Decision Support Tools
Tools and implementations that make uncertainty visible and usable in power system decision-making.
This page presents tools and open-source implementations developed as part of my research on decision-making under uncertainty in power systems. The work spans interactive demonstrations of how uncertainty behaves in real systems, and research implementations that translate stochastic optimization methods into reproducible, operationally relevant code.
A common thread across these tools is interpretability: making it possible to understand not only what the optimal decision is, but how uncertainty shapes it and what the associated risk exposure looks like.
Interactive demonstration
An interactive tool that reconstructs empirical forecast error distributions from three years of Belgian wind and solar data. Users can condition on any forecast value and observe how the shape of uncertainty changes — including the departure from Gaussian behavior near capacity bounds and under specific forecast regimes. The tool fits Beta and Gaussian approximations client-side and computes CVaR directly from conditioned samples, illustrating the risk implications of forecast uncertainty in a real operational context.
Research implementations
Open-source implementations of stochastic optimization formulations developed in my research. These accompany published work and are intended to support reproducibility and further development by the research community.
An extension of PowerModels.jl for stochastic power system optimization, developed as part of my research at KU Leuven. Implements stochastic optimal power flow for AC and hybrid AC/DC networks using Polynomial Chaos Expansion for continuous uncertainty propagation, with support for risk-based formulations and transmission switching optimization.
Additional repositories are available on GitHub. If you are working on related problems or interested in applying these methods in practice, I would be glad to discuss potential collaboration.