Optimal Sizing of a Photovoltaic System for Peak Power Shaving Application S. Majumdar Maintenance Services Schneider Electric Kolkata, India sandeepan.majumdar@gmail.com K.K Mandal Department of Power Engineering Jadavpur University Kolkata, India kkm567@yahoo.com N Chakraborty Department of Power Engineering Jadavpur University Kolkata, India chakraborty_niladri@yahoo.com Abstract— This paper is a study of variation of load of a small building and implementation of a Photovoltaic system to reduce the peak demand. Load forecasting in this paper uses input data dependent on parameters such as insolent radiation, temperature etc. A multi parameter regression model is used to model the monthly load profile of the system under study. The system under study is based on the electrical system associated with Jadavpur University Saltlake Campus, Kolkata India. An accurate PV module electrical model is presented based on Shockley Diode equation. For simulation purposes data for load temperature irradiance etc are obtained from locally measured values. Suitable assumptions are made with regards to temperature and insolent radiation so that accurate demand profile can be generated. Finally the optimal system size is determined by a Particle Swarm Optimization such that system cost and energy savings reach a breakeven over a period of few years Keywords- Particle swarm optimization; Photovoltaic systems; Regression analysis; Solar energy. I. INTRODUCTION The increase in fossil fuel prices has a significant impact on economy and standard of living of people. Renewable energy sources will find increasingly important part in the modern power generation scenario. Distributed generation (DG), defined as small-scale electricity generation, is a new concept which is considered as a solution for addressing technical, economical and environmental issues of conventional power generation. Photovoltaic (PV) systems produce DC electricity when sunlight shines on PV array. Then DC power is converted to AC Power and fed to load bus to power. Optimal sizing of PV module is the challenge that is tackled in this paper. The output of the module is fed to a DC bus and an inverter is used to supply power to the loads. Electricity demand is closely influenced by climatic parameters, there is likely to be an impact as developing countries improve their standard of living, their use of air conditioning and other weather-dependent consumption may increase the demand sensitivity to climate change [1] .This paper attempts a statistical regression method for existing studies on few parameter impacts on electricity demand, with the ultimate aim of prediction of demand profile. Energy storage devices are a fundamental part of PV system [2] .However they form the most costly part of the system in terms of installation and maintenance costs. Primary idea is to use an energy buffering system that consumes power, when supply supersedes demand. The idea is to create a scheduling pattern that mitigates the need of energy storage system, and power generated from the PV system is used to reduce the peak demand. Since the time of peak energy requirement is concurrent with hours of peak sunshine hours. II. MATHEMATICAL MODELS A. PV Model The simplest equivalent circuit of a solar cell is a current source in parallel with a diode as shown in Figure 1. Fig.1. Circuit diagram of the PV model N. Chakraborty is associated with the Dept. of Power Engineering, Jadavpur University, Kolkata-98, e-mail: chakraborty_niladri@hotmail.com K.K. Mandal is associated with Dept. of Power Engineering, Jadavpur University, Kolkata-98, e-mail: kkm567@yahoo.com Sandeepan Majumdar is a former student of Dept of Power Engineering, Jadavpur University. He is currently emplyed in Schneider Electric India Pvt Ltd., Kolkata- 98, e-mail: sandeepan.majumdar@gmail.com