multi-objective optimization analysis that requires complex calculations. Summary of the invention

[0005] In order to solve the technical problem that the experimental method for measuring mortar performance cannot intuitively analyze the influence of admixtures on mortar performance, the present invention proposes a method for optimizing the proportion of admixtures for a compatible slurry that is easy to pour and resistant to shrinkage for repairing cracks in earthen sites. This method not only solves the problem that existing research lacks analysis of the mechanical properties of this type of mortar, but also conforms to the current trend of green development in the field of building materials, and facilitates users to efficiently and conveniently obtain corresponding digital proportion optimization schemes according to their own needs.

[0006] In order to achieve the above object, the technical solution of the present invention is implemented as follows: a method for optimizing the proportion of admixtures of pourable mortar for repairing earthen ruins, the steps of which are as follows:

[0007] Step S1: collect data on the compressive strength (KY) of pourable mortar and its admixture ratio to form a data set, take the parameters that affect the mortar KY size as the analysis object of the data set, and quantify and optimize the analysis object to obtain an optimal mortar KY data set, use the Gray Wolf Optimization-Gradient Boosting Tree (GWO-GBDT) model to process the optimal mortar KY data set to determine the mortar KY prediction result, and establish a mortar KY analysis function;

[0008] Step S2: According to actual engineering requirements, at least one slurry material data of carbon emission data, cost data, and energy consumption data of materials used in mortar slurry is collected;

[0009] Step S3: establishing at least one of a carbon emission analysis function, a cost analysis function or an energy consumption analysis function of the mortar slurry that meets actual engineering requirements based on the collected slurry data;

[0010] Step S4: Minimize the comprehensive analysis function of the mortar KY analysis function and at least one of the carbon emission analysis function, the cost analysis function or the energy consumption analysis function to obtain an optimal model for the admixture ratio of the pourable mortar.

[0011] Preferably, the factors affecting the compressive strength of mortar include the proportion of effective chemical reaction components, the physical value of admixture dosage, the performance of mortar bonding materials and the curing conditions of mortar stone body; the proportion of effective chemical reaction components is the ratio between the amount of each compound participating in the chemical reaction during the mortar preparation process and the amount of the compound with the most intense chemical reaction in the mortar; the physical value of admixture dosage is the measured density data of each solid material in the mortar multiplied by the mass proportion of the solid material added to the mortar; the indicators corresponding to the performance of mortar bonding materials include water-cement ratio, mass concentration and viscosity; the indicators corresponding to the curing conditions of mortar stone body include stone body curing temperature, curing humidity and curing time. Among them, the water-cement ratio is the mass ratio of the binder to the solid admixture in the mortar.

[0012] Preferably, the method for quantifying and optimizing the analysis objects is as follows: based on the importance of the KY influence on the analysis object, the quantitative results of the importance of each analysis object are obtained, and then the objects with less importance are screened out to obtain the optimized mortar compressive strength data set.

[0013] Preferably, the data of each analysis object in the data set are collectively composed into an analysis matrix W, and the data of the compressive strength of the data set are composed into an analysis matrix t. The analysis matrix W and the analysis matrix t are used to calculate the relevant weight vectors and weight correction coefficients of the importance of each analysis object, and the estimated coefficient ∈ of the contribution of each analysis object to the compressive strength ratio is calculated; the analysis matrix t is input into the GWO-GBDT model to obtain the predicted value of the compressive strength corresponding to each analysis object, and the analysis value of the importance of the analysis object in the GWO-GBDT model is calculated by the estimated coefficient ∈, and the analysis value is normalized to obtain the quantitative value of the importance of each analysis object, and multiple analysis objects with the smallest ranking in the quantitative value are screened out, and the remaining analysis objects form the preferred mortar KY data set;

[0014] The preferred mortar compressive strength data set must retain at least one indicator, namely, the proportion of effective chemical reaction components of raw materials and the physical value of admixture dosage; the mortar bonding material performance and mortar stone body curing conditions must each retain at least one indicator as an analysis object.

[0015] Preferably, the calculation method of the estimated coefficient ∈ is: ∈ = β′ 2 γ′ 2 ;

[0016] The relevant weight vector β′ and weight correction coefficient γ′ of the importance of each analysis object satisfy:

[0017] W=PΔQ′

[0018] V=PQ′

[0019] β′=(V′V) -1 V′t=QP′t

[0020] γ′=(V′V) -1 V′W=QP′W

[0021] Where P represents the eigenvector matrix corresponding to the matrix WW′, Q represents the eigenvector matrix of the matrix W′W, Q′ is the transpose of the eigenvector matrix Q, Δ represents the diagonal matrix composed of the square root values ​​of the eigenvalues ​​of the matrices WW′ and W′W, V represents the best fitting orthogonal approximation matrix of the analysis matrix W, and β′ is the relevant weight vector of the importance of each analysis object;

[0022] The calculation method of the quantitative value of the importance of each analysis object is:

[0023] The KY prediction value y corresponding to each analysis object in the analysis matrix t is calculated by the GWO-GBDT model P and average Quantitative value W of the importance of different analysis objects j satisfy:

[0024] logit(y P )=ln[y P / (1-y P )]

[0025]

[0026]

[0027]

[0028] Among them, logit(y P ) is the predicted value y P The natural logarithm of is the natural logit(y P ), R 2 is the determination coefficient of the analytical model, y is the measured value, S x is the standard deviation of each analysis object in the analysis model, β M is the analytical value of the importance of different analysis objects in the model, W j For analysis value The quantified value of the importance of each analysis object j after normalization calculation; n represents the total number of analysis objects.

[0029] Preferably, the GWO algorithm in the gray wolf optimization-gradient boosting tree model selects the values ​​of the positions of the top three gray wolf individuals in the wolf pack in terms of fitness as the positions of the simulated wolves, selects random search or shrinkage distance by judging the distance between the simulated wolves and the prey, calculates and updates the positions of each gray wolf, and thus obtains the optimal hyperparameters; the optimal hyperparameters optimize the hyperparameters in the GBDT model, and the hyperparameters of the GBDT model optimized by GWO include the maximum number of weak learners, step size and subsampling range in the boosting framework parameters; the GBDT model processes the preferred mortar compressive strength data set, and accumulates the learning results and the obtained weights to obtain the mortar compressive strength prediction results.

[0030] Preferably, the carbon emission data, cost data, and energy consumption data of each material in the mortar include the production process and the transportation process;

[0031] The carbon emission analysis function g CEC The calculation formula is:

[0032]

[0033] Among them, g CEC represents the carbon emission value corresponding to any preferred mortar raw material ratio, l represents the amount of raw materials used in the mortar, Q r Indicates the content of each raw material in the mortar, CEC r Indicates the CEC coefficient of each raw material, LD r Indicates the transport distance of each material, CEC Tr Indicates the unit CEC coefficient during transportation;

[0034] The cost analysis function g MC The calculation formula is:

[0035]

[0036] Among them, g MC It represents the cost value corresponding to any of the mortar raw material ratios involved in the optimization, MC r Indicates the MC corresponding to each material production process, MC Tr Indicates the MC corresponding to each material transportation process;

[0037] The energy consumption analysis function g NH The calculation formula is:

[0038]

[0039] Among them, g NH It represents the energy consumption value corresponding to any ratio of raw materials involved in the optimization of mortar, NH r represents the energy intensity of each raw material, NH Tr Represents the unit energy consumption corresponding to the transportation process of each raw material.

[0040] Preferably, the method for obtaining the optimal model of admixture ratio for pourable mortar is:

[0041] B1: Calculate and obtain the mortar KY analysis function g in the data set respectively KY And the analysis function g CEC , g MC and g NH ;

[0042] B2: Use the weighted sum method to construct a comprehensive analysis function H = v1·g KY +v2·g CEC +v3·g MC +v4·g NH ;

[0043] Among them, v1, v2, v3 and v4 represent the weight values ​​of mortar compressive strength, carbon emissions, cost and energy consumption respectively and satisfy:

[0044]

[0045] B3: Use the multi-objective beetle beard algorithm or the multi-objective gray wolf optimization algorithm to calculate the Pareto frontier solution set of mortar ratio, and use the VIKOR method to calculate the benefit ratio C for each solution in the solution set d The smallest result is the optimal mix ratio of mortar.

[0046] Preferably, the benefit ratio C d The calculation method is:

[0047]

[0048] Where d represents any solution in the Pareto frontier solution set, u represents the number of target values, and Respectively represent the analysis function g KY , g CEC , g MC and g NH The e-th ideal and non-ideal solution of the single-objective optimization of any objective in , F de Denotes the analytical function g corresponding to solution d KY , g CEC , g MC and g NH The value of any target in G d represents the group analysis index, and They represent the group analysis indexes with ideal solution and non-ideal solution as reference, L d represents individual analysis indicators, and They represent the individual regret values ​​with reference to the ideal solution and the non-ideal solution, respectively. d The mortar ratio corresponding to the solution with the smallest value is the optimal ratio.

[0049] Preferably, the KY analysis function is determined by the KY prediction value obtained after the GWO-GBDT model prediction, and the KY prediction value of the mortar after 28 days of curing is used as the data required for calculation when establishing the KY analysis function of the admixture ratio optimization model;

[0050] The weight value is determined using the entropy weight method + CRITIC method;

[0051] The error BE between the obtained Pareto front solutions t for:

[0052]

[0053] Among them, BE trepresents the Pareto frontier solution point f0(s t ) and g0(x t ) in x t , δ represents the actual Pareto frontier, and K represents the number of Pareto frontier points found by the multi-objective optimization algorithm.

[0054] Compared with the prior art, the present invention has the following beneficial effects:

[0055] The compressive strength data of pourable mortar for earthen site restoration and the material ratio composition data set corresponding to its admixtures were collected. The parameters affecting the compressive strength of the mortar, such as the proportion of effective chemical reaction components, the physical value of the admixture dosage, the properties of the mortar bonding materials, and the curing conditions of the mortar stone body, were set as the analysis objects of the model data set. The importance of each analysis object was quantitatively analyzed and screened. Based on the above analysis, the Grey Wolf Optimization-Gradient Boosting Tree (GWO-GBDT) algorithm was used to establish the prediction result of the mortar compressive strength. When modeling and analyzing the mortar compressive strength, the influencing factors such as the chemical reaction inside the mortar, the physical indicators of the mortar materials, the properties of the adhesive, and the curing conditions of the stone body were considered. The previous modeling method only considered the dosage and curing time of each mortar material, and the indicators affecting the mechanical properties of the mortar were comprehensively considered. Linear analysis models are established in combination with the environmental performance, economic performance and energy consumption of each mortar admixture, so as to optimize the mortar material ratio that meets the performance standard and is environmentally friendly and economical; the Pareto frontier solution set of the mortar ratio is calculated using the multi-objective beetle whisker algorithm (MOBAS) and the multi-objective gray wolf optimization algorithm (MOGWO), and the multi-criteria compromise solution sorting method (VIKOR) is used to optimize the pourable mortar material ratio for earthen site maintenance with good performance, low pollution, low cost and low energy consumption. The invention establishes a method for optimizing the model analysis object by calculating the importance of each analysis object on the mortar KY, so that the availability of the analysis object used for inverse modeling is more intuitive. The invention can flexibly select the indicators for multi-objective optimization analysis according to the actual needs of the user, so as to maximize the engineering efficiency. The invention takes into account the performance of mortar slurry materials as well as the environmental protection and economy of the materials, so that users can choose the appropriate decision-making plan according to their own needs under the background of green development of the current engineering construction industry, and provide an efficient, convenient and scientific method for material selection and performance analysis of earthen site restoration projects. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0057] Figure 1 It is a flow chart of the present invention.

[0058] Figure 2 Schematic diagram of the relationship between ideal solutions, non-ideal solutions and Pareto frontier solution sets.

[0059] Figure 3 The figure is a flow chart of calculating the mortar KY prediction results using the GWO-GBDT model of the present invention.

[0060] Figure 4 This is a comparison chart of the predicted KY value of QM mortar predicted and analyzed using the GWO-GBDT model in the present invention and the measured value.

[0061] Figure 5 This is a diagram showing the optimization analysis results of the QM mortar admixture ratio design of the present invention.

[0062] Figure 6 This is a comparison chart of the SPF mortar KY predicted value and the measured value predicted and analyzed using the GWO-GBDT model in the present invention.

[0063] Figure 7 It is a comparison chart of the actual data of the SPF mortar of the present invention according to the design requirements and the Parato solution.

[0064] Figure 8 This is a diagram showing the optimization analysis results of the SPF mortar admixture ratio design of the present invention. DETAILED DESCRIPTION

[0065] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0066] like Figure 1 As shown, a method for optimizing the proportion of compatible slurry admixtures for repairing cracks in earthen ruins that is easy to pour and resistant to shrinkage is presented. In order to intuitively compare the differences in the optimization methods of different slurries under different actual working conditions, two different embodiments are given, involving QM (lime) mortar and SPF (potassium silicate) mortar.

[0067] Example 1

[0068] A method for optimizing the admixture ratio of pourable mortar for earthen site restoration is provided, taking QM mortar as an example, and the specific steps are as follows:

[0069] Step S1: collect the data of compressive strength (KY) of pourable mortar and its admixture ratio to form a data set, take the parameters that affect the size of mortar KY as the analysis object of the data set, and quantify and optimize the analysis object to obtain the optimal mortar KY data set, use the Gray Wolf Optimization-Gradient Boosting Tree (GWO-GBDT) model to process the optimal mortar KY data set to determine the mortar KY prediction result, and establish the mortar KY analysis function. A total of 233 groups of QM mortar data are used in this example, and the materials used include QL (quicklime), F (fly ash), S (soil), polyvinyl alcohol solution (PVA) and sticky rice mortar (GR).

[0070] The main reaction mechanism of QM mortar is that QL in the mortar reacts quickly with water to generate Ca(OH)2, which provides an alkaline environment and micro-expansion for the slurry. It subsequently reacts with water in the slurry and CO2 in the air to generate dense CaCO3, which provides the slurry with later strength. The active CaSiO3 and Ca2Al2(SiO3)5 in fly ash F react with Ca(OH)2 in an alkaline environment to generate hydrated CaCO3 and CASH (hydrated calcium aluminosilicate), which provide early strength for the slurry. The inactive part in F and C in S improve the integrity and strength of the slurry through their own particle exchange and agglomeration. PVA and GR can not only provide the slurry with the water required for the mixing process and chemical reaction process, but the content and arrangement of the effective ingredients of PVA and GR themselves can also affect the slurry performance.

[0071] Factors affecting the KY of mortar include the ratio of effective chemical reaction components (RR), the physical value of admixture dosage (DM), the properties of mortar bonding materials (AP) and the curing conditions of mortar stone (CC). Among them, RR refers to the ratio between the amount of each compound substance participating in the chemical reaction during the mortar preparation process and the amount of the compound substance with the most intense chemical reaction in the mortar. Specifically for QM mortar, it mainly includes CaO in QL, SiO2 and Al2O3 in F, and H2O in PVA and GR, and the most intense chemical reaction is CaO. Therefore, H2O-CaO is used in QM mortar to represent the relative amount of CaO and H2O in the process of reacting to generate Ca(OH)2; SiO2-CaO and Al2O3-CaO are used to represent the relative amount of CaO and the other two compounds in the process of calcium-based admixtures reacting with SiO2 and Al2O3 in F respectively; in addition, the indicator m(CaO) is used to reflect the amount of calcium in the system, that is, the mass proportion of CaO.

[0072] The calculation method of DM is to multiply the measured density of each solid material in the mortar by the mass proportion of the material added to the mortar. LThe calculation result of DM corresponding to lime-based materials is expressed as follows. Considering that the hydration reaction of QL will be completed quickly in a short time, the lime admixture will mainly exist in the form of Ca(OH)2 in the mortar. The physical value DM of the admixture dosage considering the hydration reaction is L The calculation formula is as follows:

[0073]

[0074] Where ρ(QL) represents the density of QL in the admixture, M(CaO) and M(Ca(OH)2) represent the amount of CaO and Ca(OH)2, i.e. 44 g / mol and 62 g / mol, respectively, and V(CaO) / V(Ca(OH)2) represents the relative volume change after CaO generates Ca(OH)2, which is about 1 / 2. In the formula, M(CaO), M(Ca(OH)2) and V(CaO) / V(Ca(OH)2) are all fixed values. It is only necessary to measure ρ(QL) and set m(QL) / m(QL+F+C) according to the test requirements to calculate DM. L Similarly, DM F and DM S They represent the calculation results of DM corresponding to F and DM corresponding to C, that is, in the calculation and analysis, only the density values ​​of F and C and their mass proportions in the mortar need to be measured to calculate DM. F and DM S .

[0075] The indicators corresponding to AP include water-cement ratio (WS), mass concentration (QC) and viscosity (VIS). QM In mortar, it is used to express the mass ratio of binders PVA and GR to solid admixtures; VIS QM Indicates the viscosity of adhesives PVA and GR. This index is used to indirectly reflect the difference in viscosity of the effective components in the adhesive due to the different molecular arrangements. The corresponding indicators of CC include stone curing temperature (MT), curing humidity (MH) and curing time (MA). Generally speaking, MA QM The calculation starts from the time when the mortar reaches the final setting time and has the mechanical strength to meet the requirements of demolding. The day of demolding is recorded as the 0th day of the age. The internal reaction of the mortar body is relatively active in the early stage of curing, and gradually stabilizes in the later stage. Therefore, the setting of MA usually requires denseness at the beginning and sparseness at the end. In addition, because the MT and MH conditions used in QM mortar are the same, that is, the curing temperature of the mortar body is set to 25°C and the relative humidity is set to 40% to 50%, there is no difference in this embodiment and it is no longer considered as an analysis object.

[0076] From the above analysis, we can obtain all the analysis objects that affect the compressive strength of the quicklime data set. There are 10 groups of analysis objects in the data set. Table 1 shows the numerical distribution of each indicator in the QM mortar through the mean, maximum, minimum and median values.

[0077] Table 1 Numerical statistics of various indicators in QM mortar

[0078]

[0079] The quantitative optimization of the model analysis object is based on the importance of the impact on the analysis object KY. After obtaining the quantitative results of the importance of each analysis object, the objects with less importance are screened out to obtain the optimized mortar KY data set. The analysis and calculation process of the importance of the model analysis object is as follows:

[0080] In the first step, the data of each analysis object in the data set are combined into an analysis matrix W, whose structure is m×p, where m represents the number of samples in the data set, and p represents the number of analysis objects. In this embodiment, its structure is 233×10. The KY data of the data set are combined into an analysis matrix t, whose structure is m×1=233×1; the calculation method of the estimated coefficient ∈ of the contribution of the analysis object to the KY ratio is: ∈=β′ 2 γ′ 2 ;

[0081] The relevant weight vector β′ and weight correction coefficient γ′ of the importance of each analysis object satisfy:

[0082] W=PΔQ′

[0083] V=PQ′

[0084] β′=(V′V) -1 V′t=QP′t

[0085] γ′=(V′V) -1 V′W=QP′W

[0086] Where P represents the eigenvector matrix corresponding to the matrix WW′, Q represents the eigenvector matrix of the matrix W′W, Q′ is the transpose of the eigenvector matrix Q, Δ represents the diagonal matrix composed of the square root values ​​of the eigenvalues ​​of the matrices WW′ and W′W, V represents the best fitting orthogonal approximation matrix of the analysis matrix W, and β′ is the relevant weight vector of the importance of each analysis object;

[0087] The second step is to calculate the KY prediction value y corresponding to each analysis object in the analysis matrix t through the GWO-GBDT model. P and average Quantitative value W of the importance of different analysis objects j satisfy:

[0088] logit(y P )=ln[y P / (1-y P )]

[0089]

[0090]

[0091]

[0092] Among them, logit(y P ) is the predicted value y P The natural logarithm of is the natural logit(y P ), R 2 is the determination coefficient of the analytical model, y is the measured value, S x is the standard deviation of each analysis object in the analysis model, β M is the analytical value of the importance of different analysis objects in the model, W j For analysis value The quantitative value of the importance of each analysis object j after normalization calculation; n represents the total number of analysis objects. Table 2 gives the quantitative value W of each analysis object in the QM mortar data j Calculation results. It can be found in Table 2 that the indicators related to QL and PVA are often quantified values ​​W j Higher value.

[0093] Table 2 Quantitative analysis results of the importance of the analysis objects in the PM mortar dataset

[0094]

[0095] The process of optimizing the analysis object of the data set also needs to meet the following conditions: first, at least one analysis object that can reflect the performance indicators of various raw materials in the mortar must be retained, that is, at least one indicator of RR and DM in each raw material must be retained as the analysis object reflecting the material performance indicator; second, at least one indicator in AP and CC must be retained as the analysis variable of the model. On the basis of meeting the above conditions, the quantitative value W is screened out. j The last 1 to 5 analysis variables are the remaining components of the data set finally selected by the KY prediction model. S Although m(CaO) which reflects the dosage of multiple admixtures is less important, it still needs to be retained. After optimization, 8 groups of analysis objects remain, namely index numbers X-2, X-10, X-8, X-3, X-7, X-4, X-1 and X-6.

[0096] The parameters that need to be set in GWO are the number of wolf packs and the maximum number of iterations, where the setting range of the number of wolf packs is 10 to 100, and the setting range of the maximum number of iterations is 10 to 1000. The GWO algorithm is used to optimize the hyperparameters in the GBDT model, so that the results of the GBDT model predicting the mortar KY value by analyzing the performance indicators of various raw materials are more accurate. In this example, the number of wolf packs in the GWO algorithm is set to 30, and the maximum number of iterations is set to 300. The hyperparameters of the GBDT model that need to be optimized by GWO include the maximum number of weak learners, step size, and subsampling range in the boosting framework parameters, and the optimization results are 455, 57, and 0.45, respectively. The maximum feature parameter, maximum depth, and minimum number of node samples in the weak learner parameters are optimized to 75, 86, and 32, respectively. The schematic diagram of the mortar KY prediction results obtained by the GWO-GBDT model is shown in the figure. Figure 3 As shown. The GWO algorithm selects the position of the top three gray wolves in the wolf pack as the position of the simulated wolf. By judging the distance between the simulated wolf and the prey, it selects random search or shrinkage distance, calculates and updates the position of each gray wolf, and thus obtains the optimal hyperparameters. The GBDT model is constructed using the optimal hyperparameters. The GBDT model is constructed using the Boosting framework, processes the mortar data set, and accumulates the learning results and the obtained weights to obtain the mortar KY prediction results. The results of the predicted calculation in the QM mortar data set and the deviation from the experimental value are shown in Figure 4 As shown, the R of the training set and the test set 2 They are 0.992 and 0.988 respectively, which shows that the data predicted and analyzed by the GWO-GBDT model have extremely high accuracy and can be used in subsequent ratio optimization calculation and analysis.

[0097] Step S2: According to actual engineering requirements, at least one slurry material data of carbon emission data, cost data, and energy consumption data of materials used in mortar slurry is collected;

[0098] In this embodiment, it is necessary to collect CEC, MC and NH data of each material in the mortar based on the calculated KY prediction value. The CEC, MC and NH data of each material in the mortar need to be calculated separately for the production process (PP) and the transportation process (TP). The CEC data generated by material PP is based on the retrieval of the greenhouse gas emission database or the self-test calculation and analysis. The unit is set to kgCO2 / t. The range value of the CEC coefficient generated by the use of electricity is set to 0.55~0.60kgCO2 / (kW.h). The MC data generated by PP needs to be analyzed according to the different attributes of the material source: industrial materials that can be used directly after purchase need to set MC according to their ex-factory price at the manufacturer. Materials collected on site need to calculate the labor remuneration and equipment use costs to be paid during the on-site collection process to set MC, and materials that need secondary processing before use after purchase need to calculate their production M When C is used, MC should be set based on the ex-factory price of the manufacturer and the additional costs such as electricity and water generated during the processing. The collection of NH data generated by PP also needs to be analyzed according to the different attributes of the material source: the NH data of industrial materials is set according to the implicit energy of the finished product, and its unit is MJ / t. The NH data of materials collected on site needs to consider the energy consumption generated by the electricity used during the use of the machine, and its unit is MJ / (kW.h). Materials that need secondary processing after purchase should consider both the implicit energy of the raw materials when they leave the factory and the energy consumption generated by the electricity used during the secondary processing. The two together constitute its NH data. The CEC, MC and NH data generated in TP are directly related to the transportation mileage of logistics. The range of the CEC indicator setting of TP is 0.1~0.4kgCO2 / (t.km); the MC of TP is set according to the average price of multiple logistics and transportation companies at the test site; the NH data of TP is set to 2~4MJ / (t.km);

[0099] In this embodiment, the CEC coefficient of the material is obtained by collecting data from CEC-related databases such as the China Carbon Accounting Database and the Northwest Greenhouse Gas Emission Inventory Guide. However, some materials such as PVA are estimated by analyzing their preparation process or obtained through experimental measurements. The CEC coefficient generated by the electricity used in the production process is set to 0.589 kgCO2 / (kW .h), the CEC coefficient data of each material transportation process is set to 0.2kgCO2 / (t.km), and its specific value is set according to the distance of transportation. To analyze the MC and NH of each material in the slurry, it is necessary to classify the materials first and then determine the collection method of relevant data of each material. Among them, QL and F belong to industrial materials, and the MC corresponding to their PP can be set according to their ex-factory prices at the manufacturer, while NH can be set according to the embodied energy of the finished product; S belongs to natural collection materials, and the MC corresponding to its PP needs to be set based on the integration of manual work efficiency and work remuneration during multiple field collection processes. Considering that heavy machinery cannot be used for the collection of ruins soil, the energy consumption generated in its production process is set to 0; PVA and GR belong to secondary processing materials, and their use requires secondary processing involving electricity and water. Therefore, the MC corresponding to their PP needs to increase the cost of electricity and water required for secondary processing on the basis of the ex-factory price. At the same time, the NH corresponding to the PP of these materials also needs to consider the embodied energy of the materials when the raw materials leave the factory, and the energy consumption generated by electricity and water during the secondary processing. The MC corresponding to the TP process of the material is set according to the average price of multiple logistics and transportation companies in the test location, and the average freight rate per kilometer follows the principle of decreasing with distance. The energy consumption generated by the transportation process of each material is set to 3.0MJ / (t.km). The CEC, MC and NH of each admixture of QM mortar in the PP process and TP process are shown in Table 3.

[0100] Table 3 CEC coefficient, MC and NH parameters of materials in QM mortar

[0101]

[0102] The “\” in the table indicates that the indicator can be ignored, and the brackets indicate the mass concentration QC of these secondary processing materials when they are purchased after industrial production, and they need to be diluted by secondary processing when used specifically.

[0103] Step S3: establishing at least one of a carbon emission analysis function, a cost analysis function or an energy consumption analysis function of the mortar slurry that meets actual engineering requirements based on the collected slurry data;

[0104] The carbon emission analysis function g CEC The calculation formula is:

[0105]

[0106] Among them, g CEC represents the carbon emission value corresponding to any preferred mortar raw material ratio, l represents the amount of raw materials used in the mortar, Q r Indicates the content of each raw material in the mortar, CEC r Indicates the CEC coefficient of each raw material, LD rIndicates the transport distance of each material, CEC Tr Indicates the unit CEC coefficient during transportation;

[0107] The cost analysis function g MC The calculation formula is:

[0108]

[0109] Among them, g MC It represents the cost value corresponding to any of the mortar raw material ratios involved in the optimization, MC r Indicates the MC corresponding to each material production process, MC Tr Indicates the MC corresponding to each material transportation process;

[0110] The energy consumption analysis function g NH The calculation formula is:

[0111]

[0112] Among them, g NH It represents the energy consumption value corresponding to any ratio of raw materials involved in the optimization of mortar, NH r represents the energy intensity of each raw material, NH Tr Represents the unit energy consumption corresponding to the transportation process of each raw material.

[0113] Step S4: Minimize the comprehensive analysis function of the mortar KY analysis function and at least one of the carbon emission analysis function, the cost analysis function or the energy consumption analysis function to obtain an optimal model for the admixture ratio of the pourable mortar.

[0114] Among them, the analytical function of the compressive strength of the QM mortar data set is determined by the KY predicted value obtained after the GWO-GBDT model prediction. Considering that the curing time MA corresponding to the KY data in the data set varies greatly, and there are more data with MA of 28 days, the KY predicted value of the sample after 28 days of curing is considered to be used as the data required for calculation when establishing the KY part of the analytical function of the admixture ratio optimization model, with a total of 33 sets of data.

[0115] The CEC, MC and NH data of the mortar are obtained through the relevant calculation formulas in step S3.

[0116] The following conditions must be met simultaneously during the optimization of the ratio:

[0117] (I) The mortar KY analysis model should satisfy the maximization of material KY, and the CEC, MC and NH analysis functions should satisfy the numerical minimization of material CEC, MC and NH data;

[0118] (II) The four models and analysis functions should be mutually constrained. It is impossible to have the optimal solution at the same time. In the optimization process, data with at least two indicators reaching the optimal solution at the same time should be discarded to avoid the situation where the Pareto frontier solution cannot be obtained.

[0119] Based on the above conditions, after screening the data involved in the optimal model of admixture ratio in the example, it was found that there were 31 groups of data that met the optimization conditions.

[0120] The process of selecting the optimal grouting repair mortar admixture ratio needs to be achieved through the following steps:

[0121] B1: Calculate and obtain the mortar KY analysis function g in the data set respectively KY And the analysis function g CEC , g MC and g NH ;

[0122] B2: Use the weighted sum method to construct a comprehensive analysis function H = v1·g KY +v2·g CEC +v3·g MC +v4·g NH ;

[0123] Among them, v1, v2, v3 and v4 represent the weight values ​​of mortar compressive strength, carbon emissions, cost and energy consumption respectively and satisfy:

[0124]

[0125] The analysis weight value is determined using the entropy weight method + CRITIC (Criteria Importance Through IntercrieriaCorrelation) method. This method can take into account the correlation between the data KY, CEC, MC and NH as well as the impact of data fluctuations on the weight value.

[0126] After weight calculation, the weight values ​​of the four groups of functions were found to be v1=0.3321, v2=0.2382, v3=0.2265, and v4=0.2032 respectively.

[0127] B3: Use the multi-objective beetle beard algorithm or the multi-objective gray wolf optimization algorithm to calculate the Pareto frontier solution set of mortar ratio, and use the VIKOR method to calculate the benefit ratio C for each solution in the solution set d The smallest result is the optimal mortar ratio, the benefit ratio C d The calculation process is as follows:

[0128] In this example, MOBAS is used to minimize the comprehensive analysis function H to analyze the optimal mortar ratio. iThe step size of the over-initialized beetle is randomly generated from a uniform distribution from 0 to 1. After 1000 iterations, the Pareto frontier solution is obtained. The error BE between the obtained Pareto frontier solutions t Satisfies the following formula:

[0129]

[0130] Among them, BE t represents the Pareto frontier solution point f0(x t ) and g0(x t ) in x t , δ represents the actual Pareto frontier, and K represents the number of Pareto frontier points found by the multi-objective optimization algorithm.

[0131] The error BE of the algorithm in this example is calculated t The value is only 1.56E-04, which shows that the error between the Pareto front solution points generated by the algorithm is small enough and can be used in the example calculation.

[0132] Calculate the profit ratio C of each solution in the Pareto frontier solution set using the VIKOR method d The smallest result is the optimal mortar ratio, the benefit ratio C d The calculation process is as follows:

[0133]

[0134] Where d represents any solution in the Pareto frontier solution set, u represents the number of target values, and Respectively represent the analysis function g KY , g CEC , g MC and g NH The e-th ideal and non-ideal solution of the single-objective optimization of any objective in , F de Denotes the analytical function g corresponding to solution d KY , g CEC , g MC and g NH The value of any target in G d represents the group analysis index, and They represent the group analysis indexes with ideal solution and non-ideal solution as reference, L d represents individual analysis indicators, and They represent the individual regret values ​​with reference to the ideal solution and the non-ideal solution, respectively. d The mortar ratio corresponding to the solution with the smallest value is the optimal ratio.

[0135] By calculating and They are 0.025, 0.017, 0.982 and 0.154 respectively. After substituting 31 sets of data, we can calculate the benefit ratios C of different solutions. d Value, benefit ratio C d The results of the 10 groups with the smallest ratios are shown in Table 4.

[0136] Table 4 C of each mix ratio in QM mortar data calculated by VIKOR method d Value sorting

[0137]

[0138] The parallel coordinate diagram is used to show the comparison between the final optimization result and other proportions. The calculation results of each proportion of the QM mortar data set after MOBAS optimization are as follows: Figure 5 As shown in Table 5. The vertical axis represents each objective function, each curve represents a non-dominated solution in the Pareto solution set, and the red line represents the ratio with the highest relative closeness. Figure 5 The red line in the middle corresponds to the result of the optimal mortar ratio.

[0139] Table 5 The best optimization scheme for QM mortar admixtures

[0140]

[0141] In this example, the proportion of QL in the solid material of the mortar is 0.3, which meets the requirement that the mortar produces a slight expansion to fill the crack interface to improve the compatibility with the ruins. Compared with the highest KY ratio, the CEC of the optimal ratio is reduced by 50.13%, MC is reduced by 58.15%, and NH is reduced by 16.2%. At the same time, the optimal ratio of the slurry increases the use of F by 12% and reduces the use of S by 14%, which not only increases the utilization rate of waste products in building materials, but also reduces the use of precious ruins soil. The results show that this ratio optimization method can effectively optimize the mortar ratio with good performance, low cost and environmental friendliness.

[0142] Example 2

[0143] A method for optimizing the proportion of compatible slurry admixtures for repairing cracks in earthen ruins that is easy to pour and resistant to shrinkage, taking SPF mortar as an example, the specific steps are as follows:

[0144] Step S1: Collect data on the compressive strength (KY) of pourable mortar and its admixture ratio to form a data set, take the parameters that affect the KY of the mortar as the analysis object of the model data set, and quantify and optimize the analysis object, and use the Grey Wolf Optimization-Gradient Boosting Tree (GWO-GBDT) model to determine the mortar KY prediction result.

[0145] There are a total of 197 groups of mortar data used in this example, and the materials used include SPF (potassium silicate), F, S and sodium fluorosilicate (RS).

[0146] The main reaction mechanism of SPF mortar is the gelation between silicate in SPF reagent and various particles in F and S, and their combination forms a relatively stable inorganic cement structure; the molar number of silicate and potassium salt in SPF reagent will be different after being stirred and heated in a water bath (hereinafter referred to as the SPF component number, represented by K2O-SiO2), which makes the effective components contained in SPF reagents with different component numbers that gel with particles in F and S different. In addition, the mass concentration QC and viscosity VIS of SPF reagent can directly or indirectly affect the content and arrangement of its own effective components. After mixing, F and S will produce particle exchange and granulation, which improves the integrity and strength of the slurry; RS can improve the stability of the gel formed by SPF reagent and other materials through cementation.

[0147] In this embodiment, F and S do not react significantly with SPF, so the influence of the ratio RR of the effective chemical reaction components between the admixture and the binder is not considered in the mortar. F and DM S Used to represent DM data of F and S, WS SPF Used to express the mass ratio of the binder SPF to the solid admixture, WS RS Used to express the mass ratio of RS to solid admixture, QC SPF Used to indicate the concentration of SPF, VIS SPF Used to express the viscosity of SPF, MT SPF MH SPF and MA SPF They represent the curing temperature, curing humidity and curing time of the SPF mortar stone body. The relevant numerical statistics of the analysis object and KY data in this example are shown in Table 6.

[0148] Table 6 Numerical statistics of variables in the SPF mortar dataset

[0149]

[0150] The selection of the model analysis object requires a quantitative analysis of its respective importance first, and then based on this result, the variables with less importance are screened out to obtain the optimal analysis object. The calculation steps of the analysis object weight are the same as those in the quicklime mortar example. The difference is that the structure of the data composition analysis object analysis matrix W in the data set is 197×10, and the structure of the KY data analysis matrix t is 197×1. Based on the same analysis matrix Wj The analysis results of the importance of different analysis objects in the SPF mortar example obtained by calculation and analysis methods are shown in Table 7.

[0151] Table 7 Quantitative analysis results of the importance of the analysis objects in the SPF mortar dataset

[0152]

[0153] Based on the rules mentioned in the lime mortar example, there are 8 groups of analysis objects left after optimization, namely X--4, X--6, X--10, X--1, X--5, X--2, X--7 and X--3.

[0154] In this example, the number of wolves in the GWO algorithm is set to 40, the maximum number of iterations is set to 400, and the hyperparameters of the GBDT model that need to be optimized by GWO include the maximum number of weak learners, step size and subsampling range in the boosting framework parameters, and the optimization results are 808, 87 and 0.28 respectively. The maximum feature parameter, maximum depth and minimum number of node samples in the weak learner parameters are 52, 36 and 61 respectively. The results of the SPF mortar data set after model calculation and the deviation from the experimental value are shown in the figure. Figure 5 As shown, the R of the training set and the test set 2 They are 0.985 and 0.922 respectively, which shows that the data predicted and analyzed by the GWO-GBDT model have extremely high accuracy and can be used in subsequent ratio optimization calculation and analysis.

[0155] Step S2: According to actual engineering requirements, at least one slurry material data of carbon emission data, cost data, and energy consumption data of materials used in mortar slurry is collected;

[0156] In this example, it is necessary to collect the CEC and MC indicators of each material in the mortar based on the calculated KY prediction value. The CEC and MC data of each material in the mortar need to be calculated separately for the PP process and TP process. The CEC value of the material is obtained by collecting data from CEC-related databases such as the China Carbon Accounting Database and the Northwest Greenhouse Gas Emission Inventory Guide. The CEC generated by the electricity used in the production process is set to 0.573kgCO2 / (kW . h), the CEC data of each material TP process is set to 0.3kgCO2 / (t .km), and its specific value is set according to the distance of transportation. To analyze the MC of each material in the slurry, it is necessary to classify the materials first and then determine the method of collecting relevant data of each material. Among them, F belongs to industrially manufactured materials, and its PP process MC can be set according to its ex-factory price at the manufacturer; S belongs to naturally collected materials, and its PP process MC needs to be set based on the integrated setting of manual work efficiency and work remuneration during multiple field collection processes. Considering that heavy machinery cannot be used for the collection of site soil, the NH generated during its production process is set to 0; SPF reagents and RS are secondary processing materials, and their use requires secondary processing involving electricity and water. Therefore, the calculation of their PP process MC needs to add the cost of electricity and water required for secondary processing on the basis of the ex-factory price. The CEC and MC of each admixture in the SPF mortar in the PP process and TP process are shown in Table 8.

[0157] Table 8 CEC and MC parameters of SPF mortar dataset materials

[0158]

[0159] The “\” in the table indicates that the index can be ignored, and the brackets indicate the QC of these secondary processed materials when they are purchased after industrial production, and they need to be diluted by secondary processing when they are used specifically.

[0160] Step S3: Based on the collected slurry data, establish at least one of the carbon emission analysis function, cost analysis function or energy consumption analysis function of the mortar slurry that meets the actual engineering needs. CEC The analytical function g of the cost MC The calculation formula is shown in the QL mortar example and will not be repeated here.

[0161] Step S4: Minimize the comprehensive analysis function of the mortar KY analysis function and at least one of the carbon emission analysis function, the cost analysis function or the energy consumption analysis function to obtain an optimal model for the admixture ratio of the pourable mortar.

[0162] The analytical function of KY for the SPF mortar dataset is determined by the KY prediction value obtained by the GWO-GBDT model. Considering that the curing time corresponding to the KY data in the dataset is quite different, and the MA SPF There are more data for 28 days, so the KY predicted value of the sample after 28 days of curing is considered as the data required for calculation when establishing the KY part analysis function of the admixture ratio optimization model, with a total of 87 sets of data.

[0163] The CEC and MC data of the mortar are obtained through the S3 related calculation formula in the QL mortar example. The mortar ratio optimization process needs to meet the following conditions at the same time:

[0164] The following conditions must be met simultaneously during the optimization of the ratio:

[0165] (I) The KY analysis function of mortar should satisfy the maximization of material KY, and the analysis functions of CEC, MC and NH should satisfy the minimization of material CEC data, MC data and NH data;

[0166] (II) The four models and analysis functions should be mutually constrained. It is impossible to have the optimal solution at the same time. In the optimization process, data with at least two indicators reaching the optimal solution at the same time should be discarded.

[0167] Based on the above conditions, after screening the data involved in the optimal model of admixture ratio in the example, it was found that there were 81 groups of data that met the optimization conditions. The process of selecting the optimal admixture ratio of grouting repair mortar needs to be achieved through the following steps:

[0168] B1: Calculate and obtain the KY prediction model g in the data set respectively KY and CEC and g MC ;

[0169] B2: Use the weighted sum method to construct the comprehensive analysis function H as follows:

[0170] H=v1·g KY +v2·g CEC +v3·g MC

[0171] The weight values ​​of each data KY, CEC and MC are v1, v2 and v3 respectively and satisfy The weights of the analysis functions are determined using the entropy weight + CRITIC method. After weight calculation, the weight values ​​of the three groups of functions are obtained as follows: v1 = 0.4152, v2 = 0.3366, and v3 = 0.2482.

[0172] B3: Use the multi-objective beetle whisker algorithm (MOBAS) or multi-objective grey wolf optimization algorithm (MOGWO) to calculate the Pareto frontier solution set of mortar ratio. Use the VIKOR method to calculate the benefit ratio C of each solution in the solution set. d The smallest result is the optimal mix ratio of mortar.

[0173] In this example, the optimal mortar ratio is analyzed by using the multi-objective grey wolf optimization algorithm (MOGWO) to minimize the comprehensive analysis function H, where v i The distance between the wolf pack and the prey is randomly generated by initializing it from a uniform distribution between 0 and 1. After 1000 iterations, the Pareto frontier solution is obtained as follows: Figure 7 As shown, the error BE between the obtained Pareto front solutions t Satisfies the following formula:

[0174]

[0175] Among them, BE t represents the Pareto frontier solution point f0(x t ) and g0(x t ) in x t , δ represents the actual Pareto frontier, and K represents the number of Pareto frontier points found by the multi-objective optimization algorithm.

[0176] The BE of this algorithm in this example is calculated t The value is only 9.32E-04, which indicates that the error between the Pareto frontier points generated by the algorithm is small enough and can be used in the example calculation.

[0177] For each solution in the solution set, use the VIKOR method to calculate its benefit ratio C d The smallest result is the optimal mortar ratio, C d The calculation process is the same as that in the quicklime mortar example.

[0178] By calculating and They are 0.055, 0.023, 0.962 and 0.176 respectively. After substituting 81 sets of data into the calculation, different solutions of C are obtained. d Value, C d The results of the 10 groups with the smallest values ​​are shown in Table 9. The comparison between the actual data of SPF mortar according to the design requirements and the Parato solution is shown in Figure 6 shown.

[0179] Table 9 C of each mix ratio in SPF mortar calculated by VIKOR method d Value sorting

[0180]

[0181] The parallel coordinate diagram is used to show the comparison between the final optimization result and other proportions. The calculation results of each proportion of the SPF mortar dataset after MOGWO optimization are as follows: Figure 8 As shown in Table 10. The vertical axis represents each objective function, each curve represents a non-dominated solution in the Pareto solution set, and the red line represents the ratio with the highest relative closeness. Figure 8 The red line in the middle corresponds to the result of the optimal mortar ratio.

[0182] Table 10 The best optimization scheme for SPF mortar admixtures

[0183]

[0184] In the analysis of SPF mortar, the best mortar is SPF-S. Compared with the ratio with the highest compressive strength value, its CEC is reduced by 42.5% and MC is reduced by 48.7%. At the same time, the use of SPF has decreased by 18.7%. Although the use of F with the highest mass ratio of compressive strength can reduce CEC and slurry costs, controlling the use of SPF is the best solution to optimize the slurry CEC and economic goals. Then the use of SPF is an important factor affecting the CEC and MC in this slurry. In the research of relevant scientific researchers, SPF-S slurry not only meets the physical and mechanical properties requirements for repairing the ruins, but also has good frost resistance, acid and alkali resistance, and water disintegration resistance.

[0185] The above results fully demonstrate that the proposed method for optimizing the proportion of admixtures can not only calculate solutions with better environmental and economic benefits, but also reduce the use of high-emission and high-cost materials in the preferred solution. At the same time, the optimized proportion results meet the actual engineering needs of the restoration of earthen ruins.

[0186] 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 principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for optimizing the proportion of compatible slurry admixtures for repairing cracks in earthen ruins that is suitable for irrigation and resistant to shrinkage, characterized in that: The steps are as follows: Step S1: collect data on the compressive strength of pourable mortar and its admixture ratio to form a data set, take the parameters that affect the compressive strength of the mortar as the analysis object of the data set, and quantify and optimize the analysis object to obtain an optimal mortar compressive strength data set, use the Gray Wolf Optimization-Gradient Boosting Tree model to process the optimal mortar compressive strength data set to determine the mortar KY prediction result, and establish a mortar KY analysis function; Step S2: According to actual engineering requirements, at least one slurry material data of carbon emission data, cost data, and energy consumption data of materials used in mortar slurry is collected; Step S3: establishing at least one of a carbon emission analysis function, a cost analysis function or an energy consumption analysis function of the mortar slurry that meets actual engineering requirements based on the collected slurry data; Step S4: Minimize the comprehensive analysis function of the mortar KY analysis function and at least one of the carbon emission analysis function, the cost analysis function or the energy consumption analysis function to obtain an optimal model for the admixture ratio of the pourable mortar.

2. The method for optimizing the proportion of the compatible slurry admixture for repairing cracks in earthen ruins that is easy to irrigate and resistant to shrinkage according to claim 1 is characterized in that: The factors affecting the compressive strength of mortar include the proportion of effective chemical reaction components, the physical value of the amount of admixture used, the performance of mortar bonding materials and the curing conditions of mortar stone body; the proportion of effective chemical reaction components is the ratio between the amount of substance of each compound participating in the chemical reaction during the mortar preparation process and the amount of substance of the compound with the most violent chemical reaction in the mortar; the physical value of the amount of admixture used is the measured density data of each solid material in the mortar multiplied by the mass proportion of the solid material added to the mortar; the indicators corresponding to the performance of mortar bonding materials include water-cement ratio, mass concentration and viscosity; the indicators corresponding to the curing conditions of mortar stone body include stone body curing temperature, curing humidity and curing time; among which, the water-cement ratio is the mass ratio of the binder to the solid admixture in the mortar.

3. The method for optimizing the proportion of the compatible slurry admixture for repairing cracks in earthen ruins that is easy to irrigate and resistant to shrinkage according to claim 2 is characterized in that: The method for quantifying and optimizing the analysis objects is as follows: based on the importance of the impact on the compressive strength of the analysis objects, the quantified results of the importance of each analysis object are obtained, and then the objects with less importance are screened out to obtain the optimized mortar compressive strength data set.

4. The method for optimizing the proportion of the compatible slurry admixture for repairing cracks in earthen ruins that is easy to irrigate and resistant to shrinkage according to claim 3 is characterized in that: The data of each analysis object in the data set are combined into an analysis matrix W, and the data of the compressive strength of the data set are combined into an analysis matrix t. The relevant weight vectors and weight correction coefficients of the importance of each analysis object are calculated using the analysis matrix W and the analysis matrix t, and the estimated coefficient ∈ of the contribution of each analysis object to the compressive strength ratio is calculated; The analysis matrix t is input into the GWO-GBDT model to obtain the predicted value of the compressive strength corresponding to each analysis object. The analysis value of the importance of the analysis object in the GWO-GBDT model is calculated by estimating the coefficient ∈. The analysis value is normalized to obtain the quantitative value of the importance of each analysis object. The multiple analysis objects with the smallest ranking in the quantitative value are screened out, and the remaining analysis objects form the optimized mortar KY data set. The preferred mortar compressive strength data set must retain at least one indicator, namely, the proportion of effective chemical reaction components of raw materials and the physical value of admixture dosage; the mortar bonding material performance and mortar stone body curing conditions must each retain at least one indicator as an analysis object.

5. The method for optimizing the proportion of admixtures of compatible slurry for repairing cracks in earthen ruins that is easy to irrigate and resistant to shrinkage according to claim 4, characterized in that: The calculation method of the estimated coefficient ∈ is: ∈ = β′ 2 γ′ 2 ; The relevant weight vector β′ and weight correction coefficient γ′ of the importance of each analysis object satisfy: W=PΔQ′ V=PQ′ β′=(V′V) -1 V′t=QP′t y′=(V′V) -1 V′W=QP′W Where P represents the eigenvector matrix corresponding to the matrix WW′, Q represents the eigenvector matrix of the matrix W′W, Q′ is the transpose of the eigenvector matrix Q, △ represents the diagonal matrix composed of the square root values ​​of the eigenvalues ​​of the matrices WW′ and W′W, V represents the best fitting orthogonal approximation matrix of the analysis matrix W, and β′ is the relevant weight vector of the importance of each analysis object; The calculation method of the quantitative value of the importance of each analysis object is: The KY prediction value y corresponding to each analysis object in the analysis matrix t is calculated by the GWO-GBDT model P and average Quantitative value W of the importance of different analysis objects j satisfy: logit(y P )=ln[y P / (1-y P )] Among them, logit(y P ) is the predicted value y P The natural logarithm of is the natural logarithm logit(y P ), R 2 is the determination coefficient of the analytical model, y is the measured value, S x is the standard deviation of each analysis object in the analysis model, β M is the analytical value of the importance of different analysis objects in the model, W j For analysis value The quantitative value of the importance of each analysis object after normalization calculation; n represents the total number of analysis objects.

6. The method for optimizing the proportion of admixtures of compatible slurry for repairing cracks in earthen ruins that is easy to irrigate and resistant to shrinkage according to any one of claims 1 to 5, characterized in that: The GWO algorithm in the gray wolf optimization-gradient boosting tree model selects the values ​​of the positions of the top three gray wolves in the wolf pack as the positions of the simulated wolves, and selects random search or shrinking distance by judging the distance between the simulated wolves and the prey, calculates and updates the positions of each gray wolf, thereby obtaining the optimal hyperparameters; the optimal hyperparameters optimize the hyperparameters in the GBDT model, and the hyperparameters of the GBDT model optimized by GWO include the maximum number of weak learners, step size and subsampling range in the boosting framework parameters; the GBDT model processes the preferred mortar compressive strength data set, and accumulates the learning results and the obtained weights to obtain the mortar compressive strength prediction results.

7. The method for optimizing the proportion of admixtures of compatible slurry for repairing cracks in earthen ruins that is easy to irrigate and resistant to shrinkage according to claim 6, characterized in that: Carbon emission data, cost data, and energy consumption data of each material in the mortar, including the production process and transportation process; The carbon emission analysis function g CEC The calculation formula is: Among them, g CEC represents the carbon emission value corresponding to any preferred mortar raw material ratio, l represents the amount of raw materials used in the mortar, Q r Indicates the content of each raw material in the mortar, CEC r Indicates the CEC coefficient of each raw material, LD r Indicates the transport distance of each material, CEC Tr Indicates the unit CEC coefficient during transportation; The cost analysis function g MC The calculation formula is: Among them, g MC It represents the cost value corresponding to any of the mortar raw material ratios involved in the optimization, MC r Indicates the MC corresponding to each material production process, MC Tr Indicates the MC corresponding to each material transportation process; The energy consumption analysis function g NH The calculation formula is: Among them, g NH It represents the energy consumption value corresponding to any ratio of raw materials involved in the optimization of mortar, NH r represents the energy intensity of each raw material, NH Tr Represents the unit energy consumption corresponding to the transportation process of each raw material.

8. The method for optimizing the proportion of the compatible slurry admixture for repairing cracks in earthen ruins that is easy to irrigate and resistant to shrinkage according to claim 7, characterized in that: The method for obtaining the optimal model of admixture ratio for pourable mortar is: B1: Calculate and obtain the mortar KY analysis function g in the data set respectively KY And the analysis function g CEC , g MC and g NH ; B2: Use the weighted sum method to construct a comprehensive analysis function H = v1·g KY +v2·g CEC +v3·g MC +v4·g NH ; Among them, v1, v2, v3 and v4 represent the weight values ​​of mortar compressive strength, carbon emissions, cost and energy consumption respectively and satisfy: B3: Use the multi-objective beetle beard algorithm or the multi-objective gray wolf optimization algorithm to calculate the Pareto frontier solution set of mortar ratio, and use the VIKOR method to calculate the benefit ratio C for each solution in the solution set d The smallest result is the optimal mix ratio of mortar.

9. The method for optimizing the proportion of admixtures of compatible slurry for repairing cracks in earthen ruins that is easy to irrigate and resistant to shrinkage according to claim 7, characterized in that: The benefit ratio C d The calculation method is: Where d represents any solution in the Pareto frontier solution set, u represents the number of target values, and Respectively represent the analysis function g KY , g CEC , g MC and g NH The e-th ideal and non-ideal solution of the single-objective optimization of any objective in , F de Denotes the analytical function g corresponding to solution d KY , g CEC , g MC and g NH The value of any target in G d represents the group analysis index, and They represent the group analysis indexes with ideal solution and non-ideal solution as reference, L d represents individual analysis indicators, and They represent the individual regret values ​​with reference to the ideal solution and the non-ideal solution, respectively. d The mortar ratio corresponding to the solution with the smallest value is the optimal ratio.

10. The method for optimizing the proportion of admixtures of compatible slurry for repairing cracks in earthen ruins that is easy to irrigate and resistant to shrinkage according to claim 9, characterized in that: The KY analysis function is determined by the KY prediction value obtained after the GWO-GBDT model prediction, and the KY prediction value of the mortar after 28 days of curing is used as the data required for calculation when establishing the KY analysis function of the admixture ratio optimization model; The weight value is determined using the entropy weight method + CRITIC method; The error BE between the obtained Pareto front solutions t for: Among them, BE t represents the Pareto frontier solution point f0(x t ) and g0(x t ) in x t , δ represents the actual Pareto frontier, and K represents the number of Pareto frontier points found by the multi-objective optimization algorithm.

Images