Optimal scheduling in smart grids with prediction error probability limitation

Abstract

Smart grid is an electrical grid that uses information and communications technology to improve e ciency, reliability, economics, and sustainability of production and distribution of electricity, which results in much more utilization of this technology in the future. Renewable sources will be utilized more in the future. However, due to its random nature, suppliers are suspicious while integrating these sources into grid. In this thesis, we study demand response, supply management, and power scheduling in a smart grid in the presence of renewable energy sources aiming to increase e ciency and reliability of the grid by limiting the probability of error coming from the prediction of renewable sources. This novel method brings improvement in terms of welfare compared to other methods, which are already available in the literature. In our case, the energy provider bounds the value of maximum prediction error via error probability of prediction. In this decision, the provider compares the cost of spillage and de cit cost of energy production, which are greater than scheduled power production. The goal is to maximize the total social welfare, de ned in terms of consumer utility. Furthermore, it has been shown in this research that adding battery to the grid brings an extra improvement in terms of total welfare.