(Georgia Institute of Technology, 2021-07-26)
Jung, Yun Joon
Many global cities have announced ambitious net-zero energy consumption targets or net-zero CO2 emissions plans. It is well recognized that this can only be realized through a mix of measures such as efficiency improvements at the sites of consumption and decentralized energy generation, storage and delivery mechanisms. This transition will not happen without major changes to energy supply networks, especially in the way they enable frictionless inclusion of renewable energy sources and local supply, for instance through microgrids.
At the urban scale, buildings constitute the major consumers of electricity and their integration through building-to-building and building-to-grid controls is crucial to realize efficient energy sharing in urban energy networks. Over the last decade, the building energy simulation domain has moved its focus from traditional local studies to urban energy studies. The main objective of this thesis is to make a contribution to this growing research domain, especially in enabling the simulation of energy supply networks in a robust manner and at a large scale. It is possible to simulate such networks with customized software but considering that there is no systematic way to specify urban energy models (especially with multiple concurrent control topologies), the simulation software has to be hand-customized which leads to opaque simulations that moreover are hard to use for rapid variant explorations. The thesis argues that this can be overcome by the development of an urban prosumer (UP) schema that facilitates the specification and automated mapping of an urban energy network into simulations, focusing on the effective specification of controls outside the software.
At a high level, the UP schema is comprised of a physical and a logical layer. The physical layer conceptualizes existing urban energy networks using directed graphs for energy transport between nodes. The logical layer conceptualizes how the dynamic processing (reasoning) of sensor data leads to instructions to a set of actuators that execute the control. In doing so, two levels of control are distinguished: (a) “private” (mostly rule-based) control such as the internal HVAC system following temperature setpoints, (b) “public” control that is exposed to the rest of the network and thus within the scope of the UP schema. Public control can be either rule-based or optimal control, the latter driven by an appropriate optimality criterion, defined at a network scale. In design situations, the optimality criterion is not limited to control variables but can also include design parameters, such as building design parameters, solar installation sizes, community battery size, and the number of EV charging stations. Mixed-integer non-linear programming (MINLP) is used to solve optimal control problems. The genetic algorithm is employed to solve design optimization problems.
The case studies using the UP schema for ten Georgia Tech campus buildings are presented. The purpose of the case studies is to prove that the UP schema can facilitate simulations involving different levels of controls. The simulations target optimal energy decisions for the selected campus buildings in the presence of PV and electricity battery. Additionally, three residential buildings in California are chosen to investigate how the design and control parameters act together to avoid the power outage situation with the embedded UP schema in the simulation platform.