Dam deformation prediction modeling method, prediction method, equipment and storage medium
By combining variational autoencoders and Gaussian mixture models with extended long short-term memory networks, dimensionality reduction and clustering of dam monitoring data are performed. This solves the problems of complexity in multidimensional data processing and insufficient utilization of temperature data in traditional methods, and achieves high accuracy and high reliability in dam deformation prediction.
Patent Information
- Application Number
- CN202511616691.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-06
- Publication Date
- 2026-02-03
AI Technical Summary
Traditional methods struggle to effectively capture the deep nonlinear relationship between long-term structural response and monitoring factors when processing multidimensional dam data, and they also have difficulty accurately extracting and utilizing effective information from complex temperature data, resulting in insufficient accuracy and reliability in dam deformation prediction.
A variational autoencoder was used to reduce the dimensionality of temperature data. A Gaussian mixture model and an extended long short-term memory network were combined to cluster and model the dam monitoring data, and a dam deformation prediction model was constructed. By combining temperature feature vectors and water level information, a neural network model was trained using similar datasets to improve prediction accuracy.
By using dimensionality reduction and cluster analysis, the effective utilization rate of temperature data was improved, computational losses were reduced, and the characteristic relationships between environmental data were explored in depth, thereby improving the accuracy and reliability of dam deformation prediction.
Smart Images
Figure CN121456515A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The application belongs to the technical field of deformation prediction, and particularly relates to a dam deformation prediction modeling method based on a VAE (Variational Autoencoder)-GMM (Gaussian Mixture Model)-xLSTM (Extended Long Short-Term Memory), a prediction method, equipment and a storage medium. BACKGROUND
[0002] As a key structure in water conservancy projects, the structural safety of a dam directly affects the stability of the downstream area and the sustainable development of the social economy. Therefore, it is crucial to carry out structural response monitoring of the dam. Structural response monitoring mainly involves real-time or periodic observation and analysis of the structural response of the dam body and the surrounding environment, so as to identify potential safety risks in a timely manner and take appropriate measures.
[0003] Traditional structural response analysis methods mainly include statistical analysis, finite element analysis and shallow machine learning methods. However, these traditional methods have the following challenges:
[0004] (1) The complexity of processing multi-dimensional data makes it difficult to analyze the data characteristics of the dam in different operating periods;
[0005] (2) Information is not fully utilized, making it difficult to effectively capture the deep nonlinear relationship between long time series structural response and monitoring factors;
[0006] (3) It is difficult to accurately extract and utilize effective information from complex temperature data.
[0007] These challenges limit the practical application of traditional methods and make it difficult to meet the requirements of high reliability. Therefore, there is an urgent need to research and develop new structural response monitoring and prediction technologies to improve the accuracy and reliability of monitoring. SUMMARY
[0008] The purpose of the present application is to provide a dam deformation prediction modeling method, a prediction method, equipment and a storage medium to solve the problem of difficulty in feature analysis caused by the complexity of processing multi-dimensional data in traditional methods, the difficulty in effectively capturing the deep nonlinear relationship between long time series structural response and monitoring factors, and the difficulty in accurately extracting and utilizing effective information from complex temperature data.
[0009] The present application solves the above technical problems by the following technical solutions: a dam deformation prediction modeling method, comprising:
[0010] Obtaining dam monitoring data, the dam monitoring data comprising temperature data, water level information, aging factors and dam displacement; the aging factor refers to a stepwise regression equation of an aging parameter, and the aging parameter refers to a time change rate from a monitoring time to a dam operation starting time;
[0011] Performing dimension reduction processing on the temperature data to obtain a low-dimensional temperature feature vector;
[0012] Clustering the low-dimensional temperature feature vector and the water level information to obtain N environment data sets;
[0013] Combining the N environment data sets with corresponding aging factors and dam displacements to obtain N sample data sets;
[0014] Constructing N neural network models, each neural network model corresponding to a sample data set;
[0015] For each neural network model, training and testing it by using the corresponding sample data set to obtain a dam deformation prediction model.
[0016] Further, before the dimension reduction processing, the dam monitoring data is also preprocessed, and the preprocessing includes cleaning, smoothing and denoising, and normalization.
[0017] Further, a variational autoencoder is used to perform dimension reduction processing on the temperature data, and the variational autoencoder comprises an encoder; and the specific process of the dimension reduction processing comprises:
[0018] Performing time alignment on the temperature data to obtain a multi-dimensional temperature data sequence;
[0019] Mapping the multi-dimensional temperature data sequence to an implicit variable distribution by using the encoder, and randomly sampling the implicit variable distribution to obtain a low-dimensional temperature feature vector, while introducing a KL divergence to regularize the implicit variable distribution.
[0020] Further, a Gaussian mixture model is used to cluster the low-dimensional temperature feature vector and the water level information.
[0021] Further, the neural network model is an extended long short-term memory network.
[0022] Further, for each neural network model, training and testing it by using the corresponding sample data set, comprising:
[0023] Dividing the sample data set into a training set and a test set;
[0024] Training the neural network model by using the training set to obtain a trained neural network model;
[0025] obtaining a prediction result of the trained neural network model on the test set;
[0026] integrating the prediction results of the N trained neural network models on the test set in time sequence to obtain a displacement prediction sequence;
[0027] evaluating the performance of each trained neural network model according to the displacement prediction sequence and a label sequence; wherein the label sequence is obtained by integrating the expected outputs of the N test sets in time sequence.
[0028] Based on the same concept, the present application also provides a dam deformation prediction method, comprising:
[0029] obtaining dam deformation influencing factors, the dam deformation influencing factors including temperature data, water level information and time effect factors;
[0030] performing dimension reduction processing on the temperature data to obtain a low-dimensional temperature feature vector;
[0031] combining the low-dimensional temperature feature vector and the water level information to obtain new environment data;
[0032] calling N dam deformation prediction models; the N dam deformation prediction models are constructed by using the dam deformation prediction modeling method in any one of claims 1-7;
[0033] selecting a required dam deformation prediction model from the N dam deformation prediction models according to the distance between the new environment data and each environment data set;
[0034] combining the new environment data and the time effect factors to obtain to-be-predicted data;
[0035] inputting the to-be-predicted data into the required dam deformation prediction model to obtain a prediction result.
[0036] Further, the specific calculation formula of the distance between the new environment data and each environment data set is:
[0037] ;
[0038] wherein, denotes the distance between the new environment data and the i th environment data set; denotes the new environment data; denotes the mean of the i th environment data set; ∑ denotes the covariance matrix of the i th environment data set; ∑ -1 denotes the inverse matrix of the covariance matrix of the i th environment data set.
[0039] Based on the same concept, the present invention also provides an electronic device, including a memory, a processor, and a computer program / instructions stored in the memory, wherein the processor executes the computer program / instructions to implement the dam deformation prediction modeling method or dam deformation prediction method as described above.
[0040] Based on the same concept, the present invention also provides a computer-readable storage medium having a computer program / instruction stored thereon, which, when executed by a processor, implements the dam deformation prediction modeling method or dam deformation prediction method as described above.
[0041] Compared with the prior art, the beneficial effects of the present invention are as follows:
[0042] This invention first performs dimensionality reduction on temperature data, extracting key information while preserving the essential structure and potential patterns of the temperature data, thereby improving the effective utilization rate of temperature data and reducing the difficulty of subsequent data processing. Then, it performs cluster analysis on environmental data (including low-dimensional temperature features and water level information), reducing computational losses caused by complex temperature data, while also deeply exploring the characteristic relationships between environmental data, which is beneficial to improving the performance of subsequent models. Finally, it trains a neural network model using the sample dataset corresponding to each cluster, improving model performance by training the model with data of similar features, thereby improving prediction accuracy and reliability. Attached Figure Description
[0043] To more clearly illustrate the technical solution of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only one embodiment of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0044] Figure 1 This is a flowchart of the dam deformation prediction modeling method in an embodiment of the present invention;
[0045] Figure 2 This is a schematic diagram of the temperature data dimensionality reduction processing results in an embodiment of the present invention;
[0046] Figure 3 This is a schematic diagram illustrating the environmental data clustering effect in an embodiment of the present invention;
[0047] Figure 4 This is a schematic diagram comparing test results in an embodiment of the present invention;
[0048] Figure 5 This is a flowchart of the dam deformation prediction method in an embodiment of the present invention. Detailed Implementation
[0049] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments. 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.
[0050] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.
[0051] Example 1
[0052] like Figure 1 As shown, the dam deformation prediction modeling method provided in this embodiment of the invention includes the following steps:
[0053] Step 1: Obtain dam monitoring data.
[0054] In this embodiment, the dam monitoring data includes temperature data, water level information (e.g., reservoir water level), time-dependent factors, and dam displacement. The temperature data includes ambient temperature, water temperature, concrete temperature within the dam body, and deep foundation temperature. The time-dependent factors refer to the stepwise regression equation of the time-dependent parameters, which are the rate of change of the monitoring time from the start of dam operation to the start of operation. The specific expression is as follows:
[0055] (1)
[0056] (2)
[0057] in, Indicates the timeliness parameter. Indicates the monitoring time. Indicates the start time of dam operation. Indicates timeliness factors, , All of these represent statistical coefficients. When constructing the sample dataset, only the time-sensitivity parameter needs to be obtained. Statistical coefficients , It is automatically determined during neural network model training.
[0058] Factors influencing dam deformation typically include hydraulic and temperature factors, i.e., environmental factors. However, for concrete dams, the longer the time, the more prone the concrete is to failure. Therefore, aging is also an important factor affecting dam deformation. This invention incorporates aging factors into the monitoring data, thereby improving the accuracy of dam deformation prediction.
[0059] At each monitoring moment, a set of dam monitoring data can be acquired, including water level information, ambient temperature, water temperature, concrete temperature inside the dam, deep temperature of the dam foundation, time-dependent factors calculated based on the current monitoring moment, and dam displacement.
[0060] For example, a dam located on the eastern source of the Pihe River, a tributary of the Huai River, primarily serves flood control, while also providing irrigation, power generation, and navigation. Construction of the dam began in 1954, with a maximum height of 75.9 m and a crest axis length of 510 m. It is equipped with an automatic monitoring system capable of long-term monitoring of data such as deformation, settlement, stress-strain, and temperature (including air, water, and internal temperatures). The dam has 20 cycloidal (PL) and 3 inverted cycloidal (IP) lines installed on its body and foundation for displacement monitoring, and numerous thermometers are also installed inside the dam for temperature monitoring.
[0061] To improve data quality, dam monitoring data undergoes preprocessing. In this embodiment, preprocessing includes cleaning (such as missing value imputation and outlier removal), smoothing and denoising, and normalization (such as Z-score standardization). Preprocessing ensures data stability and improves the training effect of subsequent models.
[0062] Step 2: Perform dimensionality reduction on the temperature data to obtain a low-dimensional temperature feature vector.
[0063] Temperature data includes ambient temperature, water temperature, concrete temperature within the dam body, and deep temperature of the dam foundation. The complexity of this temperature data makes it difficult to extract and utilize effective information. To address this problem, this invention performs dimensionality reduction on the temperature data. Lower-dimensional data reduces the difficulty of subsequent data processing and improves the effective utilization rate of the temperature data. In a specific embodiment of this invention, a variational autoencoder (VAE) is used to perform dimensionality reduction on the temperature data. The variational autoencoder includes an encoder, and the specific process of dimensionality reduction includes:
[0064] Step 2.1: Time-align the temperature data to obtain a multidimensional temperature data sequence;
[0065] Step 2.2: Use an encoder to map the multidimensional temperature data sequence into a latent variable distribution, and randomly sample the latent variable distribution to obtain a low-dimensional temperature feature vector. At the same time, introduce KL divergence to regularize the latent variable distribution.
[0066] The latent variable space has a much smaller dimension than the multidimensional temperature data sequence, exhibiting compression properties. Introducing KL divergence during random sampling ensures the continuity and interpretability of the latent space. During dimensionality reduction, only the encoder of the variational autoencoder is applied; no decoder is used. The principal components output by the variational autoencoder are the most representative combinations of temperature features from the multidimensional temperature data sequence, specifically including:
[0067] Principal Component 1: The trend characteristics of concrete temperature and deep foundation temperature within the dam body reflect the thermal stability and temperature gradient of the internal structure of the dam body, revealing the potential causes of structural stress changes.
[0068] Principal component 2: The seasonal coupling variation characteristics of water temperature and ambient temperature characterize the heat exchange behavior between water and air under environmental influences, and are an important indicator for judging external disturbances.
[0069] Principal component 3: Abnormal temperature fluctuation signal characteristics, used to identify sudden temperature anomalies, such as local thermal expansion and contraction caused by sudden changes in temperature.
[0070] After dimensionality reduction, multiple temperature components can be obtained; typically, a representative temperature component is selected. For example... Figure 2 As shown, after dimensionality reduction, the first temperature component accounts for 98.13%, and the second temperature component accounts for 1.64%. Therefore, only the first temperature component needs to be selected after dimensionality reduction.
[0071] This invention employs a variational autoencoder for dimensionality reduction and utilizes a learnable nonlinear generative model to construct a continuous and probabilistic low-dimensional latent space, which preserves the complex temperature data distribution structure and supports sample reconstruction and new sample generation.
[0072] Step 3: Cluster the low-dimensional temperature feature vector and water level information to obtain N environmental datasets.
[0073] In this embodiment, a Gaussian Mixture Model (GMM) is used to cluster the low-dimensional temperature feature vector and water level information, resulting in N clusters. Each cluster represents an environmental dataset, and each environmental data point in each dataset includes the low-dimensional temperature feature vector and water level information at a single sampling time. The contour coefficient and DB index are used as criteria for selecting the number of clusters; a larger contour coefficient and a smaller DB index indicate better clustering results.
[0074] Figure 3 The diagram shows the clustering results, which yielded four clusters, or four environmental datasets. The environmental data (including low-dimensional temperature feature vectors and water level information) in each environmental dataset have similar characteristics. Using data with similar characteristics to train the neural network model is beneficial for obtaining a better-performing dam deformation prediction model.
[0075] This invention uses Gaussian Mixture Model (GMM) for clustering, which can flexibly identify cluster structures of arbitrary shapes and overlaps through soft partitioning and Gaussian distribution modeling, making it suitable for clustering tasks with complex data distributions.
[0076] Step 4: Combine the N environmental datasets with the corresponding time-sensitive factors and dam displacements to obtain N sample datasets.
[0077] Each sample in each dataset corresponds to a sampling time. Each sample includes a low-dimensional temperature feature vector, water level information, time-related factors, and dam displacement at the corresponding sampling time. The input vector consists of the low-dimensional temperature feature vector, water level information, and time-related factors, and the expected output is the dam displacement.
[0078] Step 5: Construct N neural network models, each corresponding to a sample dataset.
[0079] In this embodiment, the neural network model is an Extended Long Short-Term Memory (xLSTM) network. The xLSTM network includes original LSTM memory cells with constant error polling and gating, novel Stacked Long Short-Term Memory (sLSTM) network memory cells with exponential gating, and Mixed Long Short-Term Memory (mLSTM) network memory cells. sLSTM provides a novel memory hybridization technique, while mLSTM is fully parallelizable, featuring novel matrix memory cell states and a new covariance update rule. mLSTM and sLSTM in the residual block generate xLSTM blocks, and the stacked xLSTM blocks constitute the xLSTM architecture.
[0080] xLSTM integrates various extensions of LSTM structures, possessing stronger time series modeling and parameter representation capabilities. It can more effectively capture complex long dependencies and achieve more accurate time series prediction.
[0081] Step 6: For each neural network model, train and test it using the corresponding sample dataset to obtain the dam deformation prediction model.
[0082] In a specific embodiment of the present invention, training and testing the corresponding neural network model using a sample dataset includes:
[0083] Step 6.1: Divide the sample dataset into a training set and a test set in an 8:2 ratio;
[0084] Step 6.2: Train the neural network model using the training set to obtain the trained neural network model;
[0085] Step 6.3: Obtain the prediction results of the trained neural network model on the test set;
[0086] Step 6.4: Integrate the prediction results of the N trained neural network models on the test set in chronological order to obtain the displacement prediction sequence;
[0087] Step 6.5: Evaluate the performance of each trained neural network model based on the displacement prediction sequence and the label sequence; whereby the label sequence is obtained by integrating the expected outputs of N test sets in chronological order.
[0088] Figure 4 A diagram showing the comparison of test results is provided. Figure 4 It can be seen that the predicted values (i.e., the displacement prediction sequence) are close to the true values (label sequence), and most of them are within the range of 95%, indicating that the dam deformation prediction model constructed by the method of this invention has high prediction accuracy.
[0089] Example 2
[0090] like Figure 5 As shown, the dam deformation prediction method provided in this embodiment of the invention includes the following steps:
[0091] Step 1: Obtain the influencing factors of dam deformation, including temperature data, water level information, and time-related factors;
[0092] Step 2: Use VAE to perform dimensionality reduction on the temperature data to obtain a low-dimensional temperature feature vector;
[0093] Step 3: Combine the low-dimensional temperature feature vector and water level information to obtain new environmental data;
[0094] Step 4: Call N dam deformation prediction models; the N dam deformation prediction models are constructed using the dam deformation prediction modeling method in Embodiment 1 of this invention;
[0095] Step 5: Based on the distance between the new environmental data and each environmental dataset, select the required dam deformation prediction model from N dam deformation prediction models;
[0096] Step 6: Combine the new environmental data with time-sensitive factors to obtain the data to be predicted;
[0097] Step 7: Input the data to be predicted into the required dam deformation prediction model to obtain the prediction results.
[0098] In this embodiment, the specific formula for calculating the distance between the new environmental data and each environmental dataset in step 5 is as follows:
[0099] (3)
[0100] in, This represents the distance between the new environmental data and the i-th environmental dataset; This represents new environmental data (i.e., low-dimensional temperature feature vectors and water level information). ∑ represents the mean of the i-th environmental dataset, i.e., the center of each environmental dataset; ∑ represents the covariance matrix of the i-th environmental dataset, describing the dispersion and correlation of the data distribution in that environmental dataset; ∑ -1 Let represent the inverse matrix of the covariance matrix of the i-th environmental dataset; The vector represents the deviation between the new environmental data and the mean of the i-th environmental dataset, with the superscript T indicating matrix transpose.
[0101] Example 3
[0102] This invention also provides an electronic device, which includes a memory, a processor, and a computer program / instructions stored in the memory. The processor executes the computer program / instructions to implement the dam deformation prediction modeling method or dam deformation prediction method of this invention.
[0103] Although not shown, the electronic device includes a processor that can perform various appropriate operations and processes based on programs and / or data stored in read-only memory (ROM) or loaded from a storage portion into random access memory (RAM). The processor can be a multi-core processor or may contain multiple processors. In some embodiments, the processor may include a general-purpose main processor and one or more specialized coprocessors, such as a central processing unit, graphics processing unit (GPU), neural network processor (NPU), digital signal processor (DSP), etc. Various programs and data required for device operation are also stored in RAM. The processor, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.
[0104] The processor and memory described above are used together to execute programs / instructions stored in the memory. When the program / instructions are executed by the computer, they can implement the methods, steps, or functions described in the above embodiments.
[0105] Although not shown, embodiments of the present invention also provide a computer-readable storage medium having a computer program / instruction stored thereon, which, when executed by a processor, implements the dam deformation prediction modeling method or dam deformation prediction method of the present invention.
[0106] Readable storage media include both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient media, such as modulated data signals and carrier waves.
[0107] The above description only discloses specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or modifications that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.
Claims
1. A method for predicting and modeling dam deformation, characterized in that, The modeling method includes: Acquire dam monitoring data, which includes temperature data, water level information, time-dependent factors, and dam displacement; the time-dependent factors refer to the stepwise regression equation of the time-dependent parameters, and the time-dependent parameters refer to the rate of change of the monitoring time from the start of dam operation. The temperature data is subjected to dimensionality reduction processing to obtain a low-dimensional temperature feature vector; Clustering the low-dimensional temperature feature vector and water level information yields N environmental datasets; The N environmental datasets are combined with the corresponding time-sensitive factors and dam displacements to obtain N sample datasets; Construct N neural network models, each corresponding to a sample dataset; For each neural network model, it is trained and tested using the corresponding sample dataset to obtain the dam deformation prediction model.
2. The dam deformation prediction modeling method according to claim 1, characterized in that, Before dimensionality reduction, the dam monitoring data is preprocessed, including cleaning, smoothing and denoising, and normalization.
3. The dam deformation prediction modeling method according to claim 1, characterized in that, The temperature data is subjected to dimensionality reduction using a variational autoencoder, which includes an encoder; the specific process of the dimensionality reduction includes: The temperature data is time-aligned to obtain a multidimensional temperature data sequence; The multidimensional temperature data sequence is mapped to a latent variable distribution using an encoder, and the latent variable distribution is randomly sampled to obtain a low-dimensional temperature feature vector. At the same time, KL divergence is introduced to apply regularization constraints to the latent variable distribution.
4. The dam deformation prediction modeling method according to claim 1, characterized in that, A Gaussian mixture model was used to cluster the low-dimensional temperature feature vector and water level information.
5. The dam deformation prediction modeling method according to claim 1, characterized in that, The neural network model is an Extended Long Short-Term Memory Network.
6. The dam deformation prediction modeling method according to any one of claims 1 to 5, characterized in that, For each neural network model, it is trained and tested using the corresponding sample dataset, including: The sample dataset is divided into a training set and a test set; The neural network model is trained using the training set to obtain the trained neural network model; Obtain the prediction results of the trained neural network model on the test set; The prediction results of N trained neural network models on the test set are integrated in chronological order to obtain the displacement prediction sequence. The performance of each trained neural network model is evaluated based on the displacement prediction sequence and the label sequence; wherein the label sequence is obtained by integrating the expected outputs of N test sets in chronological order.
7. A method for predicting dam deformation, characterized in that, The prediction method includes: The factors influencing dam deformation are obtained, including temperature data, water level information, and time-related factors. The temperature data is subjected to dimensionality reduction processing to obtain a low-dimensional temperature feature vector; The low-dimensional temperature feature vector and water level information are combined to obtain new environmental data; N dam deformation prediction models are invoked; the N dam deformation prediction models are constructed using the dam deformation prediction modeling method as described in any one of claims 1 to 7; Based on the distance between the new environmental data and each environmental dataset, the required dam deformation prediction model is selected from N dam deformation prediction models; By combining new environmental data with time-sensitive factors, the data to be predicted is obtained; The data to be predicted is input into the required dam deformation prediction model to obtain the prediction results.
8. The method for predicting dam deformation according to claim 7, characterized in that, The specific formula for calculating the distance between the new environmental data and each environmental dataset is as follows: ； in, This represents the distance between the new environmental data and the i-th environmental dataset; This indicates new environmental data; Let represent the mean of the i-th environmental dataset; ∑ represents the covariance matrix of the i-th environmental dataset; ∑ -1 Let represent the inverse of the covariance matrix of the i-th environmental dataset.
9. An electronic device comprising a memory, a processor, and a computer program / instructions stored in the memory, characterized in that, The processor executes the computer program / instructions to implement the dam deformation prediction modeling method as described in any one of claims 1 to 6 or the dam deformation prediction method as described in claim 7 or 8.
10. A computer-readable storage medium having a computer program / instructions stored thereon, characterized in that, When the computer program / instruction is executed by the processor, it implements the dam deformation prediction modeling method as described in any one of claims 1 to 6 or the dam deformation prediction method as described in claim 7 or 8.