摘要
为提高往复压缩机的故障诊断精度,结合变分模态分解(Variational Mode Decomposition,VMD)、SDP变换和卷积神经网络(Convolution Neural Networ,CNN),提出基于VMD-SDP融合图像和CNN的往复压缩机故障诊断新方法。方法第一步通过VMD将信号自适应分解成6个本征模态函数分量(Intrinsic Mode Functions,IMF),第二步通过SDP变换将6个IMF分量变换成极坐标下的图像,从而得到VMD-SDP融合图像,第三步通过CNN对VMD-SDP融合图像进行识别,得到最终的诊断结果。往复压缩机诊断实例结果表明,所提方法在耗时更少的情况下,得到100%的诊断精度,比其他几种方法更具优势。
In order to improve fault diagnosis accuracy of reciprocating compressors,a new fault diagnosis method based on the combination of variational mode decomposition(VMD),SDP transformation and convolution neural network(CNN)was proposed.Firstly,the signal was decomposed into six intrinsic mode functions(IMF)adaptively by VMD.Then,the six IMF were transformed into polar images by SDP transformation to obtain VMD-SDP fusion images.Finally,the VMD-SDP fusion images were recognized by CNN to obtain the final diagnosis results.The diagnosis result of reciprocating compressor shows that the proposed method can obtain 100%diagnosis accuracy with less computer time consuming,which is more advantageous than other methods.
作者
王海峰
王则林
WANG Haifeng;WANG Zelin(School of Electronic Information Engineering,Nantong Vocational College,Nantong 226007,Jiangsu,China;School of Information Science and Technology,Nantong University,Nantong 226019,Jiangsu,China)
出处
《噪声与振动控制》
CSCD
北大核心
2023年第4期116-121,共6页
Noise and Vibration Control
基金
国家自然科学基金青年资助项目(62004108)。
作者简介
王海峰(1980-),男,江苏省如皋市人,讲师,专业方向为人工智能算法及应用、智能故障诊断技术、软件设计与开发。E-mail:ntwhf0513@163.com。
