Visualizing the Dynamics of the Electricity Grid through Diffusion Maps

Abstract

The United States electricity grid today is not adequate to meet the needs of the 21st century. Due to intermittent resources such as wind and solar, threat of cyber attacks, and increasing demand, the next-generation electricity grid must be more flexible and resilient. Models of the electricity grid are used to design, manage, and analyze the grid. Current models of the electricity grid can produce large, high-dimensional data sets that are hard to analyze to due their complexity. This thesis seeks to reduce the complexity of the day-ahead economic unit commitment and dispatch model for the Pennsylvania-Jersey-Maryland electricity grid through dimensionality reduction tools, namely diffusion maps. By using diffusion maps, a new representation of the data in a lower dimensional space was found. This indicates that one can use fewer variables to run the simulation. Through this analysis, one can gain a better understanding of the system’s dynamics, which is useful in preparing for the next-generation electricity grid. For this particular application it was found that the system’s important heterogeneous parameters were the generators’ maximum capacities and variable costs.