基于隐马尔可夫模型的煤矿胶带机异常时间点预测*

南京大学学报(自然科学版) ›› 2013, Vol. 49 ›› Issue (2): 217–225.

基于隐马尔可夫模型的煤矿胶带机异常时间点预测*

李小斌^1·2**，任世锦¹，李世银³

出版日期:2015-10-28 发布日期:2015-10-28
作者简介: (1.江苏师范大学计算机科学与技术学院，徐州，221116;2.中国矿业大学计算机科学与技术学院，徐州，221008;
3.中国矿业大学信息与电气工程学院，徐州，221008)
基金资助:
国家"863”计划(2008AA062200)，江苏省产学研联合创新资金(BY2009 114)，徐州市工业科技计划(XX001)，徐
州师范大学重点基金(10XLA13 )

Predication of abnormal accident occurrence time for
mine belt conveyor based on Hidden Markov Model

Li Xiao-Bin^1’2,Ren Shi-Jin¹,Li ShI一Yin²

Online:2015-10-28 Published:2015-10-28
About author: (1 .School of Computer Science and Technology,Xuzhou Normal University,Xuzhou,221116,China;
2. School of Computer Science and Technology ,China University of Mining and Technology，Xuzhou,221008,China;
3. School of Electronic Information and Electrical Enginecring,China University of Mining and Technology,
Xuzhou，221008,China)

摘要/Abstract

摘要： 煤矿胶带输送机的保护可以保障煤矿生产的平稳高效.针对如何有效地对胶带机发生异常的
时刻的预测，提出了一种基于隐马尔可夫(Hiddcn Markov Model, HMM)和其改进型隐式半马尔可夫模
型( Hidden Semi-Markov Model, HSMM)的胶带输送机异常时刻预测的方法.通过对胶带输送机保护传
感器采集的时间序列进行特征提取，建立对应的HMM模型及HSMM模型，对胶带机异常发生时刻进
行预测.在实际生产数据集上的实验表明，HMM和HSMM模型可以有效地对异常事件发生的时间点进行预测.

Abstract: In a mine enterprise generous foods and materials are transmitted by belt conveyor. So it is of great im
portance to guarantee the belt conveyor running stable and efficient.There are many abnormal accidents in produc
tion when belt conveyor is running.Two distinct abnormal accident，i. e. pile coal lightly accident and pile coal severi-
ty accident are inspected in this paper.The pile coal lightly accident should be removed immediately when it hap-
pened,otherwise it will canse the coal severity accident happen. While the pile coal severity accident would cause the
mine production interruption immediately. So how to keep the mine conveyor continuous running is valuable for mine
enterprise production.This paper puts forward a method to predict the abnormal accident occurrence time based on
Hidden Markov Model(HMM)and Hidden Semi-Markov Model( HSMM). Firstly the paper gives a thorough theory
description about the structure and inference for HMM and HSMM. Especially a stated occurrence time predication
algorithm is given based on the HSMM. Large amount of time series is collected through belt conveyor protection
sensors in Pin Ding Shan Mine Enterprise. After feature extraction the corresponding HMM or HSMM model could
be built on these datasets.The accident occurrence time is able to be predicted based on the HMM model or HSMM
model. Experiments carried on the actual production datasets illustrate that HSMM model can effectively predict ab-
normal accident’s occurrence time.

李小斌^1·2**，任世锦¹，李世银³. 基于隐马尔可夫模型的煤矿胶带机异常时间点预测*[J]. 南京大学学报(自然科学版), 2013, 49(2): 217–225.

Li Xiao-Bin^1’2,Ren Shi-Jin¹,Li ShI一Yin². Predication of abnormal accident occurrence time for
mine belt conveyor based on Hidden Markov Model
[J]. Journal of Nanjing University(Natural Sciences), 2013, 49(2): 217–225.

参考文献

[1]Si X S, Wang W, Hu C H，et al. Remaining useful life estimation一A review on the statistical data
driven approaches. European Journal of Operational Research,2011,213(1):1一14.
[2]Lin W ,Yang L L，Xu B L. Speaker identification in Chinese whispered speech based on modified-
MFCC. Journal of Nanjing University(Natural Sciences) , 2006 ,42 ( 1 ) ; 54一62.(林玮，杨莉
莉，徐柏龄.基于修正MFCC参数汉语耳语音的话者识别.南京大学学报(自然科学),2006,2 (1):54一62).
[3]Bunks C, McCarthy D, Al-Ani T.Condition-based maintenance of machines using hidden Markov mod-
els. Mechanical Systems and Signal Processing, 2000，14:597一612.
[4]Baruah P, Chinnam R B. HMMs for diagnostics and prognostics in machining processes. International
Journal of Production Rescarch,2005,43:1275一1293.
[5]Chinnam R B, Baruah P. Autonomous diagnostics and prognostics through competitive learning driven
HMNPbased clustering. Proceedings of the lnternx tional Joint Conference on Neural Networks, 2003，2464:2466一2471.
[6]Camci F, Chinnam R B. Healtlrstate estimation and prognostics in machining processes, IEEE Transac-
dons on Automation Science and Engineering, 2010，7:581一597.
[7]Ocak H，Loparo K A，Discenzo F M. Online tracking of bearing wear using wavelet packet de-
composition and probabilistic modeling; A method for bearing prognostics. Journal of Sound and Vi- bration, 2007，302:951一961.
[8]Tai A H , Ching W K , Chan L Y. Detection of ma- chine failure; Hidden Markov Model approach.
Computers & Industrial Engineering, 2009，57: 608一619.
[9]Zhou Z J，Hu C H,Xu D L,et al. A model for real time failure prognosis based on hidden Markov mod
e1 and belief rule base. European Journal of Opera tional Research,2010,207:269一283.
[10]Yu S Z. Hidden semi-Markov models. Artificial Intelligence, 2010，174;215一243.
[11]Dong M, He D, Banerjee P，et al. Equipment health diagnosis and prognosis using hidden semi-
Markov models. The international Journal of Ad- vanced Manufacturing Technology, 2006，30:738 ~749.
[12]Dong M，He D. Hidden semi-Markov model- based methodology for multi-sensor equipment
health diagnosis and prognosis. European Journal of Operational Rescarch,2007，178:858一878.
[13]Dung M, He D. A segmental hidden semi-Markov model(HSMM)-based diagnostics and prognostics
framework and methodology. Mechanical Systems and Signal Processing,2007,21;2248一2266.
[14]Dong M. A novel approach to equipment health management based on auto-regressive hidden
semi-Markov model(AR-HSMM).Science in China Serices F; information Sciences，2008(51): 1291一1304.
[15]Dung M,Peng Y. Equipment PHM using non-station-ary segmental hidden semi-Markov model. Robotics
and Computer-integrated Manufacturing,2010,27(3): 581一590.
[16]Peng Y，Dong M. A hybrid approach of HMM and grey model for age-dependent health predir
tion of engineering assets. Expert Systems with Applications, 201 1，38(10):12946一12953.
[17]Muhammad M,Majid A，Amin M. Reliability and availability evaluation for a multi-state system
subject to minimal repair. Journal of Applied Sci- ences,2011，11(11):2036一2041.
[18]Rabiner L R. A tutorial on hidden Markov models and selected applications in speech recognition.
Proceedings of the IEEE,1989，77:257一286
[19]Xia Z Y,Yang B,Yang C H,et al. Equipment fault di- agnosis based on state space model and improved hid-
den semi-Nkzrkov model. Nkzthematics in Practice and Theory,2011,41(3):100一107.(夏震宇，杨波，杨
春辉等.基于状态空间模型与改进hsmm相结合的装备故障诊断方法研究.数学的实践与认识，2011,41(3):100一107).
[20]O’Connell J，HØjsgaard S. Hidden semi Markov models for multiple observation sequences;The
mhsmm package for R. Journal of Statistical Soft- ware,2011，39(4):1一22.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed