基于格拉姆角场和PCNN-BiGRU模型的故障诊断方法及其应用  

Fault Diagnosis Method and its Application Based on Gramian Angular Fieldand PCNN-Bigru Model

作　　者：盛世龙 王淑青[1] 王云鹤 翟宇胜 刘冬 SHENG Shi-long;WANG Shu-qin;WANG Yun-he;ZHAI Yu-sheng;LIU Dong(School of Electrical and Engineering,Hubei University of Technology,Wuhan 430068,Hubei Province,China;School of Power and Mechanical Engineering,Wuhan University,Wuhan 430062,Hubei Province,China;College of Energy and Power,North China University of Water Resources and Electric Power,Zhengzhou 450045,Henan Province,China)

机构地区：[1]湖北工业大学电气与电子工程学院,湖北武汉430068 [2]武汉大学动力与机械学院,湖北武汉430072 [3]华北水利水电大学能源与动力工程学院,河南郑州450045

出　　处：《中国农村水利水电》2025年第2期121-128,共8页China Rural Water and Hydropower

基　　金：国家自然科学基金项目(52309111)。

摘　　要：研究提出了一种基于信号处理和深度学习技术的水电机组故障诊断方法。首先,利用VMD对水电机组的原始信号进行分解和重构,以实现信号的降噪,并得到本征模态函数(IMF);随后,通过格拉姆角场(GAF)变换,将IMF转换为GASF和GADF图像。然后将所有图像数据输入到双通道并行二维卷积神经网络与双向门控循环单元(PCNNBiGRU)模型中进行训练。该模型通过CNN提取特征图,并将其输入到BiGRU中,以保持对时间特征的敏感度并剔除冗余信息;最后,为验证该方法的有效性,结合实际电站机组样本数据进行比较试验,对所提方法提供高效、准确的水电机组故障诊断解决方案进行了验证。This study proposes a fault diagnosis method for hydroelectric units based on signal processing and deep learning technology.Firstly,the original signals of the hydroelectric units are decomposed and reconstructed using Variational Mode Decomposition(VMD)to achieve signal denoising and obtain the Intrinsic Mode Functions(IMFs).Subsequently,the IMFs are transformed into Gramian Angular Field(GAF)and Gramian Angular Difference Field(GADF)images through Gramian Angular Field transformation.Then,all the image data is input into a Parallel Dual-Channel Two-Dimensional Convolutional Neural Network with Bidirectional Gated Recurrent Unit(PCNN-BiGRU)model for training.This model uses CNN to extract feature maps,which are then input into BiGRU to maintain sensitivity to temporal features and eliminate redundant information.Finally,to validate the effectiveness of this method,comparative experiments are conducted using actual samples from power plant units,confirming that the proposed method provides an efficient and accurate solution for fault diagnosis of hydroelectric units.

关 键 词：水电机组 VMD 格拉姆角场 故障诊断 并行CNN BiGRU 

分 类 号：TV734.1[水利工程—水利水电工程] TK05[动力工程及工程热物理]