基于LSTM-DAE的化工故障诊断方法研究

张敬川, 田慧欣

北京化工大学学报(自然科学版) ›› 2021, Vol. 48 ›› Issue (2) : 108-116.

PDF(1441 KB)
欢迎访问北京化工大学学报(自然科学版),今天是 2025年5月8日 星期四
Email Alert  RSS
PDF(1441 KB)
北京化工大学学报(自然科学版) ›› 2021, Vol. 48 ›› Issue (2) : 108-116. DOI: 10.13543/j.bhxbzr.2021.02.014
机电工程和信息科学

基于LSTM-DAE的化工故障诊断方法研究

  • 张敬川1, 田慧欣1,2
作者信息 +

Fault diagnosis of chemical process based on long short-term memory (LSTM)-denoising auto-encoder (DAE)

  • ZHANG JingChuan1, TIAN HuiXin1,2
Author information +
文章历史 +

摘要

现代化工过程愈加精密化、复杂化,使得化工过程数据呈现高度非线性、强耦合等特点,传统的故障诊断模型难以学习此类数据的有效特征表示,且无法挖掘隐含的时间序列信息。针对上述问题,提出了一种长短期记忆(long short-term memory,LSTM)网络与降噪自编码器(denoising auto-encoder,DAE)结合的LSTM-DAE化工故障诊断方法,用基于LSTM的特殊编码方式代替传统DAE模型的全连接网络编码方式,并结合高斯噪声处理和全连接网络解码层,搭建出基于LSTM-DAE的特征提取模型,最后由Softmax分类器输出故障诊断结果。该方法结合了DAE与LSTM的优点,确保了编码特征获取的有效性。使用田纳西-伊斯曼(Tennessee-Eastman,TE)过程数据设计所提方法与其他5类模型的对比实验,实验结果表明:在多故障诊断效果上,基于LSTM-DAE的化工故障诊断方法的训练集正确率达到了96.02%,测试集正确率达到了97.31%,平均误报率仅为0.65%,平均漏检率仅为3.19%,在6类模型中为最优;在单故障诊断效果上,基于LSTM-DAE的化工故障诊断方法能够提高对故障14的分辨能力,并缩短对故障18的检测延迟时间,有效降低了漏检率,表明所提方法能够在实际化工过程中进行有效的故障诊断。

Abstract

Modern chemical processes are becoming increasingly precise and complicated. It is difficult for traditional fault diagnosis (FD) models to learn the feature representation of raw data for high dimensional, nonlinear, and tightly coupled data. Furthermore, traditional FD models cannot extract the hidden time series information inside the raw data. Therefore, a novel FD model of chemical processes called LSTM-DAE based on long short-term memory (LSTM) and a denoising auto-encoder (DAE) has been proposed. By changing the full-connected encoding network of the traditional DAE model to a novel LSTM-DAE encoding network, and combining the Gaussian noise and fully-connected decoding network, a feature-extracting LSTM-DAE model has been established, with the final FD results given by a Softmax classifier. The proposed model combines the advantages of both DAE and LSTM, which ensures high efficacy in feature extraction. The experimental results for the Tennessee-Eastman (TE) process showed that in terms of multi-fault FD performance, the accuracy of the training set was 96.02%, the accuracy of the test set was 97.31%, the mean false alarm rate (FAR) was only 0.65%, and the mean miss detection rate (MDR) was only 3.19%, which is the best among all the six FD models. In terms of single-fault FD performance, the LSTM-DAE model can improve the resolution capability of fault 14 and reduce the delay time of fault 18, which reduces the MDR. The above analysis indicates that the proposed LSTM-DAE model can efficiently detect faults in actual chemical processes.

关键词

故障诊断 / 田纳西-伊斯曼过程 / 降噪自编码器 / 长短期记忆网络 / Softmax分类器

Key words

fault diagnosis / Tennessee-Eastman process / denoising auto-encoder / long short-term memory network / Softmax classifier

引用本文

导出引用
张敬川, 田慧欣. 基于LSTM-DAE的化工故障诊断方法研究[J]. 北京化工大学学报(自然科学版), 2021, 48(2): 108-116 https://doi.org/10.13543/j.bhxbzr.2021.02.014
ZHANG JingChuan, TIAN HuiXin. Fault diagnosis of chemical process based on long short-term memory (LSTM)-denoising auto-encoder (DAE)[J]. Journal of Beijing University of Chemical Technology, 2021, 48(2): 108-116 https://doi.org/10.13543/j.bhxbzr.2021.02.014

参考文献

[1] 胡志新. 基于深度学习的化工故障诊断方法研究[D]. 杭州:杭州电子科技大学, 2018. HU Z X. Research on chemical fault diagnosis methods based on deep learning[D]. Hangzhou:Hangzhou Dianzi University, 2018. (in Chinese)
[2] GE Z Q, SONG Z H, GAO F R. Review of recent research on data-based process monitoring[J]. Industrial & Engineering Chemistry Research, 2013, 52(10):3543-3562.
[3] VENKATASUBRAMANIAN V, RENGASWAMY R, YIN K, et al. A review of process fault detection and diagnosis Part I:quantitative model-based methods[J]. Computers and Chemical Engineering, 2003, 27(3):293-311.
[4] VENKATASUBRAMANIAN V, RENGASWAMY R, KAVURI S N. A review of process fault detection and diagnosis Part Ⅱ:qualitative models and search strategies[J]. Computers and Chemical Engineering, 2003, 27(3):313-326.
[5] VENKATASUBRAMANIAN V, RENGASWAMY R, KAVURI S N, et al. A review of process fault detection and diagnosis Part Ⅲ:process history based methods[J]. Computers and Chemical Engineering, 2003, 27(3):327-346.
[6] 杜海莲, 苗诗瑜, 杜文霞,等. 改进PCA方法在化工过程中的故障诊断研究[J]. 山东科技大学学报(自然科学版), 2017, 36(5):16-22. DU H L, MIAO S Y, DU W X, et al. Research on fault diagnosis of chemical process based on improved PCA method[J]. Journal of Shandong University of Science and Technology (Natural Science), 2017, 36(5):16-22. (in Chinese)
[7] 韩敏, 张占奎. 基于改进核主成分分析的故障检测与诊断方法[J]. 化工学报, 2015, 66(6):2139-2149. HAN M, ZHANG Z K. Fault detection and diagnosis method based on modified kernel principal component analysis[J]. CIESC Journal, 2015, 66(6):2139-2149. (in Chinese)
[8] LEE J M, QIN S J, LEE I B. Fault detection of non-linear processes using kernel independent component analysis[J]. The Canadian Journal of Chemical Engineering, 2007, 85(4):526-536.
[9] 吕鹏飞, 闫云聚, 荔越. 基于马氏距离的改进核Fisher化工故障诊断研究[J]. 自动化学报, 2020,46(11):2379-2391. LV P F, YAN Y J, LI Y. Research on fault diagnosis of improved kernel fisher based on Mahalanobis distance in the field of chemical industry[J]. Acta Automatica Sinica, 2020, 46(11):2379-2391. (in Chinese)
[10] 陈剑雪. ACO-BP算法在化工过程故障诊断中的应用[J]. 化工自动化及仪表, 2012, 39(7):872-875. CHEN J X. Application of ACO-BP algorithm to fault diagnosis in chemical process[J]. Control and Instruments in Chemical Industry, 2012, 39(7):872-875. (in Chinese)
[11] 张鑫, 胡瑾秋, 张来斌,等. 基于RS和SVM的化工过程高精度故障诊断方法[J]. 石油学报(石油加工), 2017, 33(4):777-784. ZHANG X, HU J Q, ZHANG L B, et al. High-accuracy fault diagnosis of chemical processes based on RS and SVM[J]. Acta Petrolei Sinica (Petroleum Processing Section), 2017, 33(4):777-784. (in Chinese)
[12] 冀丰偲, 余云松, 张早校. LDA_SVM方法在化工过程故障诊断中的应用[J]. 高校化学工程学报, 2020, 34(2):487-494. JI F C, YU Y S, ZHANG Z X. Application of LDA and SVM method in fault diagnosis of chemical process[J]. Journal of Chemical Engineering of Chinese Universities, 2020, 34(2):487-494. (in Chinese)
[13] 谭莉, 于春梅. 基于PCA-LVQ神经网络的化工过程故障诊断[J]. 工业控制计算机, 2016, 29(11):86-87. TAN L, YU C M. Chemical process fault diagnosis based on PCA-LVQ neural network[J]. Industrial Control Computer, 2016, 29(11):86-87. (in Chinese)
[14] 薄翠梅, 乔旭, 张广明,等. 基于ICA-SVM的复杂化工过程集成故障诊断方法[J]. 化工学报, 2009, 60(9):2259-2264. BO C M, QIAO X, ZHANG G M, et al. ICA-SVM based fault diagnosis method for complex chemical process[J]. CIESC Journal, 2009, 60(9):2259-2264. (in Chinese)
[15] 冯倩玉. 基于降噪正交自编码器的TE过程故障诊断[D]. 上海:上海交通大学, 2018. FENG Q Y. Fault diagnosis based on denoising orthogonal auto-encoder in TE process[D]. Shanghai:Shanghai Jiao Tong University, 2018. (in Chinese)
[16] ZHENG S D, ZHAO J S. A new unsupervised data mining method based on the stacked autoencoder for chemical process fault diagnosis[J]. Computers and Chemical Engineering, 2020, 135:106755.
[17] 张祥, 崔哲, 董玉玺,等. 基于VAE-DBN的故障分类方法在化工过程中的应用[J]. 过程工程学报, 2018, 18(3):590-594. ZHANG X, CUI Z, DONG Y X, et al. Application of fault classification method based on VAE-DBN in chemical process[J]. The Chinese Journal of Process Engineering, 2018, 18(3):590-594. (in Chinese)
[18] WANG Y L, PAN Z F, YUAN X F, et al. A novel deep learning based fault diagnosis approach for chemical process with extended deep belief network[J]. ISA Transactions, 2020, 96:457-467.
[19] 程换新, 王建庆. 改进深度置信网络对TE过程故障诊断研究[J]. 电子测量技术, 2019, 42(9):117-120. CHENG H X, WANG J Q. Improved DBN for TE process fault diagnosis[J]. Electronic Measurement Technology, 2019, 42(9):117-120. (in Chinese)
[20] 黄家华, 钱龙, 易军,等. 基于参数优化的深度信念网络TE过程故障诊断[C]//第30届中国过程控制会议(CPCC 2019). 昆明,2019:15. HUANG J H, QIAN L, YI J, et al. Fault diagnosis in deep belief network based on parameter optimization for TE process[C]//Chinese Process Control Conference 2019. Kunming, 2019:15. (in Chinese)
[21] 衷路生, 吴春磊. 基于AC-CNN模型的过程故障识别[J]. 计算机工程与设计, 2020, 41(2):542-549. ZHONG L S, WU C L. Fault recognition based on AC-CNN model[J]. Computer Engineering and Design, 2020, 41(2):542-549. (in Chinese)
[22] 程诚, 任佳. 基于自适应卷积核的改进CNN数值型数据分类算法[J]. 浙江理工大学学报, 2019, 41(5):657-664. CHENG C, REN J. Improved CNN classification algorithm based on adaptive convolution kernel for numerical data[J]. Journal of Zhejiang Sci-Tech University, 2019, 41(5):657-664. (in Chinese)
[23] 王翔, 柯飂挺, 任佳. 样本重构多尺度孪生卷积网络的化工过程故障检测[J]. 仪器仪表学报, 2019, 40(11):181-188. WANG X, KE L T, REN J. Chemical industrial process fault detection based on sample reconstruction multi-scale siamese CNN[J]. Chinese Journal of Scientific Instrument, 2019, 40(11):181-188. (in Chinese)
[24] 程诚, 任佳. 一种基于雷达图表示的数值型数据的CNN分类方法[J]. 信息与控制, 2019, 48(4):429-436. CHENG C, REN J. A classification method of CNN for numerical data based on radar chart representation[J]. Information and Control, 2019, 48(4):429-436. (in Chinese)
[25] 衷路生, 夏相明. 基于深度残差网络的化工过程故障诊断[J].过程工程学报,2020,20(12):1483-1490. ZHONG L S, XIA X M. Fault diagnosis for chemical processes based on deep residual network[J]. The Chinese Journal of Process Engineering, 2020, 20(12):1483-1490. (in Chinese)
[26] HOCHREITER S, SCHMIDHUBER J. Long short-term memory[J]. Neural Computation, 1997, 9(8):1735-1780.
[27] YUAN M, WU Y T, LIN L. Fault diagnosis and remaining useful life estimation of aero engine using LSTM neural network[C]//IEEE/CSAA International Conference on Aircraft Utility Systems (AUS). Beijing, 2016:135-140.
[28] LU W N, LI Y P, CHENG Y, et al. Early fault detection approach with deep architectures[J]. IEEE Transactions on Instrumentation and Measurement, 2018, 67(7):1679-1689.
[29] 王路瑶, 吴斌, 杜志敏,等. 基于长短期记忆神经网络的数据中心空调系统传感器故障诊断[J]. 化工学报, 2018, 69(S2):252-259. WANG L Y, WU B, DU Z M, et al. Sensor fault detection and diagnosis for data center air conditioning system based on LSTM neural network[J]. CIESC Journal, 2018, 69(S2):252-259. (in Chinese)
[30] ZHAO H T, SUN S Y, JIN B. Sequential fault diagnosis based on LSTM neural network[J]. IEEE Access, 2018, 6:12929-12939.
[31] GONG J, KIM H. RHSBoost:improving classification performance in imbalance data[J]. Computational Statistics and Data Analysis, 2017, 111:1-13.

基金

天津市自然科学基金(18JCYBJC22000);天津市企业科技特派员项目(19JCTPJC47600)
PDF(1441 KB)

2152

Accesses

0

Citation

Detail

段落导航
相关文章

/