Browse Prior Art Database

Method for Estimating Equipment Reliability

IP.com Disclosure Number: IPCOM000113802D
Original Publication Date: 1994-Oct-01
Included in the Prior Art Database: 2005-Mar-27
Document File: 6 page(s) / 189K

Publishing Venue

IBM

Related People

Bradley, J: AUTHOR

Abstract

Disclosed is a model and control process for estimating the reliability, or anticipating the performance, of a system or equipment. The model or process starts with the definition of a statistical data base for each component within a mechanical tool. A program then delineates a set of rules in a program to generate anticipated failure times of the tool based on statistical data. The statistical data and generated failure times, along with actual data collected from the mechanical tool, predict the reliability, anticipating the performance of the system, equipment, or tool.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 43% of the total text.

Method for Estimating Equipment Reliability

      Disclosed is a model and control process for estimating the
reliability, or anticipating the performance, of a system or
equipment.  The model or process starts with the definition of a
statistical data base for each component within a mechanical tool.  A
program then delineates a set of rules in a program to generate
anticipated failure times of the tool based on statistical data.  The
statistical data and generated failure times, along with actual data
collected from the mechanical tool, predict the reliability,
anticipating the performance of the system, equipment, or tool.

      This model and process, called RELI for Reliability Estimates
of Equipment's Life Images, is used during the stages of developing a
production tool to provide an estimate, concerning product flow,
downtime, uptime, thruput, product loss, and efficiency.  This method
can be used to facilitate design improvements and to predict cost
estimates, part replacements, planned maintenance, levels of parts
inventories, etc.  This method can also be used after a system is
installed and operating to determine the RAS (Reliability,
Availability, and Serviceability) efficiency of the machine, and to
predict where efforts should be placed to create a better product and
to increase thruput.

      Without this process, "stress tests" are used to provide
answers to such questions.  These tests are lengthy runs on each unit
at different tolerance environments, such as accelerated speed,
extreme temperatures, or deviations in levels of vibrations.  While
the results of such tests are used to make changes to the design of
the system being built, the applications of these results to advance
the design of other tools is minimal.

      Simulation and modeling can assist in answering most of these
questions.  However, the concept of simulation is to monitor the
process flow of a machine over a short period, such as 24 hours, or
for a period derived from a particular occurrence.  While an
assumption is made that this kind of simulation is done during the
"good" life of the machine, data is not provided concerning the
desired length of a burn-in process, the variations of individual
units relative to the life distribution curves of each component,
when to anticipate the beginning of a period in which the unit is
wearing out, and the expected availability and performance of the
system in six months, five years, and ten years.  There are
simulation systems that can perform life studies on mechanical tools,
but while these systems are statistical and analytical, they are not
oriented to process flow.

      Fig. 1 provides an overview of the RELI prediction package.  At
1, a statistical data base is developed for each component within a
mechanical tool, including parameters such as the MTTF (Mean Time To
Failure, MTR (Mean Time for Repair), the shape or standard deviation
of a failure distribution curve, etc. ...