Effect of Home Energy Management System Penetration Levels on Bi-Level Optimal Energy Management in Distribution Networks

In this paper, the effect of increasing penetration of deep reinforcement learning (DRL) based home energy management systems (HEMS) in distribution networks with bi-level optimal energy management (OEM) is presented. It is shown how the HEMS penetration levels affect dynamic pricing, distribution system operators’ profits as well as energy costs of the customers with smart homes. Smart home (SH) penetration levels of approximately 25%, 50%, and 100% in a DN with 55 homes are discussed. Two pricing mechanisms are studied. In the first method, the DRL agent representing the distribution system operator (DSO) generates a dynamic price signal of its own whereas in the second method, it generates a signal to manipulate the actual market price. The results suggest that while the increase in SH penetration can help improve DSO profits, it has no direct effect on consumer energy cost and thermal comfort.