基于多尺度特征能量-核极限学习机的双循环流化床气流堵塞故障智能诊断

doi:10.11949/0438-1157.20181334

摘要/Abstract

摘要：

为减轻双循环流化床结块与堵塞故障对生物质气化反应的负面影响，提出基于多尺度特征能量-核极限学习机(kernel extreme learning machine, KELM)的故障诊断模型。首先对故障状态下压力信号采用小波分解获得多尺度信号，然后提取各尺度特征能量作为特征向量，最后将其输入经遗传算法优化的核极限学习智能诊断模型，实现双循环流化床气流堵塞故障的智能诊断。通过对公开的轴承故障数据集和双循环流化床冷态实验系统数据的分类识别分析，并与基于变分模态分解和样本熵特征提取的KELM诊断模型进行比较，结果表明：本模型具有较高的故障诊断精度(82.5%)，能够有效提取故障特征，用于双循环流化床气流堵塞的高效分类识别。

关键词: 故障诊断, 循环流化床, 多尺度, 遗传算法, 核极限学习机

Abstract:

In order to alleviate the negative impact of dual circulating fluidized bed agglomeration and plugging failure on biomass gasification reaction, a fault diagnosis model based on multi-scale feature energy-kernel extreme learning machine (KELM) was proposed. Firstly, wavelet decomposition is used to obtain the multi-scale signal in the fault state, then the feature energy of each scale is extracted as the feature vector, and finally input into the nuclear limit learning intelligent diagnosis model optimized by genetic algorithm to realize the intelligent diagnosis of dual circulating fluidized bed airflow blockage faults. Through the classification and analysis of the open bearing fault data set and the dual-circulating fluidized bed cold experimental system data, and compared with the KELM diagnostic model based on variational mode decomposition and sample entropy feature extraction, the results show that the model has higher fault diagnosis accuracy (82.5%), which can effectively extract fault characteristics and be used for efficient classification and identification of dual circulating fluidized bed airflow blockage.

Key words: fault diagnosis, circulating fluidized bed, multiscale, genetic algorithm, kernel extreme learning machine

中图分类号:

TK 229

杨新, 麻哲瑞, 申赫男, 陈鸿伟. 基于多尺度特征能量-核极限学习机的双循环流化床气流堵塞故障智能诊断[J]. 化工学报, 2019, 70(7): 2616-2625.

Xin YANG, Zherui MA, Henan SHEN, Hongwei CHEN. Fault diagnosis of airflow jamming fault in double circulating fluidized bed based on multi-scale feature energy and KELM[J]. CIESC Journal, 2019, 70(7): 2616-2625.

图/表 14

参考文献 36

1	陈鸿伟, 刘焕志, 李晓伟, 等. 双循环流化床颗粒循环流率试验与BP神经网络预测[J]. 中国电机工程学报, 2010, 30(32): 25-29.
	ChenH W, LiuH Z, LiX W, et al. Experimental research on solids circulation rate in a double fluidized bed and BP neural network prediction[J]. Proceedings of the CSEE, 2010, 30(32): 25-29.
2	姜华伟. 基于风帽压力波动的流化床气固流态化特征研究[D]. 北京: 华北电力大学, 2013.
	JiangH W. Research on gas-solid fluidization characteristics of fluidized beds based on pressure fluctuations in wind caps[D]. Beijing: North China Electric Power University, 2013.
3	林伟国, 张鹏, 陈磊, 等. 流化床反应器结块故障的声纹特征提取及监测技术[J]. 化工学报, 2012, 63(9): 2851-2858.
	LinW G, ZhangP, ChenL, et al. Voiceprint extraction and monitoring of fluidized bed reactor agglomeration fault[J]. CIESC Journal, 2012, 63(9): 2851-2858.
4	CaiB, HuangL, XieM. Bayesian networks in fault diagnosis[J]. IEEE Transactions on Industrial Informatics, 2017, 13(5): 2227-2240.
5	RokkamR G, SowinskiA, FoxR O, et al. Computational and experimental study of electrostatics in gas-solid polymerization fluidized beds[J]. Chemical Engineering Science, 2013, 92: 146-156.
6	杨大为, 冯辅周, 赵永东, 等. VMD样本熵特征提取方法及其在行星变速箱故障诊断中的应用[J]. 振动与冲击, 2018, 37(16): 198-205.
	YangD W, FengF Z, ZhaoY D, et al. A VMD sample entropy feature extraction method and its application in planetary gearbox fault diagnosis[J]. Journal of Vibration and Shock, 2018, 37(16): 198-205.
7	龙英, 何怡刚, 张镇, 等. 基于小波变换和ICA特征提取的开关电流电路故障诊断[J]. 仪器仪表学报, 2015, 36(10): 2389-2400.
	LongY, HeY G, ZhangZ, et al. Switched current circuit fault diagnosis based on wavelet transform and ICA feature extraction[J]. Chinese Journal of Scientific Instrument, 2015, 36(10): 2389-2400.
8	MohantyS, GuptaK K, RajuK S. Hurst based vibro-acoustic feature extraction of bearing using EMD and VMD[J]. Measurement, 2018, 117: 200-220.
9	史历程, 赵骁, 赵群飞, 等. 基于小波能谱熵和集成经验分解的传感器故障诊断耦合算法研究[J].动力工程学报, 2018, 38(8): 624-632.
	ShiL C, ZhaoX, ZhaoQ F, et al. Study on computing algorithm of sensor fault diagnosis based on WEE and EEMD[J]. Journal of Chinese Society of Power Engineering, 2018, 38(8): 624-632.
10	梁绍华, 郑立刚, 周昊, 等. 基于支持向量机与高斯分布估计的低NO_x排放[J]. 化工学报, 2009, 60(1): 223-229.
	LiangS H, ZhengL G, ZhouH, et al. Low NO_x emissions based on support vector machine and Gaussian estimation of distribution[J]. CIESC Journal, 2009, 60(1): 223-229.
11	CaiB, ZhaoY, LiuH, et al. A data-driven fault diagnosis methodology in three-phase inverters for PMSM drive systems[J]. IEEE Transactions on Power Electronics, 2017, 32(7): 5590-5600
12	CaiB P, LiuH L, XieM. A real-time fault diagnosis methodology of complex systems using object-oriented Bayesian networks[J]. Mechanical Systems and Signal Processing, 2016, 80: 31-44.
13	张秦梫, 宋辉, 姜勇, 等. 基于OS-ELM的变压器局部放电模式识别[J]. 高电压技术, 2018, 44(4): 1122-1130.
	ZhangQ Q, SongH, JiangY, et al. Partial discharge pattern recognition of transformer based on OS-ELM [J]. High Voltage Engineering, 2018, 44(4): 1122-1130.
14	王亚, 周孟然, 陈瑞云,等.到基于多层正则极限学习机的煤矿突水光谱判别方法[J].光学学报, 2018, 38(7): 367-376.
	WangY, ZhouM R, ChenR Y, et al. Identification method of coal mine water inrush spectrum based on multilayer regularization extreme learning machine[J]. Acta Optica Sinica, 2018, 38(7): 367-376.
15	牛培峰, 王枭飞, 刘楠, 等. ASOS-ELM建模方法及在汽轮机热耗率预测中的应用[J]. 化工学报, 2018, 69(9): 3924-3931.
	NiuP F, WangX F, LiuN, et al. Modeling method of ASOS-ELM and its application in prediction of heat rate of steam turbine[J]. CIESC Journal, 2018, 69(9): 3924-3931.
16	HuangG B, ZhouH, DingX, et al. Extreme learning machine for regression and multiclass classification[J]. IEEE Transactions on System, Man and Cybernetics, Part B (Cybernetics), 2012, 42(2): 513-529.
17	李琨, 韩莹, 佘东生, 等. 基于IFOA-KELM-MEA模型的游梁式抽油机采油系统井下工况的短期预测[J]. 化工学报, 2017, 68(1): 188-198.
	LiK, HanY, SheD S, et al. IFOA-KELM-MEA model based transient prediction on down-hole working conditions of beam pumping units[J]. CIESC Journal, 2017, 68(1): 188-198.
18	RauberT W, Oliveira-SantosT, BoldtF D A, et al. Kernel and random extreme learning machine applied to submersible motor pump fault diagnosis[C]// 2017 International Joint Conference on Neural Networks. Anchorage: IEEE Computational Intelligence Society, 2017: 3347-3354.
19	ChenZ, WuL, ChengS, et al. Intelligent fault diagnosis of photovoltaic arrays based on optimized kernel extreme learning machine and Ⅰ-Ⅴ characteristics[J]. Applied Energy, 2017, 204: 912-931.
20	JiangY, WuJ, ZongC. An effective diagnosis method for single and multiple defects detection in gearbox based on nonlinear feature selection and kernel-based extreme learning machine[J]. Journal of Vibroengineering, 2014, 16(1): 499-512.
21	DragomiretskiyK, ZossoD. Variational mode decomposition[J]. IEEE Transactions on Signal Processing, 2014, 62(3): 531-544.
22	LiuR N, YangB Y, ZioE, et al. Artificial intelligence for fault diagnosis of rotating machinery: a review[J]. Mechanical Systems and Signal Processing, 2018, 108: 33-47.
23	NasimiE, GabbarH A. Signal de-noising methods for fault diagnosis and troubleshooting at CANDU (R) stations[J]. Nuclear Engineering and Design, 2014, 280: 481-492.
24	BhattacharyyaA, SharmaM, PachoriR B, et al. A novel approach for automated detection of focal EEG signals using empirical wavelet transform[J]. Neural Computing & Applications, 2018, 29(8): 47-57.
25	WangD, ZhaoY, YiC, et al. Sparsity guided empirical wavelet transform for fault diagnosis of rolling element bearings[J]. Mechanical Systems and Signal Processing, 2018, 101: 292-308.
26	YangT Y, LeuL P. Multiresolution analysis on identification and dynamics of clusters in a circulating fluidized bed[J]. American Institute Of Chemical Engineers, 2009, 55(3): 612-629.
27	张永俊, 王嘉骏, 顾雪萍, 等. 气固搅拌流化床中压力脉动特性[J]. 化工学报, 2016, 67(2): 494-503.
	ZhangY J, WangJ J, GuX P, et al. Pressure fluctuation in gas-solid agitated fluidized bed[J]. CIESC Journal, 2016, 67(2): 494-503.
28	YumM K, JungK Y, KangH C, et al. Effect of a ketogenic diet on EEG: analysis of sample entropy[J]. Seizure, 2008, 17(6): 561-566.
29	HuangG, HuangG B, SongS, et al. Trends in extreme learning machines: a review[J]. Neural Networks, 2015, 61: 32-48.
30	HuangG, HuangG B. An insight into extreme learning machines: random neurons, random features and kernels[J]. Cognitive Computation, 2014, 6(3): 376-390.
31	AlencarA S C, RochanetoA R, GomesJ P P. A new pruning method for extreme learning machines via genetic algorithms[J]. Applied Soft Computing Journal, 2016, 44: 101-107.
32	SmithW A, RandallR B. Rolling element bearing diagnostics using the Case Western Reserve University data: a benchmark study[J]. Mechanical Systems and Signal Processing, 2015, 64/65: 100-131.
33	刘长良, 武英杰, 甄成刚, 等. 基于变分模态分解和模糊C均值聚类的滚动轴承故障诊断[J]. 中国电机工程学报, 2015, 35(13): 3358-3365.
	LiuC L, WuY J, ZhenC G, et al. Rolling bearing fault diagnosis based on variational mode decomposition and fuzzy C means clustering[J]. Proceedings of the CSEE, 2015, 35(13): 3358-3365.
34	WangB, HuX, WangW, et al. Fault diagnosis using improved pattern spectrum and fruit fly optimization algorithm-support vector machine[J]. Advances in Mechanical Engineering, 2018, 10(11): 1-10.
35	俞啸. 数据驱动的滚动轴承故障特征分析与诊断方法研究[D]. 徐州: 中国矿业大学, 2017.
	YuX. Rolling bearing fault feature analysis and diagnosis method based on data-driven model[D]. Xuzhou: China University of Mining and Technology, 2017.
36	MaoW T, HeL, YanY J, et al. Online sequential prediction of bearings imbalanced fault diagnosis by extreme learning machine[J]. Mechanical Systems and Signal Processing, 2017, 83(15): 450-473.

模型		诊断精度
BPNN	79.28%	61.80%
SVM	90.77%	76.53%
ELM	95.38%	70.25%
KELM	96.33%	83.00%
VMD-SE-KLEM	93.89%	84.00%
WD-SE-KELM	85.67%	92.67%
VMD-FE-KELM	94.33%	95.00%
WD-FE-KELM	95.78%	95.33%

模型		诊断精度
BPNN	79.28%	61.80%
SVM	90.77%	76.53%
ELM	95.38%	70.25%
KELM	96.33%	83.00%
VMD-SE-KLEM	93.89%	84.00%
WD-SE-KELM	85.67%	92.67%
VMD-FE-KELM	94.33%	95.00%
WD-FE-KELM	95.78%	95.33%

K	IMF1	IMF2	IMF3	IMF4	IMF5	IMF6	IMF7	IMF8	IMF9	IMF10
1	4.33×10^-5	—	—	—	—	—	—	—	—	—
2	3.71×10^-5	7.99	—	—	—	—	—	—	—	—
3	3.71×10^-5	7.99	40.41	—	—	—	—	—	—	—
4	3.65×10^-5	7.97	21.12	40.41	—	—	—	—	—	—
5	3.64×10^-5	7.96	17.46	25.68	40.49	—	—	—	—	—
6	3.60×10^-5	7.96	17.23	24.77	33.13	40.91	—	—	—	—
7	3.62×10^-5	7.91	13.06	19.55	25.72	33.48	41.00	—	—	—
8	3.60×10^-5	7.92	13.44	19.97	25.91	33.33	40.44	47.22	—	—
9	3.60×10^-5	7.88	11.41	17.10	21.94	27.46	34.07	40.53	47.24	—
10	2.95×10^-5	4.04	7.96	11.80	17.24	22.02	27.49	33.92	40.51	47.23

K	IMF1	IMF2	IMF3	IMF4	IMF5	IMF6	IMF7	IMF8	IMF9	IMF10
1	4.33×10^-5	—	—	—	—	—	—	—	—	—
2	3.71×10^-5	7.99	—	—	—	—	—	—	—	—
3	3.71×10^-5	7.99	40.41	—	—	—	—	—	—	—
4	3.65×10^-5	7.97	21.12	40.41	—	—	—	—	—	—
5	3.64×10^-5	7.96	17.46	25.68	40.49	—	—	—	—	—
6	3.60×10^-5	7.96	17.23	24.77	33.13	40.91	—	—	—	—
7	3.62×10^-5	7.91	13.06	19.55	25.72	33.48	41.00	—	—	—
8	3.60×10^-5	7.92	13.44	19.97	25.91	33.33	40.44	47.22	—	—
9	3.60×10^-5	7.88	11.41	17.10	21.94	27.46	34.07	40.53	47.24	—
10	2.95×10^-5	4.04	7.96	11.80	17.24	22.02	27.49	33.92	40.51	47.23

系统状态	特征提取方法	尺度1/IMF2	尺度2/IMF3	尺度3/IMF4	尺度4/IMF5	尺度5/IMF6	尺度6/IMF7	尺度7/IMF8
边壁堵塞	VMD-SE	0.11167	0.29565	0.53492	0.46485	0.23394	0.06911	0.00174
	WD-SE	0.31554	0.43322	0.66762	0.48942	0.15439	0.05994	0.00892
	VMD-FE	0.29395	0.83809	0.74122	0.12259	0.01358	0.00084	0.00069
	WD-FE	0.10558	0.46791	0.85912	0.17637	0.02507	0.00716	0.00637
结块	VMD-SE	0.03594	0.00046	0.00062	0.00227	0.00133	0.00029	0.00059
	WD-SE	0.96752	0.11002	0.11603	0.08652	0.06287	0.15519	0.05303
	VMD-FE	0.02199	0.02447	0.04116	0.00386	0.00013	0.00005	0.00166
	WD-FE	0.46983	0.43053	0.60514	0.23291	0.07700	0.10832	0.39473