Optimal Condition-Based Maintenance Replacement based on Logical Analysis of Data (LAD)

CATEGORY: CONDITION BASED MAINTENANCE
Proceedings of the International MultiConference of Engineers and Computer Scientists 2013 Vol II, IMECS 2013, March 13 - 15, 2013, Hong Kong
Optimal Condition-Based Maintenance Replacement based on Logical Analysis of Data (LAD)
Alireza Ghasemi, Sasan Esameili
A. Ghasemi is an Assistant Professor with Department of Industrial Engineering of Dalhousie University, Halifax, NS, Canada (alireza.ghasemi@dal.ca ).
S. Esmaeili is a graduated Master’s student form Department of Industrial Engineering of Dalhousie University. Abstract
This paper develops equipment optimal condition based replacement model, using Logical Analysis of Data (LAD). LAD is a powerful classification method that does not relying on any statistical theory which enables LAD to overcome the usual problems concerning the statistical properties of the data. LAD profits from a straightforward procedure and self-explanatory results.
In this paper, our objective is to develop an optimal replacement method by taking its working condition (condition monitoring data) into consideration using LAD. Using equipment’s survival probability and associated costs of scheduled and non-scheduled replacements, an optimal replacement method is introduced. The proposed method is applied on a hypothetical problem and its easy to understand approach and its high performance is shown. Analysis of performance of the proposed methods reveals that the methods provide self-explanatory results that are greatly beneficial to maintenance practitioners.
Introduction
Since its introduction, LAD has been applied for the analysis of data in different fields such as medicine, biotechnology, economics, finance, politics, properties, oil exploration, manufacturing and maintenance . Recently LAD was used for diagnosis of equipment failure. LAD has proved to be a promising technique that provides interpretable results that are comparable to most pioneer techniques in the field of diagnostics in CBM.  improved LAD methodology to predict equipment’s chance of survival at each observation moment when new data on attributes of the equipment is available. It showed that LAD provides comprehensible results that are greatly beneficial to maintenance practitioners in prognosticating fault in machinery. In this work, we will introduce an optimal replacement model that minimizes the maintenance cost of equipment considering its condition monitoring data, using LAD.
Free Full Text Source: http://www.iaeng.org/publication/IMECS2013/IMECS2013_pp830-833.pdf 

Tolkku – a toolbox for decision support from condition monitoring data

CATEGORY: CONDITION MONITORING
25th International Congress on Condition Monitoring and Diagnostic Engineering, Journal of Physics: Conference Series 364 (2012)
Tolkku – a toolbox for decision support from condition monitoring data
Olli Saarela 1,3 , Mikko Lehtonen 1, Jari Halme 1, Antti Aikala 1 and Kimmo Raivio 2
olli.saarela@vtt.fi
1 VTT Technical Research Centre of Finland, P.O.Box 1000, FI-02044 VTT, Finland
2 Aalto University, School of Science and Technology, P.O.Box 15400,
FI-00076 Aalto, Finland
Abstract.
Describes a software toolbox designed for condition monitoring and diagnosis of machines. This toolbox implements both new methods and prior art and is aimed at practical down-to-earth data analysis work.
The target is to improve knowledge of the operation and behaviour of machines and processes throughout their entire life-cycles. The toolbox supports different phases of condition based maintenance with tools that extract essential information and automate data processing. The paper discusses principles that have guided toolbox design and the implemented toolbox structure. Case examples are used to illustrate how condition monitoring applications can be built using the toolbox. In the first case study the toolbox is applied to fault detection of industrial centrifuges based on measured electrical current. The second case study outlines an application for centralized monitoring of a fleet of machines that supports organizational learning.
Free Full Text Source: http://iopscience.iop.org/1742-6596/364/1/012044