Abstract:
In this paper, we investigate a data driven approach based on o-Support Vector Regression (o-SVR) to identify the dependence of present sample of power-line frequency on past few samples. In standard practice, the Phasor Measurement Units (PMUs) measure the frequency samples continuously from various bus locations in the power grid and transmit them at a fixed rate, typically at 25 samples/sec, to the Phasor Data Concentrator (PDC). Objective of the proposed strategy is to reduce the sampling rate at a PMU or transmission rate of the fixed-rate samples from a PMU to the PDC such that any impending disturbance in the power system can be detected early without compromising stability of the power system. We evaluate the performance of our proposed model by quantifying the sample data rate reduction and obtaining the average prediction error during the steady state as well as disturbed state conditions in the power gird.