Real-time optimal urban & natural flood control for Wetropolis and Leeds' Aire River

Motivated by the Boxing Day 2015 floods, we aim to improve flood mitigation using active flood control given rainfall and river level data as well as rainfall predictions. Another option is to create more water storage capacity in upstream tributaries using natural flood management, the effects of which we will explore in combination with active control. Mathematical control for transient and nonlinear phenomena such as flooding goes beyond established control techniques for linear time-invariant systems. We will explore the application of optimal control techniques for strongly nonlinear flooding events: Pontryagin’s theory of optimal control as gold standard, which requires adjoint modeling, as well as fast (linear) approximations thereof, such as State-Dependent Ricatti Equation –SDRE– control and Model Predictive Control) [2,3]. All these techniques have shortcomings in their current form in flood alleviation applications that need to be overcome. Our aim is to establish optimal flood control theory for generic use in mitigating (urban) flooding. Flood control options include: (i) the use of moveable weirs for hydraulic control, (ii) the (de-)activitation of bypass channels or canals, and (iii) the use of moveable weirs and underpass dams to create temporary reservoirs by hydraulic constriction, as controls.