Steel fiber reinforced concrete material design method based on multiple performance requirements

Abstract

The invention discloses a steel fiber reinforced concrete material design method based on multiple performance requirements, which comprises the following steps: acquiring original data of parameters and performance of steel fiber reinforced concrete, and determining a training data set; training the initial convolutional neural network model according to the training data set to obtain a convolutional neural network performance prediction model; using a convolutional neural network performance prediction module to construct a virtual function of the steel fiber reinforced concrete performance as a target function; obtaining constraint conditions of the basic performance parameters corresponding to the target function; based on a genetic algorithm, performing concrete mix proportion optimization on the target function under each constraint condition, screening an optimal scheme meeting a preset performance index, and determining an optimal solution set; an optimal solution set is screened through a TOPSIS method, and an optimal scheme of steel fiber concrete material design is obtained; various properties of the steel fiber reinforced concrete are improved, waste of raw materials is avoided, and sustainable development of building materials is achieved.

Description

Technical Field

[0001] The invention belongs to the technical field of steel fiber concrete design, and in particular relates to a steel fiber concrete material design method based on multiple performance requirements. Background Art

[0002] As one of the basic construction materials, concrete has a wide range of uses, including buildings, roads, bridges and other key infrastructure. With the further improvement of engineering requirements, the focus of concrete design has shifted from the single pursuit of high strength to taking into account work performance, mechanical properties and durability at the same time. In this context, fiber concrete, especially steel fiber concrete, is widely used in the construction of important infrastructure such as roads, bridges and airports due to its excellent crack resistance, tensile strength and toughness.

[0003] The working performance and mechanical properties of steel fiber concrete not only depend on the basic composition of concrete, such as the ratio of cement, water, fine aggregate and coarse aggregate, but also are affected by many factors such as fiber volume content, fiber aspect ratio, fiber type and fiber spatial distribution. Studies have shown that uniform fiber dispersion helps improve the mechanical properties and durability of concrete, especially in preventing the generation and expansion of cracks. Therefore, optimizing the design of steel fiber concrete materials to meet diverse performance requirements has become one of the key issues in engineering applications.

[0004] Traditional fiber concrete mix design methods often focus on empirical formulas and test data, which are difficult to adapt to complex and personalized application requirements. Summary of the invention

[0005] In view of this, an object of the present invention is to provide a steel fiber concrete material design method based on multiple performance requirements to solve the above technical problems.

[0006] In order to achieve the above object, the present invention provides the following technical solutions:

[0007] A steel fiber concrete material design method based on multiple performance requirements, including:

[0008] Obtain the original data of steel fiber concrete parameters and properties and determine the training data set;

[0009] The initial convolutional neural network model is trained according to the training data set to obtain a convolutional neural network performance prediction model;

[0010] The convolutional neural network performance prediction module is used to construct a virtual function of steel fiber concrete performance as the objective function;

[0011] Obtain the constraint conditions of the basic performance parameters corresponding to the objective function; wherein the constraint conditions satisfy the working performance and mechanical performance indicators of the designed steel fiber concrete to achieve the best;

[0012] Based on genetic algorithm, the concrete mix ratio is optimized for the objective function under each constraint condition, the optimal solution that meets the preset performance indicators is selected, and the optimal solution set is determined;

[0013] The TOPSIS method is used to screen the optimal solution set and obtain the optimal design solution for steel fiber concrete materials.

[0014] Furthermore, the original data of steel fiber concrete parameters and performance are obtained to determine the training data set, including:

[0015] The original data of steel fiber concrete parameters and properties are obtained by literature research or experimental data acquisition; the original data includes the mix design parameters and slump, compressive strength and splitting tensile strength of steel fiber concrete;

[0016] The original data is cleaned to obtain a cleaned data set, wherein the cleaned data set includes 5 types of input parameters and 3 types of output parameters;

[0017] The cleaned data set is divided into a training set and a test set according to a ratio of 8:2; wherein the training data set includes a training set and a test set.

[0018] Furthermore, the initial convolutional neural network model is trained according to the training data set to obtain a convolutional neural network performance prediction model, including:

[0019] Constructing an initial convolutional neural network model; wherein the construction of the initial convolutional neural network model includes an input layer, two convolutional layers, two ReLU activation layers for nonlinear activation of the output of the convolutional layers, a Dropout layer that randomly inhibits neurons with a probability of a preset value to reduce overfitting, a fully connected layer for generating prediction values, and a regression layer as an output layer;

[0020] Obtain a training set to train the initial convolutional neural network model; wherein the loss function during model training uses mean square error, and an adaptive time estimation method is used as an optimization algorithm;

[0021] After the training is completed, the test set is obtained to predict the results of the initial convolutional neural network model after the training is completed; the determination coefficient R2, root mean square error RMSE, mean absolute error MAE and mean deviation error MBE are used to evaluate the prediction effect of the initial convolutional neural network model after the training is completed.

[0022] When the prediction results meet the preset conditions, the convolutional neural network performance prediction model is determined.

[0023] Furthermore, the constraints of the basic performance parameters corresponding to the objective function include: constraints of water-cement ratio parameters, constraints of sand ratio parameters, constraints of steel fiber volumetric dosage parameters, constraints of steel fiber aspect ratio parameters and constraints of water reducing agent dosage parameters.

[0024] Furthermore, the TOPSIS method was used to screen the optimal solution set and obtain the optimal design scheme of steel fiber concrete materials, including:

[0025] Get the original decision matrix;

[0026] According to the vector normalization method, the original decision matrix is ​​converted into a normalized decision matrix;

[0027] Based on the normative decision matrix and the corresponding attribute weights, a weighted normative matrix is ​​constructed;

[0028] Obtain the positive ideal solution and negative ideal solution of each attribute in the weighted normative matrix, and calculate the Euclidean distance of each solution in the optimal solution set to the positive ideal solution and negative ideal solution respectively;

[0029] Calculate the comprehensive evaluation index of each solution based on the corresponding Euclidean distance;

[0030] The scheme with the highest comprehensive evaluation index is obtained, and the optimal design scheme of steel fiber concrete material is obtained.

[0031] The beneficial effects of the present invention are:

[0032] 1. This invention combines deep learning methods with steel fiber concrete material design. Compared with traditional methods, it is not only more accurate and efficient, but also has stronger generalization ability;

[0033] 2. The present invention simultaneously considers the slump, compressive strength and splitting tensile strength of steel fiber concrete, which is more in line with engineering requirements compared to the traditional method that only considers compressive strength.

[0034] Other advantages, objectives and features of the present invention will be described in the following description and will be apparent to those skilled in the art to some extent, or they may be taught from the practice of the present invention. The purposes and other advantages of the present invention may be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0035] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0037] Figure 1 A multi-objective optimization design diagram of a steel fiber concrete material design method based on multiple performance requirements in an embodiment of the present invention;

[0038] Figure 2 A convolutional neural network architecture diagram in a steel fiber concrete material design method based on multiple performance requirements in an embodiment of the present invention;

[0039] Figure 3 A comparison diagram of slump prediction results in a steel fiber concrete material design method based on multiple performance requirements in an embodiment of the present invention;

[0040] Figure 4 A comparison chart of prediction results of compressive strength in a steel fiber concrete material design method based on multiple performance requirements in an embodiment of the present invention;

[0041] Figure 5 A comparison diagram of prediction results of splitting tensile strength in a steel fiber concrete material design method based on multiple performance requirements in an embodiment of the present invention;

[0042] Figure 6 This is a comparison chart of evaluation indexes of a steel fiber concrete slump prediction model in a steel fiber concrete material design method based on multiple performance requirements in an embodiment of the present invention;

[0043] Figure 7 This is a residual comparison diagram of a steel fiber concrete slump prediction model in a steel fiber concrete material design method based on multiple performance requirements in an embodiment of the present invention;

[0044] Figure 8 This is a comparison chart of evaluation indexes of a steel fiber concrete compressive strength prediction model in a steel fiber concrete material design method based on multiple performance requirements in an embodiment of the present invention;

[0045] Fig. 9 This is a graph comparing the errors of a steel fiber concrete compressive strength prediction model in a steel fiber concrete material design method based on multiple performance requirements in an embodiment of the present invention;

[0046] Fig.10 This is a comparison chart of evaluation indexes of a prediction model for splitting tensile strength of steel fiber concrete in a steel fiber concrete material design method based on multiple performance requirements in an embodiment of the present invention;

[0047] Fig.11 This is a comparison diagram of the error of the prediction model of the splitting tensile strength of steel fiber concrete in a steel fiber concrete material design method based on multiple performance requirements in an embodiment of the present invention;

[0048] Fig.12A comparison diagram of steel fiber concrete test results and prediction results in a steel fiber concrete material design method based on multiple performance requirements in an embodiment of the present invention;

[0049] Fig.13 It is a schematic diagram of a three-dimensional Pareto front and an optimal solution in a steel fiber concrete material design method based on multiple performance requirements in an embodiment of the present invention. DETAILED DESCRIPTION

[0050] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0051] See also Figure 1 The present invention proposes a steel fiber concrete material design method based on multiple performance requirements, including:

[0052] Obtain the original data of steel fiber concrete parameters and properties and determine the training data set;

[0053] Obtain the original data of steel fiber concrete parameters and properties and determine the training data set, including:

[0054] The original data of steel fiber concrete parameters and properties are obtained by literature research or experimental data acquisition; the original data includes the mix design parameters and slump, compressive strength and splitting tensile strength of steel fiber concrete;

[0055] The original data is cleaned to obtain a cleaned data set, wherein the cleaned data set includes 5 types of input parameters and 3 types of output parameters;

[0056] The cleaned data set is divided into a training set and a test set according to a ratio of 8:2; wherein the training data set includes a training set and a test set;

[0057] The initial convolutional neural network model is trained according to the training data set to obtain a convolutional neural network performance prediction model;

[0058] The initial convolutional neural network model is trained according to the training data set to obtain a convolutional neural network performance prediction model, including:

[0059] Constructing an initial convolutional neural network model; wherein the construction of the initial convolutional neural network model includes an input layer, two convolutional layers, two ReLU activation layers for nonlinear activation of the output of the convolutional layers, a Dropout layer that randomly inhibits neurons with a probability of a preset value to reduce overfitting, a fully connected layer for generating prediction values, and a regression layer as an output layer;

[0060] Obtain a training set to train the initial convolutional neural network model; wherein the loss function during model training uses mean square error, and an adaptive time estimation method is used as an optimization algorithm;

[0061] After the training is completed, the test set is obtained to predict the results of the initial convolutional neural network model after the training is completed; the determination coefficient R2, root mean square error RMSE, mean absolute error MAE and mean deviation error MBE are used to evaluate the prediction effect of the initial convolutional neural network model after the training is completed.

[0062] When the prediction result meets the preset conditions, the convolutional neural network performance prediction model is determined;

[0063] The convolutional neural network performance prediction module is used to construct a virtual function of steel fiber concrete performance as the objective function;

[0064] Obtain the constraint conditions of the basic performance parameters corresponding to the objective function; wherein the constraint conditions satisfy the working performance and mechanical performance indicators of the designed steel fiber concrete to achieve the best;

[0065] The constraints of the basic performance parameters corresponding to the objective function include: the constraints of the water-cement ratio parameter, the constraints of the sand ratio parameter, the constraints of the steel fiber volumetric dosage parameter, the constraints of the steel fiber aspect ratio parameter and the constraints of the water reducing agent dosage parameter;

[0066] Based on genetic algorithm, the concrete mix ratio is optimized for the objective function under each constraint condition, the optimal solution that meets the preset performance indicators is selected, and the optimal solution set is determined;

[0067] Under a given objective function, the genetic algorithm optimizes the concrete mix ratio, finds the best solution that meets the predetermined performance indicators, and forms the optimal solution set, namely the Pareto frontier;

[0068] The TOPSIS method is used to screen the optimal solution set and obtain the optimal design scheme of steel fiber concrete material;

[0069] The TOPSIS method is used to screen the optimal solution set and obtain the optimal design scheme of steel fiber concrete materials, including:

[0070] Get the original decision matrix;

[0071] According to the vector normalization method, the original decision matrix is ​​converted into a normalized decision matrix;

[0072] Based on the normative decision matrix and the corresponding attribute weights, a weighted normative matrix is ​​constructed;

[0073] Obtain the positive ideal solution and negative ideal solution of each attribute in the weighted normative matrix, and calculate the Euclidean distance of each solution in the optimal solution set to the positive ideal solution and negative ideal solution respectively;

[0074] Calculate the comprehensive evaluation index of each solution based on the corresponding Euclidean distance;

[0075] Obtain the solution with the highest comprehensive evaluation index and the optimal design solution for steel fiber concrete materials;

[0076] The working principle of the above technical solution is as follows: To ensure the normal operation of the model and sufficient data storage space, the software and hardware configurations adopted by the present invention are as follows: Windows 11 operating system is equipped with Python 3.7.2 and Matlab R2021b, which can provide powerful programming and scientific computing support; it is equipped with AMD Ryzen 75800H with Radeon Graphics processor and NVIDIA GeForce RTX 3050-6G graphics processor to ensure computing power and graphics processing capabilities; in addition, it is equipped with 16GB of memory and 520GB solid-state hard drive to ensure multi-tasking capabilities and efficient data access;

[0077] Step 1: Dataset acquisition and preprocessing:

[0078] (1) Based on literature research, 294 sets of steel fiber concrete mix design parameters and slump, compressive strength, and splitting tensile strength data were obtained;

[0079] (2) Clean the original data, including processing outliers and null values, and delete the data;

[0080] (3) The data set has 5 input parameters and 3 output parameters, which are divided into training set and test set in a ratio of 8:2;

[0081] Step 2: Construction of convolutional neural network model:

[0082] The convolutional neural network architecture constructed by the present invention is shown in FIG. Figure 2 As shown;

[0083] Step 3: Verification and analysis of convolutional neural network model results:

[0084] (1) Results of each performance prediction model

[0085] The training set and test set results of the slump prediction model based on CNN are as follows: Figure 3 As shown ( Figure 3The first picture is the training set, and the second picture is the test set); it can be seen that the RMSE of the training set is 7.69mm, and the RMSE of the test set is 9.87mm, the errors are less than 10%, and the prediction accuracy of the model is high; the R2 of the training set and the test set are 0.98 and 0.97 respectively, and the degree of fitting is high; in addition, the MAE of the training set and the test set are 5.02mm and 7.23mm respectively, and the MBE are -1.67mm and 0.15mm respectively; it shows that the slump prediction model established by the present invention can accurately map the relationship between various factors and slump, and the results of the training set and the test set are not much different, indicating that the generalization ability is strong;

[0086] Figure 4 ( Figure 4 The first figure is the training set, and the second figure is the test set) shows the training set and test set results of the CNN-based compressive strength prediction model. It can be seen that the RMSE of the training set is 2.14MPa, and the RMSE of the test set is 2.48MPa, with an error of less than 5%, which is within the allowable error range of compressive strength (<15%); the R2 of the training set and the test set are 0.98 and 0.96 respectively, with a high degree of fit; in addition, the MAE of the training set and the test set are 1.21MPa and 1.69MPa respectively, and the MBBE are 0.13MPa and 0.80MPa respectively;

[0087] Figure 5 ( Figure 5 The first figure is the training set, and the second figure is the test set) shows the training set and test set results of the CNN-based splitting tensile strength prediction model. By comparison, the RMSE of the training set and the test set are 0.18MPa and 0.22MPa, respectively, with an error of less than 5%, both within the allowable error range of splitting tensile strength (<15%); the R2 of the training set and the test set are 0.97 and 0.94, respectively, with a high degree of fit; in addition, the MAE of the training set and the test set are 0.11MPa and 0.19MPa, respectively, and the MBE are 0.02MPa and 0.11MPa, respectively; the results show that the model can effectively reflect the relationship between various factors and splitting tensile strength, and exhibits a high generalization ability;

[0088] (2) Comparison with traditional models

[0089] Based on the same data set, the present invention uses traditional machine learning algorithms - BP neural network, support vector machine, random forest and radial basis neural network to predict the slump, compressive strength and tensile strength of steel fiber concrete, and compares the results with the convolutional neural network model established by the present invention;

[0090] Figure 6The comparison results of evaluation indicators of steel fiber concrete slump prediction model are shown; it can be seen that the model performance is ranked as follows: CNN>SVR>RBF>BPNN>RF; the RMSE and MAE values ​​of the CNN model are much smaller than those of other models, and the R2 is the largest, and the model accuracy is the highest; this is due to the weight sharing mechanism. The number of parameters of the CNN model is greatly reduced compared with the fully connected network, which reduces the model complexity and overfitting risk; from the MBE point of view, the performance of each model is positive and negative, but the MBE values ​​of CNN, RF and BPNN are relatively small, indicating that the prediction deviation is small; by comparing the training set and the test set, it can be found that Figure 6 ( Figure 6 The first picture is the training set, and the second picture is the test set). The shapes of the CNN models are similar and the area formed is the smallest. The RMSE, MAE, and R2 values ​​of the CNN models are not much different, and the model has the best generalization ability.

[0091] Combination Figure 7 It can be seen that the residuals (true value - predicted value) of the SVR model and the CNN model fluctuate less and are close to 0, indicating that the model performance is better than other models; however, the SVR model has more outliers (dots in the figure) and the model prediction reliability is poor; therefore, the CNN model is the best choice among the five models;

[0092] Figure 8 ( Figure 8 The first figure is the training set, and the second figure is the test set) shows the comparison results of the evaluation indicators of the steel fiber concrete compressive strength prediction model; it can be seen that the model performance is ranked as follows: CNN>SVR>RBF>RF>BPNN; due to the sufficient compressive strength data, the RF model is no longer the worst performing model at this time, and the BPNN model performs the worst due to its own defects; the RMSE and MAE of the CNN model are smaller than those of other models, and the R2 is the largest, and the model accuracy is the highest; from the perspective of MBE, the error of the CNN model is overestimated and is within the allowable error range of compressive strength; by comparing the training set and the test set, it can be found that Figure 8 The shapes of the CNN models are more similar, and the RMSE, MAE, and R2 values ​​of the CNN models are not much different. The model has the lowest risk of overfitting and the best generalization ability;

[0093] Combination Fig. 9 It can be seen that the box heights of the SVR model and the CNN model are smaller, which means that the residual fluctuation is smaller and the residual fluctuates around zero. The model performance is better than other models; however, the SVR model has more outliers and the model prediction reliability is poor; therefore, the CNN model is the best choice among the five models;

[0094] Fig.10 ( Fig.10The first figure is the training set, and the second figure is the test set) shows the comparison results of the evaluation indicators of the steel fiber concrete splitting tensile strength prediction model; it can be seen that similar to the results of the steel fiber concrete slump prediction model, the model performance is ranked as follows: CNN>SVR>RBF>BPNN>RF, among which the RF model performs the worst because of the small amount of splitting tensile strength data; the RMSE and MAE of the CNN model are smaller than those of other models, and the R2 is the largest, and the model accuracy is the highest; from the perspective of MBE, the error of the CNN model is overestimated and is within the allowable error range of the splitting tensile strength; by comparing the training set and the test set, it can be found that Fig.10 The shapes of the CNN models are more similar, and the RMSE, MAE, and R2 values ​​of the CNN models are not much different. The model has the lowest risk of overfitting and the best generalization ability;

[0095] Combination Fig.11 It can be seen that the box heights of the SVR model and the CNN model are smaller, which means that the residual is passively smaller and the residual fluctuates around zero. The model performance is better than other models; however, the SVR model has more outliers and the model prediction reliability is poor; therefore, the CNN model is the best choice among the five models;

[0096] (3) Test verification results

[0097] The test results of slump, compressive strength and splitting tensile strength of steel fiber concrete were compared with the prediction results of the model based on convolutional neural network, and the errors were calculated, such as Fig.12 As shown ( Fig.12 The first figure corresponds to slump, the second figure corresponds to compressive strength, and the third figure corresponds to splitting tensile strength); it can be seen that the slump error of steel fiber concrete is no more than 13%, the compressive strength error is no more than 11%, and the splitting compressive strength error is no more than 8%, all of which are within the allowable error range, and the prediction accuracy and reliability of the model are verified;

[0098] Step 4: Design method of steel fiber concrete material based on multiple performance requirements:

[0099] (1) The performance prediction model of slump, compressive strength and splitting tensile strength of steel fiber concrete is established through convolutional neural network, and the respective virtual functions are further constructed, which are represented as "virtualFunction_SF", "virtualFunction_FCU" and "virtualFunction_FST" respectively; Matlab software is used to implement the construction of virtual functions. The specific code is shown in Table 1 below. The code composition of each virtual function is similar; these virtual functions can accept the input parameters generated by the optimization algorithm, standardize the input data through the preprocessing step, and reshape it into a data format suitable for the convolutional neural network model; the slump, compressive strength and splitting tensile strength of steel fiber concrete are predicted based on the convolutional neural network, and the corresponding performance prediction values ​​are output; the prediction results are restored to the original data range after denormalization to ensure the practicality and interpretability of the results;

[0100] Among them, the virtual function code is:

[0101] Slump:

[0102]

[0103] Compressive strength:

[0104]

[0105] Splitting tensile strength:

[0106]

[0107]

[0108] (2) Based on the provisions of Steel Fiber Concrete (JG / T 3064-2015) and Code for Design of Ordinary Concrete Mix (JGJ55-2019), as well as the basic requirements for concrete mix and research results of relevant literature; to avoid extremely unreasonable results, ensure that the water-cement ratio, sand ratio, steel fiber volume content, steel fiber aspect ratio and water reducer dosage must be within the reasonable range determined by the specification or experiment; the range of the factors is shown in the following table:

[0109] Table 1- Constraints for multi-objective optimization design of steel fiber reinforced concrete

[0110]

[0111] (3) Taking the minimum and maximum values ​​of slump as 70 mm and 90 mm, the minimum and maximum values ​​of compressive strength as 30 MPa and 40 MPa, and the minimum and maximum values ​​of splitting tensile strength as 3 MPa and 4 MPa, as an example;

[0112] The three-dimensional Pareto frontier obtained by multi-objective optimization based on genetic algorithm and the optimal solution selected by TOPSIS method, such as Fig.13 As shown in the figure; the blue points in the figure represent all potential optimal solution sets, and each point represents a concrete mix ratio scheme, and its performance on three different performance targets; the green points highlight the optimal solutions that meet the various performance requirements, showing that these solutions not only have superior corresponding performance, but are also feasible within the preset slump range; the red points represent the optimal solution with the best comprehensive performance selected from all possible solutions by the TOPSIS method; this optimal solution balances multiple performance targets and is considered to be the solution closest to the ideal solution among all the considered factors according to the evaluation criteria of the TOPSIS method;

[0113] Table 2 shows the optimal solution set selected by the TOPSIS method, including concrete mix solutions No. 1 to 3; all of these solutions successfully meet the performance requirements, and solution No. 3 is identified by the TOPSIS method as the solution with the best comprehensive performance, with the highest compressive strength, and splitting tensile strength and slump meeting the requirements;

[0114] Table 2-Optimal solution set after TOPSIS method selection

[0115]

[0116]

[0117] The beneficial effects of the above technical scheme are as follows: by collecting a large amount of data and constructing a prediction model based on a deep learning method, the working performance and mechanical properties of steel fiber concrete can be accurately predicted, and a steel fiber concrete material design method based on multi-performance requirements is proposed in combination with a multi-objective optimization algorithm; it can not only improve the various properties of steel fiber concrete, but also avoid the waste of raw materials caused by a large amount of trial mixing work, and provide a theoretical basis for achieving sustainable development of building materials; this scheme also constructs a steel fiber concrete performance database based on literature and experimental data statistics, and establishes a prediction model for the slump, compressive strength and splitting tensile strength of steel fiber concrete based on a convolutional neural network. The model has high accuracy (R2 reaches 0.97, 0.96 and 0.94 respectively), and has better generalization ability than traditional models (BP neural network, support vector machine, random forest and radial basis neural network); finally, a material design method for steel fiber concrete based on multi-performance requirements is proposed by combining genetic algorithm and TOPSIS (Technique for Order Preference by Similarity to an Ideal Solution) method, which is used to solve the problem that traditional fiber concrete mix design methods often focus on empirical formulas and experimental data and are difficult to adapt to complex and personalized application needs.

[0118] Finally, it should be noted that the above preferred embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail through the above preferred embodiments, those skilled in the art should understand that various changes can be made in form and details without departing from the scope defined by the claims of the present invention.

Claims

1. A steel fiber concrete material design method based on multiple performance requirements, characterized in that: include: Obtain the original data of steel fiber concrete parameters and properties and determine the training data set; The initial convolutional neural network model is trained according to the training data set to obtain a convolutional neural network performance prediction model; The convolutional neural network performance prediction module is used to construct a virtual function of steel fiber concrete performance as the objective function; Obtain the constraint conditions of the basic performance parameters corresponding to the objective function; wherein the constraint conditions satisfy the working performance and mechanical performance indicators of the designed steel fiber concrete to achieve the best; Based on genetic algorithm, the concrete mix ratio is optimized for the objective function under each constraint condition, the optimal solution that meets the preset performance indicators is selected, and the optimal solution set is determined; The TOPSIS method is used to screen the optimal solution set and obtain the optimal design solution for steel fiber concrete materials.

2. A steel fiber concrete material design method based on multiple performance requirements according to claim 1, characterized in that: Obtain the original data of steel fiber concrete parameters and properties and determine the training data set, including: The original data of steel fiber concrete parameters and properties are obtained by literature research or experimental data acquisition; the original data includes the mix design parameters and slump, compressive strength and splitting tensile strength of steel fiber concrete; The original data is cleaned to obtain a cleaned data set, wherein the cleaned data set includes 5 types of input parameters and 3 types of output parameters; The cleaning data set is divided into a training set and a test set according to a ratio of 8:2; wherein the training data set includes the training set and the test set.

3. A steel fiber concrete material design method based on multiple performance requirements according to claim 1, characterized in that: The initial convolutional neural network model is trained according to the training data set to obtain a convolutional neural network performance prediction model, including: Constructing an initial convolutional neural network model; wherein the construction of the initial convolutional neural network model includes an input layer, two convolutional layers, two ReLU activation layers for nonlinear activation of the output of the convolutional layers, a Dropout layer that randomly inhibits neurons with a probability of a preset value to reduce overfitting, a fully connected layer for generating prediction values, and a regression layer as an output layer; Obtain a training set to train the initial convolutional neural network model; wherein the loss function during model training uses mean square error, and an adaptive time estimation method is used as an optimization algorithm; After the training is completed, the test set is obtained to predict the results of the initial convolutional neural network model after the training is completed; the determination coefficient R2, root mean square error RMSE, mean absolute error MAE and mean deviation error MBE are used to evaluate the prediction effect of the initial convolutional neural network model after the training is completed. When the prediction results meet the preset conditions, the convolutional neural network performance prediction model is determined.

4. The steel fiber concrete material design method based on multiple performance requirements according to claim 1 is characterized in that: The constraints of the basic performance parameters corresponding to the objective function include: constraints on water-cement ratio parameters, constraints on sand ratio parameters, constraints on steel fiber volumetric dosage parameters, constraints on steel fiber aspect ratio parameters and constraints on water reducing agent dosage parameters.

5. The steel fiber concrete material design method based on multiple performance requirements according to claim 1 is characterized in that: The TOPSIS method is used to screen the optimal solution set and obtain the optimal design scheme of steel fiber concrete materials, including: Get the original decision matrix; According to the vector normalization method, the original decision matrix is ​​converted into a normalized decision matrix; Based on the normative decision matrix and the corresponding attribute weights, a weighted normative matrix is ​​constructed; Obtain the positive ideal solution and negative ideal solution of each attribute in the weighted normative matrix, and calculate the Euclidean distance of each solution in the optimal solution set to the positive ideal solution and negative ideal solution respectively; Calculate the comprehensive evaluation index of each solution based on the corresponding Euclidean distance; The scheme with the highest comprehensive evaluation index is obtained, and the optimal design scheme of steel fiber concrete material is obtained.

Images