Development of energy aggregators for virtual communities: The energy efficiency-flexibility nexus for demand response

The implementation of load management and demand response programs is motivating utilities to propose demand-side management to incentivize customers to modify their energy consumption during critical events, providing energy flexibility to the grid. With the widespread use of distributed energy resources, energy aggregators have grown significantly to manage the portfolio of residential buildings. This paper presents a control-oriented implementation to define the energy flexibility potential of a portfolio of residential buildings within the same energy aggregator. The data-driven methodology combines the resistance-capacitance thermal network model of the households with building integrated photovoltaics/thermal and air source heat pump models to assess the modification of the baseline of power demand at the energy aggregator level. The optimization is performed through a model predictive control (MPC) framework for day-ahead predictions. Parametric analysis and scenario investigation are exploited to define the optimal configuration for both the households and the grid operator. The presented methodology provides insight for energy efficiency and flexibility about the operating and design of energy aggregators in a demand response program. A Canadian virtual community is used as a case study to estimate the performance of the presented methodology. The optimization of load management at the aggregator level is presented for a representative demand response (DR) event during the winter season for two different energy pricing structures. Numerical results show the potential of the proposed methodology to guide the design and operation of virtual clusters integrating renewable technologies towards energy efficiency, showing a reduction of 40% in total energy consumption and a peak reduction of around 32%.