Flexibility Forecast at Local Energy Community Level

Abstract

Large-scale integration of intermittent renewable energy resources into power systems increases the need for flexibility services such as frequency and voltage control. In the future, system operators need to utilize more flexible energy resources from all levels of the system in order to fulfill flexibility needs. Aggregated customers in form of a local energy community (LEC) are potential resources which can provide a part of the required flexibility. In this regard, accurate forecasting of flexible capacities of a LEC is essential. This paper proposes a methodology to estimate the flexibility of a LEC based on the LEC's predicted consumption. In addition, the paper suggests a novel prediction method which is based on a three-branch architecture using recurrent neural networks (RNN) and long short-term memory (LSTM) units to forecast the consumption of the LEC considering its temporal dependencies. Finally, the proposed prediction methods are implemented on a case study and the results are compared with each other.