基于BP-ART混合神经网络的电路故障诊断新方法

Journal of Systems Engineering and Electronics ›› 2010, Vol. 32 ›› Issue (4): 873-876.

基于BP-ART混合神经网络的电路故障诊断新方法

王安娜，刘坐乾，杨铭如，曲延华

（东北大学信息科学与工程学院，辽宁沈阳 110004）

出版日期:2010-04-23 发布日期:2010-01-03

Novel method for circuit fault diagnosis based on the BP-ART hybrid neural network

WANG An-na，LIU Zuo-qian，YANG Ming-ru，QU Yan-hua

(School of Information Science and Engineering, Northeastern Univ., Shenyang 110004, China)

Online:2010-04-23 Published:2010-01-03

摘要/Abstract

摘要：

建立了基于误差反向传播(back propagation， BP)神经网络和自适应共振理论(adaptive resonate theory， ART)神经网络的电路故障诊断模型，提出了BP神经网络和ART神经网络相结合的电路故障诊断方法，以ART网络为主，识别新故障，以BP网络为辅，识别多类故障，并对传统的ART神经网络竞争机制加以改进，有效地解决了复杂电路故障诊断的难题。实验表明，基于BP和改进ART神经网络相结合的电路故障诊断方法具有自适应性好、训练时间短、准确性高等特点。

Abstract:

A circuit fault diagnosis model based on the back propagation (BP) neural network and adaptive resonance theory neural network is established,and then circuit fault diagnosis method of the BP and ART hybrid neural network is brought forwold. Adaptive resonate theory (ART) network is used to identify new fault, and the BP network is used to identify multi-class faults, the competition method of ART neural network is improved. The method solve the problem of complex circuit fault diagnosis effectively. The experiments show that circuit fault diagnosis method based on BP and improved ART hybrid neural network has characteristics of good adaptability, short training time and high accuracy.

王安娜，刘坐乾，杨铭如，曲延华. 基于BP-ART混合神经网络的电路故障诊断新方法[J]. Journal of Systems Engineering and Electronics, 2010, 32(4): 873-876.

WANG An-na，LIU Zuo-qian，YANG Ming-ru，QU Yan-hua. Novel method for circuit fault diagnosis based on the BP-ART hybrid neural network[J]. Journal of Systems Engineering and Electronics, 2010, 32(4): 873-876.