摘要
研究了试验设计和近似模型技术,确定了回转体艇型曲线的6个控制参数,根据系列回转体阻力数值计算结果,建立了水下航行器阻力近似计算的响应面模型及径向基神经网络模型。通过系列模型试验验证了多种近似模型的拟合精度,与工程估算方法以及CFD方法进行了比较。结果表明,4阶响应面模型及径向基神经网络模型精度远高于工程经验公式,能够正确反映艇型参数对阻力结果的影响,为后续艇型优化设计带来一定的参考价值。
The design of experiment （DOE） and the approximative model are studied. The six parameters which determine the shape of rotatory submersible are confirmed. According to resistance numerical calculating result of the serial rotatory, the response surface model （RSM） and the radical basis function （RBF） of the resistance approximative calculation of submersibles＇ are established. The match precisions of several approximative models are verified by the experiments of serial models. It is compared with the engineering prediction method and CFD method. The result shows that the four-order polynomial response surface model and the radical basis function have higher level of precision than the engineering formula. The two formers can reflect the influence of the shape parameters on the resistant result, which has valuable reference for the later optimal design of the submersible shape.
出处
《船舶工程》
北大核心
2015年第6期18-21,26,共5页
Ship Engineering
基金
山东交通学院博士科研启动基金资助项目
关键词
回转体艇型
阻力估算
近似模型
模型试验
rotatory submersible
resistance prediction
approximative model
model experiment
