In load rejection transitional processes in pumped-storage plants (PSPs), the process of closing pump turbines, including guide vane (GVCS) and ball valve closing schemes (BVCS), is crucial for controlling pulsating pressures and water hammer. Extreme pressures generated during the load rejection process may result in fatigue damage to turbines, and cracks or even bursts in the penstocks. In this study, the closing schemes for pump turbine guide vanes and ball valves are optimized to minimize water hammer and pulsating pressures. A model is first developed to simulate water hammer pressures and to estimate pulsating pressures at the spiral case and draft tube of a pump turbine. This is combined with genetic algorithms (GA) or non-dominated sorting genetic algorithm II (NSGA-II) to realize single- or multi-objective optimizations. To increase the applicability of the optimized result to different scenarios, the optimization model is further extended by considering two different load-rejection scenarios: full load-rejection of one pump versus two pump turbines, simultaneously. The fuzzy membership degree method provides the best compromise solution for the attained Pareto solutions set in the multi-objective optimization. Employing these optimization models, robust closing schemes can be developed for guide vanes and ball valves under various design requirements.

Reconstructing transient states presents significant challenges, particularly within complex pipe networks. These challenges arise due to nonlinear behaviours, inherent uncertainties in the system, and limitations in data availability. This work proposed a novel approach employing Physics-Informed Neural Networks (PINN) to reconstruct transient states in pipe networks, even with limited sensor data. To integrate the complex topology of pipe network systems into neural networks, the method integrates the PINN framework with an efficient elastic water column (EWC) model which can be simply formulated across diverse pipe network configurations. The results showed the proposed PINN method can accurately reconstruct the pressure and flow variation at unmonitored locations, even provided with noisy data at a limited number of locations. One of its advantages lies in its ability to effectively capture extreme values that hold potential significance for pipe infrastructure, providing a promising avenue for pipe failure analysis and enhanced safety management. Laboratory experiments have also been conducted to validate the efficacy and reliability of this method, thus further underlining its potential for real-world applications.

Real-time burst localization in water distribution systems is essential for minimizing water loss, infrastructure damage, and service disruptions. However, traditional methods often struggle with scalability, accuracy, and responsiveness, particularly in large-scale networks. This study proposes a novel two-stage Ensemble Kalman Filter (EnKF) framework integrated with an efficient Elastic Water Column (EWC) model for transient-based burst localization. The EWC model reduces computational complexity by requiring fewer discretization points, significantly lowering the dimensionality of the state space. Moreover, a two-stage localization strategy is introduced: the first stage identifies the likely burst pipes, and the second stage pinpoints the precise burst location. This hierarchical approach further improves both computational efficiency and localization accuracy in complex networks. Numerical case studies on networks of different scales demonstrate the method's robustness, convergence, and accuracy, indicating its potential for real-time and accurate burst localization in practical water distribution systems.

Hydraulic transients pose significant risks to pipeline systems, including potential leaks and bursts. While traditional numerical methods are widely used for transient analysis, their reliance on detailed initial/boundary conditions limits their applicability in real-world scenarios. To overcome these challenges, this paper proposes a novel Filtering Physics-Informed Neural Operator (FPINO) framework for accurate and near real-time reconstruction of hydraulic transients from sparse observations. FPINO extends recent developments in Physics-Informed Neural Operators (PINOs), a class of machine learning models that incorporate physical laws directly into operator learning, to transient hydraulic applications. The FPINO framework introduces a time-weighted information filtering strategy that prioritizes the relevant observations, thereby accelerating convergence and improving prediction accuracy. Additionally, FPINO employs a hybrid loss function that integrates physics-based constraints with data-driven learning, ensuring both physical fidelity and adaptability to complex system behaviors. Numerical and experimental results show that FPINO outperforms the baseline PINO approach in reconstructing diverse transient scenarios. These results highlight the potential of FPINO as a robust and practical tool for real-time monitoring and analysis of hydraulic transients in pipeline systems.