Embed Size (px)
Citation preview
Eric Prebys
APC/LARP
The
trut
h is
out t
here
!
2/9/2009
The primary goal of this workshop was to provide the CERN Directorate with information and recommendations to define the scope of activities and final schedule and plan 2009/2010 running of the LHC.
This planning is obviously dominated by the consequences incident of September 19.
Questions about how this incident occurred were examined to the extent that they impact repair and remediation; however, questions regarding the decision process that allowed to occur were deferred to the external review committee.
2/9/2009E. Prebys, AEM
When will the LHC start up? The DG has made a strong commitment to a physics run of the
LHC “in 2009”, provided it does not threaten the long term viability of the accelerator.
Run through the winter? How much does this cost?
How many sectors to warm up? There are currently four sectors warm, and a plan to install new
pressure relief flanges with dramatically increased capacity. To install these new relief flanges on the remaining four sectors
would require a warm up, which has it’s own risks and scheduling headaches.
The baseline plan is to leave the remaining sectors cold, and effect an intermediate pressure relief enhancement by replacing clamps with springs on a number of existing flanges.
Energy Quickly became clear the options were 4 or 5 TeV/beam.
2/9/2009E. Prebys, AEM
After the close of the meeting, Steve Myers made all the workshop talks public on the CERN Indico site, which you can access with this shortcut: http://tinyurl.com/Chamonix2009 Of particular interest is Steve’s summary talk at the end
(very) Partial glossary of terms: “Consolidation” – CERNglish word referring to a set of
actions and/or repairs to address known shortcomings of a system.
“MLI” – multi-layer insulation, which was blown all over the tunnel and sucked into the beam pipe
“DN200” – 200 mm pressure relief flange (“bandaid du jour”).
“QPS” – quench protection system “MCI” – maximum credible incident
2/9/2009E. Prebys, AEM
On September 19th, sector 3-4 was being ramped to 9.3 kA, the equivalent of 5.5 TeV All other sectors had already been ramped to this level Sector 3-4 had previously only been ramped to 7 kA (4.1 TeV)
At 11:18AM, a quench developed in the splice between dipole C24 and quadrupole Q24 Not initially detected by quench protection circuit Power supply tripped at .46 sec Discharge switches activated at .86 sec
Within the first second, and arc formed at the site of the quench Helium pressure rose beyond .13 MPa and ruptured into the
insulation vacuum. Vacuum also degraded in the beam pipe
The pressure at the vacuum barrier reached ~10 bar (design value 1.5 bar). The force was transferred to the magnet stands, which broke.
2/9/2009
*see talk by Philippe LeBrun, Monday AME. Prebys, AEM
Incident on 19th of September 2008 => failure of some supports of SSS in sector 3-4 due to longitudinal loads
*O. Capatina, Tuesday PM
2/9/2009E. Prebys, AEM
Under the rules at the time, there could have been people in adjacent sectors or service areas, who might have been killed or seriously hurt.
It could have occurred in, eg, the dispersion suppression (DS) region, for which spares would have been an issue.
It could have happened in the triplet (which BTW has much better designed joints), and damaged the experiments.
If the arc had cut into both He lines, it would have been 40 kg/s of He, a factor of 20 over design This MCI drives the new pressure relief specification
2/9/2009
*see talk Jim Strait, Wednesday PM
E. Prebys, AEM
Why did the joint fail? Quality control problems Inherent weakness in joint design (no clamps)
Why wasn’t it detected in time? There was indirect (calorimetric) evidence of an ohmic
heat loss, but these data were not routinely monitored The bus quench protection circuit had a threshold of 1V, a
factor of ~1000 too high to detect the quench in time. Why did it do so much damage?
The pressure relief system was designed around an MCI Helium release of 2 kg/s, a factor of ten below what occurred.
2/9/2009E. Prebys, AEM
Cable Junction Box Cross-section
Upper Tin/Silver Soldering alloy Layer
Inter-Cable Tin/Silver Soldering Alloy Layer
Superconducting Cable in Copper
Stabilizer
Upper Copper Profile
Lower Copper U Profile
Lower Tin/Silver Soldering Alloy Layer
Completed Junction
2/9/2009
*see talk by Arjan Verweij, Tuesday PME. Prebys, AEM
Theory: A resistive joint of about 220 nTheory: A resistive joint of about 220 n with bad with bad electrical and thermal contacts with the stabilizerelectrical and thermal contacts with the stabilizer
No electrical contact between wedge and U-profile with the bus on at least 1 side of the
joint
No bonding at joint with the U-profile and the
wedge
A. Verweij
• Loss of clamping pressure on the joint, and between joint and stabilizer
• Degradation of transverse contact between superconducting cable and stabilizer
• Interruption of longitudinal electrical continuity in stabilizer Problem: this is where
the evidence used to be
2/9/2009E. Prebys, AEM
Use calorimetric data to find suspicious joints, then check with precision resistance measurements: 1 sigma = 20 nOhms (design value .6 nOhm) Thermal runaway @ ~85 nOhms @ 5TeV Two additional bad joints found, both of them inside
magnets, which were removed. All new joints photographed and check with
ultrasound.
Cryodipole 2334 right splice
0
20
40
60
80
100
0 20 40 60 80 100Time (a.u.)
Tra
nsm
itte
d p
ow
er (
a.u
.)
right splice, far leftright splice, center leftright splice, center rightright splice, far rightacceptance limit
Cryodipole 2334 left splice
0
20
40
60
80
100
0 20 40 60 80 100Time (a.u.)
Tra
ns
mit
ted
po
wer
(a.
u.)
left splice, far leftleft splice, center leftleft splice, center rightleft splice, far rightacceptance limit
GOOD BAD
2/9/2009E. Prebys, AEM
-10 mK
S1-2 S2-3 S3-4 S4-5
S5-6 S6-7 S7-8 S8-1
7 kA 7 kA 7 kA 9.3 kA
7 kA 7 kA 8.5 kA 7 kA
+40 mK
-10 mK
Rel
ativ
e te
mp
+40 mK
Rel
ativ
e te
mp
1-2 hour flat tops
6 February 2009
Smoking gun?
In the future, calorimetric data will be routinely monitored
2/9/2009
*See talk by Nuria Catalan Lasheras, Monday AME. Prebys, AEM
Old quench protection circuit triggered at 1V on bus. New QPS triggers at .3 mV
Factor of 3000 Should be sensitive down to 25 nOhms (thermal runaway at 7
TeV) Can measure resistences to <1 nOhm
Concurrently installing improved quench protection for “symmetric quenches” A problem found before September 19th
Worrisome at >4 TeV
2/9/2009
*See talks by Arjan Verveij and Reiner Denz, Tuesday PM
E. Prebys, AEM
2 kg/s
20 kg/s
40 kg/s
DP1
1.52
2.53
3.54
4.5
0 20 40 60 80 100 120
Vac
encl
osur
e P
[bar
]
Vac enclosure He T [K]
2 kg/s20 kg/s
40 kg/s
DP
1
1.1
1.2
1.3
1.4
1.5
1.6
0 20 40 60 80 100 120
Vac
encl
osur
e P
[bar
]
Vac enclosure He T [K]
New configuration on cold sectors: Turn several existing flanges into pressure reliefs (while cold). Also reinforce stands to hold ~3 bar
New configurationon warm sectors: new flanges(12 DN200 + 8 DN100)
(DP: Design Pressure)
L. Tavian2/9/2009
*Vittorio Parma and Ofelia Capatina, Tuesday PM E. Prebys, AEM
15 SSS (MQ)1 not removed (Q19) 14 removed
8 cold mass revamped (old CM, partial de-cryostating for cleaning and careful inspection of supports and other components)
6 new CMs In this breakdown there is
consideration about timing (SSS cryostating tales long time; variants problems).
42 Dipoles (MBs)3 not removed
(A209,B20,C20)39 removed
9 Re-used (old CM, no decryostating –except one?)
30 new CMs New cold masses are
much faster to prepare than rescuing doubtful dipoles)
2/9/2009
*Lucio Rossi (magnets) and Francesco Bertinelli (connections), Tues AM
E. Prebys, AEM
Most dirty beam pipes in removed magnets Several must be cleaned in tunnel
Soot: successive passes with wet and dry sponges MLI: little vacuum cleaner
2/9/2009
*Vincent Baglin, Tuesday AM
E. Prebys, AEM
Magnet reinstallation begins this week. Work on flanges begins in a few weeks. Decision of Directorate (announced today)
Do NOT warm up remaining four sectors Start up in late September, collisions in late October Run through winter until next fall (with short break for
Christmas, of course) Raise energy to 5 TeV as soon as safe (symmetric QPS?) Collect ~200 pb-1 at 5+5
200 pb-1 @ 5+5 TeV ~ Tevatron Run II so far Possible ion run at the end
2/9/2009
*See Friday talks, particularly Steve Myers’ summary
E. Prebys, AEM
We shouldn’t forget how incredibly well things were going before the incident. Beam was circulated in both directions within a few hours Beam was captured within a day Optics measurements uncovered a few minor wiring problems.
Right
Wrong
2/9/2009
*Thursday AM
E. Prebys, AEM
Will NOT happen this year Problem with dipoles “forgetting their training”
after the period of storage Only occurs with one out of the three vendors (Noel) This was discovered before the turn on Project LOTS of quenches to go beyond 6 TeV
6.5 TeV: ~10 per sector 7.0 TeV: ~100 per sector
Can only do 2-3 quenches/day 1 to 2 months of quenching to get to 7 TeV Risk of damage?
2/9/2009
*Arjan Verveij, Monday AM and Ezio Todesco, Thursday PM E. Prebys, AEM
