Development of a supervisory controller for residential energy management problems

In recent years, the infrastructure that supplies energy to residential areas has started to evolve into a multi-source system, just like in automotive industry in which hybrid electric vehicles (HEVs) have been replacing conventional gasoline vehicles. Multi energy source systems considered as a potential solution for carbon emission problems despite their challenges in their operation due to increased complexity. In this paper, a control design approach successfully applied in the automotive industry is used to solve a residential energy management problem. First, a dynamic programming method is applied to obtain optimal control actions for the representative demand profiles and then by using these results, a causal supervisory controller is developed. It is found that the developed baseline controller performs 1-2% better daily in its initial form in terms of operational costs, compared to available heuristic strategies.