A fault diagnosis method based on improved kernel Fisher

doi:10.11949/j.issn.0438-1157.20161000

Abstract

Abstract:

A fault diagnosis method of kernel Fisher projection was proposed to solve issues of nonlinear data, complex classes, and difficult fault-diagnosing in chemical processes. The proposed method provided a uniform solution for partial sample mix-up induced by a large difference in category distances and nebulous classification of different category's boundary data after projection of the original data sample. First, an improved category distance was used to change sample distribution in the projection space so that the sample had a good projection. Then, boundary data was screened out by a defined threshold parameter and classified by improved K-Nearest Neighbor (K-NN) algorithm, which none-boundary data was classified by Mahalanobis distance. Simulation in a TE process showed that training accuracy was increased by 2.25% and testing accuracy was increased by 2.41% in the first experiment, whereas training accuracy was increased by 4.75% and testing accuracy was increased by 6.75% in the second experiment. Therefore, the method improved both fault diagnosis time and accuracy.

Key words: kernel Fisher, fault diagnosis, improved K-NN algorithm, experimental verification, optimizing

摘要：

针对化工过程故障数据呈非线性分布，数据类别复杂，难以进行故障诊断的问题，提出一种改进核Fisher的故障诊断方法。对于原始样本数据投影后，样本出现因类间距离存在很大差异性而导致部分类别样本存在混叠的现象，以及不同类别的边界数据归类模糊问题，给出了统一的解决办法。该方法首先采用改进类间距的方法来改变样本投影空间的分布，使得样本具有较好的投影效果，然后通过定义阈值参数来筛选出边界数据，对于边界数据，采用改进的K近邻（K-NN）算法来分类，对于非边界数据，采用马氏距离来分类。最后在TE过程中进行了仿真验证，结果表明方法在兼顾了故障诊断时间的同时，有效提高了故障诊断精度。

关键词: 核Fisher, 故障诊断, 改进K-NN算法, 实验验证, 优化

CLC Number:

TP277

MA Liling, XU Fafu, WANG Junzheng. A fault diagnosis method based on improved kernel Fisher[J]. CIESC Journal, 2017, 68(3): 1041-1048.

马立玲, 徐发富, 王军政. 一种基于改进核Fisher的故障诊断方法[J]. 化工学报, 2017, 68(3): 1041-1048.

References 30

[1]	LENNOX B, HIDEN H G, MONTAGUE G A, et al. Application of multivariate statistical process control to batch operations[J]. Computers and Chemical Engineering, 2000, 24(2/3/4/5/6/7):291-296.
[2]	文成林, 吕菲亚, 包哲静, 等. 基于数据驱动的微小故障诊断方法综述[J]. 自动化学报, 2016, 42(9):1285-1299. WEN C L, LÜ F Y, BAO Z J, et al. A review of data driven-based incipient fault diagnosis[J]. Acta Automatica Sinica, 2016, 42(9):1285-1299.
[3]	李晗, 萧德云. 基于数据驱动的故障诊断方法综述[J]. 控制与决策, 2011, 26(1):1-9, 16. LI H, XIAO D Y. Survey on data driven fault diagnosis methods[J]. Control and Decision, 2011, 26(1):1-9, 16.
[4]	于春梅, 杨胜波, 陈馨, 等. 多元统计方法在故障诊断中的应用综述[J]. 计算机工程与应用, 2007, 43(8):205-208. YU C M, YANG S B, CHEN X, et al. Survey of multivariate statistical methods for fault diagnosis[J]. Computer Engineering and Applications, 2007, 43(8):205-208.
[5]	张可, 周东华, 柴毅, 等. 复合故障诊断技术综述[J].控制理论与应用, 2015, 32(9):1143-1157. ZHANG K, ZHOU D H, CHAI Y, et al. Review of multiple fault diagnosis methods[J]. Control Theory & Applications, 2015, 32(9):1143-1157.
[6]	苏中鲜. 基于数据驱动的工业过程故障诊断方法综述[J]. 软件导刊, 2016, 15(3):149-150. SU Z X. Review of fault diagnosis based on data-driven in industry process[J]. Software Guide, 2016, 15(3):149-150.
[7]	SCHOLKOPF B, MIKA S, BURGES C J C, et al. Input space versus feature space in kernel-based methods[J]. IEEE Transactions on Neural Networks, 1999, 10(5):1000-1016.
[8]	HU W P, HU H F. Heterogeneous face recognition based on modality-independent kernel Fisher discriminant analysis joint sparse auto-encoder[J]. Electronics Letters, 2016, 52(21):1753-1755.
[9]	FRANCK D. A one-class kernel fisher criterion for outlier detection[J]. IEEE Transactions on Neural Networks and Learning Systems, 2015, 26(5):982-994.
[10]	BO L F, WANG L, JIAO L C. Feature scaling for kernel fisher discriminant analysis using leave-one-out cross validation[J]. Neural Computation, 2006, 18(4):961-978.
[11]	VAN GESTEL T, SUYKENS J A K, LANCKRIET G, et al. Bayesian framework for least-squares support vector machine classifiers, Gaussian processes, and kernel Fisher discriminant analysis[J]. Neural Computation, 2002, 14(5):1115-1147.
[12]	VOLKER R. Kernel Fisher discriminants for outlier detection[J]. Neural Computation, 2006, 18(4):942-960.
[13]	JIANKE Z, STEVEN C H H, MICHAEL R L. Face annotation using transductive kernel Fisher discriminant[J]. IEEE Transactions on Multimedia, 2008, 10(1):86-96.
[14]	LIU Q S, LU H Q, MA S D. Improving kernel Fisher discriminant analysis for face recognition[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2004, 14(1):42-49.
[15]	SANG W L, ZEUNGNAM B. Representation of a Fisher criterion function in a kernel feature space[J]. IEEE Transactions on Neural Networks, 2010, 21(2):333-339.
[16]	WON K J, SAUNDERS C, PRÜGEL-BENNETT A. Evolving Fisher kernels for biological sequence classification[J]. Evolutionary Computation, 2013, 21(1):83-105.
[17]	祝志博, 宋执环. 故障分离——一种基于FDA-SVDD的模式分类算法[J]. 化工学报, 2009, 60(8):2010-2016. ZHU Z B, SONG Z H. Fault isolation:an FDA-SVDD based pattern classification algorithm[J]. CIESC Journal, 2009, 60(8):2010-2016.
[18]	SHI H T, LIU J C, WU Y H, et al. Fault diagnosis of nonlinear and large-scale processes using novel modified kernel Fisher discriminant analysis approach[J]. International Journal of Systems Science, 2016, 47(5):1095-1109.
[19]	李锋, 王家序, 汤宝平, 等.有监督不相关局部Fisher判别分析故障诊断[J]. 振动工程学报, 2015, 8(4):657-665. LI F, WANG J X, TANG B P, et al. Fault diagnosis method based on supervised uncorrelated local Fisher discriminant analysis[J]. Journal of Vibration Engineering, 2015, 8(4):657-665.
[20]	王晶, 刘莉, 曹柳林, 等. 基于核Fisher包络分析的间歇过程故障诊断[J]. 化工学报, 2014, 65(4):1317-1326. WANG J, LIU L, CAO L L, et al. Fault diagnosis based on kernel Fisher envelope surface for batch processes[J]. CIESC Journal, 2014, 65(4):1317-1326.
[21]	赵旭, 阎威武, 邵惠鹤, 等. 基于核Fisher判别分析方法的非线性统计过程监控与故障诊断[J]. 化工学报, 2007, 58(4):951-956. ZHAO X, YAN W W, SHAO H H, et al. Nonlinear statistical process monitoring and fault diagnosis based on kernel Fisher discriminant analysis[J]. Journal of Chemical Industry and Engineering(China), 2007, 58(4):951-956.
[22]	苏祖强, 汤宝平, 刘自然, 等. 基于正交半监督局部Fisher判别分析的故障诊断[J]. 机械工程学报, 2014, 50(18):7-13. SU Z Q, TANG B P, LIU Z R, et al. Fault diagnosis method based on orthogonal semi-supervised local Fisher discriminant analysis[J]. Journal of Mechanical Engineering, 2014, 50(18):7-13.
[23]	MIKA S, RATSCH G, WESTON J, et al. Fisher discriminant analysis with kernels[C]//Proceeding of the IEEE International Workshop on Neural Networks for Signal Processing, 1999:41-48.
[24]	彭陈松. 基于核函数Fisher判别的数据分类算法研究[D]. 杭州:浙江理工大学, 2011. PENG C S. Data classification algorithm based on kernel Fisher discriminant[D]. Hangzhou:Zhejiang Sci-Tech University, 2011.
[25]	王炜, 郭小明, 王淑艳, 等. 关于核函数选取的方法[J]. 辽宁师范大学学报(自然科学版), 2008, 31(1):1-4. WANG W, GUO X M, WANG S Y, et al. On the method of kernel-function selection in support vector machine[J]. Journal of Liaoning Normal University (Natural Science Edition), 2008, 31(1):1-4.
[26]	武优西, 郭磊, 柴欣, 等. 基于优化算法的核函数参数选择的研究[J]. 计算机应用与软件, 2010, 27(1):137-140. WU Y X, GUO L, CHAI X, et al. On parameter selection of kernel function based on optimisation algorithm[J]. Computer Applications and Software, 2010, 27(1):137-140.
[27]	张翔, 肖小玲, 徐光祐, 等. 一种确定高斯核模型参数的新方法[J]. 计算机工程, 2007, 33(12):52-53, 56. ZHANG X, XIAO X L, XU G Y, et al. A new method for determining the parameter of Gaussian kernel[J]. Computer Engineering, 2007, 33(12):52-53, 56.
[28]	DAI G, YEUNG D Y, QIAN Y T. Face recognition using a kernel fractional-step discriminant analysis algorithm[J]. Pattern Recognition, 2007, 40(1):229-243.
[29]	DOWNS J J, VOGEL E F. A plant-wide industrial process control problem[J]. Computers and Chemical Engineering, 1993, 17(3):245-255.
[30]	吕宁, 于晓洋. 基于二阶互信息特征选取的TE过程故障诊断[J]. 化工学报, 2009, 60(9):2252-2258. LÜ N, YU X Y. Fault diagnosis in TE process based on feature selection via second order mutual information[J]. CIESC Journal, 2009, 60(9):2252-2258.