Title: The%20LHC%20from%20construction%20to%20commissioning
1The LHC from construction to commissioning
- Lucio Rossi
- CERN Accelerator Technology Dept
2Particle Physics is one of the two frontiers of
the physics spectrum
Why accelerators? To investigate Particle Physics
Accelerators
Microscopes
Binoculars
Optical, radio télescopes
Particle physics looks at matter in its smallest
dimensions and accelerators are very fine
microscope or, better, atto-scope! ? h/p
_at_LHC T 1 TeV ? ? ? 10-18 m
3back to Big Bang
- Trip back toward the Big Bang t?s?1/E2Gev
- T ? 1 ps for single particle creation
- T ? 1 ?s for collective phenomena QGS
(Quark-Gluon Soup)
- But we are left with the task of explaining how
the rich complexity that developed in the ensuing
13.7 billion years came about - Which is a much more complex task!
4Why do we need technology at the edge?
2 routes to new knowledge about the fundamental
structure of the matter
5CERNs particle accelerator chain
From LINAC to LHC
?
?
?
?
2004 The 20 member states
6The push for energy giant size
1500 tonnes of top quality SC cables
1800 Power Converter from 60 A to 24 kA
Bdip? 8.3 T Rdip ? 3 kmLdip ? 15 m ?
1232Ltunnel 27 km
15000 MJ of magnetic energy
1800 HTS Leads 11 kW_at_1.9 K
7The LHCFor 20 years it has been a viewgraph
8Rationale for superconductivity in accelerators
- Circular Accelerators
- Ebeam 0.3 B r GeV T m
- ? superconducting bending and focussing
magnets - high-energy hadron synchrotrons
- Linear Accelerators
- Ebeam E L MeV MV/m m
- ? superconducting acceleration cavities
- high-energy linacs
9Rationale for superconductivity in accelerators
- Electromagnets (d.c.)
- Resistive
- Power per unit length P/L rCu j B
- Total power (Joule) P rCu j B r rCu
j Ebeam - Superconducting
- Total power (refrigeration) P r
- for given operating
- temperature, power
- consumption is only
- cryogenic, i.e. independent
- of field
10Rationale for superconductivity in accelerators
- The LHC has a circumference of 26.7 km, out of
which some 20 km of main superconducting magnets
operating at 8.3 T. Cryogenics will consume about
40 MW electrical power from the grid. - If the LHC were not superconducting
- If it used resistive magnets operating at 1.8 T
(limited by iron saturation), the circumference
would have to be about 100 km, and the electrical
consumption 900 MW (a good-size nuclear
power plant), leading to prohibitive capital
and operation costs.
11Cost structure of the LHC
Total cost 4 G
Magnetcryogenics 66
12The power of the LHC beams
The LHC beam energy is 350 MJ. Already at
injection the beam can damage a magnet.
13LHC tunnel 2002
14LHC tunnel 2006
15Regular arc Magnets
1232 main dipoles 3700
multipole corrector magnets
16Regular arc Magnets
392 main quadrupoles 2500 corrector magnets
Installed dipole
SSS being transported
17Connection via service module and jumper
Supply and recovery of helium with 26 km long
cryogenic distribution line
Static bath of superfluid helium at 1.9 K in
cooling loops of 110 m length
18Insulation vacuum for the cryogenic distribution
line
19quench protection, power converters for orbit
correctors and instrumentation (beam, vacuum
cryogenics)
20Inner triplet magnets
21Accelerator magnets issues
- In iron dominated magnets the pole shape dictates
field quality - In superconducting magnets the conductor position
dictates the accuracy of the field. - Coils not self-supporting
- Beam will circulate 500 Millions times in the
LHC ! Field accuracy 10-100 ppm - Necessity to have all dipoles equal in length
within 100 ppm (1.5 mm over 15 m of the LHC
dipole length !)
22The hystorical outlook
23The LHC superconductor 7000 km of Cu/Nb-Ti cable
24Critical Current of LHC inner cable
25Magnetization for LHC NbTi
26Rutherford cables
Needs for 10-20 kA cable for protection Needs
very high packing factor 90 !! Needs a system
simple that keep strands
resistive contact Rc at cross-over point
The strand are fully transposed BUT field changes
over a period ! Ends problems Junctions BICC
? dB/dt
induced eddy currents in the loop I ? -dB/dt and
I ? 1/Rc
superconducting path in the strands
27Controlling the contact resistance
CERN has developed the controlled oxidation
method Value too low gives field errors Too
high may give instability
28Number of tests in 2003 at CERN
29QA Laboratory Equipment( 300 K tests)
30QA Laboratory Equipment
31Cable for main dipoles delivery
International permanent review
Multiple suppliers helped however only a few
were really critical for the project
32Coils - Winding
- Accurate positioning?for quench?for field
accuracy - Winding
- Curing at 185 C
- 3-D ends. Quasi-impregnation
33Critical ProcessWinding-Curing-Coil
30 µm
34Dipole cross section Collars
Collars and collaring are the main controllers of
the final coil shape Fine blanking of special
austenitic strips (?rlt1.005).
- Collaring press
- gt 20 MN/m
- 700 mm wide
- 450 mm high
- Beam accuracy 20 ?m
35QA Magnetic measurements
Introduced first to steer the FQ toward beam
dynamics targets.
Team of 3 physicists dedicated to on-time
analysis covering also holidays periods. 4 hours
to react yes-no-hold 2 days max time to
accept/reject hold-on cases
It has also helped to detect a number of
defects. It has also been used to detect subtle
electrical shorts
Courtesy of E. Todesco and C. Vollinger
36Steering the production through Field Quality
AT-MAS
To get it right we need model predicting position
and deformation at 10-20 micron
37Dipole cross section yoke shrinking cylinder
38Dipole -end partend plate
Also MQ
39Dipole-end part Bus Bars
Also MQ
40Dipole -end partShrinking cyilinder
Also MQ
41 Dipole -end partCu HXT
Also MQ
42 Dipole -end partCorrector Magnets
Also MQ
43Dipole -end partCold foot, Bellows and N-line
Also MQ
44(No Transcript)
45Snapshot at IndustrySuperconducting poles
46Snapshot at IndustryAperture assembly
47Snapshot at Industrycollaring process
48Snapshot at IndustryCold Mass
Laser Tracking measurements
49Snapshot at IndustryCold Mass waiting
50Snapshot at Industrydelivery to CERN
51Snapshot at Industry MAIN Quadrupole
526000 of various corrector magnets
MCDO Decapole Octupole Magnets
MCS Sextupole Magnets
MO Octupole Magnets
MQSXA Quadrupole Octupole Sextupole Magnets
53Superconducting LHC dipoles the long route
1988-98 short models and six prototypes for each
of the three generation design were built by
industry/CERN
1999 3x30 pre-series magnets were ordered from
three firms.
Three contracts for the fabrication of 1146 (30
spares) magnets have been signed March 2002
54Integrated supply chain management
CERN took care of most components
- Benefits
- Technical homogeneity
- Quality assurance
- Economy of scale
- Security of supply
- Balanced industrial return
- Risks drawbacks
- Responsibility interface
- Additional workload
- Liability for delays (just in time!)
- Transport, storage logistics we have moved
120,000 tonnes around Europe (5 TIRs a day for 5
years)
55LHC components industrial products
Industrial production
Art!
Courtesy of Ph. Lebrun
56Series production of LHC components
Courtesy of Ph. Lebrun
5790 main supply contracts worldwide
Europe in details
58Personnel training in Coil production
Courtesy of Jeumont, France
59LHC Learning Curvecollared coil production
Courtesy of Babcock Noell, Germany
60 Log Linear model learning
curves FIrm 3
Fit Crawford model
Collared Coil production 800-1000h Cold Mass
Production 500-1000h
Work done with P. Fessia
61Comparison with other industries
Industry r
Complex machine tools for new models 75-85
Repetitive electrical operations 75-85
Shipbuilding 80-85
LHC magnets 80-85
RHIC 85
Aerospace 85
Purchased Parts 85-88
Repetitive welding operations 90
Repetitive electronics manufacturing 90-95
Repetitive machining or punch-press operations 90-95
Raw materials 93-96
LHC Main Dipoles
62Learning phases? Different drivers?
63Number of toolingnumber of firm
- The number of tooling has been correct (the goal
was 3-4/week per company - Less companies (for example 2 instead of three)
would not have lessened significantly the unit
price and would have led to high risk
64Dipole performance today
65The great success the dipole delivery
review
66A milestone achieved
67Magnet work flow at CERNbefore beam test
Cold test
Arrival CM
Cryostat
Reception
Cryostat
Cold test
Reception
Arrival CM
Beam screen
Extra-work (30 dip)
Store 4 areas
Stripping Fiducials
Beam screen
Store 4 areas
Extra-work (30 dip)
Stripping Fiducials
Alignment
Tunnel
Store limited number
Connections
Connections
Alignment
Tunnel
Store limited number
1st dipole lowered 7th March 2005
Hardware Commissioning
Beam test
Beam test
Hardware Commissioning
December 2007
Start January 2007
68Experienced important setbacks QRL
- In Dec 2001 the contract to single supplier
awarded to Air Liquide - Installation started 21 July 2003
- Suspended after a few months conflicts between
AL and subcontractors - Re-started in spring 2004 and then suspended in
May for leaks, when a good part of 7-8 was
already installed. The delays were already
consistent - Endoscopic examination in June 2004 revealed
broken support. Work were stopped
69QRL
Sliding table
70Welding defects inspection of 18 service
modules in August 2004
Temper colour 100 Scaled surface 100
End crater pipe 100 Root concavity 30
Root porosity 50 Incomplete root penetration 45
Inspections of other elements at CERN and at
factory will continue
71QRL handled
- The company, near to drop the contract, has
re-committed itself to the project. - CERN took over the dismounting, repair and
re-installation for all 7-8 - CERN has supported follow in subcontracting and
coped (also financially) in a high rate plan to
catch up the delays. - AL and CERN has reviewed all design (tables,
weldings, alignment, bellow instability, etc.) - The first octant completed 1.8 months delay
- The last only less than 1 year in delay
- However the consequence are still there (for
example the ICs team was widely diverted to QRL
repair and then IC did not start with the wanted
quality. The rate requested for installation and
IC has been double than foreseen
72Another item that has caused delay DFB
Production in Protvino now finished
CERN re-insourced activty will end in May 2007
73Main work today IC
- Late start, accelerated rate
- 200 people in the tunnel
- 100 contractor
- 100 CERNassociated for managing, QA, repair,
in-sourcing of special WPs
74IC - 2
- New technologies introduced
- u.s. welding for single wire
- Induction heating for large cablegood but
problems in passing to high rate - Some IC of 7-8 must be reopened since we have
doubts - The 200 people in double shift are barely seen in
the 27 km tunnel!
75IC status
- Today we are working, as IC in 7 octants !!!
- In Sept 06 work done was 15. At half April the
work done has passed 2/3 !! - Arc (DS) finished by end of August 2007
- Long straight section by end of September 2007
(triplet last repair is not integrated)
76The plague electrical NC Q4R8 quench heater
circuit problem
- Abnormal quench heater voltage decay trace and
abnormal resistance of the heater after its first
powering in the tunnel - Quench heater is short to ground
- Magnet will be used as is thanks to the quench
heater redundancy - Quench heater power supply and interlocks had to
be modified
Arcing across open circuit
Good circuit traces
77A short to ground of the main dipole circuit at
warm It has been repaired with acrobatic
intervention
- A short to ground of the main dipole circuit
during the high voltage qualification
- The fault has been localized inside a dipole
magnet on the compensation loop of the dipole
bus-bar - Confirmation by endoscope after opening the
interconnection - The fault has been repaired in situ
Electrical Quality Assurance of LHC circuits, D.
Bozzini, V. Chareyre CERN/AT/MEL
78Interim short circuits to ground on the main
defocusing quadrupole circuit
- Detection of interim short circuits to ground on
the main defocusing quadrupole circuit during the
first phase of the cool-down (300 K to 80 K)
Electrical Quality Assurance of LHC circuits, D.
Bozzini, V. Chareyre CERN/AT/MEL
- The location of the fault was estimated with
accuracy of /- 10 meters - The short has disappeared after the restart of
the cool-down
79Q9R3 A short to ground of the quadrupole circuit
at warm
- Detection of the short circuits to ground on the
quadrupole circuit during the warm electrical
quality assurance tests - Drawings analysis has revealed that the busbars
in SSS 500 series are to long by 40mm - Corrective action (shortening of the busbar
extremities) is underway on the 32 SSS-500
80Resistance of the DFLS.5L8 circuit
- An abnormal resistance has been measured on one
corrector circuit of a stand-alone magnet that
supposed to be superconducting
Electrical Quality Assurance of LHC circuits, D.
Bozzini, V. Chareyre CERN/AT/MEL
- Powering of the circuit is blocked
- Further diagnostics and analysis is underway
81HV breakdown of 4 corrector circuits at cold
- 4 corrector circuits had a discharge during
voltage ramping-up to 600 V. The breakdowns have
been identified among the circuits.
Routing through the cold-mass. Tested at cold on
magnet test bench
Instrumentation Feed-through System tested at 3.1
kV
Standard connector for instrumentation tested at
1kV
Current leads tested at warm in gaseous He at 1kV
Electrical Quality Assurance of LHC circuits, D.
Bozzini, V. Chareyre CERN/AT/MEL
Lead to bus-bar connection
Suspected area
Bus-bar interconnection
Bus-bar interconnection
- Powering of the 4 concerned circuits is blocked
- Further diagnostics and analysis is underway
82HWC
- ElQA after pressure test
- Cool down
- ElQA at 80 K
- ElQA at 4.2 K on some circuit and DFB
- Full ElQA at 1.9 K
- Power test
- 4 Octants full
- 4 Octant reduced at 2 KA (11.85 kA nominal)
- Also HWC has been strongly compressed (long term
QRL effects). Strong participation of associated
Courtesy of Roberto Saban
83HWC - global view
84j
?
j
?
?
- All magnets isolated and active cooling in
thermal shields and cryogenic lines - ?Turbine filter exchange with thermal shield
cooling stop
85j Tuning of cold compressors turbines with
temporary stop of magnet cooling
k Stop of active cooling in weekend with only on
call activity limited to secure hardware l Stop
of magnet cooling for logic improvement in 1.8K
refrigeration unit m Random emergency stop in
cryogenic surface building with stop of sector 78
cooling ? Micro-electrical stop followed by
utility stops
861.9 K cooldown along the arc
87Follow-up of Helium inventory
Based on 29 l/m
88Major events towards 1.9K
- Utilities
- Power (400V failure on 30Mar07, with cascade
effect on cooling water 2h45, control networks
45, mobile vacuum pumping units. Situation
restored 10h45 without re-cooling the
cryo-magnets - Cryogenic part
- Progressive set-up of procedures to pump-down to
15mbar, while keeping DFBs with 4.5K conditions - Continued upgrades in instrumentation (Level
gauges, Heaters, ) but more efforts required to
improve reliability and availability - About to start testing the magnet temperature
control loops for 1.9K operation - 1.8K Refrigeration unit trip (frequency drive)
- Difficulties to restore 1.9 K conditions after a
stop - First tests of CRYO Start and CRYO
Maintain signals, to be continued as verbal
powering authorisation could have soon some
limits! - Clients
- DFBAs Cryo cmg underway with cooling of current
leads under tuning - ELQA_at_1.9K Underway, as in fact 3-4K is enough!
89Cool-down milestones
- OK for HW Cmg on DFBMs Stand-alone magnets
- Early troubleshooting on other DFBs continued
- Entire arc 7-8 below 2.0K on April 5th, 2007
- Difficulties to restore 1.9 K conditions after a
stop - Efforts to be continued (instrumentation, logic,
procedures) while preparing for powering
90CPP - cryogenic actions
- Current activities
- 1.8 K units pumping down to 15 mbar
- Supercritical helium distribution temperature
- Heat loads and inventory assessment
- Valve sizing assessment
- CRYO_START and CRYO_MAINTAIN validation
- DFBMA commissioning
- Level sensors in-situ calibration
- Temperatures mapping on chimneys with different
control parameters - Added heating power on chimneys to avoid
condensation - Analysis of phase separator and stand-alone level
gauges measurements - Analysis of possible line Y interruption around
Q9
911.8 K units pumping down to 15 mbar - actions on
the 1.9 K
- Cold compressors trip during pump down
- Cause
- Cells phase separator with liquid helium
perturbing the pumping flow during lambda
transition (up to 60 pumping flow increase) - Consequences
- Delays in the recovery of nominal conditions
after stops - Action plan
- Reduce the quantity of liquid in the return
module phase separators - Interlock to close the supply valve after a trip
(debugging) - Avoid connecting cold compressors at saturation
pressure (eq. saturation temperature) higher than
the colder cold mass temperature
92The real problem Supercritical helium
distribution temperatures
We should be working here
and were there!
Courtesy of PhL
93Supercritical helium distribution temperatures
- Supercritical helium temperature above
specification - Cause
- EX-LEP refrigerator subcooler not adapted for LHC
operation as seen last autumn - Consequences
- Difficulties for some transients adaptation
- DFBs feed limitations
- Action plan
- Preparation for new cryoplant connection
(process logic and interlocks) - Sector 7-8 fed by new cryoplant
- Consolidation of ex-LEP subcooler being
investigated for implementation before autumn
2007 after sector 7-8 warm-up
94Heat loads preliminary assessment on sub-sectors
95Valve sizing assessment JT valve
- Valve opening
- Measured around 1 3
- Expected 13 23
- Possible causes
- Lower stand-by heat loads gt valve outside design
range - Valve sizing incorrect gt manufacturer error
- -gt request modification of flow plug
- Consequences
- Difficult control of magnets temperature
- Increased risk of oscillations and perturbations
on the whole system - Action plan
- Test bench measurements on a spare valve
- Discussion with valve manufacturer
- Assessment of heat loads to confirm flow rate
- Valve characterization with additional electrical
heating in the cold masses - Valve consolidation (plug exchange) for sector
4-5 before cooldown - Valve consolidation (plug exchange) for sector
7-8 after warm-up
96DFBMA commissioning
- Level sensors readings and condensation on top
flanges - Different level sensors show different readings
- 120 A assemblies cannot be cool down to nominal
operating conditions (20 K) due to condensation
(below 5 deg C) - Minimum temperature achieved 26 and 36 K (4 and 8
leads assemblies) - Possible causes
- Supercritical helium temperature
- Position of level sensor near helium feed
- Convection paths in the chimneys
- Pt sensor precision at 20 K
- Warmer helium intake from adjacent chimneys for
larger cooling flows - Consequences
- Delay in nominal cryogenic conditions for magnet
powering - Action plan
- Improve supercritical helium distribution
temperature - Calibrate the level sensors against geometrical
changes in the vessel / leads - Increase level to observe possible cooling flow
reduction - Verification of Pt sensors calibration curve
- Additional heating on chimneys top flange
97Where we are today
- Struggling for cryogenics
- The program of HWC is delayed of about 1 month
- At least two octants to warm up (for many
reasons) - But closing the rings will be done on schedule
(triplet impact to be assessed) - Have we lost the engineering run in 2007 ?
- We dont care we keep this goal.