赵海洋;王金东;刘树林;陈桂娟;
ZHAO Hai-yang,WANG Jin-dong,LIU Shu-lin,CHEN Gui-juan(Daqing Petroleum Institute,Daqing 163318,China)

• 1:大庆石油学院
• 2:大庆石油学院
• 3:大庆石油学院
• 4:大庆石油学院 黑龙江大庆163318

摘要(Abstract)：

结合神经网络和支持向量机的优点,针对实际应用的不同阶段,提出了一套基于神经网络和支持向量机的复合故障诊断技术。该技术不但可以融入新的故障信息,而且可以使故障诊断模型始终处于最优识别状态,并以往复压缩机气缸系统常见故障的诊断为实例,验证了该技术的有效性。
According to different advantages of Artificial Neural Network and Support Vector Machine,a compound fault diagnosis technique based on Artificial Neural Network and Support Vector Machine is introduced for different stages of application. This technique can not only amalgamate new fault information,but also keep the fault diagnosis model on the optimized identification state.Take the fault diagnosis of reciprocating compressor cylinder system as an example,effectiveness of the technique is validated.

关键词(KeyWords)： 神经网络;支持向量机;故障诊断;往复压缩机
artificial neural network;support vector machine;fault diagnosis;reciprocating compressor

Abstract:

Keywords:

基金项目(Foundation):

作者(Authors): 赵海洋;王金东;刘树林;陈桂娟;
ZHAO Hai-yang,WANG Jin-dong,LIU Shu-lin,CHEN Gui-juan(Daqing Petroleum Institute,Daqing 163318,China)

参考文献(References)：

• [1]Cortes C,Vapnik V.Support vector networks[J].Machine Learning,1995,(20):273-397.
• [2]Vapnik V N,张学工(译).统计学习理论的本质[M].北京:清华大学出版社:2000.102-105.
• [3]丛爽.面向MAYLAB工具箱的神经网络理论与应用[M].合肥:中国科学技术大学出版社,1998.46-48.
• [4]王金东,赵海洋,宋义,等.2D12往复压缩机故障诊断系统及其应用[J].振动工程学报(增刊),2004,17(S):564-566.
• [5]王金东,张嘉钟,杨敬源.2D12往复压缩机气缸内压力检测技术[J].压缩机技术,2002,(5):21-23.
• [6]姜谙男,冯夏庭.由工程实例获取隧洞围岩最大变形的支持向量机方法[J].东北大学学报(自然科学版),2004,(25):793-795.