摘要
模拟元件的非线性属性导致电路特征集异常庞大,当故障电路响应出现一定带宽的随机分量与主频分量相互叠加,信号频谱中的故障特征频率及其2、3次谐波附近的边频带均出现显著增长。由于提升小波算法的预测和更新原理与故障信息紧密相关,预测器和更新器设计可替代小波基的选取。本文首先将响应信号经过提升小波变换,对蕴含大量故障信息的高频细节部分进行Hilbert调制分析,从包络谱中剔除载频的常规分量,随后通过特征频率识别技术实现数据降维处理和故障定位。最后通过一个实例验证了该方法完全胜任模拟电路故障特征频率识别。
The inherent nonlinear properties of analog component results in large circuit feature set,when the ran-dom components with certain bandwidth and main frequency components are superimposed in analog circuit response,the fault feature frequency and the sideband around the second and third harmonic are all increased in signal spectrum.For predicting and updating principle in lifting wavelet algorithm closely related to fault information,the predictor and updater are designed to replace the wavelet basis selection.The response signal is decomposed by lifting wavelet,the high fre-quency parts contained a large number of fault information is analyzed by Hilbert modulation,the signal spectrum enve-lope removed the conventional components for carrier frequency effectively,the fault data dimension reduction and fault location are realized by feature frequency identification technology.Finally,an illustration verifies this method fully qualified for the fault feature frequency identification in analog circuit.
出处
《电子测量与仪器学报》
CSCD
2011年第8期700-703,共4页
Journal of Electronic Measurement and Instrumentation
基金
国家自然科学基金(编号:61001049)资助项目
北京自然科学基金(编号:4112012)资助项目
