基于灰色神经网络组合模型的故障预测

doi:10.3969/j.issn.1001-506X.2020.01.32

摘要/Abstract

摘要：

针对装备故障预测存在有效样本少、模型预测精度低等问题，集成灰色理论和神经网络方法，提出基于灰色神经网络的故障预测组合模型。基于新信息优先原理和重构背景值方法优化灰色GM（1，1）模型的初始值与背景值，利用Levenberg-Marquardt算法改进反向传播神经网络模型；采用组合预测思想，将多方法融合改进灰色模型和神经网络模型，分别构建基于权重分配、基于误差修正和基于结构优化的3种灰色神经网络组合模型。以某雷达发射机的故障预测为例，验证上述方法在故障预测中的有效性。结果表明，灰色神经网络组合模型的预测精度优于单一预测模型，可用于装备的故障预测和预测性维修。

关键词: 故障预测, 灰色模型, 神经网络, 组合模型

Abstract:

To solve the problems of equipment's failure prediction, including insufficient effective samples and low accuracy of the prediction model, the grey theory and neural network method are integrated in this study, and the combined models are proposed. On the basis of the new information priority principle and the reconstruction background value method, the initial and background values of the GM(1, 1) model are optimized, and the back propagation neural network model is improved by the Levenberg-Marquardt algorithm. By using combined forecasting theorem, the improved grey model and neural network model are integrated with multiple approaches and three combined models are established based on weight allocation, error correction and structural optimization respectively. Selecting the failure prediction of a radar transmitter as an example, the effectiveness of the proposed methods is verified. The results show that compared with the existing prediction model, more accurate failure prediction results can be obtained with the proposed combined models. Therefore, it is helpful for equipment's failure prediction and predictive maintenance.

Key words: failure prediction, grey model, neural network, combined model

中图分类号:

TB114.3

黄魁, 苏春. 基于灰色神经网络组合模型的故障预测[J]. 系统工程与电子技术, 2020, 42(1): 238-244.

Kui HUANG, Chun SU. Failure prediction based on combined model of grey neural network[J]. Systems Engineering and Electronics, 2020, 42(1): 238-244.

图/表 11

图1

图2

图3

图4

图5

表1

表2

表3

图6

表4

表5

参考文献 26

1	BAPTISTA M , SANKARARAMAN S , MEDEIROS I P D , et al. Forecasting fault events for predictive maintenance using data-driven techniques and ARMA modeling[J]. Computers and Industrial Engineering, 2018, 115, 41- 53. doi: 10.1016/j.cie.2017.10.033
2	DE FARIA H , COSTA J G S , OLIVAS J L M . A review of monitoring methods for predictive maintenance of electric power transformers based on dissolved gas analysis[J]. Renewable and Sustainable Energy Reviews, 2015, 46, 201- 209.
3	HE Y H , GU C C , CHEN Z X , et al. Integrated predictive maintenance strategy for manufacturing systems by combining quality control and mission reliability analysis[J]. International Journal of Production Research, 2017, 55 (19): 5841- 5862. doi: 10.1080/00207543.2017.1346843
4	KAN M S , TAN A C C , MATHEW J . A review on prognostic techniques for non-stationary and non-linear rotating systems[J]. Mechanical Systems and Signal Processing, 2015, (62/63): 1- 20.
5	ZHU S P , HUANG H Z , PENG W , et al. Probabilistic physics of failure-based framework for fatigue life prediction of aircraft gas turbine discs under uncertainty[J]. Reliability Engineering and System Safety, 2016, 146, 1- 12. doi: 10.1016/j.ress.2015.10.002
6	JIN G , MATTHEWS D , FAN Y , et al. Physics of failure-based degradation modeling and lifetime prediction of the momentum wheel in a dynamic covariate environment[J]. Engineering Failure Analysis, 2013, 28 (3): 222- 240.
7	ERIK E K , JOHN D S . Physics of failure as a basis for solder elements reliability assessment in wind turbines[J]. Reliability Engineering and System Safety, 2012, 108, 100- 107. doi: 10.1016/j.ress.2012.06.020
8	GUAN S , YAN L H , PENG C . Application of regression algorithm of LS-SVM in tool wear prediction[J]. China Mechanical Engineering, 2015, 26 (2): 217- 222.
9	FINK O , ZIO E , WEIDMANN U . Predicting component reliability and level of degradation with complex-valued neural networks[J]. Reliability Engineering and System Safety, 2014, 121, 198- 206. doi: 10.1016/j.ress.2013.08.004
10	LI X , DING Q , SUN J Q . Remaining useful life estimation in prognostics using deep convolution neural networks[J]. Reliability Engineering and System Safety, 2018, 172, 1- 11. doi: 10.1016/j.ress.2017.11.021
11	赵建忠, 徐廷学, 叶文, 等. 基于数据融合和改进MUGM (1, m, w)的导弹装备故障预测[J]. 系统工程与电子技术, 2015, 37 (4): 832- 837.
	ZHAO J Z , XU T X , YE W , et al. Missile equipment fault prediction based on data fusion and improved MUGM (1, m, w)[J]. Systems Engineering and Electronics, 2015, 37 (4): 832- 837.
12	SHI M J , JIANG L , FU Y H . Study on prediction methods for the fault state of rotating machinery based on dynamic grey model and metabolism grey model[J]. Wireless Personal Communications, 2018, 102 (4): 3615- 3627. doi: 10.1007/s11277-018-5395-0
13	ZHANG Y W , ZOU Y J , LIU Z F , et al. Prediction of ship main engine exhaust gas temperature using AR model[J]. Applied Mechanics and Materials, 2015, 697, 244- 248.
14	PHAM H T , YANG B S . Estimation and forecasting of machine health condition using ARMA/GARCH model[J]. Mechanical Systems and Signal Processing, 2010, 24 (2): 546- 558. doi: 10.1016/j.ymssp.2009.08.004
15	ZHOU H W , HUANG J Q , LU F . Reduced kernel recursive least squares algorithm for aero-engine degradation prediction[J]. Mechanical Systems and Signal Processing, 2017, 95, 446- 467. doi: 10.1016/j.ymssp.2017.03.046
16	QIAN Y N , YAN R Q . Remaining useful life prediction of rolling bearings using an enhanced particle filter[J]. IEEE Trans.on Instrumentation and Measurement, 2015, 64 (10): 2696- 2707. doi: 10.1109/TIM.2015.2427891
17	KAYACAN E , BARIS U , OKYAY K . Grey system theory-based models in time series prediction[J]. Expert Systems with Applications, 2010, 37 (2): 1784- 1789.
18	YIN K D , GENG Y , LI X M . Improved grey prediction model based on exponential grey action quantity[J]. Journal of Systems Engineering and Electronics, 2018, 29 (3): 560- 570. doi: 10.21629/JSEE.2018.03.13
19	刘思峰, 谢乃明. 灰色系统理论及其应用[M]. 北京: 科学出版社, 2012.
	LIU S F , XIE N M . Grey system theory and application[M]. Beijing: Science Press, 2012.
20	YE Z , KIM M K . Predicting electricity consumption in a building using an optimized back-propagation and Levenberg-Marquardt back-propagation neural network:case study of a shopping mall in China[J]. Sustainable Cities and Society, 2018, 42, 176- 183. doi: 10.1016/j.scs.2018.05.050
21	SU L , ZHA Z , LU X , et al. Using BP network for ultrasonic inspection of flip chip solder joints[J]. Mechanical Systems and Signal Processing, 2013, 34 (1/2): 183- 190.
22	王明涛. 确定组合预测权系数最优近似解的方法研究[J]. 系统工程理论与实践, 2000, 20 (3): 104- 109. doi: 10.3321/j.issn:1000-6788.2000.03.018
	WANG M T . Study on method of calculating optimal approximate solution about weight coefficients of combined forecasting methods[J]. System Engineering Theory and Practice, 2000, 20 (3): 104- 109. doi: 10.3321/j.issn:1000-6788.2000.03.018
23	CUI K , JING X . Research on prediction model of geotechnical parameters based on BP neural network[J]. Neural Computing and Applications, 2018, 29 (6): 1- 11.
24	MARKOPOULOS A P , GEORGIOPOULOS S , MANOLAKOS D E . On the use of back propagation and radial basis function neural networks in surface roughness prediction[J]. Journal of Industrial Engineering International, 2016, 12 (3): 389- 400. doi: 10.1007/s40092-016-0146-x
25	LIU H , TIAN H Q , LIANG X F , et al. Wind speed forecasting approach using secondary decomposition algorithm and Elman neuralnetworks[J]. Applied Energy, 2015, 157, 183- 194. doi: 10.1016/j.apenergy.2015.08.014
26	AHMADI M A , EBADI M , HOSSEINI S M . Prediction breakthrough time of water coning in the fractured reservoirs by implementing low parameter support vector machine approach[J]. Fuel, 2014, 117, 579- 589.

序号	电压值/kV
1	1.278
2	1.652
3	1.419
4	2.136
5	1.834
6	2.337
7	2.629
8	3.241
9	3.463
10	3.327
11	3.684
12	4.219
13	4.943
14	4.767
15	5.593
16	5.875
17	6.598

序号	实际值	DI-GM		LM-BP		GM-BP1		GM-BP2		GM-BP3
序号	实际值	预测值	误差	预测值	误差	预测值	误差	预测值	误差	预测值	误差
13	4.943	4.693	-0.250	4.626	-0.317	4.659	-0.284	4.741	-0.202	4.474	-0.469
14	4.767	5.077	0.310	4.841	0.074	4.956	0.189	5.113	0.346	4.931	0.164
15	5.593	5.711	0.118	5.753	0.160	5.732	0.139	5.395	-0.198	5.434	-0.159
16	5.875	6.300	0.425	5.517	-0.358	5.900	0.025	5.747	-0.128	5.989	0.114
17	6.598	6.950	0.352	6.828	0.230	6.888	0.290	6.694	0.096	6.600	0.002

模型名称	平均绝对误差/kV	平均相对误差/%
DI-GM	0.291 1	5.25
LM-BP	0.227 6	4.08
GM-BP1	0.185 5	3.41
GM-BP2	0.193 9	3.70
GM-BP3	0.181 6	3.55

序号	实际值	LM-RBF		LM-Elman		LS-SVM
序号	实际值	预测值	误差	预测值	误差	预测值	误差
13	4.943	4.683	-0.260	4.751	-0.192	5.116	0.173
14	4.767	5.142	0.375	5.058	0.291	5.373	0.606
15	5.593	6.107	0.514	5.937	0.344	5.475	-0.118
16	5.875	5.561	-0.314	5.741	-0.134	6.051	0.176
17	6.598	6.819	0.221	6.452	-0.146	6.837	0.239

模型名称	平均绝对误差/kV	平均相对误差/%
LM-RBF	0.336 8	6.20
LM-Elman	0.221 4	4.13
LS-SVM	0.262 4	4.99