Research Papers

Online Degradation Assessment and Adaptive Fault Detection Using Modified Hidden Markov Model

[+] Author and Article Information
Seungchul Lee

Department of Mechanical Engineering, University of Michigan-Ann Arbor, 1210 H. H. Dow, 2300 Hayward Street, Ann Arbor, MI 48109-2136seunglee@umich.edu

Lin Li1

Department of Mechanical Engineering, University of Michigan-Ann Arbor, 1035 H. H. Dow, 2300 Hayward Street, Ann Arbor, MI 48109-2136lilz@umich.edu

Jun Ni

Department of Mechanical Engineering, University of Michigan-Ann Arbor, 1023 H. H. Dow, 2300 Hayward Street, Ann Arbor, MI 48109-2136junni@umich.edu


Corresponding author.

J. Manuf. Sci. Eng 132(2), 021010 (Apr 01, 2010) (11 pages) doi:10.1115/1.4001247 History: Received May 18, 2009; Revised February 08, 2010; Published April 01, 2010; Online April 01, 2010

Online condition monitoring and diagnosis systems play an important role in the modern manufacturing industry. This paper presents a novel method to diagnose the degradation processes of multiple failure modes using a modified hidden Markov model (MHMM) with variable state space. The proposed MHMM is combined with statistical process control to quickly detect the occurrence of an unknown fault. This method allows the state space of a hidden Markov model to be adjusted and updated with the identification of new states. Hence, the online degradation assessment and adaptive fault diagnosis can be simultaneously obtained. Experimental results in a turning process illustrate that the tool wear state can be successfully detected, and previously unknown tool wear processes can be identified at the early stages using the MHMM.

Copyright © 2010 by American Society of Mechanical Engineers
Your Session has timed out. Please sign back in to continue.



Grahic Jump Location
Figure 1

Basic form of a HMM

Grahic Jump Location
Figure 2

Block diagram of the proposed modified HMM algorithm

Grahic Jump Location
Figure 3

Markov chain with an unknown state S5

Grahic Jump Location
Figure 4

Original observable signals and the HMM with the four known states

Grahic Jump Location
Figure 5

HMM algorithm serving to estimate the states

Grahic Jump Location
Figure 6

Posterior probability of P{q(n)=Si∣O(1),…,O(n)}

Grahic Jump Location
Figure 7

Result of a wrong state estimation

Grahic Jump Location
Figure 8

Posterior probability, but no wiggling is shown

Grahic Jump Location
Figure 9

Test bed of the turning process with coolant supply (Ft: thrust force; Fc: cutting force)

Grahic Jump Location
Figure 10

(a) Normal turning process with coolant and (b) different tool wear modes without coolant

Grahic Jump Location
Figure 13

Control chart with MHMM

Grahic Jump Location
Figure 14

The estimated states via various algorithms: (a) HMM, (b) neural network, (c) GMM, and (d) K-means

Grahic Jump Location
Figure 11

Estimated states: vertical dashed lines indicate the true states, while the solid lines represent the estimated states

Grahic Jump Location
Figure 12

Control chart with conventional HMM




Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In