The increasing number of rooftop solar panels and home-scale batteries in residential areas enables an emerging energy market for homeowners to take part in. By optimizing the energy usage at home, the surplus amount can be exported to the grid for selling to intermediaries like aggregators at a higher price than a feed-in tariff set by the government. Though not trading with the power utility, the resulting reverse power flow may cause the service voltage to increase and violate the set upper limit. To avoid this problem, the utility asks all grid-connected inverters to reduce the reverse power by installing volt-watt functions. Unfortunately, this kind of function may reduce the profits of the homeowners since the exported energy cannot match the amount committed to the aggregator. Our research addresses this issue by proposing a double-layer optimization for each home energy management system. The first layer optimizes the one-day home energy usage with consideration of uncertainty in the predicted data of load and solar irradiance to estimate the highest home profit as possible. Then, the second layer tries to reduce the high reverse power with minimal impact on the profit found in the first layer. A simulation of a distribution system shows that the proposed method is able to flatten the power profile flowing reversely to the grid and reduce the drawbacks of volt-watt functions. Homeowners are recommended to practice the proposed method when trading the surplus energy.
Related works The literature includes various tech-niques for managing home and commercial building energy, and how to achieve benefit from the demand response program [2]. Since the system contains a wide range of home appliances at multiple power levels (washing machines, cloth dryers, and dishwashers), flexible and deferrable (electric vehicles and batteries), thermal-related (water heaters and air conditioners), there are a vast number of variables needed for modeling and taking part in the optimization. Some techniques are mixed-integer linear