Weak Fault Feature Extraction of the Rotating Machinery Using Flexible Analytic Wavelet Transform and Nonlinear Quantum Permutation Entropy

Addressing the challenges posed by the nonlinear and non-stationary vibrations in rotating machinery,where weak fault characteristic signals hinder accurate fault state representation,we propose a novel feature extraction method that combines the Flexible Analytic Wavelet Transform(FAWT)with Nonlinear Quantum Permutation Entropy.FAWT,leveraging fractional orders and arbitrary scaling and translation factors,exhibits superior translational invariance and adjustable fundamental oscillatory characteristics.This flexibility enables FAWT to provide well-suited wavelet shapes,effectively matching subtle fault components and avoiding performance degradation associated with fixed frequency partitioning and low-oscillation bases in detecting weak faults.In our approach,gearbox vibration signals undergo FAWT to obtain sub-bands.Quantum theory is then introduced into permutation entropy to propose Nonlinear Quantum Permutation Entropy,a feature that more accurately characterizes the operational state of vibration simulation signals.The nonlinear quantum permutation entropy extracted from sub-bands is utilized to characterize the operating state of rotating machinery.A comprehensive analysis of vibration signals from rolling bearings and gearboxes validates the feasibility of the proposed method.Comparative assessments with parameters derived from traditional permutation entropy,sample entropy,wavelet transform(WT),and empirical mode decomposition(EMD)underscore the superior effectiveness of this approach in fault detection and classification for rotating machinery.

Cloud Resource Integrated Prediction Model Based on Variational Modal Decomposition-Permutation Entropy and LSTM

Predicting the usage of container cloud resources has always been an important and challenging problem in improving the performance of cloud resource clusters.We proposed an integrated prediction method of stacking container cloud resources based on variational modal decomposition(VMD)-Permutation entropy(PE)and long short-term memory(LSTM)neural network to solve the prediction difficulties caused by the non-stationarity and volatility of resource data.The variational modal decomposition algorithm decomposes the time series data of cloud resources to obtain intrinsic mode function and residual components,which solves the signal decomposition algorithm’s end-effect and modal confusion problems.The permutation entropy is used to evaluate the complexity of the intrinsic mode function,and the reconstruction based on similar entropy and low complexity is used to reduce the difficulty of modeling.Finally,we use the LSTM and stacking fusion models to predict and superimpose;the stacking integration model integrates Gradient boosting regression(GBR),Kernel ridge regression(KRR),and Elastic net regression(ENet)as primary learners,and the secondary learner adopts the kernel ridge regression method with solid generalization ability.The Amazon public data set experiment shows that compared with Holt-winters,LSTM,and Neuralprophet models,we can see that the optimization range of multiple evaluation indicators is 0.338∼1.913,0.057∼0.940,0.000∼0.017 and 1.038∼8.481 in root means square error(RMSE),mean absolute error(MAE),mean absolute percentage error(MAPE)and variance(VAR),showing its stability and better prediction accuracy.

Short-Term Prediction of Photovoltaic Power Generation Based on LMD Permutation Entropy and Singular Spectrum Analysis

The power output state of photovoltaic power generation is affected by the earth’s rotation and solar radiation intensity.On the one hand,its output sequence has daily periodicity;on the other hand,it has discrete randomness.With the development of new energy economy,the proportion of photovoltaic energy increased accordingly.In order to solve the problem of improving the energy conversion efficiency in the grid-connected optical network and ensure the stability of photovoltaic power generation,this paper proposes the short-termprediction of photovoltaic power generation based on the improvedmulti-scale permutation entropy,localmean decomposition and singular spectrum analysis algorithm.Firstly,taking the power output per unit day as the research object,the multi-scale permutation entropy is used to calculate the eigenvectors under different weather conditions,and the cluster analysis is used to reconstruct the historical power generation under typical weather rainy and snowy,sunny,abrupt,cloudy.Then,local mean decomposition(LMD)is used to decompose the output sequence,so as to extract more detail components of the reconstructed output sequence.Finally,combined with the weather forecast of the Meteorological Bureau for the next day,the singular spectrumanalysis algorithm is used to predict the photovoltaic classification of the recombination decomposition sequence under typical weather.Through the verification and analysis of examples,the hierarchical prediction experiments of reconstructed and non-reconstructed output sequences are compared.The results show that the algorithm proposed in this paper is effective in realizing the short-term prediction of photovoltaic generator,and has the advantages of simple structure and high prediction accuracy.

Microseismic signal denoising by combining variational mode decomposition with permutation entropy

Remarkable progress has been achieved on microseismic signal denoising in recent years,which is the basic component for rock-burst detection.However,its denoising effectiveness remains unsatisfactory.To extract the effective microseismic signal from polluted noisy signals,a novel microseismic signal denoising method that combines the variational mode decomposition(VMD)and permutation entropy(PE),which we denote as VMD–PE,is proposed in this paper.VMD is a recently introduced technique for adaptive signal decomposition,where K is an important decomposing parameter that determines the number of modes.VMD provides a predictable eff ect on the nature of detected modes.In this work,we present a method that addresses the problem of selecting an appropriate K value by constructing a simulation signal whose spectrum is similar to that of a mine microseismic signal and apply this value to the VMD–PE method.In addition,PE is developed to identify the relevant effective microseismic signal modes,which are reconstructed to realize signal filtering.The experimental results show that the VMD–PE method remarkably outperforms the empirical mode decomposition(EMD)–VMD filtering and detrended fl uctuation analysis(DFA)–VMD denoising methods of the simulated and real microseismic signals.We expect that this novel method can inspire and help evaluate new ideas in this field.

Morphology Similarity Distance for Bearing Fault Diagnosis Based on Multi-Scale Permutation Entropy

Bearings are crucial components in rotating machines,which have direct effects on industrial productivity and safety.To fast and accurately identify the operating condition of bearings,a novel method based on multi⁃scale permutation entropy(MPE)and morphology similarity distance(MSD)is proposed in this paper.Firstly,the MPE values of the original signals were calculated to characterize the complexity in different scales and they constructed feature vectors after normalization.Then,the MSD was employed to measure the distance among test samples from different fault types and the reference samples,and achieved classification with the minimum MSD.Finally,the proposed method was verified with two experiments concerning artificially seeded damage bearings and run⁃to⁃failure bearings,respectively.Different categories were considered for the two experiments and high classification accuracies were obtained.The experimental results indicate that the proposed method is effective and feasible in bearing fault diagnosis.

INVESTIGATION OF THE SPATIAL AND TEMPORAL DISTRIBUTION OF EXTREME HIGH TEMPERATURE IN CHINA WITH DETRENDED FLUCTUATION AND PERMUTATION ENTROPY

With temperatures increasing as a result of global warming,extreme high temperatures are becoming more intense and more frequent on larger scale during summer in China.In recent years,a variety of researches have examined the high temperature distribution in China.However,it hardly considers the variation of temperature data and systems when defining the threshold of extreme high temperature.In order to discern the spatio-temporal distribution of extreme heat in China,we examined the daily maximum temperature data of 83 observation stations in China from 1950 to 2008.The objective of this study was to understand the distribution characteristics of extreme high temperature events defined by Detrended Fluctuation Analysis(DFA).The statistical methods of Permutation Entropy(PE)were also used in this study to analyze the temporal distribution.The results showed that the frequency of extreme high temperature events in China presented 3 periods of 7,10—13 and 16—20 years,respectively.The abrupt changes generally happened in the 1960s,the end of 1970s and early 1980s.It was also found that the maximum frequency occurred in the early 1950s,and the frequency decreased sharply until the late 1980s when an evidently increasing trend emerged.Furthermore,the annual averaged frequency of extreme high temperature events reveals a decreasing-increasing-decreasing trend from southwest to northeast China,but an increasing-decreasing trend from southeast to northwest China.And the frequency was higher in southern region than that in northern region.Besides,the maximum and minimum of frequencies were relatively concentrated spatially.Our results also shed light on the reasons for the periods and abrupt changes of the frequency of extreme high temperature events in China.

Adaptive Bearing Fault Diagnosis based on Wavelet Packet Decomposition and LMD Permutation Entropy

Bearing fault signal is nonlinear and non-stationary, therefore proposed a fault feature extraction method based on wavelet packet decomposition （WPD） and local mean decomposition （LMD） permutation entropy, which is based on the support vector machine （SVM） as the feature vector pattern recognition device Firstly, the wavelet packet analysis method is used to denoise the original vibration signal, and the frequency band division and signal reconstruction are carried out according to the characteristic frequency. Then the decomposition of the reconstructed signal is decomposed into a number of product functions （PE） by the local mean decomposition （LMD） , and the permutation entropy of the PF component which contains the main fault information is calculated to realize the feature quantization of the PF component. Finally, the entropy feature vector input multi-classification SVM, which is used to determine the type of fault and fault degree of bearing The experimental results show that the recognition rate of rolling bearing fault diagnosis is 95%. Comparing with other methods, the present this method can effectively extract the features of bearing fault and has a higher recognition accuracy

基于tSNE多特征融合的JTC轨旁设备故障检测

无绝缘轨道电路(Jointless Track Circuit,JTC)的轨旁设备在室外长期运营过程中,其可靠性会逐渐降低,进而给列车行车安全带来严重威胁。以轨道电路读取器(Track Circuit Reader,TCR)感应电压为基础,针对JTC故障诊断研究中轨旁设备故障类型复杂和故障特征提取不充分等问题,提出一种基于t分布随机邻域嵌入(t-distribution Stochastic Neighbor Embedding,tSNE)多特征融合的JTC轨旁设备故障检测模型。首先,根据不同轨旁设备故障对TCR感应电压信号的影响,分析各轨旁设备的故障特性。其次,提取TCR感应电压信号的方差、有效值、峰值因子等幅值域特征,以及排列熵、散布熵特征构成原始故障特征集。为了去除其中的冗余信息,得到具有较高判别性的融合流形特征,利用tSNE算法进行特征融合。最后输入深度残差网络(Deep Residual Network,DRN)得到故障检测混淆矩阵,实现轨旁设备故障定位。实验结果表明:tSNE算法融合后的特征在异类和同类故障样本之间分别有较大的类间间距和较小的类内间距,相比主成分分析(Principal Component Analysis, PCA)、随机相似性嵌入(Stochastic Proximity Embedding, SPE)、随机邻域嵌入(Stochastic Neighbor Embedding,SNE)算法具有更优的融合特征提取效果。此外,结合DRN可以有效识别多种轨旁设备故障,达到98.28%的故障检测准确率。通过现场信号进行实例验证,结果表明该故障检测模型能满足铁路现场对室外设备进行故障定位的实际需求。

小净距隧道掘进爆破及其振动响应规律研究

为了研究爆破荷载作用下小净距隧道中夹岩区的动力稳定性问题,依托小龙门隧道爆破工程,开展了现场爆破振动监测试验。通过改进的变分模态分解(variational mode decomposition,VMD)与多尺度排列熵(multi-scale permutation entropy,MPE)算法对爆破振动信号进行消噪处理,基于此分析了掏槽孔与周边孔爆破在后行洞左拱腰(非中夹岩区)、右拱腰(中夹岩区)中产生的振动特征差异。结果表明:采用改进的自适应VMD-MPE算法可以有效消除振动信号中的噪声,并降低了主观决策的影响;此外,相对于非中夹岩区,中夹岩对爆破振动具有明显的放大效应,其质点峰值振速明显大于非中夹岩区,但中夹岩区的振动衰减速度更快;同时,通过对比非中夹岩区与中夹岩区各测点振动频率特征可以发现,中夹岩区小于40 Hz的低频振动能量占比较大,更易引起支护结构的共振,发生损伤与破坏的风险更高,应重点关注;受“转角削弱”作用以及地震波传播路径的影响,在比例距离SD小于等于11.57 m·kg^(1/3)范围内,周边孔爆破在掌子面后方围岩中产生的振速大于掏槽孔。

基于小波散射变换和MFCC的双特征语音情感识别融合算法

为了充分挖掘语音信号频谱包含的情感信息以提高语音情感识别的准确性,提出了一种基于小波散射变换和梅尔频率倒谱系数(Mel-frequency cepstral coefficient,MFCC)的排列熵加权和偏差调整规则的语音情感识别融合算法(PEW-BAR)。算法首先获取语音信号的小波散射特征和梅尔频率倒谱系数的相关特征;然后按尺度维度扩展小波散射特征,利用支持向量机得到情感识别的后验概率并获得排列熵,并使用排列熵对后验概率进行加权;最后采用一种偏差调整规则进一步融合MFCC的相关特征的识别结果。实验结果表明,在EMODB、RAVDESS和eNTERFACE05数据集上,与传统的基于小波散射系数的语音情感识别方法相比,该算法将ACC分别提高了2.82%、2.85%和5.92%,将UAR分别提升了3.40%、2.87%和5.80%,IEMOCAP上提高了6.89%。

基于改进变分模态分解和优化堆叠降噪自编码器的轴承故障诊断

针对滚动轴承在噪声干扰下故障特征难以提取的问题,提出一种改进变分模态分解(VMD)和复合缩放排列熵(CZPE)的特征提取新方法,并利用优化堆叠降噪自编码器(SDAE)进行故障分类。首先,提出由“余弦相似度—峭度—包络熵”新综合评价指标自适应优化分解参数的改进VMD方法,并通过该指标筛选分解后的本征模态函数(IMF)分量;然后,为提取更全面的故障特征,引入新的复合缩放排列熵对各有效IMF的故障特征进行量化;最后,提出一种基于鼠群优化算法(RSO)与麻雀搜索算法(SSA)的混合算法优化SDAE网络超参数,将故障特征输入优化后SDAE网络中得到分类结果。采用美国CWRU轴承数据集进行验证,实验结果表明该方法能全面稳定地提取背景噪声下的故障特征,且与其他方法相比具有更好的抗噪性能和更高的故障诊断准确率。

基于信号图像化和CNN-ResNet的配电网单相接地故障选线方法

配电网发生单相接地故障时,零序电流呈现较强的非线性与非平稳性,故障选线较为困难,针对此问题,提出一种基于信号图像化和卷积神经网络-残差网络的配电网单相接地故障选线方法。首先,利用排列熵优化变分模态分解算法的参数,将零序电流信号分解成一系列固有模态函数;其次,引入新的数据预处理方式,将固有模态函数转成二维图像,获得零序电流信号的时频特征图;最后,利用一维卷积神经网络提取零序电流信号的相关性和特征,利用残差网络提取时频特征图的特征,将两个网络融合,构建混合卷积神经网络结构,实现故障选线。仿真与实验结果表明,该方法能够在高阻接地、采样时间不同步、强噪声等情况下准确地选择出故障线路,可满足配电网对故障选线准确性和可靠性的需求。

室内地埋供热管道泄漏的声学监测及其信号去噪方法研究

为了更精确地对室内地埋供热管道泄漏声学监测信号进行主要特征频率提取和分析,需先对测量信号去噪。本文采用变分模态分解(variational mode decomposition,VMD),对距离漏水点5 cm和40 cm测量点的原始声信号分别进行模态分解,计算出各模态分量的排列熵并将其作为噪声信号剔除的依据,最后对信号进行重构,并与经验模态分解(empirical mode decomposition,EMD)、集合经验模态分解(ensemble empirical mode decomposition,EEMD)的处理结果进行对比。发现相比于其他两种分解降噪方法,VMD能够更好地解决模态混叠问题,能更为精确地将噪声信号去除,达到更好的去噪效果。

一种针对驻留转换雷达的信号分选算法

复杂电磁环境下驻留转换雷达的信号分选是制约电子侦察技术发展的瓶颈,由于其复杂多变的成形规律,信号子周期难以被检测;脉冲重复周期调制类型不明,脉冲序列无法成功提取;提取的多个子周期脉冲序列无法合并为一部雷达信号,造成虚警。针对以上问题,提出一种针对驻留转换雷达的信号分选算法,利用双门限提高子周期检测概率,通过霍夫变换结合迭代自组织聚类判断信号是否属于驻留转换雷达,改进搜索算法以提取脉冲序列,依据排列熵指标将多脉冲序列合并为一部驻留转换雷达信号。仿真实验结果表明,所提算法在25%脉冲丢失率的复杂电磁环境下,可以将三部驻留转换雷达信号成功分选。

基于VMD-多尺度排列熵和SVM的船用空压机故障诊断方法

船用机械振动信号存在非线性、非平稳性问题,故障特征难提取,通过变分模态分解(variational mode decomposition,VMD)多尺度排列熵(multiscale permutation entropy,MPE)与支持向量机(support vector machine,SVM)融合的故障诊断方法,对振动信号进行研究。以空压机为例,首先,模拟6种空压机工况,对各工况的热工参数进行测试,分析各工况热工参数的变化程度,并对采集的振动信号进行频域分析。然后通过VMD对振动信号进行分解,得到一系列固有模态分量,计算与原始信号的互相关系数筛选敏感固有模态分量。最后计算出敏感固有模态分量的多尺度排列熵,将其作为特征向量,输入到SVM中,进行故障辨识。实验结果表明:VMD多尺度排列熵与SVM融合的空压机故障辨识方法,能有效地识别故障类型,整体准确率可保持在98.6667%,将此方法与其他方法进行对比,证明此方法有效。

基于PE-CEEMD-SVD的Φ-OTDR信号降噪方法

为实现相位敏感光时域反射仪中相位信号的精确测量,提出了一种基于排列熵算法的互补集合经验模态分解联合奇异值分解的新型降噪方法(PE-CEEMD-SVD)。首先,对含有噪声的相位信号进行CEEMD分解,得到一系列频率不同的IMF分量;然后,将PE算法和相关系数机制相结合,保留较大相关的有用分量,对较小相关的噪声分量使用SVD算法进行二次降噪;最后将两次降噪后保留下来的有用分量进行重构。仿真和实验结果表明,相较于EMD、EEMD和CEEMD降噪方法,该方法可获得更高信噪比的信号,有利于相位信号的精确测量。

CEEMDAN-WPE-CLSA超短期风电功率预测方法研究

提出了一种结合自适应噪声完全集合经验模态分解、加权排列熵、卷积神经网络、长短期记忆网络和自注意力机制的超短期风电功率预测方法。首先,利用自适应噪声完全集合经验模态分解将原始风电功率时间序列自适应分解为一系列的模态分量,降低原始序列的非线性和波动性;其次,根据加权排列熵计算各模态分量间的相似性并对相似的分量进行重组,以修正自适应噪声完全集合经验模态分解的过度分解问题,使得修正后的模态分量更具规律性;最后,将重组后的分量输入卷积长短期记忆网络进行时序建模,并利用自注意力机制对卷积长短期记忆网络的神经元权重进行重新分配,提高了卷积长短期记忆网络对输入特征不确定性的适应能力。在此基础上,明确了自注意力机制和自适应噪声完全集合经验模态分解、加权排列熵在风电功率预测中的作用机制,以及风电功率信号包含的重要物理信息,证明了自适应噪声完全集合经验模态分解、加权排列熵以及自注意力机制在风电功率信号模态分解和长短期记忆网络隐层输出权重分配中的有效性。

RCMNAAPE在旋转机械故障诊断中的应用

针对精细复合多尺度排列熵(RCMPE)无法充分提取旋转机械振动信号中的故障信息,从而导致旋转机械故障识别准确率不稳定这一缺陷,提出了一种基于精细复合多尺度归一化幅值感知排列熵(RCMNAAPE)、拉普拉斯分数(LS)和灰狼算法优化支持向量机(GWO-SVM)的旋转机械故障诊断方法。首先,利用幅值感知排列熵替换了RCMPE中的排列熵,提出了RCMNAAPE,并将其用于提取旋转机械振动信号的故障特征生成特征样本;随后,采用了LS从原始的高维故障特征向量中筛选出较少的能够更准确描述故障状态的特征,构造敏感特征样本;最后,将低维的故障特征向量输入由灰狼算法优化的支持向量机中进行了训练和测试,完成了旋转机械样本的故障识别和分类,利用滚动轴承和齿轮箱故障数据集将RCMNAAPE-LS-GWO-SVM与其他故障诊断方法进行了对比分析,并开展了评估。研究结果表明:基于RCMNAAPE-LS-GWO-SVM的故障诊断方法能够有效识别旋转机械的各类故障,其识别准确率高于其他对比的故障诊断方法,其中滚动轴承故障的识别准确率达到99.33%,齿轮箱故障的识别准确率达到98.67%。虽然,该方法的特征提取效率不佳,平均特征提取时间分别为153.02 s和163.98 s,仅优于精细复合多尺度模糊熵(RCMFE),但其综合性能更加优异。

基于改进小波阈值—CEEMDAN的变压器局部放电超声波信号白噪声抑制方法

为了有效去除变压器局部放电超声信号中的白噪声干扰,提高后续局部放电模式识别及定位的准确性,提出了一种基于改进小波阈值和自适应噪声完全集合经验模态分解(complete ensemble empirical mode decomposition with adaptive noise,CEEMDAN)的变压器局部放电超声波信号去噪方法。首先,通过对放电信号进行CEEMDAN分解得到一系列由高频到低频的本征模函数IMF(intrinsic mode function);然后,利用多尺度排列熵(multi-scale permutation entropy,MPE)算法计算各阶IMF分量的排列熵PE(permutation entropy),根据各IMF的排列熵值确定信号的去噪阈值与有效阈值。对高于去噪阈值的IMF分量采用改进小波阈值法进行去噪处理,对低于有效阈值的IMF分量视为基线漂移进行剔除。最后,通过重构去噪分量与剩余分量来获得去噪后的超声波信号。仿真和实验结果均表明,文中所提出的去噪算法大大提高了信号的信噪比,并保留了原始超声波信号中的有效信息,对提高后续利用超声波信号进行局部放电模式识别及定位的精确性具有重要意义。

基于变分模态分解的休息态虚拟现实晕动症脑电自动检测

虚拟现实晕动症的存在是制约VR技术行业进一步发展的关键因素,研究虚拟现实晕动症相关的神经活动及对其准确检测是解决此问题的前提,此前研究缺少对休息态虚拟现实晕动症神经活动的研究。因此,本研究利用虚拟现实晕动症暴露任务前后休息态脑电信号,提出虚拟现实晕动症脑电特征作为指标实现对虚拟现实晕动症的检测。首先,通过统计分析对所选的5个电极即Fp1、Fp2、F8、T7及T8的脑电信号分别进行变分模态分解,并从选中的模态分量中提取样本熵、排列熵及中心频率。然后,通过统计检验和ReliefF算法进行两个阶段的特征选择。最后,将选择的特征向量送入支持向量机中进行分类,进而实现对虚拟现实晕动症的自动检测。结果表明,此方法准确率、灵敏度及特异度分别达到了98.3%、98.5%及98.1%,ROC曲线下的面积值达到了1,优于其他方法,证明了此方法在虚拟现实晕动症脑电信号自动检测方面优势与有效性。