Page d'accueil // SnT // News & E... // PhD Defense: Enhancing Smart Grid Resilience and Reliability by Using and Combining Simulation and Optimization Methods

PhD Defense: Enhancing Smart Grid Resilience and Reliability by Using and Combining Simulation and Optimization Methods

twitter linkedin facebook email this page
Add to calendar
Conférencier : Nikolaos Antoniadis (SerVal Research Group)
Date de l'événement : mercredi 26 mai 2021 14:00 - 17:00

Please click on this link and join the online PhD defense.

Members of the defense committee:

  • Prof. Dr. Nicolas NAVET, University of Luxembourg, Chairman
  • Prof. Dr Olivier BARAIS, University of Rennes 1, France, Vice-chairman
  • Prof. Dr Yves LE TRAON, University of Luxembourg, Supervisor
  • Prof. Dr Sylvain KUBLER, Université de Lorraine, France, Member
  • Dr. Maxime CORDY, University of Luxembourg, Member



Modern electrical grids include numerous digital technologies for producing, transmitting, distributing, and supplying electricity. The electrical grids that achieve the most reliable, efficient, and less environmental impact operation using the above technologies combined with renewable energy sources are characterized as smart grids. The study of electricity networks aims at the continuous and uninterrupted production, transmission, and distribution of electricity under the safest operating conditions. Therefore, an electrical grid is designed and studied in a multifaceted way to highlight the weaknesses and reduce possible disturbances.

One of the most critical disturbances that can occur in energy grids is overload. Overloads on an electrical system are dangerous, as they can cause overheating or an electric arc. Cables in an electrical grid have a maximum ampacity, i.e., current capacity, that can safely flow. If an excessive number of devices, such as electric vehicles, are connected to a circuit, the electrical current will overheat the cables. If the cable insulation melts, an electric arc can be generated and cause a fire in the overheating area, even inside a wall. In order to avoid overloads, fuses are installed in the circuits. If the current exceeds a specific value, the fuse is activated, drops, and opens the circuit, thus interrupting electricity flow. However, even if they are below the safety limits, sustained overloads could also damage the wires. Smart grid operators could change the state of each grid's fuse or could remotely curtail the over-producing/over-consuming users so that, with the minimum interruption, any potential overload could be prevented. Nevertheless, making the most appropriate decisions is a complicated decision-making task, mainly due to contractual and technical obligations.

The present dissertation studies the overloading prevention problem in terms of smart grids' reliability and resilience and evaluates real-world topology in a Luxembourg city district. To this end, it suggests solution methods that can suggest optimal countermeasures to operators facing potential overloading incidents. Specifically, the dissertation has three main axes:

The first axis regards the deterministic overloading prevention problem. Given the topology and the energy data of a microgrid at the current time, the potential overloading incidents are detected, and the optimal countermeasures are calculated for the next measurement interval. The grid operators can apply the proposed actions to recover the grid from the disturbance. Into the thesis, the problem is defined and formulated as a Multiobjective Mixed Integer Quadratically Constrained Program. The dissertation also suggests a solution method using a combinatorial optimization approach with a state-of-the-art exact solver.

The second axis focuses on reliability analysis through simulation after a potential overloading incident. Smart grid operators would be of great use to ensure stability after a potential overload for a planning horizon, as the future electrical values are unknown. To evaluate the robustness of the topology reconfiguration after a disturbance, like an overload, reliability analysis through simulation is employed.

The third axis proposes the single-stage stochastic overloading prevention problem. It differs from the deterministic problem as the optimal countermeasures are calculated for a measurement horizon, e.g. 24 hours. The dissertation defines the corresponding single-stage stochastic program and proposes a simheuristic method to solve it.

Overall, this thesis presents a fully-edge study on reliability optimization for smart grids to provide the appropriate countermeasures after a potential overloading disturbance. The present approach has been developed in collaboration with an industrial partner and evaluated on real-world topology.