Search

Ground Penetrating Radar (GPR) is a non-destructive geophysical technique utilizing electromagnetic pulses to detect subsurface material properties. The analysis of regions of interest (ROIs) in GPR images often entails the identification of hyperbolic reflection regions of underground targets through accurate segmentation, a crucial preprocessing step. Currently, this represents a research gap. In the hyperbolic reflection region, manual segmentation not only demands professional expertise but is also time-consuming and error-prone. Automatic segmentation can aid in accurately determining the location and depth of the reflection region, thereby enhancing data interpretation and analysis. This study presents a deep residual Convolutional Neural Network (Res-CNN) that integrates skip connections within an encoder-decoder framework for ROI-binarized hyperbolic segmentation. The proposed framework includes designed downsampling and upsampling modules that facilitate feature computation sharing between these two modules through skip connections within network blocks. In the evaluation of both simple and complex models, our method attained PSNR, SSIM, and FSIM values of 57.1894, 0.9933, and 0.9336, and 58.4759, 0.9958, and 0.9677, respectively. Compared to traditional segmentation methods, the proposed approach demonstrated clearer segmentation results, enabling intelligent and effective identification of the ROI region containing abnormal hyperbolic reflection waves in GPR images. [ABSTRACT FROM AUTHOR]

Data assimilation for unconfined seepage analysis has faced significant challenges due to hybrid causes, such as sparse measurements, heterogeneity of porous media, and computationally expensive forward models. To address these bottlenecks, this paper introduces a physics-informed neural network (PINN) model to resolve the data assimilation problem for seepage analysis of unsaturated earth–rockfill dams. This strategy offers a solution that decreases the reliance on numerical models and enables an accurate and efficient prediction of seepage parameters for complex models in the case of sparse observational data. For the first attempt in this study, the observed values are obtained by random sampling of numerical solutions, which are then contributed to the synchronous constraints in the loss function by informing both the seepage control equations and boundary conditions. To minimize the effects of sharp gradient shifts in seepage parameters within the research domain, a residual adaptive refinement (RAR) constraint is introduced to strategically allocate training points around positions with significant residuals in partial differential equations (PDEs), which could facilitate enhancing the prediction accuracy. The model's effectiveness and precision are evaluated by analyzing the proposed strategy against the numerical solutions. The results indicate that even with limited sparse data, the PINN model has great potential to predict seepage data and identify complex structures and anomalies inside the dam. By incorporating coupling constraints, the validity of our PINN model could lead to theoretically viable applications of hydrogeophysical inversion or multi-parameter seepage inversion. The results show that the proposed framework can predict the seepage parameters for the entire research domain with only a small amount of observation data. Furthermore, with a small amount of observation data, PINNs are able to obtain more accurate results than purely data-driven DNNs. [ABSTRACT FROM AUTHOR]

For better stability, safety and water resource management in a dam, it is important to evaluate the amount of seepage from the dam body. This research is focused on machine learning approach to predict the amount of seepage from Pakistan’s Earth and rock fill Tarbela Dam during 2003 to 2015. The data of temperature, rainfall, water inflow, sediment inflow, reservoir level collected during 2003 to 2015 served as input while the seepage from dam during this period was the output. Artificial Neural Network (ANN), Random Forest (RF), Support Vector Machine (SVM), and CatBoost (CB), have been used to model the input-output relationship. The algorithms used to predict the dam seepage reported a high R2 scores between actual and predicted values of average seepage, suggesting their reliability in predicting the seepage in the Tarbela Dam. Moreover, the CatBoost algorithm outperformed, by achieving an R2 score of 0.978 in training, 0.805 in validation, and 0.773 in testing phase. Similarly, RMSE was 0.025 in training, 0.076 in validation, and 0.111 in testing phase. Furthermore, to understand the sensitivity of each parameter on the output (average seepage), Shapley Additive Explanations (SHAP), a model explanation algorithm, was used to understand the affect of each parameter on the output. A comparison of SHAP used for all the machine learning models is also presented. According to SHAP summary plots, reservoir level was reported as the most significant parameter, affecting the average seepage in Tarbela Dam. Moreover, a direct relationship was observed between reservoir level and average seepage. It was concluded that the machine learning models are reliable in predicting and understanding the dam seepage in the Tarbela Dam. These Machine Learning models address the limitations of humans in data collecting and analysis which is highly prone to errors, hence arriving at misleading information that can lead to dam failure.

Subsurface mapping can be implemented by the migration imaging of ground penetrating radar (GPR) data. Finite difference (FD)-based migration can effectively achieve fast imaging in media with lateral heterogeneity. However, this migration inherently suffers from insufficient accuracy in determining the true morphology of any steep-obliquity interface. In this letter, we propose a new algorithm based on crossed differential operator (CDO) and reverse-time migration (RTM) strategy to refocus the reflections back into its precise location. By employing the implicit Crank–Nicolson scheme without any operator approximations, wave field can be extrapolated nearly along the characteristic directions of wave propagation based on the nonapproximate equations. We determined the potential of the algorithm using synthetic and field data. For synthetic data, we consider a multigeometry scenario, including different angle interfaces, different-sized spheres, and lateral heterogeneity. In particular, we focus on extrapolation accuracy, amplitude–frequency focusing ability, entropy analysis, and imaging errors. In the field data test, we found that the finer morphology of the steeper-obliquity interfaces can be determined, the interpretability and amount of discernible details can also be improved. This improvement makes CDO-based migration a very promising tool for the finer imaging of shallow subsurface. [ABSTRACT FROM AUTHOR]