To reduce greenhouse gas emissions due to fossil fuels, renewable energy generators have been deployed in electric grids in a large scale. Among renewable energy technologies such as hydroelectric, photovoltaic (PV), wind, geothermal, biomass, and tidal systems, grid-connected solar PV continues to be the fastest growing power generation technology; however, solar energy generation tends to be unsteady due to the diurnal cycle of the solar geometry and clouds, which necessitates the usage of storage devices. This talk focuses on the problem of determining the size of battery storages used in grid-connected PV systems. In our setting, electricity is generated from PV and is used to supply the demand from loads. Excess electricity generated from the PV can be stored in a battery to be used later on, and electricity can be purchased from the electric grid if the PV generation and battery discharging cannot meet the demand. The objective is to minimize the electricity purchase from the electric grid while at the same time choosing an appropriate battery size (denoted as E_{max}^c). We propose an upper bound on E_{max}^c, and show that the upper bound is achievable for certain scenarios. For the case with ideal PV generation and constant loads, we characterize the exact value of E_{max}^c, and also show how the storage size changes as the constant load changes; these results are validated via simulations.