Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Stochastic Online Feedback Optimization for Power Systems
The objective of the Thesis is to build a stochastic online feedback optimization algorithm taking into account the uncertainty from a dynamic state estimation, and use to solve a power grid optimal power flow.
Keywords: convex optimization, Kalman filter, power systems
Online Feedback Optimization is a novel way of controlling a system, and drive it to an optimal point defined by an optimization problem (e.g. maximum efficiency). It can be used to solve the AC Optimal Power Flow in real-time in power grids. In this context, it consists in continuously driving the controllable power injections and loads towards the optimal set-points in time-varying conditions based on real-time measurements performed on the grid. Most implementations in the literature assume noise-free full state measurements, which may be unrealistic for certain networks, like distribution grids. Instead, a recent approach (https://arxiv.org/pdf/1909.02753.pdf) allows to use any set of available measurements by connecting the online feedback optimization to a dynamic State Estimation, and certifies stability of the interconnection and convergence in expectation.
Online Feedback Optimization is a novel way of controlling a system, and drive it to an optimal point defined by an optimization problem (e.g. maximum efficiency). It can be used to solve the AC Optimal Power Flow in real-time in power grids. In this context, it consists in continuously driving the controllable power injections and loads towards the optimal set-points in time-varying conditions based on real-time measurements performed on the grid. Most implementations in the literature assume noise-free full state measurements, which may be unrealistic for certain networks, like distribution grids. Instead, a recent approach (https://arxiv.org/pdf/1909.02753.pdf) allows to use any set of available measurements by connecting the online feedback optimization to a dynamic State Estimation, and certifies stability of the interconnection and convergence in expectation.
The goal of this project is to extend those results to include the uncertainty of the estimation in the online optimization scheme, and build a stochastic online feedback optimization. Additionally, this stochastic optimization can be robustified against unknown probability distribution using tools from distributionally robust optimization.
The goal of this project is to extend those results to include the uncertainty of the estimation in the online optimization scheme, and build a stochastic online feedback optimization. Additionally, this stochastic optimization can be robustified against unknown probability distribution using tools from distributionally robust optimization.