Multiobjective Optimal Unit Sizing of a Grid-Connected Photovoltaic System in Consideration of Its Probabilistic Characteristics

Abstract

A probabilistic approach is presented to optimal unit sizing of a grid-connected photovoltaic system without storage batteries. In consideration of probabilistic characteristics of solar insolation and electricity demand, the surface area of photovoltaic array, capacity of receiving device, and electric contract demand are determined so as to minimize the expected values of annual total cost and annual energy consumption subject to the annual loss of power supply probability. This optimization problem is considered as a multiobjective one, and a discrete set of Pareto optimal solutions is derived numerically using the weighting method. A case study is carried out on an on-site system with the option of reverse power flow into the grid, using data on solar insolation obtained through real measurement. The trade-off relationship between the two objectives are clarified and optimal values of the above design items are determined in relation to the annual loss of power supply probability and the capital unit cost of photovoltaic array. Results by this probabilistic approach are compared with those by a deterministic approach.