Intelligent design method of concrete faced rockfill dam profile based on data feature fusion

Abstract

The invention relates to the technical field of artificial intelligence, in particular to a concrete faced rockfill dam profile intelligent design method based on data feature fusion, and the method comprises the steps: building a learning prediction model to predict the height and slope ratio of a dam body through collecting initial data and historical data such as feature water levels, anti-seismic attributes, stone parameters and the like; dam body contour parameters are calculated, and the area proportion of each partition is predicted; thirdly, the dam body contour parameters, the feature water level, the anti-seismic attribute and the partition area proportion are fused into a multi-dimensional feature vector, the multi-dimensional feature vector is input into the generative adversarial network model, and a partition design image of the concrete faced rockfill dam is generated. According to the method, intelligent design of the concrete faced rockfill dam profile scheme is achieved, the design efficiency is greatly improved, and the intelligent level of concrete faced rockfill dam design in China is further improved.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to an intelligent design method for a panel rockfill dam profile based on data feature fusion. Background Art

[0002] Dams are the core of hydropower engineering construction. Concrete face rockfill dams and core wall dams are the two main types of earth-rock dams. Among them, face rockfill dams have the characteristics of safety, economy and good adaptability, and have become the mainstream dam type of water retaining structures for hydropower facilities such as conventional hydropower and pumped storage power stations. The conventional design of face rockfill dams has gradually matured. In China alone, the number of existing dams designed and built has exceeded 98,000. However, the design mode of concrete face rockfill dams, which is highly dependent on the manual work of engineers, faces the development difficulties of difficult experience inheritance and efficiency improvement. With the advancement and implementation of pumped storage power stations and hydropower development projects in recent years, the digital and intelligent design of face rockfill dams is an inevitable choice to achieve high-quality and efficient engineering design and match the development needs of engineering intelligent construction. It is necessary to transform the design intelligently to solve the current problems.

[0003] With the development of artificial intelligence technology, cutting-edge AI technologies represented by generative AI are reshaping the engineering design industry, bringing efficient design tools to the industry, effectively reusing existing design data, and improving design quality. Generative AI has good deep extraction and learning of existing data features, as well as the ability to generate new designs. Therefore, generative AI imitates human engineers to learn the existing design data of panel rockfill dams, learn past panel rockfill dam design experience, analyze complex drawings, combine demand texts, integrate data and experience knowledge, and then realize a new design.

[0004] Patent document with publication number CN117094059A discloses a design method for a panel rockfill dam with a zigzag-type dam crest, wherein the rockfill dam comprises a dam body, the elevation of the panel crest of different sections of the dam body is the sum of the static water level in normal use and the reserved settlement there, the upstream panel slope of the dam body is unique and fixed, and the upstream panel intersects the front edge of the wave-breaking wall bottom plate with a zigzag line; the center line of the dam crest road is designed according to the reserved settlement of the dam crest and the panel slope, a zigzag-type wave-breaking wall and dam crest road model is established, an upstream panel overall model is established, a panel model is constructed according to the wave-breaking wall bottom plate surface, and a panel rockfill dam with a zigzag-type dam crest is constructed.

[0005] This shows that the following problems exist: due to the many subjective judgments and diverse design methods in the design process of the existing technology, it is difficult to form a unified standard process, resulting in low consistency and repeatability of the design results. Summary of the invention

[0006] To this end, the present invention provides an intelligent design method for panel rockfill dam profile based on data feature fusion, so as to overcome the problems in the prior art design process that there are many subjective judgments and diverse design methods, it is difficult to form a unified standard process, and the consistency and repeatability of the design results are low.

[0007] To achieve the above object, the present invention provides a method for intelligent design of a concrete face rockfill dam profile based on data feature fusion, comprising: Step S1, collecting initial data and foundation surface parameters required for the design of a concrete face rockfill dam, wherein the initial data includes characteristic water level, seismic properties and stone material parameters, and also collecting historical data corresponding to the initial data; Step S2, constructing a learning prediction model according to the initial parameters and the historical data to predict the dam body height, upstream dam slope ratio and downstream dam slope ratio of the rockfill dam; Step S3, calculating according to the dam body height, the upstream dam slope ratio, the downstream dam slope ratio and the foundation surface parameters to obtain dam body contour parameters, and predicting the target partition area ratios of the main stacking area, the secondary stacking area, the drainage area and the mold increasing area according to the dam body contour parameters; Step S4, drawing an input image according to the dam body contour parameters, and fusing the input image, the characteristic water level, the seismic properties and the area ratios of each target partition to obtain a multi-dimensional feature vector; Step S5, using the multi-dimensional feature vector as a model input to construct a generative adversarial network model, and generating a panel rockfill dam zoning design image according to the generative adversarial network model.

[0008] Furthermore, the process of step S2 includes: Normalizing the characteristic water level, the earthquake resistance property, the stone material parameter and the historical data to obtain a first processing result; Performing feature extraction on the first processing result to obtain a first feature result; Characterizing the first characteristic result to obtain a characterization result; The characterization results are used as inputs of the learning prediction model to predict the dam body height, the upstream dam slope ratio and the downstream dam slope ratio.

[0009] Furthermore, the process of using the first feature result as the input of the learning prediction model to predict the dam body height, the upstream dam slope ratio and the downstream dam slope ratio includes: Performing correlation analysis on the first feature result to obtain an analysis result; The dam body height and the slope ratio corresponding to the historical data are used as target variables; Using a part of the analysis results and the target variable as a training set, and another part of the analysis results and the target variable as a validation set; Training the learning prediction model according to the training set to obtain a first initial training model; Verifying the initial training model according to the verification set to obtain a first target training model; The dam body height, the upstream dam slope ratio and the downstream dam slope ratio are predicted according to the target training model.

[0010] Furthermore, the process of calculating according to the dam body height, the upstream dam slope ratio, the downstream dam slope ratio and the foundation surface parameters to obtain the dam body contour parameters includes: The ratio of the dam body height to the upstream dam slope length is taken as the upstream dam slope ratio, the ratio of the dam body height to the downstream dam slope length is taken as the downstream dam slope ratio, and the width of the dam body bottom in contact with the foundation is taken as the foundation surface width; The sum of the upstream dam slope length, the downstream dam slope length and the foundation surface width is taken as the dam crest width; The dam body profile parameters are determined according to the dam body height, the dam top width, the upstream dam slope ratio and the downstream dam slope ratio using a geometric method.

[0011] Furthermore, the process of predicting the area ratio of the dam body partitions of the main pile, the secondary pile, the drainage and the increased mold according to the dam body contour parameters includes: The dam body is divided into a main pile area, a secondary pile area, a drainage area and an increased mold area according to the dam body contour parameters; Calculate the area of ​​the main pile area, the secondary pile area, the drainage area and the mold increase area according to a geometric method; Comparing the area of ​​each of the regions with the total area of ​​the dam to obtain the area ratio of each sub-region; Collecting the partition area ratio data of the historical data, and analyzing the partition area ratio data to obtain a second analysis result; The second analysis result is used as an input of a prediction model, and the prediction model is trained using the partition area ratio data to obtain the area ratio of each target partition.

[0012] Furthermore, the process of step S4 includes: Converting the characteristic water level, the seismic resistance attribute and the partition area ratio into numerical features to obtain a conversion result; Fusing the conversion result and the input image to obtain a fusion result; Extracting multimodal features from the fusion result to obtain a second feature result; Performing normalization and feature alignment processing on the second feature result to obtain a second processing result; The second processing results are stacked along the data channel to obtain the multi-dimensional feature vector.

[0013] Furthermore, the process of step S5 includes: Counting the cross-sectional profile parameters of the input image, the initial data, the coordinate parameters of each of the subareas and the cross-sectional area ratio parameters to obtain the distribution rules of different parameters; According to the distribution law, the upstream dam slope ratio, the downstream dam slope ratio, and the area of ​​each region are augmented and expanded by parameterized adjustment to construct a multi-source heterogeneous training data set and a test data set; Performing feature data fusion on the training data set and the test data set to construct a corresponding input feature data set and an output feature data set; Taking the input feature data set as an input parameter, and using the training data set to train the generative adversarial network model to obtain a second initial training model; Verifying the second initial training model according to the verification data set to obtain a second target training model; The model is trained according to the second objective to generate the zoning design image of the concrete face rockfill dam.

[0014] Furthermore, the process of augmenting and expanding the upstream dam slope ratio, the downstream dam slope ratio, and the area of ​​each region by parameterized adjustment according to the distribution law to construct a multi-source heterogeneous training data set and a test data set includes: Determining parameter thresholds of the upstream dam slope ratio, the downstream dam slope ratio and the area of ​​each region; defining the adjustment step length of the upstream dam slope ratio, the downstream dam slope ratio and the area of ​​each region; Establishing a combination rule according to the parameter threshold and the adjustment step size; Generate a series of parameter value combinations according to the combination rule; Performing geometric transformation and simulation on the input image according to the generated parameter value combination to generate a target cross-sectional image; Assigning corresponding parameter values ​​to each of the target images as label information; The training data set and the test data set are constructed according to the target input image and the label information.

[0015] Furthermore, the process of using the verification data set to verify the second initial training model to obtain the second target training model includes: Creating a binary mask using the profile parameters of the input image, wherein the profile area is marked as 1 and the remaining profile areas are marked as 0; Filtering the output feature according to the mask to obtain a filtering result; The generated image is determined according to the filtering result to be valid only within the panel cross-section contour.

[0016] Further, the process of filtering the output feature according to the mask to obtain a filtering result includes: in, is the original output height H, width W, and channel C of the generative network, represents the element-wise multiplication Hadamard product, is the effective generated image after mask filtering, is a binary mask matrix that satisfies: .

[0017] Compared with the prior art, the beneficial effect of the present invention is that the present invention reduces a large amount of manual calculation and drawing work in traditional design through automated and intelligent methods, and significantly shortens the design cycle. By using historical data and machine learning models for prediction, it is possible to design based on a large amount of case experience, reduce human errors, and improve the accuracy and reliability of the design. By learning the prediction model, key parameters such as dam height and dam slope ratio can be predicted more accurately, thereby optimizing the dam structure and improving engineering stability. The design scheme can be quickly adjusted according to different initial data and foundation surface parameters to adapt to different engineering environments and requirements. The labor cost and repetitive work in the design process are reduced, and the overall design cost is reduced through intelligent design methods.

[0018] In particular, normalization eliminates the impact of different dimensions and data ranges, allowing the model to better learn data features and thus improve prediction accuracy. Normalized data can speed up model convergence and reduce the time and resources required for calculations. Design predictions based on historical data and machine learning models can reduce human errors and design risks and improve engineering safety.

[0019] In particular, through correlation analysis, the most relevant features to the target variables (dam height and slope ratio) were screened out, reducing the interference of noise and irrelevant variables, thereby improving the prediction accuracy of the model. Using training sets and validation sets for model training and validation, respectively, helps to avoid overfitting and ensure that the model has good generalization ability. The intelligent prediction model can quickly output design parameters, greatly shortening the time required for traditional design methods and accelerating project progress.

[0020] In particular, the use of geometric methods to calculate the dam body contour parameters can ensure the accuracy of the calculation results and avoid human errors. The use of geometric methods to calculate the dam body contour parameters can ensure the accuracy of the calculation results and avoid human errors. By improving design efficiency and accuracy, the design cost can be reduced.

[0021] In particular, by predicting the area ratio of the partitions, the material configuration can be optimized. By optimizing the material configuration, the cost of materials can be reduced and the profitability of the project can be improved. The stability of the dam body can also be improved and the project risk can be reduced. By predicting the area ratio of the partitions, the material composition of the dam body can be better understood, thereby optimizing the dam structure and improving the stability of the project.

[0022] In particular, by fusing features from different modalities, the performance of the GAN model can be improved. By fusing features from different modalities, the robustness of the GAN model can be improved. By analyzing the impact of different modal features on the generated images, the interpretability of the GAN model can be improved.

[0023] In particular, through parameter adjustment and augmentation, diverse training and test data sets can be created, which helps improve the generalization ability of the model and enables it to handle more diverse design scenarios. By performing feature fusion and parameter adjustment on the data set, the amount of data required for model training can be reduced while improving training efficiency. Data augmentation can increase the diversity of training data, thereby improving the generalization ability of the model. By training the generative adversarial network model, high-quality panel rockfill dam zoning design images can be generated.

[0024] In particular, the automated parameterization and simulation process greatly reduces the time of manual design, allowing designers to quickly explore multiple design options and optimize the design process. Multi-source heterogeneous data sets contain a variety of possible profile designs, which helps to avoid overfitting during model training and ensures that the model can handle diverse design requirements. By simulating different parameter combinations, the optimal or most unfavorable design parameters can be identified, thereby guiding actual engineering design and improving the quality and safety of the final design.

[0025] In particular, by focusing only on the area within the profile, the model can more accurately learn design-related features and reduce interference from external noise or irrelevant information. By excluding the area outside the profile, unnecessary calculations are reduced, computational efficiency is improved, and resource consumption is reduced. Ensuring that the generated zoning design image of the face rockfill dam contains only valid information within the profile helps to generate a design solution that better meets actual engineering needs.

[0026] In particular, mask filtering can ensure that the generated image is only valid within the panel profile, avoiding interference from irrelevant areas, thereby improving the accuracy and reliability of the generated image. Mask filtering can reduce the amount of image information that the generated network needs to process, thereby reducing the amount of calculation and computing time. The area of ​​the generated image can be controlled by adjusting the type and parameters of the mask, such as generating an area of ​​a specific shape or size. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 A schematic diagram of a flow chart of an intelligent design method for a face rockfill dam profile based on data feature fusion provided by an embodiment of the present invention; Figure 2 A schematic flow chart of step S2 in the method for intelligent design of a concrete face rockfill dam profile based on data feature fusion provided in an embodiment of the present invention; Figure 3 A schematic flow chart of step S4 in the method for intelligent design of a concrete face rockfill dam profile based on data feature fusion provided in an embodiment of the present invention; Figure 4 A schematic flow chart of step S5 in the method for intelligent design of a concrete face rockfill dam profile based on data feature fusion provided in an embodiment of the present invention; Figure 5 The text-image fusion and intelligent design generation results provided by the embodiments of the present invention. DETAILED DESCRIPTION

[0028] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0029] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0030] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the drawings. This is merely for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.

[0031] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0032] See also Figure 1 As shown, an embodiment of the present invention provides a method for intelligent design of a concrete face rockfill dam profile based on data feature fusion, comprising: Step S1, collecting initial data and foundation surface parameters required for the design of a concrete face rockfill dam, wherein the initial data includes characteristic water level, seismic properties and stone material parameters, and also collecting historical data corresponding to the initial data; Step S2, constructing a learning prediction model according to the initial parameters and the historical data to predict the dam body height, upstream dam slope ratio and downstream dam slope ratio of the rockfill dam; Step S3, calculating according to the dam body height, the upstream dam slope ratio, the downstream dam slope ratio and the foundation surface parameters to obtain dam body contour parameters, and predicting the target partition area ratios of the main stacking area, the secondary stacking area, the drainage area and the mold increasing area according to the dam body contour parameters; Step S4, drawing an input image according to the dam body contour parameters, and fusing the input image, the characteristic water level, the seismic properties and the area ratios of each target partition to obtain a multi-dimensional feature vector; Step S5, using the multi-dimensional feature vector as a model input to construct a generative adversarial network model, and generating a panel rockfill dam zoning design image according to the generative adversarial network model.

[0033] Specifically, collect characteristic water level data, including design flood level, normal water storage level, dead water level, etc. Collect seismic property data, such as earthquake intensity, seismic motion parameters, etc. Collect stone material parameters, including planned excavation volume of material source, elastic modulus, porosity, dry density, and moisture content of rockfill and core wall clay. Collect parameters such as topography, geology, and geotechnical engineering properties of the foundation surface. Collect historical design case data corresponding to the above initial data, including dam body height, dam slope ratio, partition area ratio, etc. Determine the key features that affect dam body height and dam slope ratio. Select appropriate machine learning algorithms, such as random forest, support vector machine, neural network, etc. Use historical data to train the model, with the input features being the initial data and foundation surface parameters, and the output being the dam body height, upstream dam slope ratio, and downstream dam slope ratio. Verify the prediction accuracy of the model through methods such as cross-validation. Calculate the dam body contour parameters based on the predicted dam body height, upstream and downstream dam slope ratio, and foundation surface parameters. Using the calculated dam contour parameters, the target partition area ratios of the main pile area, secondary pile area, drainage area and mold augmentation area are predicted through regression analysis or other prediction methods. Draw the dam profile input image according to the dam contour parameters. Digitize the input image and extract image features. Convert non-image data such as characteristic water level, seismic properties and target partition area ratio into numerical features. Merge all features into a multidimensional feature vector. Use the multidimensional feature vector as input to build a generative adversarial network (GAN) model. Design and train the generator and discriminator. Through iterative training, the generator can generate realistic panel rockfill dam partition design images. Adjust the model parameters to optimize the quality and accuracy of the generated images. Use the trained generator to generate panel rockfill dam partition design images based on the multidimensional feature vector. Compare the generated design image with the actual design standard to verify the rationality and accuracy of the design.

[0034] Specifically, through automation and intelligent methods, a large amount of manual calculation and drawing work in traditional design has been reduced, significantly shortening the design cycle. Using historical data and machine learning models for prediction, designs can be made based on a large amount of case experience, reducing human errors and improving the accuracy and reliability of the design. By learning the prediction model, key parameters such as dam height and dam slope ratio can be predicted more accurately, thereby optimizing the dam structure and improving the stability of the project. The design scheme can be quickly adjusted according to different initial data and foundation surface parameters to adapt to different engineering environments and requirements. Labor costs and repetitive work in the design process are reduced, and the overall design cost is reduced through intelligent design methods.

[0035] Specifically, if Figure 2 As shown, the process of step S2 includes: Step S21, normalizing the characteristic water level, the earthquake resistance property, the stone material parameter and the historical data to obtain a first processing result; Step S22, performing feature extraction on the first processing result to obtain a first feature result; Step S23, characterizing the first feature result to obtain a characterization result; Step S24, using the characterization result as input of the learning prediction model to predict the dam body height, the upstream dam slope ratio and the downstream dam slope ratio.

[0036] Specifically, the collected characteristic water level, seismic requirements, and rockfill material source parameters are normalized. For continuous numerical data (such as uniaxial compressive strength, porosity, etc.), minimum-maximum normalization or Z-score normalization is used. For categorical data (such as seismic resistance level), one-hot encoding is used for conversion. The normalized data set is recorded as the first processing result. Analyze the normalized data and select key features that have a greater impact on dam design. Key features can be determined by statistical tests, feature importance scoring, etc. The selected key feature set is recorded as the first feature result. Floating point numbers are used to represent key parameters (such as uniaxial compressive strength, porosity, etc.). One-hot vectors are used to represent key categories (such as seismic resistance level). The refined feature representation data set is recorded as the representation result. The representation result is used as the input of the learning prediction model. Select a suitable machine learning algorithm, such as regression model, neural network, etc. Input the representation result into the model for training. Divide the data set into a training set and a test set. Use the training set to train the model. Use the test set to evaluate the model performance and perform model tuning. The trained model is used to predict the new data to obtain the predicted values ​​of the dam height, upstream dam slope ratio, and downstream dam slope ratio.

[0037] Specifically, normalization eliminates the impact of different dimensions and data ranges, allowing the model to better learn data features and thus improve prediction accuracy. Normalized data can speed up model convergence and reduce the time and resources required for calculations. Design predictions based on historical data and machine learning models can reduce human errors and design risks and improve engineering safety.

[0038] Specifically, the process of using the first feature result as the input of the learning prediction model to predict the dam body height, the upstream dam slope ratio and the downstream dam slope ratio includes: Performing correlation analysis on the first feature result to obtain an analysis result; The dam body height and the slope ratio corresponding to the historical data are used as target variables; Using a part of the analysis results and the target variable as a training set, and another part of the analysis results and the target variable as a validation set; Training the learning prediction model according to the training set to obtain a first initial training model; Verifying the initial training model according to the verification set to obtain a first target training model; The dam body height, the upstream dam slope ratio and the downstream dam slope ratio are predicted according to the target training model.

[0039] Specifically, the correlation between features is calculated using methods such as the Pearson correlation coefficient. A correlation matrix is ​​constructed to show the degree of correlation between different features. Feature pairs with high correlation are identified (for example, the correlation coefficient is greater than a certain threshold, such as 0.8). Highly correlated features are deleted or merged to reduce feature redundancy. Principal component analysis is performed on the cleaned feature data. Eigenvalues ​​and eigenvectors are calculated to determine the principal components. Principal components are selected based on the eigenvalues, and usually the principal components with a cumulative explained variance of a certain proportion (such as 95%) are selected. The original feature vector is projected onto the selected principal component to obtain the reduced-dimensional feature vector. The key features after PCA dimensionality reduction are used as the input of the model. The dam height and the slope ratio parameters of the upstream and downstream dam slopes are used as the output of the model. The corresponding dam height, upstream dam slope ratio, and downstream dam slope ratio are extracted from the historical data as target variables. The processed data is divided into training sets and test sets, usually using methods such as 70-30, 80-20, or cross-validation. An integrated learning model is built, and the model components include artificial neural network model, bagging algorithm, random forest, and support vector regression. Determine the structure and parameters of the model, such as the number of neural network layers, the number of neurons, the number of decision trees, etc. The parameter prediction formula is as follows:

[0040] In the formula is the predicted value, is the feature vector of the i-th sample, It is decision trees, K is the total number of all decision trees.

[0041] Specifically, through correlation analysis, the most relevant features to the target variables (dam height and slope ratio) were screened out, reducing the interference of noise and irrelevant variables, thereby improving the prediction accuracy of the model. Using training sets and validation sets for model training and validation, respectively, helps to avoid overfitting and ensure that the model has good generalization ability. The intelligent prediction model can quickly output design parameters, greatly shortening the time required for traditional design methods and accelerating project progress.

[0042] Specifically, the process of calculating according to the dam body height, the upstream dam slope ratio, the downstream dam slope ratio and the foundation surface parameters to obtain the dam body contour parameters includes: The ratio of the dam body height to the upstream dam slope length is taken as the upstream dam slope ratio, the ratio of the dam body height to the downstream dam slope length is taken as the downstream dam slope wave ratio, and the width of the dam body bottom in contact with the foundation is taken as the foundation surface width; The sum of the upstream dam slope length, the downstream dam slope length and the foundation surface width is taken as the dam crest width; The dam body profile parameters are determined according to the dam body height, the dam top width, the upstream dam slope ratio and the downstream dam slope ratio using a geometric method.

[0043] Specifically, the upstream dam slope length (L1) was calculated using the following formula: L1 = H / m1; The downstream dam slope length (L2) is calculated using the following formula: L2 = H / m2; Calculate the crest width (W) using the following formula: W = L1 + L2 + B; Using geometric methods, the dam body profile parameters are calculated based on the dam body height (H), dam crest width (W), upstream dam slope ratio (m1) and downstream dam slope ratio (m2), such as: Upstream dam slope line: the straight line segment from the dam top to the foundation surface, with a length of L1.

[0044] Downstream dam slope line: the straight line segment from the dam top to the foundation surface, with a length of L2.

[0045] Dam top line: A straight line segment parallel to the ground, with a length of B1.

[0046] Bottom line of dam: a straight line segment parallel to the ground, with a length of B2.

[0047] Cross-sectional area of ​​dam body: The area of ​​the cross-sectional area of ​​dam body can be calculated by integration.

[0048] Geometry method examples: Triangle method: treat the cross section of the dam as a triangle, calculate the base and height of the triangle, and then calculate the area of ​​the triangle.

[0049] Trapezoid method: treat the cross section of the dam as a trapezoid, calculate the upper base, lower base and height of the trapezoid, and then calculate the area of ​​the trapezoid.

[0050] Numerical integration method: Divide the cross section of the dam into multiple small units, such as triangles or trapezoids, and then calculate the area of ​​each unit, and finally add up the areas of all units to get the cross-sectional area of ​​the dam. Use existing engineering software, such as AutoCAD or Civil 3D, to calculate the dam contour parameters. You can also use programming languages, such as Python or MATLAB, to write codes to calculate the dam contour parameters, which will not be repeated here.

[0051] Specifically, using geometric methods to calculate the dam body contour parameters can ensure the accuracy of the calculation results and avoid human errors. Using geometric methods to calculate the dam body contour parameters can ensure the accuracy of the calculation results and avoid human errors. By improving design efficiency and accuracy, design costs can be reduced.

[0052] Specifically, the process of predicting the area ratio of the main pile, secondary pile, drainage and increased mold of the dam body according to the dam body contour parameters includes: The dam body is divided into a main pile area, a secondary pile area, a drainage area and an increased mold area according to the dam body contour parameters; Calculate the area of ​​the main pile area, the secondary pile area, the drainage area and the mold increase area according to a geometric method; Comparing the area of ​​each of the regions with the total area of ​​the dam to obtain the area ratio of each sub-region; Collecting the partition area ratio data of the historical data, and analyzing the partition area ratio data to obtain a second analysis result; The second analysis result is used as an input of a prediction model, and the prediction model is trained using the partition area ratio data to obtain the area ratio of each target partition.

[0053] Specifically, the main heap area is the main part of the dam body and plays the main role of water retaining. The secondary heap area is located upstream of the main heap area and plays a transition and anti-seepage role. The drainage area is located upstream of the dam body and is used to drain water and reduce pore water pressure. The mold-increasing area is located downstream of the dam body and is used to improve the stability of the dam body. The dam body is divided into different areas according to the dam body contour and the preset partition boundaries. The dam body can also be divided into grids, and the grids are assigned to different areas according to the location of the grid center point. Each area is divided into multiple small units, such as triangles or trapezoids, and then the area of ​​each unit is calculated. Finally, the area of ​​all units is added to obtain the total area of ​​each area. The area of ​​each area is compared with the total area of ​​the dam body to obtain the area ratio of each area. Collect the partition area ratio data of historical design cases. Analyze the distribution law of historical data, such as mean, variance, maximum and minimum values. Key factors affecting the partition area ratio, such as dam body height, slope ratio, stone parameters, etc. Select a suitable machine learning algorithm, such as regression model, decision tree or neural network. Divide the historical data into training set and test set. Use the training set to train the prediction model, with input features such as dam height, slope ratio, stone material parameters, etc., and output as the partition area ratio. Use the test set to evaluate the performance of the model and perform model tuning. Use the trained prediction model to predict the target partition area ratio based on the new dam height, slope ratio, stone material parameters, etc.

[0054] Specifically, by predicting the area ratio of the partitions, the material configuration can be optimized. By optimizing the material configuration, the cost of materials can be reduced and the profitability of the project can be improved. The stability of the dam can also be improved and the project risk can be reduced. By predicting the area ratio of the partitions, the material composition of the dam can be better understood, thereby optimizing the dam structure and improving the stability of the project.

[0055] Specifically, if Figure 3 As shown, the process of step S4 includes: Step S41, converting the characteristic water level, the seismic resistance attribute and the partition area ratio into numerical features to obtain a conversion result; Step S42, fusing the conversion result and the input image to obtain a fusion result; Step S43, extracting multimodal features from the fusion result to obtain a second feature result; Step S44, normalizing and aligning the second feature result to obtain a second processing result; Step S45: stack the second processing results along the data channel to obtain the multi-dimensional feature vector.

[0056] Specifically, characteristic water levels (such as design flood level, normal water storage level, dead water level), seismic properties (such as earthquake intensity, seismic motion parameters) and partition area ratios are converted into numerical features. For features that can be directly converted to numerical values, such as water levels and area ratios, their numerical values ​​can be used directly. For categorical features, such as seismic resistance level, unique hot encoding can be used to convert them into numerical vectors. For numerical features, normalization methods can be used to scale them to a specific range, such as between 0 and 1. Fuse the numerical features with the input image. Concatenate the numerical feature vector with the image feature vector to form a longer feature vector. Perform weighted summation on the numerical feature vector and the image feature vector to generate a new feature vector. Extract multimodal eigenstates from the fused feature vector. The following methods can be used: Convolutional Neural Networks: Use convolutional neural networks to extract image features such as edges, textures, and shapes.

[0057] Recurrent Neural Networks: Use recurrent neural networks to extract sequence features, such as time series data or text data.

[0058] Use normalization methods, such as L2 normalization, to scale the feature vector to unit length. Use feature alignment methods, such as feature transformation or feature selection, to align features of different modalities. Stack the normalized and aligned multimodal features along the data channel to form a multidimensional feature vector.

[0059] Specifically, by fusing features from different modalities, the performance of the GAN model can be improved. By fusing features from different modalities, the robustness of the GAN model can be improved. By analyzing the impact of different modal features on the generated images, the interpretability of the GAN model can be improved.

[0060] Specifically, if Figure 4 As shown, the process of step S5 includes: Step S51, counting the cross-sectional profile parameters of the input image, the initial data, the coordinate parameters of each of the subareas and the cross-sectional area ratio parameters to obtain the distribution rules of different parameters; Step S52, according to the distribution law, the upstream dam slope ratio, the downstream dam slope ratio, and the area of ​​each region are augmented and expanded by using parameterized adjustment to construct a multi-source heterogeneous training data set and a test data set; Step S53, performing feature data fusion on the training data set and the test data set to construct a corresponding input feature data set and an output feature data set; Step S54, using the input feature data set as an input parameter, and using the training data set to train the generative adversarial network model to obtain a second initial training model; Step S55, verifying the second initial training model according to the verification data set to obtain a second target training model; Step S56, training the model according to the second objective to generate the zoning design image of the concrete face rockfill dam.

[0061] Specifically, calculate the length, curvature and direction of the profile, and record the coordinates of each profile point. Analyze the distribution of characteristic water levels, seismic properties and stone parameters, such as mean, variance, maximum and minimum values. Record the coordinate position of each partition area, such as vertex coordinates and center coordinates. Calculate the area ratio of each partition and compare it with the total area of ​​the dam. According to the statistical distribution law, parameterize the upstream dam slope ratio, downstream dam slope ratio and the area of ​​each area, for example: Random perturbation: Randomly change the slope ratio and area ratio within a certain range to simulate different design options.

[0062] Interpolation and Scaling: Interpolate and scale section profiles to generate images of different sizes and shapes.

[0063] Rotation and Translation: Rotate and translate the image to increase the diversity of the data.

[0064] Data augmentation can be conveniently performed using existing data augmentation tools, such as ImageDataGenerator or albumentations. Fuse the augmented image data with the corresponding initial data (characteristic water level, seismic properties, and stone parameters) to form an input feature dataset. Extract feature vectors from the image using a convolutional neural network or other image feature extraction method. Concatenate the image feature vector with the initial data vector to form an input feature dataset. Fuse the augmented image data with the corresponding partition area ratios to form an output feature dataset. Label each partition area, such as using pixel-level labels or instance segmentation labels. Extract feature vectors from the annotation information, such as the area ratio of each partition. Concatenate the image feature vector with the partition area ratio vector to form an output feature dataset. Select an appropriate generative adversarial network model, such as BigGAN. Select an appropriate loss function, such as L1 loss or L2 loss, and adversarial loss. Train the generative adversarial network model using the training dataset and validate it using the validation dataset. Convert the new input data (characteristic water level, seismic properties, stone parameters, and partition area ratios) into a multidimensional feature vector. Use the trained GAN model to generate panel rockfill dam zoning design images based on the input feature vectors. Evaluate the quality and accuracy of the generated images, for example using manual visual inspection or quantitative metrics.

[0065] Specifically, through parameter adjustment and augmentation, diverse training and test data sets can be created, which helps improve the generalization ability of the model and enables it to handle more diverse design scenarios. By performing feature fusion and parameter adjustment on the data set, the amount of data required for model training can be reduced while improving training efficiency. Data augmentation can increase the diversity of training data, thereby improving the generalization ability of the model. By training the generative adversarial network model, high-quality panel rockfill dam zoning design images can be generated.

[0066] Specifically, the process of augmenting and expanding the upstream dam slope ratio, the downstream dam slope ratio, and the area of ​​each region by using parameterized adjustment according to the distribution law to construct a multi-source heterogeneous training data set and a test data set includes: Determining parameter thresholds of the upstream dam slope ratio, the downstream dam slope ratio and the area of ​​each region; defining the adjustment step length of the upstream dam slope ratio, the downstream dam slope ratio and the area of ​​each region; Establishing a combination rule according to the parameter threshold and the adjustment step size; Generate a series of parameter value combinations according to the combination rule; Performing geometric transformation and simulation on the input image according to the generated parameter value combination to generate a target cross-sectional image; Assigning corresponding parameter values ​​to each of the target images as label information; The training data set and the test data set are constructed according to the target input image and the label information.

[0067] Specifically, determine the minimum (U_min) and maximum (U_max) values ​​of the upstream dam slope ratio (U), for example, U_min=1:1.5, U_max=1:2.5. Determine the minimum (D_min) and maximum (D_max) values ​​of the downstream dam slope ratio (D), for example, D_min=1:1.5, D_max=1:3. Determine the minimum (A_min) and maximum (A_max) values ​​of each area (A), for example, A_min=1000 square meters, A_max=5000 square meters. Set the adjustment step size (U_step) of the upstream dam slope ratio, for example, U_step=0.1. Set the adjustment step size (D_step) of the downstream dam slope ratio, for example, D_step=0.1. Set the adjustment step size (A_step) of the area, for example, A_step=100 square meters. Create a rule to generate all possible parameter combinations according to the parameter thresholds and step sizes of the upstream dam slope ratio, the downstream dam slope ratio, and the area of ​​the area.

[0068] Rule example: For each U value, go from U_min to U_max with a step length of U_step, traverse all D values ​​from D_min to D_max with a step length of D_step, and then for each UD combination, traverse all A values ​​from A_min to A_max with a step length of A_step. Write a script in a programming language (such as Python) to generate a list of parameter value combinations based on the combination rules. For each input image, perform geometric transformations based on the parameter value combinations to simulate different dam slope ratios and regional areas. Use image processing software or custom algorithms to implement these transformations, such as adjusting the shape and size of the image. For each generated target profile image, store its corresponding parameter value (U, D, A) as label information.

[0069] Specifically, the automated parameterization and simulation process greatly reduces the time of manual design, allowing designers to quickly explore multiple design options and optimize the design process. Multi-source heterogeneous data sets contain a variety of possible profile designs, which helps to avoid overfitting during model training and ensures that the model can handle diverse design requirements. By simulating different parameter combinations, the optimal or most unfavorable design parameters can be identified, thereby guiding actual engineering design and improving the quality and safety of the final design.

[0070] Specifically, the process of using the verification data set to verify the second initial training model to obtain the second target training model includes: Creating a binary mask using the profile parameters of the input image, wherein the profile area is marked as 1 and the remaining profile areas are marked as 0; Filtering the output feature according to the mask to obtain a filtering result; The generated image is determined according to the filtering result to be valid only within the panel cross-section contour.

[0071] Specifically, the mask formula is as follows:

[0072]

[0073] In the formula, is the contour mask matrix, is the value of each point in the matrix; There are different feature matrices, namely the main pile, secondary pile, drainage area, mold increase area, and scaling ratio. are the normalized scalar values ​​of different features.

[0074] Specifically, for each input image, the parameters of the profile are first identified, which define the shape and position of the concrete face rockfill dam profile. Using these parameters, a binary mask with the same size as the input image is generated. In this mask, the pixels inside the profile are set to 1, while the pixels outside the profile are set to 0. This can be achieved through image processing techniques such as edge detection and region filling. After the model processes the input image, a series of feature maps are output, which represent the model's understanding of the input image at different levels. These output feature maps are filtered using the created binary mask. The specific operation is to multiply the mask by each feature map pixel by pixel, so that only the features corresponding to the pixels with a value of 1 in the mask will be retained, and the features corresponding to the pixels with a value of 0 will be set to 0.

[0075] Specifically, by focusing only on the area within the profile, the model can more accurately learn design-related features and reduce interference from external noise or irrelevant information. By excluding the area outside the profile, unnecessary calculations are reduced, computational efficiency is improved, and resource consumption is reduced. Ensuring that the generated zoning design image of the face rockfill dam contains only valid information within the profile helps generate a design solution that better meets actual engineering needs.

[0076] Specifically, the process of filtering the output feature according to the mask to obtain a filtering result includes:

[0077] in, is the original output height H, width W, and channel C of the generative network, represents the element-wise multiplication Hadamard product, is the effective generated image after mask filtering, is a binary mask matrix that satisfies: .

[0078] Specifically, the original output of the generative network (height H, width W, channel C) is multiplied element-by-element (Hadamard product) with the binary mask matrix. The result of the multiplication will generate a new image in which only the profile area retains the original output features, and the rest of the area is set to 0. This new image is the effective generated image after mask filtering. Analyze the effective generated image after mask filtering to ensure that the generated image is only valid within the panel profile. According to the analysis results, adjust the parameters of the generative adversarial network model to optimize the quality and accuracy of the generated image to ensure that the generated image meets the design requirements.

[0079] Specifically, through mask filtering, it can be ensured that the generated image is only valid within the panel profile, avoiding interference from irrelevant areas, thereby improving the accuracy and reliability of the generated image. Mask filtering can reduce the amount of image information that the generated network needs to process, thereby reducing the amount of calculation and computing time. The area of ​​the generated image can be controlled by adjusting the type and parameters of the mask, such as generating an area of ​​a specific shape or size.

[0080] like Figure 5 As shown, in this embodiment, a concrete rockfill panel dam in China is used for case study. The seismic base intensity of the project is VII, the design earthquake is 0.305g, and the ultimate seismic capacity is about 0.6g. The main rockfill materials of the project include basalt, tuffaceous basalt, etc. According to the design requirements of the rockfill dam, the seismic design requirements, and the mechanical parameters and volume of the original rock.

[0081] The intelligent design process of a concrete face rockfill dam is as follows: 1. Intelligent prediction of dam contour parameters: The prediction of dam body contour parameters adopts machine learning algorithm, and intelligent mapping generates key parameters of surface zoning design. Among them, the input features are dam height 240m, saturated compressive strength of main rockfill 80 MPa, saturated compressive strength of secondary rockfill 50 MPa, main rockfill source quantity 11.9 million m3 and secondary rockfill source quantity 6.4 million m3, a total of 5 features; the output features are upstream and downstream slope ratio of 1.4, and area ratio of secondary rockfill: drainage: main rockfill: augmentation mold = 0.24:0.12:057:0.07.

[0082] 2. Intelligent generation design of rockfill dam profile partitions: (1) Input data construction: According to the data feature expression method of specific implementation methods 1 and 2, text-image feature fusion is realized. In the design of the concrete rockfill panel dam, the input image feature is the outer contour of the rockfill dam section determined according to the design requirements, and the input text feature is the normalized area of ​​the main rockfill, secondary rockfill, drainage, and mold augmentation areas of the project (secondary rockfill: drainage: main rockfill: mold augmentation = 0.24:0.12.:057:0.07).

[0083] (2) Intelligent design: A generative adversarial network model is constructed according to the method in Specific Implementation Method 2, and the intelligent generative model is trained and tested. After the trained model receives the image-text input data features, it quickly generates an intelligent design result expressed in pixels, which includes the layout position and size of the main pile, secondary pile, drainage and mold augmentation area in the profile of the rockfill dam.

[0084] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

[0085] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An intelligent design method for a concrete face rockfill dam profile based on data feature fusion, characterized in that: include: Step S1, collecting initial data and foundation surface parameters required for the design of a concrete face rockfill dam, wherein the initial data includes characteristic water level, seismic properties and stone material parameters, and also collecting historical data corresponding to the initial data; Step S2, constructing a learning prediction model according to the initial parameters and the historical data to predict the dam body height, upstream dam slope ratio and downstream dam slope ratio of the rockfill dam; Step S3, calculating according to the dam body height, the upstream dam slope ratio, the downstream dam slope ratio and the foundation surface parameters to obtain dam body contour parameters, and predicting the target partition area ratios of the main stacking area, the secondary stacking area, the drainage area and the mold increasing area according to the dam body contour parameters; Step S4, drawing an input image according to the dam body contour parameters, and fusing the input image, the characteristic water level, the seismic properties and the area ratios of each target partition to obtain a multi-dimensional feature vector; Step S5, using the multi-dimensional feature vector as a model input to construct a generative adversarial network model, and generating a panel rockfill dam zoning design image according to the generative adversarial network model.

2. The method for intelligent design of a concrete face rockfill dam profile based on data feature fusion according to claim 1 is characterized in that: The process of step S2 includes: Normalizing the characteristic water level, the earthquake resistance property, the stone material parameter and the historical data to obtain a first processing result; Performing feature extraction on the first processing result to obtain a first feature result; Characterizing the first characteristic result to obtain a characterization result; The characterization results are used as inputs of the learning prediction model to predict the dam body height, the upstream dam slope ratio and the downstream dam slope ratio.

3. The method for intelligent design of a concrete face rockfill dam profile based on data feature fusion according to claim 2 is characterized in that: The process of using the first feature result as the input of the learning prediction model to predict the dam body height, the upstream dam slope ratio and the downstream dam slope ratio includes: Performing correlation analysis on the first feature result to obtain an analysis result; The dam body height and the slope ratio corresponding to the historical data are used as target variables; Using a part of the analysis results and the target variable as a training set, and another part of the analysis results and the target variable as a validation set; Training the learning prediction model according to the training set to obtain a first initial training model; Verifying the initial training model according to the verification set to obtain a first target training model; The dam body height, the upstream dam slope ratio and the downstream dam slope ratio are predicted according to the target training model.

4. The method for intelligent design of a concrete face rockfill dam profile based on data feature fusion according to claim 3 is characterized in that: The process of calculating according to the dam body height, the upstream dam slope ratio, the downstream dam slope ratio and the foundation surface parameters to obtain the dam body contour parameters comprises: The ratio of the dam body height to the upstream dam slope length is taken as the upstream dam slope ratio, the ratio of the dam body height to the downstream dam slope length is taken as the downstream dam slope ratio, and the width of the dam body bottom in contact with the foundation is taken as the foundation surface width; The sum of the upstream dam slope length, the downstream dam slope length and the foundation surface width is taken as the dam crest width; The dam body profile parameters are determined according to the dam body height, the dam top width, the upstream dam slope ratio and the downstream dam slope ratio using a geometric method.

5. The method for intelligent design of a concrete face rockfill dam profile based on data feature fusion according to claim 4 is characterized in that: The process of predicting the area ratio of the main pile, secondary pile, drainage and increased mold of the dam body according to the dam body contour parameters includes: The dam body is divided into a main pile area, a secondary pile area, a drainage area and an increased mold area according to the dam body contour parameters; Calculate the area of ​​the main pile area, the secondary pile area, the drainage area and the mold increase area according to a geometric method; Comparing the area of ​​each of the regions with the total area of ​​the dam to obtain the area ratio of each sub-region; Collecting the partition area ratio data of the historical data, and analyzing the partition area ratio data to obtain a second analysis result; The second analysis result is used as an input of a prediction model, and the prediction model is trained using the partition area ratio data to obtain the area ratio of each target partition.

6. The method for intelligent design of a concrete face rockfill dam profile based on data feature fusion according to claim 5 is characterized in that: The process of step S4 includes: Converting the characteristic water level, the seismic resistance attribute and the partition area ratio into numerical features to obtain a conversion result; Fusing the conversion result and the input image to obtain a fusion result; Extracting multimodal features from the fusion result to obtain a second feature result; Performing normalization and feature alignment processing on the second feature result to obtain a second processing result; The second processing results are stacked along the data channel to obtain the multi-dimensional feature vector.

7. The method for intelligent design of a concrete face rockfill dam profile based on data feature fusion according to claim 6 is characterized in that: The process of step S5 includes: Counting the cross-sectional profile parameters of the input image, the initial data, the coordinate parameters of each of the subareas and the cross-sectional area ratio parameters to obtain the distribution rules of different parameters; According to the distribution law, the upstream dam slope ratio, the downstream dam slope ratio, and the area of ​​each region are augmented and expanded by parameterized adjustment to construct a multi-source heterogeneous training data set and a test data set; Performing feature data fusion on the training data set and the test data set to construct a corresponding input feature data set and an output feature data set; Taking the input feature data set as an input parameter, and using the training data set to train the generative adversarial network model to obtain a second initial training model; Verifying the second initial training model according to the verification data set to obtain a second target training model; The model is trained according to the second objective to generate the zoning design image of the concrete face rockfill dam.

8. The method for intelligent design of a concrete face rockfill dam profile based on data feature fusion according to claim 7 is characterized in that: The process of augmenting and expanding the upstream dam slope ratio, the downstream dam slope ratio, and the area of ​​each region by using parameterized adjustment according to the distribution law to construct a multi-source heterogeneous training data set and a test data set includes: Determining parameter thresholds of the upstream dam slope ratio, the downstream dam slope ratio and the area of ​​each region; defining the adjustment step length of the upstream dam slope ratio, the downstream dam slope ratio and the area of ​​each region; Establishing a combination rule according to the parameter threshold and the adjustment step size; Generate a series of parameter value combinations according to the combination rule; Performing geometric transformation and simulation on the input image according to the generated parameter value combination to generate a target cross-sectional image; Assigning corresponding parameter values ​​to each of the target images as label information; The training data set and the test data set are constructed according to the target input image and the label information.

9. The method for intelligent design of a concrete face rockfill dam profile based on data feature fusion according to claim 8 is characterized in that: The process of using the verification data set to verify the second initial training model to obtain the second target training model includes: Creating a binary mask using the profile parameters of the input image, wherein the profile area is marked as 1 and the remaining profile areas are marked as 0; Filtering the output feature according to the mask to obtain a filtering result; The generated image is determined according to the filtering result to be valid only within the panel cross-section contour.

10. The method for intelligent design of a concrete face rockfill dam profile based on data feature fusion according to claim 9, characterized in that: The process of filtering the output feature according to the mask to obtain a filtering result includes: in, is the original output height H, width W, and channel C of the generative network, represents the element-wise Hadamard product, is the effective generated image after mask filtering, is a binary mask matrix that satisfies: .

Images