基于经验小波变换的模拟电路健康预测方法研究

发布时间：2018-01-07 03:04

本文关键词：基于经验小波变换的模拟电路健康预测方法研究　出处：《合肥工业大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着电子技术的快速发展,模拟电路结构日趋复杂,电路故障发生率越来越高。由于模拟电路在运行过程中,其故障表现为非线性和不确定性,一般故障特征提取方法很难准确并有效的提取电路的故障特征,利用新的技术进行模拟电路故障特征提取成为必然的趋势。经验小波变换(EWT)是近几年来发展起来的一种新的自适应信号分析方法,其主要用于处理非平稳、不确定随机信号。为了获得模拟电路最优的故障特征,本文将EWT方法运用到模拟电路故障特征提取中,提出了一种基于经验小波变的模拟电路故障特征提取新方法。为了验证该方法,本文通过特定的电路进行实验仿真,提取出了电路不同故障模式下故障信号的基频分量,并结合信号的能量熵,成功的提取出了电路的故障特征。仿真结果表明了本文所提方法的可行性和有效性。目前,国内外大多数学者都集中在模拟电路故障诊断等领域的研究,而对模拟电路健康预测方面的研究比较少见。现有的故障预测方法所提取出的故障特征数据不能很好的反应电路元件的退化程度,无法有效的对模拟电路的健康状态进行准确预测。由此,本文在经验小波变换方法的基础上,提出了表征电路元件参数退化程度的电路健康指数(HI)的概念,研究了模拟电路故障健康指数与故障特征之间的量化关系,建立了模拟电路健康状态的评价机制,结合BP神经网络预测模型,完成了模拟电路的健康预测。最后通过相关电路进行了实验验证,实验结果表明:基于EWT的模拟电路健康方法预测出的数据与元实际数据误差较小,能够比较准确的对元件故障退化情况进行预测,非常适合用于实际模拟电路的健康预测。
[Abstract]:With the rapid development of electronic technology, the structure of analog circuits is becoming more and more complex, and the incidence of circuit faults is becoming higher and higher. General fault feature extraction methods are difficult to extract the circuit fault features accurately and effectively. Using new techniques to extract fault features of analog circuits becomes an inevitable trend. Empirical wavelet transform (EWT) is a new adaptive signal analysis method developed in recent years. It is mainly used to deal with non-stationary and uncertain random signals. In order to obtain the optimal fault characteristics of analog circuits, this paper applies the EWT method to the fault feature extraction of analog circuits. A new method of fault feature extraction for analog circuits based on empirical wavelet transform is proposed. In order to verify this method, experimental simulation is carried out through specific circuits. The fundamental frequency components of the fault signal in different fault modes of the circuit are extracted and combined with the energy entropy of the signal. The simulation results show that the proposed method is feasible and effective. At present, most scholars at home and abroad focus on analog circuit fault diagnosis and other areas of research. However, the research on analog circuit health prediction is rare. The fault characteristic data extracted by existing fault prediction methods can not well reflect the degradation degree of circuit components. The health state of analog circuits can not be predicted accurately. Therefore, this paper based on empirical wavelet transform method. The concept of circuit health index (HI), which is used to characterize the degradation of circuit component parameters, is proposed, and the quantitative relationship between fault health index and fault characteristics of analog circuits is studied. The health evaluation mechanism of analog circuit is established, and the model of BP neural network is used to predict the health of analog circuit. The experimental results show that the data predicted by the health method of analog circuit based on EWT has less error than the actual data, and it can accurately predict the degradation of component faults. It is very suitable for health prediction of practical analog circuits.
【学位授予单位】：合肥工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN710

【相似文献】