您好,欢迎来到中国测试科技资讯平台!

首页> 数字期刊群 >本期导读>小波变换和CNN涡旋压缩机故障诊断

小波变换和CNN涡旋压缩机故障诊断

469    2023-08-15

免费

全文售价

作者:苏莹莹, 毛海旭

作者单位:沈阳大学机械工程学院,辽宁 沈阳 110000


关键词:故障诊断;振动信号;小波变换;卷积神经网络


摘要:

针对传统单尺度信号分析难以有效解决涡旋压缩机故障诊断中的故障特征信息多尺度耦合问题,提出一种基于小波变换和卷积神经网络的涡旋压缩机故障诊断方法。首先将采集到的振动信号进行连续小波变换生成时频图,并对时频图进行网格化规范处理,将预处理后的时频图作为特征图输入Alexnet卷积神经网络,通过不断调节网络参数,得出最为理想的神经网络模型,以此实现对涡旋压缩机故障类型的辨识诊断。结果表明,该方法针对涡旋压缩机故障类型的识别准确率达到94.6%,与传统多尺度排列熵、信息熵熵距的故障诊断方法相比,该故障识别方法具有更高的准确率。


Fault diagnosis of scroll compressor based on wavelet transform and CNN
SU Yingying, MAO Haixu
School of Mechanical Engineering, Shenyang University, Shenyang 110000, China
Abstract: In order to solve the problem that traditional single-scale signal analysis is difficult to effectively solve the problem of multi-scale coupling of fault feature information in the fault diagnosis of scroll compressors, a fault diagnosis method based on wavelet transform and convolutional neural network(CNN) is proposed. Firstly, the collected vibration signal is analyzed by continuous wavelet transform to generate time-frequency diagram. And the generated time-frequency diagram is gridded and normalized. Then, it has to be inputtd to Alexnet convolutional neural network, and the network parameters are adjusted to obtain the most ideal network model, so as to realize the identification and diagnosis fault types of scroll compressors. The results show that the recognition accuracy reaches 94.6%, with higher accuracy than the traditional methods of multi-scale permutation entropy and distance of the information entropy.
Keywords: fault diagnosis;vibration signal;wavelet transform;convolutional neural networks
2023, 49(4):92-97  收稿日期: 2021-07-10;收到修改稿日期: 2021-10-20
基金项目:
作者简介: 苏莹莹(1983-),女,辽宁沈阳市人,硕士生导师,副教授,博士,主要从事故障诊断的研究
参考文献
[1] 蒋龙陈, 王红军, 张顺利, 等. 基于卷积神经网络的燃气轮机故障诊断研究[J]. 北京信息科技大学学报(自然科学版), 2020, 35(5): 25-29
[2] 刘涛, 邵华. 基于振动信号的变频涡旋压缩机故障诊断[J]. 振动. 测试与诊断, 2012, 32(1): 141-145+168
[3] 鄢仁武, 林穿, 高硕勋, 等. 基于小波时频图和卷积神经网络的断路器故障诊断分析[J]. 振动与冲击, 2020, 39(10): 198-205
[4] 周林春, 陈春俊. 复数卷积神经网络滚动轴承故障诊断研究[J]. 中国测试, 2020, 46(11): 109-115
[5] 刘涛, 马转霞, 杜楠. 多尺度排列熵在涡旋压缩机故障诊断中的应用[J]. 兰州理工大学报, 2018, 44(1): 42-46
[6] ANURAG C, TAUHEED M, SHAHAB F. Convolutional neural network based bearing fault diagnosis of rotating machine using thermal images[J]. Measurement, 2021, 176(4): 109196
[7] 孟强斌. 基于时频图和卷积神经网络的水电机组故障诊断研究[D]. 西安: 西安理工大学, 2020.
[8] MALLA C, RAI A,  KAUL V, et al. Rolling element bearing fault detection based on the complex Morlet wavelet transform and performance evaluation using artificial neural network and support vector machine[J]. Noise & Vibration Worldwide, 2019, 50(9-11): 313-327
[9] ZHANG K, ROBINSON N, LEE S, et al. Adaptive transfer learning for EEG motor imagery classification with deep convolutional neural network[J]. Neural Networks, 2021, 136: 1-10
[10] 杨寿勇, 张海阳, 李成, 等. 基于卷积神经网络模型的微藻种类识别[J]. 环境科学与技术, 2020, 43(S2): 158-164
[11] 赵志坚, 茆志伟, 张进杰, 等. 基于复Morlet变换和改进Alexnet神经网络的柴油机气门间隙异常故障诊断方法[J]. 北京化工大学学报(自然科学版), 2021, 48(4): 64-70
[12] 李泽宇. 变频涡旋压缩机振动信号分析[D]. 兰州: 兰州理工大学, 2012.
[13] 邵华. 基于LabVIEW的变频涡旋压缩机振动测试系[D]. 兰州: 兰州理工大学, 2010.
[14] 胡边, 谭丕成, 叶源, 等. 基于声学特性的灯泡贯流式水轮发电机组噪声信号采集系统设计[J]. 中国测试, 2021, 47(3): 139-143
[15] 刘思洋, 黄坚, 刘桂雄, 等. Alexnet两光照下多类别法定货币识别技术[J]. 中国测试, 2019, 45(9): 118-122
[16] 杜楠. 基于信息熵熵距的涡旋压缩机故障诊断[D]. 兰州: 兰州理工大学, 2016.