Comparison of Dynamic Mode Decompositon and Iterative Matrix Pencil Method for Power System Modal Analysis

This paper compares two modal analysis techniques developed for handling large sets of data, Dynamic Mode Decomposition and the Iterative Matrix Pencil method. The two methods are compared by analyzing large-scale synthetic grids and using metrics such as the average cost function and computation time. Dynamic Mode Decomposition is shown to be able to quickly identify modes, but suffering inaccuracies in reproduced data. These difficulties are explored in various case studies and are compared to the Iterative Matrix Pencil method. Sampling frequency is also considered in order to examine the impact of larger data sets when comparing the two methodologies.