Cement strength prediction algorithm based on big data

Abstract

The invention relates to the technical field of data processing, in particular to a cement strength prediction algorithm based on big data. The system comprises a data acquisition unit, and the data acquisition unit is used for collecting concrete mix proportion data in a historical engineering project and establishing a database; the model processing unit is used for training and verifying the prediction model based on a multivariate adaptive regression spline algorithm; the predictive analysis unit is used for solving a slump predictive value on the basis of the matching information of the concrete to be adopted and the trained model so as to predict the strength of the cement; and the optimization updating unit is used for carrying out iterative optimization on the concrete proportion based on a gradient descent method and according to the prediction result.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and more specifically, to a cement strength prediction algorithm based on big data. Background Technology

[0002] The cement strength prediction algorithm based on big data is a cutting-edge technology that utilizes modern data science techniques and machine learning methods to accurately predict the strength of cement products. The core operation of this algorithm involves in-depth analysis of massive amounts of historical data, covering various key parameters throughout the entire cement production process, such as the chemical composition of raw materials, specific production process settings, and various external environmental conditions. It also includes strength test data of finished cement products. Through comprehensive analysis of this data, the algorithm constructs a highly accurate predictive model capable of effectively anticipating the future strength performance of cement products.

[0003] Existing cement strength prediction algorithms based on big data have limitations when dealing with complex cement strength prediction problems. For example, they struggle to capture the intricate relationships between cement strength and various influencing factors, and their processing capacity is insufficient when faced with large amounts of data, making them susceptible to noise and outliers. Due to the nonlinear, multivariate, uncertain, and multi-time-delay characteristics of the cement production process, traditional prediction methods may fail to accurately capture all influencing factors, resulting in poor prediction model accuracy. Therefore, this paper proposes a cement strength prediction algorithm based on big data. Summary of the Invention

[0004] The purpose of this invention is to provide a cement strength prediction algorithm based on big data, addressing the limitations of existing big data-based cement strength prediction algorithms in handling complex cement strength prediction problems. For example, these algorithms struggle to capture the complex relationships between cement strength and various influencing factors, and their processing capacity is insufficient when faced with large amounts of data, making them susceptible to noise and outliers. Due to the nonlinear, multivariate, uncertain, and multi-time-delay characteristics of the cement production process, traditional prediction methods may fail to accurately capture all influencing factors, resulting in poor accuracy of the prediction model.

[0005] To achieve the above objectives, the present invention aims to provide a cement strength prediction algorithm based on big data, including a data acquisition unit, which is used to collect concrete mix proportion data from historical engineering projects and establish a database.

[0006] A model processing unit, which trains and validates a prediction model based on a multivariate adaptive regression spline algorithm;

[0007] The predictive analysis unit calculates the slump prediction value based on the concrete mix design information to be used and the trained model, thereby predicting the strength of the cement.

[0008] An optimization and update unit is used to iteratively optimize the concrete mix design based on the gradient descent method and the prediction results.

[0009] As a further improvement to this technical solution, the data acquisition unit includes a database module, which is used to organize and store the collected concrete mix proportion data.

[0010] As a further improvement to this technical solution, the model processing unit establishes a prediction model based on the multivariate adaptive regression spline algorithm, which provides a solid foundation for the prediction analysis unit. The specific steps for training and validating the prediction model are as follows:

[0011] S3.1 Clean and standardize the collected data to ensure its accuracy and comparability;

[0012] S3.2 Training the prediction model using a linear regression algorithm to enable the prediction model to accurately predict concrete slump;

[0013] S3.3 Input the reserved test dataset into the trained model and evaluate the model using mean squared error.

[0014] As a further improvement to this technical solution, the linear regression algorithm in S3.2 is specifically as follows:

[0015] X = [x1, x2, ..., x] n ];

[0016] y i =β0 + β1x1 + β2x2 + ... + β n x n ;

[0017] Where X represents the input feature matrix; n represents n features: y i Represents the true value of concrete slump; β0 represents the intercept term; β n This indicates that it corresponds to the nth feature x.

[0018] As a further improvement to this technical solution, in S3.3, the mean square error is specifically as follows:

[0019]

[0020] Where L(β) represents the mean squared error function; m represents the number of samples; This represents the predicted value of concrete slump.

[0021] As a further improvement to this technical solution, the prediction and analysis unit is used to predict the slump of the new concrete mix and the strength of the cement. The specific steps of the prediction and analysis are as follows:

[0022] S6.1 Input the new feature vector X new =[x new1 x new2 , ..., x newn ];

[0023] S6.2 Predict the slump of the new concrete mix using a trained prediction model;

[0024] S6.3 Predict the strength of cement based on the predicted slump value.

[0025] As a further improvement to this technical solution, in S6.2, the concrete slump predicted by the new mix proportion is specifically as follows:

[0026]

[0027] in, This represents the predicted value of the concrete slump based on the new mix design.

[0028] As a further improvement to this technical solution, in S6.3, the prediction of cement strength is specifically as follows:

[0029]

[0030] in, The predicted cement strength values ​​are represented by: γ0 representing the intercept term; and γ1 representing the coefficient corresponding to the slump.

[0031] As a further improvement to this technical solution, the gradient descent method in the optimization update unit is specifically as follows:

[0032]

[0033] Where, β (j) This represents the j-th parameter; α represents the learning rate; The loss function L represents the parameter β. (j) The gradient.

[0034] As a further improvement to this technical solution, the iterative optimization in the optimization and update unit specifically includes:

[0035]

[0036] in, Indicates parameter β (j) The value prior to the current iteration; This represents the updated value.

[0037] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0038] 1. This cement strength prediction algorithm based on big data ensures data integrity by storing concrete mix proportion data in a database. The prediction model is trained and validated using a multivariate adaptive regression spline algorithm, enabling it to handle large volumes of high-dimensional data, making it particularly suitable for complex prediction problems.

[0039] 2. This big data-based cement strength prediction algorithm uses the proposed concrete mix design information and a trained model to calculate the slump prediction value, thereby predicting the cement strength. This significantly shortens the testing cycle, reduces costs, and improves prediction accuracy. The gradient descent method is used, and the concrete mix design is iterated based on the prediction results to make the prediction model more accurate, thus improving the reliability of the prediction results. Attached Figure Description

[0040] Figure 1 This is an overall flowchart of the present invention;

[0041] The meanings of the labels in the diagram are as follows:

[0042] 1. Data acquisition unit; 2. Model processing unit; 3. Predictive analysis unit; 4. Optimization and update unit. Detailed Implementation

[0043] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0044] Example

[0045] Please see Figure 1 As shown, a cement strength prediction algorithm based on big data is provided, including a data acquisition unit 1, which is used to collect concrete mix proportion data from historical engineering projects and establish a database.

[0046] In this example, the data acquisition unit 1 includes a database module, which is used to organize and store the collected concrete mix proportion data.

[0047] The cement strength prediction algorithm based on big data also includes a model processing unit 2, which trains and validates the prediction model based on a multivariate adaptive regression spline algorithm.

[0048] Multivariate Adaptive Regression Splines (MARS) is a nonlinear regression method used for modeling and prediction. Based on the concept of spline functions, MARS can adaptively regress nonlinear relationships in data. MARS can handle multiple independent variables (features), enabling it to cope with complex datasets containing multiple input variables. The basic idea of ​​MARS is to construct a series of basis functions to approximate the nonlinear relationships of the input features, and then combine these basis functions to form the final model.

[0049] In this example, the model processing unit 2 establishes a prediction model based on the multivariate adaptive regression spline algorithm, which provides a solid foundation for the prediction analysis unit 3. The specific steps for training and validating the prediction model are as follows:

[0050] S3.1 Clean and standardize the collected data;

[0051] S3.2 Training the prediction model using a linear regression algorithm to enable the prediction model to accurately predict concrete slump;

[0052] In this example, the linear regression algorithm in S3.2 is specifically as follows:

[0053] X = [x1, x2, ..., x] n ];

[0054] y i =β0 + β1x1 + β2x2 + ... + β n x n ;

[0055] Where X represents the input feature matrix; n represents n features; y i Represents the true value of concrete slump; β0 represents the intercept term; β n This indicates that it corresponds to the nth feature x.

[0056] S3.3 Input the reserved test dataset into the trained model and evaluate the model using mean squared error.

[0057] Mean Square Error (MSE) is a statistical measure of the difference between predicted and actual values. It is calculated as the average of the sum of the squares of the differences between each predicted value and its corresponding actual value. In statistics, MSE is used to assess the degree of difference between the estimated quantity and the estimated quantity.

[0058] In this example, the mean square error in S3.3 is specifically as follows:

[0059]

[0060] Where L(β) represents the mean squared error function; m represents the number of samples; This represents the predicted value of concrete slump.

[0061] The cement strength prediction algorithm based on big data also includes a prediction analysis unit 3. The prediction analysis unit 3 calculates the slump prediction value based on the concrete mix information to be used and the trained model, and then predicts the strength of the cement.

[0062] In this example, the prediction and analysis unit 3 is used to predict the slump of the new concrete mix and the strength of the cement. The specific steps of the prediction and analysis are as follows:

[0063] S6.1 Input the new feature vector X new =[x new1 x new2 , ..., x newn ];

[0064] S6.2 Predict the slump of the new concrete mix using a trained prediction model;

[0065] In this example, the concrete slump predicted by the new mix design in S6.2 is specifically as follows:

[0066]

[0067] in, This represents the predicted value of the concrete slump based on the new mix design.

[0068] S6.3 Predict the strength of cement based on the predicted slump value.

[0069] In this example, in S6.3, predicting the strength of the cement specifically involves:

[0070]

[0071] in, γ represents the predicted cement strength value; γ0 represents the intercept term; γ1 represents the coefficient corresponding to the slump.

[0072] The cement strength prediction algorithm based on big data also includes an optimization and update unit 4, which iteratively optimizes the concrete mix ratio based on the gradient descent method and the prediction results.

[0073] Gradient descent is an optimization algorithm widely used in machine learning and deep learning to find model parameters that minimize the loss function. It is an iterative method that gradually approaches the minimum point of the loss function by continuously updating the parameter vector.

[0074] In this example, the gradient descent method in the optimization update unit 4 is specifically as follows:

[0075]

[0076] Where, β (j) This represents the j-th parameter; α represents the learning rate; The loss function L represents the parameter β. (j) The gradient.

[0077] In this example, the iterative optimization in the optimization update unit 4 specifically involves:

[0078]

[0079] in, Indicates parameter β (j) The value prior to the current iteration; This represents the updated value.

[0080] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A cement strength prediction algorithm based on big data, characterized in that: include Data acquisition unit (1), the data acquisition unit (1) is used to collect concrete mix proportion data in historical engineering projects and establish a database; Model processing unit (2), which trains and validates the prediction model based on the multivariate adaptive regression spline algorithm; The prediction analysis unit (3) calculates the slump prediction value based on the concrete mix proportion information to be used and the trained model, and then predicts the strength of the cement. The optimization and update unit (4) optimizes the concrete mix ratio iteratively based on the gradient descent method and the prediction results.

2. The cement strength prediction algorithm based on big data according to claim 1, characterized in that: The data acquisition unit (1) includes a database module, which is used to organize and store the collected concrete mix proportion data.

3. The cement strength prediction algorithm based on big data according to claim 1, characterized in that: The model processing unit (2) establishes a prediction model based on the multivariate adaptive regression spline algorithm, which provides a solid foundation for the prediction analysis unit (3). The specific steps for training and validating the prediction model are as follows: S3.1 Clean and standardize the collected data; S3.2 Training the prediction model using a linear regression algorithm to enable the prediction model to accurately predict concrete slump; S3.3 Input the reserved test dataset into the trained model and evaluate the model using mean squared error.

4. The cement strength prediction algorithm based on big data according to claim 1, characterized in that: In S3.2, the linear regression algorithm is specifically as follows: X=[x1,x2,...,x n ]; y i =β0+β1x1+β2x2+...+β n x n ; Where X represents the input feature matrix; n represents n features; y i Represents the true value of concrete slump; β0 represents the intercept term; β n This indicates that it corresponds to the nth feature x.

5. The cement strength prediction algorithm based on big data according to claim 1, characterized in that: In S3.3, the mean square error is specifically as follows: Where L(β) represents the mean squared error function; m represents the number of samples; This represents the predicted value of concrete slump.

6. The cement strength prediction algorithm based on big data according to claim 1, characterized in that: The prediction and analysis unit (3) is used to predict the slump of the new concrete mix and the strength of the cement. The specific steps of the prediction and analysis are as follows: S6.1 Input the new feature vector X new =[x new1 x new2 , ..., x newn ]; S6.2 Predict the slump of the new concrete mix using a trained prediction model; S6.3 Predict the strength of cement based on the predicted slump value.

7. The cement strength prediction algorithm based on big data according to claim 1, characterized in that: In S6.2, the concrete slump predicted by the new mix proportion is specifically as follows: in, This represents the predicted value of the concrete slump based on the new mix design.

8. The cement strength prediction algorithm based on big data according to claim 1, characterized in that: In S6.3, the predicted strength of cement is specifically as follows: in, γ represents the predicted cement strength value; γ0 represents the intercept term; γ1 represents the coefficient corresponding to the slump.

9. The cement strength prediction algorithm based on big data according to claim 1, characterized in that: The gradient descent method in the optimization and update unit (4) is specifically as follows: Where, β (j) This represents the j-th parameter; α represents the learning rate; The loss function L represents the parameter β. (j) The gradient.

10. The cement strength prediction algorithm based on big data according to claim 1, characterized in that: The iterative optimization in the optimization and update unit (4) specifically includes: in, Indicates parameter β (j) The value prior to the current iteration; This represents the updated value.

Images