摘要
首先分析了神经网络正向推理在故障诊断领域中存在的缺陷，进而引入了神经网络混合双向推理算法，并介绍了其在故障诊断中的应用，仿真结果表明，利用混合双向推理可以在保证诊断精度的前提下，减少征兆的输入数目。
The shortcoming of the forward inference based on neural networks used in the field of fault diagnosis is discussed in the paper, then the algorithm of bi-directional inference based on neural networks and its use in fault diagnosis are introduced. The emulation results showed: The number of symptoms inputted could be reduced without sacrificing the accuracy of fault diagnosis.
出处
《五邑大学学报（自然科学版）》
CAS
1998年第2期1-4,共4页
Journal of Wuyi University(Natural Science Edition)
基金
广东省自然科学基金
关键词
故障诊断
神经网络
双向推理
Fault Diagnosis
Neural Networks
Bi-directional inference
Forward Infernce
