Abstract
Energy flexibility refers to the ability of a building or neighbourhood to activate its on-site flexibility sources (building’s thermal mass, heat storage tanks, batteries, EV charging) while safeguarding user needs and comfort. FLEXor is an optimization tool for flexible energy use, storage and generation in buildings, that generates optimal load profiles in response to grid signals, such as dynamic energy prices, thus facilitating energy planning and optimal system dimensioning and control It comprises interconnected models that handle user inputs, activate sub-models, and define system topology. These include component models (building envelope, hot water tank, heat sources, EV charging, PV systems and electric batteries), functional models (energy costs, fixed energy demand), and a model for typical demand profiles It is implemented in Python and uses the Pyomo software package for formulating, solving and analyzing optimization models.The tool is available in a backend solution, run on a SINTEF server, while their frontend is accessible in two ways: via a Web App, and via API (Application Programming Interface) for more advanced users and for use by other software. FLEXor is a three-level model. The top model reads the user input, collects and organizes the input parameters and input data, sets up and activates the second- and third-level models, and collects and organizes the results. This model is not interchangeable. The second-level model sets up the system topology in the building used to define the connections between the energy components in the building. This model is interchangeable. The third-level models have several different purposes. These include component models, such as the building envelope, domestic hot water (DHW) tanks, electric batteries, heat sources (e.g. direct electric heaters, district heating, and heat pumps), electric vehicle charging, and onsite PV systems; functional models to calculate energy costs, fixed energy demand, the Linear Time-Invariant (LTI) model structure; and a model to calculate energy demand profiles via PROFet. The starting point is given by typical (non-flexible) energy demand load profiles taken from PROFet, based on a statistical analysis of hourly measurements from several buildings classified in different categories. The flexibility sources are modelled as internal variables (the model’s states) such as indoor temperature, tank temperature, battery state of charge, and are subject to boundary conditions and constraints that represent user comfort and user needs, such as a comfort band for indoor temperature, a lower bound for the hot water tank's temperature, and the charging of electric vehicles within the connection time and capacity. All the component models (third-level) in FLEXor are designed to be self-standing. Thus, they are self-contained, and do not include the control and/or optimization of other components. The models are designed to be i) linear, ii) in state space form (when applicable), and iii) transparent. This allows the high-level model to be fast, lean, relatively simple, and able to leave a component out of the optimization process if necessary.