0
Research Papers

Machine Tool Condition Monitoring Based on an Adaptive Gaussian Mixture Model

[+] Author and Article Information
Jianbo Yu

School of Mechatronic Engineering and Automation,  Shanghai University, #149 Yan Chang Road, Shanghai 200072, People's Republic of Chinajianboyu@shu.edu.cn

J. Manuf. Sci. Eng 134(3), 031004 (Apr 25, 2012) (13 pages) doi:10.1115/1.4006093 History: Received March 18, 2011; Revised January 30, 2012; Published April 24, 2012; Online April 25, 2012

Indirect, online tool wear monitoring is one of the most difficult tasks in the context of industrial machining operation. The challenge is how to construct an effective model that can consistently exemplify the degradation propagation of tool performance (i.e., tool wear) based on a continuous acquisition of multiple sensor signals. This paper proposes an adaptive Gaussian mixture model (AGMM) to provide a comprehensible and robust indication (i.e., Kullback–Leibler (KL) divergence) for quantifying tool performance degradation. Based on dynamic learning rate, parameter updating, and merge and split of Gaussian components, AGMM is capable of online adaptively learning the dynamic changes of tool performance in its full life. Furthermore, the performance changes of tools are quantified by measuring the distance between two density distributions approximated by the AGMM and the baseline GMM trained by the normal data, respectively. Experimental results of its application in a machine tool test demonstrate the effectiveness of the AGMM-based KL-divergence indication for assessment of tool performance degradation.

FIGURES IN THIS ARTICLE
<>
Copyright © 2012 by American Society of Mechanical Engineers
Your Session has timed out. Please sign back in to continue.

References

Figures

Grahic Jump Location
Figure 2

Flow chart of AGMM algorithm

Grahic Jump Location
Figure 1

AGMM-based prognostics system for tool performance degradation assessment

Grahic Jump Location
Figure 3

Distance calculation between two Gaussian components

Grahic Jump Location
Figure 4

Signals of one running sample (entry-milling-exit cutting procedure) from two sensors: (a) AE sensor on table and (b) vibration sensor on table

Grahic Jump Location
Figure 5

Inconsistent degradation patterns of RMS of AE and vibration sensors on table for the run cases 1–4 from subgroup G11: (a) RMS and (b) WE1

Grahic Jump Location
Figure 6

Scatter plot of training data with two principal components along with data distribution estimation of GMM for four baseline data sets: (a) Databsg11, (b) Databsg12, (c) Databsg21, and (d) Databsg22

Grahic Jump Location
Figure 7

Scatter plot of samples data with two principal components along with data distribution estimation of AGMM in the full life of tools for four subgroups: (a) G11, (b) G12, (c) G21, and (d) G22

Grahic Jump Location
Figure 8

KL-divergence monitoring charts for full life of tools from four subgroups: (a) G11, (b) G12, (c) G21, and (d) G22

Grahic Jump Location
Figure 9

LLP monitoring charts for full life of tools from four subgroups: (a) G11, (b) G12, (c) G21, and (d) G22

Grahic Jump Location
Figure 10

KL-divergence charts for the run case 1 at different running sample time points: (a) the ninth running sample is finished, and (b) the last running sample is finished

Grahic Jump Location
Figure 11

KL-divergence charts for the run case 3 at different running sample time points: (a) the ninth running sample is finished, and (b) the last running sample is finished

Tables

Errata

Discussions

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