In this paper, we focused on the advancement of Dam Monitoring Software that incorporates the Finite Element Method (FEM), as these large infrastructure constructions are crucial for ensuring a dependable water supply, irrigation, flood control, renewable electric energy generation, and safe operation, which is of utmost importance to any country. However, the material properties and geotechnical environments of dams can change (deteriorate) over time, while the standards and legal norms that govern them become more and more rigorous, so in order to accurately assess the state of a dam and detect any concerning behavior, the software must be updated as well. The custom-developed FEM solver, unlike many commercial alternatives, is adaptable and can be reconfigured to function within a Dam Monitoring System. In this paper, we present the procedure for interpolating numerical values at measurement points, when the position of the measurement point does not align with the node of the element, allowing for additional instrument locations to be added to the monitored system without the need for remeshing the numerical model. This procedure is used to compare the actual pore pressures and temperature values of the concrete dam structure with the prediction of the numerical model, and the agreement is much greater with the new interpolation algorithm in comparison to the nearest nodal values, with the average relative difference for pore pressure reduced from 8.89% to 8.10%, justifying this implementation.

Dam monitoring data are nonlinear and nonstationary time series. Most existing data-driven dam displacement models are developed independently for each measuring point, disregarding the fact that a dam is a complex structure composed of various interconnected elements that form a unified whole. Regardless of the dam type, all points on the dam are exposed to the same external environmental influences. To account for the correlation between displacement time series at different points, this paper proposes a novel fine-tuned deep-learning nonlinear autoregressive (NAR) model based on a Long Short-Term Memory (LSTM) network for predicting dam tangential displacement, and a new method for generating source data to train the base model. The models for three measuring points were developed and tested on experimental data collected over a period of slightly more than twelve years. Compared with the model without fine-tuning, the proposed approach achieves an average mean square error (MSE) reduction of 80.68% on the training set and 65.79% on the test set, as well as an average mean absolute error (MAE) reduction of 51.05% and 52.62%, respectively. Furthermore, the proposed model outperforms Random Forest (RF), Support Vector Regression (SVR), and Multi-Layer Perceptron (MLP) models for dam displacement prediction.

Complex hydropower systems are vital infrastructure in modern societies. Because of the risks associated with the operation of hydropower plants and dam structures, it is necessary to assess the dynamic resilience of these systems by considering the effects of faulty equipment and/or extreme environmental events. Common approach is to evaluate the performance of system components when exposed to a set of scenarios, combining various equipment faults in adverse conditions (e.g. flooding, earthquake). The evaluation is usually carried out with specific needs: (1) establishing design criteria; (2) providing real-time operation of the hydropower system. However, these specific needs may limit the application in complex circumstances, as they require repetition of evaluation system performance with a widespread of input datasets. The authors propose a novel methodology for interactive dynamic resilience assessment based on machine learning techniques and big data. The methodology is based on producing an extensive computational dataset for various scenario assessment and providing a machine learning-based tool for rapid search within the dataset previously simulated. These scenarios are then evaluated interactively, without subsequent simulations, against user-defined resilience metrics. This way, an operation manager may evaluate the dynamic resilience of the hydropower system for various operational constraints in timely manner. The Pirot hydropower system located in Serbia is analysed and findings are presented. The results indicate that the proposed search methodology enhances the traditional simulation model for dynamic resilience assessment by reducing the time required for series of simulation runs, thereby enabling faster improvements in system resilience.