PhD Machine Learning for Outliers in Grids Operations and Intraday Energy Markets

Updated: over 1 year ago
Job Type: Temporary
Deadline: 25 Sep 2022

As part of your 4-year PhD project, you will develop novel Artificial Intelligence(AI)-based algorithms for coupling electric mobility and energy systems. You will join the Austrian Institute of Technology (AIT) as an employee (PhD contract) and enrol as a PhD at TU Delft. You will be based in Vienna, Austria and travel around two times a year to the Netherlands for research visits of multiple weeks. You will benefit from the professional development programs and facilities of AIT and TU Delft at the same time.

Your PhD project is to develop a self-learning algorithm that detects anomalies in grid & market data and distinguishes them from transferable policies. You will work with data from Phasor Measurement Units (PMU) and market anomalies. Subsequently, you will develop an unsupervised learning algorithm to detect outliers that have never been seen before. You investigate how incrementally to improve the algorithm's performance under system changes. For example, a change in the distribution grid data could relate to a detected outlier, cyber-attack or planned maintenance. You will design an algorithm using the network topology, transfer learning, General Adversarial Networks, and building the algorithmic blocks with modules that consider reasonable perturbations from expert knowledge.

AIT and TU Delft are launching the "AI research in energy system planning and operations" program forming a new expert team combining the strength of both institutions. In this program, we combine ground-breaking AI methods with the reliable theory of the physical energy system. The area of data-driven scientific computing promises to incorporate statistics, time-frequency analysis, low-dimensional model reductions, and other techniques to extract information from data. With AI, we make such valuable information for managing and planning complex energy systems. At TU Delft, your academic supervisors are Dr. Jochen Cremer (daily supervisor) and Assoc. Prof. Dr. Peyman Mohajerin Esfahani, Delft Center for Systems and Control. At AIT, your supervisors are Tara Esterl, Fabian Leimgruber and Dr. Mark Stefan.



Similar Positions