基于AE-ANFIS的船舶柴油机故障诊断  

Fault Diagnosis for Marine Diesel Engines Base on AE-ANFIS

作　　者：姜苗 向阳[1,2] JIANG Miao;XIANG Yang(School of Naval Architecture,Ocean and Energy Power Engineering,Wuhan University of Technology,Wuhan 430063,China;Key Laboratory of High Performance Ship Technology,Ministry of Education,Wuhan University of Technology,Wuhan 430063,China)

机构地区：[1]武汉理工大学船海与能源动力工程学院,武汉430063 [2]武汉理工大学高性能船舶技术教育部重点实验室,武汉430063

出　　处：《噪声与振动控制》2023年第5期188-195,238,共9页Noise and Vibration Control

基　　金：工业和信息化部高技术船舶资助项目(MC-201917-C09);工业和信息化部绿色智能内河船舶创新资助项目(20201g0079)。

摘　　要：柴油机作为船舶主要动力设备,在船舶行业应用极其广泛,但其工作环境恶劣,极易发生故障。为减小船舶航行时柴油机故障带来的经济损失,有必要对其进行故障诊断。通过柴油机实验台架模拟不同类型故障,并在柴油机缸盖处使用振动加速度传感器采集故障信号,选取在1缸缸盖处采集的信号作为样本数据进行数据分析。由于采集的原始信号是多激励源合成信号,其中包含传播噪声、环境噪声,为降低噪声对识别精度影响,首先使用变分模态分解(VariationalModeDecomposition,VMD)对信号进行分解降噪,把原始一维数据分解成能反映柴油机运行状态的多维数据;接着使用自编码器(Auto-Encode,AE)对分离信号进行特征提取,以降低分解信号间的干扰,提高识别准确率;再使用自适应模糊神经网络(Adaptive Fuzzy Neural Network,ANFIS)建立故障诊断模型,并将所提取特征作为诊断模型输入;最后根据诊断模型的识别准确度评价以上方法的可行性。The diesel engine,as the key power equipment of a ship,is frequently used in the marine industry,but it is always working in an adverse environment and is susceptible to failure.It is necessary to carry out fault diagnosis in order to reduce the economic losses caused by diesel engine failures during ship navigation.In this paper,different types of faults are simulated by using the diesel engine experimental setup,and the fault signals are collected at the cylinder head of the diesel engine using vibration acceleration sensors.Since the original signal collected is a composite signal from multiple excitation sources,which contains transmission noise and environmental noise,Variational Mode Decomposition(VMD)is used to decompose the signal and reduce the influence of noise on the recognition accuracy.And the original one-dimensional data is decomposed into multi-dimensional data which can reflect the operating status of the diesel engine.Then,the Auto-Encoder(AE)is used to extract features from the separated signals to reduce the interference among the decomposed signals and improve the recognition accuracy.Finally,the Adaptive Fuzzy Neural Network(ANFIS)is used to build a fault diagnosis model,and the extracted features are used to input to the diagnosis model.The feasibility of the above method is evaluated according to the recognition accuracy of the diagnostic model.

关 键 词：故障诊断 柴油机 变分模态分解 自编码器 自适应模糊神经网络 

分 类 号：TP206.3[自动化与计算机技术—检测技术与自动化装置]