Method and device for determining concrete proportion

Abstract

The invention provides a concrete proportion determination method and device, and the method comprises the steps: carrying out the preprocessing of an obtained initial concrete related data set, and obtaining a sample data set needed by model training; according to the sample data set, training an initial prediction model by adopting a random grid search method based on a pipeline mechanism to obtain a target strength prediction model; packaging the target strength prediction into a target function for Pareto frontier solution; according to the target function, the target strength, the performance parameters of the raw materials required by the concrete, the original cost and the preset constraint condition, a multi-target optimization algorithm is adopted for proportion prediction, and the optimal proportion is obtained. Therefore, through data-driven model training, accurate objective function optimization and efficient proportion prediction, the problems of time consumption, labor consumption, low precision and the like in the existing concrete mix proportion design can be effectively solved.

Description

Technical Field

[0001] The present application relates to the technical field of construction engineering, and in particular to a method and device for determining a concrete mix ratio. Background Art

[0002] Commercial concrete is made by mixing raw materials such as cement, water, sand, stone, and admixtures in a certain proportion. It is widely used in buildings, roads, bridges and other fields. It is the most widely used building material in the world. Compressive strength and fluidity are one of the most important properties of concrete, and the mix ratio of concrete is the main factor determining its performance.

[0003] The conventional concrete mix design method is based on empirical formulas and is ultimately determined through a large number of trial mix experiments and tests. However, this method requires a curing period of at least 28 days from the trial mix experiment to the strength test results, so it is time-consuming, labor-intensive and has low accuracy. Summary of the invention

[0004] In view of this, the purpose of this application is to provide a method and device for determining concrete mix ratio, which can effectively solve the problems of time-consuming, labor-consuming and low precision existing in the existing concrete mix ratio design through data-driven model training, precise objective function optimization and efficient mix ratio prediction.

[0005] The embodiment of the present application provides a method for determining a concrete mix ratio, the method comprising:

[0006] Preprocess the acquired initial concrete-related data set to obtain the sample data set required for model training;

[0007] According to the sample data set, a random grid search method based on a pipeline mechanism is used to train the initial prediction model to obtain a target intensity prediction model;

[0008] Encapsulating the target intensity prediction as an objective function of Pareto front solution;

[0009] According to the objective function, target strength, performance parameters of raw materials required for concrete, original cost, and preset constraints, a multi-objective optimization algorithm is used to predict the mix ratio to obtain the optimal mix ratio.

[0010] Optionally, obtain an initial concrete-related dataset by following the steps below:

[0011] Identify multiple target systems storing concrete related data and determine the type of data stored in each target system;

[0012] Establish data extraction links based on the data types stored in each target system;

[0013] According to the data extraction link, the concrete related data stored in the multi-target system are extracted in association to obtain a plurality of concrete related data subsets, and a concrete related data set is constructed according to the plurality of concrete related data subsets.

[0014] Optionally, the obtained initial concrete-related data set is preprocessed to obtain a sample data set required for model training, including:

[0015] Performing data missing identification on each concrete-related data subset in the initial concrete-related data set to screen out a target data subset with data missing;

[0016] According to the remaining concrete-related data subsets except the target data subset, each target data subset is interpolated and enhanced;

[0017] A sample data set is constructed based on the fully processed target data subset and the remaining concrete-related data subsets.

[0018] Optionally, the concrete-related data includes at least one of the following: raw material data, mix ratio data, and concrete performance data;

[0019] The raw material data includes at least one of the following: fly ash fineness, water requirement, 28-day activity, cement standard consistency, setting time, flexural and compressive strength at each age, crushed stone grading, mud content, crushing index, needle-like proportion, admixture type, density, solid content, net slurry fluidity, pH value, water reduction rate, sand fineness modulus, stone powder content, mud block content, MB value, mineral powder fineness, fluidity ratio, and 28-day activity.

[0020] Optionally, the initial prediction model consists of an XGBoost model and multiple weakly supervised learners packaged by a bootstrap integration strategy.

[0021] Optionally, the training of the initial prediction model using a random grid search method based on a pipeline mechanism according to the sample data set to obtain a target intensity prediction model includes:

[0022] The raw material data and the corresponding mix ratio data in the sample data set are used as input, and the strength at each age is used as output. The random grid search method based on the pipeline mechanism is used to optimize the hyperparameters of the initial prediction model to obtain the optimized hyperparameter combination; wherein the initial prediction model introduces an attention mechanism;

[0023] The target intensity prediction model is established using the optimized hyperparameter combination.

[0024] Optionally, the method of using a multi-objective optimization algorithm to predict the mix ratio according to the objective function, the target strength, the performance parameters of the raw materials required for the concrete, the original cost, and the preset constraints to obtain the optimal mix ratio includes:

[0025] According to the objective function, the target strength, the performance parameters of the raw materials required for the concrete and the original cost, a mix ratio prediction is performed by a multi-objective optimization algorithm under preset constraints to determine a plurality of initial prediction results;

[0026] The multi-attribute decision-making method is used to select the best prediction result from multiple initial prediction results;

[0027] The optimal prediction result is subjected to reverse normalization processing to obtain the optimal ratio.

[0028] The embodiment of the present application also provides a device for determining a concrete mix ratio, the device comprising:

[0029] A preprocessing module is used to preprocess the acquired initial concrete-related data set to obtain a sample data set required for model training;

[0030] A training module, used to train the initial prediction model according to the sample data set by adopting a random grid search method based on a pipeline mechanism to obtain a target intensity prediction model;

[0031] A packaging module, used for packaging the target intensity prediction into an objective function for solving the Pareto front;

[0032] The prediction module is used to predict the mix ratio using a multi-objective optimization algorithm according to the objective function, target strength, performance parameters of raw materials required for concrete, original cost, and preset constraints to obtain the optimal mix ratio.

[0033] Optionally, the determining device further includes an acquisition module, and the acquisition module is used to:

[0034] Identify multiple target systems storing concrete related data and determine the type of data stored in each target system;

[0035] Establish data extraction links based on the data types stored in each target system;

[0036] According to the data extraction link, the concrete related data stored in the multi-target system are extracted in association to obtain a plurality of concrete related data subsets, and a concrete related data set is constructed according to the plurality of concrete related data subsets.

[0037] Optionally, when the preprocessing module is used to preprocess the acquired initial concrete-related data set to obtain a sample data set required for model training, the preprocessing module is used to:

[0038] Performing data missing identification on each concrete-related data subset in the initial concrete-related data set to screen out a target data subset with data missing;

[0039] According to the remaining concrete-related data subsets except the target data subset, each target data subset is interpolated and enhanced;

[0040] A sample data set is constructed based on the fully processed target data subset and the remaining concrete-related data subsets.

[0041] Optionally, the concrete-related data includes at least one of the following: raw material data, mix ratio data, and concrete performance data;

[0042] The raw material data includes at least one of the following: fly ash fineness, water requirement, 28-day activity, cement standard consistency, setting time, flexural and compressive strength at each age, crushed stone grading, mud content, crushing index, needle-like proportion, admixture type, density, solid content, net slurry fluidity, pH value, water reduction rate, sand fineness modulus, stone powder content, mud block content, MB value, mineral powder fineness, fluidity ratio, and 28-day activity.

[0043] Optionally, the initial prediction model consists of an XGBoost model and multiple weakly supervised learners packaged by a bootstrap integration strategy.

[0044] Optionally, when the training module is used to train the initial prediction model according to the sample data set using a random grid search method based on a pipeline mechanism to obtain a target intensity prediction model, the training module is used to:

[0045] The raw material data and the corresponding mix ratio data in the sample data set are used as input, and the strength at each age is used as output. The random grid search method based on the pipeline mechanism is used to optimize the hyperparameters of the initial prediction model to obtain the optimized hyperparameter combination; wherein the initial prediction model introduces an attention mechanism;

[0046] The target intensity prediction model is established using the optimized hyperparameter combination.

[0047] Optionally, when the prediction module is used to predict the mix ratio using a multi-objective optimization algorithm according to the objective function, the target strength, the performance parameters of the raw materials required for the concrete, the original cost, and the preset constraints to obtain the optimal mix ratio, the prediction module is used to:

[0048] According to the objective function, the target strength, the performance parameters of the raw materials required for the concrete and the original cost, a mix ratio prediction is performed by a multi-objective optimization algorithm under preset constraints to determine a plurality of initial prediction results;

[0049] The multi-attribute decision-making method is used to select the best prediction result from multiple initial prediction results;

[0050] The optimal prediction result is subjected to reverse normalization processing to obtain the optimal ratio.

[0051] An embodiment of the present application also provides an electronic device, comprising: a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate through the bus, and when the machine-readable instructions are executed by the processor, the steps of the determination method as described above are performed.

[0052] An embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the determination method described above are executed.

[0053] The embodiments of the present application provide a method and device for determining a concrete mix ratio. The method comprises: preprocessing an acquired initial concrete-related data set to obtain a sample data set required for model training; training an initial prediction model based on the sample data set using a random grid search method based on a pipeline mechanism to obtain a target strength prediction model; encapsulating the target strength prediction as an objective function solved by a Pareto frontier; and using a multi-objective optimization algorithm to predict the mix ratio based on the objective function, the target strength, the performance parameters of the raw materials required for the concrete, the original cost, and preset constraints to obtain an optimal mix ratio.

[0054] In this way, this application can effectively solve the problems of time-consuming, labor-consuming and low precision in the existing concrete mix design through data-driven model training, precise objective function optimization and efficient mix ratio prediction. This solution improves work efficiency by reducing the number of experiments, and can effectively reduce the waste of raw materials by providing a more reasonable mix ratio scheme, thereby reducing costs; this solution also enhances the accuracy and reliability of the design. By implementing this solution, the scientificity and efficiency of the concrete production process can be significantly improved, and the technological progress of the industry can be promoted.

[0055] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0057] Figure 1 A flow chart of a method for determining a concrete mix ratio provided in an embodiment of the present application;

[0058] Figure 2 A schematic diagram of a process for obtaining an initial concrete-related data set provided for this application;

[0059] Figure 3 An example of experimental results of a concrete mix ratio scheme provided for this application;

[0060] Figure 4 A schematic diagram of the structure of a device for determining a concrete mix ratio provided in an embodiment of the present application;

[0061] Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0062] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application usually described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, each other embodiment obtained by those skilled in the art without making creative work belongs to the scope of protection of the present application.

[0063] Commercial concrete is made by mixing raw materials such as cement, water, sand, stone, and admixtures in a certain proportion. It is widely used in buildings, roads, bridges and other fields. It is the most widely used building material in the world. Compressive strength and fluidity are one of the most important properties of concrete, and the mix ratio of concrete is the main factor determining its performance.

[0064] The conventional concrete mix design method is based on empirical formulas and is ultimately determined through a large number of trial mix experiments and tests. However, this method requires a curing period of at least 28 days from the trial mix experiment to the strength test results, so it is time-consuming, labor-intensive and has low accuracy.

[0065] Based on this, the embodiments of the present application provide a method and device for determining concrete mix ratio, which can effectively solve the problems of time-consuming, labor-consuming and low precision existing in the existing concrete mix ratio design through data-driven model training, precise objective function optimization and efficient mix ratio prediction.

[0066] See also Figure 1 , Figure 1 This is a flow chart of a method for determining a concrete mix ratio provided in an embodiment of the present application. Figure 1 As shown in , the determination method provided in the embodiment of the present application includes:

[0067] S101. Preprocess the acquired initial concrete-related data set to obtain a sample data set required for model training.

[0068] S102: According to the sample data set, a random grid search method based on a pipeline mechanism is used to train the initial prediction model to obtain a target intensity prediction model.

[0069] S103: Encapsulate the target intensity prediction as an objective function for solving the Pareto front.

[0070] S104, according to the objective function, the target strength, the performance parameters of the raw materials required for the concrete, the original cost, and the preset constraints, a multi-objective optimization algorithm is used to perform a mix ratio prediction to obtain an optimal mix ratio.

[0071] The embodiment of the present application provides an implementation process of a method for determining a concrete mix ratio. First, the acquired initial concrete-related data is preprocessed to provide a complete and structured sample data set for subsequent model training; then, the initial prediction model is trained using a random grid search method based on a pipeline mechanism to ensure the connection between the steps and find the best model configuration, thereby obtaining a concrete strength prediction model; then, the trained target strength prediction model is encapsulated into an objective function, which can be used to solve the Pareto frontier, so that the strength prediction model is converted into an objective function that can be used for optimization algorithms; finally, according to the encapsulated objective function, target strength, performance parameters of concrete raw materials, original cost and preset constraints, a multi-objective optimization algorithm is used to predict the optimal mix ratio of concrete, so that a balance can be found between different objectives to obtain the optimal concrete mix ratio. Therefore, this solution can effectively solve the problems of time-consuming, labor-consuming and low precision in the existing concrete mix ratio design through data-driven model training, precise objective function optimization and efficient mix ratio prediction.

[0072] Before executing step S101, an initial concrete-related data set needs to be obtained.

[0073] In one embodiment provided in the present application, an initial concrete-related data set is obtained by the following steps:

[0074] S201, determining multiple target systems storing concrete-related data, and determining the type of data stored in each target system.

[0075] S202, establishing a data extraction link according to the data type stored in each target system;

[0076] S203, performing association extraction on the concrete related data stored in the multi-target system according to the data extraction link to obtain a plurality of concrete related data subsets, and constructing a concrete related data set according to the plurality of concrete related data subsets.

[0077] Regarding step S201, in this step, it is first necessary to determine which systems store data related to concrete, and clarify the type of data stored in each system.

[0078] These target systems may include ERP systems, integrated DCS control systems, press control systems, quality management systems, etc. The data types in each system may include material composition, test results, production records, etc.

[0079] Here, ERP system: As the core of data collection, ERP (Enterprise Resource Planning) system is responsible for managing various resources and business processes of the enterprise. In concrete production, ERP system usually contains information such as raw material management, production plan, mix ratio data, logistics, etc.

[0080] DCS control system: DCS (distributed control system) is mainly used for real-time monitoring and control of operations in the concrete production process, especially equipment control and data collection on the production line, such as raw material feeding, mixing, and transportation.

[0081] Press control system: This system is mainly used to control the operation of the press in concrete quality inspection. It is usually used to test the compressive strength of concrete samples and generate mechanical properties data of concrete products.

[0082] Quality management system: This system is used to record and manage the quality inspection results of products during the concrete production process, including key quality indicators such as slump and compressive strength.

[0083] Different target systems generally store different types of data.

[0084] With respect to step S202, in this step, association clues are determined according to the data types stored in each target system, and a data extraction link is established according to the association clues.

[0085] For example, the associated clues may include factory entry time, waybill number, and other clues, which are used to match and associate data from different sources.

[0086] The data stored in different target systems may include raw material data, mix ratio data, and concrete performance data.

[0087] In this way, associating data through associative clues (data extraction links can be established) can include: associating raw material information in the ERP system with actual usage through factory entry time or waybill number. Comparing the designed mix ratio in the ERP system with the mix ratio used in actual production. Associating the actual usage of concrete with product quality test data (such as compressive strength, etc.) recorded in the quality management system.

[0088] Regarding step S203, in this step, concrete-related data is extracted from multiple target systems according to the established data extraction link, and the data subsets from different systems are unified into a complete data set to determine the concrete-related data set.

[0089] For example, this solution can integrate raw material quality, proportion usage, actual usage and concrete product quality data into a unified link, forming a complete and transparent data link. This can achieve data interconnection and interoperability from the time raw materials enter the factory to concrete production, usage and product quality, and ensure that data throughout the process is traceable.

[0090] And by building a data extraction link, users can extract key fields such as "entry time" or "waybill number" with one click. This makes it easy to obtain all relevant data of the entire production process, facilitating production monitoring, quality management and data analysis.

[0091] For examples, see Figure 2 , Figure 2 A schematic diagram of the process of obtaining an initial concrete-related data set provided in this application. Figure 2 As shown, the acquisition process includes: S1: determine the actual strength test result; S2: group according to the standard ratio; S3: determine whether there are multiple groups of data, if so, execute step S4, if not, execute step S5; S4: obtain the data corresponding to the task with the largest production volume; S5: obtain the ratio information; S6: determine whether there are multiple groups of data, if so, execute step S7, if not, execute step S5; S7: obtain the latest data; S8: obtain the ratio material information.

[0092] For example, the proportion material information may include material information of cement, material information of sand, and material information of stone, etc., and the extraction principle of each material data is limited.

[0093] Continuing with step S101, the concrete-related data includes at least one of the following: raw material data, mix ratio data, and concrete performance data. Others may also be included.

[0094] The raw material data includes at least one of the following: fly ash fineness, water requirement, 28-day activity, cement standard consistency, setting time, flexural and compressive strength at each age, crushed stone grading, mud content, crushing index, needle-like proportion, admixture type, density, solid content, net slurry fluidity, pH value, water reduction rate, sand fineness modulus, stone powder content, mud block content, MB value, mineral powder fineness, fluidity ratio, and 28-day activity.

[0095] The concrete performance data may include strength at different ages and fluidity.

[0096] Continuing with step S101, the preprocessing may include operations such as removing outliers, supplementing missing values, and standardization.

[0097] In one embodiment provided in the present application, the preprocessing of the acquired initial concrete-related data set to obtain the sample data set required for model training includes:

[0098] S1011, performing data missing identification on each concrete-related data subset in the initial concrete-related data set, and screening out target data subsets with data missing.

[0099] S1012: Perform interpolation and enhancement processing on each target data subset according to the remaining concrete-related data subsets except the target data subset.

[0100] S1013: construct a sample data set based on the target data subset that has been processed and the remaining concrete-related data subsets.

[0101] With respect to step S1011, in this step, data missing identification is performed on each subset in the initial concrete data set, and target data subsets with missing data are screened out.

[0102] For example, suppose the concrete-related data subset contains concrete mix information and strength test results of cement, sand, gravel, water, and admixtures. If some records are missing the amount of admixtures used or the proportion of certain materials, these data subsets will be marked as target data subsets with missing data.

[0103] With respect to step S1012, in this step, the missing contents in the target data subset are inferred based on the complete data in the remaining concrete-related data subsets except the target data subset, and the missing contents in the target data subset are filled or enhanced based on the inference results.

[0104] Here, the adaptive MixUp method can be used to interpolate and enhance the target data subset. The basic principle is to perform weighted summation of similar data to predict the value of missing data (i.e., unlabeled data), and generate pre-labeled data from unlabeled data through a multi-layer perceptron regression model. The model is iterated based on the Pseudo-Huber loss function, and the mask matrix is ​​introduced to achieve smooth calculation of the loss function in the case of missing data until the most appropriate weight value λ is obtained, the predicted value of the missing data is generated, and the final data enhancement model is obtained, thereby supplementing the target data subset.

[0105] After each target data subset is supplemented, the target data subset can be added to the remaining concrete-related data subsets as reference data for the next target data subset, thereby achieving expansion while supplementing and supplement while expanding.

[0106] Regarding step S1013, in this step, after completing the interpolation and enhancement processing of the target data subset, the processed target data subset is combined with the remaining non-missing concrete data subsets to construct a complete sample data set. This sample data set will contain all pre-processed data and can be used for subsequent model training.

[0107] Among them, the sample data set can be divided into a training set (70%), a test set (20%) and a validation set (10%).

[0108] For step S102, the initial prediction model is composed of an XGBoost model and a plurality of weakly supervised learners packaged by a bootstrap integration strategy.

[0109] The weakly supervised learner may include: SVM, RF, MLP, etc.

[0110] It should be noted that due to the high dimension of XGBoost and too many features of commercial concrete data, it is difficult to optimize the model parameters. A random grid search method based on a pipeline mechanism is used to optimize the hyperparameters, taking into account both training efficiency and accuracy.

[0111] In an embodiment provided in the present application, the initial prediction model is trained according to the sample data set using a random grid search method based on a pipeline mechanism to obtain a target intensity prediction model, including:

[0112] S1021. The raw material data and the corresponding mix ratio data in the sample data set are taken as input, and the strength at each age is taken as output. The random grid search method based on the pipeline mechanism is used to optimize the hyperparameters of the initial prediction model to obtain the optimized hyperparameter combination.

[0113] S1022. Establish a target intensity prediction model using the optimized hyperparameter combination.

[0114] For step S1021, the initial prediction model introduces an attention mechanism.

[0115] It should be noted that due to the large number of factors affecting concrete performance, pure data-driven methods may produce results that violate common sense. Therefore, an attention mechanism is introduced to constrain the influence weights of different features based on expert experience and industry knowledge to improve the accuracy and interpretability of the model.

[0116] In step S1021, the raw material data and mix ratio data in the sample data set are used as the input of the model, and the concrete strength at different ages is used as the output of the model. The random grid search method based on the pipeline mechanism is used to optimize the hyperparameters and obtain the optimal hyperparameter combination.

[0117] Raw material data include but are not limited to fly ash fineness, water requirement, 28-day activity, cement standard consistency, setting time, flexural and compressive strength at each age, crushed stone grading, mud content, crushing index, needle-like proportion, admixture type, density, solid content, net slurry fluidity, pH value, water reduction rate, sand fineness modulus, stone powder content, mud block content, MB value, mineral powder fineness, fluidity ratio, 28-day activity, etc.

[0118] Mix proportions refer to the proportions of different materials in concrete, which will affect the final concrete strength.

[0119] Age strength refers to the strength performance of concrete in different time periods (such as 7 days, 28 days, etc.).

[0120] Regarding step S1022, in this step, after finding the optimized hyperparameter combination, these parameters are used to train the final target intensity prediction model.

[0121] In this way, based on the known production formula (concrete mix ratio) and the properties of the raw materials used, the trained target strength prediction model can effectively and accurately predict the concrete strength.

[0122] Regarding step S103, in this step, the trained target intensity prediction model is encapsulated into an objective function and used to solve the Pareto front.

[0123] It should be noted that the Pareto front is a concept used in multi-objective optimization, which means that when optimizing multiple objectives simultaneously, it is impossible to further improve one objective without compromising the solution set of other objectives.

[0124] Regarding step S104, in this step, according to the encapsulated objective function, target strength, performance parameters of concrete raw materials, original cost of concrete raw materials and preset constraints, a multi-objective optimization algorithm is used to find a balance between different objectives to obtain the optimal concrete mix ratio.

[0125] The target intensity is pre-specified, and the constraints may include a cost cap or environmental protection requirements, etc.

[0126] The multi-objective optimization algorithm can adopt MOEA / D and NSGAII etc.

[0127] In one embodiment provided in the present application, the multi-objective optimization algorithm is used to predict the mix ratio according to the objective function, the target strength, the performance parameters of the raw materials required for the concrete, the original cost, and the preset constraints to obtain the optimal mix ratio, including:

[0128] S1041. According to the objective function, the target strength, the performance parameters of the raw materials required for concrete and the original cost, a mix ratio prediction is performed using a multi-objective optimization algorithm under preset constraints to determine a plurality of initial prediction results.

[0129] S1042. Use a multi-attribute decision-making method to select the best prediction result from multiple initial prediction results.

[0130] S1043. Perform reverse normalization processing on the optimal prediction result to obtain an optimal ratio.

[0131] For step S1041, in this step, a multi-objective optimization algorithm is used to predict the concrete mix ratio based on the objective function (target strength prediction model), the target strength of concrete (pre-specified value), the performance parameters of raw materials (such as cement strength, aggregate particle size, etc.), the original cost and preset constraints (such as cost cap or environmental protection requirements) to generate multiple initial prediction results.

[0132] Here, each result represents a possible raw material combination and ratio scheme.

[0133] For step S1042, a multi-attribute decision-making method, such as the TOPSIS method, is used to score each ratio according to attributes such as strength and cost, and finally the optimal ratio result is screened out.

[0134] Regarding step S1043, in this step, the optimal ratio result may be normalized (such as all materials are expressed in relative proportions). For practical use, these normalized data need to be converted back to the actual material usage. Through the reverse normalization process, the actual optimal ratio is finally obtained.

[0135] To further understand the effect of this solution, please refer to the example. Figure 3 , Figure 3 An example of experimental results of a concrete mix ratio scheme provided for this application. The basic information and results used here include: Sample size: 2663 rows; Modules: 1. Concrete performance indicators; 2. Mix ratio indicators (including water, cement, limestone, slag, sand, fly ash, granite, water reducer, factory grade); 3. Each material indicator. Prediction target: 28-day compressive strength of concrete; Model: XGBoost; Result: r 2 is 0.8, the average deviation is 1.08MPa, and the prediction deviation of the 134 samples in the test set is within plus or minus 2MPa, accounting for 82%. Figure 3 As shown above, the test data set results using XGBoost, the blue line represents the target value, and the red line represents the predicted value. Figure 3 The following figure is an example of the comparison between the predicted results and the actual results. In the figure below, the blue line represents the actual value, the green line represents the predicted value within the range of ±2, and the red dot represents the predicted value outside the range of ±2. Therefore, this solution can determine the optimal concrete mix ratio to meet the expected requirements faster and more accurately.

[0136] In this way, this application can effectively solve the problems of time-consuming, labor-consuming and low precision in the existing concrete mix design through data-driven model training, precise objective function optimization and efficient mix ratio prediction. This solution improves work efficiency by reducing the number of experiments, and can effectively reduce the waste of raw materials by providing a more reasonable mix ratio scheme, thereby reducing costs; this solution also enhances the accuracy and reliability of the design. By implementing this solution, the scientificity and efficiency of the concrete production process can be significantly improved, and the technological progress of the industry can be promoted.

[0137] Based on the same inventive concept, a determination device corresponding to the determination method is also provided in the embodiment of the present application. Since the principle of solving the problem by the device in the embodiment of the present application is similar to the above-mentioned determination method in the embodiment of the present application, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be repeated.

[0138] See also Figure 4 , Figure 4 This is a schematic diagram of a device for determining a concrete mix ratio provided in an embodiment of the present application. Figure 4 As shown in , the determining device 400 includes:

[0139] A preprocessing module 410 is used to preprocess the acquired initial concrete-related data set to obtain a sample data set required for model training;

[0140] A training module 420 is used to train the initial prediction model according to the sample data set by using a random grid search method based on a pipeline mechanism to obtain a target intensity prediction model;

[0141] A packaging module 430, for packaging the target intensity prediction into an objective function for solving the Pareto front;

[0142] The prediction module 440 is used to predict the mix ratio using a multi-objective optimization algorithm according to the objective function, target strength, performance parameters of raw materials required for concrete, original cost, and preset constraints to obtain the optimal mix ratio.

[0143] Optionally, the determining device 400 further includes an acquisition module (not shown in the figure), and the acquisition module is used to:

[0144] Identify multiple target systems storing concrete related data and determine the type of data stored in each target system;

[0145] Establish data extraction links based on the data types stored in each target system;

[0146] According to the data extraction link, the concrete related data stored in the multi-target system are extracted in association to obtain a plurality of concrete related data subsets, and a concrete related data set is constructed according to the plurality of concrete related data subsets.

[0147] Optionally, when the preprocessing module 410 is used to preprocess the acquired initial concrete-related data set to obtain a sample data set required for model training, the preprocessing module 410 is used to:

[0148] Performing data missing identification on each concrete-related data subset in the initial concrete-related data set to screen out a target data subset with data missing;

[0149] According to the remaining concrete-related data subsets except the target data subset, each target data subset is interpolated and enhanced;

[0150] A sample data set is constructed based on the fully processed target data subset and the remaining concrete-related data subsets.

[0151] Optionally, the concrete-related data includes at least one of the following: raw material data, mix ratio data, and concrete performance data;

[0152] The raw material data includes at least one of the following: fly ash fineness, water requirement, 28-day activity, cement standard consistency, setting time, flexural and compressive strength at each age, crushed stone grading, mud content, crushing index, needle-like proportion, admixture type, density, solid content, net slurry fluidity, pH value, water reduction rate, sand fineness modulus, stone powder content, mud block content, MB value, mineral powder fineness, fluidity ratio, and 28-day activity.

[0153] Optionally, the initial prediction model consists of an XGBoost model and multiple weakly supervised learners packaged by a bootstrap integration strategy.

[0154] Optionally, when the training module 420 is used to train the initial prediction model according to the sample data set using a random grid search method based on a pipeline mechanism to obtain a target intensity prediction model, the training module 420 is used to:

[0155] The raw material data and the corresponding mix ratio data in the sample data set are used as input, and the strength at each age is used as output. The random grid search method based on the pipeline mechanism is used to optimize the hyperparameters of the initial prediction model to obtain the optimized hyperparameter combination; wherein the initial prediction model introduces an attention mechanism;

[0156] The target intensity prediction model is established using the optimized hyperparameter combination.

[0157] Optionally, when the prediction module 440 is used to predict the mix ratio using a multi-objective optimization algorithm according to the objective function, the target strength, the performance parameters of the raw materials required for the concrete, the original cost, and the preset constraints to obtain the optimal mix ratio, the prediction module 440 is used to:

[0158] According to the objective function, the target strength, the performance parameters of the raw materials required for the concrete and the original cost, a mix ratio prediction is performed by a multi-objective optimization algorithm under preset constraints to determine a plurality of initial prediction results;

[0159] The multi-attribute decision-making method is used to select the best prediction result from multiple initial prediction results;

[0160] The optimal prediction result is subjected to reverse normalization processing to obtain the optimal ratio.

[0161] See also Figure 5 , Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 5 As shown in , the electronic device 500 includes a processor 510 , a memory 520 and a bus 530 .

[0162] The memory 520 stores machine-readable instructions executable by the processor 510. When the electronic device 500 is running, the processor 510 communicates with the memory 520 via the bus 530. When the machine-readable instructions are executed by the processor 510, the above-mentioned figure can be executed. Figure 3 The specific implementation of the steps in the method embodiment shown can be found in the method embodiment, and will not be repeated here.

[0163] The present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the computer program can execute the above-mentioned Figure 3 The specific implementation of the steps in the method embodiment shown can be found in the method embodiment, and will not be repeated here.

[0164] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0165] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of devices or units can be electrical, mechanical or other forms.

[0166] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0167] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0168] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present application can essentially or in other words, the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0169] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The protection scope of the present application is not limited thereto. Although the present application is described in detail with reference to the above-mentioned embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-mentioned embodiments within the technical scope disclosed in the present application, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A method for determining a concrete mix ratio, characterized in that: The determination method comprises: Preprocess the acquired initial concrete-related data set to obtain the sample data set required for model training; According to the sample data set, a random grid search method based on a pipeline mechanism is used to train the initial prediction model to obtain a target intensity prediction model; Encapsulating the target intensity prediction as an objective function of Pareto front solution; According to the objective function, target strength, performance parameters of raw materials required for concrete, original cost, and preset constraints, a multi-objective optimization algorithm is used to predict the mix ratio to obtain the optimal mix ratio.

2. The determination method according to claim 1, characterized in that: Obtain the initial concrete-related dataset by following the steps below: Identify multiple target systems storing concrete related data and determine the type of data stored in each target system; Establish data extraction links based on the data types stored in each target system; According to the data extraction link, the concrete related data stored in the multi-target system are extracted in association to obtain a plurality of concrete related data subsets, and a concrete related data set is constructed according to the plurality of concrete related data subsets.

3. The determination method according to claim 1, characterized in that: The preprocessing of the acquired initial concrete-related data set to obtain a sample data set required for model training includes: Performing data missing identification on each concrete-related data subset in the initial concrete-related data set to screen out a target data subset with data missing; According to the remaining concrete-related data subsets except the target data subset, each target data subset is interpolated and enhanced; A sample data set is constructed based on the fully processed target data subset and the remaining concrete-related data subsets.

4. The determination method according to claim 1, characterized in that: The concrete-related data includes at least one of the following: raw material data, mix ratio data, and concrete performance data; The raw material data includes at least one of the following: fly ash fineness, water requirement, 28-day activity, cement standard consistency, setting time, flexural and compressive strength at each age, crushed stone grading, mud content, crushing index, needle-like proportion, admixture type, density, solid content, net slurry fluidity, pH value, water reduction rate, sand fineness modulus, stone powder content, mud block content, MB value, mineral powder fineness, fluidity ratio, and 28-day activity.

5. The determination method according to claim 1, characterized in that: The initial prediction model consists of an XGBoost model and multiple weakly supervised learners packaged by a bootstrap ensemble strategy.

6. The determination method according to claim 1, characterized in that: The initial prediction model is trained according to the sample data set by using a random grid search method based on a pipeline mechanism to obtain a target intensity prediction model, including: The raw material data and the corresponding mix ratio data in the sample data set are used as input, and the strength at each age is used as output. The random grid search method based on the pipeline mechanism is used to optimize the hyperparameters of the initial prediction model to obtain the optimized hyperparameter combination; wherein the initial prediction model introduces an attention mechanism; The target intensity prediction model is established using the optimized hyperparameter combination.

7. The determination method according to claim 1, characterized in that: The method uses a multi-objective optimization algorithm to predict the mix ratio according to the objective function, the target strength, the performance parameters of the raw materials required for the concrete, the original cost, and the preset constraints to obtain the optimal mix ratio, including: According to the objective function, the target strength, the performance parameters of the raw materials required for the concrete and the original cost, a mix ratio prediction is performed by a multi-objective optimization algorithm under preset constraints to determine a plurality of initial prediction results; The multi-attribute decision-making method is used to select the best prediction result from multiple initial prediction results; The optimal prediction result is subjected to reverse normalization processing to obtain the optimal ratio.

8. A device for determining concrete mix ratio, characterized in that: The determining device comprises: A preprocessing module is used to preprocess the acquired initial concrete-related data set to obtain a sample data set required for model training; A training module, used to train the initial prediction model according to the sample data set by adopting a random grid search method based on a pipeline mechanism to obtain a target intensity prediction model; A packaging module, used for packaging the target intensity prediction into an objective function for solving the Pareto front; The prediction module is used to predict the mix ratio using a multi-objective optimization algorithm according to the objective function, target strength, performance parameters of raw materials required for concrete, original cost, and preset constraints to obtain the optimal mix ratio.

9. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate through the bus, and the machine-readable instructions are executed by the processor to perform the steps of the determination method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the determination method according to any one of claims 1 to 7 are executed.

Images