摘要
引起水电机组振动的原因很复杂,而且水电机组的振动故障往往是多故障同时发生,使得故障诊断很困难,目前主要是应用基于模式识别的神经网络来进行故障分类,但如何选择故障诊断的神经网络一直是个难点。文章研究了3种人工神经网络,即反向传播网络(BPN)、概率神经网络(PNN)和学习矢量量化网络(LVQ)对水电机组振动故障诊断性能的影响。结果表明,人工神经网络的结构和算法,包括相关训练参数的选择对故障诊断性能有着重要影响。比较而言,学习矢量量化网络和概率神经网络在分类能力方面要比反向传播网络好一些,概率神经网络在计算负载方面比学习矢量量化网络要更胜一筹。
As the complexity of vibrant reasons and the multi-fault occurred simultaneously, the vibration fault diagnosis is very difficulty. The vibration fault diagnosis of hydraulic generating set is distinguished by artificial neural network（ANN） based on patterns recognition at present. But the selecting of neural network is always difficult. Three types of artificial neural networks were investigated including back propagation network （BPN）, probabilistic neural network （PNN） and learning vector quantization network （LVQ） on the performance of a hydroelectric generator set＇s fault diagnosis. The results indicate that the ANN architecture and corresponding algorithm including the choice of relative training parameters play a critical role in determining the performance of hydroelectric generator set＇s fault diagnosis. By comparison, LVQ network and PNN network are better than BPN network in classification ability, and PNN network is better than the others in computation load.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2006年第12期1711-1714,共4页
Chinese Journal of Scientific Instrument
基金
国家自然科学基金(90410019)资助项目
