基于CapsNet模型的过程故障识别研究  

Research on Process Fault Identification Based on CapsNet Model

作　　者：衷路生 卢文涛 Zhong Lusheng;Lu Wentao(School of Electrical and Automation Engineering,East China Jongtong University,Nanchang 330013,China)

机构地区：[1]华东交通大学电气与自动化工程学院,江西南昌330013

出　　处：《华东交通大学学报》2020年第4期33-40,共8页Journal of East China Jiaotong University

基　　金：国家自然科学基金项目(61863012,61263010);江西省科技厅项目(20181BBE50020,20161BBE50082,20161BAB202067)。

摘　　要：为了实现过程故障的识别诊断,文章使用CapsNet模型训练数据。首先,运用网络模型的空间特性,以向量的形式对训练数据进行特征表示、归一化处理。然后,进行卷积操作,在动态一致路由更新上进行故障分类。最后,增加重构模块来对输入数据矩阵反馈修正,降低损失误差,使网络快速收敛。同时,在每一层网络进行特征可视化,能清楚看到每一层网络特征图的变化。实验结果表明,文章模型的过程故障识别性能优于其他神经网络模型。In order to realize the recognition and diagnosis of process faults,this paper uses CapsNet model to train data.Firstly,using the spatial characteristics of the network model,the training data was characterized and normalized in the form of vectors.Then,a convolution operation was performed to classify the faults on the dynamic consistent routing update.Finally,the reconstitution module was added to modify the input data matrix,reduce the loss error and make the network converge quickly.At the same time,feature visualization wasper formed on each layer of the network,and the changes in the feature map of each layer were clearly seen.The experimental results show that the process fault recognition performance of this model is better than other neural network models.

关 键 词：CapsNet 动态路由更新 故障分类 CNN 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]