All-smelting waste slag based concrete design method based on target performance

Abstract

The invention discloses a target performance-based full-smelting waste slag-based concrete design method, which comprises the following steps of: performing induced carbonization on steel slag coarse aggregate to obtain a carbonization result to construct a carbonization result prediction model, and determining a design mortar ratio, a concrete volume expansion rate threshold value and coarse aggregate grading according to the target performance. The method comprises the following steps: determining the maximum mixing amount of steel slag coarse aggregate according to a set CaO content index, preparing full-smelting waste slag-based concrete, determining the concrete strength, constructing a full-smelting waste slag-based concrete strength prediction model, and determining a full-smelting waste slag-based concrete design objective function and design constraint conditions; and according to the target performance, the design target function and the design constraint condition, carrying out optimization search to determine a full-smelting waste residue-based concrete design scheme. According to the method, the mixing amount of the smelting waste slag can be increased, design parameters can be dynamically adjusted according to target performance of different projects, and key technical support can be provided for industrial and large-scale application of the full-smelting waste slag-based concrete.

Description

Technical Field

[0001] This invention relates to the field of concrete design technology, and in particular to a design method for all-metallurgical waste slag-based concrete based on target performance. Background Technology

[0002] Driven by both the global low-carbon development strategy and the green transformation of the construction industry, the high-value utilization of smelting waste has become one of the key paths to solve resource shortages and environmental pressures. Using steel slag as aggregate to prepare all-smelting waste-based concrete can not only dispose of a large amount of industrial solid waste, but also reduce the mining of natural sand and gravel resources, which has significant environmental and economic benefits. However, the free calcium oxide in steel slag aggregate is prone to delayed volume expansion in the hydration environment, which leads to concrete cracking and seriously threatens the long-term durability and volume stability of the structure. Reasonable carbonation treatment, admixture selection and gradation selection are of great strategic significance for promoting the sustainable development of the construction industry.

[0003] However, current traditional technologies for preparing concrete using steel slag face numerous bottlenecks: First, existing processes typically employ a conservative low-dosage strategy, significantly limiting the utilization efficiency of solid waste; second, traditional concrete mix design methods heavily rely on engineers' experience and numerous repetitive trial mixes, resulting in low efficiency, high costs, and difficulty in ensuring quality control; furthermore, the lack of an effective quantitative correlation model between the induced carbonation process of steel slag aggregate and its macroscopic properties in concrete leads to a disconnect between aggregate pretreatment and concrete design, hindering collaborative optimization. Therefore, this invention proposes a design method for all-metallurgical waste slag-based concrete based on target performance. By constructing a carbonation result prediction model and a concrete strength prediction model, it deeply integrates material design with artificial intelligence and optimization algorithms, thereby accurately and efficiently deriving concrete design schemes that meet specific target performance at the optimal cost. This not only provides core technological support for the efficient and high-value utilization of metallurgical waste slag but also promotes the sustainable development of the construction industry, possessing significant practical value and broad market prospects. Summary of the Invention

[0004] The purpose of this invention is to provide a design method for all-metallurgical waste slag-based concrete based on target performance.

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

[0006] This invention includes the following steps:

[0007] Carbonized steel slag coarse aggregate is obtained by combining different induced carbonization indices to induce carbonization of steel slag coarse aggregate. The carbonization results and physical properties of the carbonized steel slag coarse aggregate are measured. A carbonization result prediction model is constructed based on the induced carbonization indices and the carbonization results.

[0008] The target performance of the all-metallurgical waste slag-based concrete is determined. Based on the target performance, the design mortar mix ratio, the concrete volume expansion rate threshold, and the coarse aggregate porosity range are determined. The design mortar is prepared according to the design mortar mix ratio, and the design mortar strength is measured. The coarse aggregate particle size distribution range is determined according to the coarse aggregate porosity range. The coarse aggregate porosity is adjusted by the coarse aggregate particle size distribution.

[0009] The CaO content of the coarse aggregate of the carbonized steel slag was determined, and the maximum admixture of the steel slag coarse aggregate was determined based on the design mortar mix ratio, the concrete volume expansion rate threshold, and the CaO content.

[0010] Carbonized steel slag coarse aggregates with different admixtures and different induced carbonization indices were added to the design mortar to prepare concrete based on smelting waste slag and the concrete strength was measured. A prediction model for the strength of concrete based on smelting waste slag was constructed.

[0011] The design objective function and design constraints of the all-metallurgical waste slag-based concrete are determined, a set of initial induced carbonization indices and initial steel slag coarse aggregate content are determined, and the design scheme of the all-metallurgical waste slag-based concrete is determined by optimization search based on the target performance, the design objective function and the design constraints.

[0012] Furthermore, the method for constructing a carbonization result prediction model includes:

[0013] The range of induced carbonization indices was determined, and a combination of variables for induced carbonization indices was designed using response surface methodology. Steel slag coarse aggregate was induced to carbonize under different variable combinations, and the carbonization results and physical properties of the carbonized steel slag coarse aggregate were measured under different variable combinations. The induced carbonization indices included carbonization temperature, carbonization humidity, carbonization time, carbonization pressure, and CO2 concentration. The carbonization results included aragonite structure coverage and carbonization efficiency. The physical properties included elastic modulus and porosity.

[0014] Polynomial fitting was performed on the carbonization results and physical indices to determine the relationships between aragonite structure coverage and carbonization efficiency, carbonization efficiency and elastic modulus, and carbonization efficiency and porosity.

[0015] The carbonization results and corresponding carbonization indices are combined into a carbonization composite set. Random forest is used to randomly divide the carbonization composite set into a first training set and a first test set in a ratio of 6:4. The carbonization result prediction model is trained using the first training set and the performance of the carbonization result prediction model is tested using the first test set.

[0016] The carbonization result prediction model is specifically a multilayer perceptron, including an input layer, a hidden layer, and an output layer. The input layer processes the input induced carbonization index and outputs induced carbonization features. The hidden layer includes two fully connected hidden layers. The first hidden layer captures the high-order interaction and nonlinear relationship between induced carbonization features and aragonite structure coverage, and the second hidden layer further abstracts and combines induced carbonization features. The output layer is connected to the two fully connected hidden layers to perform regression prediction on aragonite structure coverage. The carbonization efficiency is predicted through the embedded aragonite structure coverage-carbonization efficiency relationship, and finally, the aragonite structure coverage and carbonization efficiency are output.

[0017] The carbonization result prediction model uses mean squared error loss to evaluate the difference between the predicted and actual values, and uses the Adam optimizer to update and optimize the network's weight parameters.

[0018] Furthermore, the method for determining the design mortar mix ratio, the concrete volume expansion rate threshold, and the coarse aggregate porosity range includes:

[0019] The target strength, target workability, and volume stability of the concrete based on smelting waste slag were determined; the target workability of the mortar included fluidity, yield stress, and setting time.

[0020] Based on the slurry volume method and water-cement ratio strength criterion, several mortar mix proportions were initially designed. Mortars that meet the target workability were made into standard specimens and their strength was tested. The mortar with the highest strength was selected as the design mortar strength, and the corresponding mix proportion was the design mortar mix proportion.

[0021] Based on the concrete volume stability, the threshold for concrete volume expansion rate and the minimum carbonation efficiency are determined. The volume expansion rate of all-smelting waste slag-based concrete with steel slag coarse aggregate of different porosities is measured. A porosity-concrete volume expansion rate relationship is established by fitting. The minimum porosity of coarse aggregate is determined based on the concrete volume expansion rate threshold. The maximum porosity of coarse aggregate is determined based on the minimum carbonation efficiency and the carbonation efficiency-porosity relationship, thus obtaining the range of coarse aggregate porosity. The range of coarse aggregate particle size distribution is determined based on the range of coarse aggregate porosity. The coarse aggregate porosity is adjusted by the coarse aggregate particle size distribution.

[0022] Furthermore, the method for determining the maximum content of steel slag coarse aggregate includes:

[0023] The CaO content of the coarse aggregate of the carbide steel slag was determined, and the active CaO content was calculated based on the CaO content; the CaO content includes free CaO content and total CaO content.

[0024] The total amount of active silicon and aluminum is calculated based on the design mortar mix ratio, and the molar ratio of Ca(OH)2 to SiO2 and Al2O3 in the pozzolanic reaction is determined. After converting to the mass ratio, the amount of free CaO that can be consumed is calculated. The expression is as follows:

[0025] ω CaO,active =ω CaO,total -ω CaO,inert

[0026]

[0027] Where ω CaO,active The active CaO content, ω CaO,total For the total CaO content, ω CaO,inert For free CaO content, C consume To determine the amount of consumable free CaO, m CaO , This refers to the mass ratio of CaO, SiO2, and Al2O3 in the volcanic ash reaction. The molar ratio of SiO2 to Al2O3 in the volcanic ash reaction, S active For active SiO2 content, A active Content of active Al2O3;

[0028] The volumetric expansion rate of standard concrete specimens based on smelting waste slag with different steel slag coarse aggregate parameters was determined. The relationship between the net increase in active calcium and the volumetric expansion rate of concrete was fitted. The net increase in active calcium corresponding to the concrete volumetric expansion rate threshold was taken as the equilibrium value of the net increase in active calcium. The maximum adsorption of steel slag coarse aggregate was calculated based on the calcium balance condition, and the expression is as follows:

[0029]

[0030] Where m zlag,max The maximum admixture content of steel slag coarse aggregate per cubic meter, C critical This represents the net increase in the balance value of active calcium.

[0031] Furthermore, the method for constructing the strength prediction model for steel slag coarse aggregate includes:

[0032] Different mortars were prepared and their strength was determined. Carbonized steel slag coarse aggregates with different admixtures and different carbonization induction indices were added to the mortars to prepare standard test blocks of concrete based on smelting waste slag and the strength of the standard test blocks of concrete based on smelting waste slag was determined.

[0033] The induced carbonization index corresponding to each test block is input into the carbonization result prediction model to obtain the corresponding carbonization result prediction value. The carbonization result prediction value is input into the carbonization efficiency-elastic modulus relationship and the carbonization efficiency-porosity relationship to obtain the porosity and elastic modulus of the corresponding carbonized steel slag coarse aggregate.

[0034] The porosity, elastic modulus, mortar strength, steel slag coarse aggregate content, and carbonization results of the carbonized steel slag coarse aggregate were combined to form a strength comprehensive set. Random forest was used to randomly divide the strength comprehensive set into a second training set and a second test set at a ratio of 6:4. The second training set was used to train the strength prediction model of the concrete based on smelting waste slag, and the second test set was used to test the performance of the strength prediction model of the concrete based on smelting waste slag.

[0035] The strength prediction model for concrete based on smelting waste slag includes an input layer, a feature extraction layer, a Bayesian prediction layer, and an output layer. The input layer receives comprehensive strength data preprocessing to obtain comprehensive strength features. The feature extraction layer uses a multimodal attention gating mechanism to assign different attention weights to the comprehensive strength features, performs weighted fusion, and outputs fused strength features. The Bayesian prediction layer uses two fully connected Bayesian hidden layers to perform high-order nonlinear transformations on the fused strength features and maps them to the output layer. The output layer processes the mapping results through linear neurons to predict the strength of concrete based on smelting waste slag.

[0036] The strength prediction model for concrete based on smelting waste slag uses an evidence lower bound loss function to evaluate the difference between the predicted and actual values, employs a variational inference algorithm to optimize the model, and uses the Adam optimizer to update the variational distribution parameters. The expression is as follows:

[0037]

[0038] in Let θ be the lower bound loss function for evidence, and θ be the variational distribution parameter. To perform arithmetic on the variational posterior distribution q(ω|θ) Find the average value. For the current training data The likelihood function under the weight ω of the Bayesian neural network, where KL(·) is the likelihood function and p(ω) is the prior distribution.

[0039] Furthermore, the method for determining the design scheme of all-metallurgical waste slag-based concrete includes:

[0040] Based on the target performance prediction accuracy, carbonation cost, and concrete stability, the design objective function for all-metallurgical waste slag-based concrete is determined, and its expression is:

[0041]

[0042] in For design scheme The corresponding design objective function values ​​are: w1 is the intensity weight, w2 is the carbonization operation cost weight, w3 is the stability weight, and f... cu,target For the target strength of concrete, For design scheme The predicted strength of concrete For design scheme The cost of carbonization operations, For design scheme The stability of concrete is denoted by η, which is the comprehensive cost coefficient, and α1, α2, α3, α4, and α5 are the weights of the carbonation index. For design scheme The carbonization operation corresponds to the temperature, humidity, time, pressure, and CO2 concentration, ξ t The time sensitivity coefficient, β1 represents the CO2 concentration sensitivity coefficient, and β2, β3 represent the stability index weights. For design scheme E corresponds to the carbonization efficiency, porosity, and aragonite structure coverage of coarse aggregate made from carburized steel slag. min For the minimum carbonization efficiency of the carbonization operation, p agg,max The maximum porosity of coarse aggregate in steel slag carbide;

[0043] The range of coarse aggregate porosity, the maximum amount of steel slag coarse aggregate, and the minimum carbonization efficiency are used as design constraints.

[0044] A set of initial induced carbonization indices and initial admixtures for steel slag coarse aggregate were determined. A tangent-flying animal optimization algorithm was then used to perform an optimization search based on the target performance, the design objective function, and the design constraints to determine the design scheme for all-smelting waste slag-based concrete. The specific steps are as follows:

[0045] Based on the initial induced carbonization index, the carbonization result prediction model is called to obtain the predicted aragonite structure coverage and predicted carbonization efficiency. Based on the predicted carbonization efficiency, the carbonization efficiency-elastic modulus relationship and the carbonization efficiency-porosity relationship are called to determine the predicted porosity and predicted elastic modulus of the carbonized steel slag coarse aggregate. Based on the predicted aragonite structure coverage, predicted carbonization efficiency, predicted porosity, predicted elastic modulus and initial admixture, the whole smelting waste slag-based concrete strength prediction model is called to obtain the predicted concrete strength under the corresponding design scheme.

[0046] Calculate the objective function value under the corresponding design scheme. Then check if all constraints are met, and update the population position using the following expression:

[0047]

[0048] in The position of particle i is updated in the (k+1)th iteration, ω(k) is the inertia weight in the kth iteration, K is the maximum number of iterations, and x best c1 and c2 are learning factors with values ​​ranging from (0,1)d, representing the current globally optimal solution. Let θ be the Levy flight vector, tan(θ) be the tangent flight operator, φ be a random angle within [-π / 2, π / 2], and x be the tangent flight vector. r1 x r2 Let r1 and r2 be the positions of random particles, and φ be the step size scaling factor. To comply with A multidimensional random vector that is normally distributed. To comply with A multidimensional random vector with a normal distribution, σ u σ y Let Γ(·) be the standard deviation corresponding to the normal distribution, Γ(·) be the gamma function, and τ be the characteristic index of the Levy distribution;

[0049] The population after position update is subjected to simulated binary crossover and polynomial mutation operations to obtain a new population, generating a new design scheme and calculating the corresponding design objective function. The iteration is repeated until the rate of change of the design objective function after 5 consecutive iterations is less than 0.5% or the preset maximum number of iterations is reached, at which point the iteration stops and the global optimal solution is output. The particle size distribution of steel slag coarse aggregate is adjusted so that the overall porosity of the steel slag coarse aggregate meets the optimal porosity of steel slag coarse aggregate corresponding to the global optimal solution, and the corresponding optimal particle size distribution of steel slag coarse aggregate is output. The global optimal solution and the optimal particle size distribution of steel slag coarse aggregate are used as the design scheme for all-smelting waste slag-based concrete. The global optimal solution includes the optimal induced carbonation index, the optimal steel slag coarse aggregate content, the optimal steel slag coarse aggregate porosity, and the design mortar mix ratio.

[0050] Furthermore, the all-smelting waste slag-based concrete comprises: coarse aggregate of steel slag carbide and supersulfur cement mortar; the coarse aggregate of steel slag carbide comprises three particle sizes: 5-10mm, 10-20mm, and 20-25mm; the supersulfur cement mortar comprises cementitious materials, fine aggregate of steel slag carbide, water, and admixtures; the cementitious materials comprise mineral powder, desulfurized gypsum, cement, and steel slag powder, wherein the mineral powder, desulfurized gypsum, and steel slag powder are 100-300 mesh, and the cement is 400-600 mesh; the fineness modulus of the fine aggregate of steel slag carbide is 1.8-2.8; and the admixtures are one or more combinations of water-reducing agents and expanding agents.

[0051] The beneficial effects of this invention are:

[0052] This invention is a design method for all-metallurgical waste slag-based concrete based on target performance. Compared with the prior art, this invention has the following technical advantages:

[0053] This invention, through preliminary experiments, mortar mix design screening, determination of the maximum content of steel slag coarse aggregate, optimization of search and model construction steps, can improve data preprocessing capabilities and enhance model adaptability in the design of all-metallurgical waste slag-based concrete, thereby improving the efficiency and accuracy of all-metallurgical waste slag-based concrete design. Optimizing the design method of all-metallurgical waste slag-based concrete can improve the scientific nature of the formula, improve the volume stability and mechanical properties of steel slag coarse aggregate, increase the content of metallurgical waste slag, and dynamically adjust design parameters according to the target performance of different projects, providing key technical support for the industrialization and large-scale application of all-metallurgical waste slag-based concrete. Attached Figure Description

[0054] Figure 1 This is a flowchart illustrating the steps of the design method for all-metallurgical waste slag-based concrete based on target performance according to the present invention. Detailed Implementation

[0055] The present invention will be further described below through specific embodiments. The illustrative embodiments and descriptions herein are used to explain the present invention, but are not intended to limit the present invention.

[0056] The present invention provides a method for designing concrete based on smelting waste slag, which includes the following steps:

[0057] like Figure 1 As shown, this embodiment includes the following steps:

[0058] Carbonized steel slag coarse aggregate is obtained by combining different induced carbonization indices to induce carbonization of steel slag coarse aggregate. The carbonization results and physical properties of the carbonized steel slag coarse aggregate are measured. A carbonization result prediction model is constructed based on the induced carbonization indices and the carbonization results.

[0059] The target performance of the all-metallurgical waste slag-based concrete is determined. Based on the target performance, the design mortar mix ratio, the concrete volume expansion rate threshold, and the coarse aggregate porosity range are determined. The design mortar is prepared according to the design mortar mix ratio, and the design mortar strength is measured. The coarse aggregate particle size distribution range is determined according to the coarse aggregate porosity range. The coarse aggregate porosity is adjusted by the coarse aggregate particle size distribution.

[0060] The CaO content of the coarse aggregate of the carbonized steel slag was determined, and the maximum admixture of the steel slag coarse aggregate was determined based on the design mortar mix ratio, the concrete volume expansion rate threshold, and the CaO content.

[0061] Carbonized steel slag coarse aggregates with different admixtures and different induced carbonization indices were added to the design mortar to prepare concrete based on smelting waste slag and the concrete strength was measured. A prediction model for the strength of concrete based on smelting waste slag was constructed.

[0062] The design objective function and design constraints of the all-metallurgical waste slag-based concrete are determined, a set of initial induced carbonization indices and initial steel slag coarse aggregate content are determined, and the design scheme of the all-metallurgical waste slag-based concrete is determined by optimization search based on the target performance, the design objective function and the design constraints.

[0063] In this embodiment, the method for constructing a carbonization result prediction model includes:

[0064] The range of induced carbonization indices was determined, and a combination of variables for induced carbonization indices was designed using response surface methodology. Steel slag coarse aggregate was induced to carbonize under different variable combinations, and the carbonization results and physical properties of the carbonized steel slag coarse aggregate were measured under different variable combinations. The induced carbonization indices included carbonization temperature, carbonization humidity, carbonization time, carbonization pressure, and CO2 concentration. The carbonization results included aragonite structure coverage and carbonization efficiency. The physical properties included elastic modulus and porosity.

[0065] Polynomial fitting was performed on the carbonization results and physical indices to determine the relationships between aragonite structure coverage and carbonization efficiency, carbonization efficiency and elastic modulus, and carbonization efficiency and porosity.

[0066] The carbonization results and corresponding carbonization indices are combined into a carbonization composite set. Random forest is used to randomly divide the carbonization composite set into a first training set and a first test set in a ratio of 6:4. The carbonization result prediction model is trained using the first training set and the performance of the carbonization result prediction model is tested using the first test set.

[0067] The carbonization result prediction model is specifically a multilayer perceptron, including an input layer, a hidden layer, and an output layer. The input layer processes the input induced carbonization index and outputs induced carbonization features. The hidden layer includes two fully connected hidden layers. The first hidden layer captures the high-order interaction and nonlinear relationship between induced carbonization features and aragonite structure coverage, and the second hidden layer further abstracts and combines induced carbonization features. The output layer is connected to the two fully connected hidden layers to perform regression prediction on aragonite structure coverage. The carbonization efficiency is predicted through the embedded aragonite structure coverage-carbonization efficiency relationship, and finally, the aragonite structure coverage and carbonization efficiency are output.

[0068] The carbonization result prediction model uses mean squared error loss to evaluate the difference between the predicted value and the true value, and uses the Adam optimizer to update and optimize the weight parameters of the network.

[0069] In the actual evaluation, the range of induced carbonization indicators was determined as follows: carbonization temperature 20-80℃, carbonization humidity 50%-80%RH, carbonization time 2-24h, carbonization pressure 0.1-0.5MPa, and CO2 concentration 20%-100%. Response surface methodology was used to design combinations of induced carbonization indicator variables for induced carbonization, and the corresponding carbonization results and physical indicators were measured. The specific steps are as follows:

[0070] After removing dust and impurities from the surface of the steel slag coarse aggregate, it was washed three times with deionized water and dried to constant weight in a vacuum drying oven at 60℃. The treated specimen was fixed on a conductive stage and a 5nm thick gold film was sputtered to enhance conductivity. The aggregate surface was observed using a scanning electron microscope, with a focus on identifying aragonite structures (determining aragonite structure coverage). Indentation was performed at test points within a 10-50μm range inward from the aggregate surface in both aragonite-rich and aragonite-free areas of the steel slag coarse aggregate. The results were calculated according to the Oliver-Pharr method. The elastic modulus was calculated using the phenolphthalein titration method. The steel slag coarse aggregate was cut along the center, sprayed with phenolphthalein alcohol solution, and the average depth of the uncolored area (carbonized zone) was measured as the carbonization efficiency. The steel slag coarse aggregate was crushed under low pressure (0-0.2MPa) to measure the large pores with a pore size >50nm, and crushed under high pressure (0.2-200MPa) to measure the medium pores with a pore size 2-50nm. The porosity of steel slag coarse aggregate with different particle sizes was determined. The overall porosity of steel slag coarse aggregate was determined by adjusting the gradation of steel slag coarse aggregate with different particle sizes.

[0071] Binomial fitting was performed on the carbonization results and physical indices to determine the relationship between aragonite structure coverage and carbonization efficiency, and the relationship between carbonization efficiency and elastic modulus.

[0072] In the carbonization result prediction model, the input layer has 5 nodes, the first hidden layer contains 128 neurons with the ReLU activation function, the second hidden layer contains 64 neurons with the ReLU activation function, and the output layer contains 2 neurons.

[0073] In this embodiment, the method for determining the design mortar mix ratio, the concrete volume expansion rate threshold, and the coarse aggregate porosity range includes:

[0074] The target strength, target workability, and volume stability of the concrete based on smelting waste slag were determined; the target workability of the mortar included fluidity, yield stress, and setting time.

[0075] Based on the slurry volume method and water-cement ratio strength criterion, several mortar mix proportions were initially designed. Mortars that meet the target workability were made into standard specimens and their strength was tested. The mortar with the highest strength was selected as the design mortar strength, and the corresponding mix proportion was the design mortar mix proportion.

[0076] Based on the concrete volume stability, the threshold for concrete volume expansion rate and the minimum carbonation efficiency are determined. The volume expansion rate of all-smelting waste slag-based concrete with steel slag coarse aggregate of different porosities is measured. A porosity-concrete volume expansion rate relationship is established by fitting. The minimum porosity of coarse aggregate is determined based on the concrete volume expansion rate threshold. The maximum porosity of coarse aggregate is determined based on the minimum carbonation efficiency and the carbonation efficiency-porosity relationship, thus obtaining the range of coarse aggregate porosity. The range of coarse aggregate particle size distribution is determined based on the range of coarse aggregate porosity. The coarse aggregate porosity is adjusted by the coarse aggregate particle size distribution.

[0077] In actual evaluation, taking the design of a concrete based on smelting waste slag as an example, the target strength of the concrete based on smelting waste slag is determined to be C40, the target workability of mortar (flowability 180-220mm, plastic viscosity > 5.5Pa·s, mortar yield stress > 110Pa, initial setting time > 120min, final setting time < 480min) and the volume stability of the concrete based on smelting waste slag (concrete volume expansion rate < 5%, carbonation efficiency (calculated according to carbonation depth) > 1.5mm);

[0078] Based on the slurry volume method and water-cement ratio strength criterion, five sets of net mortar ratios with water-cement ratios between 0.35 and 0.45 were initially designed. Mortar mixes that meet the target workability of the mortar were prepared and screened. If no mortar mix that meets the target workability of the mortar is found, the water-cement ratio is fixed and the dosage of water-reducing agent and the proportion of admixtures are adjusted until the corresponding mortar meets the target workability of the mortar. Standard specimens are made and tested for compressive strength after curing under standard conditions for 28 days. The highest strength value of the specimen (40.2 MPa) is selected as the design mortar strength, and the corresponding mix ratio is the design mortar mix ratio (water:cement mass ratio of 7:20, of which the mass ratio of mineral powder, desulfurized gypsum, cement and steel slag powder in the cementitious materials is 5:3:1:1).

[0079] Using coarse aggregates of carbonized steel slag with different porosities and designed mortar mix proportions, all-smelting waste slag-based concrete was prepared. The 28-day volume expansion rate was measured, and the relationship between porosity and concrete volume expansion rate (negative correlation) was obtained by fitting. Substituting the expansion rate threshold of 0.05% into the "porosity-concrete volume expansion rate relationship" yielded a minimum aggregate porosity of 8.5%. Substituting the minimum carbonation efficiency (minimum carbonation depth) of 1.5mm into the carbonation efficiency-porosity relationship (negative correlation) yielded a maximum aggregate porosity of 18%. The porosity range of coarse aggregate was taken as 8%-18%.

[0080] In this embodiment, the method for determining the maximum admixture content of steel slag coarse aggregate includes:

[0081] The CaO content of the coarse aggregate of the carbide steel slag was determined, and the active CaO content was calculated based on the CaO content; the CaO content includes free CaO content and total CaO content.

[0082] The total amount of active silicon and aluminum is calculated based on the design mortar mix ratio, and the molar ratio of Ca(OH)2 to SiO2 and Al2O3 in the pozzolanic reaction is determined. After converting to the mass ratio, the amount of free CaO that can be consumed is calculated. The expression is as follows:

[0083] ω CaO,active =ω CaO,total -ω CaO,inert

[0084]

[0085] Where ω CaO,active The active CaO content, ω CaO,total For the total CaO content, ω CaO,inert For free CaO content, C consume To determine the amount of consumable free CaO, m CaO , This refers to the mass ratio of CaO, SiO2, and Al2O3 in the volcanic ash reaction. The molar ratio of SiO2 to Al2O3 in the volcanic ash reaction, S active For active SiO2 content, A active Content of active Al2O3;

[0086] The volumetric expansion rate of standard concrete specimens based on smelting waste slag with different steel slag coarse aggregate parameters was determined. The relationship between the net increase in active calcium and the volumetric expansion rate of concrete was fitted. The net increase in active calcium corresponding to the concrete volumetric expansion rate threshold was taken as the equilibrium value of the net increase in active calcium. The maximum adsorption of steel slag coarse aggregate was calculated based on the calcium balance condition, and the expression is as follows:

[0087]

[0088] Where m zlag,max The maximum admixture content of steel slag coarse aggregate per cubic meter, C critical This represents the net increase in the balance value of active calcium.

[0089] In the actual assessment, the total amount of active silica and aluminum was calculated based on the design mortar mix ratio, and the molar ratio of Ca(OH)2 to SiO2 and Al2O3 in the pozzolanic reaction was determined to be 1:0.6:0.2, which is converted to a mass ratio of CaO:SiO2:Al2O3 of 56:60:51. The net increase in active calcium corresponding to a concrete volume expansion rate of ≤0.05% was taken as the balance value of the net increase in active calcium, and the maximum parameter of steel slag coarse aggregate per cubic meter was calculated to be 1500 kg.

[0090] In this embodiment, the method for constructing a prediction model for the contribution strength of steel slag coarse aggregate includes:

[0091] Different mortars were prepared and their strength was determined. Carbonized steel slag coarse aggregates with different admixtures and different carbonization induction indices were added to the mortars to prepare standard test blocks of concrete based on smelting waste slag and the strength of the standard test blocks of concrete based on smelting waste slag was determined.

[0092] The induced carbonization index corresponding to each test block is input into the carbonization result prediction model to obtain the corresponding carbonization result prediction value. The carbonization result prediction value is input into the carbonization efficiency-elastic modulus relationship and the carbonization efficiency-porosity relationship to obtain the porosity and elastic modulus of the corresponding carbonized steel slag coarse aggregate.

[0093] The porosity, elastic modulus, mortar strength, steel slag coarse aggregate content, and carbonization results of the carbonized steel slag coarse aggregate were combined to form a strength comprehensive set. Random forest was used to randomly divide the strength comprehensive set into a second training set and a second test set at a ratio of 6:4. The second training set was used to train the strength prediction model of the concrete based on smelting waste slag, and the second test set was used to test the performance of the strength prediction model of the concrete based on smelting waste slag.

[0094] The strength prediction model for concrete based on smelting waste slag includes an input layer, a feature extraction layer, a Bayesian prediction layer, and an output layer. The input layer receives comprehensive strength data preprocessing to obtain comprehensive strength features. The feature extraction layer uses a multimodal attention gating mechanism to assign different attention weights to the comprehensive strength features, performs weighted fusion, and outputs fused strength features. The Bayesian prediction layer uses two fully connected Bayesian hidden layers to perform high-order nonlinear transformations on the fused strength features and maps them to the output layer. The output layer processes the mapping results through linear neurons to predict the strength of concrete based on smelting waste slag.

[0095] The strength prediction model for concrete based on smelting waste slag uses an evidence lower bound loss function to evaluate the difference between the predicted and actual values, employs a variational inference algorithm to optimize the model, and uses the Adam optimizer to update the variational distribution parameters. The expression is as follows:

[0096]

[0097] in Let θ be the lower bound loss function for evidence, and θ be the variational distribution parameter. To perform arithmetic on the variational posterior distribution q(ω|θ) Find the average value. For the current training data The likelihood function under the weight ω of the Bayesian neural network, where KL(·) is the likelihood function and p(ω) is the prior distribution;

[0098] In practical evaluation, the intensity aggregation set data is preprocessed, standardized at the feature extraction layer, and an attention weight is calculated for each feature using an attention mechanism (dynamically emphasizing the most critical features for intensity prediction and suppressing interference from irrelevant or noisy features). Each feature is then concatenated with its corresponding attention weight, as shown in the expression:

[0099]

[0100] Where α i h is the attention weight for feature i. i W is the embedding representation of feature i. h b is a learnable weight matrix. h Let v be a learnable bias vector and v be a learnable parameter vector.

[0101] The Bayesian prediction layer uses two hidden layers with 128 and 64 neurons respectively, which conform to a Gaussian distribution. The fused strength features are subjected to high-order nonlinear transformation and mapping. The probability distribution of concrete strength is predicted through the output layer, and the predicted strength of all-metallurgical waste slag-based concrete and the corresponding confidence level are output.

[0102] In this embodiment, the method for determining the design scheme of all-metallurgical waste slag-based concrete includes:

[0103] Based on the target performance prediction accuracy, carbonation cost, and concrete stability, the design objective function for all-metallurgical waste slag-based concrete is determined, and its expression is:

[0104]

[0105] in For design scheme The corresponding design objective function values ​​are: w1 is the intensity weight, w2 is the carbonization operation cost weight, w3 is the stability weight, and f... cu,target For the target strength of concrete, For design scheme The predicted strength of concrete For design scheme The cost of carbonization operations, For design scheme The stability of concrete is denoted by η, which is the comprehensive cost coefficient, and α1, α2, α3, α4, and α5 are the weights of the carbonation index. For design scheme The carbonization operation corresponds to the temperature, humidity, time, pressure, and CO2 concentration, ξ t The time sensitivity coefficient, β1 represents the CO2 concentration sensitivity coefficient, and β2, β3 represent the stability index weights. For design scheme E corresponds to the carbonization efficiency, porosity, and aragonite structure coverage of coarse aggregate made from carburized steel slag. min For the minimum carbonization efficiency of the carbonization operation, p agg,max The maximum porosity of coarse aggregate in steel slag carbide;

[0106] The range of coarse aggregate porosity, the maximum amount of steel slag coarse aggregate, and the minimum carbonization efficiency are used as design constraints.

[0107] A set of initial induced carbonization indices and initial admixtures for steel slag coarse aggregate were determined. A tangent-flying animal optimization algorithm was then used to perform an optimization search based on the target performance, the design objective function, and the design constraints to determine the design scheme for all-smelting waste slag-based concrete. The specific steps are as follows:

[0108] Based on the initial induced carbonization index, the carbonization result prediction model is called to obtain the predicted aragonite structure coverage and predicted carbonization efficiency. Based on the predicted carbonization efficiency, the carbonization efficiency-elastic modulus relationship and the carbonization efficiency-porosity relationship are called to determine the predicted porosity and predicted elastic modulus of the carbonized steel slag coarse aggregate. Based on the predicted aragonite structure coverage, predicted carbonization efficiency, predicted porosity, predicted elastic modulus and initial admixture, the whole smelting waste slag-based concrete strength prediction model is called to obtain the predicted concrete strength under the corresponding design scheme.

[0109] Calculate the objective function value under the corresponding design scheme. Then check if all constraints are met, and update the population position using the following expression:

[0110]

[0111] in The position of particle i is updated in the (k+1)th iteration, ω(k) is the inertia weight in the kth iteration, K is the maximum number of iterations, and x best c1 and c2 are learning factors with values ​​ranging from (0,1)d, representing the current globally optimal solution. Let θ be the Levy flight vector, tan(θ) be the tangent flight operator, φ be a random angle within [-π / 2, π / 2], and x be the tangent flight vector. r1 x r2 Let r1 and r2 be the positions of random particles, and φ be the step size scaling factor. To comply with A multidimensional random vector that is normally distributed. To comply with A multidimensional random vector with a normal distribution, σ u σ y Let Γ(·) be the standard deviation corresponding to the normal distribution, Γ(·) be the gamma function, and τ be the characteristic index of the Levy distribution;

[0112] The population after position update is subjected to simulated binary crossover and polynomial mutation operations to obtain a new population, generating a new design scheme and calculating the corresponding design objective function. The iteration is repeated until the rate of change of the design objective function after 5 consecutive iterations is less than 0.5% or the preset maximum number of iterations is reached, at which point the iteration stops and the global optimal solution is output. The particle size distribution of steel slag coarse aggregate is adjusted so that the overall porosity of the steel slag coarse aggregate meets the optimal porosity of steel slag coarse aggregate corresponding to the global optimal solution, and the corresponding optimal particle size distribution of steel slag coarse aggregate is output. The global optimal solution and the optimal particle size distribution of steel slag coarse aggregate are used as the design scheme for all-smelting waste slag-based concrete. The global optimal solution includes the optimal induced carbonation index, the optimal steel slag coarse aggregate content, the optimal steel slag coarse aggregate porosity, and the design mortar mix ratio.

[0113] In the actual evaluation, a set of initial induced carbonization indicators for steel slag coarse aggregate were determined (carbonization temperature 45℃, carbonization humidity 80%, carbonization time 12h, carbonization pressure 0.3MPa, and CO2 concentration 0.05mol / L) and an initial admixture dosage of 1000kg / m³. 3 Based on the design mortar strength of 40.2 MPa, the carbonation result prediction model, the relationship between carbonation efficiency and elastic modulus, the relationship between carbonation efficiency and porosity, and the strength prediction model of all-smelting waste slag-based concrete were used to obtain the following results: aragonite structure coverage of 66.2%, predicted carbonation efficiency (carbonation depth) of 1.8 mm, predicted porosity of 12%, predicted elastic modulus of 68.7 GPa, and predicted concrete strength of 38.3 MPa.

[0114] Taking the following weights as follows: strength deviation weight 0.5, carbonization operation cost weight 0.25, stability weight 0.25, comprehensive cost coefficient 1.2, carbonization index weights 0.25 / 0.25 / 0.15 / 0.15, stability index weights 0.4 / 0.35 / 0.25, time sensitivity coefficient 0.8, and CO2 concentration sensitivity coefficient 1.2, the initial value of the design objective function is calculated to be 0.3039.

[0115] The maximum number of iterations was set to 200, the learning factor to be 0.6 / 0.4, the Levy distribution characteristic index to be 1.5, the crossover probability / distribution index to be 0.85 / 20, and the mutation probability / distribution index to be 0.15 / 20. The group positions were iteratively updated, and the design objective function was calculated. Until the 82nd-86th iterations, the rate of change of the design objective function value was less than 0.5% for five consecutive iterations. The corresponding local optimum was output, along with the optimal induced carbonization indices (carbonization temperature 38.2℃, carbonization humidity 75%, carbonization time 10h, carbonization pressure 0.25MPa, and CO2 concentration 0.07mol / L; the particle size of the fine steel slag aggregate and steel slag powder is much smaller than that of the coarse steel slag aggregate, achieving optimal carbonization results during the carbonization operation), and the optimal coarse steel slag aggregate dosage of 1200kg / m³. 3The optimal steel slag coarse aggregate porosity of 12% (adjusted by changing the coarse aggregate particle size distribution, corresponding to a particle size distribution / mass ratio of 3:5:2 for three grades of steel slag coarse aggregate: 5-10mm, 10-20mm, and 20-25mm) and the design mortar mix ratio (where the mass content of steel slag coarse aggregate and mortar per cubic meter of concrete is 1200kg and 1140kg respectively; in the mortar, the mass ratio of water:cement material:steel slag fine aggregate is 7:20:33; in the cement material, the mass ratio of mineral powder:desulfurized gypsum:cement:steel slag powder is 5:3:1:1) are used as the design scheme for all-smelting waste slag-based concrete. The corresponding predicted concrete strength is 39.8MPa, carbonation efficiency (carbonation depth) is 2mm, and aragonite structure coverage is 78%.

[0116] In this embodiment, the all-smelting waste slag-based concrete comprises: coarse aggregate of steel slag carbide and supersulfur cement mortar; the coarse aggregate of steel slag carbide comprises three particle sizes: 5-10mm, 10-20mm, and 20-25mm; the supersulfur cement mortar comprises cementitious materials, fine aggregate of steel slag carbide, water, and admixtures; the cementitious materials comprise mineral powder, desulfurized gypsum, cement, and steel slag powder, wherein the mineral powder, desulfurized gypsum, and steel slag powder are 100-300 mesh, and the cement is 400-600 mesh; the fineness modulus of the fine aggregate of steel slag carbide is 1.8-2.8; the admixtures are one or more combinations of water-reducing agents and expanding agents.

[0117] The preparation steps for the all-metallurgical waste slag-based concrete include:

[0118] Induced carbonization: Steel slag is screened into coarse aggregate of the target particle size and rinsed with clean water to remove dust and impurities. The cleaned coarse aggregate is placed in an oven and baked to constant weight (105±5℃, 24 hours). The dried steel slag aggregate is spread in a single layer in the carbonization box. The carbonization box parameters are set according to the optimal induced carbonization index to carbonize the material.

[0119] To prepare supersulfur cement mortar: Add the cementitious materials according to the components and dry mix for 1 minute. Add steel slag fine aggregate and water according to the proportion and mix for 2 minutes to obtain supersulfur cement mortar.

[0120] Pouring and curing: Add steel slag coarse aggregate of different particle sizes according to the optimal steel slag coarse aggregate particle size distribution, and mix the freshly mixed supersulfur cement mortar and steel slag coarse aggregate evenly according to the design ratio and pour it into the mold. Use a vibrating device to vibrate for an appropriate time to remove air bubbles and obtain high density. Then let it stand to form naturally. Remove the mold and carry out standard curing to ensure that the strength and durability of the concrete develop to the best state.

[0121] 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 within the protection scope of the present invention.

Claims

1. A design method for all-metallurgical waste slag-based concrete based on target performance, characterized in that, Includes the following steps: S1. Combine different induced carbonization indices to induce carbonization of steel slag coarse aggregate to obtain carbonized steel slag coarse aggregate, measure the carbonization result and physical index of the carbonized steel slag coarse aggregate, and construct a carbonization result prediction model based on the induced carbonization indices and the carbonization result. S2. Determine the target performance of the all-metallurgical waste slag-based concrete. Based on the target performance, determine the design mortar mix ratio, concrete volume expansion rate threshold, and coarse aggregate porosity range. Prepare the design mortar according to the design mortar mix ratio and determine the design mortar strength. Determine the coarse aggregate particle size distribution range according to the coarse aggregate porosity range. The coarse aggregate porosity is adjusted by the coarse aggregate particle size distribution. S3. Determine the CaO content index of the carbonized steel slag coarse aggregate, and determine the maximum admixture of steel slag coarse aggregate based on the designed mortar mix ratio, the concrete volume expansion rate threshold and the CaO content index. S4. Carbonized steel slag coarse aggregates with different admixtures and different carbonization induction indices were added to the design mortar to prepare concrete based on smelting waste slag and the concrete strength was measured. A prediction model for the strength of concrete based on smelting waste slag was constructed. S5. Determine the design objective function and design constraints for all-metallurgical waste slag-based concrete, determine a set of initial induced carbonization indices and initial steel slag coarse aggregate content, and perform optimization search based on the target performance, the design objective function and the design constraints to determine the design scheme for all-metallurgical waste slag-based concrete.

2. The design method for all-metallurgical waste slag-based concrete based on target performance according to claim 1, characterized in that, The method for constructing a carbonization result prediction model includes: The range of induced carbonization indices was determined, and a combination of variables for induced carbonization indices was designed using response surface methodology. Steel slag coarse aggregate was induced to carbonize under different variable combinations, and the carbonization results and physical properties of the carbonized steel slag coarse aggregate were measured under different variable combinations. The induced carbonization indices included carbonization temperature, carbonization humidity, carbonization time, carbonization pressure, and CO2 concentration. The carbonization results included aragonite structure coverage and carbonization efficiency. The physical properties included elastic modulus and porosity. Polynomial fitting was performed on the carbonization results and physical indices to determine the relationships between aragonite structure coverage and carbonization efficiency, carbonization efficiency and elastic modulus, and carbonization efficiency and porosity. The carbonization results and corresponding carbonization indices are combined into a carbonization composite set. Random forest is used to randomly divide the carbonization composite set into a first training set and a first test set in a ratio of 6:

4. The carbonization result prediction model is trained using the first training set and the performance of the carbonization result prediction model is tested using the first test set. The carbonization result prediction model is specifically a multilayer perceptron, including an input layer, a hidden layer, and an output layer. The input layer processes the input induced carbonization index and outputs induced carbonization features. The hidden layer includes two fully connected hidden layers. The first hidden layer captures the high-order interaction and nonlinear relationship between induced carbonization features and aragonite structure coverage, and the second hidden layer further abstracts and combines induced carbonization features. The output layer is connected to the two fully connected hidden layers to perform regression prediction on aragonite structure coverage. The carbonization efficiency is predicted through the embedded aragonite structure coverage-carbonization efficiency relationship, and finally, the aragonite structure coverage and carbonization efficiency are output. The carbonization result prediction model uses mean squared error loss to evaluate the difference between the predicted and actual values, and uses the Adam optimizer to update and optimize the network's weight parameters.

3. The design method for all-metallurgical waste slag-based concrete based on target performance according to claim 1, characterized in that, The method for determining the design mortar mix ratio, the concrete volume expansion rate threshold, and the coarse aggregate porosity range includes: The target strength, target workability, and volume stability of the concrete based on smelting waste slag were determined; the target workability of the mortar included fluidity, yield stress, and setting time. Based on the slurry volume method and water-cement ratio strength criterion, several mortar mix proportions were initially designed. Mortars that meet the target workability were made into standard specimens and their strength was tested. The mortar with the highest strength was selected as the design mortar strength, and the corresponding mix proportion was the design mortar mix proportion. Based on the concrete volume stability, the threshold for concrete volume expansion rate and the minimum carbonation efficiency are determined. The volume expansion rate of all-smelting waste slag-based concrete with steel slag coarse aggregate of different porosities is measured. A porosity-concrete volume expansion rate relationship is established by fitting. The minimum porosity of coarse aggregate is determined based on the concrete volume expansion rate threshold. The maximum porosity of coarse aggregate is determined based on the minimum carbonation efficiency and the carbonation efficiency-porosity relationship, thus obtaining the range of coarse aggregate porosity. The range of coarse aggregate particle size distribution is determined based on the range of coarse aggregate porosity. The coarse aggregate porosity is adjusted by the coarse aggregate particle size distribution.

4. The design method for all-metallurgical waste slag-based concrete based on target performance according to claim 1, characterized in that, The method for determining the maximum content of steel slag coarse aggregate includes: The CaO content of the coarse aggregate of the carbide steel slag was determined, and the active CaO content was calculated based on the CaO content; the CaO content includes free CaO content and total CaO content. The total amount of active silicon and aluminum is calculated based on the design mortar mix ratio, and the molar ratio of Ca(OH)2 to SiO2 and Al2O3 in the pozzolanic reaction is determined. After converting to the mass ratio, the amount of free CaO that can be consumed is calculated. The expression is as follows: oh CaO,active =ω CaO,total -oh CaO,inert Where ω CaO,active The active CaO content, ω CaO,total For the total CaO content, ω CaO,inert For free CaO content, C consume To determine the amount of consumable free CaO, m CaO , This refers to the mass ratio of CaO, SiO2, and Al2O3 in the volcanic ash reaction. The molar ratio of SiO2 to Al2O3 in the volcanic ash reaction, S active For active SiO2 content, A active This refers to the content of active Al2O3; The volumetric expansion rate of standard concrete specimens based on smelting waste slag with different steel slag coarse aggregate parameters was determined. The relationship between the net increase in active calcium and the volumetric expansion rate of concrete was fitted. The net increase in active calcium corresponding to the concrete volumetric expansion rate threshold was taken as the equilibrium value of the net increase in active calcium. The maximum adsorption of steel slag coarse aggregate was calculated based on the calcium balance condition, and the expression is as follows: Where m zlag,max The maximum admixture content of steel slag coarse aggregate per cubic meter, C critical This represents the net increase in the balance value of active calcium.

5. The design method for all-metallurgical waste slag-based concrete based on target performance according to claim 1, characterized in that, The method for constructing a prediction model for the contribution strength of steel slag coarse aggregate includes: Different mortars were prepared and their strength was determined. Carbonized steel slag coarse aggregates with different admixtures and different carbonization induction indices were added to the mortars to prepare standard test blocks of concrete based on smelting waste slag and the strength of the standard test blocks of concrete based on smelting waste slag was determined. The induced carbonization index corresponding to each test block is input into the carbonization result prediction model to obtain the corresponding carbonization result prediction value. The carbonization result prediction value is input into the carbonization efficiency-elastic modulus relationship and the carbonization efficiency-porosity relationship to obtain the porosity and elastic modulus of the corresponding carbonized steel slag coarse aggregate. The porosity, elastic modulus, mortar strength, steel slag coarse aggregate content, and carbonization results of the carbonized steel slag coarse aggregate were combined to form a strength comprehensive set. Random forest was used to randomly divide the strength comprehensive set into a second training set and a second test set at a ratio of 6:

4. The second training set was used to train the strength prediction model of the concrete based on smelting waste slag, and the second test set was used to test the performance of the strength prediction model of the concrete based on smelting waste slag. The strength prediction model for concrete based on smelting waste slag includes an input layer, a feature extraction layer, a Bayesian prediction layer, and an output layer. The input layer receives comprehensive strength data preprocessing to obtain comprehensive strength features. The feature extraction layer uses a multimodal attention gating mechanism to assign different attention weights to the comprehensive strength features, performs weighted fusion, and outputs fused strength features. The Bayesian prediction layer uses two fully connected Bayesian hidden layers to perform high-order nonlinear transformations on the fused strength features and maps them to the output layer. The output layer processes the mapping results through linear neurons to predict the strength of concrete based on smelting waste slag. The strength prediction model for concrete based on smelting waste slag uses an evidence lower bound loss function to evaluate the difference between the predicted and actual values, employs a variational inference algorithm to optimize the model, and uses the Adam optimizer to update the variational distribution parameters. The expression is as follows: in Let θ be the lower bound loss function for evidence, and θ be the variational distribution parameter. To perform arithmetic on the variational posterior distribution q(ω|θ) Find the average value. For the current training data The likelihood function under the weight ω of the Bayesian neural network, where KL(·) is the likelihood function and p(ω) is the prior distribution.

6. The design method for all-metallurgical waste slag-based concrete based on target performance according to claim 1, characterized in that, The method for determining the design scheme of concrete based on smelting waste slag includes: Based on the target performance prediction accuracy, carbonation cost, and concrete stability, the design objective function for all-metallurgical waste slag-based concrete is determined, and its expression is: in For design scheme The corresponding design objective function values ​​are: w1 is the intensity weight, w2 is the carbonization operation cost weight, w3 is the stability weight, and f... cu,target For the target strength of concrete, For design scheme The predicted strength of concrete For design scheme The cost of carbonization operations, For design scheme The stability of concrete is denoted by η, which is the comprehensive cost coefficient, and α1, α2, α3, α4, and α5 are the weights of the carbonation index. For design scheme The carbonization operation corresponds to the temperature, humidity, time, pressure, and CO2 concentration, ξ t The time sensitivity coefficient, β1 represents the CO2 concentration sensitivity coefficient, and β2, β3 represent the stability index weights. For design scheme E corresponds to the carbonization efficiency, porosity, and aragonite structure coverage of coarse aggregate made from carburized steel slag. min For the minimum carbonization efficiency of the carbonization operation, p agg , max The maximum porosity of coarse aggregate in steel slag carbide; The range of coarse aggregate porosity, the maximum amount of steel slag coarse aggregate, and the minimum carbonization efficiency are used as design constraints. A set of initial induced carbonization indices and initial admixtures for steel slag coarse aggregate were determined. A tangent-flying animal optimization algorithm was then used to perform an optimization search based on the target performance, the design objective function, and the design constraints to determine the design scheme for all-smelting waste slag-based concrete. The specific steps are as follows: Based on the initial induced carbonization index, the carbonization result prediction model is called to obtain the predicted aragonite structure coverage and predicted carbonization efficiency. Based on the predicted carbonization efficiency, the carbonization efficiency-elastic modulus relationship and the carbonization efficiency-porosity relationship are called to determine the predicted porosity and predicted elastic modulus of the carbonized steel slag coarse aggregate. Based on the predicted aragonite structure coverage, predicted carbonization efficiency, predicted porosity, predicted elastic modulus and initial admixture, the whole smelting waste slag-based concrete strength prediction model is called to obtain the predicted concrete strength under the corresponding design scheme. Calculate the objective function value under the corresponding design scheme. Then check if all constraints are met, and update the population position using the following expression: in The position of particle i is updated in the (k+1)th iteration, ω(k) is the inertia weight in the kth iteration, K is the maximum number of iterations, and x best c1 and c2 are learning factors with values ​​ranging from (0,1)d, representing the current globally optimal solution. Let θ be the Levy flight vector, tan(θ) be the tangent flight operator, φ be a random angle within [-π / 2, π / 2], and x be the tangent flight vector. r1 x r2 Let r1 and r2 be the positions of random particles, and φ be the step size scaling factor. To comply with A multidimensional random vector that is normally distributed. To comply with A multidimensional random vector with a normal distribution, σ u σ y Let Γ(·) be the standard deviation corresponding to the normal distribution, Γ(·) be the gamma function, and τ be the characteristic index of the Levy distribution; The population after position update is subjected to simulated binary crossover and polynomial mutation operations to obtain a new population, generating a new design scheme and calculating the corresponding design objective function. The iteration is repeated until the rate of change of the design objective function after 5 consecutive iterations is less than 0.5% or the preset maximum number of iterations is reached, at which point the iteration stops and the global optimal solution is output. The particle size distribution of steel slag coarse aggregate is adjusted so that the overall porosity of the steel slag coarse aggregate meets the optimal porosity of steel slag coarse aggregate corresponding to the global optimal solution, and the corresponding optimal particle size distribution of steel slag coarse aggregate is output. The global optimal solution and the optimal particle size distribution of steel slag coarse aggregate are used as the design scheme for all-smelting waste slag-based concrete. The global optimal solution includes the optimal induced carbonation index, the optimal steel slag coarse aggregate content, the optimal steel slag coarse aggregate porosity, and the design mortar mix ratio.

7. The design method for all-metallurgical waste slag-based concrete based on target performance according to claim 1, characterized in that, The all-smelting waste slag-based concrete comprises: coarse aggregate of steel slag carbide and supersulfur cement mortar; the coarse aggregate of steel slag carbide comprises three particle sizes: 5-10mm, 10-20mm, and 20-25mm; the supersulfur cement mortar comprises cementitious materials, fine aggregate of steel slag carbide, water, and admixtures; the cementitious materials comprise mineral powder, desulfurized gypsum, cement, and steel slag powder, wherein the mineral powder, desulfurized gypsum, and steel slag powder are 100-300 mesh, and the cement is 400-600 mesh; the fineness modulus of the fine aggregate of steel slag carbide is 1.8-2.8; the admixtures are one or more combinations of water-reducing agents and expanding agents.

Images