In the global context of the electric power grid modernization, storage of electricity is a crucial issue. Nowadays, energy storage systems (ESS) are used more and more in positive energy buildings in conjunction with new Low Voltage Direct Current (LVDC) grids. However, the impact of renewable energy sources (RES) on ESS is not well known. The main objective of this article is to determine a systematic methodology to study energy data from a positive energy building in order to determine the impact on ESS dedicated to be included in smart-grids. The aim is to obtain comparative results in normalized working conditions and determine charge/discharge cycles. Clustering methods were compared to choose the more adapted one to treat the stored data of energy production and consumption during more than three years in our experimental platform in LAAS-CNRS, Toulouse. Each type of cycle will help further study in order to estimate its impact on efficiency and lifetime of ESS and then choose the more adapted element for each application.