Διδακτορικές Διατριβές
Μόνιμο URI για αυτήν τη συλλογήhttps://repository.ihu.gr/handle/11544/47642
Περιηγούμαι
Πλοήγηση Διδακτορικές Διατριβές ανά Συγγραφέα "Παπαδοπούλου, Σημίρα"
A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
Α Β Γ Δ Ε Ζ Η Θ Ι Κ Λ Μ Ν Ξ Ο Π Ρ Σ Τ Υ Φ Χ Ψ Ω
Τώρα δείχνει 1 - 1 από 1
- Αποτελέσματα ανά σελίδα
- Επιλογές ταξινόμησης
Τεκμήριο Optimal energy management strategies for autonomous and interconnected microgrids with renewable energy production and storage(ΔΙΠΑΕ, 2025-10-31) Τρίγκας, Δημήτριος; Παπαδοπούλου, Σημίρα; Σχολή Μηχανικών, Τμήμα Μηχανικών Παραγωγής και ΔιοίκησηςThis thesis presents an integrated study on the energy management of autonomous multi-node microgrids, developed and analyzed using as a basis a real small-scale experimental installation. The primary objective is to design and demonstrate supervisory control strategies that maximize renewable energy utilization, minimize dependence on fossil-based backup units and ensure continuous supply to all loads. The study addresses the coordination of distributed renewable generation, energy storage, hydrogen systems and flexible demand through hierarchical energy management, with emphasis on the development and application of advanced predictive control frameworks. The research adopts a hierarchical control structure, with node-level strategies balancing renewable generation and storage and supervisory strategies coordinating energy exchanges across nodes. The first supervisory method explored is based on a Finite State Machine (FSM). By discretizing battery’s state of charge (SOC) into operational zones and defining transitions corresponding to import, export, or autonomous modes, the FSM provides a transparent and computationally efficient way to manage power exchanges. Results show that FSM-based coordination reduces renewable curtailment and backup systems utilization compared to isolated node operation. However, as a rule-based approach, it cannot anticipate future fluctuations or optimize decisions over time horizons longer than the present operating step. To overcome these limitations, the thesis develops a Nonlinear Model Predictive Control (NMPC) framework, which constitutes its central contribution. NMPC relies on forecasts of renewable generation and load profiles to optimize energy exchanges among nodes over a prediction horizon. By explicitly including the operational limits of the energy subsystems, NMPC ensures feasible operation while minimizing fossil fuel usage and renewable curtailment. Comparative studies demonstrate that isolated operation of each node, while capable of maintaining self-sufficiency, often results in inefficient use of storage resources and higher diesel reliance. NMPC coordination, by contrast, enables surplus renewable energy to be absorbed by neighboring nodes or storage systems and allows deficits to be shared across the network, significantly improving the efficiency. Building on this framework, the thesis explores the refinement of NMPC through the tuning of weight coefficients. These weights balance the importance assigned to SOC deviation from reference trajectories and control effort in the optimization problem. By employing a forecast-driven tuning mechanism, the NMPC adapts to changing operating conditions and improves flexibility in the management of renewable variability. Although the transition from local rule-based management to NMPC coordination represents the primary advance, the incorporation of weight tuning illustrates how predictive controllers can be further enhanced to achieve even greater efficiency. The thesis also examines the integration of electric vehicles (EVs) as flexible loads in the multi-node microgrid. Uncontrolled charging increases demand peaks and increases fossil fuel dependence, while optimal charging scheduling aligns EV demand with renewable availability. The results demonstrate that this approach reduces peak loads and improves renewable absorption, turning EVs into controllable resources that actively support system efficiency and stability. Finally, the study develops a digital twin of the experimental microgrid, designed as a digital replica that links operational data with a virtual representation. Except of providing monitoring capabilities, the digital twin enables predictive scenario analysis, that allows alternative control strategies and future conditions to be tested before deployment. This functionality enhances resilience, improves decision-making and complements the predictive supervisory control framework. The results of the thesis demonstrate that predictive supervisory management substantially improves the operation of autonomous multi-node microgrids. FSM-based control offers a first level of coordination, but NMPC provides a decisive improvement by proactively scheduling distributed resources and enabling effective energy exchange among nodes. Compared with isolated operation, NMPC achieves higher renewable utilization, reduced diesel usage and more balanced exploitation of both short and long-term storage. The addition of weight tuning further strengthens flexibility, while the integration of EVs and the digital twin expand the functionality and resilience of the system. All these contributions establish a comprehensive framework for the optimal and sustainable management of future distributed energy systems.
