Solution of hydrodynamic problems by complex modeling methods

Abstract

Abstract Pollution of the environment, especially the marine environment, is currently a rather serious problem. The impact of noise on the fragile biota of the Arctic with an increase in shipping in this region is quite strong. The manufacture of new low-noise devices at an enterprise engaged in the creation of surface and underwater vessels is a very costly process. Computer simulation systems can reduce enterprise costs. The author proposes a system for complex modeling of low-noise devices that could predict the approximate behaviour of a viscous incompressible fluid (gas) in the form of vorticity per single object, described by Navier-Stokes differential equations for the case of fluid motion in a limited space. The system includes: a computer model in the well-established FlowSimulation system and a machine learning model. According to the results of the experimental study, the author found that the model from FlowSimulation has a clearer picture, but for FlowSimulation this is the limit of possibilities, and for the machine learning system there is the possibility of further development to improve the image. In the future, the author intends to continue experiments on training the model in order to obtain a better image.