Storage in Power Systems: Frequency Control, Scheduling of Multiple Applications, and Computational Complexity

In this talk, I present different methods improving storage profitability and reducing computational burden that I developed in my thesis.

First, I introduce methods enabling energy-constrained storage units to participate in frequency control. These methods take advantage of the higher ramping rate of storage units, and rely on slower conventional generators to take care of the bias of the frequency control signal. The most elaborated method relies on distributed optimization and considers a set of generators and storage units.

Building on these frequency control methods, I propose multitasking algorithms that improve the storage profitability by combining revenue streams from different applications, including frequency control. These algorithms rely on model predictive control or stochastic dual dynamic programming.

The last part of this talk focuses on a different problem: reducing the computational complexity of convex multi-period stochastic DC OPF problems. The method includes an iterative algorithm approximating a multi-dimensional convex function that can be used outside of the field of power systems. A 118-bus case study shows that the method achieves between 25 and 81% of the cost improvement of the stochastic OPF (compared to the deterministic OPF), but requires only 12 to 41% of the increase in computation time required by the stochastic OPF.
Olivier Megel obtained his B.S. ('08) and M.S. ('11) degrees in Mechanical Engineering from EPFL Lausanne, Switzerland, and his Ph.D. degree ('17) in Electrical Engineering from ETH Zurich (Swiss Federal Institute of Technology), Switzerland. His research interests include optimization and control strategies for storage units in power systems.