摘要
分析了大机组组合故障难以正确识别的原因和常见组合故障识别方法的不足 ,提出了基于识别后验熵的组合故障优化判别规则 ,改进了概率神经网络的分类能力。最后 ,对模拟组合样本和大机组高频组合故障进行识别 。
In this paper, we first discussed the reason why it is difficult to detect correctly the meshed faults of rotor, and the shortcomings of the existing methods for detecting meshed faults. Then,an optimal judging rule for the meshed fault based on the identifying of post priori entropy is proposed, the classifying ability of probabilistic neural networks is improved. Finally, this method is used in simulation meshed fault samples and practical high frequency meshed faults of rotor, and good results are obtained.
出处
《机械科学与技术》
CSCD
北大核心
2000年第1期85-87,共3页
Mechanical Science and Technology for Aerospace Engineering
