Post by Intelligent Data Analysis Lab | FEE CTU

🌊⚡ How can we bridge the gap between complex physical water modeling and the real-time control of massive hydropower cascades? 📜 In their paper, "Towards Real Time Control of Water Engineering with Nonlinear Hyperbolic Partial Differential Equations," recently accepted to the Journal of Computational Science, Fabio Difonzo, Michael Holst, Morteza Kimiaei, our Vyacheslav Kungurtsev, and Songqiang Qiu explore the immense theoretical and practical hurdles of Mixed-Integer PDE-Constrained Optimization for shallow water equations. To bypass current computational limits, the authors propose a novel three-tiered architecture that combines offline HPC Machine Learning to catalog reference solutions, meso-scale approximate dynamic programming, and real-time Model Order Reduction. This integrated framework lays a principled roadmap for bringing explicit, real-time PDE control to large-scale infrastructure. Preprint: https://lnkd.in/d7DZWBK2 #Hydropower #DifferentialEquations #HPC #IDA #CTU