Building smart grid : optimal coordination of consumption with decentralized energy generation and storage

This work describes an implementation of an office and/or personal smart grid for environmentally friendly buildings. These can be equipped with a local energy source (e.g. photovoltaic panels or combined heat-power units), energy storage devices (batteries, electric hot water boilers, heating and ventilation systems including air conditioning), a building energy management system with sensors (e.g. providing the room temperatures) and household appliances acting as actuators (in general, split into groups of schedulable and non-schedulable ones). The idea behind this work has been to develop an automatic control system which would optimally decide for the end-customer when to buy, sell or store electric energy with the objective to minimize his total costs. At the same time, it fulfills all constraints in terms of the limits on power allowed to be taken from the grid. For the user, besides the optimal scheduling of household appliances and selling energy to the grid whenever it becomes profitable, the developed prototype gives also the opportunity to monitor the local generation, storage and consumption in real time. On the other side, the electric utility can produce a real-time pricing signal which reflects daily peaks in consumption. Making such a price signal available for this building energy management system will lead to a decrease of power consumption from the grid. In this project, a model predictive control approach to the energy optimization problem in a building has been proposed. Different scenarios have been run and the results are discussed here