Évènements

Modeling for control of Multi Energy Systems based on a physical description in Modelica: application to a city district

Date : 13/04/2021

Tuesday 13th April at 14:00 via Microsoft Teams at
Click here to join the meeting
Learn More | Meeting options

Speaker: Joy El Feghali, PhD student at Paris-Saclay University

Title: Modeling for control of Multi Energy Systems based on a physical description in Modelica: application to a city district

Abstract: The environmental pressure associated with the development of control possibilities of energy networks implies a great research effort on the production field side to allow a large renewable energy penetration and limited use of fossil energy production means. In parallel, from the consumption side, optimization at the utility level (residential, tertiary, and industrial sectors) can lead to an increase in the global energy efficiency and a decrease (or at least a better temporal dispatch) of the energy needs. The energy transition requires, therefore, more flexibility both on the production and consumption sides. This need has to be supported by new control systems.
In contrary to day-ahead optimizations for which it is possible to use steady-state or quasi-steady-state models, we are interested in shorter time horizons to investigate the contribution to ancillary services. In that case, dynamical models well-suited to time horizons from a second to an hour are required. These dynamical models are still difficult to obtain. Indeed, if we choose an on-site identification strategy, we are faced with nonlinearities and coupling aspects. On the opposite, starting from physical modeling leads to very complex and often untractable models, which cannot be used for control purposes and control law design.
This talk presents a methodology to reduce the complexity of multi-energy networks’ models, modeled with Modelica language, using an energy-based reduction method that simplifies components with small energy contributions. This methodology relies on using the Modelica language, an object-oriented modeling language, and the tools developed in particular in the EDF R&D. This method is tested on a model of a district located in the suburbs of Paris.

Bio: Joy El Feghali received a MSc. in Control, Signal and Image processing from Paris-Saclay University in 2019. She started her Ph.D. in December 2019 at Paris-Saclay University in collaboration with EDF R&D. Her research interests are model reduction methods of complex systems modeled with Modelica language and control techniques for multi-energy systems.