GRA-NGBoost based System Inertia Prediction for High Wind Power-Penetrated Grids

With the increasingly penetration of wind power generation, the system inertia presents nonlinear and time-varying characteristics. Accurately predicting the variation tendency of the system inertia is helpful for operators to identify the time period when the system inertia is less than the critical level, and to take corresponding remedial measures in advance. Driven by the virtual inertia control technology, the virtual inertia provided by renewable units also needs to be taken into account. This paper proposes a GRA-NGBoost based system inertia prediction method for high wind power-penetrated grids. Case study on the modified IEEE RTS-79 system show that the proposed method has high accuracy, low sample dependence, and good fitting capability of the system inertia, which is of great significance for operators to prepare appropriate unit schedule plan under high risk low-inertia operating scenarios.