Toward nearly Zero Energy Buildings through Electrical Storage and Mathematical Optimization

Abstract

With the nearly Zero Energy Buildings (nZEBs) EU Directive (2010/31/EU) currently in force, all new buildings shall simultaneously reduce their primary energy consumption (energy from utility grids) and increase their energy share from Renewable Energy Sources (RES). Based on the fact that nZEBs are commonly addressed during their design and construction phase, this Thesis proposes a novel mathematical optimization approach, which attempts to maintain a low import and export energy profile (i.e., net grid electrical energy), daily, in an adaptive manner, and hence, allowing the building to further reduce its primary energy consumption throughout the year. For this purpose, a Linear Programming (LP) model is developed for allowing the battery’s optimum daily dispatch. The LP model is assisted by tools such as Artificial Neural Networks (ANN) for forecasting the next day’s hourly load consumption and Photovoltaic (PV) generation, and Genetic Algorithm (GA) for optimally driving LP and maintaining the building’s daily net grid electrical energy as close to zero as possible (i.e., nearly zero). Moreover, for addressing the non-linear and complex nature of the battery the proven freeware System Advisor Model (SAM) of National Renewable Energy Laboratory (NREL) is integrated with the proposed approach. Using real data of PV generation and load consumption of a building, in Cyprus, the obtained results show that the daily hourly profile of the import and export energies is smoothed and flattened; thus, achieving, a nearly zero grid energy of the building. The results suggest that this method is superior than a conventional rule-based battery dispatch and can lead to the reduction of the annual aggregated grid usage by 53% and to the increase of the building’s RES share by 60%, when compared to a no storage scenario. Finally, the proposed approach further decreases the primary energy consumption of the building, when compared to the no storage scenario and a battery dispatch approach driven by a conventional rule-based algorithm. Based on these findings, the proposed paradigm provides a tool contributing to the enhancement of the daily building’s energy consumption; thus, supporting the nZEB philosophy, in addition to the design measures taken for nZEBs so far.