DocumentCode :
1982873
Title :
Simulation of improved BP algorithm in the fault diagnosis of analog circuit
Author :
Xue, Zhi-qiang ; Li, Yi ; Cao, Yan
Author_Institution :
Dept. 4th, New Star Res. Inst. of Appl. Tech, Hefei, China
fYear :
2011
fDate :
16-18 Sept. 2011
Firstpage :
3148
Lastpage :
3150
Abstract :
In view of the complexity of analog circuits, to address the drawbacks of the traditional BP algorithm, such as slow speed of constringency and proneness to lapse into local minima, an improved approach based on affixation momentum & factor adjust and the learning rate is proposed and applied to its fault diagnosis. The results show that the model improves the efficiency of fault diagnosis compared with the traditional method. After the data concerning faults is entered into the trained neural network, the types of the faults can be judged based upon the output, and then the accurate localization of the faulty module can be realized. Thus it can be concluded that the improved BP algorithm is of practical value.
Keywords :
analogue integrated circuits; backpropagation; circuit complexity; electronic engineering computing; fault diagnosis; neural nets; BP algorithm; analog circuit; complexity; fault diagnosis; faulty module; learning rate; neural network; Analog circuits; Artificial neural networks; Biological neural networks; Capacitance; Circuit faults; Fault diagnosis; Training; analog circuit; fault diagnosis; improved BP Algorithm; neural network;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Electrical and Control Engineering (ICECE), 2011 International Conference on
Conference_Location :
Yichang
Print_ISBN :
978-1-4244-8162-0
Type :
conf
DOI :
10.1109/ICECENG.2011.6057502
Filename :
6057502
Link To Document :
بازگشت