Wide-area PMU-ANN based monitoring of low frequency oscillations in a wind integrated power system

The wind integration is on the rise in modern grids to generate cleaner energy due to increasing environmental concerns. This further escalates the problem of low frequency oscillations (LFOs) in power system. Thus, real-time monitoring of oscillations becomes even more important in present interconnected power systems. The conventional methods are offline and time consuming. In this work, a wide-area based method employing Phasor Measurement Units (PMUs) and Artificial Neural Network (ANN) is proposed to predict the system oscillatory status in real-time. The PMUs are optimally placed using modified Integer Linear Programming. The PMU data is dimensionally reduced using Principal Component Analysis before using it to train the ANN. The ANN predicts the LFO related information like damping ratio and frequency and an index to access the mode's localness. The proposed methodology is verified on IEEE New England benchmark system. The suggested method is very fast and accurate in predicting the required information in real-time for different operating conditions involving topological variations with very less computational requirement.