基于改进卷积神经网络的柴油机故障诊断方法研究

摘要
图/表
参考文献(16)
相关文章 (15)

全文: PDF (1977 KB) HTML (1 KB)
输出: BibTeX | EndNote (RIS)

摘要针对现有基于卷积神经网络(Convolutional Neural Network，CNN)的柴油机故障诊断方法易过拟合、收敛速度、处理小样本数据时诊断精度低的问题，提出了一种基于改进CNN的“端到端”柴油机故障诊断方法。该方法在CNN架构上，采用指数线性单元（Exponential Linear Units，ELU）作为激活函数及小批量训练方法加速模型收敛，用全局平均池化(global average pooling，GAP)代替全连接层以降低过拟合风险。基于台架试验的诊断结果表明，所提方法进行柴油机典型故障诊断的精度达到99.18%；与未改进模型及现有基于CNN的柴油机故障诊断算法相比，该方法在处理小样本数据集时仍保持最高识别精度。

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	张俊红1
	2
	孙诗跃1
	朱小龙1
	周启迪1
	戴胡伟1
	林杰威1

关键词 ：柴油机, 故障诊断, 深度学习, 卷积神经网络

Abstract：Aiming at the problems of slow model convergence and low diagnosis accuracy when processing small sample data based on the existing Convolutional Neural Network (CNN) diesel engine fault diagnosis method, a diesel engine fault diagnosis method based on improved CNN is proposed. In the convolutional neural network architecture, Exponential Linear Units (ELU) are used as the activation function and the small batch training method accelerates the model convergence, and the global average pooling replaces the fully connected layer to reduce the risk of overfitting. Experimental data analysis shows that the accuracy of the method proposed for diesel engine typical fault diagnosis reaches 99.18%; compared with the unimproved model and the existing CNN-based diesel engine fault diagnosis algorithm, this method still maintains the highest accuracy when dealing with small sample data sets.

Key words： diesel engine fault diagnosis deep learning CNN

收稿日期: 2020-11-09 出版日期: 2022-03-28

引用本文:

张俊红1,2，孙诗跃1，朱小龙1，周启迪1，戴胡伟1，林杰威1. 基于改进卷积神经网络的柴油机故障诊断方法研究[J]. 振动与冲击, 2022, 41(6): 139-146.
ZHANG Junhong1,2，SUN Shiyue1，ZHU Xiaolong1，ZHOU Qidi1，DAI Huwei1，LIN Jiewei1. Diesel engine fault diagnosis based on an improved convolutional neural network. JOURNAL OF VIBRATION AND SHOCK, 2022, 41(6): 139-146.

链接本文:

http://jvs.sjtu.edu.cn/CN/ 或 http://jvs.sjtu.edu.cn/CN/Y2022/V41/I6/139

[1] ZHAO H, ZHANG J, JIANG Z, et al. A new fault diagnosis method for a diesel engine based on an optimized vibration mel frequency under multiple operation conditions[J]. Sensors (Switzerland), 2019,19(11).
[2] 王凤利, 邢辉, 邱赤东, 等. 基于改进自适应EEMD的柴油机气缸磨损诊断[J]. 内燃机学报, 2017,35(01): 89-95.
    Wang Fengli, Xing Hui, Qiu Chidong, et al. Diesel engine cylinder wear diagnosis based on improved adaptive EEMD[J]. Journal of Internal Combustion Engine, 2017,35(01): 89-95.
[3] 孔运,王天杨,褚福磊.自适应TQWT滤波器算法及其在冲击特征提取中的应用[J].振动与冲击,2019,38(11):9-16+23.    Kong Yun,Wang Tianyang,Chu Fulei.Adaptive TQWT filter algorithm and its application in shock feature extraction[J].Vibration and Shock,2019,38(11):9-16+23.
[4] XU X, ZHAO Z, XU X, et al. Machine learning-based wear fault diagnosis for marine diesel engine by fusing multiple data-driven models[J]. KNOWLEDGE-BASED SYSTEMS, 2020,190(105324).
[5] SHAO H, JIANG H, WANG F, et al. An enhancement deep feature fusion method for rotating machinery fault diagnosis[J]. Knowledge-Based Systems, 2017,119.
[6] 张永祥, 王宇, 姚晓山. 基于加窗和卷积神经网络的柴油机拉缸故障诊断[J]. 车用发动机, 2019(06): 84-89.
    Zhang Yongxiang, Wang Yu, Yao Xiaoshan. Diesel engine cylinder pulling fault diagnosis based on windowing and convolutional neural network[J]. Automotive Engine, 2019(06): 84-89.
[7] JIANG Z, LAI Y, ZHANG J, et al. Multi-Factor Operating Condition Recognition Using 1D Convolutional Long Short-Term Network[J]. Sensors (Basel, Switzerland), 2019,19(24): 5488.
[8] 张康, 陶建峰, 覃程锦, 等. 随机丢弃和批标准化的深度卷积神经网络柴油机失火故障诊断[J]. 西安交通大学学报, 2019,53(08): 159-166.
    Zhang Kang, Tao Jianfeng, Qin Chengjin, et al. Diesel engine misfire fault diagnosis based on random discard and batch standardization with deep convolutional neural network[J]. Journal of Xi'an Jiaotong University, 2019,53(08): 159-166.
[9] 周飞燕, 金林鹏, 董军. 卷积神经网络研究综述[J]. 计算机学报, 2017,40(06): 1229-1251.
    Zhou Feiyan, Jin Linpeng, Dong Jun. Summary of Convolutional Neural Network Research[J]. Chinese Journal of Computers, 2017,40(06): 1229-1251.
[10] 唐智川, 张克俊, 李超, 等. 基于深度卷积神经网络的运动想象分类及其在脑控外骨骼中的应用[J]. 计算机学报, 2017,40(06): 1367-1378.
    Tang Zhichuan, Zhang Kejun, Li Chao, et al. Classification of motor imagination based on deep convolutional neural network and its application in brain-controlled exoskeleton[J]. Chinese Journal of Computers, 2017,40(06): 1367-1378.
[11] ZHANG W, PENG G, LI C, et al. A New Deep Learning Model for Fault Diagnosis with Good Anti-Noise and Domain Adaptation Ability on Raw Vibration Signals[J]. SENSORS, 2017,17(4252).
[12] KESKAR N S, MUDIGERE D, NOCEDAL J, et al. On Large-Batch Training for Deep Learning: Generalization Gap and Sharp Minima[J]. arXiv, 2016: 16.
[13] SCHIRRMEISTER R T, SPRINGENBERG J T, FIEDERER L D J, et al. Deep Learning With Convolutional Neural Networks for EEG Decoding and Visualization[J]. HUMAN BRAIN MAPPING, 2017,38(11): 5391-5420.
[14] HSIAO T, CHANG Y, CHOU H, et al. Filter-based deep-compression with global average pooling for convolutional networks[J]. JOURNAL OF SYSTEMS ARCHITECTURE, 2019,95: 9-18.
[15] 刘昱. 基于振动分析的柴油机燃油系统与配气机构故障诊断研究[D]. 天津大学, 2016.
    Liu Yu. Fault diagnosis of diesel engine fuel system and valve train based on vibration analysis[D]. Tianjin University, 2016.
[16] YAGUO L, ZHENGJIA H, YANYANG Z, et al. Fault diagnosis of rotating machinery based on multiple ANFIS combination with GAs[J]. Mechanical Systems and Signal Processing, 2007,21(5): 2280-2294