A data-mining based optimal demand response program for smart home with energy storages and electric vehicles

•Using data clustering methods to recognize controllable appliances.•Comparing DBSCAN clustering method with previous conventional techniques.•Proposing an optimal demand response scheme for a smart home.•Deploying energy storage systems to increase energy efficiency of homes. In recent years, modern appliances with high electricity demand have played a significant role in residential energy consumption. Despite the positive impact of these appliances on the quality of life, they suffer from major drawbacks, such as serious environmental concerns and high electricity bills. This paper introduces a consolidated framework of load management to alleviate those drawbacks. Initially, benefiting from a demonstrative analysis of home energy consumption data, controllable and responsive appliances in smart home are identified. Then, the energy consumption pattern is reduced and shifted using flexible load models and better utilization of existing energy storage systems. This can be achieved through data mining approaches, i.e., density-based spatial clustering of application with noise (DBSCAN) method. In this technique, no sensor for detection or measurement instruments will be required, whose deployment incur cost to the system or increase security risk for consumers. In the following, one scheduling of using controllable appliances, which is formulated by convex optimization, is considered for the demand response (DR) program, provided that this plan doesn't affect customers’ priority and convenience. In the last stage, the deployment of energy storage systems, such as plug-in hybrid electric vehicles (PHEVs) and battery energy storage systems (BESS), is introduced to lower the energy cost and improve the performance of the proposed DR model. Simulation results of this demand response are compared with conventional k-clustering methods to confirm the economic superiority of the DBSCAN clustering technique using the data of a residential unit during three different scenarios.