摘要
采用了模糊神经网络模型,对柴油机缸套磨损故障以及缸套破坏性磨损故障进行了诊断研究.通过缸套磨损故障的模拟实验,获取柴油机机身振动和铁谱分析等多源多维故障信息,并对融合故障信息进行预处理,解决了模糊神经网络输入矢量的模糊特性化、输出矢量的隶属函数及网络的学习训练问题,对缸套不同磨损故障进行了诊断.研究表明,这种基于多信息的诊断方法减小了故障诊断的不确定性,提高了诊断精度.
Fault diagnosis for diesel cylinder liner wear based on the fuzzy neural networkis studied, and the monitoring of cylinder liner under the extreme wear condition is discussed. The multi-source and multi-dimension wear fault information of diesel cylinder liner, which includes the cylinder block surface vibration and the lubricant iron spectrum analyses is acquired by imitated experiment, and then is preprocessed. Thus problems of the fuzzy neural network model, such as the fuzzy input method, the membership function of the output vectors, the choice of the training samples and the learning and training function of the networkare solved, and the different wear fault of cylinder liner are diagnosed. The study based on the multi-information showsthe uncertainty of the fault diagnosis isdecreased moderately,and the accuracy of the fault diagnosis is increased greatly.
出处
《海军工程大学学报》
CAS
北大核心
2005年第1期67-70,75,共5页
Journal of Naval University of Engineering
基金
海军工程大学科研基金资助项目(HGDJJ03012)
关键词
故障诊断
缸套
磨损
模糊神经网络
多信息
fault diagnosis
cylinder liner
wear
fuzzy neural network
multi-information
