Files
Abstract
Driven by economic and environmental policies, there is an increase in distributed energy resources (DER) penetration, such as wind farms and rooftop solar panels in the electrical grid system. With advancement in smart technology such as smart meters and smart inverters, the DERs are transforming the traditional consumers into prosumers (Producers and Consumers) that can actively contribute to the power grid operation by providing grid services. This work will investigate the grid management and coordination strategies for distributed photovoltaics (PVs) and smart buildings in distribution power networks.This research includes two parts. The first part is to establish a centralized optimization framework to effectively coordinate the operations of distributed PVs and Heating, ventilation, and air conditioning (HVAC) unit in smart buildings in the distribution network to minimize the total network losses. The proposed control strategy has been compared with a basic thermostat control logic to demonstrate its effectiveness.
The second part of this research proposes a distributed optimization approach to coordinating distributed PVs and building aggregators. It is usually not feasible for a system operator to directly model and control individual HVAC unit in each building from the grid operation’s perspective. Also, the privacy concerns of the customers and other parties in the network need to be considered. In this regard, a decentralized optimization framework is proposed in this work for distributed PVs and building aggregators. The optimization problem is divided as grid operations main problem and aggregators sub-problem. A modified Benders decomposition algorithm is proposed to solve the model. The objective of the main problem is to minimize the total network losses while maintaining the nodal voltage in the distribution network. The sub-problem represents the operations of buildings. The objective of sub-problem is to minimize the active power consumption by optimally operating the HVAC units over a time period. The Lagrangian dual extracted from sub-problems is used to update the main problem to converge to an optimal solution using the modified Benders
decomposition algorithm which is based on the classical Benders decomposition technique.
All the models have been implemented in MATLAB with YALMIP tool box and solved using commercial solvers such as Gurobi or CPLEX to obtain the optimal solutions. Case studies and comparisons have been conducted to verify the effectiveness of the proposed models.