多通道多余物微弱信号检测方法研究

发布时间：2018-03-20 16:41

本文选题：多余物　切入点：微弱信号　出处：《哈尔滨工业大学》2016年硕士论文　论文类型：学位论文

【摘要】：作为军用、航天、国防设备中的关键基础元件,密封电子元器件能否稳定、可靠地工作直接关系到设备和系统的可靠性。多余物作为密封电子元器件内部的杂质,由于其运动的随机性和不可控性,严重危害着元器件的可靠性。现有多余物检测装置及方法对0.02mg以上多余物识别准确率高达90%,但对质量小于0.02mg多余物产生的微弱信号检测的无能为力。因此开展多余物微弱信号的研究对于发展多余物检测技术,消除设备可靠性隐患至关重要。针对传统多余物检测设备难以对多余物微弱信号实现有效检测的问题,本文设计多余物微弱信号采集模块。分析微弱信号成因,对信号采集通道的各个环节进行重新设计。选用三通道高灵敏度谐振式声音传感器,设计三通道多级调理电路和四通道数据采集电路实现多余物微弱信号的采集。从电源、传输线、阻抗匹配、PCB抗干扰设计、屏蔽与接地等各个角度对噪声进行抑制,满足了信号采集过程中的低噪声要求,为信号的处理及检测奠定了良好的基础。针对PIND试验信号中多种噪声对干扰多余物微弱信号检测的问题,对声发射信号及PIND试验信号特性进行分析,确定试验信号中最重要的噪声种类为背景噪声和环境噪声。对于背景噪声,选择时频分析见长的小波分析方法,提出改进的阈值函数进行噪声滤除。对于环境噪声,借鉴语音增强的方法,提出基于卡尔曼滤波的环境噪声消除方法,并对去噪的结果进行分析。针对单一传感器捕捉的多余物信号微弱,甚至存在误差、丢失和畸变等问题,本文采用三通道声音传感器对PIND信号进行接收。通过选择合适的数据融合方式,分析传统加权融合方法中存在的不足,采用新的基于量子遗传算法的多通道数据算法对三通道多余物微弱信号数据进行融合。最后通过仿真实验对融合算法的有效性进行验证。针对幅值过于微小、湮没于噪声中,经软硬件降噪后仍无法使信噪比得到显著提升的多余物微弱信号,采用随机共振的方法进行信号的增强与提取。先对非周期随机共振及其响应结果、表征方式及产生机制进行研究。通过建立多余物信号模型,研究各参数对双稳态系统的影响,确定适合多余物微弱信号检测的条件及参数。实验表明,该方法对0.01mg多余物的检测准确率为66.7%。本文研究的多余物微弱信号检测方法,为多余物检测灵敏度的提高提供了依据和方法,所取得的成果也为其它领域微弱信号检测提供了一定参考。
[Abstract]:As the military, aerospace, defense equipment in key infrastructure components, sealing electronic components can stable and reliable work is directly related to the reliability of equipment and systems. Excess sealing as impurities of electronic components, due to the randomness of the movement and uncontrollability, seriously endanger the reliability of components. The existing multi residue detection device and the method of 0.02mg above foreignmaterial recognition accuracy rate is as high as 90%, but weak signal detection on the quality of the remainder is less than 0.02mg. So to study on the remainder incapable of action for the development of weak signal multi residue detection technology, eliminating equipment reliability is very important. In view of the traditional hidden remainder detection device is difficult to realize weak signal redundances the effective detection problem, this paper design foreignmaterial weak signal acquisition module. Analysis of causes of weak signal, the signal acquisition channel A link is designed again. The three channel high sensitivity resonant sound sensor, the design of three channel multi channel data acquisition and conditioning circuit four circuit foreignmaterial weak signal acquisition. From the power supply, transmission line, impedance matching, PCB anti-jamming design, to suppress the noise of each point of shielding and grounding, meet low noise signal acquisition process requirements, laid a good foundation for the processing and detection signal. To solve the problem of multiple PIND test signal in the noise interference in the detection of weak signal redundancy, the signal and PIND signal characteristics of acoustic emission test analysis, determine the most important types of noise test signal in the background noise and the environment noise. For background noise, choose good time-frequency analysis method of wavelet analysis, put forward the improved threshold function for noise filtering. For environmental noise, from the language of Yin Zengqiang The method of environmental noise elimination method based on Calman filter, and the denoising results were analyzed. For the remainder of single sensor to capture weak signal, and even the existence of errors, loss and distortion, this paper adopts three channel sound sensor signal of PIND receiver. By selecting the appropriate data fusion methods, analysis of the lack of the traditional weighted fusion method, the new algorithm based on multi channel data quantum genetic algorithm on the three channel foreignmaterial weak signal data fusion. Finally, through the simulation experiment on the effectiveness of the fusion algorithm is verified. The amplitude is too small, buried in the noise, the noise reduction of hardware and software are still unable to make the letter weak signal noise ratio redundances has been significantly improved, using the method of stochastic resonance enhancement and signal extraction. The aperiodic stochastic resonance and its response The characterization and mechanism were studied. Through the establishment of multi residue signal model, study the effects of each parameter on the bistable system, suitable for multi residue detection of weak signal conditions and parameters. Experimental results show that the detection of 0.01mg redundances accuracy of this new method of weak signal detection method for the remainder of this paper 66.7., provide a basis and method for redundances improved sensitivity, the results obtained will provide some reference for other areas of weak signal detection.

【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TN911.23

【相似文献】