Use of data mining techniques for mine machine condition monitoring

Author

Hui Hu

Keywords and Phrases

Vital Information Management System (VIMS); Major factor analysis

Abstract

"Mining technology plays an increasingly important role in mining. As a result, the cost of maintenance in general and repairs in particular contributes more and more to the overall mine operation cost.

In this work advanced data mining techniques are used to build models that predict machinery operation to facilitate improvements in its performance and reduce unscheduled maintenance and repair costs. These involve downloading machine condition and performance data from on-board monitoring systems and discovering the knowledge contained in this data.

DB2 Universal Database Management System and Intelligent Miner of IBM were selected to store, preprocess, mine and interpret the condition and performance data collected from Caterpillar off-highway trucks. The raw data collected in the field was transferred and preprocessed using different interfaces and techniques to make efficient data mining possible. Statistical and data mining algorithms of the Intelligent Miner software package were studied and used as tools for knowledge discovery.

The Major Factor Analysis, a statistical technique, was identified to allow discovery of important information contained within Vital Information Management System data. It gives a prospective view of the key performance indicators for the studied machines.

Decision Tree and Neural Network Classification tools were found to facilitate construction of a predictive model of machine condition. This model allows the prediction of machine failures or malfunctions within certain time horizons"--Abstract, page iii.

Advisor(s)

Golosinski, Tad S.

Committee Member(s)

Bullock, Richard Lee, 1929-
Grayson, R. Larry
Tien, Jerry C.
Newkirk, Joseph William

Department(s)

Mining Engineering

Degree Name

Ph. D. in Mining Engineering

Comments

Accompanying CD-ROM, available at Missouri S&T Library, contains Appendix for chapters 4-9.

Publisher

University of Missouri--Rolla

Publication Date

Fall 2003

Pagination

xiv, 116 pages; CD-ROM

Note about bibliography

Includes bibliographical references (pages 112-115).

Rights

© 2003 Hui Hu, All rights reserved.

Document Type

Dissertation - Restricted Access

File Type

text

Language

English

Subject Headings

Data mining

Thesis Number

T 8345

Print OCLC #

55056170

Recommended Citation

Hu, Hui, "Use of data mining techniques for mine machine condition monitoring" (2003). Doctoral Dissertations. 1498.
https://scholarsmine.mst.edu/doctoral_dissertations/1498