面向实际化工过程故障诊断的强化深度卷积神经网络模型构建与应用

doi:10.16085/j.issn.1000-6613.2023-1283

摘要/Abstract

摘要：

基于数据驱动的故障诊断技术可以帮助操作人员及时有效发现和检测异常情况，是当前工业与大数据融合的热点领域之一。深度卷积神经网络（deep convolutional neural networks，DCNN）是最常用的基于数据驱动的故障诊断模型，但其激活过程存在正负值计算不匹配以及信息流通效率低导致的参数冗余问题。本文提出一种基于最大平滑单元（maximum smoothing unit，MSF）函数的新激活机制克服传统激活函数的缺点，并且引入注意力机制（attention mechanism）结合门控循环单元（gated recurrent unit，GRU）提升DCNN的信息流通效率克服参数冗余问题，以综合提升传统DCNN模型的故障诊断性能。强化深度卷积神经网络（enhanced deep convolutional neural networks，EDCNN）的现有模型表现出显著提高的故障诊断性能，这在工业致动器控制系统和工业酸性气体吸收过程中的应用得到了验证。两个过程的平均故障诊断率均超过99.0%。

关键词: 故障诊断, 强化深度卷积神经网络, 过程控制, 系统工程, 激活函数

Abstract:

Data-driven fault diagnosis technologies can help operators find and detect process abnormalities in a timely and effective manner, having emerged as one of the hot topics in the current integration of industry and big data. The deep convolutional neural network (DCNN) approach is the most commonly used data-driven fault diagnosis model, but its activation process suffers from the mismatch of positive and negative values and the problem of parameter redundancy resulted by inefficient information flow. In this study, a novel activation mechanism based on the maximum smoothing unit (MSF) function was proposed to overcome the shortcomings of the previous activation functions, and the attention mechanism combined with the gated recurrent unit (GRU) was introduced to overcome the problem of parameter redundancy by improving the efficiency of information flow in DCNN. The as-established model of enhanced deep convolutional neural networks (EDCNN) exhibited significantly improved performance, which was verified by its applications in two industrial processes, the industrial actuator control system and the industrial acid gas absorption process. The average fault diagnosis rate in both processes exceeded 99.0%.

Key words: fault diagnosis, enhanced deep convolutional neural networks, process control, systems engineering, activation function

中图分类号:

TQ021.8

张佳鑫, 张淼, 戴一阳, 董立春. 面向实际化工过程故障诊断的强化深度卷积神经网络模型构建与应用[J]. 化工进展, 2024, 43(9): 4833-4844.

ZHANG Jiaxin, ZHANG Miao, DAI Yiyang, DONG Lichun. Design and application of enhanced deep convolutional neural networks model for fault diagnosis in practical chemical processes[J]. Chemical Industry and Engineering Progress, 2024, 43(9): 4833-4844.

图/表 22

参考文献 19

1	ZHU Zhiqin, LEI Yangbo, QI Guanqiu, et al. A review of the application of deep learning in intelligent fault diagnosis of rotating machinery[J]. Measurement, 2023, 206: 112346.
2	李国友, 张新魁, 才士文, 等. 基于改进KFDA与DE优化SOM的故障诊断模型及其化工过程诊断[J]. 化工进展, 2022, 41(4): 1793-1801.
	LI Guoyou, ZHANG Xinkui, CAI Shiwen, et al. Study on fault diagnosis model and chemical process fault diagnosis based on improved KFDA and DE optimized SOM[J]. Chemical Industry and Engineering Progress, 2022, 41(4): 1793-1801.
3	宋冰, 郑城风, 侍洪波, 等. 基于VAE-OCCA的质量相关故障检测方法研究[J]. 化工学报, 2023, 74(4): 1630-1638.
	SONG Bing, ZHENG Chengfeng, SHI Hongbo, et al. Research on quality-related fault detection method based on VAE-OCCA[J]. CIESC Journal, 2023, 74(4): 1630-1638.
4	姚羽曼, 罗文嘉, 戴一阳. 数据驱动方法在化工过程故障诊断中的研究进展[J]. 化工进展, 2021, 40(4): 1755-1764.
	YAO Yuman, LUO Wenjia, DAI Yiyang. Research progress of data-driven methods in fault diagnosis of chemical process[J]. Chemical Industry and Engineering Progress, 2021, 40(4): 1755-1764.
5	张焕, 张庆, 于纪言. 卷积神经网络中激活函数的性质分析与改进[J]. 计算机仿真, 2022, 39(4): 328-334.
	ZHANG Huan, ZHANG Qing, YU Jiyan. Analysis and improvement of properties of activation functions in convolutional neural networks[J]. Computer Simulation, 2022, 39(4): 328-334.
6	谢星怡, 张正江, 闫正兵, 等. 基于信号特征提取和卷积神经网络的轴承故障诊断研究[J]. 计算机测量与控制, 2023, 31(10): 21-27.
	XIE Xingyi, ZHANG Zhengjiang, YAN Zhengbing, et al. Research on bearing fault diagnosis based on signal feature extraction and convolutional neural network[J]. Computer Measurement & Control, 2023, 31(10): 21-27.
7	WEN Long, LI Xinyu, GAO Liang, et al. A new convolutional neural network-based data-driven fault diagnosis method[J]. IEEE Transactions on Industrial Electronics, 2018, 65(7): 5990-5998.
8	JIANG Guoqian, WANG Jing, WANG Lijin, et al. An interpretable convolutional neural network with multi-wavelet kernel fusion for intelligent fault diagnosis[J]. Journal of Manufacturing Systems, 2023, 70: 18-30.
9	周中, 闫龙宾, 张俊杰, 等. 基于自注意力机制与卷积神经网络的隧道衬砌裂缝智能检测[J/OL]. 铁道学报, 2023, .
	ZHOU Zhong, YAN Longbin, ZHANG Junjie, et al. Intelligent detection of tunnel lining cracks based on self-attention mechanism and convolutional neural network[J/OL]. Journal of the China Railway Society, 2023, .
10	尹梓诺, 马海龙, 胡涛. 基于联合注意力机制和一维卷积神经网络-双向长短期记忆网络模型的流量异常检测方法[J].电子与信息学报, 2023, 45(10): 3719-3728.
	YIN Zinuo, MA Hailong, HU Tao. Traffic anomaly detection method based on joint attention mechanism and 1-dimensional convolutional neural network-bidirectional long short-term memory network model[J]. Journal of Electronics & Information technology, 2023, 45(10): 3719-3728.
11	VINOD Nair, HINTON Geoffrey E. Rectified linear units improve restricted boltzmann machines[C]//Proceedings of the 27th International Conference on International Conference on Machine Learning. Haifa, Israel: ACM, 2010: 807-814.
12	RAMACHANDRAN Prajit, ZOPH Barret, LE Quoc V. Searching for activation functions[EB/OL]. (2017-10-16). .
13	GOODFELLOW Ian J, David WARDE-FARLEY, MIRZA Mehdi, et al. Maxout networks[C]//International Conference on Machine Learning. Atlata, USA: PMLR, 2013: 1319-1327.
14	陈文华, 黄伟稀. 基于卷积神经网络的海上风电机组齿轮箱故障诊断[J]. 电子设计工程, 2022, 30(21): 6-10.
	CHEN Wenhua, HUANG Weixi. Fault diagnosis of offshore wind turbine gearbox based on convolutional neural network[J]. Electronic Design Engineering, 2022, 30(21): 6-10.
15	DENG Lu, ZHANG Yangxue, DAI Yiyang, et al. Integrating feature optimization using a dynamic convolutional neural network for chemical process supervised fault classification[J]. Process Safety and Environmental Protection, 2021, 155: 473-485.
16	HENDRYCKS Dan, GIMPEL Kevin. Gaussian error linear units (GELUs)[EB/OL]. （2016-6-27). .
17	MATTEN L V D, HINTON G. Visualizing data using t-SNE[J]. Journal of Machine Learning Research, 2008, 9 (11): 2579-2605.
18	ZHANG Zhanpeng, ZHAO Jinsong. A deep belief network based fault diagnosis model for complex chemical processes[J]. Computers & Chemical Engineering, 2017, 107: 395-407.
19	YAO Yuman, ZHANG Jiaxin, LUO Wenjia, et al. A hybrid intelligent fault diagnosis strategy for chemical processes based on penalty iterative optimization[J]. Processes, 2021, 9(8): 1266.

编号	符号	描述	范围阈值	单位
1	P57-03	P1-果汁压力（阀门入口）	0~1000	kPa
2	P57-04	P2-果汁压力（阀门出口）	0~1000	kPa
3	T57-03	T-果汁温度（阀门入口）	0~150	℃
4	FC57-03PV	PV-过程值（果汁流量，第2个蒸发器出口）	0~100	m³/h
5	FC57-03CV	CV-控制值（控制器输出）	0~100²	—
6	FC57-03X	X-伺服电机杆位移	0~100²	—

编号	符号	描述	范围阈值	单位
1	P57-03	P1-果汁压力（阀门入口）	0~1000	kPa
2	P57-04	P2-果汁压力（阀门出口）	0~1000	kPa
3	T57-03	T-果汁温度（阀门入口）	0~150	℃
4	FC57-03PV	PV-过程值（果汁流量，第2个蒸发器出口）	0~100	m³/h
5	FC57-03CV	CV-控制值（控制器输出）	0~100²	—
6	FC57-03X	X-伺服电机杆位移	0~100²	—

故障编号	描述	类型
1	阀门堵塞	控制阀故障
2	阀塞或阀座沉淀
3	阀塞或阀座腐蚀
4	阀门或衬套摩擦力增加
5	外部漏电（套管、盖子、端子漏电）
6	内部泄漏（阀门密封性）
7	介质蒸发或临界流量
8	扭转伺服电机的活塞杆	伺服电机故障
9	伺服电机外壳或端子的密封性
10	伺服电机隔膜穿孔
11	伺服电机弹簧故障
12	电空转换器故障	定位器故障
13	杆位移传感器故障
14	压力传感器故障
15	定位器反馈故障
16	定位器供给压降	一般故障/外部故障
17	阀门上出现意外的压力变化
18	旁通阀完全或部分打开
19	流量传感器故障

故障编号	描述	类型
1	阀门堵塞	控制阀故障
2	阀塞或阀座沉淀
3	阀塞或阀座腐蚀
4	阀门或衬套摩擦力增加
5	外部漏电（套管、盖子、端子漏电）
6	内部泄漏（阀门密封性）
7	介质蒸发或临界流量
8	扭转伺服电机的活塞杆	伺服电机故障
9	伺服电机外壳或端子的密封性
10	伺服电机隔膜穿孔
11	伺服电机弹簧故障
12	电空转换器故障	定位器故障
13	杆位移传感器故障
14	压力传感器故障
15	定位器反馈故障
16	定位器供给压降	一般故障/外部故障
17	阀门上出现意外的压力变化
18	旁通阀完全或部分打开
19	流量传感器故障

项目		预测
项目		阳性（P）	阴性（N）
实际	真（T）	TP	TN
实际	假（F）	FP	FN