News
Driving Energy Flexibility Forward: UL’s Innovative Forecast Models in OPENTUNITY
The OPENTUNITY project aims to empower end-users by leveraging their flexibility assets within the ecosystem, supporting local system operators and creating new revenue streams. Aggregators play a key role by gathering flexibility from various sources and offering it to energy markets, optimizing asset activation when bids are accepted. This is where the University of Ljubljana (UL) contributes significantly.
UL’s main involvement in OPENTUNITY is in Task 4.4, focusing on developing a method for optimal flexibility selection from units of various sizes and types. The outcome will be an algorithm for optimal selection based on market participation (DSO or TSO) and minimal cost criteria.
We’ve adopted a bottom-up approach, splitting the task into:
- Baseline and flexibility forecast models for EV fleet (AVANT) and HEMS fleet (AMIBIT).
- Optimal selection algorithm.
Currently, we’re focused on developing forecast models for EV and HEMS fleets using Python, Keras, and Tensorflow. The analysis of charging point (CP) data for EV fleets shows significant flexibility potential during night hours by postponing charging sessions.
The difference between charging session length can be seen in the figure below.
This flexibility potential is further illustrated in the figure below.
In the coming months, UL will continue to delve into baseline and flexibility forecasting for EV and HEMS fleets, marking the initial step towards developing the optimal selection innovation and empowering end-users within the flexibility ecosystem.
Embarking on AI-Powered Non-Intrusive Load Monitoring (NILM) Algorithms
Embarking on AI-Powered Non-Intrusive Load Monitoring (NILM) Algorithms
11/10/2023
ETRA‘s role in this project extends beyond mere coordination; it also involves the delivery of innovative technologies and solutions to empower grid operators. One significant stride in this journey is ETRA’s work in the development of Artificial Intelligence-based Non-Intrusive Load Monitoring (NILM) algorithms. This algorithm belongs to the OPENFLEX innovation.
These advanced algorithms harness data derived from overall household energy consumption, enabling the inference of active appliances and their respective energy usage at any given moment. What sets NILM apart is its ability to achieve disaggregated energy consumption insights without the need for submetering, offering cost-effective benefits to end-users.
ETRA has conducted extensive research in this field, meticulously assessing the State of the Art. The most promising avenue for its work is the development of semi-supervised algorithms, an area that remains relatively uncharted in the realm of AI. Within the OPENTUNITY project, ETRA is committed to expanding knowledge in this domain, ultimately delivering new technology to European households built on this innovative and promising approach.