Improved load approximation models and tractable solution schemes for home energy management

Home energy management systems (HEMS) allow residential consumers to schedule appliances in accordance to time varying prices to reduce electricity cost. The cost reduction may come with discomfort — delayed start in a process or interruptions in another. Scheduling algorithms that implement such load control strategies typically consider average or rated power consumption of appliances, underestimating the occurrence of peaks and valleys of the actual profile. This paper presents an approximated model that considers the notion of cycle of operation, with power consumption varying across different cycles. Additionally, consumption models for air conditioners and electric vehicles are proposed. These load models are used in a HEMS multi-objective optimization problem that minimizes cost and discomfort. The formulated problem is also approximately solved using Lagrangian decomposition and continuous variable problem formulation. Application of the proposed solution schemes to representative prosumer set-ups are presented and the resulting power consumption profiles and cost are studied. Results show how the proposed approaches result in lower electricity costs compared to the typical scheduling approaches that use the average power consumption model. Investigation on the cost metrics and computation provides insights on practical implementation of such HEMS scheduling approaches.