摘要
根据水轮机组振动故障与振动征兆之间复杂的非线性关系,总结了适用于水轮发电机组振动故障诊断的频谱特征表和振动部位幅值特征表;针对C-均值聚类易产生误分类问题,提出基于动态核聚类分析的水轮机组故障模式分类方法;对振动信号频率、振动信号幅值特征、振动部位进行分析,获得振动频谱征兆隶属度值,在此基础上,建立了基于故障分层的水轮机组运行状态自动诊断模型.工程应用实例表明:该模型的诊断效率是可行的,诊断结果具有较高的可信度.
Based on the complicated relationships between the symptoms and the defects of hydro - generator unit, the frequency spectrum properties and amplitude properties applied to fault diagnosis for hydrogenerator unit were presented. In terms of false classification of Fuzzy C-Means, a dynamic kernel based clustering method was presented to distinguish fault pattern. Membership grade of symptoms of vibration frequency spectrum was acquired by analyzing frequency of vibration signals, amplitude properties and vibration position. A model of automatic diagnosis to hydroelectric power based on faults classification was built. Examples of project application indicated： Diagnosis efficiency of the model is very feasible and diagnosis result possesses superior reliability.
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2005年第9期47-49,52,共4页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
江西省科技厅攻关项目(200210200211)
江西省教育厅科技项目(赣教计字[2001]387).
关键词
水轮机组
故障诊断
模式识别
核
聚类分析
hydro-generator unit
fault diagnosis
pattern classification
kernel
cluster analysis
