复杂环境下船舶轴系推力信号辨识模型分析

Analysis of Identification Model of Ship Shafting Thrust Signal in Complex Environment

为解决复杂环境下的船舶轴系推力信号辨识问题,提出一种基于时频分析和深度学习的轴系推力信号辨识模型。该模型首先以短时傅里叶变换(STFT)方法对轴系推力信号与环境干扰信号的时频特征进行提取,以频段分析方法对两种信号的频率(周期)和能量等关键特征进行计算;然后以循环神经网络(RNN)方法对模型中2种信号的关键特征进行充分训练,得到经泛化后的深度学习辨识模型;最后,基于实验室模拟实船的轴系推力信号与环境干扰信号数据对模型进行仿真试验和验证。验证结果表明,该模型在复杂环境干扰下施加恒定推力与动态推力时均具备良好的辨识能力,可为船舶轴系推力信号辨识技术的研究提供参考。

In order to solve the problem of ship shafting thrust signal identification in complex environment,a shafting thrust signal identification model based on time-frequency analysis and deep learning is proposed.First,the short-time fourier transform(STFT)method is used to extract the time-frequency characteristics of shafting thrust signal and environmental disturbance signal,then the frequency(period)and energy of the two signals are calculated by frequency band analysis method,and then the key characteristics of the two signals in the model are fully trained by recurrent neural network(RNN)method,and the generalized deep learning identification model is obtained.Finally,the model is simulated and verified based on the data of shafting thrust signal and environmental disturbance signal simulated in laboratory.The verification results show that the model has good identification ability when applying constant thrust and dynamic thrust under complex environment disturbance,which provides reference for the research of ship shafting thrust signal identification.