TBM monthly footage estimation method and device based on two-stage data enhancement
Through a two-stage data augmentation method and machine learning model, the accuracy problem of TBM monthly footage estimation was solved, accurate estimation of the entire tunnel/section was achieved, and the prediction accuracy and robustness of the model were improved.
Patent Information
- Application Number
- CN202510869772.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-26
- Publication Date
- 2025-10-17
AI Technical Summary
Existing technologies are unable to accurately estimate the average monthly advance of a hard rock tunnel boring machine (TBM) throughout a tunnel or section, especially under complex and variable geological conditions, as influencing factors are not fully considered.
A two-stage data enhancement method is used to construct a geological label prediction model. Reliable pseudo labels are generated through Monte Carlo Dropout and pseudo labeling technology. The training data set is enhanced by combining random oversampling and SMOTE-NC algorithms, and monthly footage estimation is performed using machine learning models such as KAN.
The model's prediction accuracy and robustness have been improved, enabling accurate estimation of the average monthly advances of TBMs under complex geological conditions, taking into account key factors such as TBM type, excavation diameter, tunnel depth, and surrounding rock grade.
Smart Images
Figure CN120804697A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to a data enhancement and data-driven TBM monthly footage prediction method, belonging to the technical field of hard rock TBM and machine learning combination. BACKGROUND
[0002] Hard rock tunnel boring machine (TBM) method has the advantages of fast excavation speed, high construction quality, small environmental disturbance, etc. compared with the traditional drilling and blasting method, so it is more and more applied in major tunnel (cave) / roadway projects such as transportation, water conservancy, energy, and mining. However, the performance of TBM is greatly affected by the stratum (when the stratum condition is good, the excavation speed is fast; when the stratum condition is poor, construction disasters may occur frequently), and the rock-machine interaction is complex, which brings challenges to accurately predicting the TBM monthly footage. Before the tunnel (cave) construction, accurately predicting the average monthly footage of TBM is extremely important for method selection, TBM selection, construction period and cost estimation, and engineering planning, etc.
[0003] Currently, the academic field mainly focuses on the construction progress prediction during the TBM excavation process in order to optimize the TBM excavation control parameters or predict the construction progress of TBM under specific geological conditions (lithology, UCS, RMR, etc.). The engineering field relies on standards, specifications, etc. to roughly estimate the monthly footage under different surrounding rock grades, without considering the influence of key factors such as excavation diameter, buried depth, TBM type, etc. Whether in the academic field or the engineering field, there is no method for predicting the average monthly footage of TBM for the whole tunnel (cave) / section. SUMMARY
[0004] In order to solve the problem that the current TBM monthly footage prediction method cannot accurately predict the average monthly footage of the whole tunnel (cave) / section before construction, the present application provides a TBM monthly footage prediction method based on two-stage data enhancement and a device. Based on the extensive collection of TBM construction cases at home and abroad, the method establishes a data-driven TBM monthly footage prediction model, and the input of the model considers key influencing factors such as TBM type, excavation diameter, tunnel (cave) buried depth, surrounding rock grade proportion, etc., which can accurately predict the average monthly footage of TBM for the whole tunnel (cave) / section.
[0005] In order to effectively solve the above problems, the technical solutions provided by the present application are described in detail as follows: In a first aspect, a TBM monthly footage prediction method based on two-stage data enhancement is provided, comprising: obtaining an original data set; the original data set includes TBM type, proportion of each level of surrounding rock along the tunnel, average buried depth of the tunnel, excavation diameter, purpose of the tunnel, and average monthly footage; The original data set is divided into a complete data set and a data set without a geological label, wherein the complete data set is divided into a first training set and a test set; and the data set without a geological label is a data set in which the proportion of surrounding rock grades is missing; A geological label prediction model is constructed and trained; The trained geological label prediction model is used to add pseudo-geological labels to the case without a geological label, and the pseudo-labeled data is added to the first training set to obtain a second training set; The second training set is subjected to data enhancement to obtain a third training set; A machine learning-based monthly footage prediction model is trained based on the third training set, and the optimal prediction model is screened out by using a verification set; The optimal prediction model is used to predict the monthly footage of the TBM.
[0006] In a possible implementation manner, the proportion of each level of surrounding rock along the tunnel is classified into I, II, III, IV and V according to the BQ or HC classification standard commonly used in China, and the sum of the proportions of each level of surrounding rock is 100%; the average monthly footage is obtained by dividing the TBM excavation distance by the TBM construction time, wherein the construction time is from the start of TBM excavation to the end of TBM breakthrough.
[0007] In a possible implementation manner, the geological label prediction model comprises an input layer, a full connection layer, a neural network layer with Monte Carlo Dropout, and an output layer; wherein the activation function of the full connection layer is ReLU, and the activation function of the output layer is softmax.
[0008] Further, the training method of the geological label prediction model comprises: The geological label prediction model is preliminarily trained based on the first training set to obtain a pre-trained prediction model; The data set without a geological label is input into the pre-trained prediction model, the Monte Carlo Dropout method is used to repeatedly predict the data N times, so that each sample without a geological label obtains N sets of repeated prediction results; the mean and the standard deviation of the N sets of repeated prediction results are calculated, and the standard deviation is used to measure the reliability of the prediction results; The prediction results are sorted from high to low according to the reliability, a reliability threshold is set, and reliable samples are screened out; meanwhile, the mean of the prediction results is used as the pseudo-label of the sample without a geological label to enhance the reliability of the prediction results; The reliable pseudo-labeled samples are added to the first training set, and the pre-trained prediction model is retrained; The above steps are repeated until the number of iterations is reached or no new reliable pseudo-labeled samples are added, to obtain a second training set containing pseudo-labeled samples and a trained geological label prediction model.
[0009] In a possible implementation, the data enhancement method comprises random oversampling and an SMOTE-NC algorithm.
[0010] In a possible implementation, the monthly footage estimation model comprises a classification and regression tree, a k-nearest neighbor, a support vector regression, a random forest, an XGBoost, an AdaBoost, a LightGBM, a DNN, and a KAN.
[0011] In a second aspect, a TBM monthly footage estimation device based on two-stage data enhancement is provided, comprising: An acquisition module is configured to acquire an original data set, wherein the original data set comprises a TBM type, a proportion of each level of surrounding rock along a tunnel, an average buried depth of the tunnel, an excavation diameter, a tunnel purpose, and an average monthly footage. A division module is configured to divide the original data set into a complete data set and a data set without a geological label, wherein the complete data set is divided into a first training set and a test set, and the data set without a geological label is a data set in which a proportion of surrounding rock grade is missing. A framework module is configured to build and train a geological label prediction model. A first enhancement module is configured to add a pseudo-geological label to a case without a geological label by using the trained geological label prediction model, and add the pseudo-labeled data to the first training set to obtain a second training set. A second enhancement module is configured to perform data enhancement on the second training set to obtain a third training set. A screening module is configured to train a monthly footage estimation model based on machine learning based on the third training set, and screen an optimal estimation model by using a validation set. An output module is configured to perform TBM monthly footage estimation by using the optimal estimation model.
[0012] In a third aspect, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the TBM monthly footage estimation method according to the first aspect when executing the program.
[0013] In a fourth aspect, a non-transitory computer-readable storage medium is provided, which stores a computer program, wherein the computer program is executed by a processor to implement the TBM monthly footage estimation method according to the first aspect.
[0014] In a fifth aspect, a computer program product is provided, comprising a computer program, wherein the computer program is executed by a processor to implement the TBM monthly footage estimation method according to the first aspect.
[0015] Compared with the prior art, the present application has the following beneficial effects: The method is not limited to one or several specific geological types, and can realize the prediction of the average monthly tunneling length of the whole tunnel (hole) TBM under complex and variable geological conditions. The method fully considers the key factors such as TBM type, excavation diameter, tunnel (hole) burial depth, and proportion of each level of surrounding rock grade in model input, has high accuracy, and is closer to the actual situation. The method adopts two-stage data enhancement, so that the prediction accuracy and robustness of the model are improved. BRIEF DESCRIPTION OF DRAWINGS
[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor on the basis of these drawings.
[0017] Figure 1 The flowchart of the TBM monthly progress prediction method based on two-stage data enhancement of the present application; Figure 2 The performance of the machine learning model in the embodiment of the present application on the test set in different data enhancement stages.
[0018] Figure 3 The performance comparison of the AdaBoost model, the deep neural network (DNN) model and the KAN model in the embodiment of the present application on the test set.
[0019] Figure 4 The structural schematic diagram of the TBM monthly progress prediction device based on two-stage data enhancement and KAN neural network provided by the embodiment of the present application; Figure 5 The structural schematic diagram of the electronic device provided by the embodiment of the present application. DETAILED DESCRIPTION
[0020] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.
[0021] REFERENCE Figure 1 A TBM monthly progress prediction method based on two-stage data enhancement, comprising the following steps: S100, obtaining an original data set; the original data set comprises a TBM type, a proportion of each level of surrounding rock along a tunnel, an average buried depth of the tunnel, an excavation diameter, a tunnel purpose, and an average monthly footage.
[0022] In a possible implementation manner, the proportion of each level of surrounding rock along the tunnel is classified into Class I, Class II, Class III, Class IV, and Class V according to a BQ (GB / T 50218-2014) or HC (GB 28376-2012 Tunnel Fire Protection Board) classification standard commonly used in China, and the sum of the proportions of each level of surrounding rock is 100%.
[0023] In a possible implementation manner, the average buried depth of the tunnel is obtained by a tunnel longitudinal section drawing estimation.
[0024] In a possible implementation manner, the average monthly footage is obtained by dividing a TBM excavation distance by a TBM construction time, wherein the construction time is from the start of TBM excavation to the end of TBM breakthrough of a section.
[0025] In a possible implementation manner, the tunnel purpose is a hydraulic tunnel and a non-hydraulic tunnel, the former generally adopts an HC classification system, and the latter generally adopts a BQ classification system.
[0026] S200, dividing the original data set into a complete data set and a data set without a geological label, wherein the complete data set is divided into a first training set and a test set; and the data set without a geological label is a data set in which the proportion of each level of surrounding rock is missing.
[0027] S300, constructing and training a geological label prediction model.
[0028] In a possible implementation manner, the TBM type comprises an open type, a single-shield TBM, and a double-shield TBM.
[0029] In a possible implementation manner, the geological label prediction model comprises, in sequence, an input layer, a full connection layer, a neural network layer with Monte Carlo Dropout, and an output layer; wherein an activation function of the full connection layer is ReLU, and an activation function of the output layer is softmax.
[0030] The prediction model performs MC-Dropout inference through the neural network with Monte Carlo Dropout, the output of the model is the proportion of each level of surrounding rock, and the activation function of the output layer is “softmax” to ensure that the sum of the proportions of each level of surrounding rock is 100%.
[0031] In a possible implementation manner, a training method of the geological label prediction model comprises: S301, preliminarily training the geological label prediction model based on the first training set to obtain a pre-training prediction model. S302, input the geological label-free dataset into the pre-trained prediction model, use the Monte Carlo Dropout method to repeatedly predict the data N times (the Dropout layer is in the active state), so that each geological label-free sample obtains N sets of repeated prediction results; calculate the mean and standard deviation of the N sets of repeated prediction results to measure the reliability of the prediction results (the smaller the standard deviation, the more reliable the prediction results); S303, sort the reliability of the prediction results from large to small, set a reliability threshold, and select reliable samples; at the same time, use the mean of the prediction results as the pseudo-label of the geological label-free sample to enhance the reliability of the prediction results; S304, add reliable pseudo-labeled samples to the first training set and retrain the pre-trained prediction model; S305, repeat the above steps until the iteration number is reached or no new reliable pseudo-labeled samples are added, and obtain the second training set containing pseudo-labeled samples and the trained geological label prediction model.
[0032] Further, the determination method of the reliability threshold is a fixed value (such as 0.03), or the standard deviation of the first M% (such as the first 20%) of all predicted pseudo-labels (sorted from small to large).
[0033] S400, use the trained geological label prediction model to add pseudo-geological labels to the geological label-free cases, and add the pseudo-labeled data to the first training set to obtain the second training set.
[0034] S500, perform data augmentation on the second training set to obtain a third training set.
[0035] In a possible implementation manner, the data augmentation includes random oversampling and SMOTE-NC algorithm.
[0036] S600, train a machine learning-based month footage prediction model based on the third training set, and use a validation set to screen an optimal prediction model.
[0037] In a possible implementation manner, the month footage prediction model includes commonly used machine learning methods such as classification and regression tree (CART), k-nearest neighbor (KNN), support vector regression (SVR), random forest (RF), AdaBoost, XGBoost, LightGBM (LGBM), deep neural network DNN, and KAN.
[0038] In a possible implementation manner, the month footage prediction model uses a grid search method to find optimal hyperparameters during training.
[0039] S700, use the optimal prediction model to perform TBM month footage prediction.
[0040] In a possible implementation manner, the optimal estimation model is KAN.
[0041] The content of the present application will be further described below in combination with more specific embodiments.
[0042] The TBM monthly footage estimation method based on two-stage data enhancement includes the following steps: Step 1: Collect TBM tunnel (hole) case data at home and abroad, including TBM type (open type, single shield or double shield TBM), proportion of each level of surrounding rock along the tunnel (hole) (proportion of each level of surrounding rock under BQ / HC surrounding rock classification standard), average buried depth of tunnel (hole), excavation diameter, purpose of tunnel (hole), average monthly footage of TBM in the whole tunnel (hole) / section, to obtain an original data set.
[0043] Specifically, there are a total of 275 cases in the original data set, of which 177 cases come from China.
[0044] Step 2: The original data set is further divided into a complete data set and a data set without geological labels (the proportion of surrounding rock grades is missing or foreign tunnels (holes) do not use BQ / HC surrounding rock classification standard).
[0045] Specifically, there are 121 complete cases, and the rest are incomplete cases, wherein the complete data set is divided into a first training set (training set 0, 106 cases) and a test set (15 cases).
[0046] Step 3: Construct and train a geological label prediction model.
[0047] Specifically, the geological label prediction model iteratively adds pseudo-geological labels to cases without geological labels. The number of iterations is set to 30, the threshold value is set to 0.025, and the threshold value is multiplied by a decay factor of 0.9 after each iteration. Reliable pseudo-label data is added to the training set 0 to participate in the training of the geological label (proportion of each level of surrounding rock) prediction model, and a preliminary enhanced training set 1 (the number of cases reaches 227) is obtained.
[0048] Step 4, use the trained geological label prediction model to add pseudo-geological labels to cases without geological labels, and add pseudo-label data to the first training set (training set 0) to obtain a second training set (training set 1); Step 5: Enhance the training set 1 by using random oversampling and SMOTE-NC algorithm in turn to obtain a training set 2 (the number of cases reaches 704).
[0049] It should be noted that the purpose of performing random oversampling operation before SMOTE-NC oversampling is to ensure that there are enough neighbor interpolation samples during SMOTE-NC oversampling.
[0050] Step 6: Based on the training set 2, the TBM monthly footage prediction model is respectively established based on the machine learning model, and the performance comparison is carried out on the test set, and the optimal prediction model is screened out as the KAN neural network model.
[0051] In a possible implementation manner, the machine learning model comprises a classification and regression tree (CART), a k-nearest neighbor (KNN), a support vector regression (SVR), a random forest (RF), an XGBoost (XGB), an AdaBoost (ADB) and a LightGBM (LGBM), a deep neural network (DNN) and a KAN (Kolmogorov-Arnold network).
[0052] In a possible implementation manner, the TBM monthly footage prediction model adopts a grid search method to find the optimal hyperparameter in the training process.
[0053] Step 7: The optimal prediction model is used for TBM monthly footage prediction.
[0054] Figure 2 The prediction effect (R2) of each machine learning model provided in the embodiment of the application on the test set in different data enhancement stages is shown in the following table. Figure 2 It can be seen that, after the enhancement of the two-stage data enhancement method, the average R2 of the machine learning model on the test set is improved from 0.62 to 0.72. Figure 3 After the two-stage data enhancement, the performance of the AdaBoost model, the DNN model and the KAN model on the test set is compared, and it can be seen that the KAN model performs best, and the R2 reaches 0.89.
[0055] The following describes the TBM monthly footage prediction device based on the two-stage data enhancement provided by the application, and the TBM monthly footage prediction device based on the two-stage data enhancement described in the following can be correspondingly referred to the TBM monthly footage prediction method based on the two-stage data enhancement described in the foregoing.
[0056] Figure 4 is a structural schematic diagram of the TBM monthly footage prediction device based on the two-stage data enhancement provided by the embodiment of the application, as shown in Figure 4 It includes an acquisition module 41, a division module 42, a framework module 43, a first enhancement module 44, a second enhancement module 45, a screening module 46 and an output module 47, wherein: The acquisition module 41 is used for acquiring an original data set; the original data set comprises a TBM type, a proportion of each level of surrounding rock along a tunnel, an average buried depth of the tunnel, an excavation diameter, a tunnel purpose and an average monthly footage; The dividing module 42 is configured to divide the original data set into a complete data set and a data set without a geological label, wherein the complete data set is divided into a first training set and a test set; and the data set without a geological label is a data set in which the proportion of surrounding rock grades is missing. The framework module 43 is configured to build and train a geological label prediction model. The first enhancement module 44 is configured to add pseudo-geological labels to the data set without a geological label by using the trained geological label prediction model, and add the pseudo-labeled data to the first training set to obtain a second training set. The second enhancement module 45 is configured to perform data enhancement on the second training set to obtain a third training set. The screening module 46 is configured to train a machine learning-based monthly footage prediction model based on the third training set, and screen an optimal prediction model by using a verification set. The output module 47 is configured to perform TBM monthly footage prediction by using the optimal prediction model.
[0057] Figure 5 An example of an entity structure diagram of an electronic device is shown in Figure 5 As shown, the electronic device can include a processor 510, a communications interface 520, a memory 530, and a communications bus 540, wherein the processor 510, the communications interface 520, and the memory 530 can communicate with each other through the communications bus 540. The processor 510 can invoke the logical instructions in the memory 530 to execute the two-stage data enhancement-based TBM monthly footage prediction method.
[0058] In addition, the logical instructions in the memory 530 described above can be implemented in the form of a software function unit and sold or used as an independent product, which can be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the present application or parts of the present application that essentially contribute to the prior art or the parts of the technical solutions can be embodied in the form of a software product, which is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the embodiments of the present application. The aforementioned storage medium includes a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.
[0059] In another aspect, the present application also provides a computer program product comprising a computer program, the computer program being stored in a non-transitory computer-readable storage medium, and the computer program being executable by a processor to enable a computer to perform the two-stage data augmentation based TBM monthly footage prediction method provided by the above methods.
[0060] In yet another aspect, the present application also provides a non-transitory computer-readable storage medium having stored thereon a computer program, the computer program being executable by a processor to implement the two-stage data augmentation based TBM monthly footage prediction method provided by the above methods.
[0061] The device embodiments described above are merely illustrative, wherein the units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, i.e., can be located in one place, or can be distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the present embodiment scheme according to actual needs. Those skilled in the art can understand and implement without creative labor.
[0062] From the above description of the embodiments, those skilled in the art can clearly understand that the embodiments can be realized by means of software plus necessary universal hardware platforms, and of course can also be realized by hardware. Based on such understanding, the above technical solutions, essentially or in other words, the part that contributes to the prior art, can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes a number of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods described in each embodiment or some parts of the embodiments.
[0063] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement to some technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.
Claims
1. A two-stage data enhancement-based TBM monthly footage estimation method is characterized by: include: Get the original dataset; The original data set includes TBM type, proportion of surrounding rock masses along the tunnel, average tunnel depth, excavation diameter, tunnel purpose, and average monthly footage. The original data set is divided into a complete data set and a data set without geological labels, wherein the complete data set is divided into a first training set and a test set; the data set without geological labels is a data set with missing data on the proportion of surrounding rock grades; Build and train a geological label prediction model; Using the trained geological label prediction model, pseudo geological labels are added to cases without geological labels, and the pseudo labeled data are added to the first training set to obtain the second training set. Perform data augmentation on the second training set to obtain the third training set; A machine learning-based monthly progress estimation model was trained based on the third training set, and the optimal estimation model was selected using the validation set. Use the optimal estimation model to estimate TBM monthly footage.
2. The method according to claim 1, characterized in that The proportion of surrounding rock masses of various levels along the tunnel is divided into Class I, II, III, IV and V according to the BQ or HC classification standards commonly adopted in China. The average monthly advance is obtained by dividing the TBM excavation distance by the TBM construction time, where the construction time is from the start of TBM excavation to the end of the TBM breakthrough section.
3. The method according to claim 1, characterized in that The geological label prediction model includes an input layer, a fully connected layer, a neural network layer with Monte Carlo Dropout, and an output layer; wherein the activation function of the fully connected layer is ReLU, and the activation function of the output layer is softmax.
4. The method according to claim 3, characterized in that The training method of the geological label prediction model includes: Preliminary training of the geological label prediction model based on the first training set to obtain a pre-trained prediction model; The dataset without geological labels is input into the pre-trained prediction model, and the Monte Carlo Dropout method is used to repeatedly predict the data N times, so that each sample without geological labels obtains N sets of repeated prediction results. The mean and standard deviation of the N sets of repeated prediction results are calculated, and the reliability of the prediction results is measured by the standard deviation. The reliability of the prediction results is sorted from large to small, and a reliability threshold is set to screen out reliable samples. At the same time, the mean of the prediction results is used as a pseudo-label for samples without geological labels to enhance the reliability of the prediction results. Add reliable pseudo-labeled samples to the first training set and retrain the pre-trained prediction model; Repeat the above steps until the number of iterations is reached or no new reliable pseudo-labeled samples are added, to obtain a second training set containing pseudo-labeled samples and a trained geological label prediction model.
5. The method according to claim 1, wherein The data enhancement method includes random oversampling and SMOTE-NC algorithm.
6. The method according to claim 1, wherein The monthly progress estimation models include classification and regression trees, k-nearest neighbors, support vector regression, random forest, XGBoost, AdaBoost, LightGBM, DNN and KAN.
7. The TBM monthly footage prediction device based on two-stage data enhancement is characterized by: include: Acquisition module, used to obtain the original data set; The original data set includes TBM type, proportion of surrounding rock masses along the tunnel, average tunnel depth, excavation diameter, tunnel purpose, and average monthly footage. A partitioning module is used to divide the original data set into a complete data set and a data set without geological labels, wherein the complete data set is divided into a first training set and a test set; the data set without geological labels is a data set with missing data on the proportion of surrounding rock grades; Framework module, used to build and train geological label prediction models; The first enhancement module is used to add pseudo geological labels to cases without geological labels using the trained geological label prediction model, and add the pseudo labeled data to the first training set to obtain a second training set; A second enhancement module is used to perform data enhancement on the second training set to obtain a third training set; A screening module is used to train a monthly footage estimation model based on machine learning based on the third training set, and to select the optimal estimation model using the validation set; The output module is used to estimate the monthly TBM footage using the optimal estimation model.
8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the TBM monthly progress estimation method based on two-stage data enhancement as described in any one of claims 1 to 6 is implemented.
9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for estimating monthly TBM progress based on two-stage data enhancement as described in any one of claims 1 to 6 is implemented.
10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the method for estimating monthly TBM progress based on two-stage data enhancement as described in any one of claims 1 to 6 is implemented.