【作者】 窦唯； 孙明； 刘树林；

【Author】 DOU Wei, SUN Ming ,LIU Shu-lin(Dept of Petroleum Mechanical Engineering, Daqing Petroleum Institute, Daqing,Heilongjiang 163318; Qiqiha’er Engineering Corporation,Ha’erbin Railway Bureau,Qiqiha’er, Heilongjiang 161000)

【机构】 大庆石油学院机械科学与工程学院； 哈尔滨铁路局齐齐哈尔工程总公司；

【摘要】 针对往复压缩机在线监测时难以提取异常特征的实际情况,提出一种适于往复压缩机的在线状态监测方法。该方法是基于生物免疫系统反面选择机理,并结合人工神经网络进行监测的一种方法。该方法能有效的提取压缩机的异常特征,提高压缩机的在线状态监测的准确率。在线监测往复压缩机示功图的实例结果表明,该方法能准确的监测出往复压缩机主要故障所引起的异常,表明该方法是可行的,并且具有较好的在线性、准确性及鲁棒性。更多还原

【Abstract】 It is rather difficult to extract abnormal characteristics when on-line reciprocating compressors are monitored. According to this practical situation, a new approach which suits for on-line state monitoring of reciprocating compressors is presented.This approach is based on negative selection mechanism of the natural immune system and combined with artificial neural network to be used for monitoring.This approach is effective to extract abnormal characteristics of reciprocating compressors and improve the accuracy of on-line state monitoring of reciprocating compressors. The result by the example of reciprocating compressors dynamometer card of on-line state monitoring shows this approach can exactly detect the degree of abnormality of reciprocating compressors. It shows this approach is feasible and has better on-line quality, accuracy and robustness.更多还原