螺旋桨式水下推进器推力的大小决定着水下机器人的水动性能。为快速准确预测螺旋桨的敞水性能，建立一种RBF神经网络螺旋桨敞水性能估计器模型。利用几种型号螺旋桨敞水仿真值作为训练样本，对网络规模进行调整。在此基础上，广义逆动态地调整各网络间的连接权重，进行网络参数的优化。通过不断地迭代优化达到学习精度要求，最终得到一种高维优化的神经网络敞水性能估计器。同时，利用CFD仿真软件对螺旋桨进行模拟仿真和数值计算。对比分析RBF神经网络螺旋桨敞水性能估计器模型预测的敞水系数与CDF仿真的敞水系数，结果表明两者之间的差距较小，故RBF神经网络敞水性能估计器模型满足准确预测和快速性的要求，能够作为螺旋桨敞水系数的有效预测方式之一。
The hydrodynamic performance of the underwater robot is determined by the thrust of the propeller-type underwater propeller. In order to predict the open water performance of the propeller quickly and accurately, an open water performance estimator model which is based on RBF neural network needs to be established. The network size is adjusted by using several types of propeller open water simulation values as training samples.On this basis, the connection weights between the networks have been adjusted, and the network parameters have also been optimized. After achieving the learning accuracy requirements through continuous iterative optimization, a high-dimensional optimized neural network open water performance estimator is finally obtained. By comparing and analyzing the open water coefficient predicted by the RBF neural network propeller open water performance estimator model and the open water coefficient simulated by CDF, the result shows that the gap between them is small. Therefore, the RBF neural network open water performance estimator model can meet the requirements of accuracy and rapidity of prediction, and can also be used as one of the effective prediction methods of propeller open water coefficient.
2022,44(5): 54-58 收稿日期：2021-03-12
DOI：10.3404/j.issn.1672-7649.2022.05.011
分类号：U661
作者简介：祝建平(1987-)，男，硕士研究生，讲师，主要从事工业产品设计方面工作
参考文献：
[1] 王超, 黄胜, 解学参. 基于CFD方法的螺旋桨水动力性能预报[J]. 海军工程大学学报, 2008, 20(4): 107-112
WANG Chao, H UANG Sheng, XIE Xue-shen. Hydrodynamic performance prediction of some propeller based on CFD[J]. Journal of Naval University of Engineering, 2008, 20(4): 107-112
[2] 孙群, 范佘明. 基于BP神经网络的螺旋桨敞水性能数值预报的修正方法[C]// 第十二届中国国际船艇展暨高性能船学术报告会. 2007.
[3] 翟鑫钰, 陆金桂. 基于神经网络的螺旋桨敞水性能预测[J10L]. 南京工业大学学报(自然科学版): 1-[12, 2022-3-27].
[4] 彭翔, 田中文, 何珍, 等. 不同湍流模型在螺旋桨敞水性能预报中的应用分析[J]. 舰船科学技术, 2019, 41(21): 46-49+53
PENG Xiang, TIAN Zhong-wen, HE Zhen, et al. Application and analysis of different turbulence models in the prediction of open water performance of a propeller[J]. Ship Science and Technology, 2019, 41(21): 46-49+53
[5] 张力为. 导管螺旋桨的空化及噪声性能数值分析[D]. 武汉: 武汉理工大学, 2019.
[6] 孙承亮, 赵江滨. 基于CFD的对转螺旋桨水动力性能分析[J]. 舰船科学技术, 2019, 41(3): 36-40
[7] 李家盛, 张振果, 华宏星. 基于有限元和面元法的弹性螺旋桨流固耦合特性分析[J]. 振动与冲击, 2018, 37(21): 14-21
LI Jiasheng, ZHANG Zhenguo, HUA Hongxing. Hydro-elastic analysis for dynamic characteristics of marine propellers using finite element method and panel method[J]. Journal of Vibration and Shock, 2018, 37(21): 14-21
[8] 袁健, 丁晓阳. 基于数值计算的舰船可调螺距螺旋桨水动力性能分析[J]. 舰船科学技术, 2017, 39(18): 13-15
YUAN Jian, DING Xiao-yang. Hydrodynamic performance analysis of controllable pitch propeller based on numerical calculation[J]. Ship Science and Technology, 2017, 39(18): 13-15
[9] 王淑生, 王超, 仇宝云, 等. 基于CFD轴流泵叶轮在流场中的有限元分析[J]. 机械工程与自动化, 2016(6): 82-83+86
[10] 李雪芹, 陈科, 刘刚. 基于ANSYS的复合材料螺旋桨叶片有限元建模与分析[J]. 复合材料学报, 2017, 34(4): 591-598
LI Xue-qin, CHEN Ke, LIU Gang. Finite element modeling and analysis of composite propeller blade based on ANSYS[J]. Acta Materiae Compositae Sinica, 2017, 34(4): 591-598
[11] 王本武, 李庆刚. 轴流泵叶片有限元分析[J]. 科协论坛(下半月), 2013(3): 21-23
