Can retention control be self-adaptive?

A key innovation of the proposed solution is the use of machine learning to forecast stormwater inflow and automatically control retention capacity in real time. This allows for a faster response to changing rainfall conditions, reduces the risk of sewer system overload and local flooding, and makes better use of the available reservoir capacity. The system can be pre-trained on synthetic data and then self-adapt based on actual rainfall-runoff events, without the need for constant supervision by specialists.
The solution is designed for local, underground retention systems that can integrate with existing facilities or be implemented as new installations. The collected water can be used for irrigating green spaces and for firefighting purposes. Underground reservoirs save space, reduce evaporation, and protect the water from light exposure, thereby minimizing the growth of microorganisms and the deterioration in water quality typical of open reservoirs.
The system’s localized nature reduces the need for integration with large-scale water supply infrastructure and facilitates its phased implementation. Solutions of this type are envisioned as the future standard for smart stormwater management. The presentation aims to encourage local governments, investors, and designers to apply these solutions in pilot projects as part of an implementation-oriented doctoral program—carried out in collaboration with the Wrocław University of Environmental and Life Sciences and Amiblu Poland, and with the support of the Ministry of Science and Higher Education.

An advocate of modern solutions that support environmental protection and the rational management of natural resources. An enthusiast of applying machine learning methods in engineering.
He contributed to the development of Amiblu’s solutions for stormwater retention. Currently, as part of his doctoral studies at the Wrocław University of Environmental and Life Sciences, he is developing solutions in the field of automation and sewer network management using neural network models.
An Amiblu Group expert in jacking and microtunneling technologies. He leads the team responsible for engineering work and project management related to the construction of primary cooling circuits for power plants in Ostrołęka, Grudziądz, Adamów, and Gdańsk. He leads a team of Amiblu engineers responsible for the implementation of GRP-based products in Central and Eastern Europe. He also has several years of experience in the repair and protection of concrete structures in the environmental engineering sector.