空气动力学学报2024，Vol.42Issue(5)：53-63,11.DOI:10.7638/kqdlxxb-2023.0053
基于神经网络与CFD相结合的扑翼推进性能优化
Optimization of flapping hydrofoil propulsion performance based on combined neural network and CFD
摘要
Abstract
In order to improve the propulsion performance of existing underwater hydrofoil robots,the Taguchi experiments,neural networks and CFD are combined to systematically study the effects of aspect ratio,heaving amplitude,pitching amplitude and flapping frequency on the propulsion performance of a three-dimensional NACA 0012 hydrofoil.First,the parameter combinations for CFD simulation are determined by the Taguchi method.Next,CFD simulations are performed and the results are used as the training set of the neural network.Then,the neural network is trained and used to predict the CFD result.Finally,the mechanism for the optimal propulsion performance at the optimized parameter combinations is analyzed.The results show that the aspect ratio,heaving amplitude,pitching amplitude,and flapping frequency can significantly affect the propulsion performance of the hydrofoil,among which,the flapping frequency(aspect ratio)has the greatest(least)influence on the propulsion performance.The maximum propulsion efficiency of the hydrofoil can reach 55.43%after the optimization of the neural network.Further analysis of the flow field structure around the hydrofoil with different parameters reveals that,under the optimal parameters there forms a stable vortex on the hydrofoil surface,which can stay on the hydrofoil surface for a long time during the flapping process.This is the intrinsic reason for the better propulsion performance at the optimal parameters.关键词
仿生扑翼/田口实验/神经网络/推进效率/数值模拟Key words
bionic hydrofoil/Taguchi experiments/neural network/propulsion efficiency/numerical simulation分类
航空航天引用本文复制引用
宋振宇,朱建阳,董璐..基于神经网络与CFD相结合的扑翼推进性能优化[J].空气动力学学报,2024,42(5):53-63,11.基金项目
国家自然科学基金(51975429) （51975429）
