阅读说明：本技术 一种基于大数据的混凝土原材料及配合比推荐方法 (Concrete raw material and mixing proportion recommendation method based on big data ) 是由 张鸿 黄涛 杨秀礼 张永涛 张国志 屠柳青 程茂林 朱明清 涂同珩 夏昊 徐杰 于 2020-12-14 设计创作，主要内容包括：本发明公开了一种基于大数据的混凝土原材料及配合比推荐方法,包括：S1,建立有关混凝土数据的历史数据库；S2,对S1中有关混凝土各原材料的性能参数数据做标准化处理；S3,利用混合学习器建立混凝土各项性能参数的预测模型；S4,根据施工要求,筛选出满足条件的供应商,并列出各原材料供应商的组合情况；S5,利用预测模型,建立穷举预测数据库；S6,以混凝土各项性能参数指标为边界条件,在穷举预测数据库中筛选出各原材料总成本最低的供应商和混凝土配合比的组合数据,并推荐给混凝土施工单位。本发明将多种回归算法和穷举预测相结合,不仅推荐出符合混凝土性能要求下最经济的配合比,还推荐出各原材料的供应商,推荐结果更全面。(The invention discloses a concrete raw material and mixing proportion recommendation method based on big data, which comprises the following steps: s1, establishing a historical database of the concrete data; s2, standardizing the performance parameter data of each raw material of the concrete in the S1; s3, establishing a prediction model of each performance parameter of the concrete by using the hybrid learner; s4, screening suppliers meeting the conditions according to the construction requirements, and listing the combination condition of each raw material supplier; s5, establishing an exhaustive prediction database by using a prediction model; and S6, screening out combined data of suppliers and concrete mix proportions with the lowest total cost of raw materials from an exhaustive prediction database by taking each performance parameter index of the concrete as a boundary condition, and recommending the combined data to a concrete construction unit. The invention combines various regression algorithms and exhaustive prediction, not only recommends the most economic mix proportion meeting the performance requirements of concrete, but also recommends suppliers of various raw materials, and the recommendation result is more comprehensive.)

1. A concrete raw material and mixing proportion recommendation method based on big data is characterized by comprising the following steps:

s1, establishing a historical database containing cost unit price data of each raw material in the concrete, performance parameter data of each raw material, concrete mixing ratio data and each performance parameter data of the concrete based on the raw material suppliers, raw material production and test data of the concrete accumulated in the finished concrete construction project;

s2, standardizing the performance parameter data of each raw material in the historical database to obtain new performance parameter data of each raw material;

s3, integrating a plurality of learners to obtain a hybrid learner, wherein the hybrid learner comprises a Gaussian process regression, a random forest and a gradient lifting decision tree;

s4, taking concrete mixing ratio data in a historical database and performance parameter data of each raw material after standardization as input values, taking each performance parameter data of the concrete as output values, and establishing a prediction model of each performance parameter of the concrete by using a hybrid learner;

s5, preliminarily screening suppliers of raw materials meeting the conditions according to the actual conditions and the regional characteristics of project construction, and listing all combination conditions of the suppliers of the raw materials of the concrete;

s6, exhaustively predicting the performance parameter data of the concrete under different supplier combinations and mix proportions by using a prediction model of each performance parameter of the concrete and the performance parameter data of each raw material and the mix proportion data after screening as variables and by using an exhaustion method, and establishing an exhaustion prediction database;

and S7, inputting cost unit price data of each raw material into an exhaustive prediction database by taking each performance parameter index of the concrete as a boundary condition, screening combined data of suppliers of each raw material of the concrete and the mix proportion of the concrete, wherein each performance parameter of the concrete reaches the standard and the total cost of each raw material is the lowest, and recommending the combined data to a concrete construction unit.

2. The method for recommending concrete raw materials and mix ratios based on big data as claimed in claim 1, wherein the way of standardizing the performance parameter data of each raw material in the historical database in S2 is to subtract the mean value of all the data of the performance parameter from the data of each performance parameter of each raw material and divide the data by the variance of all the data of the performance parameter to obtain new performance parameter data of each raw material, and the processed performance parameter data of each raw material is used as an input value and input into the hybrid learner.

3. The big data based concrete raw material and mixing proportion recommendation method according to claim 2, it is characterized in that in the process of establishing a prediction model of each performance parameter of concrete, the performance parameter data of each raw material, the concrete mixing ratio data and each performance parameter data of the concrete after standardized processing are randomly divided into a 75 percent training set and a 25 percent testing set, and the performance parameter data of the raw materials and the concrete mixing ratio data after the standardization treatment are used as input values, the performance parameter data of each item of concrete are used as output values, training regression models of various performance parameters of the concrete by using a learner matched with the various performance parameters of the concrete on a training set, testing regression accuracy on a testing set, and gradually adjusting the hyper-parameters of the training model according to the regression precision, and finally training to obtain an accurate prediction model of each performance parameter of the concrete.

4. The big data based concrete raw material and mix proportion recommendation method according to claim 3, wherein the regression accuracy of the regression model is tested by taking the mean square error as a test index.

5. The big data based concrete raw material and mix proportion recommendation method according to claim 1, wherein in S6, the maximum value and the minimum value of the screened concrete mix proportion data are multiplied by 110% and 90%, and are used as the upper limit and the lower limit of the exhaustive prediction concrete mix proportion, and an exhaustive prediction database is established.

6. The big data based concrete raw material and mix proportion recommendation method of claim 5, wherein the concrete mix proportion data between the upper limit and the lower limit is divided into ten equal parts as sampling intervals in the exhaustive prediction process.

7. The big data based concrete raw material and mix proportion recommendation method of claim 1, wherein the raw materials of the concrete comprise cement, fly ash, mineral powder, admixture and coarse and fine aggregates.

8. The big data based concrete raw material and mix proportion recommendation method of claim 1, wherein the concrete performance parameters comprise slump, concrete strength, chloride ion diffusion coefficient and corrosion resistance coefficient, and electric flux.

9. The big data-based concrete raw material and mix proportion recommendation method according to claim 8, wherein the prediction models of various performance parameters of the concrete comprise a slump prediction model, a concrete strength prediction model, a chloride ion diffusion coefficient prediction model, a corrosion resistance coefficient prediction model and an electric flux prediction model.

Technical Field

The invention relates to the field of concrete construction. More specifically, the invention relates to a concrete raw material and mixing proportion recommendation method based on big data.

Background

The concrete is used as the most basic material in the main structure of civil engineering, the quality and the dosage of the concrete are important for the quality control and the cost control of engineering, wherein, the selection of raw materials and the design of the mixing ratio are the key factors influencing the quality and the cost of finished concrete, the raw materials and the design of the mixing ratio are mutually coupled, the design of the reasonable and economic concrete mixing ratio is established on the basis of the raw materials, and the concrete raw materials are the basis for realizing the product quality and embodying the economic benefit. In the raw material supply link, due to the lack of information resources, a large amount of time is needed to search for raw material resources after a project enters a field, the most suitable supplier cannot be really selected, part of raw materials (such as additives) are generally used at low price, the adaptability is poor, the quality is unstable, in the mix proportion design link, due to the fact that the quality of the selected raw materials is unstable and the experience of workers is insufficient, in order to ensure the strength and the durability of concrete, the use amount of cement and other glue materials is generally required to be increased, the mix proportion is not economical, and the cost of the concrete is increased. Therefore, the selection of proper raw material manufacturers is of great importance to the design link of the mix proportion, and the reasonable use amount of the mix proportion directly influences the performance and the cost of the finished concrete.

At present, in the related field, the design of the mix proportion of concrete raw materials and the prediction of concrete performance are realized by using historical data of a concrete production test through an intelligent optimization and machine learning method, but the research on the prediction of the concrete performance does not consider the design of the mix proportion and the optimization selection of a material manufacturer, the research on the design of the mix proportion of raw materials usually needs to take the performance of raw materials selected in advance as a premise, and in actual construction, the production cost of the concrete is not only related to the mix proportion, but also closely related to the supply price of the raw materials.

Disclosure of Invention

The invention aims to provide a concrete raw material and mixing ratio recommendation method based on big data, which combines multiple regression algorithms and exhaustive prediction, can recommend the most economic mixing ratio meeting the concrete performance requirements and the suppliers of all raw materials, has more comprehensive recommendation results, and can replace fussy test work, improve the working efficiency and save resources by establishing a prediction model of each performance parameter of concrete by using a hybrid learner.

To achieve these objects and other advantages in accordance with the present invention, there is provided a big data based concrete raw material and mix proportion recommendation method, comprising:

s1, establishing a historical database containing cost unit price data of each raw material in the concrete, performance parameter data of each raw material, concrete mixing ratio data and each performance parameter data of the concrete based on the raw material suppliers, raw material production and test data of the concrete accumulated in the finished concrete construction project;

s2, standardizing the performance parameter data of each raw material in the historical database to obtain new performance parameter data of each raw material;

s3, integrating a plurality of learners to obtain a hybrid learner, wherein the hybrid learner comprises a Gaussian process regression, a random forest and a gradient lifting decision tree;

s4, taking concrete mixing ratio data in a historical database and performance parameter data of each raw material after standardization as input values, taking each performance parameter data of the concrete as output values, and establishing a prediction model of each performance parameter of the concrete by using a hybrid learner;

s5, preliminarily screening suppliers of raw materials meeting the conditions according to the actual conditions and the regional characteristics of project construction, and listing all combination conditions of the suppliers of the raw materials of the concrete;

s6, exhaustively predicting the performance parameter data of the concrete under different supplier combinations and mix proportions by using a prediction model of each performance parameter of the concrete and the performance parameter data of each raw material and the mix proportion data after screening as variables and by using an exhaustion method, and establishing an exhaustion prediction database;

and S7, inputting cost unit price data of each raw material into an exhaustive prediction database by taking each performance parameter index of the concrete as a boundary condition, screening combined data of suppliers of each raw material of the concrete and the mix proportion of the concrete, wherein each performance parameter of the concrete reaches the standard and the total cost of each raw material is the lowest, and recommending the combined data to a concrete construction unit.

Preferably, in step S2, the performance parameter data of each raw material in the history database is normalized by subtracting the mean value of all the data of the performance parameter from the data of each performance parameter of each raw material and dividing the result by the variance of all the data of the performance parameter to obtain new performance parameter data of each raw material, and the processed performance parameter data of each raw material is input to the hybrid learner as an input value.

Preferably, in the process of establishing the prediction model of each performance parameter of the concrete, the performance parameter data, the concrete mixing ratio data and each performance parameter data of the concrete after standardized processing are randomly divided into 75% of a training set and 25% of a testing set, the performance parameter data and the concrete mixing ratio data of the raw materials after standardized processing are used as input values, each performance parameter data of the concrete is used as an output value, a learner matched with each performance parameter of the concrete is used on the training set to train the regression model of each performance parameter of the concrete, the regression precision is tested on the testing set, the hyperparameter of the training model is gradually adjusted according to the regression precision, and finally the prediction model of each performance parameter of the concrete is obtained through training.

Preferably, the regression accuracy of the regression model is measured by using a mean square error as a measurement index.

Preferably, in S6, the maximum value and the minimum value of the screened concrete mix proportion data are multiplied by 110% and 90% respectively, and are used as the upper limit and the lower limit of the exhaustive prediction concrete mix proportion, and the exhaustive prediction database is established.

Preferably, the concrete mix ratio data between the upper limit and the lower limit is divided equally into ten equal parts as the sampling interval in the exhaustive prediction process.

Preferably, the raw materials of the concrete include cement, fly ash, mineral powder, admixture, coarse aggregate and fine aggregate.

Preferably, the performance parameters of the concrete include slump, concrete strength, chloride ion diffusivity and corrosion resistance coefficient, and electric flux.

Preferably, the prediction models of various performance parameters of the concrete comprise a slump prediction model, a concrete strength prediction model, a chloride ion diffusion coefficient prediction model, a corrosion resistance coefficient prediction model and an electric flux prediction model.

The invention at least comprises the following beneficial effects:

1. the invention realizes that the performance parameters of the concrete raw materials and the mix proportion of the concrete are used as input values, the data of each performance parameter of the concrete are used as output values, and the prediction model of each performance parameter of the concrete is established.

2. The invention combines various regression algorithms and exhaustive prediction, not only recommends the most economic mix proportion meeting the performance requirements of concrete, but also recommends suppliers of various raw materials, and the recommendation result is more comprehensive.

Additional advantages, objects, and features of the invention will be set forth in part in the description which follows and in part will become apparent to those having ordinary skill in the art upon examination of the following or may be learned from practice of the invention.

Drawings

FIG. 1 is a schematic flow chart of the present invention for establishing a prediction model of various performance parameters of concrete;

FIG. 2 is a schematic flow chart of exhaustive prediction of suppliers and recommended mix proportion of raw materials for concrete according to the present invention.

Detailed Description

The present invention is further described in detail below with reference to the attached drawings so that those skilled in the art can implement the invention by referring to the description text.

It should be noted that in the description of the present invention, the terms "lateral", "longitudinal", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. indicate orientations or positional relationships based on those shown in the drawings, and are only for convenience of description and simplicity of description, and do not indicate or imply that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and thus, should not be construed as limiting the present invention.

As shown in fig. 1-2, the invention provides a concrete raw material and mix proportion recommendation method based on big data, which comprises the following steps:

s1, establishing a historical database containing cost unit price data of each raw material in the concrete, performance parameter data of each raw material, concrete mixing ratio data and each performance parameter data of the concrete based on the raw material suppliers, raw material production and test data of the concrete accumulated in the finished concrete construction project;

s2, standardizing the performance parameter data of each raw material in the historical database to obtain new performance parameter data of each raw material;

s3, integrating a plurality of learners to obtain a hybrid learner, wherein the hybrid learner comprises a Gaussian process regression, a random forest and a gradient lifting decision tree;

s4, taking concrete mixing ratio data in a historical database and performance parameter data of each raw material after standardization as input values, taking each performance parameter data of the concrete as output values, and establishing a prediction model of each performance parameter of the concrete by using a hybrid learner;

s5, preliminarily screening suppliers of raw materials meeting the conditions according to the actual conditions and the regional characteristics of project construction, and listing all combination conditions of the suppliers of the raw materials of the concrete;

s6, exhaustively predicting the performance parameter data of the concrete under different supplier combinations and mix proportions by using a prediction model of each performance parameter of the concrete and the performance parameter data of each raw material and the mix proportion data after screening as variables and by using an exhaustion method, and establishing an exhaustion prediction database;

and S7, inputting cost unit price data of each raw material into an exhaustive prediction database by taking each performance parameter index of the concrete as a boundary condition, screening combined data of suppliers of each raw material of the concrete and the mix proportion of the concrete, wherein each performance parameter of the concrete reaches the standard and the total cost of each raw material is the lowest, and recommending the combined data to a concrete construction unit.

In the technical scheme, the raw materials in the concrete are various, the suppliers of each raw material are also various, so that one raw material performance parameter in a historical database has various data, the mix proportion of the concrete is the quality of each raw material contained in each cubic meter of concrete, and each corresponding performance parameter of the concrete also has various data, the invention establishes the historical database according to the information about the concrete accumulated in the finished project, carries out standardized processing on the performance parameter data of each raw material in the historical database to obtain new performance parameter data of each raw material, and inputs the new performance parameter data of each raw material and the concrete mix proportion data into a hybrid learner as one part of the input value of the hybrid learner and the concrete mix proportion data, establishes a prediction model by taking each performance parameter data of the concrete as the output value, and the hybrid learner comprises Gauss process regression, random forest and gradient promotion decision tree, the learners and the learners are independent from each other, a plurality of learners are used for predicting various performance parameter models of the concrete, the prediction models correspond to different suppliers, the performance parameters of various raw materials of the provided concrete are different, so that the performance parameters of various raw materials of the concrete in the prediction models have a plurality of groups of data output of various corresponding performance parameters of the concrete, in a concrete construction project, certain requirements on the performance of various raw materials of the concrete are provided, firstly, according to the requirements of the project construction on various raw materials of the concrete and the region where the project construction is located, suppliers, which do not meet the conditions and are located far away from the project construction region, of various raw materials of the concrete are preliminarily screened, the raw materials of the concrete provided by the suppliers which meet the conditions are reserved, and the suppliers of various raw materials of the concrete which meet the conditions are combined, the performance parameter data and the concrete mixing proportion data of the screened raw materials are used as variables, the performance parameter data of the concrete of different suppliers are exhaustively predicted, the cost unit price data of the raw materials are input into the exhaustively predicted data, the data of the supplier and the mixing proportion with the performance parameters of the concrete up to the standard and the lowest total cost are obtained through secondary screening, and the data are recommended to a concrete construction unit, so that the concrete construction unit can know how to directly select the raw materials of the concrete with lower cost and up to the standard quality, meanwhile, the concrete test work can be reduced, the work efficiency is improved, and resources are saved.

In a coagulationIn the project of the concrete construction, in order to obtain the combined data of suppliers of various raw materials and the mix proportion of the concrete, wherein various performance parameters of the concrete reach the standard and the total cost of the raw materials of the concrete is the lowest, and purchase the raw materials of the concrete according to the combined data, firstly, 5 suppliers of cement, fly ash, mineral powder and additives meeting the conditions are preliminarily screened according to the actual situation and the regional characteristics of the project construction, 1 supplier of coarse aggregate and fine aggregate respectively exists, namely the combined situation of the suppliers of the raw materials of the concrete has 5, 1, 625 groups, the specific weight of each raw material has 10, namely the mix proportion corresponding to 6 raw materials has 10 groups⁶In that case, a group of raw material suppliers is combined by 10⁶Group mix ratio, 625 group supplier mix results in 625 x 10⁶The combination ratio of the groups and the data of various performance parameters of the concrete are used as conditions, and 10 conforming to the conditions are secondarily screened out according to the requirements of the project on various performances of the concrete⁶The group combination ratio and the data of various performance parameters of the concrete are according to 10⁶Calculating the unit price and the mix proportion of the raw materials provided by each raw material supplier corresponding to the group data 10⁶And corresponding cost in the group data is recommended to the project responsible person by adopting the combination condition with the lowest cost.

In another embodiment, the performance parameter data of each raw material in the historical database is normalized in S2 by subtracting the mean value of all the data of the performance parameter from the data of each performance parameter of each raw material and dividing the result by the variance of all the data of the performance parameter to obtain new performance parameter data of each raw material, and inputting the processed performance parameter data of each raw material as an input value into the hybrid learner.

In the technical scheme, the performance parameter data of the raw materials are subjected to standardized processing, so that the performance parameter data of the raw materials can be unified, larger data difference is avoided, abnormal data can be checked, and the data reading efficiency of the prediction model is improved.

In another technical scheme, in the process of establishing the prediction model of each performance parameter of the concrete, the performance parameter data of each raw material after standardized processing, the concrete mixing ratio data and each performance parameter data of the concrete are randomly divided into 75% of training sets and 25% of testing sets, the performance parameter data of each raw material after standardized processing and the concrete mixing ratio data are used as input values, each performance parameter data of the concrete is used as an output value, a learner matched with each performance parameter of the concrete is used on the training sets to train regression models of each performance parameter of the concrete, regression accuracy is tested on the testing sets, the hyperparameters of the training models are gradually adjusted according to the regression accuracy, and finally the accurate prediction model of each performance parameter of the concrete is obtained through training.

In the technical scheme, data of input values are randomly divided into a training set and a testing set, in order to test the accuracy of a prediction model, in the process of establishing the prediction model of each performance parameter of concrete, firstly, performance parameter data of raw materials after standardization processing are used as input values, each performance parameter data of the concrete is used as output values, a corresponding regression model is trained on the training set by using a trainer corresponding to each performance parameter of the concrete, the regression accuracy of the regression model is tested on the testing set, the regression hyper-parameter is repeatedly adjusted according to the accuracy, and finally, the accurate prediction model of each performance parameter of the concrete is obtained through training, so that the data in an exhaustive prediction database in S6 are more accurate.

In another technical solution, the regression accuracy of the regression model is measured by using a mean square error as a measurement indicator.

In the technical scheme, the mean square error is a measure for representing the degree of difference between the training set and the test set, and the smaller the mean square error is, the better the accuracy of the data representing various performance parameters of the concrete described by the prediction model is, and the numerical value of the mean square error can directly represent the accuracy of the prediction model.

In another technical scheme, in S6, the maximum value and the minimum value of the screened concrete mix proportion data are multiplied by 110% and 90% respectively, and the maximum value and the minimum value are used as the upper limit and the lower limit of the exhaustive prediction concrete mix proportion, and a exhaustive prediction database is established.

In the technical scheme, exhaustion is one of the most common methods for solving the problem by a computer, firstly, a mathematical model is required to be established, the model can be a group of variables and conditions which the variables need to satisfy, when various performance parameters of the concrete are solved by adopting the exhaustion method, the mathematical model is performance parameter data and concrete mixing ratio data of various screened raw materials, the concrete mixing ratio is determined in an approximate value range according to the related knowledge of the concrete mixing ratio, namely, the maximum value of the concrete mixing ratio data is multiplied by 110 percent, the minimum value is multiplied by 90 percent, the concrete mixing ratio data is sequentially valued in the range, whether the obtained value satisfies the conditions of the mathematical model is judged according to the basic requirements of the concrete mixing ratio, all the values which accord with the conditions are found until the values which accord with the conditions are found, and an exhaustion prediction database is established by the numerical values which accord with the conditions, the concrete mixing proportion is determined in an approximate value range, and the value is taken in the range, so that the calculated amount in exhaustive prediction can be reduced, and the working efficiency is improved.

In another technical scheme, the concrete mixing ratio data between the upper limit and the lower limit is equally divided into ten equal parts as sampling intervals in an exhaustive prediction process.

In the technical scheme, the too large sampling interval can lead to incomplete exhaustion results, the too small sampling interval can lead to overlarge calculated amount, the exhaustion space is equally divided into ten parts, and the calculated amount can not be increased under the condition of complete exhaustion results.

In another technical scheme, the raw materials of the concrete comprise cement, fly ash, mineral powder, an additive, coarse aggregate and fine aggregate.

In the technical scheme, the concrete comprises six raw materials of cement, fly ash, mineral powder, an additive, coarse aggregate and fine aggregate, and the performance parameters of the cement comprise: specific surface area S_c(m²/kg), density ρ_c(g/cm³) Standard water consumption for thickening P_c(%), compressive Strength R_c(MPa) mass m contained in per cubic meter of concrete_c(ii) a The performance parameters of the fly ash include: specific surface area S_F(m²/kg), density ρ_F(g/cm³) Water demand ratio W_F(％)、Coefficient of activity H₂₈(%), the loss on ignition X (%), the mass m contained in each cubic meter of concrete_F(ii) a The performance parameters of the mineral powder comprise: specific surface area S_K(m²/kg), density ρ_k(g/cm³) Water demand ratio W_F(%), coefficient of activity H₂₈(%), the loss on ignition X (%), the mass m contained in each cubic meter of concrete_k(ii) a The performance parameters of the admixture comprise: water reduction rate W_R(%), solid content S (%), density rho_aCompressive strength ratio R_S(%), mass m in each cubic meter of concrete_a(ii) a The performance parameters of the coarse aggregate include: bulk density ρ_L1(kg/m³) Crush value f₁(MPa), maximum particle diameter d₁(mm) and mud content ω c₁(%), the content of mud cake ω c,₁(%), bulk porosity V_L1(%), mass m in each cubic meter of concrete_G1(ii) a The performance parameters of the fine aggregate include: bulk density ρ_L2(kg/m³) Crush value f₂(MPa), maximum particle diameter d₂(mm) and mud content ω c₂(%), the content of mud cake ω c,₂(%), void volume V_L2(%), mass m in each cubic meter of concrete_G2The concrete is composed of a plurality of raw materials, each raw material has a plurality of performance parameters, each performance parameter of the concrete is related to the performance parameter of each raw material in the concrete, and when each performance prediction model of the concrete is established, not only the data of each raw material of the concrete but also the data of the performance parameter corresponding to each raw material are required to be considered, so that the data in the prediction model of each performance parameter of the concrete are more comprehensive.

In another technical scheme, the performance parameters of the concrete comprise slump, concrete strength, chloride ion diffusion coefficient, corrosion resistance coefficient and electric flux.

In another technical scheme, the prediction model of each performance parameter of the concrete comprises a slump prediction model, a concrete strength prediction model, a chloride ion diffusion coefficient prediction model, a corrosion resistance coefficient prediction model and an electric flux prediction model.

In the technical scheme, the performance parameters of the concrete are various, and the parameters are different from one another in characteristics, so that a learner which is adaptive to the parameter characteristics is required to perform prediction modeling on the performance parameters of the concrete, the performance parameter data of the standardized raw materials is used as an input value, the concrete slump is used as an output, a gaussian process regressor is trained on a training set, the regression accuracy is tested on a test set, the gaussian process regression superparameter is repeatedly adjusted according to the accuracy, a slump prediction model with high accuracy is finally obtained through training, and the final gaussian process regression superparameter is kernel-RBF, alpha-0.001, n _ seconds _ optimizer-1, norm _ y-false, copy _ X _ in-true; taking standardized raw material data as input, taking concrete strength as output, training a Gaussian process regressor on a training set, testing regression accuracy on a testing set, repeatedly adjusting Gaussian process regression hyper-parameters according to the accuracy, finally training to obtain a concrete strength prediction model with higher accuracy, wherein the final Gaussian process regression hyper-parameters are kernel-polymeric, alpha-0.0005, n _ thresholds _ optimum-0, normal _ y-false, copy _ X _ train-true; taking standardized raw material data as input, taking a concrete chloride ion diffusion coefficient as output, training a random forest regressor on a training set, testing regression precision on a testing set, repeatedly adjusting a random forest regression hyperparameter according to the precision, and finally training to obtain a chloride ion diffusion coefficient prediction model with higher precision, wherein the final random forest regressor hyperparameter is bootstrap ═ False, max _ features ═ 0.2, min _ samples _ leaf ═ 1, min _ samples _ split ═ 4, and n _ estimators ═ 100; taking standardized raw material data as input, taking the concrete corrosion resistance coefficient as output, training a gradient lifting tree regressor on a training set, testing regression accuracy on a testing set, repeatedly adjusting the super-parameter of the gradient lifting tree according to the accuracy, and finally training to obtain a corrosion resistance coefficient prediction model with higher accuracy, wherein the super-parameter of the gradient lifting tree regressor is learning _ rate ═ 0.1, max _ depth ═ 6, max _ features ═ sqrt', min _ samples _ leaf ═ 5, min _ samples _ split ═ 14, and n _ estimators ═ 50; taking the standardized raw material data as input, taking the electric flux of the concrete as output, training a random forest regressor on a training set, testing regression accuracy on a testing set, repeatedly adjusting the super-parameters of the random forest according to the accuracy, finally training to obtain an electric flux prediction model with higher accuracy, wherein the super-parameters of the random forest regressor are bootstrap, max _ features, 0.05, min _ samples _ leaf, 3, min _ samples _ split, 8 and n _ estimators, 100, and according to the steps, the prediction model of each performance parameter of the concrete can be obtained.

While embodiments of the invention have been described above, it is not limited to the applications set forth in the description and the embodiments, which are fully applicable in various fields of endeavor to which the invention pertains, and further modifications may readily be made by those skilled in the art, it being understood that the invention is not limited to the details shown and described herein without departing from the general concept defined by the appended claims and their equivalents.