基于独立分量分析的航空发动机振动信号盲源分离

阐述了基于峭度的盲源分离开关算法,对仿真信号进行分离,验证了该算法的可行性.并将该算法与FFT分离法结合,对某型双转子航空发动机高、低压转子实测振动信号进行了盲源分离实践,取得了很好的分离效果,从而为双转子发动机振动监测与故障诊断提供了一种可行的信号处理方法.     

基于小波分析的航空发动机故障诊断方法研究

小波变换是一种多分辨率的时频分析方法,应用在振动信号处理、故障诊断方面较传统方法优越。在阐述小波分析理论基础上,对某型航空发动机的振动信号进行分析,采用小波分解和信号重构的方法,提取了噪声掩盖下振动信号中的故障信息,根据航空发动机整机振动的典型故障特征频率,判断该发动机是否发生故障。     

小型航空活塞发动机喷油异常的故障诊断

针对小型航空活塞发动机出现的喷油异常故障,基于发动机的缸内压力和缸盖振动信号,采用一种变分模态分解和布谷鸟搜索优化支持向量机相结合的故障诊断方法对发动机喷油异常故障进行诊断。该方法使用变分模态分解对发动机的缸内压力信号和缸盖振动信号进行处理得到本征模态函数,对本征模态函数进行奇异值分解和能量特征提取,将缸内压力和缸盖振动的数据集输入布谷鸟搜索算法优化的支持向量机模型中进行训练和测试。结果表明:该方法较好地识别出发动机喷油异常的故障,其中缸内压力和缸盖振动信号的故障识别分类准确率分别为95.32%和92.47%,验证了该方法的有效性。      

齿轮故障识别的密度峰值聚类欠定盲源分离算法

为提高盲源分离算法在振动源数目估计问题中的噪声鲁棒性，提出了一种基于密度峰值聚类的欠定盲源分离算法。对预处理后的信号提取单源点，通过密度峰值聚类对单源点进行聚类得到混合矩阵的估计值。通过基于压缩感知模型对源信号进行重构，得到分离信号。为验证所提算法分离准确性和噪声鲁棒性，用所提算法对不同信噪比下的仿真信号进行分离，结果显示：在信噪比不低于4 dB时，所提算法均可以准确分离出源信号，算法的准确性和鲁棒性得到验证。设计旋转部件故障诊断试验台对所提算法在实际应用中的有效性进行验证，对实测复合故障振动信号进行分离，试验结果表明该算法成功分离出观测信号中的锥齿轮和行星齿轮单一故障特征，有助于工程中旋转部件故障诊断。     

局域均值分解中PF获取方法的改进

航空发动机故障振动信号为典型的多分量信号,局域均值分解方法可以对其分离。在研究了经典的局域均值分解方法后,发现相邻的两个乘积函数(Production Function,PF)中存在模态混淆现象,表现为较高频PF中包含较低频PF的频率成分。针对这一问题,对PF的定义进行了改进,滤除了混淆在其中的低频信号,使信号分离更加充分。仿真信号的分析表明,改进后的算法分离效果优于原算法。试验信号的分析表明,方法可以用于航空发动机故障特征提取。     

基于GA-AANN神经网络的SDQ算法的航空发动机传感器数据预处理

为实现对输入健康管理系统的航空发动机传感器数据进行数据鉴定、故障诊断以及去除噪声信号干扰,提出了一种航空发动机传感器数据预处理方法。针对双通道传感器航空涡扇发动机,搭建了以合理性检验模块和解析冗余检验模块为主要内容的SDQ算法模型,利用遗传算法优化的AANN神经网络实现传感器的解析冗余检验。采用蒙特卡罗仿真方法,将改进的SDQ算法与一种基于最小二乘法的SDQ算法进行对比仿真验证。结果表明,本文提出的SDQ算法在发动机稳态条件下对阶跃故障和漂移故障隔离的平均正确率分别提高了1.7%和19.1%,在发动机动态条件下对阶跃故障和漂移故障隔离的平均正确率分别提高了12.5%和33.8%。且在多传感器故障诊断和除噪方面性能优异,处理后的传感器信号平均信噪比提高了8.27dB。     

航空发动机中介轴承振动故障分析方法

针对中介轴承振动故障信号微弱、较难提取的问题,提出一种基于小波变换的航空发动机中介轴承故障诊断方法:首先对中介轴承故障信号进行小波分解,得到各层细节信号,并对细节信号进行重构;然后对重构信号进行频谱变换,从频谱图上清晰观察出中介轴承的故障特征频率。对真实发动机中介轴承故障信号进行的实例分析表明,本文方法具有较好的降噪能力,较频谱分析更能突出中介轴承的故障特征。     

并行鲁棒观测器对发动机减小转速信号的故障分离

通过对减小转速故障信号产生机理的研究 ,提出了一种基于多重模型假设下的并行鲁棒观测器残差输出的故障分离方法 ,建立了实际系统的正常工作状态和 4种故障工作模式 ,将该方法应用于某型航空涡扇发动机导致减小转速信号产生的传感器的故障检测与分离。数字仿真和误差分析结果显示 ,文中提出的方法可靠性高 ,故障定位准确 ,对发动机减小转速信号的故障分离行之有效。      

航空发动机振动信号盲分离研究

根据航空发动机振源信号的相互独立性,阐述了利用盲分离技术对航空发动机振动信号进行分离的原理和实现.利用Fast ICA算法对某型涡轮风扇发动机转子产生的振动信号进行了盲分离实践.     

基于小波包能量谱的航空发动机故障特征提取方法

使用成熟的旋转机械振动信号频谱分析方法,对采集到的多台航空发动机试车振动信号进行分析,找出能反应出发动机转子不对中、转子不平衡和转静件碰磨3种故障频谱特征的振动信号.再使用小波包将信号分解为不同的频段,之后分别计算能反应出故障信息的特征频段的能量,将它们组成用来区分上述3种故障的特征向量,为以后的航空发动机故障的模式识别做准备.     

Creating complex signal samples from a band-limited real signal

A very efficient method of creating complex signal samples from a band-limited real signal is presented. Because the method employs a simple mixer followed by one analog-to-digital (A/D) converter, plus a finite-duration impulse response (FIR) filter for image band rejection, there is no phase distortion in the resulting sampled signal. The method is more efficient than competing methods based on infinite-duration impulse response (IIR) filters     

Modified Costas signal

A modification to the Costas signal is suggested. It involves an increase of the frequency separation /spl Delta/f beyond the inverse of the subpulse duration t/sub b/ combined with adding linear FM (LFM) with bandwidth B, in each subpulse. Specific relationships between /spl Delta/f and B will prevent autocorrelation grating lobes, that would normally appear when t/sub b//spl Delta/f>1.     

相关法测量中的噪声信号源和扫频信号源的性能比较

时差法相关测量中信号的带宽越宽，互相关函数的峰值点的峰值优势越大，其测量的抗干扰性能越好。通过理论分析和数学仿真研究了两种宽带信号源——噪声信号和扫频信号，并利用相关分析法测量两信号问时差时的性能。结论是：噪声信号频带范围宽，带宽不受所采用信号长短的限制，测量分辨力高，但其频谱与噪声干扰频谱相似，抗干扰能力差；扫频信号的频率和带宽可人为控制，频谱特征明显，抗干扰性能好。     

Decentralized CFAR signal detection

The authors develop the theory of CA-CFAR (cell-averaging constant false-alarm rate) detection using multiple sensors and data fusion, where detection decisions are transmitted from each CA-CFAR detector to the data fusion center. The overall decision is obtained at the data fusion center based on some k out of n fusion rule. For a Swerling target model I embedded in white Gaussian noise of unknown level, the authors obtain the optimum threshold multipliers of the individual detectors. At the data fusion center, they derive an expression for the overall probability of detection while the overall probability of false alarm is maintained at the desired value for the given fusion rules. An example is presented showing numerical results     

Sensor systems signal processing platform

A field programmable gate array-based computing platform for high-speed sensors such as short-range radars is presented. The circuit performs necessary A/D conversions, raw data stream compression and target detection, and constructs a message structure suitable for external displays. In the prototype, USB is the connecting path to a laptop computer. An elementary pulse radar is utilized as an application example. An application interest would be in collision avoidance systems. Observed data transfer rates with real radar input signals were 36 Mbytes/s when typical target detection algorithms were applied.     

Development of random signal radars

Development of random signal radar (RSR) over the past 30 years is described. Conventional methods of implementing RSR are summarized such as correlation, spectrum analysis, and anticorrelation. Some typical RSR systems are introduced, for example, noise frequency modulation CW radar, random binary phase-coded CW radar, etc., and their merits and demerits are also pointed out. Finally, RSR development trends are analyzed     

Galileo AltBOC signal multiresolution acquisition strategy

Herein, some critical aspects related to the acquisition of the future Galileo signals are discussed. In particular, the Alternative Binary Offset Carrier (AItBOC) modulation that will be used by the Galileo satellites to broadcast navigation signals on the E5 band is considered, addressing acquisition issues only partially analyzed in previous works. The implementation of an acquisition section for the AltBOC signals is not straightforward, since several different receiver architectures can be used and remarkable differences are required with respect to the conventional signal processing used in GPS receivers. The main problems that must be handled (risk of false lock, resolution of the search space, and computational burden) are outlined in the following. An innovative technique, called multiresolution acquisition and tailored to the AltBOC signals, is then proposed as an effective solution to previous problems. As demonstrated by means of simulations, this novel strategy can be successfully used in coherent dual-band AltBOC receiver architectures with a feasible implementation complexity and it leads to remarkable advantages in terms of acquisition time.     

Recent advances in analog signal processing

Analog signal processors which perform convolution, correlation, matched filtering, and Fourier transformation are considered. Specifically, the authors focus on one- and two-dimensional signal processors using charge-coupled-device, surface-acoustic-wave, acoustooptic, and superconductive components. The device performances are compared in terms of parameters such as signal bandwidth, time-bandwidth product, and dynamic range. These analog components are compared with existing and planned digital processor designs. Potential applications are highlighted in wideband spread-spectrum packet-radio communication, radar range-Doppler imaging, wideband compressive receivers, and low-bit-rate image-coding systems     

GNSS receiver for GLONASS signal reception

This paper deals with the latest version of Experimental GNSS receiver built at the Czech Technical University and describes integration of GLONASS signal processing to the receiver. The new FPGA platform Virtex-D Pro by Xilinx is used and enables integration of whole digital signal processing of GNSS receiver into the single chip. The RE unit of the receiver is capable of processing all GLONASS frequency of the Li and L2 bands in two independent RE channels; each channel can process one band. The frequency selection of the appropriate satellite is accomplished in a digital correlator. The development flow of the GLONASS correlator is discussed herein. The complexity of the GLONASS correlator with complexity of GPS correlator is compared. The developed GLONASS correlator was tested in Simuelink tool during development. The next test was carried out using GLONASS simulator and real GLONASS satellite signal.     

Collapsing losses of peak signal integrators

Collapsing losses are computed for systems in which the peak return of K samples of noise plus one sample of signal-plus-noise are integrated over N looks. The statistical approach, collapsing losses, and an application are described. The peak integrator is found to have substantially lower collapsing losses than conventional systems in which the average, not the peak, is integrated