小波-Fisher的地下建筑通风设备故障诊断  

Fault diagnosis for underground construction ventilator based on wavelet packet-Fisher

作　　者：胡凤涛 刘海卿[1] 赵秀杰[2] 

机构地区：[1]辽宁工程技术大学建筑工程学院,辽宁阜新123000 [2]安徽工程大学人文学院,安徽合肥230000

出　　处：《辽宁工程技术大学学报（自然科学版）》2014年第10期1338-1341,共4页Journal of Liaoning Technical University (Natural Science)

基　　金：辽宁省自然科学基金资助项目(20102091)

摘　　要：将小波包分析与Fisher判别分析法相结合应用于地下建筑通风设备故障诊断问题中,运用小波包频道能量分解技术提取了不同频带反映地下建筑通风机不同工作状态的特征向量,以此作为Fisher判别分析的故障样本.通过分析计算建立了建筑通风机故障诊断的Fisher判别函数,并利用回代估计方法进行检验.结果表明:该模型准确有效,判别能力较稳定,不需要优化网络结构,为地下建筑通风机故障诊断提供一种新的预测方法.In order to diagnosis the fault of building ventilator, this paper, in combination with the data of typical vibration signal, establishes a wavelet packet-Fisher discriminant analysis model. The state features of typical vibration signal are extracted by wavelet packet technology, and those features are selected as discriminant factor of Fisher discriminant analysis. A wavelet packet-Fisher discriminant analysis model is established through training a large set of learning samples from a series of building ventilator fault projects. The results show that this method has excellent discriminant potency and a low cross-validated misjudgment ratio, no need to optimize structure and provides a new prediction method for the fault diagnosis of building ventilator.

关 键 词：地下建筑 通风设备 小波包 FISHER判别分析 故障诊断 

分 类 号：TU28[建筑科学—建筑设计及理论]