A Multi-Agent system for Smart Grid simulations: exploiting flexibility from thermal behaviour of residential buildings

In order to orientate inside the vastness of literature concerning the Smart Grid world, simulation tools will occupy the enabling role for testing different scenarios and strategies. It is needed to have the easier to use and more complete tools to assess at best the analysis of what new technologies and policies could represent in the electric systems of the future. Flexible, modular and scalable systems must be preferred in order to avoid extreme sectorialisation of the research. The ability to assess multiple different issues while researching, and doing that with simplicity and the desired level of detail, will be a key feature for simulation systems. This thesis aims at contributing in the preliminary stages of designing and testing for a Smart Grid agent-based simulator. Micro-services and Object Orient Programming concepts, coupled with MultiAgent-based modelling, will ensure the right flexibility to the tool. The main contributions of this thesis project are, the creation of the framework from scratch and the implementation and testing of two main modules that will assess different aspects of the grid. The first one will be used to study and OPF based solution for balancing real-time power values at the primary substation of a distribution net. The second one will assess the flexibility management of residential buildings in the heating season by exploiting their thermal performances and behaviours. These two modules are encapsulated in a framework that, as a first prototype, will lay the foundations of a much more complete and flexible simulation tool, that will be able to assess multiple strategies concerning the various fields of the Smart Grid world.