| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 2.73 MB | Adobe PDF |
Advisor(s)
Abstract(s)
In the energy sector, prosumers are becoming relevant entities for energy management systems since they can share energy with their citizen energy community (CEC). Thus, this paper proposes a novel methodology based on demand response (DR) participation in a CEC context, where unsupervised learning algorithms such as convolutional neural networks and k-means are used. This novel methodology can analyze future events on the grid and balance the consumption and generation using end-user flexibility. The end-users’ invitations to the DR event were according to their ranking obtained through three metrics. These metrics were energy flexibility, participation ratio, and flexibility history of the end-users. During the DR event, a continuous balancing assessment is performed to allow the invitation of additional end-users. Real data from a CEC with 50 buildings were used, where the results demonstrated that the end-users’ participation in two DR events allows reduction of energy costs by EUR 1.31, balancing the CEC energy resources.
Description
Keywords
Citizen energy community Demand response End-user participation Energy flexibility Unsupervised learning
Citation
Publisher
MDPI