基于机器学习算法的舰面流场预警系统研究

Early Warning System of Warship Flow Field Based on Machine Learning

中文关键词:  神经网络  支持向量机  数值模拟  舰面流场  预警系统

英文关键词:neural network  support vector machine  numerical simulation  warship flow field  early warning system

中文摘要:

      飞行甲板是直升机海上作战的重要平台，为了提高直升机的作战效能及舰面操作的安全性，研究开发了一套舰面流场预警系统。该系统主要通过分析数值计算所得的流场数据，以舰面关键位置的压强作为输入，利用神经网络算法和支持向量机训练出一套模型，准确地预测出舰面流场的主要特征（包括来流速度大小、方向和起降点涡结构等），从而实现舰船表面危险气流场的实时预报。预测结果表明，该预警系统的预测准确性较高，能够快速地预测来流情况及旋涡的大小、位置，为飞行员的舰面安全起降提供了参考，对加强国防建设有着重要意义。

英文摘要:

      The flight deck is an important platform for helicopter maritime operations.In order to improve the combat effectiveness of helicopters and the safety of operations on warship,a kind of early warning system for dangerous flow field on warship is proposed.The system mainly analyzes the flow field data obtained by numerical calculations,which takes the pressure of key positions on ship as input and uses neural network algorithm as well as support vector machine to train the model.It can accurately predict the important characteristics of the ships flow field including the magnitude,direction of incoming flow and the vortex structure around take-off and landing points,so as to realize the real-time forecast of the dangerous airflow field on the warship surface.The prediction results show that the early warning system has high accuracy and can quickly predict the situation of incoming flow,and the size and position of vortices.It provides a reference for pilots to accomplish the safe take-off and landing on warship,and has great significance for strengthening national defense construction.