Projects

Consortia grants

List of projects awarded to Dr. Aguilar-López and his group.

Funded by TKI Deltatechnologie · 2023–2027

ReInDeeR

Resilience for inundation based on deep learning and risk. Since 2017, Dutch flood defences have statutory safety standards that can differ by a factor of three to ten between opposite stretches of the same river, including the IJssel. Assessments treat each dike ring on its own, so they miss the system effect of an interacting flood defence. ReInDeeR tests whether design should take a lower upstream defence into account.

TU Delft leads the project with Rijkswaterstaat and HKV. Selected flood-defence scenarios are simulated so that each one carries as much information as possible, and a deep-learning surrogate of the whole system is trained on those runs. The aim is to quantify the effect of variable standards on the IJssel, and to answer the same questions for many more scenarios in a feasible time.

TKI Deltatechnologie

Funded by Horizon Europe · 2022–2026

MULTICLIMACT

TU Delft’s role is to develop, test and demonstrate fiber-optic sensing for the structural health of fixed and movable flood defences during high water, heavy rain and drought. Cables go into dikes and onto temporary barriers. Interrogators read strain and temperature. Models help decide whether a change is a warning.

On 13 October 2025, at the Groene Rivier in Roermond, the group installed a strain cable across a BoxBarrier wall of 25 boxes. Screwless clamps held the cable at the joints. One box was partly emptied on purpose. The system located that box and saw the first movement about fifteen minutes before failure, at about 400 microstrain — on the order of 0.12 mm of elongation over roughly 60 cm of cable.

Partners on the demonstrator included fibrisTerre, Waterschap Limburg, Dutch Water Prevention and RINA. An earlier prototype was tested at Flood Proof Holland.

Roermond demonstrator · MULTICLIMACT

Funded by STOWA · 2026–2028 · with HZ and Deltares

BEAVER

Burrowing Experimental Analysis of Vulnerability of Embankments due to Rodents. With Vana Tsimopoulou and Danny Janssen at HZ University of Applied Sciences, and with Deltares and STOWA, the group is preparing a real-scale test at Flood Proof Holland: a dike with pre-cast animal burrows, instrumented, loaded, and brought to failure.

The aim is to see how burrowing changes stability in real time, using the same combination of large-scale experiment, fiber-optic sensing, finite-element models and machine learning as the rest of the group.

Funded by NWA-NWO Idea Generator · 2021–2023

Fiber-Deep-Crack

Dike crack detection using distributed temperature sensing and deep learning. Cracks in the clay cover, from wetting and drying, roots and burrows, weaken the layer that is meant to keep water out of the core. Visual inspection and trenching are infrequent and labour-intensive, and they do not cover long stretches of dike.

The project treats drought, rather than high water, as the monitoring problem. A fiber-optic cable reads temperature along the dike as a proxy for moisture, while a thermal camera records the surface. A convolutional neural network is trained on both, at a crack-prone dike in Flood Proof Holland, so that cracks can be found along kilometres of defence in time to act. HKV is the societal partner.