A novel data-driven method for detection and localization of power system events causing violation of pre-defined ROCOF limits

Abstract

For reliable operation of a power system, its situational awareness should be accurate. Similarly, for power system planning, correct post-mortem analysis of system events is required. A novel event assessment method based on the energy of system frequency and generator's rotational speed measurements has been proposed in this paper which can be used for both the real-time as well as off-line event assessment purposes. In this paper, event detection is carried out based on the magnitude of the second difference of the energy of machine's rotational speed/system frequency measurements. A detection threshold has been designed for this purpose. It successfully assesses multiple events occurring in small time interval. For event localization, a novel indicator named 'Sharpness Index' has been proposed in this paper. Performance of the proposed approach has been compared with a few existing methods and found to be more accurate and comprehensive. This proposed event assessment technique has been tested for various test cases simulated using PSS/E (Power System Simulator for Engineering) as well as real PMU (phasor measurement unit) measurements.