Russian Scientists Use Neural Networks to Boost the Efficiency of Hydrodiodes
Researchers at Omsk Polytechnic University are training artificial intelligence to determine the ideal geometry of hydrodiodes, devices that control liquid flow. The project shows how neural networks are entering even highly specialized fields of engineering.
The breakthrough was achieved by Sergei Kaigorodov, an associate professor at the Department of Hydromechanics and Transport Machines, together with his team. They use neural networks to analyze data collected from hydrodiode experiments and train AI models to identify optimal values for different parameters.
According to Kaigorodov, the efficiency of a hydrodiode depends largely on the geometry of the channel through which liquid flows. Even small changes directly affect performance, making it essential to fine-tune the geometric parameters.
The researcher secured a grant of 1.5 million rubles (about 17 647 USD) to advance the project. The funding will support numerical studies using specialized software for hydrodynamic calculations, as well as the development of experimental prototypes and testing on dedicated stands.
