Self-adaptive Controllers for Renewable Energy Communities Based on Transformer Loading Estimation

Paper i proceeding, 2022

In this paper, self-adaptive controllers for renewable energy communities based on data-driven approach are proposed to mitigate the voltage rise and transformer congestion at the community level. In the proposed approach, the transformer loading percentage is estimated by the trained data-driven model, which uses the extreme gradient boosting regression algorithm based on a measurement set acquired from critical coupling points of the communities. To avoid voltage rise issues, the droop control parameters (i.e., voltage threshold for P - V, Q - V curves) are adaptively tuned based on the solar irradiance availability and estimated transformer loading. The proposed approach has been tested in the IEEE European LV distribution network. Results showed that the control approach could effectively reduce 22.2 % of the total overloaded instances, while still keeping voltage magnitude in the operation range. This method can help DSOs manage voltage violation and congestion without further communication.