Availability and Reliability Issues for ILC Page: 4 of 4
This article is part of the collection entitled: Office of Scientific & Technical Information Technical Reports and was provided to Digital Library by the UNT Libraries Government Documents Department.
The following text was automatically extracted from the image on this page using optical character recognition software:
determine what broke and when its repair was finished.
The recovery time started from this time. In this way extra
downtime caused by things which broke during the
original repair were counted in the recovery time in a
manner similar to the simulation. The results are shown in
Figure 1 where the hundreds of downtimes are averaged
into 3 points. There is a large scatter in the recovery to
downtime ratio. The measured slope is 0.32 0.08. This
compares to the 20% that was used to simulate the ILC
The simulation was written in the MATLAB scripting
language. It was written in a general way with nothing
ILC specific in the code. All of the machine dependent
input data was contained in a spreadsheet which was read
by the MATLAB program. The spreadsheet included a set
of macros which aided in handling the large amount of
data needed to describe the number, locations, MTBFs,
etc. of the components. The MATAB program output
results to another spreadsheet which allowed one to
examine in detail which devices caused the downtime. It
also provided summaries of which type of device caused
the downtime and how much each region of the
accelerator contributed to the downtime.
This software could be useful (and has been used) in
the design of other accelerators. The MATLAB code,
spreadsheets and instructions are available at http://www-
SOME ILC RESULTS
Simulation results have been used to help make several
ILC design decisions and to establish an unavailability
budget for the various systems and components.
Use for design decisions
AvailSim was used to compare the uptimes of versions
of the ILC with slightly different designs. We could then
decide whether the loss in uptime was worth the cost
savings. For example, putting both DRs in a single tunnel
decreased the uptime by less than one percent while
saving the cost of a >6 km tunnel. This was cost effective
so the baseline ILC design has both DRs in a single
tunnel. In contrast, if all the klystrons, modulators, and
power supplies were in the same tunnel as the linac
instead of in a separate tunnel, the uptime was reduced by
about 14%. This availability was too low to be acceptable,
but it was judged too expensive and too large a risk to try
to improve the reliability of individual components
sufficiently to regain the 14% loss. Hence, the baseline
ILC design has two tunnels for the linac.
Many similar comparisons have been mad to aid in
other design decisions.
For a given set of MTBFs and MTTRs, availSim gives
the total downtime and the downtime caused by each type
of component. Using this information the MTBFs were
tuned to improve the downtime from the 30% originally
predicted by using the MTBFs in Table 1 to the desired
15%. One could have accomplished the same availability
improvement by increasing all the MTBFs by a factor of
2, but instead the MTBFs of the devices that were
dominating the downtime were increased by a larger ratio
while other device MTBFs were left unchanged. The
result is a set of MTBFs that if achieved will allow the
ILC to have an 85% uptime. Water cooled magnets and
their power supplies and power supply controllers are the
components which need the largest improvement (a factor
of 10-20). This is not surprising as these devices are
typically single points of failure and there are thousands
of them in the ILC design. Hardware R&D projects have
been started to develop high availability versions of some
of the devices that need significantly improved MTBFs.
The availability simulation has been a valuable tool in
the design of the ILC. It has been used to make major
design decisions and to determine what components need
to have reliabilities much greater than has typically been
achieved in operating accelerators. The tool will continue
to guide the ILC design and help determine where efforts
should be concentrated on design and high availability
R&D. It is a general purpose tool that can be used for
This article can be searched. Note: Results may vary based on the legibility of text within the document.
Tools / Downloads
Get a copy of this page or view the extracted text.
Citing and Sharing
Basic information for referencing this web page. We also provide extended guidance on usage rights, references, copying or embedding.
Reference the current page of this Article.
Himel, T.; Nelson, J.; Phinney, N.; /SLAC; Ross, M. & /Fermilab. Availability and Reliability Issues for ILC, article, June 27, 2007; [Menlo Park, California]. (digital.library.unt.edu/ark:/67531/metadc878283/m1/4/: accessed January 21, 2019), University of North Texas Libraries, Digital Library, digital.library.unt.edu; crediting UNT Libraries Government Documents Department.