A Bootstrap Method for Statistical Power System Mode Estimation and Probing Signal Selection

Near real-time measurement-based electromechanical-mode estimation offers considerable potential for many future power-system operation and control strategies. Recent research has investigated the use of low-level pseudo random noise (PRN) probing signals injected into power systems to estimate the low-frequency electromechanical modes. Because of the random nature of a power system, estimating the modes from a single probing experiment is very difficult. Ideally, one would use a Monte-Carlo approach with multiple independent probing experiments resulting in a mode estimation distribution. Then, one could state that the mode is within a region of the complex plane. Unfortunately, conducting multiple probing experiments is prohibitive for most power-system applications. This paper presents a methodology for estimating the mode distribution based on one probing test using a Bootstrap algorithm. The proposed method is applied to both simulation data and actual-system measurement data to illustrate its performance and application. It is demonstrated that the method can provide valuable information to PRN tests and guide future PRN probing signal design and selection