in fluctuating renewable energy, such as wind and solar, and the increase in
electricity demand driven by the electrification of heat and transport sectors
challenge traditional power system operation. Power system operators routinely
perform system security assessments to ensure a safe and stable operation. With
growing uncertainty, the complexity of this computationally intensive task
further increases calling for the development of new approaches.
Deep learning has
shown tremendous potential in a wide range of scientific and technological
applications, recently publicized for the game of Go with the computer program AlphaGo
winning against the world champion. However, the application of deep
learning in power systems has been limited so far. One promising direction is
deep learning for feature extraction. The goal is to map the high-dimensional
power system state to a significantly lower-dimensional feature space more
suitable for a classifier to discriminate between safe and unsafe operation. The
resulting tool could aid operators in maintaining a stable power system
Please see the attached proposal for more information.
Power system modeling or machine learning, good knowledge of Matlab, Python or another programming language, high interest in deep learning.