Concrete temperature field reconstruction method, system, electronic device and storage medium

Abstract

The present invention belongs to the field of data processing and relates to a method, system, electronic device and storage medium for reconstructing a concrete temperature field, comprising: S1, storing the acquired temperature and coordinates of the temperature measuring points in a temperature measuring point database; S2, dividing the temperature measuring point database into a training database and a test database; S3, obtaining an LSTM model output result through an LSTM model; S4, obtaining t j The interpolated temperature of the temperature measurement points corresponding to all temperature measurement points in the test database at all times; S5, obtain t through the DNN model j The temperature data training results of the temperature measurement points in the test database at all times; S6, based on t j The temperature data training results of the temperature measurement points in the database are constantly tested to determine whether the iteration stop condition has been met. If the iteration stop condition has not been met, S3 to S5 are repeated; if the stop condition is met, the iteration ends and the temperature field reconstruction is completed. This solves the problem of low accuracy in concrete temperature field reconstruction.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and specifically discloses a concrete temperature field reconstruction method, system, electronic equipment and storage medium. Background Art

[0002] With the advancement of sensor technology, the precision of sensors used to monitor temperature in hydropower projects is increasing, and their number and variety are also increasing. In particular, the application of distributed fiber optic technology has transformed traditional point-based temperature acquisition into linear sensing, greatly increasing the fineness and density of temperature data. Consequently, acquiring or reconstructing the true temperature field of the monitored object has become a current research hotspot in hydraulic engineering.

[0003] Currently, spatial interpolation methods are primarily used to construct the true temperature field of concrete based on monitored temperatures. As is well known, temperature is a physical quantity that characterizes the degree of hotness or coldness of an object and exhibits correlations in both spatial and temporal distributions. When interpolating the global temperature field based on a limited number of temperature monitoring points, only considering the spatial correlation of temperature while ignoring its temporal correlation leads to insufficient interpolation accuracy. Therefore, to improve the accuracy of temperature field interpolation, it is essential to develop spatiotemporal interpolation methods that integrate both temporal and spatial information of temperature data. In recent years, several researchers have introduced the temporal dimension based on spatial interpolation methods, proposing a series of spatiotemporal data interpolation methods, including spatiotemporal Kriging interpolation, spatiotemporal anti-weighted interpolation, and interpolation methods that account for both spatiotemporal and temporal heterogeneity. Among these, the spatiotemporal Kriging interpolation method is the most widely used. However, its variogram model requires consideration of both temporal and spatial dimensions, resulting in high complexity and the need for manual intervention and subjectivity in the selection of the variogram model, leading to low accuracy in the obtained temperature field. Furthermore, the interpolation accuracy of other spatiotemporal interpolation methods is also low, similarly resulting in inaccurate temperature fields. Therefore, conducting research on new methods for reconstructing the real spatiotemporal temperature field is of great significance for overcoming the current problem of insufficient temperature field interpolation accuracy, obtaining the real spatiotemporal evolution law of the concrete temperature field, and improving the temperature control quality. Summary of the Invention

[0004] The object of the present invention is to provide a concrete temperature field reconstruction method, system, electronic device and storage medium to solve the problem of low accuracy in concrete temperature field reconstruction.

[0005] The specific scheme of the present invention is as follows:

[0006] A concrete temperature field reconstruction method comprises the following steps:

[0007] S1. storing the acquired temperature and coordinates of the temperature measuring points in a temperature measuring point database;

[0008] S2. Split the temperature measurement point database into a training database and a test database;

[0009] S3, according to the training database t j-l ~t j The temperature of the temperature measurement point in the time period is obtained through the LSTM model output, where l is the length of the time period;

[0010] S4, based on the training database j The temperature of the temperature measuring point at the moment and the coordinates of the temperature measuring point in the test database are obtained by the OK space interpolation method. j Constantly test the interpolated temperature of all temperature measurement points in the database;

[0011] S5. Based on the output of the LSTM model and the interpolated temperature of the temperature measurement point, t is obtained through the DNN model. j Constantly test the temperature data training results of the temperature measurement points in the database;

[0012] S6, based on t j The temperature data training results of the temperature measurement points in the database are constantly tested to determine whether the iteration stop condition is met. If the iteration stop condition is not met, S3 to S5 are repeated; if the stop condition is met, the iteration is terminated to complete the temperature field reconstruction.

[0013] In some embodiments, if the iteration stop condition is not met, repeatedly performing S3 to S5 includes:

[0014] If the iteration stopping condition is not met, S3 to S5 are repeatedly executed using the Adam optimization method and the gradient descent method.

[0015] In some embodiments, the gradient descent method includes updating the LSTM model parameters based on the first-order moment estimation and the second-order moment estimation of the stochastic gradient, and the formulas for the first-order moment estimation and the second-order moment estimation are respectively:

[0016] ,

[0017] Among them, P k is the first-order moment estimate of the kth iteration; P k-1 is the first-order moment estimate of the k-1th iteration; v k is the second-order moment estimate of the kth iteration; v k-1 is the second-order moment estimate of the k-1th iteration; θ k is the LSTM model parameter of the kth iteration; θ k-1 is the LSTM model parameter of the k-1th iteration; is the gradient of the loss function at the kth iteration; ω1 is the exponential decay rate of the first-order moment estimate; ω2 is the exponential decay rate of the second-order moment estimate; α is the learning rate; ε is a constant.

[0018] In some embodiments, the iteration stopping condition is to set t j The temperature data training results of the temperature measurement points in the test database at the moment are the same as those in the test database at t j The error value calculated by comparing the temperature measurement point data information at the moment is less than the set threshold.

[0019] In some embodiments, the calculation of the error value includes: using the mean square error MSE as the loss function L(W) to calculate the error value, and the formula of the loss function L(W) is:

[0020] ,

[0021] Among them, W is the weight coefficient matrix; n is the number of samples; y i is the ith measured value; f(x i ) is the i-th predicted value.

[0022] In some embodiments, the iteration stopping condition is a preset maximum number of iterations.

[0023] In some embodiments, a pre-processing step is further included, wherein the pre-processing step comprises:

[0024] Normalize the temperatures and coordinates of abnormal and missing temperature measurement points.

[0025] The present invention also relates to a concrete temperature field reconstruction system, which is used in the above-mentioned concrete temperature field reconstruction method, comprising:

[0026] The acquisition module is used to obtain the temperature and coordinates of the temperature measurement points and store them in the temperature measurement point database;

[0027] A pre-processing module is used to normalize the temperature and coordinates of abnormal and missing temperature measurement points;

[0028] A database segmentation module is used to segment the temperature measurement point database into a training database and a test database;

[0029] LSTM-DNN network training module is used to train the j-l ~t j The temperature of the temperature measurement point in the time period is obtained through the LSTM model to obtain the LSTM model output result based on the t j The temperature of the temperature measuring point at time t and the coordinates of the temperature measuring point in the test database are obtained by the OK space interpolation method j The interpolated temperature of the temperature measurement points corresponding to all temperature measurement points in the database is tested at all times. The t is obtained through the DNN model based on the output results of the LSTM model and the interpolated temperature of the temperature measurement points. jThe temperature data training results of the temperature measurement points in the database are constantly tested to determine whether the LSTM-DNN network has reached the iteration stop condition. If the iteration stop condition has not been reached, the iterative training continues. If the stop condition has been reached, the iterative training ends and the temperature field reconstruction is completed.

[0030] The present invention also relates to an electronic device, comprising: a processor and a memory; the memory is used to store executable instructions of the processor, and the processor is configured to execute the above-mentioned concrete temperature field reconstruction method by executing the executable instructions.

[0031] The present invention also relates to a computer storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for reconstructing a concrete temperature field is implemented.

[0032] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0033] The present invention firstly calculates the j-l ~t j The temperature of the temperature measurement point in the time period is trained through the LSTM model to learn the time variation characteristics of the temperature. j The temperature of the temperature measuring point at time t and the coordinates of the temperature measuring point in the test database are obtained by the OK space interpolation method j The interpolated temperatures of the temperature measurement points corresponding to all temperature measurement points in the test database are used at all times to compensate for the spatial variation characteristics of temperature that disappear in the LSTM model. The temporal variation characteristics of temperature trained by the LSTM model and the spatial variation characteristics of temperature obtained by the spatial interpolation method are simultaneously trained by the DNN model to simultaneously train the temporal variation characteristics and spatial variation characteristics of temperature. Finally, the precise temperature of the temperature measurement points in the test database is reconstructed. By fully integrating the temporal variation characteristics and spatial variation characteristics of temperature, the problem of insufficient temperature field interpolation accuracy is solved, and the accuracy of temperature field reconstruction is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 This is a flow chart of a concrete temperature field reconstruction method in Example 1 of the present invention.

[0035] Figure 2 This is a flowchart of LSTM-DNN network training in Example 1 of the present invention.

[0036] Figure 3 This is a block diagram of a concrete temperature field reconstruction system in Example 1 of the present invention. DETAILED DESCRIPTION

[0037] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations.

[0038] Example 1

[0039] A concrete temperature field reconstruction method, such as Figure 1 As shown, the following steps are included:

[0040] S1. Obtaining data information of temperature measurement points and boundary coordinates of the temperature field reconstruction area;

[0041] The data information of the temperature measurement point and the boundary coordinates of the temperature field reconstruction area are obtained by monitoring the concrete. The data information of the temperature measurement point includes the temperature of the temperature measurement point and the coordinates of the temperature measurement point.

[0042] S2. After pre-processing the acquired data information of the temperature measurement points and the boundary coordinates of the temperature field reconstruction area, the data information of the temperature measurement points is stored in a temperature measurement point database.

[0043] After normalizing the temperatures of abnormal and missing temperature measurement points, save them to the temperature database;

[0044] After normalizing the coordinates of the temperature measurement points and the boundary coordinates of the temperature field reconstruction area, save them to the coordinate database;

[0045] After one-to-one correspondence between the temperature measurement point temperatures in the temperature database and the temperature measurement point coordinates in the coordinate database, the temperatures are stored in the temperature measurement point database T(x, y, z, t). The expression of the temperature measurement point database is:

[0046] ,

[0047] in, Indicates that at the mth point (x, y, z, t), t n The data information of the temperature measurement point at the moment.

[0048] S3. Split the temperature measurement point database into a training database and a test database.

[0049] For example, the temperature measurement point database is divided into two parts according to the ratio of 4:1, namely, the training database T tr Accounting for 80% of the temperature measurement point database, the test database T te It accounts for 20% of the measurement point database.

[0050] S4. Train the LSTM-DNN network based on the training database and the test database to complete the temperature field reconstruction.

[0051] The LSTM-DNN network consists of a series of long short-term memory recurrent neural networks (LSTMs) and deep neural networks (DNNs).

[0052] LSTM-DNN network training, such as Figure 2 As shown, including:

[0053] S41, according to the training database t j-l ~t j The temperature of the temperature measurement point in the time period is obtained through the LSTM model output result;

[0054] In the training database, j-l ~t j The temperature of the temperature measurement point in the time period is input into the LSTM model, where l is the length of the time period. After the LSTM model is trained on the time variation characteristics of the temperature field, the LSTM model output results are output.

[0055] S42, based on the training database j The temperature of the temperature measuring point at the moment and the coordinates of the temperature measuring point in the test database are obtained by the OK space interpolation method. j Constantly test the interpolated temperature of all temperature measurement points in the database;

[0056] Based on the training database j The temperature of the temperature measuring point at the moment, the temperature measurement point coordinates in the test database are used to predict the temperature using the OK space interpolation method to obtain t j The interpolated temperature of the temperature measurement points corresponding to all temperature measurement points in the test database is tested at all times.

[0057] S43, based on the LSTM model output results and the temperature measurement point interpolation temperature, obtain t through the DNN model j Constantly test the temperature data training results of the temperature measurement points in the database;

[0058] The output results of the LSTM model and the interpolated temperatures of the temperature measurement points corresponding to all temperature measurement points in the test database are input into the DNN model. The DNN model simultaneously trains the temporal variation characteristics and spatial variation characteristics of the temperature measurement points, and fully integrates the temporal variation characteristics and spatial variation characteristics of the temperature.

[0059] S44, based on t jThe temperature data training results of the temperature measurement points in the database are constantly tested to determine whether the iteration stop condition is met. If the iteration stop condition is not met, S41 to S43 are repeatedly executed; if the stop condition is met, the iteration is terminated to complete the temperature field reconstruction.

[0060] t j The temperature data training results of the temperature measurement points in the test database at the moment are the same as those in the test database at t j The temperature measurement point data information at each moment is compared to calculate the error value. If the error value is greater than or equal to the set threshold, the weight and bias matrix are continuously updated through the Adam optimization method and the gradient descent method, and S41 to S43 are repeated. If the error value is less than the set threshold, the iteration is terminated to reconstruct the accurate temperature of the temperature measurement point in the test database and complete the temperature field reconstruction.

[0061] The mean square error (MSE) is used as the loss function L(W) to calculate the error value. The formula of the loss function L(W) is:

[0062] ,

[0063] Among them, W is the weight coefficient matrix; n is the number of samples; y i is the ith measured value; f(x i ) is the i-th predicted value.

[0064] Or if the number of iterations does not reach the preset maximum number of iterations, S41 to S43 are repeatedly executed through the Adam optimization method and the gradient descent method; if the number of iterations reaches the preset maximum number of iterations, the iteration is terminated, the precise temperature of the temperature measurement point in the test database is reconstructed, and the temperature field reconstruction is completed.

[0065] The gradient descent method includes updating the LSTM model parameters based on the first-order moment estimation and the second-order moment estimation of the stochastic gradient. The formulas for the first-order moment estimation and the second-order moment estimation are:

[0066] ,

[0067] Among them, P k is the first-order moment estimate of the kth iteration; P k-1 is the first-order moment estimate of the k-1th iteration; v k is the second-order moment estimate of the kth iteration; v k-1 is the second-order moment estimate of the k-1th iteration; θ k is the LSTM model parameter of the kth iteration; θ k-1 is the LSTM model parameter of the k-1th iteration; is the gradient of the loss function at the kth iteration; ω1 is the exponential decay rate of the first-order moment estimate; ω2 is the exponential decay rate of the second-order moment estimate; α is the learning rate; ε is a small constant.

[0068] The present invention firstly calculates the j-l ~t j The temperature of the temperature measurement point in the time period is trained through the LSTM model to learn the time variation characteristics of the temperature. j The temperature of the temperature measuring point at time t and the coordinates of the temperature measuring point in the test database are obtained by the OK space interpolation method j The interpolated temperatures of the temperature measurement points corresponding to all temperature measurement points in the test database are used at all times to compensate for the spatial variation characteristics of temperature that disappear in the LSTM model. The temporal variation characteristics of temperature trained by the LSTM model and the spatial variation characteristics of temperature obtained by the spatial interpolation method are simultaneously trained by the DNN model to simultaneously train the temporal variation characteristics and spatial variation characteristics of temperature. Finally, the precise temperature of the temperature measurement points in the test database is reconstructed. By fully integrating the temporal variation characteristics and spatial variation characteristics of temperature, the problem of insufficient temperature field interpolation accuracy is solved, and the accuracy of temperature field reconstruction is improved.

[0069] The present invention also relates to a concrete temperature field reconstruction system, which is used in the above-mentioned concrete temperature field reconstruction method. Figure 3 As shown, including:

[0070] The acquisition module is used to obtain the temperature and coordinates of the temperature measurement points and store them in the temperature measurement point database;

[0071] A pre-processing module is used to normalize the temperature and coordinates of abnormal and missing temperature measurement points;

[0072] A database segmentation module is used to segment the temperature measurement point database into a training database and a test database;

[0073] LSTM-DNN network training module is used to train the j-l ~t j The temperature of the temperature measurement point in the time period is obtained through the LSTM model to obtain the LSTM model output result based on the t j The temperature of the temperature measuring point at time t and the coordinates of the temperature measuring point in the test database are obtained by the OK space interpolation method j The interpolated temperature of the temperature measurement points corresponding to all temperature measurement points in the database is tested at all times. The t is obtained through the DNN model based on the output results of the LSTM model and the interpolated temperature of the temperature measurement points. jThe temperature data training results of the temperature measurement points in the database are constantly tested to determine whether the LSTM-DNN network has reached the iteration stop condition. If the iteration stop condition has not been reached, the iterative training continues. If the stop condition has been reached, the iterative training ends and the temperature field reconstruction is completed.

[0074] The present invention also relates to an electronic device, comprising: a processor and a memory; the memory is used to store executable instructions of the processor, and the processor is configured to execute the above-mentioned concrete temperature field reconstruction method by executing the executable instructions.

[0075] The present invention also relates to a computer storage medium having a computer program stored thereon. When the computer program is executed by a processor, the above-mentioned method for reconstructing a concrete temperature field is implemented.

[0076] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A method for reconstructing a concrete temperature field, characterized in that: The following steps are involved: S1. storing the acquired temperature and coordinates of the temperature measuring points in a temperature measuring point database; S2. Split the temperature measurement point database into a training database and a test database; S3, according to the training database t j-l ~t j The temperature of the temperature measurement point in the time period is obtained through the LSTM model output, where l is the length of the time period; S4, based on the training database j The temperature of the temperature measuring point at the moment and the coordinates of the temperature measuring point in the test database are obtained by the OK space interpolation method. j Constantly test the interpolated temperature of all temperature measurement points in the database; S5. Based on the LSTM model output and the temperature measurement point interpolation temperature, t is obtained through the DNN model. j Constantly test the temperature data training results of the temperature measurement points in the database; S6, based on t j Constantly test the temperature data training results of the temperature measurement points in the database to determine whether the iteration stop condition is met. If the iteration stop condition is not met, repeat S3 to S5; If the stopping condition is reached, the iteration ends and the temperature field reconstruction is completed.

2. A concrete temperature field reconstruction method according to claim 1, characterized in that: If the iteration stop condition is not met, then S3 to S5 are repeatedly executed, including: If the iteration stopping condition is not reached, S3 to S5 are repeatedly executed by the gradient descent method.

3. A concrete temperature field reconstruction method according to claim 2, characterized in that: The gradient descent method includes updating the LSTM model parameters based on the first-order moment estimation and the second-order moment estimation of the stochastic gradient, and the formulas of the first-order moment estimation and the second-order moment estimation are respectively: ， Among them, P k is the first-order moment estimate of the kth iteration; P k-1 is the first-order moment estimate of the k-1th iteration; v k is the second-order moment estimate of the kth iteration; v k-1 is the second-order moment estimate of the k-1th iteration; θ k is the LSTM model parameter of the kth iteration; θ k-1 is the LSTM model parameter of the k-1th iteration; is the gradient of the loss function at the kth iteration; ω1 is the exponential decay rate of the first-order moment estimate; ω2 is the exponential decay rate of the second-order moment estimate; α is the learning rate; ε is a constant.

4. The method for reconstructing a concrete temperature field according to claim 1, wherein: The iteration stopping condition is to set t j The temperature data training results of the temperature measurement points in the test database at the moment are the same as those in the test database at t j The error value calculated by comparing the temperature measurement point data information at the moment is less than the set threshold.

5. A concrete temperature field reconstruction method according to claim 4, characterized in that: The calculation of the error value includes: using the mean square error MSE as the loss function L(W) to calculate the error value, and the formula of the loss function L(W) is: ， Among them, W is the weight coefficient matrix; n is the number of samples; y i is the ith measured value; f(x i ) is the i-th predicted value.

6. The method for reconstructing a concrete temperature field according to claim 1, characterized in that: The iteration stopping condition is the maximum value of the preset number of iterations.

7. A concrete temperature field reconstruction system, characterized in that: A concrete temperature field reconstruction method for use in any one of claims 1 to 6, comprising: The acquisition module is used to obtain the temperature and coordinates of the temperature measurement points and store them in the temperature measurement point database; A database segmentation module is used to segment the temperature measurement point database into a training database and a test database; LSTM-DNN network training module is used to train the j-l ~t j The temperature of the temperature measurement point in the time period is obtained through the LSTM model to obtain the LSTM model output result based on the t j The temperature of the temperature measuring point at time t and the coordinates of the temperature measuring point in the test database are obtained by the OK space interpolation method j The interpolated temperature of the temperature measurement points corresponding to all temperature measurement points in the database is tested at all times. The t is obtained through the DNN model based on the output results of the LSTM model and the interpolated temperature of the temperature measurement points. j The temperature data training results of the temperature measurement points in the database are constantly tested to determine whether the LSTM-DNN network has reached the iteration stop condition. If the iteration stop condition has not been reached, the iterative training continues. If the stop condition has been reached, the iterative training ends and the temperature field reconstruction is completed.

8. An electronic device, characterized in that: include: processor and memory; The memory is used to store executable instructions of the processor, and the processor is configured to perform a concrete temperature field reconstruction method according to any one of claims 1 to 6 by executing the executable instructions.

9. A computer storage medium, characterized in that: The computer storage medium stores a computer program, and when the computer program is executed by the processor, the method for reconstructing a concrete temperature field according to any one of claims 1 to 6 is implemented.

Images