[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Main Index]

[lcdds 271] Re: modified assembly scheme for GLD

Subject:   [lcdds 271] Re: modified assembly scheme for GLD
From:   Yasuhiro Sugimoto <yasuhiro.sugimoto@xxxxxx>
Date:   Fri, 15 Sep 2006 17:37:25 +0900

Dear Andrei, and all,

   I'm sorry for late reply. I will  discuss  on the issues one by one.
My  conclusion is that pure CMS assembly is more costly than
modified CMS assembly.

>Dear Sugimoto san and all,
>
>1) If a 400t crane is installed in the hall, one
>need to add 5M$ crane cost to the cost of extra
>volume or rock that need to be excavated to install
>this large crane. Plus, the cost of support of the crane.
>The actual cost may be three or more times higher.

   Please look at the excel  file which I attached.  I estimated
the cost of experimental halls and related things which would be
different between pure-CMS and modified-CMS assembly.
   The extra cost due to extra space occupied by the 300ton crane
is about 1.6M$.  If the experimental hall  has straight walls like
ATLAS,  the crane rails can be put at the top of the straight wall
and the increase of the wall cost is negligible. Please look at the
right side figure of the attached pdf file "exphall060915.pdf".
If the wall  is curved like CMS, I think that means the cavern is
made by NATM and  many anchor bolts are put into the  wall rock.
So the crane  rails can be supported by these anchor bolts and the
increase of the cost is also negligible.

>The pure-CMS assembly, even if it requires larger
>20m shaft, appears to be more cost effective.
>I will forward a corresponding tentative assessment
>from Jean-Luc Baldy.

     The assumption made by Jean-Luc  seems  somewhat biased
to support  pure-CMS assembly.
     The cost  increase for the larger shaft is quite large.
In my estimation, 20m shaft is  9M$ more expensive
than 16m shaft. If the ground is soft or wet,  the total shaft cost
increases and the cost difference increases.
    CMS has two 80ton cranes in the surface assembly hall.
The cost for these cranes is not included in your estimation.
In case of ILC,  we will use the support tube which could be
30ton or more. We will use many concrete shieldings. How do you
put these with 20 ton crane in the cavern?

>Could you please discuss the modified CMS assembly versus
>pure CMS assembly?

   The extra detector cost by using pure-CMS like assembly is not
negligible. We have not  estimated  the cost increase precisely,
but it would be more than 3M$.

    Although we do not want to determine the assembly procedure
just by the cost,  pure-CMS like assembly does not seem more
cost effective than modified CMS like assembly.


>2) How often do you plan to extract muon detectors
>for the maintenance? If it is a seldom procedure,
>this also can be done with 22m hall width, if the detector
>could be moved along the beamline close to the wall.
>(see scheme attached to [lcdds 260])
>
>Please discuss if 22m width would be sufficient.

    It is difficult to tell exact number of failure rate of muon detector
because the technology to be used for the muon detector
is not established.
    The cost difference between  widths of 25m and 22m  is
about 3M$. I don't think we should take a risk to save 3M$.
Please remember that 10% efficiency loss of the detector
corresponds to more than 10M$/year loss of operation cost.
     Another reason I prefer 25m  width is that there could be
many other things which occupies the space but is not drawn
in the figure,  such as  electronics hut, cables, pipes, and so on.
     Crane accessibility is also an issue. If the hall width is 25m,
the upper half of concrete shielding between the detector
and the wall  can be moved  by the 300t crane  (lower half
can be moved by air-pads).  The larger hall width lowers the
risk of destroying the inner detectors by  hitting them
with the support tube in its installation.


>3) Is GLD designed to be self shielding?
>In this case the shielding around beamline, between detector
>and the tunnel wall, shown in your schemes,
>may need to contain concrete and iron.
>Here iron (placed inside) is helping to slow down
>high energy neutrons. A pure concrete shield would need
>to be much thicker.
>
>4) Iron, contained in the shield around the beamline,
>would change the distribution of field leaking from
>the detector. It probably would shield the field
>on the beamline completely.
>
>Could you please include the iron in the shield into
>consideration of leaking magnetic field?

     The magnetic filed  inside the iron shield becomes
much weaker.  However,  outside the iron shield,
the B-field becomes much stronger. The iron works
as a flux concentrator.  Please look at the attached file
"gld_g3_FeShield.pdf".  In this calculation, I put 50cm thick
iron cylinder from R=1.5m  extending  from z=7.6m to 12.5m.
So the inner shield should be  made by non-magnetic
stainless steel of copper.


>There is still an issues of leaking field in the hall.
>What are the limits of occupational exposure to static
>magnetic fields in Japan?

   I have no idea on this issue.


>5) With respect to your note on 4th concept:
>
>"I thought that the 4th concept would produce very
>large leakage field because they do not have iron yoke.
>I actually calculated the B-field of the 4th concept,
>and found that my guess was completely wrong."
>
>Indeed, it looks that double solenoid concept, which
>returns the flux, allows reduction of the amount of iron.
>
>Is it worth discussing to apply this double solenoid concept
>to GLD to reduce amount of iron and the overall size of detector?

       The stored energy of GLD solenoid is 1.7GJ.  That of 4th concept
is  2.8GJ in my calculation and 5.7GJ in their  detector outline document.
If the double solenoid  is used for GLD, the stored energy would be  3.7GJ
extrapolating from my calculation. The cost of solenoid is  roughly
proportional to E**0.662 according to the particle data book.
So, the double solenoid is 1.67 times more expensive than single solenoid.
About 1/3 of GLD cost  comes from the solenoid cost.
So, using double solenoid  is out of question from the view point of
cost optimization. Probably you can buy another experimental hall
by the cost increase of the solenoid.


    In order to check the cost of reinforced concrete,  I  made google
search by the keywords of "concrete" and "cost" in Japanese.
Most of the hit page is on the "Aneha case".  Aneha is a Japanese
architect who made illegal  design of buildings  in order to reduce
the cost as much as possible. But the cost reduction by that illegal
design was very small fraction (much less than 1%) of the total
cost of the buildings. He lost his carrier,  some companies which
asked design to him and gave him a lot of pressure to reduce the cost
were bankrupted. Some  hotels  who bought the building  which
became unusable from  these companies  were bankrupted.
I hope ILC will not lose physics output by excessive  cost reduction.


Best regards,

Yasuhiro Sugimoto
KEK


Attachment: Exphall_cost.xls
Description: application/applefile

Attachment: exphall060915.pdf
Description: application/applefile

Attachment: gld_g3_FeShield.pdf
Description: application/applefile