Information length quantification and forecasting of power systems kinetic energy

Power systems operation and planning are facing several short-coming challenges due to the large inclusion of non-synchronous generation and the constant expansion of the electrical network. One of these challenges corresponds to the monitoring and forecasting of power systems Kinetic Energy (KE) to show on-line additional information for the Transmission System Operators (TSOs). In view of this challenge, KE monitoring requires innovative methods to analyse the continuous fluctuations in the system. Moreover, KE forecasting can foresee the status of the strength to overcome further events. In this work, we propose the use of information theory (specifically the concept of information length) as a way to provide useful insight in the power system KE variability and to demonstrate its usage as a starting point in decision making for power systems management. Additionally, a short-period forecasting using a Long Short Term Memory (LSTM) neural network model is proposed to estimate the value of information length in real time. The methodology is applied to a monthly collected data from the Nordic Power System. Results show that our method provides an effective description of the seasonal statistical variability.