·1106 工程科学学报，第39卷，第7期 征偏离度作为评价选择最优小波包基，构建了最优故 based UCISVM.Neurocomputing,2016,173:1752 障特征向量，在软故障类别数量相同的条件下，诊断率 ] Aminian M,Aminian F.Neural-network based analog-circuit fault 达到了100%，成功诊断区分了文献中未能区分的4 diagnosis using wavelet transform as preprocessor.IEEE Trans Circuits Syst II:Analog Digital Signal Process,2000,47(2): 种故障类别.在文献9]的故障诊断方法中，以信息 151 熵和峭度两个特征参数作为分类器的输入，而本文只 [8]Aminian F,Aminian M,Collins H W.Analog fault diagnosis of 提取了归一化能量作为唯一一个特征参数，却达到了 actual circuits using neural networks.IEEE Trans Instrum Meas, 更好的分类效果 2002,51(3):544 ]Peng M F.Tse C K,Shen M E,et al.Fault diagnosis of analog 5结语 circuits using systematic tests based on data fusion.Circuits Syst Signal Process,2013,32(2):525 (1)本文提出了一种基于小波包分解和GA-BP [10]Sheikhan M,Sha bani AA.PSO-optimized modular neural net- 的模型用于开关电流电路的软故障诊断检测.以开关 work trained by OWO-HWO algorithm for fault location in analog 电流电路中的六阶切比雪夫低通滤波器作为测试电 circuits.Neural Comput Appl,2013,23(2):519 路，通过GA算法优化BP神经网络的权值和阈值.诊 [01] Zhao D S,He YZ.A new test points selection method for analog 断结果表明，提出的诊断模型对开关电流电路具有良 fault dictionary techniques.Analog Integr Circuits Signal 好的故障分类性能. Process,2015,82(2):435 (2)将本文方法与其他文献方法作了比较分析. [2]Jiang YY,Wang YR,Luo H.Fault diagnosis of analog circuit based on a second map SVDD.Analog Integr Circuits Signal 这对于进一步完善开关电流电路故障诊断模型理论体 Process,2015,85(3):395 系具有参考意义 [13]Han H,Wang H J,Tian S L,et al.A new analog circuit fault diagnosis method based on improved Mahalanobis distance.J 参考文献 Electron Testing,2013,29(1):95 04] Guo J R,He Y G,Tang S X,et al.Switched-current circuits [1]Toumazou C,Hughes J B,Battersby N C.Switched-Currents,an test using pseudo-random method.Analog Integr Circuits Signal Analogue Technique for Digital Technology.London:Peter Pereg- Process,2007,52(1):47 rinus Ltd,1993 ū5] Guo J R,Cai X H,He Y G.PRBS test signature analysis of Long Y.The Study of Suitched Current Circuit Test and Fault Di- switched current cireuit/2009 1st International Conference on agnosis Methods [Dissertation].Changsha:Hunan University, Information Science and Engineering (ICISE).Nanjing,2009: 2012 627 (龙英.开关电流电路测试与故障诊断方法研究[学位论文] 06 Long Y,He Y G,Liu L,et al.Implicit functional testing of 长沙：湖南大学，2012) switched current filter based on fault signatures.Analog Integr B]Spence H.Automatic analog fault simulation /Proceedings of Au- Circuits Signal Process,2012,71 (2):293 to Test Conference Test Technology and Commercialization.Day- 07] Guo J R,He Y G,Liu M R.Wavelet neural network approach ton,1996 for testing of switched-current circuits.J Electr Test,2011,27: [4]Jantos P.Grzechca D,Rutkowski J.A global parametric faults di- 611 agnosis with the use of artificial neural networks//ECCTD 2009 08] Long Y,He Y G,Yuan L F.Fault dictionary based switched European Conference on Circuit Theory and Design.Antalya, current circuit fault diagnosis using entropy as a preprocessor. 2009:651 Analog Integr Circuits Signal Process,2011,66(1):93 [5]Tan Y H,Sun Y C.Yin X.Analog fault diagnosis using S4rans- 9] Zhang Z,Duan Z M,Long Y,et al.A new swarm-SVM-based form preprocessor and a QNN classifier.Measurement,2013,46 faul diagnosis approach for switched current cireuit by using kur- (7):2174 tosis and entropy as a preprocessor.Analog Integr Circuits Signal 6]Zhang A H,Chen C,Jiang B S.Analog circuit fault diagnosis Process,2014,81(1):289工程科学学报，第 39 卷，第 7 期 征偏离度作为评价选择最优小波包基，构建了最优故 障特征向量，在软故障类别数量相同的条件下，诊断率 达到了 100% ，成功诊断区分了文献中未能区分的 4 种故障类别． 在文献［19］的故障诊断方法中，以信息 熵和峭度两个特征参数作为分类器的输入，而本文只 提取了归一化能量作为唯一一个特征参数，却达到了 更好的分类效果． 5 结语 ( 1) 本文提出了一种基于小波包分解和 GA--BP 的模型用于开关电流电路的软故障诊断检测． 以开关 电流电路中的六阶切比雪夫低通滤波器作为测试电 路，通过 GA 算法优化 BP 神经网络的权值和阈值． 诊 断结果表明，提出的诊断模型对开关电流电路具有良 好的故障分类性能． ( 2) 将本文方法与其他文献方法作了比较分析． 这对于进一步完善开关电流电路故障诊断模型理论体 系具有参考意义． 参 考 文 献 ［1］ Toumazou C，Hughes J B，Battersby N C． Switched-Currents，an Analogue Technique for Digital Technology． London: Peter Pereg￾rinus Ltd，1993 ［2］ Long Y． The Study of Switched Current Circuit Test and Fault Di￾agnosis Methods ［Dissertation］． Changsha: Hunan University， 2012 ( 龙英． 开关电流电路测试与故障诊断方法研究［学位论文］． 长沙: 湖南大学，2012) ［3］ Spence H． Automatic analog fault simulation / / Proceedings of Au￾to Test Conference Test Technology and Commercialization． Day￾ton，1996 ［4］ Jantos P，Grzechca D，Ｒutkowski J． A global parametric faults di￾agnosis with the use of artificial neural networks / / ECCTD 2009 European Conference on Circuit Theory and Design． Antalya， 2009: 651 ［5］ Tan Y H，Sun Y C，Yin X． Analog fault diagnosis using S-trans￾form preprocessor and a QNN classifier． Measurement，2013，46 ( 7) : 2174 ［6］ Zhang A H，Chen C，Jiang B S． Analog circuit fault diagnosis based UCISVM． Neurocomputing，2016，173: 1752 ［7］ Aminian M，Aminian F． Neural-network based analog-circuit fault diagnosis using wavelet transform as preprocessor． IEEE Trans Circuits Syst II: Analog Digital Signal Process，2000，47 ( 2 ) : 151 ［8］ Aminian F，Aminian M，Collins H W． Analog fault diagnosis of actual circuits using neural networks． IEEE Trans Instrum Meas， 2002，51( 3) : 544 ［9］ Peng M F，Tse C K，Shen M E，et al． Fault diagnosis of analog circuits using systematic tests based on data fusion． Circuits Syst Signal Process，2013，32( 2) : 525 ［10］ Sheikhan M，Sha'bani A A． PSO-optimized modular neural net￾work trained by OWO--HWO algorithm for fault location in analog circuits． Neural Comput Appl，2013，23( 2) : 519 ［11］ Zhao D S，He Y Z． A new test points selection method for analog fault dictionary techniques． Analog Integr Circuits Signal Process，2015，82( 2) : 435 ［12］ Jiang Y Y，Wang Y Ｒ，Luo H． Fault diagnosis of analog circuit based on a second map SVDD． Analog Integr Circuits Signal Process，2015，85( 3) : 395 ［13］ Han H，Wang H J，Tian S L，et al． A new analog circuit fault diagnosis method based on improved Mahalanobis distance． J Electron Testing，2013，29( 1) : 95 ［14］ Guo J Ｒ，He Y G，Tang S X，et al． Switched-current circuits test using pseudo-random method． Analog Integr Circuits Signal Process，2007，52( 1) : 47 ［15］ Guo J Ｒ，Cai X H，He Y G． PＲBS test signature analysis of switched current circuit / / 2009 1st International Conference on Information Science and Engineering ( ICISE) ． Nanjing，2009: 627 ［16］ Long Y，He Y G，Liu L，et al． Implicit functional testing of switched current filter based on fault signatures． Analog Integr Circuits Signal Process，2012，71( 2) : 293 ［17］ Guo J Ｒ，He Y G，Liu M Ｒ． Wavelet neural network approach for testing of switched-current circuits． J Electr Test，2011，27: 611 ［18］ Long Y，He Y G，Yuan L F． Fault dictionary based switched current circuit fault diagnosis using entropy as a preprocessor． Analog Integr Circuits Signal Process，2011，66( 1) : 93 ［19］ Zhang Z，Duan Z M，Long Y，et al． A new swarm-SVM-based fault diagnosis approach for switched current circuit by using kur￾tosis and entropy as a preprocessor． Analog Integr Circuits Signal Process，2014，81( 1) : 289 · 6011 ·