摘要
首先论述各种状态信息和设备状态之间的对应关系，即模式分类的重要性．然后通过学习，建立故障诊断的神经网络模型，并应用于大型旋转机械的故障诊断．实验研究表明，神经网络能够较好地表达训练样本要求的决策区域，具有较强的分类能力；利用机械振动特征信息进行训练的神经网络对大型旋转机械单个故障有较好的联想能力，其识别效果令人满意，投入现场应用是可行的．
The corresponding relation between all kinds of feature information and equipment condition is dicussed firstly in this paper,that isthe importance of pattern classification. The models of the neural network are built by learning and applied to failure diagnosis of large scale rotating machinery. The experiments show that the neural network can better express the determinative regions demanded by training samples,it has strong capabilities of classifiction; the neural network trained through viblation information has a good associative capability for the single failure of large scale rotating machinery. Its classification effect is satisfactory using it on the spot is feasiable.
出处
《河北工业大学学报》
CAS
1997年第4期10-15,共6页
Journal of Hebei University of Technology
基金
河北省自然科学基金
关键词
神经网络
模式识别
故障诊断
旋转机械
Neural networks, Pattern classification, Failure diagnosis, Large scale rotating machinery, Vibration imformation
