基于CEEMDAN多尺度改进排列熵和SVM的空化噪声特征提取  

Feature extraction of cavitation noise based on CEEMDAN multi-scale improved permutation entropy and SVM

作　　者：兀成龙 高翰林 朱丹丹 李亚安[1] WU Chenglong;GAO Hanlin;ZHU Dandan;LI Ya’an(School of Navigation,Northwestern Polytechnic University,Xi’an 710072,China)

机构地区：[1]西北工业大学航海学院,西安710072

出　　处：《振动与冲击》2024年第13期190-197,216,共9页Journal of Vibration and Shock

摘　　要：当水下航行器处于高速航行时就会形成空化噪声,所产生的噪声会严重影响水下航行器的性能和安全。螺旋桨噪声包含着丰富的空化信息,是识别空化状态的有效手段。针对改进排列熵在单尺度下对原信号进行分析,无法有效区分不同空化状态,提出了将改进排列熵与自适应噪声完备经验模态分解(complete ensemble empirical mode decomposition with adaptive noise, CEEMDAN)相结合的空化噪声特征提取方法。首先,采用CEEMDAN方法对水下航行器螺旋桨的空化噪声进行分解,提取具有空化特征的固有模态函数(intrinsic mode function, IMF)分量;其次,选取相关系数最高的IMF分量并计算其多尺度改进排列熵(multi-scale improved permutation entropy, MIPE);最后,基于多尺度改进排列熵,建立支持向量机的特征分类模型。仿真和试验结果表明,该方法具有更好的可分性。An underwater vehicle is sailing at high speed to generate cavitation noise it can seriously affect performance and safety of the underwater vehicle.Propeller noise contains rich cavitation information,it is an effective means to identify cavitation status.Here,aiming at the problem of improved permutation entropy being unable to effectively distinguish different cavitation states in analyzing the original signal under a single scale,a cavitation noise feature extraction method combining improved permutation entropy to complete ensemble empirical mode decomposition with adaptive noise(CEEMDAN)was proposed.Firstly,CEEMDAN method was used to decompose cavitation noise of underwater vehicle propeller and extract intrinsic mode function(IMF)components with cavitation characteristics.Secondly,the IMF component with the highest correlation coefficient was chosen to calculate its multi-scale improved permutation entropy(MIPE).Finally,based on MIPE,a feature classification model for support vector machine(SVM)was established.The simulation and experimental results showed that the proposed method has better separability.

关 键 词：多尺度改进排列熵(MIPE) 自适应噪声完备经验模态分解(CEEMDAN) 空化噪声 特征提取 

分 类 号：TB566[交通运输工程—水声工程]