1. Thickness of the Kamaboko Tunnel Shield Wall under Different Assumptions and the Impact on Operations
Ewan Paterson
LCWS15 Meeting NOV 3, 2015
2/24/2019
LCWS 2015/jmp

2. THE “GREAT” WALL
2/24/2019
LCWS 2015/jmp

3. History of Shielding
RDR Assumed site independent twin tunnel design, one beam tunnel and one service tunnel which can be accessed by people (Rad workers) at any time.
SB 2009 Study single tunnel designs and operational “Availability” of accelerator systems in different scenarios.
Began developing RF systems for a Single Tunnel Linac.
DRFS with many 800kw klystrons in a variety of configurations all inside the tunnel.
KCS Klystron Cluster with 20 x 10 MW tubes combined in surface buildings, delivered to and distributed in the tunnel through over-moded wave guide.
By 2011 both systems were accepted as having acceptable availability but site dependent choice and more R&D still required.
2011 Kamaboko tunnel with shield wall proposed for selected Japanese deep site.
Development of KCS an DRFS stops and we are back to two tunnel equivalent in preparation for the TDR.
2013 In the TDR the choice of RF distribution system is site dependent. KCS on the surface for shallow tunnel and some form of DRFS in a deep tunnel.
Japan proposes a real (deep) site and we go with a single Kamaboko tunnel with a wall! In effect a two tunnel solution.
2/24/2019
LCWS 2015/jmp

4. Single Tunnel Today
Today we have only one site and it is deep underground. In the past the only option considered for a single tunnel was for the linac RF systems, to have the klystrons, Modulators, etc in the tunnel with all beams off for any access for maintenance, or systems commissioning or development.
In the past the Availsim assumptions had 40 khrs for the MTBF of the “new’ 10MW tubes and 50 khrs for the modulators, this meant that you needed a >10% energy overhead (installed spare cryomodules and RF stations) to maintain the availability goal.
This needs to be updated based on experience. The tubes were designed (by three manufacturers) to have lifetimes of > 100 Khrs. They are multibeam tubes with low cathode loading and low voltage and they should have lifetimes comparable to the 800 kw tubes in the DRFS system design. So far, experience indicates that this is a valid assumption and could be used in a single tunnel with only a few % energy overhead as is required in the twin tunnel models for maintenance and availability. This is also true for modern solid state and modular klystron modulators.
DO WE NEED A WALL AT ALL? YES, BUT COULD BE THINNER?
2/24/2019
LCWS 2015/jmp

5. Downtime versus Energy Overhead from TDR
2/24/2019
LCWS 2015/jmp

6. Review the Shield Wall Today
The baseline design is 3.5 meter but this is very conservative and we need to review the assumptions.18 MW being lost is unrealistic but in some future scenario perhaps one might have to consider rare accidental beam losses of megawatts some place in the machine.
The wall could be thinner but probably not much less than 3.5 m (my best guess) as long as there is the potential for high power beams.
One could have a variable thickness wall tailored to local conditions but this would be very restrictive in the future as it would be very difficult to change for as yet undefined upgrades.
Let us re-open the question of Availability in WHAT IS NEW?
1) ONE SITE AND ONE RF SYSTEM DESIGN WITH 10MW TUBES.
2) MODULATORS ARE NOW SOLID STATE WITH BUILT IN REDUNDANCY
3) 6 MORE YEARS OF GOOD EXPERIENCE WITH THE 10MW KLYSTRONS.
4) CONSIDERATION OF AN ALTERNATE SCENARIO ALLOWING ACCESS TO
EQUIPMENT IN SERVICE TUNNEL WITH EQUIPMENT ON (dark current)
BUT WITH NO PRIMARY BEAM ALLOWED IN THE 22 km’s OF LINAC.
THE 5 km CENTRAL REGION IS VERY DIFFERENT AND WILL BE
CONSIDERED SEPARATELY.
2/24/2019
LCWS 2015/jmp

7. Thin Wall Design Requirements
In the given site, now the only site under study, there are kilometers between exits and personnel safety depends on having more frequent ( few hundred meters) escape routes. This means we need some firewall between two parts of the tunnel, with periodic access between them. This separator wall has to be fireproof and reasonably air tight for smoke or oxygen deficiency hazards.
A 30 cm thick concrete wall would satisfy this problem but there are more!
Shielding against long term radiation damage to electronic components in normal operation. Could be done locally or with the wall, if thick enough to satisfy the following.
Personnel shielding from dark currents while RF processing , and testing or maintenance with all systems ON but NO primary beam.
The on-going studies of dark current behavior are important and probably determine the minimum wall thickness at (I hope) ≈ one meter!
IN THIS SCENARIO, DURING A SCHEDULED MAINTENANCE DAY
( or an unscheduled access) ALL BEAM IS OFF BUT ALL POWER
SUPPLIES AND RF ARE ON AND ACCESSABLE BEHIND THE WALL.
ONE CAN MONITOR OR LOCALLY TEST AND MAINTAIN MOST OF THE
HARDWARE BUT WITHOUT BEAM INFORMATION.
THIS SCENARIO IS WORTH CONSIDERATION
2/24/2019
LCWS 2015/jmp

8. Impact of No Beam During Access to the Service Tunnel
In the linacs the Impact not as bad as one initially thinks. PS’s and RF stay on and warm. This will shorten recovery times and reduce failures on cycling power and temperature compared to a single tunnel case.
In the linac to change klystrons or work on equipment on every two week maintenance days, the beams have to be safely turned off upstream of work location. This means no beams in the linac during most maintenance days even with two tunnels.
2 Klystrons, on average, every 2 weeks will require both replacement and checkout time. This assumes ≥ 100 khrs MBTF for klystrons and with redundancy ≥ 300 khrs, for the modulators.
A thin wall will have great impact in central region if one changes from twin tunnel, as in the TDR, to Kamaboko tunnel with a thin wall. No beams would be allowed in the e- injector or the alternate e+ source and therefore into the DR’s. This could have a significant impact during commissioning!
Perhaps there can be a different solution for these regions where the beam energy is less than or equal to 5 GeV. This is for a later discussion?
2/24/2019
LCWS 2015/jmp

9. REGION KLYSTRONS POWER SUPPLIES FOR I&C ELECTRONCS
APPROXIMATE NUMBER OF ELECTRICAL SYSTEMS IN DIFFERENT REGIONS THAT MAY REQUIRE FREQUENT MAINTENANCE
REGION KLYSTRONS POWER SUPPLIES FOR I&C ELECTRONCS
MAGNETS, MODULATORS, CORRECTORS
E- Source
E+ Source without alternate e+ source
RTML
Linacs
BDS
TOTAL
2/24/2019
LCWS 2015/jmp

10. Operations Model and Updated MTBF’s
There is a scheduled 24 hour maintenance period every 2 weeks and 2 one month shutdowns per year. These were the assumptions in the RDR and TDR discussions of availability and presumably detector push-pull would be during some of these.
The 10 MW multi-beam klystrons MTBF is 10*5 hrs. This was 4x10*4 in Availsym studies.
The klystron modulator ( now solid state and modular) will have MTBF comparable to state of the art modern powers supplies, 3 to 10 x10*5 hours. 5X10*4 assumed in TDR
The MTBF’s for ≈ 50 items used in the final report of the Availability Task Force range from 3 to 10 X 10*5. (see SB 2009 report) Let us use an average of 5 x 10*5 for everything other than the klystrons and look at a 2 week period.
2/24/2019
LCWS 2015/jmp

11. REGION KLYSTRONS POWER SUPPLIES FOR I&C ELECTRONCS
NUMBER OF ELECTRICAL SYSTEMS THAT WILL REQUIRE SCHEDULED MAINTENANCE EVERY 2 WEEKS OF ROUTINE OPERATION WITH THESE UPDATED MTBF”s
REGION KLYSTRONS POWER SUPPLIES FOR I&C ELECTRONCS
MAGNETS, MODULATORS, CORRECTORS
E- Source
E+ Source not including alternate source!
RTML
Linacs
BDS
TOTAL
Average failures in RTML’s + LINAC’s every
2 weeks or 0.3%
95% probable or .6%
99% probable or 1.0% or 0.2% or 0.2%
2/24/2019
LCWS 2015/jmp

12. Example of Availability from TDR
2/24/2019
LCWS 2015/jmp

13. OPERATION BETWEEN SCHEDULED 2 WEEK MAINTENANCE PERIODS
The impact of these failures is very different in the linacs including the RTML 5 GeV transport lines compared to those in the injectors, the RTML compressors and the BDS’s.
The failure of these small number of RF stations, quads, correctors, BPM’s in the linac’s can be adsorbed in a small additional energy overhead (≈1%) and or retuning of the optics and steering to maintain constant conditions at the end of the linac. (Note that one RF station makes ≤ 1 GeV which at 500 GeV, equals 0.2% in energy or is equal to the ∆p/p of the beam.) This retuning would be necessary, even in twin tunnel case, to maintain smooth operation.
In the injectors, sources, compressors and BDS, one does not have this flexibility and corrections have to be more locally with installed spares for redundancy. See examples from the TDR on the next slide.
2/24/2019
LCWS 2015/jmp

14. CENTRAL REGION REDUNDANCY AS LISTED IN THE TDR
E- SOURCE TO E- DR
2 Guns and 2 Laser Systems Installed
Buncher:- has a spare klystron and accelerator structure Installed
5 GeV Booster :- It has 21 cryomodules + 3 spare, plus a spare klystron Installed
Energy Compressor :- It has 2 x 5 Mw klystrons Installed
E+ SOURCE TO E+ DR
Spare Undulator Sections but not Installed? Needs review.
Spare Target, Flux Concentrator and Capture sections. Needs review.
Spare Klystrons for warm RF, Booster Linac and Energy Compressor Could be
installed as in E- Injector along with spare cryomodules.
Timing Chicane Needs design and entry in CFS layout and lattice.
ALL OF THE ABOVE NEED TO BE, UPDATED, AGREED UPON TO BE CONSISTENT WITH CFS LAYOUTS AND LATTICE, AND ARE INDEPENDENT OF THE DISCUSSION OF A THINNER SHIELD WALL OR TUNNEL CROSS-SECTION .
AUXILLIARY OR ALTERNATIVE E+ SOURCE DESIGNS WILL HAVE A LARGE LOCAL (over 1 or 2 km) IMPACT ON THE LATTER.
2/24/2019
LCWS 2015/jmp

15. UNSCHEDULED MAINTENANCE IF REQUIRED
Most of the non electrical (cryo, water pumps etc) equipment that may require frequent maintenance are accessible during beam operation and do not impact the shield wall design. They do contribute to the projected 15% downtime as shown in slide 12.
If there is a problem with lack of energy overhead, in retuning the linac or a fluctuation in numbers, then a single klystron can be replaced in one shift or less by 2 teams of 2 people. One team handling the km’s of transport both ways while the second team handles disconnect/connect and testing. With the power on/beam off scenario, the operation should be complete within this single shift.
Some other component maintenance can be also added in this shift but is not considered necessary in the operating model of one day, 3 shifts, of planned and scheduled maintenance every two weeks.
2/24/2019
LCWS 2015/jmp

16. Conclusions and Decisions?
DARK CURRENT AND PERSONELL SAFETY?
How thin can the wall be assuming we adopt beam OFF electrical systems
ON as a scenario allowing access? Are there other issues?
THICK ENOUGH TO PROTECT COMMERCIAL ELECTRONICS IN THE SERVICE TUNNEL?
Can we do without local shielding as is adopted in X-FEL to prevent long
term radiation damage to electronic components in the service tunnel.
CHOICE OF LINAC TUNNEL CROSS-SECTION?
From the presentations today can we select the best shield wall and tunnel cross-section for the 22 km that contain the RTML and LINAC’s and prepare a corresponding change request ?
RELATED FUTURE STUDIES
Study the 5 km central region containing the E+ source, E- source, DR, and BDS for updated
technical design, availability and optimum tunnel designs and cross-sections. These could be
different from the main linac along with beam operation and access scenarios.
2/24/2019
LCWS 2015/jmp