Underwater vehicle flow field result prediction method based on improved Unet + + network

Abstract

The invention provides an underwater vehicle flow field result prediction method based on an improved Unet + + network. According to the technical scheme, the method comprises the following steps: firstly, constructing a network model training data set based on CFD results under simulation parameter combination of multiple groups of underwater vehicle models; constructing an improved Unet + + network model, wherein the improved Unet + + network model is formed by sequentially cascading a down-sampling module, an up-sampling module and a layer-skipping connection module; training a network through a data set of a training model, finally, combining the obtained training model with a new simulation parameter combination label to achieve flow field result rapid regression prediction of position simulation parameters. The method has the advantages that the problem that the prediction definition precision is low in the prior art is solved through the provided improved Unt + + model, flow field regression results of more label combinations can be achieved in a second level, and the flow field result obtaining efficiency is greatly improved.

Description

technical field [0001] The invention relates to the technical field of the combination of computer fluid dynamics simulation and artificial intelligence, and relates to a method for predicting flow field results of underwater vehicles based on an improved Unet++ network. Background technique [0002] Since the 1950s, with the continuous development of the computer level and the great progress of numerical calculation methods, as an interdisciplinary subject of mathematics and fluid mechanics, Computational fluid dynamics (CFD) has developed rapidly. In recent years, the integration and development of machine learning and physical models has brought new research models to fluid mechanics and related engineering fields. [0003] At present, the computing performance of computers has doubled, and artificial intelligence technology has also developed rapidly. As the core driving force of a new round of technological revolution and industrial transformation, artificial intellige...

AI Technical Summary

Problems solved by technology

[0004] For complex simulation problems, the CFD method still consumes huge computing resources, especially the complex flow field simulation of underwater structures, which often requires technicians to spend a lot of time and effort on model processing and setting

At present, some researchers have conducted research on the regression prediction of flow field images using convolutional neural networks, but there are still problems of limited prediction accuracy and low quality of predicted images.

Images