Integrated smart appliance scheduling and HVAC control for peak residential load management

Residential consumers are significant contributors to peak energy demand because of their aggregate daily consumption as a result of typical occupancy cycles and responses to fluctuations in ambient temperature. In this work, we create an energy management framework that optimizes the use of air-conditioning (A/C) systems and energy-intensive household appliances in order to minimize the peak load for a community of homes. Specifically, we propose a two-tiered approach that first implements a centralized model predictive control system which makes thermostat set point decisions that reduce the peak of the aggregate demand profile. We then execute a scheduling calculation that optimizes the start times and cycle duration of the most energy-intensive (and flexible) smart appliances to further reduce the peak load. We rely on residential energy use data collected from 40 existing houses in Austin, TX, USA and local weather information to provide a case study demonstrating the efficacy of our framework. On average, the aggregate neighborhood load is reduced by 25.8% (15.7kW) (compared with the case using the original thermostat settings and appliance start times), while minimally increasing energy use, respecting the thermal comfort and appliance completion time preferences of individual homes, and without creating secondary demand peaks.