磨煤机组是火力发电厂的重要设备之一,为了保证生产过程的安全性和稳定性,提出一种基于聚类粗化图卷积神经网络(CC-GCN)的故障诊断方法。首先通过KNN算法在原始故障样本之间建立图结构并转换成图样本,然后利用谱聚类将图样本压缩成多级粗化图,并分别对每一级别的粗化图进行卷积操作以及特征的融合,最后基于图分类方法对故障样本进行故障诊断。在磨煤机组的两组不同运行状态的数据集上进行仿真实验,结果表明该方法不仅能有效提高故障诊断的精度,还能显著降低模型的运行时间。
The coal mill group is one of the important equipment in thermal power plants. To ensure the safety and stability of the production process, a fault diagnosis method named clustering coarsening graph convolution neural network (CC-GCN) is proposed in this paper. Firstly, the graph structure is established between the original fault samples by KNN algorithm and converted into graph samples. Then, spectral clustering is used to compress the graph samples into multi-level coarsening graphs, and convolution operations and feature fusion are performed for each level of coarsening graph respectively. Finally, fault diagnosis is performed on the fault samples based on the graph classification method. Simulation experiments are carried out on two sets of data sets with different operation conditions of the coal mill group, and the results show that this method can not only effectively improve the accuracy of fault diagnosis, but also significantly reduce the running time of the model.
[1] 郭悦. 基于迁移学习的磨煤机状态监测与故障诊断研究[D]. 北京:华北电力大学,2021.
GUO Y. Study on condition monitoring and fault diagnosis of coal mill based on transfer learning[D]. Beijing: North China Electric Power University, 2021.
[2] 袁杰,王福利,王姝,等.基于D-S融合的混合专家知识系统故障诊断方法[J].自动化学报,2017,43(9):1580-1587.
YUAN J, WANG F L, WANG S, et al. A fault diagnosis approach by D-S fusion theory and hybrid expert knowledge system[J]. Acta Automatica Sinica, 2017, 43(9):1580-1587.
[3] 池阔,康建设,张星辉,等.基于匹配稳态随机共振的轴承故障诊断方法[J].复杂系统与复杂性科学,2019,16(2):85-94.
CHI K, KANG J S, ZHANG X H, et al. Bearing fault diagnosis based matched-stable stochastic resonance[J]. Complex Systems and Complexity Science, 2019, 16(2):85-94.
[4] 尹怀仙,王凯,张铁柱,等.基于PSO-BP神经网络的城轨列车转向架轮对轴箱故障预测[J].复杂系统与复杂性科学,2015,12(4):97-103.
YIN H X, WANG K, ZHANG T Z, et al. Fault prediction based on PSO-BP neural network about wheel and axle box of bogie in urban rail train[J]. Complex Systems and Complexity Science, 2015, 12(4):97-103.
[5] LI J T, YUAN Z, MA B Y. Research on fault diagnosis model of coal mills based on FPGA[C]//2021 China Automation Congress.Kunming,China: IEEE, 2021: 2704-2709.
[6] HU Y, PING B Y, ZENG D, et al. Research on fault diagnosis of coal mill system based on the simulated typical fault samples[J]. Measurement, 2020, 161: 107864.
[7] GAO Y, LIU X Y, XIANG J. FEM simulation-based generative adversarial networks to detect bearing faults[J]. IEEE Transactions on Industrial Informatics, 2020, 16(7): 4961-4971.
[8] LI T F, ZHAO Z B, SUN C, et al. Multireceptive field graph convolutional networks for machine fault diagnosis[J]. IEEE Transactions on Industrial Electronics, 2020, 68(12): 12739-12749.
[9] VON LUXBURG U. A tutorial on spectral clustering[J]. Statistics and Computing, 2007, 17(4): 395-416.
[10] SHI J B, MALIK J. Normalized cuts and image segmentation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2000, 22(8): 888-905.
[11] HY T S, NGUYEN V B, TRAN-THANH L, et al. Temporal multiresolution graph neural networks for epidemic prediction[DB/OL]. [2022-08-20]. https://arxiv.org/pdf/2205.14831v1.pdf.
[12] WU F, SOUZA A, ZHANG T Y, et al. Simplifying graph convolutional networks[DB/OL]. [2022-08-20]. https://arxiv.org/pdf/1902.07153.pdf.
[13] DEFFERRARD M, BRESSON X, VANDERGHEYNST P. Convolutional neural networks on graphs with fast localized spectral filtering[DB/OL]. [2022-08-20]. https://arxiv.org/pdf/1606.09375.pdf.
[14] MORRIS C, RITZERT M, FEY M, et al. Weisfeiler and leman go neural: higher-order graph neural networks[DB/OL]. [2022-08-20]. https://arxiv.org/pdf/1810.02244.pdf.
[15] HAMILTON W, YING Z, LESKOVEC J. Inductive representation learning on large graphs[DB/OL]. [2022-08-20]. https://arxiv.org/pdf/1706.02216.pdf.
