The high penetration level of photovoltaic (PV) generation in distribution networks not only brings benefits like carbon savings, but also induces undesirable outcomes, like more harmonic components and voltage fluctuations. Driven by decreasing costs of energy storage, the focus of this paper is to investigate the feasibility of applying energy storage in the grid-connected PV system to mitigate its intermittency. Firstly, to appreciate the functionality of storage, a generic PV-battery-supercapacitor model was simulated in MATLAB/Simulink, and a flat load profile was obtained to enhance predictability from the network management point of view. However, the usage of supercapacitors at the residential level is limited, due to its high startup costs. Secondly, a detailed residential PV-battery model was implemented in the System Advisor Model (SAM) based on local data in Dubai. The optimal sizing of a battery system was determined by assessing two criteria: The number of excursions, and average target power, which are contradictory in optimization process. Statistical indicators show that a properly sized battery system can alleviate network fluctuations. The proposed sizing method can be also applied to other PV-storage systems. Finally, economic studies of PV-battery system demonstrated its competitiveness against standalone PV systems under appropriate tariff incentives.