Concrete working viscosity and compressive strength prediction method based on statistical analysis

Abstract

The invention relates to the technical field of concrete, in particular to a concrete working viscosity and compressive strength prediction method based on statistical analysis, and the method comprises the following steps: S1, obtaining concrete component information and proportion information; s2, obtaining mixing conditions of concrete, and inputting the mixing conditions, the component information and the proportion information into a pre-constructed flow parameter prediction model to obtain flow parameters; s3, performing viscosity calculation on the flow parameter by using a preset constant term and a reference index to obtain the working viscosity of the concrete; s4, acquiring an environmental condition of a preset future time period; s5, inputting the concrete component information, the proportion information, the working viscosity and the environmental conditions into a pre-constructed compressive strength prediction model to obtain compressive strength; through scientific data processing and model optimization, more accurate working viscosity and compressive strength prediction results can be provided, and powerful support is provided for engineering decision making, so that the construction quality and the structure safety are effectively guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of concrete, and particularly to a method for predicting the working viscosity and compressive strength of concrete based on statistical analysis. Background Art

[0002] In modern concrete projects, the working performance and mechanical properties of concrete are key factors determining its construction quality and long-term durability. The working performance is mainly reflected in the fluidity, pumpability and compactness of concrete, while the mechanical properties are mainly reflected in indicators such as compressive strength and tensile strength. Among them, the working viscosity and compressive strength are two core parameters, which respectively reflect the flow performance of concrete during construction and the load-bearing capacity after hardening.

[0003] Traditionally, evaluating the working viscosity of concrete relies on on-site tests such as slump tests, spread tests and inverted cylinder time measurements, etc. However, these methods are not only time-consuming and laborious, but also difficult to adjust in real time to meet specific engineering requirements. For the prediction of compressive strength, it usually relies on empirical formulas or laboratory tests, which also have the problems of long time consumption and high cost. In addition, the existing methods have poor adaptability to different types of concrete and changing construction conditions.

[0004] The information disclosed in this background art section is only intended to deepen the understanding of the overall background art of the present invention, and should not be regarded as an admission or any form of suggestion that this information constitutes the prior art known to those skilled in the art. Summary of the Invention

[0005] To solve the above technical problems, the present invention provides a method for predicting the working viscosity and compressive strength of concrete based on statistical analysis. Through scientific data processing and model optimization, it can provide more accurate prediction results of working viscosity and compressive strength, providing strong support for engineering decisions, thus effectively ensuring the construction quality and structural safety.

[0006] The method for predicting the working viscosity and compressive strength of concrete based on statistical analysis of the present invention includes: S1 Obtain the concrete component information and mixing ratio information; S2 Obtain the mixing conditions of the concrete, and input the mixing conditions, component information and mixing ratio information into a pre-constructed flow parameter prediction model to obtain flow parameters; S3 Use preset constant terms and reference indices to calculate the viscosity of the flow parameters to obtain the working viscosity of the concrete; S4 Obtain the environmental conditions for a preset future time period; S5 Input the concrete component information and mixing ratio information, working viscosity and environmental conditions into a pre-constructed compressive strength prediction model to obtain the compressive strength.

[0007] As a preferred embodiment of the present invention, the component information includes cement, fly ash, slag powder, silica fume, stone powder, fine aggregate, coarse aggregate, water reducing agent and water.

[0008] As a preferred embodiment of the present invention, the method for constructing the flow parameter prediction model includes: S21 Collect historical data, including concrete component information, mixing ratio information, mixing conditions and corresponding concrete flow parameters, and preprocess the historical data; S22 Screen out the key influencing characteristics of the concrete flow parameters from the historical data, introduce domain knowledge into the key influencing characteristics, construct new influencing characteristics, and combine the key influencing characteristics and the new influencing characteristics to generate a historical data set; S23 Select a machine learning model as the basic architecture of the model, use the historical data set to train and optimize the machine learning model, and obtain a flow parameter prediction model; S24 Deploy the optimized flow parameter prediction model to actual production to predict the concrete flow parameters.

[0009] As a preferred embodiment of the present invention, the flow parameters include slump, spread and inversion time.

[0010] As a preferred embodiment of the present invention, the working viscosity calculation formula in S3 is: S = w 1 + b×w 2 + c×w 3 - d×w 4 ; Wherein, S is the working viscosity, b is the slump, c is the inversion time, d is the spread, and w 1 is a constant term, and w 2 , w 3 and w 4 are the reference indices of the slump, inversion time and spread respectively.

[0011] As a preferred embodiment of the present invention, the method for setting the constant term and reference index in S3 includes: S31 Collect the concrete flow parameters and corresponding working viscosities under different component information, mixing ratio information and mixing conditions, and preprocess the concrete flow parameters to obtain an initial data set; S32 Use statistical methods to analyze the initial data set, establish the relationship between each concrete flow parameter and the working viscosity, and set the initial values of the reference index and the constant term according to the analysis results; S33 Substitute the concrete flow parameters and the initial values of the set reference index and constant term into the working viscosity calculation formula to obtain the predicted value of the concrete working viscosity; S34 Use an independent validation dataset to validate the predicted value of the concrete working viscosity. Based on the validation results, use an optimization algorithm to fine-tune the initial values of the constant term and the reference index.

[0012] As a preferred embodiment of the present invention, the environmental conditions include temperature, humidity, and curing period.

[0013] As a preferred embodiment of the present invention, the method for constructing the compressive strength prediction model includes: S51 Collect historical data, which includes concrete component information and mix ratio information, working viscosity, environmental conditions, and compressive strength; S52 Select several machine learning models as the basic models to construct the first-layer compressive strength prediction model, and train the basic models with the historical data; S53 Select a meta-model to construct the second-layer compressive strength prediction model, and train the meta-model with the predicted values of the basic models; S54 Integrate the trained basic models and the meta-model to obtain the compressive strength prediction model; S55 Use an independent test set to evaluate and optimize the compressive strength prediction model, and deploy the optimized compressive strength prediction model to actual applications to predict the compressive strength of concrete.

[0014] As a preferred embodiment of the present invention, the mixing conditions include: Mixing equipment: type, capacity, and rotation speed; Mixing time: initial mixing time and total mixing time; Temperature and humidity control: ambient temperature, material temperature, and air humidity.

[0015] As a preferred embodiment of the present invention, the method for obtaining the environmental conditions in a future time period includes: S41 Collect and organize the temperature and humidity data for a preset future time period; S42 Divide the time period according to the curing key nodes or the natural day cycle; S43 Extract the temperature and humidity characteristics in each time period, and set corresponding weights according to their importance; S44 Use time series decomposition technology to process the temperature and humidity data; S45 Use the processed temperature and humidity data and the corresponding weights as the environmental conditions in the future time period.

[0016] The beneficial effects of the present invention compared with the prior art are as follows: By systematically integrating multi-source data such as concrete component information, mixing conditions, flow parameters, and environmental conditions, and using a pre-constructed model for calculation and prediction, the work efficiency and accuracy are significantly improved. Compared with traditional on-site tests and empirical formulas, this method not only reduces the dependence on manual operations, lowers time and cost consumption, but also can be adjusted in real time to adapt to specific engineering requirements, greatly enhancing the adaptability to different types of concrete and changing construction conditions. In addition, through scientific data processing and model optimization, more accurate predictions of work viscosity and compressive strength can be provided, providing strong support for engineering decisions, thus effectively ensuring construction quality and structural safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is a schematic flow chart of the present invention; Figure 2 is a schematic flow chart of a method for obtaining environmental conditions in a future time period. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] In order to make the above objects, features, and advantages of the present invention more obvious and understandable, the following detailed description of the specific embodiments of the present invention will be made in conjunction with the accompanying drawings of the specification.

[0019] In the following description, many specific details are set forth in order to fully understand the present invention. However, the present invention can also be implemented in other ways different from those described herein. Those skilled in the art can make similar extensions without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0020] Secondly, the so-called "embodiment" herein refers to specific features, structures, or characteristics that can be included in at least one implementation manner of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment that is mutually exclusive with other embodiments. Embodiment

[0021] Referring to Figure 1 , this embodiment provides a method for predicting the work viscosity and compressive strength of concrete based on statistical analysis, including: S1 Obtain concrete component information and mix ratio information; among them, the component information includes cement, fly ash, slag powder, silica fume, stone powder, fine aggregate, coarse aggregate, water reducer, and water; The mix proportion of concrete usually includes the following main components: Cement: As the main binder of concrete; Fly ash: Mineral admixture to improve workability and long-term performance; Mineral powder (such as slag powder): Mineral admixture that improves density and impermeability; Silica fume: Fine pozzolanic material that enhances early strength and impermeability performance; Stone powder: Used as a filler to improve workability and reduce bleeding; Fine aggregate: Provides the skeletal structure of concrete and affects fluidity; Coarse aggregate: Together with fine aggregate, constitutes the skeleton of concrete and determines volume stability and bearing capacity; Water reducing agent: Reduces the mixing water consumption, improves workability and strength; Water: Participates in the hydration reaction of cement and is one of the key factors in forming concrete; The main function of step S1 is to obtain and record the mix proportion information of concrete according to specific production tasks; S2 obtains the mixing conditions of concrete, and inputs the mixing conditions, component information and mix proportion information into a pre-constructed flow parameter prediction model to obtain flow parameters; Among them, the mixing conditions include: Mixing equipment: Type: Different mixing equipment (such as compulsory mixers, self-falling mixers, etc.) has a significant impact on the mixing effect of concrete. Compulsory mixers are suitable for mixing high-fluidity concrete or concrete containing a large amount of fine aggregate, while self-falling mixers are more suitable for ordinary concrete; Capacity: The capacity of the mixer determines the amount of concrete that can be mixed each time. Being too large or too small will affect the uniformity of concrete; Rotation speed: The rotation speed of the mixer directly affects the mixing efficiency and the uniformity of concrete. Usually, it is necessary to adjust the appropriate rotation speed according to the type and mix ratio of concrete; Mixing time: Initial mixing time: Refers to the time from when all raw materials are added to the mixer until the initial mixing is completed. This time needs to ensure that all materials are fully mixed, but it should not be too long to avoid material separation or overmixing; Total mixing time: Includes the initial mixing time and additional mixing time to ensure that the concrete reaches the best work performance. The total mixing time should be adjusted according to the specific concrete mix ratio and construction requirements; Temperature control: Ambient temperature: The ambient temperature during concrete mixing will affect its setting time and work performance. Under high temperature conditions, the water evaporation rate of concrete is fast, which may lead to a decrease in fluidity; while under low temperature conditions, the setting speed of concrete slows down, and heat preservation measures may need to be taken; Material Temperature: The temperature of raw materials (such as cement, sand, gravel, water) also needs to be controlled. Especially under extreme climate conditions, the temperature of concrete can be adjusted by cooling or heating the raw materials to ensure that it is stirred within an appropriate temperature range. Air Humidity: Air humidity affects the evaporation rate of water in concrete. In a dry environment, the surface of concrete is prone to water loss, resulting in plastic shrinkage cracks; while in a humid environment, excessive moisture absorption needs to be prevented. A method for constructing a flow parameter prediction model includes the following steps: S21 Collect historical data, including concrete component information, mix ratio information, mixing conditions, and corresponding concrete flow parameters, and preprocess the historical data. As a preference of this embodiment, this step collects a large amount of experimental data or historical data, including concrete component information, mix ratio information, mixing conditions, and corresponding flow parameters. To ensure the quality of the collected data, outliers, missing values, and duplicate data in the collected data can be removed, and the collected data can be normalized or standardized.

[0022] S22 Screen out the key influencing characteristics of the concrete flow parameters from the historical data, introduce domain knowledge into the key influencing characteristics, construct new influencing characteristics, and combine the key influencing characteristics and the new influencing characteristics to generate a historical data set. In this step, characteristics that have a significant impact on the fluidity of concrete can be screened out from historical data through statistical analysis, machine learning algorithms, etc. For example, the water-binder ratio, dosage of water reducer, ratio of fine aggregate to coarse aggregate, etc. may be key characteristics; in order to construct a prediction model with stronger prediction ability, based on the physical properties and engineering background of concrete, combined with the experience of domain experts, some new characteristics can be introduced; or the key characteristics screened out and the new characteristics introduced by domain knowledge can be cross-combined to generate higher-order composite characteristics. For example, derivative characteristics such as water-binder ratio, total amount of cementitious materials, and aggregate gradation index can be calculated to better reflect the performance of concrete; finally, the key characteristics screened out and the newly generated characteristics are combined to generate a new historical data set containing all relevant information, which can not only help the model capture more potential complex relationships, but also improve the prediction ability of the model.

[0023] S23 Select a machine learning model as the basic architecture of the model, use the historical data set to train and optimize the machine learning model, and obtain a flow parameter prediction model. S24 Deploy the optimized flow parameter prediction model to actual production to predict the concrete flow parameters.

[0024] Specifically, a machine learning model is selected as the model's infrastructure, which includes recurrent neural networks, long short-term memory networks, convolutional neural networks, etc.; the generated historical dataset is divided into a training set, a validation set, and a test set, and the training set data is used to train the model, adjusting the model parameters to minimize the prediction error; through methods such as cross-validation or grid search, the hyperparameters of the model (such as learning rate, tree depth, regularization parameter, etc.) are optimized; after the model is trained, the validation set is used to evaluate the performance of the model to check for overfitting or underfitting problems; appropriate evaluation metrics can be selected, such as mean squared error (MSE), mean absolute error (MAE), coefficient of determination (R²), etc., to quantify the prediction accuracy of the model. According to the validation results, the model structure or parameters are adjusted and the model is retrained; finally, an independent test set can be used to evaluate the performance of the trained model on unseen data, and the model is tuned according to the evaluation results. The tuning methods include adjusting the model architecture, modifying the loss function, and adjusting the hyperparameters, etc.

[0025] The above process of constructing the flow parameter prediction model ensures the accuracy and reliability of the model through data preparation, feature engineering, selection of appropriate machine learning algorithms, and rigorous training and validation steps. This method first cleans and standardizes the data to improve the quality, and then uses expertise to select key features and create new features to enhance the model performance; then, a powerful deep learning model is adopted to capture complex relationships, and overfitting or underfitting is avoided by optimizing hyperparameters. Finally, an independent test set is used to verify the model effect to ensure its practicality. The overall process is simple and efficient, aiming to improve the prediction accuracy while ensuring the stability and effectiveness of the model in practical applications.

[0026] Among them, the flow parameters include slump, spread, and slump flow time. The slump is a commonly used index to evaluate the fluidity of fresh concrete, which reflects the deformation and flow ability of concrete under its own weight; from the perspective of rheology, although the size of the slump mainly reflects the yield stress of concrete, it is also indirectly related to its plastic viscosity, because the plastic viscosity describes the resistance encountered after the concrete starts to flow, and this resistance will affect the slump shape and height of the concrete. Therefore, the working viscosity of concrete can be indirectly inferred through the slump. The spread measures the ability of concrete to flow freely on a horizontal plane, and it is also closely related to the fluidity of concrete. Concrete with a larger spread usually has a lower plastic viscosity because they encounter less resistance during the flow process. The spread can also be used as a reference index for calculating the working viscosity. The slump time measures the time required for concrete to be poured out of a certain container, which reflects the flow rate and compactability of the concrete. A shorter slump time usually means that the concrete has better fluidity and lower plastic viscosity. Through the slump time, the working viscosity characteristics of the concrete can be further understood; By establishing an appropriate model, the flow parameters can be converted into an estimate of the working viscosity. This conversion not only depends on theoretical analysis but also combines a large number of experimental verifications to ensure its accuracy and reliability.

[0027] S3 uses preset constant terms and reference exponents to calculate the viscosity of the flow parameters to obtain the working viscosity of the concrete; The constant term helps to adjust the calculation result to make it closer to the actual measured value. The constant term can be used to compensate for the deviations caused by factors such as experimental conditions and material property differences; for example, under different temperature or humidity conditions, even with the same material ratio, the fluidity may be different, and by introducing an appropriate constant term, such effects can be corrected; the constant term may come from the average value, standard value of experimental data or the basic value derived from theoretical derivation; The reference exponent helps to establish the mathematical relationship between the input variable (flow parameter) and the output variable (working viscosity). Different concrete formulations and construction conditions may require different reference exponents to accurately reflect the actual situation; for example, when constructing in a high-temperature environment, the water evaporation rate accelerates, which may lead to a decrease in the fluidity of the concrete; at this time, applying a reference exponent that takes into account the temperature effect can help to more accurately predict the working viscosity, thus guiding on-site construction adjustments; Among them, the calculation formula is as follows: S = w 1 + b×w 2 + c×w 3 - d×w 4 ; Among them, S is the working viscosity, b is the slump, c is the slump time, d is the spread, and w 1 is the constant term, and w 2 , w 3 and w 4 are the reference exponents of the slump, slump time, and spread respectively; A higher slump indicates stronger fluidity of the concrete. Plastic viscosity is the viscosity of the concrete, that is, the ability of the concrete to resist flow. The two show a positive correlation; when the slump increases, the fluidity of the concrete increases, but at the same time, the viscosity of the concrete also increases, so the plastic viscosity rises because the concrete is more likely to have internal friction during flow, resulting in a higher plastic viscosity; The spread is the degree of expansion of concrete under stress during the test. It has a negative correlation with viscosity. When the spread increases, the concrete is more prone to deformation and flow, which is usually accompanied by lower internal friction. Therefore, the plastic viscosity decreases, the concrete is more fluid, and the ability to resist internal friction is lower, so the plastic viscosity decreases; The reverse drum time is affected by various factors, including the viscosity, fluidity of the concrete, and internal friction. When the concrete is more viscous or has poor fluidity, the reverse drum time usually extends because the friction between the internal particles of the concrete increases, and it takes more time for the concrete to flow from the mixing container to the target position. Therefore, as the reverse drum time increases, the plastic viscosity also increases; The method for setting the constant term and the reference index includes the following steps: S31 Collect the concrete flow parameters and the corresponding working viscosities under different component information, mixing ratios, and mixing conditions, and preprocess the concrete flow parameters to obtain an initial data set; S32 Use statistical methods to analyze the initial data set, establish the relationship between each concrete flow parameter and the working viscosity, and set the initial values of the reference index and the constant term according to the analysis results; In this embodiment, a large number of concrete experimental data under different component information, mixing ratios, and mixing conditions are collected, including slump, spread, reverse drum time, and the corresponding working viscosity values; in order to improve the quality of the data, outliers, missing values, and duplicate data can be removed, and the data can be normalized or standardized; use statistical methods to analyze the collected initial data set, such as regression analysis, correlation analysis, etc., to determine the relationship between each flow parameter and the working viscosity. According to the analysis results, the initial values of the reference index and the constant term can be initially set in combination with existing experience or industry standards.

[0028] S33 Substitute the concrete flow parameters and the initial values of the set reference index and constant term into the working viscosity calculation formula to obtain the predicted value of the concrete working viscosity; S34 Use an independent verification data set to verify the predicted value of the concrete working viscosity. Based on the verification results, use an optimization algorithm to fine-tune the initial values of the constant term and the reference index.

[0029] Specifically, the preliminary working viscosity can be obtained through model calculation. The initially set constant term and reference exponent can be substituted into the fluidity model (such as the Power Law model) to calculate the working viscosity of concrete. An independent validation dataset different from the initial dataset is used to test the performance of the model and verify the accuracy of the preliminary model. This validation dataset includes the actually measured working viscosity values, and the error between the working viscosity calculated by the model and the actually measured values in the validation dataset is compared. Based on the error in the validation dataset, the constant term and reference exponent can be fine-tuned through optimization algorithms (such as minimizing the error function, Bayesian optimization, etc.) to ensure that these two parameters can better reflect the actually measured working viscosity and improve the prediction performance of the model.

[0030] The above method ensures the accuracy and reliability of the prediction of the working viscosity of concrete by comprehensively using means such as data collection, statistical analysis, and model verification. It determines the relationship between the flow parameters and the working viscosity based on a large amount of experimental data, and optimizes the prediction model by continuously adjusting the reference exponent and the constant term. Finally, it realizes the efficient and accurate prediction of new data, which not only improves the generalization ability and stability of the model, but also provides a scientific basis for the quality control of concrete construction in actual projects.

[0031] S4 Obtain the environmental conditions for a preset future time period; Based on the actual engineering requirements and construction plan, one or more key time nodes are determined as the "preset future time period"; for example, it is a very common practice to evaluate the compressive strength of concrete at time points such as 7 days and 28 days; this is because the strength of concrete increases rapidly in the first few days, then the growth rate slows down, and reaches the design strength at about 28 days; The environmental conditions include: Temperature: The internal and environmental temperatures of concrete have a direct impact on the rate of cement hydration reaction. Higher temperatures can accelerate the hydration process, but excessive temperatures may cause early thermal cracking; Humidity: Maintaining appropriate humidity is very important for preventing the rapid evaporation of surface water in concrete, avoiding plastic shrinkage cracks, and promoting continuous hydration; Curing period: Different types of concrete may require different times to reach the maximum strength under different environmental conditions; Preferably, referring to Figure 2 , the method for obtaining the environmental conditions for the future time period includes: S41 Collect and organize the temperature and humidity data for the preset future time period; S42 Divide the time period according to the key maintenance nodes or the natural day cycle; divide the entire maintenance period into several small time periods according to the key stages of concrete maintenance (such as 7 days, 28 days, etc.) or on a daily basis; each time period corresponds to a different maintenance stage, and these stages have different effects on the development of concrete strength; S43 Extract the temperature and humidity characteristics within each time period and set corresponding weights according to their importance; Among them, the temperature and humidity characteristics include: Average value: Calculate the average temperature and humidity within each small time period; Cumulative effect: Calculate the cumulative temperature and cumulative humidity to reflect the total heat and humidity effect within a period of time; Extreme values: Record the highest temperature, lowest temperature, highest humidity and lowest humidity within each small time period; According to the characteristics of concrete strength growth, set different weights for each time period. In the initial stage (the first 3 days), due to the rapid hydration reaction, the temperature and humidity changes have a greater impact on the strength; in the middle stage (the 4th to 7th days), it is the second; in the later stage (after 7 days), although the strength growth slows down, it still needs to be appropriately concerned; S44 Use time series decomposition technology to process the temperature and humidity data, and decompose the temperature and humidity data into trend components, seasonal components and random components to better understand its long-term trend and periodic changes; S45 Take the processed temperature and humidity data and the corresponding weights as the environmental conditions for future time periods; More specifically, form the temperature and humidity characteristics within each small time period and their corresponding weights into a vector and input it into the compressive strength prediction model. The advantage of doing this is that more information about the time series can be retained, which helps to capture the dynamic change relationship between different time periods. In addition to temperature and humidity information, other relevant factors such as wind speed and sunshine intensity can also be considered to form a multi-dimensional input vector to further improve the prediction ability of the model; Through the above method, not only can the continuous fluctuating temperature and humidity data in future time periods be effectively processed, but also the influence of temperature and humidity on compressive strength at different stages can be accurately reflected, thereby improving the accuracy and reliability of compressive strength prediction. This method not only considers the influence of short-term fluctuations but also takes into account the long-term trend, which helps to more comprehensively evaluate the strength development of concrete.

[0032] S5 Input the concrete component information, mix ratio information, working viscosity and environmental conditions into the pre-constructed compressive strength prediction model to obtain the compressive strength; The component information and mix ratio information of concrete directly determine its physical and chemical properties; for example, increasing the cement dosage usually increases the strength of concrete, but may also increase costs and shrinkage; the water-cement ratio is one of the key factors affecting the strength of concrete, and a lower water-cement ratio often results in higher strength; The workability directly affects the fluidity of concrete. Appropriate fluidity ensures that the concrete can be evenly distributed during pouring and fills all the spaces within the formwork, thus forming a denser structure. This denseness has a direct positive impact on the compressive strength because fewer voids mean higher structural strength; a higher workability helps reduce water separation (bleeding) in the concrete and separation between the aggregates and the cement paste (segregation). Bleeding and segregation can lead to weak layers or holes inside the concrete, which will weaken the overall strength of the concrete; The strength development of concrete is a process of gradual enhancement over time, and this process is significantly affected by environmental conditions (such as temperature and humidity). Suitable environmental conditions help promote the hydration reaction of cement, making the internal structure of the concrete denser and thus increasing the strength; A method for constructing a compressive strength prediction model, including: S51 Collect historical data, where the historical data includes concrete component information and mix ratio information, workability, environmental conditions, and compressive strength; S52 Select several machine learning models as the base models, construct the first-layer compressive strength prediction model, and train the base models with the historical data; As a preference for this implementation, multiple complementary base models (such as support vector machines, decision trees, neural networks, etc.) can be used. Complementary base models can improve the overall performance of the prediction model, avoid overfitting, and use the training set to train all the selected base models separately. Each base model will output a predicted value, and the prediction results of each base model on the training set form the output of the first layer.

[0033] S53 Select a meta-model to construct the second-layer compressive strength prediction model, and train the meta-model with the predicted values of the base models; To improve the accuracy and robustness of the prediction model, a single meta-model (such as linear regression, ridge regression, etc.) can be set in the second layer. Using the output of the base models as input for the final prediction; for the training of the meta-model, to avoid overfitting, the prediction results of the first-layer base models in the validation set or cross-validation can be used as training data to train the meta-model, and the label is the compressive strength value of the concrete.

[0034] S54 Integrate the trained base models and the meta-model to obtain the compressive strength prediction model; The trained meta-model can combine the outputs of the base models into the final prediction. Since each base model may perform well on some data points but poorly on others, the meta-model can assign greater weights to the better-performing base models when learning how to weight the prediction values of different base models, optimizing the final output. Therefore, it can integrate the advantages of different models, effectively handle complex non-linear relationships and multi-dimensional data, and improve the prediction accuracy of the prediction model.

[0035] S55 uses an independent test set to evaluate and optimize the compressive strength prediction model, and deploys the optimized compressive strength prediction model into practical applications to predict the compressive strength of concrete.

[0036] By systematically collecting and processing historical data and using machine learning models for training and optimization, this process not only improves the accuracy and reliability of compressive strength prediction, but also enhances the model's performance in dealing with complex relationships and unseen data. This method enables the finally deployed model to efficiently and accurately predict the compressive strength of concrete in practical engineering applications, providing a scientific basis for optimizing material ratios and improving construction quality, and significantly improving engineering efficiency and safety.

[0037] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. A method for predicting concrete working viscosity and compressive strength based on statistical analysis, characterized in that: include: S1 obtains concrete component information and mix ratio information; S2 obtains the mixing conditions of the concrete, and inputs the mixing conditions, the component information and the mix ratio information into a pre-built flow parameter prediction model to obtain the flow parameters; S3 uses a preset constant term and a reference index to calculate the viscosity of the flow parameter to obtain the working viscosity of the concrete; S4 obtains environmental conditions for a preset future time period; S5: Inputting the concrete component information and mix ratio information, the working viscosity and the environmental conditions into a pre-built compressive strength prediction model to obtain the compressive strength.

2. The method for predicting concrete working viscosity and compressive strength based on statistical analysis according to claim 1, characterized in that: The component information includes cement, fly ash, mineral powder, silica fume, stone powder, fine aggregate, coarse aggregate, water reducing agent and water.

3. The method for predicting concrete working viscosity and compressive strength based on statistical analysis according to claim 1, characterized in that: The method for constructing the flow parameter prediction model comprises: S21 collects historical data, including concrete component information, mix ratio information, mixing conditions, and corresponding concrete flow parameters, and pre-processes the historical data; S22: screening out key influencing features of the concrete flow parameters from the historical data, introducing domain knowledge into the key influencing features, constructing new influencing features, and combining the key influencing features with the new influencing features to generate a historical data set; S23 selects a machine learning model as the basic architecture of the model, and uses the historical data set to train and optimize the machine learning model to obtain a flow parameter prediction model; S24 deploys the optimized flow parameter prediction model to actual production to predict the concrete flow parameters.

4. The method for predicting concrete working viscosity and compressive strength based on statistical analysis according to claim 3, characterized in that: The flow parameters include slump, spread and reversing time.

5. The method for predicting concrete working viscosity and compressive strength based on statistical analysis according to claim 4, characterized in that: The working viscosity calculation formula in S3 is: S = w1 + b × w2 + c × w3 - d × w4; Wherein, S is the working viscosity, b is the slump, c is the inverting time, d is the expansion, w1 is a constant term, and w2, w3 and w4 are reference indices of slump, inverting time and expansion respectively.

6. The method for predicting concrete working viscosity and compressive strength based on statistical analysis according to claim 5, characterized in that: The method for setting the constant term and the reference index in S3 includes: S31 collects concrete flow parameters and corresponding working viscosity under different component information, mix ratio information and mixing conditions, and preprocesses the concrete flow parameters to obtain an initial data set; S32: analyzing the initial data set by using a statistical method, establishing a relationship between each of the concrete flow parameters and the working viscosity, and setting initial values ​​of a reference index and a constant term according to the analysis results; S33: Substituting the concrete flow parameter and the set reference index and initial value of the constant term into a working viscosity calculation formula to obtain a predicted value of concrete working viscosity; S34 verifies the predicted value of concrete working viscosity using an independent validation data set, and based on the validation results, fine-tunes the initial values ​​of the constant term and the reference index using an optimization algorithm.

7. The method for predicting concrete working viscosity and compressive strength based on statistical analysis according to claim 1, characterized in that: The environmental conditions include temperature, humidity and curing period.

8. The method for predicting concrete working viscosity and compressive strength based on statistical analysis according to claim 7, characterized in that: The method for constructing the compressive strength prediction model comprises: S51 collects historical data, wherein the historical data includes concrete component information and mix ratio information, working viscosity, environmental conditions, and compressive strength; S52 selects several machine learning models as basic models, constructs a first-layer compressive strength prediction model, and trains the basic model using the historical data; S53: selecting a meta-model to construct a second-layer compressive strength prediction model, and training the meta-model by using the prediction value of the basic model; S54 integrates the trained basic model and the meta-model to obtain a compressive strength prediction model; S55 uses an independent test set to evaluate and optimize the compressive strength prediction model, deploys the optimized compressive strength prediction model to practical applications, and predicts the compressive strength of concrete.

9. The method for predicting concrete working viscosity and compressive strength based on statistical analysis according to claim 1, characterized in that: The mixing conditions include: Mixing equipment: type, capacity and speed; Stirring time: initial stirring time and total stirring time; Temperature and humidity control: ambient temperature, material temperature and air humidity.

10. The method for predicting concrete working viscosity and compressive strength based on statistical analysis according to claim 7, characterized in that: The method for obtaining the environmental conditions in the future time period includes: S41 collects and organizes temperature and humidity data for a preset future time period; S42 divides time periods according to key maintenance nodes or natural daily cycles; S43 extracts the temperature and humidity features in each time period and sets corresponding weights according to their importance; S44 uses time series decomposition technology to process temperature and humidity data; S45 uses the processed temperature and humidity data and the corresponding weights as environmental conditions for a future time period.

Images