EURO-Online login
- New to EURO? Create an account
- I forgot my username and/or my password.
- Help with cookies
(important for IE8 users)
3193. Machine Learning based surrogate models for large-scale sector-coupled energy systems
Invited abstract in session MA-19: Learning-assisted Optimization in Energy Problems, stream OR in Energy.
Monday, 8:30-10:00Room: 44 (building: 116)
Authors (first author is the speaker)
| 1. | Tim Graulich
|
| DTU Management, Technical University of Denmark | |
| 2. | Marie Münster
|
| DTU Management Engineering, Technical University of Denmark | |
| 3. | Bissan Ghaddar
|
| Ivey Business School | |
| 4. | Rasmus Bramstoft
|
| Technical University of Denmark |
Abstract
Surrogate modeling is a well-established field within engineering sciences, where computational simulation and optimization models are becoming increasingly complex. Instead of running costly high-fidelity models, surrogate models can provide accurate estimations , while reducing computation times. Currently, the energy system is becoming increasingly integrated across the energy sectors and vectors, leading to a growing demand for sector-coupled large-scale energy system models. At the same time higher shares of variable renewable energy sources require higher temporal and spatial resolutions when performing the energy system modeling. This development results in an increasing complexity of these models. One possible method to deal with this growing complexity is to build surrogate models representing a specific energy sector (e.g. hydrogen), which can then be coupled to another operational model (e.g. power system model). However, while surrogate models have been used for various applications in the energy domain, ranging from building energy performance simulation to power grid optimization, no surrogate models have been developed to represent large-scale sectoral energy system models. A big challenge of these model is the large number of input and output parameters. This talk will present ongoing research on how to use sensitivity analysis to reduce the models complexity for the surrogate model.
Keywords
- Machine Learning
- Energy Policy and Planning
Status: accepted
Back to the list of papers