Box–Cox integrated sARIMA model for day-ahead inertia forecasting

Modern power systems are characterized by low inertia due to the high penetration of renewable generation (RG). The challenges due to low inertia like high rate of change of frequency (RoCoF) can be overcome by deploying sufficient Fast-Frequency Containment Reserves (FFCRs). The optimal allocation of FFCRs necessitates accurate inertia forecasting. Inertia forecasting has got only a little research attention. Also, existing inertia forecasting models consider only the inertia support from synchronous generators. However, wind generators and motor loads also contribute to the inertia. Further, data pre-processing, transformation, and input selection techniques can be used to improve inertia forecasting accuracy. Therefore, this paper proposes a novel Box–Cox integrated seasonal autoregressive integrated moving average (sARIMA) model by considering the inertia contribution of wind generators and motor loads along with synchronous generators. Box–Cox transformation stabilizes the variance and reduces the impact of heteroscedasticity. This will help reduce white noise and obtain accurate forecasts. sARIMA model is suitable for inertial energy forecasting as inertial energy data shows seasonal patterns. The proposed model uses the Spearman correlation as input selection and the tsrobprep package as data pre-processing techniques. The performance analysis shows that the proposed model shows an annual forecasting accuracy of above 97.5%.