对转桨水动力性能实时预报方法  

Real Time Prediction Method for Hydrodynamic Performance of Contra-Rotating Propellers

作　　者：高楠 侯立勋 胡安康 侯志雅 常欣 GAO Nan;HOU Lixun;HU Ankang;HOU Zhiya;CHANG Xin(School of Naval Architecture and Ocean Engineering,Dalian Maritime University,Dalian 116026,Liaoning,China)

机构地区：[1]大连海事大学船舶与海洋工程学院,辽宁大连116026

出　　处：《船舶工程》2022年第3期61-66,共6页Ship Engineering

基　　金：辽宁省自然科学基金项目(2019ZD0161);中国博士后科学基金面上项目(2020M680935)。

摘　　要：为了实现对转桨水动力性能实时预报,基于BP神经网络构建对转桨水动力性能预报模型。首先,采用低阶速度势边界元法建立对转桨水动力性能预报模型,通过调整来流速度和前后桨转速开展对转桨水动力性能多工况计算,从而获得构建神经网络所需的样本空间。建立适用于对转桨水动力性能预报的神经网络架构,通过训练使其具备良好的泛化能力。以某组对转桨为研究对象开展水动力性能实时预报方法研究,结果表明,采用BP神经网络预报模型可获得与边界元法精度相当的预报结果,但该模型与边界元法相比计算所耗时间可以忽略不计,可有效实现对转桨水动力性能实时、快速预报。In order to realize the real-time prediction of the hydrodynamic performance of contra-rotating propellers, a hydrodynamic performance prediction model of contra-rotating propellers is built based on the back propagation(BP) neural network. Firstly, the low-order potential based boundary element method is used to establish the hydrodynamic performance prediction model of the contra-rotating propellers. By adjusting the inlet flow velocity and the rotational speeds of the front and rear propeller, the multi-condition calculations of the hydrodynamic performance of the contra-rotating propellers are carried out, and the sample space required for the construction of the neural network is obtained. A neural network framework for the hydrodynamic performance prediction of the contra-rotating propellers is established which had good generalization ability through training. A certain set of contra-rotating propellers is taken as the research object to carry out the hydrodynamic performance real-time forecast method investigation. The results show that the prediction results through the neural network model showed a good agreement with that of the boundary element method.However, the time consumption of the neural network model could be neglected compared with that of the boundary element method, but the calculation time of the model is negligible compared with the boundary element method, which can effectively realize the real-time and rapid prediction of the hydrodynamic performance of the propeller.

关 键 词：对转桨 水动力性能 边界元法 实时预报 BP神经网络 

分 类 号：U664.33[交通运输工程—船舶及航道工程]