Data-driven assessment of center of inertia and regional inertia content considering load contribution

Abstract

The inertia displacement in power grids, caused by the massive integration of renewable energy resources (RES) and responsive loads, has become one of the biggest challenges to operating and controlling power systems. Locating the centre of inertia (COI) of a region can help identify this inertia displacement. This work proposes a novel fully data-driven disturbance-based methodology to estimate both the COI and the inertia of a region, where the variation of the COI due to RES and load contribution are considered. The methodology uses a recursive form of the typicality-based data analysis (TDA) to find the pilot-bus (TDAp). The TDA methodology approximates the multi-modal distribution of active power and frequency measurements by the typicality property, to detect the COI of each region and the corresponding pilot-bus. A composite metric of correlation and cosine similarities of active power and frequency measurements is used to approximate the measurements’ unknown distribution. The window of measurements necessary for detection is attained by the convergence of the typicality variance. The frequency of the detected pilot-bus of the region and tie-lines active powers deviations are used by an auto-regressive moving average exogenous input (ARMAX) approach to estimate the inertia of a regional equivalent machine. The methodology is capable of identifying RES and load contribution, since the pilot-bus detection through the TDAp is sensible to the COI displacement by these equipments, unlike traditional methods that only consider an average of synchronous machines’ inertia and some heuristics for load contribution. The proposed methodology is tested by using the IEEE 68-bus benchmark test system, an adapted version with aggregated dynamical loads, and also with RES participation through type-3 wind generators, corroborating the effectiveness of the proposal.